Datasets:

izumi-lab

/

pile-modified

like 1

It often is desired to feed a load (such as material to be destroyed, food to be shredded, etc.) towards a processing mechanism (such as a shredding mechanism, cutting mechanism, etc.). Conventionally, feeding mechanisms have been established for continuously moving materials within a certain range of dimensions towards the processing mechanism. For example, a load within a first range of thickness could be handled continuously, but the feeding mechanism would need to be stopped and manually adjusted before trying to feed a load of a second range of thickness. It has long been commonly appreciated that care is usually required in feeding an input load into most machines as to proper orientation of pieces of the load, uniform quality of pieces of the load, etc. This has long been applicable for machines involving cutting and especially the specific case of paper shredders. It conventionally had been recognized that the operative cutting mechanisms in paper shredders were designed to accommodate a particular thickness of paper and that inputting too thick a stack of pages, for example, could damage or at least “stall” or jam the shredder. For accomplishing high-security destruction, in the past, each kind of material needing to be destroyed had a particular destruction mechanism designed to destroy the material based on its dimensions, kind of material, etc. Other than the present inventor's recent work being brought to the market, there has not yet been a destruction mechanism that would destroy paper to high-security small pieces as well as also destroy non-paper materials such as a polyester-type material (such as key tape), a thick material (such as a book), etc. Rather, conventionally no more was expected of a paper shredder than that it shred paper. Merely meeting the recent security requirements demanding yet smaller-sized residue has occupied the shredder industry, as most of the shredder industry seemingly has been unable to design products to satisfy the new high-security shredder requirements. Only a few companies have actually managed to do so with actual viable products in the marketplace. In the case of an expensive category of machine called a disintegrator, paper shredding and different types of to-be-destroyed material ultimately may be accommodated. However, within the disintegrator machine the different materials may travel non-identical feed paths. See US 2003/0201353 A1 by Lefrancois et al., titled “Dual-path office product disintegrator” published Oct. 30, 2003. Different input ports are provided for different types of input materials. Disintegrators are heavy, large-dimensioned (non-portable) machinery. Several commercially-available disintegrators actually are two machines combined, in order to perform the destruction function. Typically, a conventional shredder is typically atop a conventional disintegrator. It pre-shreds paper (and some other materials), and then feeds the-pre-shredded material to the disintegrator. This is usually necessary to obtain adequate throughput rates. In high-security destruction, before the present inventor's own work it had not been possible to destroy different to-be-destroyed materials in a single shredder or a single disintegrator. Thus, the question of feeding a significantly non-uniform load of to-be-destroyed material had not been encountered in the area of high-security destruction. Examples of conventional feeding mechanisms are mentioned. U.S. Pat. No. 3,958,737 issued May 25, 1976 to Scott (Precision Sales Corp.) for “Adjustable Feed Mechanism.” U.S. Pat. No. 5,622,330 issued Apr. 22, 1997 to Sharp et al. (ASC Machine Tools, Inc.) for “Self-adjusting Feed Stock Accumulator System.” U.S. Pat. No. 5,348,282 issued Sep. 20, 1994 to Choi et al. (Xerox Corp.) for “Self Adjusting Feed Roll.” U.S. Pat. No. 4,621,798 issued Nov. 11, 1986 to Akers (Bell & Howell Co.) for “Envelope Feeding Mechanism for Mail Sorting Machines.”

Q: Displaying weekday on Active Admin column How it's possible to display the weekday next to a date ? column :start_date #displaying for example 29 Jan. 2016 How to echo it as Friday, 29 Jan. 2016? A: You can use strftime method. column "Start Date:" do |post| post.start_date.strftime("%A, %d %b. %Y") end

Over the last six years, women's rugby has been growing in Nova Scotia and gaining recognition as a high-performance sport. "The big success right now for our Keltics program is now all the universities are offering scholarships because the level of play within Nova Scotia has improved so much," said Jack Hanratty, provincial coach for Nova Scotia Keltics. Hanratty said the growth can be shown by results, membership or through experience playing the sport. The Keltics program focuses on all three. Changing attitudes "The one thing we've noticed is that you used to play rugby because somebody you knew, or somebody in your family, had a history in rugby," said Hanratty. "And now people are seeing the game, because now it's an Olympic sport, as … a real option of being a student-athlete." This summer, the Keltics senior women's team won the silver medal at the Canadian championship tournament, finishing ahead of B.C. and Ontario. "We are continuing to improve and it's showing when we do go to compete," said Sarah Boudreau, captain of the Acadia Axewomen and a member of the Keltics program. Competing for national honours Boudreau and the Axewomen, the host team, will be competing for a medal at the national university championship event starting Thursday in Wolfville. Eight teams representing all of the U Sports conferences, including Atlantic University Sport champion St. Francis Xavier, will compete for the national banner. "This is the first year that Acadia is hosting and it's kind of cool to be able to compete for that on your own home turf," said Boudreau. "I'm feeling very confident." There will be approximately 200 athletes attending, with 25 of them having played with the Keltics. Calgary, Laval, Queen's, Guelph, Victoria and Ottawa are the other teams competing. There are four quarter-final matches on Thursday with the final on Sunday afternoon.

Musings, Inspirations & All Things Marang Tuesday, November 15, 2011 Ok, so by now I'm sure you've heard about all the changes Etsy has made on the whole changing your shop name. I had given an account on how I went about changing my shop name, here, before these newest updates were made. And since I see a lot of people still refering to that post, I thought I'd give a few links from Etsy's Storque on their updates. This should make the whole process SO much easier!! Monday, August 8, 2011 I know I've been absent and quiet for a while. I have a good excuse, I promise! I moved across the country from Santa Fe, NM to Columbus, OH. It took me longer to get everything situated in Santa Fe and then getting settled and looking for a job has been taken longer than I'd like also. But, I have been having way too much fun discovering this new city! New people, new experiences, new places to find and new things to do. I've even gotten to cross off a few things from my 100 List!! 9. Learn to water ski. I tried water skiing this past weekend and although I didn't quite get the hang of it yet, I'm very close and very sore! It even hurts to type! 78. Go to a drive-in theater. This past weekend (after the water skiing attempt!) I was treated to a relaxing evening at the drive-in movie theater. :) I've gotten a new-to-me car also. My uncle was kind enough to pass down my Grandmother's car after she passed away. The Silvery Minnow, as she is now named, is an '89 Mercury Grand Marquis. I drive a boat!! I am trying to get situated with a job and back on a 'normal' schedule of life, creating and blogging. Please stay in touch and come back to see what's new!! Tuesday, June 14, 2011 I may have skipped a few treasuries, so I apologize if your's doesn't get posted. But, here are some more treasuries that feature my items! Thank you to urbancreatures and RenewableCrafts for these great treasuries!! Wednesday, June 8, 2011 Well the promotional creations I ended up making for the Trashion Team swap are seed packets. I had small envelopes that I had already made or had ready to go in my stash (of way too many things!). From last years vegetable garden, I had a huge jar of marigold seeds waiting to be used. I added a note with a coupon code and my business cards and tied them up! A bunch of flowers just for you!!

forceinline overload Bool() : Bool = false; [I when Integer?(I)] forceinline overload I(a:Bool) : I = if (a) I(1) else I(0); [I when Integer?(I)] forceinline overload Bool(a:I) : Bool = (a != I(0)); forceinline overload equals?(a:Bool, b:Bool) : Bool = if (a) b else not b; forceinline overload lesser?(a:Bool, b:Bool) : Bool = ((not a) and b);

Q: R populate column based on several criteria I have a dataset in which I want to create a new column that takes existing values from another column when available, when no values are available, it should instead take values one year in the past. Example, I have the following data: Date Market Campaign Impressions 1/1/2017 SE Home 150 1/2/2017 SE Home 100 1/1/2017 GB Home 50 1/2/2017 GB Home 55 1/1/2018 SE Home NA 1/2/2018 SE Home NA I would like to write something that will provide me with: Date Market Campaign Impressions Future_impressions 1/1/2017 SE Home 150 150 1/2/2017 SE Home 100 100 1/1/2017 GB Home 50 50 1/2/2017 GB Home 55 55 1/1/2018 SE Home NA 150 1/2/2018 SE Home NA 100 I have managed to populate Future_impressions with the Impressions for historic months using: df$Future_impressions[is.na(df$Impressions)] <- NA But I have no idea how get the future months populated. I basically want to write something that says: - IF Impressions=NA - THEN MATCH (Date minus 12 months) AND MATCH Market and Campaign and retrieve Impressions I am a very new R user but I've searched around extensively, so I hope I haven't asked a redundant question! Thank you all so much in advance! A: Here is an option using apply df$Date <- as.Date(df$Date,'%m/%d/%Y') get.last.year <- function(dt, mk, cp){ ly <- as.POSIXlt(dt) ly$year <- ly$year - 1 ly <- as.Date(ly) x <- df[df$Date == ly & df$Market == mk & df$Campaign == cp, "Impressions"] return(x) } df$Future.impressions <- apply(df, 1, function(x) ifelse(!is.na(x[["Impressions"]]), x[["Impressions"]], get.last.year(x[["Date"]], x[["Market"]], x[["Campaign"]])))

Q: is there a significant speed difference between attributing a boolean to a variable and a string to variable in python? I am currently developing a python project where I am concerned with performance because my CPU is always using like 90-98% of its computing capacity. So I was thinking about what could I change in my code to make it faster, and noticed that I have a string variable which always receives one of two values: state = "ok" state = "notOk" Since it only has 2 values, I tought about changing it to a boolean like: isStateOk = True isStateOk = False Does it make any sense to do that? Is there a big difference, or any difference at all, in the speed of attributing a string to a variable and attributing a boolean to a variable? I should also mention that I am using this variable in like 30 if comparisons in my code, so maybe the speed of comparison would be a benefit? if (state == "ok) # Original comparison if (isStateOk) # New comparison A: That's not going to fix the program using 90-98% CPU, but technically yes, using a Boolean is better. You can also use is instead of ==: isStateOk = True if isStateOk is True: # Do stuff Edit: Nevermind, in https://hg.python.org/cpython/rev/01a7e66525c2/ they already made == True get converted by the Python interpreter to is True under the hood, so there is no performance difference to writing it either way. While it is all around a good idea to use Booleans here since the purpose of Booleans is to represent ok/not ok state, it's not going to give any type of noticeable performance improvement.

#!/usr/bin/env python3 ''' s3qlstat - this file is part of S3QL (http://s3ql.googlecode.com) Copyright © 2008 Nikolaus Rath <[email protected]> This program can be distributed under the terms of the GNU GPLv3. ''' import sys import os.path # We are running from the S3QL source directory, make sure # that we use modules from this directory basedir = os.path.abspath(os.path.join(os.path.dirname(sys.argv[0]), '..')) if (os.path.exists(os.path.join(basedir, 'setup.py')) and os.path.exists(os.path.join(basedir, 'src', 's3ql', '__init__.py'))): sys.path = [os.path.join(basedir, 'src')] + sys.path import s3ql.statfs s3ql.statfs.main(sys.argv[1:])

Tuesday, August 2, 2011 Took me about 4 evenings work to model and paint this guy, but here he is. I'm pretty pleased with how he came out. I got in a game with him the other day vs this a**hole. Merv brought a Zoraida/Collodi crew and I learned the meaning of frustration dealing with the latter. He is such a bloody pain to put down. Anyway, to cut a long story short, I lost, badly, but by no fault of the rider. He pretty much held up the entire enemy crew for 2 turns - on his own (while Seamus missed stuff with his flintlock, the gunslinger trudged through a forest and the Belles just flopped around the place being Obey-ed). On an unrelated note, some of you may recognise the fat and thin Daemon princes in the background in these pics. Saturday, July 23, 2011 Anyway, this line is tapped, so I'll be brief (stupidest line from the Matrix, if the line is bloody tapped it means someone is listening - don't arrange to meet under the bridge dumbass). I pre-ordered some shiny new Malifaux stuff last month, and it arrived bright and early (the girlfriend answered the door for it, my precious sleep was not disturbed) on Thursday. Here's what I got: I think I'll put this chap up as a prize in the Tournament I'm planning next month. Mr. Dead Rider is currently undercoated black and sitting on my painting station. Snow Storm is pretty. Very much living up to the hype, Storm's body is cast in resin with metal horns. I checked for air bubbles and only found a few superficial ones on the soles of his feet and the flat areas where his arms connect. Citadel Fineass could learn something from this. Snow is entirely metal, only her hands need to be attached, they'll need pinning and possibly a small drill bit to do the job. I reckon a paperclip is the right size. Dead Rider is entirely metal. There was very little flash, and besides the scythe arm being a bit tricky he was pretty easy to pin, about an hours work altogether (while watching an episode of Dexter). After pinning I used a tiny bit of green stuff to blend in the joins. The rocky part under the steeds right hoof is actually part of the model, so I built up a rocky outcrop out of thin cork tile. Monday, February 7, 2011 I have to do a post on this mini. I'm a fan of simple well executed colour schemes on models, but this one really takes the cake. And I thought I did well painting some Onryu with only 5 different colours... In other news GW is still promoting their flying toaster. One of the lads from out LGS has one painted up. It's actually very nice in the flesh. Tuesday, February 1, 2011 Ok Forgeworld, you did bad in the past but after this I'll forgive you. Now that's the quality we expect from the price tag. Wyrd posted some pics of their February releases today and once again they do not disappoint. Check out the paint job on the Lost Love! I want to do my spirits like that now. As regards sculpt so far Nix seems to be a crowd pleaser (and I must agree) though I think the rats and Stolen ooze character. Special mention on the release of the Malifaux rules manual. Now I can't comment on the content besides the fact that it claims to contain up to date errata'd rules for the game, so I guess that's what it'll contain, but I can say the cover looks nice. Which is the main thing. Tuesday, January 11, 2011 Seriously? What, you're serious Games Workshop? Tyberos The Red Wake is the best you lads can think up? Look at his bloody hands. Damn that's ugly. I can just imagine the board meeting; "Guys, guys!""What?""Y'know what'd be cooler than a power fist, lightning claws and chainswords...? All of them! At once! Combined!""... Incredible. We'll do it! It totally won't look stupid and our unwaveringly loyal customers will want to spend a mint on it for sure! This is totally the best idea we've had all week since the revival of the S-shaped monster's no-one liked. Quick - think up some bullshit reason for it.""Uh... 'ancient relic-gauntlets'...?""... Wow! Someone's getting a promotion!"

Introduction {#s1} ============ Dendritic cells (DCs) orchestrate immune responses, linking innate to adaptive immunity. DCs, which are important professional antigen-presenting cells (APCs) depending on their maturation state, take up a broad range of antigens and present them to T cells. Several endogenous and exogenous stimuli, such as Toll-like receptor (TLR) ligands and inflammatory cytokines, phenotypically and functionally activate DCs. Upon activation, DCs display phenotypic changes, including upregulation of the expression of MHC II and costimulatory molecules, and changes in expression patterns of chemotactic and homing receptors. Activated DCs also exhibit functional changes, including the downregulation of antigen uptake and secretion of chemokines and cytokines ([@B1]--[@B6]). DCs are classified as classical DCs (cDCs), which are further subdivided into classical type 1 DCs (cDC1s) and classical type 2 DCs (cDC2s), plasmacytoid DCs (pDCs), and monocyte-derived DCs (MoDCs) ([@B7], [@B8]). Inflammatory MoDCs are differentiated from murine bone marrow (BM) cells by treatment with granulocyte macrophage-colony stimulating factor (GM-CSF) *in vitro* ([@B9]). In contrast to the treatment with GM-CSF, BM cells treated with Flt3 ligand differentiate into cDC1s, cDC2s, and pDCs ([@B10]). GM-CSF-induced BM-derived DCs (BMDCs) are similar to *in vivo* inflammatory and TNF/iNOS-producing DCs rather than DCs in the steady state ([@B4], [@B11]). GM-CSF promotes the maturation and activation of monocytes, macrophages, DCs, and granulocytes during inflammation ([@B4], [@B12]--[@B14]). GM-CSF also acts on the myeloid cell-dependent Th17 inflammatory response, which is mediated by TNF, IL-6, IL-23, and IL1β that are produced by macrophages and MoDCs ([@B15]--[@B17]). Besides the important roles of GM-CSF *in vivo*, it was originally identified by its ability to induce inflammatory DCs from mouse BM precursor cells *in vitro* ([@B9], [@B16]). The binding of GM-CSF to its receptor initiates the activation of downstream pathways, such as MAPK, PI3K/Akt, NF-κB, and STAT5. The PI3K/Akt signaling pathway regulates the GM-CSF-induced proliferation, survival, and development of DCs ([@B17], [@B18]). Mst1, mammalian STE20-like kinase 1, is a multifunctional serine/threonine kinase highly expressed in immune organs, such as the thymus, spleen, and lymph nodes ([@B19], [@B20]). Mst1 plays important roles in cell proliferation, differentiation, apoptosis, and organ size regulation ([@B21]--[@B24]) as well as in the regulation of survival, proliferation, trafficking, and function of T cells, a type of lymphocytes in adaptive immunity ([@B19], [@B20], [@B25]--[@B30]). Furthermore, recent studies revealed that Mst1-deficient DCs promote the overproduction of IL-6, which induces Th17 differentiation in DC-specific (CD11c-Cre) conditional Mst1-knockout (KO) mice ([@B31]) and Mst1 signaling contributes to CD8^+^ DC function in mediating CD8^+^ T cell priming through the regulation of mitochondrial activity and IL-12 signaling ([@B32]). However, the roles of Mst1 in the activation and maturation of MoDCs are still poorly understood. In this study, we aimed to elucidate the role of Mst1 in GM-CSF-induced BMDCs, which mimic *in vivo* inflammatory MoDCs more closely. We found that *Mst1*^−/−^ BMDCs displayed higher expression levels of cell surface CD40, B7, and MHC II molecules as well as increased production of inflammatory cytokines than the *Mst1*^+/+^ BMDCs. Mst1 deficiency also increased the surface expression level of B7.2 in migratory monocyte-derived cells (MCs) *in vivo*. In addition, hyperactivated *Mst1*^−/−^ BMDCs stimulated *in vitro* T cell proliferation and activation to a greater extent than *Mst1*^+/+^ BMDCs. We also found that *Mst1*^−/−^ BMDCs enhanced GM-CSF-induced Akt/c-myc pathway. Inhibition of the Akt/c-myc pathway reversed the hyperactivation of BMDCs induced by the loss of Mst1. The findings of the present study establish that Mst1 is implicated in the regulation of GM-CSF-mediated Akt/c-myc pathway in DCs, suggesting that Mst1 is one of the endogenous factors that determine the activation status of GM-CSF-stimulated inflammatory DCs. To our knowledge, this is the first observation of Akt1/c-myc-mediated functions of Mst1 in the immune system. Materials and Methods {#s2} ===================== Mice ---- *Mst1*^−/−^ mice on the C57BL/6 background were kindly provided by Dr. Dae-Sik Lim (Korea Advanced Institute of Science and Technology, Daejeon, Korea) ([@B33]). Mice used in the experimental protocols were backcrossed more than twelve generations to C57BL/6 mice. *Mst1*^−/−^ and littermate control were maintained in a specific pathogen-free animal facility at Korea University, Seoul, Korea. These mice experiments were performed according to the guidelines of Korea University Institutional Animal Care and Use Committee (KUIACUC-2017-109 and 2019-0013). Chemicals and Reagents ---------------------- Murine recombinant GM-CSF was purchased from ProSpec. CpG DNA (ODN 1826) was purchased from Invivogen. The following chemicals were used: Akt1 inhibitor MK-2206 (AdooQ Bioscience), myc inhibitors 10074-G5 (Cayman) and (+)-JQ1 (Sigma-Aldrich). Antibodies against the following proteins were used in western blot analyses: Mst1 (Cell Signaling, no. 3682), Mst2 (Cell Signaling, no. 3952), phospho-Akt (Cell Signaling, no. 4060), phospho-YAP (Cell Signaling, no. 13008), YAP (Cell Signaling, no. 14074), c-myc (Cell Signaling, no. 13987), Akt (Santa Cruz, sc-1618), GAPDH (Santa Cruz, sc-32233), Lamin B (Santa Cruz, sc-6217), and β-actin (Santa Cruz, sc-47778). Cell Numbers and Phenotyping of Primary Cells in Mouse Lymphoid Organs ---------------------------------------------------------------------- Sex- and age-matched, 7- to 8-weeks old mice (female) were used in this study. Primary cells of BM, spleen, and mesenteric lymph nodes (MLN) in *Mst1*^+/+^ and *Mst1*^−/−^ littermates were used. BM cells were collected from femurs and tibias of mice. All organs were homogenized and cell suspensions were filtered through a 40-μm cell strainer for the removal of any cell clumps, and erythrocytes were removed via treatment with red blood cells lysis buffer (BioVision). For immuno-fluorescenct staining of cells, cells were resuspended in fluorescence-activated cell sorting (FACS) buffer, which consists of phosphate-buffered saline (PBS) containing 1% fetal bovine serum (FBS) and 0.05% sodium azide. We first gated on the myeloid cells based on their forward (FSC) vs. side (SSC) scatter properties, and then identify each cell type by at least two specific surface marker as follow: CD11b^+^Ly6C^+^ cells for monocytes, CD11c^+^ MHC II^+^ cells for splenic cDCs, and CD11c^+^CD103^−^CD11b^+^ cells for MCs. Generation of BMDCs ------------------- BMDCs were differentiated by culture of *Mst1*^+/+^ and *Mst1*^−/−^ mouse BM cells in the presence of 20 ng/ml GM-CSF for 8 days ([@B9]). Briefly, BM cells were isolated from the femur and tibia bones of *Mst1*^+/+^ and *Mst1*^−/−^ 8- to 10-weeks old mouse (female), followed by RBC lysis. The BM cells were then cultured at a concentration of 2 × 10^5^ cells per ml. The cells were cultured in RPMI containing 10% heat-inactivated FBS (Gibco), 2 mM glutamine, 1 mM sodium pyruvate, 10 mM HEPES, 100 U/ml penicillin, 100 μg/ml streptomycin (Corning), and 50 μM 2-mercaptoethanol (Sigma-Aldrich). The cells were supplemented with 20 ng/ml GM-CSF after 3, 5, and 7 days in the course of the 8-days culture or after 3, 5, 7, and 8 days in the course of the 10-days culture. Antibodies and Flow Cytometric Analysis --------------------------------------- The expression of cell surface molecules on CD11c^+^ BMDCs was investigated in *Mst1*^+/+^ and *Mst1*^−/−^ BMDCs after 8 days in culture unless otherwise specified. One hundred microliters of cells were blocked with Fc blocker (anti-CD16/32 antibody, 2.4G2; BD Bioscience) for 10 min at 4°C, and then incubated with an antibody cocktail for 30 min at 4°C, and washed with FACS buffer. Fluorescent antibodies (eBioscience, BD Biosciences, and Biolegend) were used as follows: CD11c (HL3), CD11b (M1/70), B220 (RA3-6B2), Ly6c (HK1.4), MHC II (NIMR-4 and AF6-120.1), CD40 (3/23), CD80 (16-10A1), CD86 (GL1), and CD131 (JORO50) monoclonal antibodies for cell surface staining. For intracellular staining of TLR9 and CD206, *Mst1*^+/+^ and *Mst1*^−/−^ BMDCs after 8 days of culture were stained for CD11c by incubation with anti-CD11c antibody; subsequently, the cells were fixed and permeabilized using a cytofix/cytoperm kit (BD Biosciences, San Jose, CA), according to the manufacturer\'s instructions, and then stained with anti-TLR9 (M9.D6) or -CD206 (C068C2) antibody. The stained cells were examined by FACS Calibur (BD Biosciences, San Diego, CA), and analyzed using Cell Quest Pro software (BD Biosciences). Mst1 siRNA-mediated knockdown BMDCs were used to analyze the expression of cell surface CD40 and MHC II molecules on CD11c^+^ BMDCs after 36 h culture in the presence of GM-CSF following microporation of Mst1 siRNA. Semi-quantitative RT-PCR ------------------------ *Mst1*^+/+^ and *Mst1*^−/−^ BMDCs after 8 days in culture were used in analysis of cytokine mRNA expression. RNAs were isolated using Ribo-EX reagent (GeneAll), as described by the manufacturer. RNA samples (1 μg) were converted to cDNA by reverse transcription using oligo (dT)18 primer and moloney murine leukemia virus (M-MLV) reverse transcriptase (Enzynomics). The PCR was performed within a range of cycles (24--37 cycles). The sequences of PCR primers used in this study are described in [Table S1](#SM1){ref-type="supplementary-material"}. In Mst1 siRNA-mediated knockdown system, BMDCs were used after 36 h culture in the presence of GM-CSF following microporation of Mst1 siRNA and media change. Cytokine Assay -------------- The quantities of IL-12p40 and IL-23 in the culture supernatants were determined by a sandwich ELISA. To enhance cytokine production, BMDCs were stimulated with 0.1 μM CpG DNA. Production of the proinflammatory cytokines IL-23 and TNF-α secreted by Mst1-knockdown BMDCs was analyzed in the culture supernatants stimulated for 3 or 30 h, respectively, with 0.1 μM CpG DNA by a sandwich ELISA. Mouse IL-12p40 and TNF-α ELISA sets and anti-mouse IL-23p19 (5B2) and anti-mouse IL-12/23 p40 (C17.8) for IL-23 ELISA were purchased from eBioscience. The assays were performed according to the manufacturer\'s instructions. Quantitation of Antigen Uptake ------------------------------ To measure antigen uptake ability, *Mst1*^+/+^ and *Mst1*^−/−^ CD11c^+^ BMDCs after 8 days in culture were incubated with FITC-labeled dextran (Sigma-Aldrich). Briefly, 2 × 10^5^ cells were equilibrated at 4°C or 37°C for 30min and then incubated at 37°C (for a negative control, BMDCs were incubated at 4°C) in medium containing 1 mg/ml FITC-dextran for 1 h. After incubation, the cells were washed twice in PBS to remove excess dextran. The quantitative uptake of FITC-dextran by BMDCs was determined by flow cytometric analysis. We measured percentage of FITC-dextran^+^ *Mst1*^+/+^ and *Mst1*^−/−^ BMDCs following incubation after FITC-dextran treatment. Mixed Leukocyte Reaction (MLR) ------------------------------ *Mst1*^+/+^ and *Mst1*^−/−^ BMDCs (I-A^b^) after 8 days in culture were replated and cultured for 24 h in the presence of GM-CSF and used as stimulators. Allogeneic CD4^+^ T cells from spleen and lymph node of BALB/c mice (I-A^d^) were isolated by positive immunomagnetic selection using MACS with CD4 MicroBeads (Miltenyi Biotec). CD4^+^ T cells were labeled for 10 min at 37°C with 1 μM CFSE. After CFSE staining of CD4^+^ T cells, 1 × 10^5^ CD4^+^ T cells were cultured with *Mst1*^+/+^ and *Mst1*^−/−^ BMDCs at a ratio of 1:10, 20, 40, and 80 (BMDCs:CD4^+^ T cells) for 4 days in order to perform proliferation assay. The proliferation activity of CD4^+^ T cell was measured as dilution of CFSE. For detection of IL-2 production, CD4^+^ T cells were cultured with *Mst1*^+/+^ and *Mst1*^−/−^ BMDCs at a ratio of 1:10 for 3 days, and then supernatants were collected to analyze through ELISA. IL-2--expressing CD4^+^ T cells were analyzed by BD Accuri C6 Plus (BD Biosciences) at a ratio of 1:20 for 3 days after cocultured with *Mst1*^+/+^ and *Mst1*^−/−^ BMDCs. Small Interfering RNA (siRNA) Transfection ------------------------------------------ The siRNA-mediated interference technique was used to silence mouse Mst1 expression. The Mst1-specific sense siRNA sequence (5′-CCG UCU UUC CUU GAA UAC UUU-3′) ([@B34]) was synthesized by ST Pharm (Seoul, Korea), and a scrambled control siRNA was synthesized by Bioneer (Daejon, Korea). siRNAs were transfected into BMDCs after 8 days in culture by Neon Transfection System (Invitrogen), according to the manufacturer\'s instructions. Western Blot Analysis --------------------- *Mst1*^+/+^ and *Mst1*^−/−^ BMDCs were harvested for cell lysis after 7 days in culture. The cells were harvested and then were lysed in lysis buffer (50 mM Tris-Cl pH 8.0, 150 mM NaCl, 1% Triton X-100, 0.1% SDS, 10 mM NaF, 1 mM Na~3~VO~4~, 0.3 mM PMSF, and protease inhibitor cocktail which was from Sigma-Aldrich). Protein concentration was measured using Pierce BCA protein assay kit (Thermo Fisher Scientific), as described by the manufacturer. Equal amounts of protein from whole-cell extracts were separated on 8--10% SDS-PAGE (Bio-Rad) and transferred onto polyvinylidene difluoride (PVDF) membranes (Merck Millipore). The PVDF membrane was then incubated in blocking buffer (Tris-buffered saline containing 0.1% Tween 20 and 5% BSA) for 1 h at room temperature. Then the membranes were incubated with appropriate primary antibody overnight at 4°C with gentle shaking, followed by 1 h of incubation at room temperature with the appropriate horseradish peroxidase-conjugated secondary antibody. The blots were visualized using Amersham ECL Prime Western Blotting Detection Reagent (GE Healthcare and Life Sciences) according to the manufacturer\'s instructions. Western blot digital images were obtained using the Fujifilm LAS-3000 imager. Treatment of DCs With Inhibitors -------------------------------- In case of the experiments using inhibitors, the *Mst1*^−/−^ and Mst1-knockdown BMDCs were treated with each specific inhibitor. *Mst1*^+/+^ and *Mst1*^−/−^ BMDCs after 7 days of differentiation in culture were replated in the presence of GM-CSF for additional 6 h in culture. Vehicle control and MK-2206 (2 μM) were added in culture medium of *Mst1*^+/+^ and *Mst1*^−/−^ BMDCs for western blotting. For flow cytometric analysis, *Mst1*^+/+^ and *Mst1*^−/−^ BMDCs after 8 days of differentiation in culture were replated in the presence of GM-CSF and the following inhibitors: MK-2206 (2 μM), 10074-G5 (25 μM), and (+)JQ-1 (250 nM), and then annexin V^−^CD11c^+^ BMDCs were examined by BD Accuri C6 Plus after 20 h of culture. In Mst1 siRNA-mediated knockdown system, BMDCs were used after 36 h culture in the presence of GM-CSF following microporation of Mst1 siRNA and media change. Vehicle control and the indicated inhibitors were added in culture medium of scrambled control-transfected and Mst1-knockdown BMDCs during the last 6 h. Statistical Analyses -------------------- Statistically significant differences of all data expressed as mean ± SD were assessed by the unpaired Student *t*-test using SigmaPlot 10 software. The statistical differences in cell numbers of *Mst1*^+/+^ and *Mst1*^−/−^ mouse lymphoid organs were analyzed by Mann-Whitney U test using IBM SPSS Statistics 25 software. A *p*-value \< 0.05 was considered statistically significant. Results {#s3} ======= Mst1-Deficiency Triggers a Hyperactivated Phenotype in BMDCs ------------------------------------------------------------ A previous report showed that Mst1 is expressed abundantly in GM-CSF-induced BMDCs compared to their precursor BM cells ([@B20]). In agreement with this report, we also observed that protein levels of Mst1 in BM cells were gradually increased in a time-dependent manner when cultured with GM-CSF, which suggests that Mst1 is involved in GM-CSFR signaling ([Figure S1A](#SM1){ref-type="supplementary-material"}). Previous reports showed that Hippo pathway phosphorylates and suppresses transcriptional coactivator YAP, the component of Hippo pathway ([@B24]). We further investigated whether the protein level of YAP was altered in the BM cell differentiation into BMDCs. The protein level of YAP was inversely correlated with Mst1 expression ([Figure S1A](#SM1){ref-type="supplementary-material"}). Furthermore, the nuclear level of YAP increased in *Mst1*^−/−^ BMDCs ([Figure S1B](#SM1){ref-type="supplementary-material"}). Thus, these results suggest that Hippo pathway is activated in GM-CSF-induced BMDCs and has indispensable roles in GM-CSF-induced activation and maturation of DCs. To clarify the involvement of Mst1 in the activation and maturation of DCs, we compared cell surface expression levels of costimulatory and MHC II molecules, activation/maturation-related cell surface markers, between *Mst1*^+/+^ and *Mst1*^−/−^ BMDCs. There was a greater proportion of *Mst1*^−/−^ BMDCs with high expression of the costimulatory molecules CD40, B7.1, and B7.2 ([Figure 1A](#F1){ref-type="fig"}). Similarly, the percentage of MHC II^hi^ *Mst1*^−/−^ BMDCs was significantly higher than that of MHC II^hi^ *Mst1*^+/+^ BMDCs ([Figure 1A](#F1){ref-type="fig"}). To further investigate the role of Mst1 in the regulation of cell surface expression of costimulatory and MHC II molecules, we compared expression levels of these cell surface molecules in *Mst1*^+/+^ and *Mst1*^−/−^ BMDCs at 4, 6, 8, and 10 days after initiation of culture. *Mst1*^−/−^ BMDCs showed an increase in the expression of the costimulatory molecules, CD40, B7.1, and B7.2, and MHC II in a time-dependent manner compared with that of *Mst1*^+/+^ BMDCs ([Figure 1B](#F1){ref-type="fig"}). To investigate whether *Mst1*^−/−^ MoDCs display phenotypic differences *in vivo*, we compared expression levels of cell surface costimulatory molecules on MCs in the MLN of *Mst1*^+/+^ and *Mst1*^−/−^ mice. The number of migratory MCs was higher in the MLN of *Mst1*^−/−^ mice ([Figure S2A](#SM1){ref-type="supplementary-material"}). Moreover, *Mst1*^−/−^ MCs showed higher expression level of B7.2 than *Mst1*^+/+^ MCs ([Figure S2B](#SM1){ref-type="supplementary-material"}). {#F1} Decreased Ag-uptake activity and the level of mannose receptor (MR) are the hallmarks of mature DCs. Dextran is captured by pinocytosis and MR-mediated endocytosis ([@B5], [@B35], [@B36]). To compare the endocytic activity between *Mst1*^+/+^ and *Mst1*^−/−^ BMDCs, we measured the activity of MR-mediated FITC-dextran uptake. *Mst1*^−/−^ BMDCs showed decreased MFI and percentage of FITC-dextran^+^ cells in a flow cytometry analysis, suggesting a reduced antigen uptake activity of *Mst1*^−/−^ BMDCs ([Figures 2A,B](#F2){ref-type="fig"}). In agreement with this data, the expression level of MR (CD206) was lower on *Mst1*^−/−^ BMDCs than on *Mst1*^+/+^ BMDCs ([Figures 2C,D](#F2){ref-type="fig"}), indicating that Mst1-deficiency induced maturation of GM-CSF-derived BMDCs with the reduced endocytic activity. {#F2} DC maturation and activation are known to affect the expression of a series of inflammatory genes; as a result, they modulate subsequent immune responses ([@B6]). Therefore, we hypothesized that Mst1-deficiency leads to an overproduction of inflammatory cytokines. To explore this hypothesis, we compared mRNA expression and secretion levels of inflammatory cytokines between *Mst1*^+/+^ and *Mst1*^−/−^ BMDCs. The mRNA expression levels of IL-6, IL-12p40, IL-23p19, and TNF-α levels were higher in *Mst1*^−/−^ BMDCs than in *Mst1*^+/+^ BMDCs ([Figure 3A](#F3){ref-type="fig"}). To confirm these increased mRNA expression levels at the protein level, we compared the production of these inflammatory cytokines by *Mst1*^+/+^ and *Mst1*^−/−^ BMDCs stimulated by CpG DNA to amplify activation. Consistent with mRNA expression levels, IL-12p40 and IL-23 production levels were notably higher in CpG-stimulated *Mst1*^−/−^ BMDCs than in *Mst1*^+/+^ BMDCs ([Figures 3B,C](#F3){ref-type="fig"}). We excluded the possibility that the increased production of inflammatory cytokines of *Mst1*^−/−^ BMDCs stimulated through TLR may be due to an increase in TLR expression of *Mst1*^−/−^ BMDCs by determining the intracellular TLR9 expression levels of *Mst1*^+/+^ and *Mst1*^−/−^ BMDCs ([Figure S3](#SM1){ref-type="supplementary-material"}). {#F3} Taken together, Mst1-deficiency in BMDCs induced higher expression levels of cell surface molecules and proinflammatory cytokines, suggesting that Mst1 negatively regulates the activation and maturation of BMDCs. *Mst1*^−/−^ BMDCs Have a Greater Allostimulatory Capacity Than *Mst1*^+/+^ BMDCs -------------------------------------------------------------------------------- To investigate whether hyperactivated *Mst1*^−/−^ BMDCs induce a stronger activation of T cells, we compared the capacity of *Mst1*^+/+^ and *Mst1*^−/−^ BMDCs to stimulate T cells in an *in vitro* allogeneic coculture. As the activation of *Mst1*^−/−^ BMDCs was enhanced in a time-dependent manner ([Figure 1B](#F1){ref-type="fig"}), we replated *Mst1*^+/+^ and *Mst1*^−/−^ BMDCs, generated after 8 days of culture, for a further 24 h in the presence of GM-CSF ([Figure 4A](#F4){ref-type="fig"}) to enhance the allogeneic activity of BMDCs. As expected, *Mst1*^−/−^ BMDCs enhanced allogeneic T cell proliferation ([Figure 4B](#F4){ref-type="fig"}). Moreover, CD4 T cells cocultured with *Mst1*^−/−^ BMDCs comprised a greater proportion of the activated CD44^hi^ T cell subset ([Figure 4C](#F4){ref-type="fig"}). As CD4 T cells strongly produce IL-2 upon activation ([@B37]), we measured IL-2 production levels of CD4 T cells stimulated by *Mst1*^+/+^ or *Mst1*^−/−^ BMDCs. The percentage of IL-2-producing T cells and secretion level of IL-2 were higher in CD4 T cells cocultured with *Mst1*^−/−^ BMDCs than with *Mst1*^+/+^ BMDCs ([Figures 4D,E](#F4){ref-type="fig"}). Accordingly, Mst1-deficiency in BMDCs induced the activation of allogeneic T cells to a greater extent than *Mst1*^+/+^ BMDCs *in vitro*. Collectively, these results show that Mst1 plays an important role in determining the phenotypic and functional activation degree of BMDCs. {#F4} Mst1 Silencing in BMDCs Exhibits the Hyperactivated Phenotype Similar to That Observed in *Mst1*^−/−^ BMDCs ----------------------------------------------------------------------------------------------------------- Next, we confirmed phenotype of the hyperactivated *Mst1*^−/−^ BMDCs in an Mst1-specific siRNA-mediated knockdown system. The efficiency of Mst1 silencing was validated by semi-quantitative RT-PCR ([Figure 5A](#F5){ref-type="fig"}, top) and western blot analysis ([Figure 5A](#F5){ref-type="fig"}, bottom). Silencing of the Mst1 gene in BMDCs increased the surface expression levels of CD40 and MHC II molecules ([Figure 5B](#F5){ref-type="fig"}). The mRNA expression levels of IL-1β, IL-6, IL-23p19, and TNF-α were increased in Mst1-knockdown BMDCs ([Figure 5C](#F5){ref-type="fig"}). In agreement with these data, secretion of these proinflammatory cytokines was enhanced in Mst1-knockdown BMDCs, with IL-23 and TNF-α levels significantly elevated ([Figures 5D,E](#F5){ref-type="fig"}). Accordingly, silencing of Mst1 promoted BMDC hyperactivation, similar to the hyperactivation observed in *Mst1*^−/−^ BMDCs. To perform these Mst1-knockdown experiments, we transfected siRNA targeting Mst1 into BMDCs differentiated after 8 days of culture, which are different from *Mst1*^−/−^ BMDCs in which Mst1 is absent in cells before the differentiation of BM cells into BMDCs. Differentiation rates and cell numbers of BM cells into BMDCs between *Mst1*^+/+^ and *Mst1*^−/−^ cells were similarly obtained ([Figures S4A,B](#SM1){ref-type="supplementary-material"}), which means that *Mst1*^−/−^ BMDCs normally differentiate from BM cells. Furthermore, both *Mst1*^+/+^ and *Mst1*^−/−^ BMDCs induced by GM-CSF had no difference in the percentages of CD11c^+^B220^+^ pDC population ([Figure S4C](#SM1){ref-type="supplementary-material"}). It convinced us to exclude the possibility that BMDC hyperactivation is due to an effect of Mst1-deficiency on cell development *in vitro*. These results are consistent with the idea that Mst1-deficiency in differentiated BMDCs, and not in precursors, gives rise to their hyperactivation. Taken together, these data suggest that endogenous Mst1 in differentiated BMDCs suppresses their hyperactivation. {#F5} Hyperactivated Phenotype Induced by Loss of Mst1 in BMDCs Is Not Due to a Change in GM-CSFR Expression ------------------------------------------------------------------------------------------------------ GM-CSF is involved in the inflammatory phenotype of DCs ([@B4], [@B12], [@B13]). The α subunit of GM-CSF receptor (GM-CSFRα) recruits GM-CSFRβc ([@B38]), which results in the initiation of GM-CSF signal transduction and activation of downstream pathways followed by regulation of the development, survival, and activation of GM-CSF-induced DCs ([@B18]). We first compared the absolute cell numbers of monocytes, the main precursor of BMDCs, in BM of *Mst1*^+/+^ and *Mst1*^−/−^ BMDCs. Cell numbers of monocytes were comparable in BM from *Mst1*^−/−^ and *Mst1*^+/+^ mice ([Figure 6A](#F6){ref-type="fig"}). Next, we checked the expression level of cell surface GM-CSFRβc, which transmit GM-CSF signaling, to check whether the hyperactivation of *Mst1*^−/−^ BMDCs was due to hyperresponsiveness of monocytes to GM-CSF. The expression of cell surface GM-CSFRβc was slightly increased in *Mst1*^−/−^ monocytes but not significantly ([Figure 6B](#F6){ref-type="fig"}). Mst1-knockdown BMDCs showed comparable mRNA expression level of GM-CSFRα ([Figure 6C](#F6){ref-type="fig"}) and the cell surface expression level of GM-CSFRβc [(Figure 6D)](#F6){ref-type="fig"}. As expected, the mRNA expression level of IL-23p19 increased in Mst1-knockdown BMDCs, regardless of the presence of GM-CSF ([Figure 6E](#F6){ref-type="fig"}). These data suggest that responsiveness to GM-CSF does not play a role in the hyperactivation of BMDCs induced by the loss of Mst1. {#F6} Mst1-Deficiency Causes Hyperactivation of BMDCs Through Enhanced Akt1/c-myc Signaling ------------------------------------------------------------------------------------- Previous reports showed that Mst1 antagonizes Akt1 activation in various cell types ([@B39]--[@B41]), including regulatory T cells in which FoxO1/3 proteins, directly and indirectly regulated by Mst1, are involved in their development ([@B29]). We hypothesized that Mst1-deficiency triggers the hyperactivation of GM-CSF-induced BMDCs by regulating Akt1 activity. To test this hypothesis, we investigated Akt1 activity in Mst1-deficient BMDCs after 7 days in culture. Consistent with our hypothesis, the phosphorylation of Akt1 increased in *Mst1*^−/−^ BMDCs compared to that in *Mst1*^+/+^ BMDCs ([Figure 7A](#F7){ref-type="fig"}). Akt1-mediated regulation of c-myc expression plays a crucial role in the determination of an inflammatory phenotype of GM-CSF-induced macrophages ([@B42]), and contributes DC development, regulating survival and maturation ([@B43]). Therefore, we compared the protein level of c-myc in *Mst1*^+/+^ and *Mst1*^−/−^ BMDCs. After 7 days in culture, the protein level of c-myc increased in *Mst1*^−/−^ BMDCs ([Figure 7A](#F7){ref-type="fig"}). We confirmed the gene induction of c-myc in Mst1-knockdown BMDCs. The mRNA expression level of c-myc also increased in Mst1-knockdown BMDCs compared to that in cells transfected with a scrambled control ([Figure 7B](#F7){ref-type="fig"}). To investigate whether Akt1 is a potent inducer of c-myc, *Mst1*^+/+^ and *Mst1*^−/−^ BMDCs after 7 days in culture were further cultured in the presence of GM-CSF and MK-2206. The treatment of BMDCs with MK-2206 decreased the phosphorylation of Akt1, and partially reversed the protein level of c-myc in *Mst1*^−/−^ BMDCs ([Figure 7C](#F7){ref-type="fig"}). Together, these results indicate that Akt1 is involved in the de novo synthesis of c-myc in Mst1-deficient DCs. {#F7} Next, to determine whether the Akt1/c-myc axis is responsible for the hyperactivation of *Mst1*^−/−^ BMDCs, we compared expression levels of the costimulatory molecules, B7.1 and B7.2, on the cell surface of *Mst1*^−/−^ BMDCs treated with a vehicle control or the indicated inhibitors. As Akt1 and c-myc are involved in the survival of DCs, we excluded the dead cells in this comparison. *Mst1*^−/−^ BMDCs treated with inhibitors of c-myc reversed the increase in expression levels of the costimulatory molecules in *Mst1*^−/−^ BMDCs although MK-2206 failed to decrease expression levels of the costimulatory molecules in *Mst1*^−/−^ BMDCs ([Figure 7D](#F7){ref-type="fig"}). To further investigate whether the Akt1/c-myc axis mediates the hyperactivation of BMDCs induced by the loss of Mst1, the mRNA expression level of IL-23p19 was compared in Mst1-knockdown BMDCs treated with the vehicle control or the indicated inhibitors. The mRNA expression level of IL-23p19 in Mst1-knockdown BMDCs treated with inhibitors of Akt1 and c-myc was downregulated compared to that in Mst1-knockdown BMDCs treated with the vehicle control ([Figure 7E](#F7){ref-type="fig"}). Therefore, these data suggest that the enhanced Akt1/c-myc signaling is responsible for the hyperactivation of *Mst1*^−/−^ BMDCs. Thus, Mst1 negatively regulates the Akt1/c-myc axis, which determines the inflammatory phenotype of GM-CSF-induced DCs. Discussion {#s4} ========== Given that DCs are used in vaccination, connecting innate and antigen-specific responses, understanding how the maturation and activation of DCs are regulated is important ([@B14], [@B44], [@B45]). Mst1 is a multifunctional serine/threonine kinase involved in cell proliferation, differentiation, apoptosis, and organ size regulation ([@B21]--[@B24]). Several recent studies have revealed crucial roles for Mst1 in the immune system; specifically, it regulates the survival, proliferation, trafficking, and function of T cells ([@B19], [@B20], [@B25]--[@B30]). Although previous studies have revealed that Mst1 is involved in the induction of reactive oxygen species to clear bacterial infection in macrophages ([@B46]) and in the production of IL-6 ([@B31]) and IL-12 ([@B32]) from DCs, the roles of Mst1 in the activation and maturation of MoDCs are still largely unknown. In the present study, we aimed to clarify the intrinsic role of Mst1 in the determination of the activation status of GM-CSF-induced inflammatory DCs. We found that *Mst1*^−/−^ BMDCs exhibited an increased expression of costimulatory and MHC II molecules and production of several inflammatory cytokines *in vitro*; moreover, the results of Mst1 knockdown in BMDCs are consistent with the idea that Mst1 suppresses the overexpression of several inflammatory cytokines and cell surface molecules in fully differentiated BMDCs. In conclusion, our results suggest that Mst1 negatively regulates the phenotypical and functional activation of GM-CSF-induced DCs. DCs have functional properties depending on their maturation status. Their distinctive intrinsic properties lead to maturation of different subsets ([@B2], [@B11], [@B47], [@B48]). Furthermore, the previous study showed that Mst1 is the negative regulator of proliferation in naïve T cells ([@B19]) and regulates development and function of regulatory T cells ([@B29]). To investigate whether the hyperactivation of *Mst1*^−/−^ GM-CSF-derived DCs is due to a differential development, we compared cell numbers and percentages of the CD11c^+^CD11b^+^ population in *Mst1*^+/+^ and *Mst1*^−/−^ BMDCs. Comparable percentages and cell numbers of the CD11c^+^CD11b^+^ population at an earlier culture time were observed ([Figures S4A,B](#SM1){ref-type="supplementary-material"}). The previous study showed that GM-CSF suppresses the differentiation of pDCs ([@B49]). Consistent with the previous study, we observed that both *Mst1*^+/+^ and *Mst1*^−/−^ BMDCs induced by GM-CSF had no apparent percentage of CD11c^+^B220^+^ pDC population ([Figure S4C](#SM1){ref-type="supplementary-material"}). Thus, these data show that *Mst1*^−/−^ BMDCs from mouse BM cells normally differentiate into CD11c^+^CD11b^+^ inflammatory DCs, which suggests that Mst1 has a redundant role in the *in vitro* differentiation of BMDCs by GM-CSF. Taken together, these data demonstrate that BMDC hyperactivation is not due to an effect of Mst1-deficiency on cell development *in vitro*. In the present study, we did not observe any abnormal death of Mst1-KO mice as a result of a spontaneous autoimmune response *in vivo*, which seems inconsistent with the *in vitro* hyperactivation of *Mst1*^−/−^ GM-CSF-induced DCs. Previous reports have reported a systemic T cell lymphopenia due to defects in homing and survival in Mst1-KO mice ([@B20], [@B25], [@B50]), and Mst1-mutated patients with impaired T cell survival that resulted in primary T cell immunodeficiency ([@B26]). However, a recent study has revealed that DC-specific (CD11c-Cre) conditional Mst1-KO mice exhibit overproduction of IL-6 by DCs, inducing Th17 differentiation and autoimmune response *in vivo* ([@B31]). We speculate that the inconsistency between the normal phenotype of Mst1-KO mice and the hyperactivation of *Mst1*^−/−^ BMDCs *in vitro* may have several explanations. First, Mst1-KO mice have severe T cell lymphopenia in peripheral lymphoid organs (data not shown), which is consistent with the previous reports ([@B19], [@B20]). Second, consistent with a previous report ([@B32]), we did not observe any critical differences in numbers and phenotypic changes of DCs in the spleen of Mst1-KO mice ([Figure S5](#SM1){ref-type="supplementary-material"}). Finally, the critical roles of GM-CSF in inflammation rather than steady state *in vivo* might explain the absence of spontaneous autoimmune responses in Mst1-KO mice. These findings shed light on our understanding of the physiological role of Mst1 in the regulation of activation status of GM-CSF-induced DCs. Mst1 dampens the hyperactivation of BMDCs by regulating the Akt1/c-myc axis rather than GM-CSFR expression. A previous report showed that Mst1 antagonizes Akt1 activation in regulatory T cells in which FoxO1/3 proteins that are directly and indirectly regulated by Mst1 act on their development ([@B29]). Thus, we hypothesized that Mst1-deficiency triggers an enhanced activity of Akt1, which results in the hyperactivation of BMDCs. As expected, phosphorylation of Akt1 was increased in *Mst1*^−/−^ BMDCs ([Figure 7A](#F7){ref-type="fig"}), and the blockade of Akt1 activity reversed the hyperactivation of Mst1-knockdown BMDCs ([Figure 7E](#F7){ref-type="fig"}). Akt1 is also known for controlling the cellular metabolism. Activation-induced T cell metabolic reprogramming ([@B51]) and glycolytic metabolism in TLR-activated DCs ([@B52]) have been suggested, with clear evidence of the involvement of cellular metabolism in immune cell function. The PI3K/Akt1-induced transcription factor, c-myc, is a regulator of cellular metabolism, especially glycolysis, and thereby of the activation of macrophages ([@B42]). Moreover, c-myc, the downstream effector of mTORC1, is involved in the development of DCs ([@B43]). Thus, we tested whether c-myc is involved in the hyperactivation of *Mst1*^−/−^ BMDCs. Consistent with our hypothesis, we demonstrated elevated protein ([Figure 7A](#F7){ref-type="fig"}) and mRNA ([Figure 7B](#F7){ref-type="fig"}) levels of c-myc in Mst1-deficient BMDCs. Inhibitors of c-myc reversed the hyperactivation of Mst1-KO ([Figure 7D](#F7){ref-type="fig"}) and knockdown ([Figure 7E](#F7){ref-type="fig"}) in BMDCs. Although previous studies showed that PI3K/Akt signaling regulates GM-CSF-induced proliferation, survival, and development of DCs ([@B18]), we observed a normal development and yield of CD11b^+^CD11c^+^ *Mst1*^−/−^ BMDCs in an *in vitro* culture ([Figure S4](#SM1){ref-type="supplementary-material"}), which means that the Mst1/Akt1/c-myc pathway has a redundant role in the proliferation and differentiation of BM precursor cells into BMDCs, whereas it is required to maintain a moderate maturation phenotype of GM-CSF-induced DCs. Collectively, Mst1-deficiency triggers the hyperactivation of BMDCs through the overactivation of GM-CSF-induced Akt1/c-myc signaling pathway. We observed that the treatment of Mst1-deficient BMDCs with Akt1 inhibitor partially decreased the protein level of c-myc ([Figure 7C](#F7){ref-type="fig"}) and also failed to reverse expression levels of the costimulatory molecules ([Figure 7D](#F7){ref-type="fig"}), which means that the blockade of Akt1 activity was not sufficient to suppress c-myc-mediated hyperactivation in GM-CSF-stimulated DCs. The recovery of increased costimulatory B7 expression levels in *Mst1*^−/−^ BMDCs might be required for complete inhibition of c-myc expression, even though the modest reduction of c-myc level was sufficient to reverse the mRNA expression level of IL-23p19, a proinflammatory cytokine. Thus, although we have elucidated one crucial mechanism underlying inhibition of hyperactivation of GM-CSF-stimulated DCs, we expect that unknown other mediators might also exist. In summary, we have demonstrated that Mst1 dampens the hyperactivation of DCs via downregulation of the Akt1/c-myc axis in response to GM-CSF, suggesting that Mst1 in mouse inflammatory DCs correlates with GM-CSF-driven disease state. The Mst1/Akt1/c-myc pathway in the regulation of DC activation will give a new insight into understanding of the way how Mst1 regulates appropriate immune responses. These findings shed light on how the maturation and activation of DCs are regulated by a novel endogenous serine/threonine kinase factor. Furthermore, since GM-CSF-induced DCs are a key player in inflammation and autoimmunity ([@B17]), Mst1 can be a new and considerable therapeutic target in the treatment of GM-CSF-derived inflammatory diseases, such as multiple sclerosis, rheumatoid arthritis, and inflammatory bowel disease ([@B17], [@B53]). Data Availability {#s5} ================= The datasets generated for this study are available on request to the corresponding author. Ethics Statement {#s6} ================ The experimental protocols adopted in this study were approved by the Institutional Animal Care and Use Committee of Korea University. Author Contributions {#s7} ==================== K-MC designed and performed all the experiments. K-MC, MK, and TK analyzed and interpreted the experimental results. H-JJ maintained the mice used in this study. E-JC played a role in discussing the results and provided crucial Mst1-related reagents. K-MC and TK collaborated on the manuscript writing. TK supervised the study and corrected the manuscript. Conflict of Interest Statement ------------------------------ The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. We thank Dr. Dae-Sik Lim (Korea Advanced Institute of Science and Technology, Korea), M.Sc. Jae Hwan Kim and Dr. Ji-Yun Lee (Korea University, Korea) for providing valuable reagents and technical assistance. **Funding.** This work was supported by the National Research Foundation of Korea (NRF) grant (NRF-2017R1A2B2009442). Supplementary Material {#s8} ====================== The Supplementary Material for this article can be found online at: <https://www.frontiersin.org/articles/10.3389/fimmu.2019.02142/full#supplementary-material> ###### Click here for additional data file. [^1]: Edited by: Elodie Segura, Institut Curie, France [^2]: Reviewed by: Loredana Saveanu, Institut National de la Santé et de la Recherche Médicale (INSERM), France; Chaohong Liu, Huazhong University of Science and Technology, China [^3]: This article was submitted to Antigen Presenting Cell Biology, a section of the journal Frontiers in Immunology

Realistically recreate the set of your children&65533;s favourite animated series, UNDERWATER! Nickelodeon brings you Dora the Explorer aquarium ornaments and accessories. Complete your collection with Dora&65533;s house, and make this detailed and colourful decoration part of your very own underwater creation. A hiding place for tiny fish or simply a decorative artifact, this piece is sure to put a big smile on your youngster&65533;s face. COLLECT THE ENTIRE SET!!!, while supplies last. (Approximately 4&65533;L x 4&65533;W x 3&65533;H).

Categories The Brain Boosting Power of Walking “All truly great thoughts are conceived by walking” Friedrich Nietzche There’s a harmonious effect that happens when we walk and think at the same time. Walking in a natural environment can improve our attention. Our working memory is also better when we walk at a pace of our own choosing. A brisk walk will produce more brain cells. What’s even more amazing is that stimulation from an enriching environment will increase the connectivity between those brain cells. Those connections will also be more stable. Our consciousness can be shaped more subtly outside in nature. It’s a much more proactive process when we are not distracted by the intrusions of an urban environment. “It is you who decides to examine a plant more closely or to focus on a far flung horizon one moment, then perhaps to lean up against a tree the next. This internally driven sequence of events will then have the additional benefit of restoring control, of giving you a longer time frame in which to develop and deepen your thoughts” Walking Therapy takes advantage of all these brain benefits and adds in even more. There’s an opportunity to think (and feel) more deeply and powerfully as we walk together side by side.

Q: Box: How to download a file to a specific folder? I'm using java sdk provided by box and I saw in the documentation, how to download a file: String fileID = "11111"; BoxFile file = new BoxFile(api, fileID); BoxFile.Info info = file.getInfo(); FileOutputStream stream = new FileOutputStream(info.getName()); file.download(stream); stream.close(); I'm able to download it, but it's putting everything inside my project structure and I want to give a specific path so I can get all my outputs there. A: the BoxFile documentation does not provide an attribute for the file path, where to store the downloaded file. But the API for the FileOutPutStream is providing a solution for this: new FileOutputStream("this/path/to/my/file.txt"); You can find all possible values for the constructor of this class under https://docs.oracle.com/javase/8/docs/api/?java/io/FileOutputStream.html

Agora que as taxas de juro dos novos créditos à habitação estão nos mínimos mais baixos de que há registo — 1,41% em maio passado, em média, segundo o Banco de Portugal —, quem compra casa procura outras vantagens além dos custos reduzidos. Maximizar o montante financiado pode ser uma delas. A maioria dos 29 grupos bancários que preveem conceder financiamento para a compra de casa não empresta mais de 80% do valor do imóvel. Todavia, nos últimos meses, algumas entidades começaram a subir o montante máximo financiado. Agora, alguns consumidores podem conseguir que um banco lhe financie a totalidade do valor de compra. O EuroBic empresta até “100% do valor do investimento”, desde que esse montante não ultrapasse “90% do valor da avaliação” do imóvel, de acordo com o preçário da instituição, atualizado a 3 de julho de 2018. Assim, o EuroBic pode emprestar 100 mil euros a um cliente que queira comprar uma habitação própria por 100 mil euros, se a avaliação bancária apontar para um valor superior a 111.111,11 euros, por exemplo. No início deste mês de julho, entrou em vigor uma recomendação do Banco de Portugal que impõe um empréstimo máximo de 90% do menor entre a avaliação e preço de compra nos financiamentos para aquisição de habitação própria. O EuroBic não está sozinho nos empréstimo superiores a 80% do valor da casa. O Crédito Agrícola e a sucursal portuguesa da Unión de Créditos Inmobiliarios (UCI) fornecem crédito até 90% do preço de aquisição, desde que o valor de avaliação bancária seja igual ou superior a esse montante. O Santander Totta também pode emprestar até 90% do preço de compra, mas atribui um teto máximo de 85% do valor de avaliação. Quem dá crédito à habitação superior a 80% do preço de compra? Foram considerados os empréstimos a taxa variável com garantia hipotecária para compra de habitação própria no regime geral. Há pouco mais de dois anos, nenhum emprestava mais de 85% do valor do imóvel. Instituição Taxa de juro Montante máximo financiado EuroBic Euribor 6 ou 12 meses + 1,49% a 3,10% 90% do valor de avaliação, até 100% do valor de compra Crédito Agrícola Euribor 12 meses + 1,20% a 2,65% 90% do valor do imóvel (o menor entre avaliação e compra) UCI Euribor 6 meses + 1,75% a 2,75% 90% do valor do imóvel (o menor entre avaliação e compra) Santander Totta Euribor 12 meses + 1,15% a 2,50% 85% do valor de avaliação, até 90% do valor de compra Banco BPI Euribor 12 meses + 1,50% a 4,10% 85% do valor do imóvel (o menor entre avaliação e compra) Banco CTT Euribor 12 meses + 1,20% a 2,10% 85% do valor do imóvel (o menor entre avaliação e compra) Caixa Geral de Depósitos Euribor a 12 meses + 1,30% a 5,35% 80% do valor de avaliação, até 90% do valor de compra Millennium bcp Euribor 12 meses + 1,25% a 3,00% 80% do valor de avaliação, até 90% do valor de compra Novo Banco Euribor 12 meses + 1,25% a 5,00% 80% do valor de avaliação, até 90% do valor de compra Fonte: bancos a 13 de julho de 2018. É possível que, em situações específicas, outros bancos além dos que estão na lista anterior emprestem mais de 80% do valor da habitação. É o caso, por exemplo, dos empréstimos dos clientes que adquirem imóveis nas carteiras dos bancos financiadores. A lista anterior inclui algumas das instituições financeiras que propõem os spreads — a margem de lucro na operação — mais baixos da banca, como o Santander Totta (desde 1,15%), o Crédito Agrícola (1,20%) e o Banco CTT (1,20%), mas exclui outras, como o Bankinter, que tem um spread mínimo de 1,15%, mas financia até 80% do valor do imóvel.

Explanation: What makes comet tails so colorful? This photograph of Comet Hyakutake was taken the night of April 18th and highlights different components of the tail. The gold and red tail features are dust, made predominately of little bits of rock and carbon. The dust tail shines by reflecting sunlight. Extending past the dust tail is the comet's ion tail, shown here glowing in blue. The ion tail is composed mostly of ions of water, carbon monoxide, and cyanogen. The ion tail glows by emitting light when elections re-combine with electrically charged ions to make uncharged molecules. The photograph was taken just north of Kansas City, Missouri, USA.

208 Okla. 192 (1953) 255 P.2d 277 BOETTCHER OIL & GAS CO. et al. v. LAMB et al. No. 35567. Supreme Court of Oklahoma. March 17, 1953. Mont R. Powell, Sam Hill, and William R. Saied, Oklahoma City, for petitioners. Kerr, Lambert, Conn & Roberts, Ada, and Mac Q. Williamson, Atty. Gen., for respondents. *193 WELCH, J. In this case it appears that on the 20th day of September, 1951, John David Lamb, deceased, was employed by the Boettcher Oil & Gas Company as a welder and mechanic; that on said day he suffered a heart injury from which he died. On November 9, 1951, Lois L. Lamb, wife of deceased, filed a claim for compensation before the State Industrial Commission on her behalf, and on behalf of two minor dependent children of deceased against Boettcher Oil & Gas Company, employer of deceased, and the State Insurance Fund. Compensation is claimed under the provisions of Title 85, chap. 2, S.L. 1951, pp. 267 to 270, inclusive, commonly called the Death Benefit Act. The trial commissioner, in substance, found: That on the 20th day of September, 1951, John David Lamb, deceased, while in the employ of Boettcher Oil & Gas Company sustained an accidental personal injury consisting of an injury to his heart; that decedent left surviving him Lois L. Lamb, his widow, David Harold Lamb, a son age 16 years, and Linda Kay Lamb, a daughter age 8 years, all of whom were dependent upon deceased for their support, and that said persons are entitled to receive compensation in the sum of $13,500 to be divided proportionately between them, each to receive $4,500. An award was entered accordingly which was sustained on appeal to the commission en banc. Boettcher Oil & Gas Company and State Insurance Fund, hereinafter referred to as petitioners, bring the case here to review this award and rely solely to vacate the award on the ground that deceased sustained no accidental injury while in the employ of Boettcher Oil & Gas Company; that his death was not caused by an accidental injury, but was due to other causes. This assignment requires a review of the evidence. The evidence shows that deceased, on the 20th day of September, 1951, while in the employ as above stated, sustained an injury to his heart caused while cranking a motor. A fellow employee testified in behalf of respondent, in substance, as follows: On the morning of the day Mr. Lamb sustained his injury he and Mr. Lamb were engaged in overhauling a motor; that they completed that task about noon of that day; that at about 1:30 p.m., Mr. Lamb attempted to start the motor for the purpose of placing it on a pump. It was necessary to crank the motor by the use of a hand crank. The motor was hard to crank, it required very heavy lifting to turn the motor over. After he had succeeded in turning the motor over about five times he stepped back and shrugged and twisted his shoulders and stated that he was suffering severe pain in his chest. The witness then started to crank the motor, but was unable to get it started, and he and Mr. Lamb eventually succeeded in starting the motor. About 20 minutes after the motor was started Mr. Lamb, this witness, and two other employees started over to another lease owned by the Boettcher Oil & Gas Company. The lease was about three quarters of a mile from the place where the motor was cranked. In arriving at that lease Mr. Lamb cranked another and smaller motor. This task he completed about 3:30 p.m. Mr. Lamb, the witness, and two other employees then started to play dominoes. Just as the game started Mr. Lamb arose, walked to the door and stated that he could not get his breath, and that he was still suffering great pain in his chest. He was then taken to a hospital at Ada, Oklahoma, for treatment *194 and examination. An intern met him at the hospital and obtained a history of the case from him, reduced it to writing and delivered it to the doctor. The intern testified at the hearing as to the history obtained which is substantially the same as above detailed. The doctor testified that he first saw John David Lamb at the Valley View Hospital at Ada, Oklahoma, at about 4 o'clock in the afternoon of September 20, 1951. He was then in a dying condition. He was pulseless, but had a slight heartbeat. He was given oxygen and other stimulants in an effort to revive him, but without success and that Mr. Lamb died; that he was furnished a history of the case by the intern, and from such history, his examination and experience as a physician, expressed the opinion that Mr. Lamb's death was caused by overexertion, that it produced a coronary occlusion which means that one or more of the arteries was blocked by an embolus. The doctor further testified: "Q. What, in your opinion, Doctor, was the thing that produced his death? A. Well, the coronary occlusion produced the death, but when you go to classifying them, you have a little bit of difficulty, because everything depends on what a person means by accident or injury. If you mean by a blow or something like that, then it — he didn't receive a blow or anything of that nature. The overexertion produced a marked increase in the heart beat, an exertion of the heart, which tore an embolus loose. If it lodges in the coronary arteries, then you may have sudden death. If it lodges away in the stomach or in the chest muscles or in the legs, then you may have a little pain for a day or two, and you are over it. But your heart, you have a different matter. "Q. Then, Doctor, if I understand your explanation, it was the strain and stress under which he was working that produced the injury to that artery? A. That's right." No other physician testified in the case. Other employees of petitioner Boettcher Oil & Gas Company, however, testified that they did not hear respondent complain of any pain in his chest until they started playing dominoes. The evidence further shows that deceased had worked for petitioner Boettcher Oil & Gas Company doing the same kind of work for about three years prior to September 20, 1951, and that he had never complained of any pain in his chest or any kind of heart trouble. We think the evidence above detailed is sufficient to sustain the finding of the commission that the injury sustained by deceased constituted an accidental injury within the meaning of the Workmen's Compensation Act. We have heretofore held that an injury sustained by a workman to his heart caused by overexertion and strain constitutes an accidental injury within the meaning of the Workmen's Compensation Act. Carden Mining & Milling Co. v. Yost, 193 Okla. 423, 144 P.2d 969; Clarksburg Paper Co. v. Roper, 196 Okla. 504, 166 P.2d 425; Gulf Oil Corporation v. Rouse, 202 Okla. 395, 214 P.2d 251. In Producers Drilling Co. v. Percival, 207 Okla. 17, 246 P.2d 374, we held: "An injury sustained by an employee consisting of a strain to his back while engaged in cranking a motor constitutes an accidental injury within the meaning of the Workmen's Compensation Act, 85 O.S. 1951, § 3, subd. 7." We think the evidence also sufficient to sustain the finding of the commission that Mr. Lamb's death occurred from a heart injury sustained by him on the 20th day of September, 1951, while in the employ of petitioner Boettcher Oil & Gas Company. Award sustained. HALLEY, C.J., JOHNSON. V.C.J., and CORN, DAVISON, O'NEAL, WILLIAMS, and BLACKBIRD, JJ., concur.

ALBANY — Some uptown residents want to halt dense development in their neighborhoods that they say isn’t following proper regulations, will drastically change the character of the area, and put stress on the city’s infrastructure. Others are pushing back, saying those in opposition simply don’t want development in their backyards, or things to change in Albany. They argue that denser development will make Albany more walkable and increase the tax base. Eagle Hill resident Vincent Riguso has launched a ‘Stop the Stories’ campaign calling for a moratorium on development in Albany. It is supported by other residents opposed to recent developments and named to reflect their concerns over building heights. He questioned why the new Unified Sustainable Development Ordinance, or USDO, isn’t keeping development consistent with existing neighborhood characteristics and isn’t considering impacts from the development on traffic, infrastructure and environmental concerns. “It feels as if the goal, in actuality, is to bring in loads of renters at the expense of those of us who have been here for decades, if not generations, and who have been contributing to the community in all kinds of ways,” Riguso wrote in a letter to city officials recently. The growing vocal opposition has prompted Walkable Albany founder Andrew Neidhardt to organize a counter rally to provide a different perspective. “We have seen recently in the city that folks are opposing developments that are creating walkable communities in our city. They’re opposing it simply because it’s a change to the way things were,” he said. “We just want to be a voice in favor of development where it’s appropriate and it’s improving walkability for people.” Both groups are expected to converge around 6 p.m. Monday at City Hall prior to the regular Common Council meeting. Many of the proposed projects in Albany have brought up similar concerns among residents – increased traffic, possible exacerbation of stormwater issues, and destruction of the character of their neighborhoods. Key among those concerns is the height of the buildings. The proposed six-story building at 1211 Western Ave. is one of several projects that has drawn the ire of the Stop the Stories group. “We don’t want to stop development. That’s the last thing we want to do,” Riguso said. He said a building of that size doesn’t belong in his neighborhood. The Western Avenue project would demolish an existing three-story office building to make way for a six-story, 137-unit apartment complex. It would be located next to the University at Albany’s administration building, which is four-stories, and is next to a gas station and strip mall along Western. Opponents also suggest that the building material creates a fire hazard because it’s a wood structure. However, Albany Fire Chief Joseph Gregory said the projects are up to federal, state and local code and he’s approved the plans for the Western Avenue project. The site is zoned for up to five stories and the developer plans to install a blue or green roof as part of an incentive that will allow another story. Albany Planning Commissioner Chris Spencer said previous zoning would have allowed for a building up to 85 feet, which is roughly eight stories. The additional story incentive offered through installing green or blue roofs – those with either vegetation on top to absorb rainfall or retention tanks to collect water until it can be released during dry weather – has become popular among developers looking to make a proposal financially viable. Each new development’s stormwater runoff can be no more than what would come from the site undeveloped during a 10-year storm, Albany Water Department Commissioner Joseph Coffey said. That means each project is doing “significant” stormwater detention, meanwhile the city continues to invest millions in stormwater management efforts from diversion and detention to new technology for monitoring and management, he said. “We have requirements that we have to meet with best management projects as part of our Combined Sewer Overflow permit, and we’re meeting those,” Coffey said. “They are factored into our reviews of every one of those projects, and if they can’t be done, we can’t approve the project.” Common Councilperson Judy Doesschate has proposed postponing the blue/green roof incentive until at least the end of June 2020 to determine whether it's benefiting the city. The planning department is proposing something similar, but the moratorium isn’t as long. “If we’re getting no additional benefit from this blue roof or green roof technology, why are we upsetting the general scheme of the USDO with regard to what we already determined were appropriate height requirements for each individual zone?” Doesschate said. She said the multi-building, mixed-use project proposed for 563 New Scotland Ave. will rise as high as five stories, with one of its buildings contrasting with the majority of the one and two-story buildings in the neighborhood. It is across from St. Peter’s Hospital. The area is zoned for buildings to be as high as 3 ½ stories, but the developer is installing a green roof and has received a variance to allow for five stories. Spencer said the planning and the water departments are assessing the benefits of the mechanisms but noted the water department does see value in them. Development overall can benefit the city, he said. “Adding development and growing the tax base has benefits for people’s property taxes, for the services we can provide and the quality we can have throughout the city,” Spencer said. “You can’t tax your way out of things, you have to grow your way out.” Doesschate cautioned that with developers receiving tax breaks, the city gets a fraction of the tax money due for several years before the payment-in-lieu-of-taxes, or PILOT, expires. The city receives at least the taxes due based on the undeveloped property’s value, but taxes on the improvements are phased in depending on the agreement.

El acuerdo es por 9 millones de dólares (1,2 millones son para Cerro Porteño y a Racing le quedarán 7,8 millones) y restan definirse detalles del contrato y ver la documentación necesaria por parte del jugador. Atlanta United, dirigido por Gerardo Martino, jugará su primera temporada en la Major League Soccer y ya contrató a otro paraguayo, Miguel Almirón, de Lanús, como una de sus principales figuras. Según Martino, los dos futbolistas paraguayos son los mejores extranjeros en el torneo argentino en la temporada 2016 y cuenta con ambos para formar una delantera demoledora en su sueño de triunfar en la millonaria liga estadounidense.

Tag Archives: IKEA What a cool idea! IKEA wanted customers to be able to view their digital catalog on-the-go, so they introduced a new sewing kit that includes conductive thread to customize your gloves or mittens to be touchscreen-compatible. That way, you can … Continue reading →

Role of glia in K+ and pH homeostasis in the neonatal rat spinal cord. Stimulation-evoked transient changes in extracellular potassium ([K+]e) and pH (pHe) were studied in the neonatal rat spinal cords isolated from 3-13-day-old pups. In unstimulated pups the [K+]e baseline was elevated and pHe was more acid than that in Ringer's solution (3.5 mM K+, pH 7.3-7.35). The [K+]e and pHe in 3-6-day-old pups was 3.91 +/- 0.12 mM and pHe 7.19 +/- 0.01, respectively, while in 10-13-day-old pups it was 4.35 +/- 0.15 mM and 7.11 +/- 0.01, respectively. The [K+]e changes evoked in the dorsal horn by a single electrical stimulus were as large as 1.5-2.5 mM. Such changes in [K+]e are evoked in the adult rat spinal cord with stimulation at a frequency of 10-30 Hz. The maximal changes of 2.1-6.5 mM were found at a stimulation frequency of 10 Hz in 3-6-day-old animals. In older animals the [K+]e changes progressively decreased. The poststimulation K(+)-undershoot was found after a single stimulus as well as after repetitive stimulation. In 3-8-day-old pups, the stimulation evoked an alkaline shift, which was followed by a smaller poststimulation acid shift when the stimulation was discontinued. In pups 3-4-days-old the stimulation evoked the greatest alkaline shifts, i.e., by as much as 0.05 pH units after a single pulse and by about 0.1 pH units during stimulation at a frequency of 10 Hz. In 5-8-day-old pups, the alkaline shift became smaller and the poststimulation acid shift increased.(ABSTRACT TRUNCATED AT 250 WORDS)

Q: Meaning of かまって in the following title At the end of each tankōbon, the manga DEAD Tube has always an extra page with a funny drawing/short story. Every time it has a different title. In the volume in question the title of the extra page is: かまって！？水野さん！ In the page, the character called Mizuno is just bored while their friends are away. What is the meaning of かまって in the title? I guess it comes from 構う, but I don't know how to interpret in this context. Is it related to this question? How would you translate the title? You can see the original page here. Thank you for your help! A: Basically, 「かまって！」 only has one usage at least in Standard Japanese. It is the "casual request" form of 「構{かま}う」, meaning: "Please pay (more) attention to me!" "Talk to me!" or "Look at me!" "Don't ignore me!" "Take (better) care of me!" It is most often uttered by children, followed by women. In your manga, a question mark is used, so I would interprete it as: "Mizuno looks as if she wanted to say 「かまって！」 to someone/people". I know that sounds wordy, but that would be the "feel" of the title to me. Thus. this is directly related to the other question that you linked to. 「かまってちゃん」 means an "attention-seeker" in colloquial Japanese.

Hurrah for Ants! Many of us are familiar with the children’s counting song, “The ants go marching one by one…” Scientists have now discovered that at least some species of ants do appear to be counting their steps when they go marching out in search of food. After training some desert ants to look for food in a specific location, scientists investigated how the ants could consistently locate the same source of food with seeming ease. It is known that ants will leave scent markers to guide their nest mates back to a food source. This is a rather laborious process, however, and slows the ants down. The scientists glued extensions on the legs of some ants, lengthening their strides. When these stilt-walking ants were released to search for food, they started out in the right direction, but overshot the food source every time. Scientists then cut the legs off other ants at the first joint, shortening their stride. Again the ants were released to search for food. This time they stopped short of the food source. The scientists concluded the scout probably lays a scent trail initially, but then the workers memorize the number of steps needed to arrive at the food. This is surprising evidence of relatively complicated communication from such a tiny creature. Ants have a variety of ways to tell each other not only of food, but also warn the colony of danger. As mentioned above, scents, or pheromones, are an integral part of ant communication. Weaver ants have two scent glands in their abdomens and four more in their heads. They can release one or more of these chemicals at a time to convey a variety of messages. Entomologists estimate they can employ between 10 and 20 chemical “phrases” or “words” to communicate with their nest mates. These pheromone messages may be left on a hard surface or released into the air. Weaver ants also appear to use their feces to mark their territory. It is suspected there are scent components at work here, too. A majority of species also communicate with sound. They can produce a high pitched squeak by rubbing a thin scraper located on their waist against a series of tiny parallel ridges on the adjacent abdomen. The signal is barely audible to humans. The squeaking is used for a variety of messages. Ironically, the ants are not influenced by the audible sound of the signal, but rather by the portion of the vibrations that come to them through the soil. Ants have amazingly complicated communication systems. Are these the product of random chance, or complex by design? We don’t have to guess; God has already told us “look to the ant” (Prov. 6) for instruction. We can only add, “Hurrah! Hurrah!”

Q: XNA Collision Detection - Vector2.Reflect - Help Calculating the Normal of a Circle Sprite - C# I'm having trouble wrapping my mind around how to calculate the normal for a moving circle in a 2d space. I've gotten as far as that I'm suppose to calculate the Normal of the Velocity(Directional Speed) of the object, but that's where my college algebra mind over-heats, any I'm working with to 2d Circles that I have the centerpoint, radius, velocity, and position. Ultimately I'm wanting to use the Vector2.Reflect Method to get a bit more realistic physics out of this exercise. thanks ahead of time. EDIT: Added some code trying out suggestion(with no avail), probably misunderstanding the suggestion. Here I'm using a basketball and a baseball, hence base and basket. I also have Position, and Velocity which is being added to position to create the movement. if ((Vector2.Distance(baseMid, basketMid)) < baseRadius + basketRadius) { Vector2 baseNorm = basketMid - baseMid; baseNorm.Normalize(); Vector2 basketNorm = baseMid - basketMid; basketNorm.Normalize(); baseVelocity = Vector2.Reflect(baseVelocity, baseNorm); basketVelocity = Vector2.Reflect(basketVelocity, basketNorm); } basePos.Y += baseVelocity.Y; basePos.X += baseVelocity.X; basketPos.Y += basketVelocity.Y; basketPos.X += basketVelocity.X; basketMid = new Vector2((basketballTex.Width / 2 + basketPos.X), (basketballTex.Height / 2 + basketPos.Y)); baseMid = new Vector2((baseballTex.Width / 2 + basePos.X), (baseballTex.Height / 2 + basePos.Y)); A: First the reflection. If I'm reading your code right, the second argument to Vector2.Reflect is a normal to a surface. A level floor has a normal of (0,1), and a ball with velocity (4,-3) hits it and flies away with velocity (4,3). Is that right? If that's not right then we'll have to change the body of the if statement. (Note that you can save some cycles by setting basketNorm = -baseNorm.) Now the physics. As written, when the two balls collide, each bounces off as if it had hit a glass wall tangent to both spheres, and that's not realistic. Imagine playing pool: a fast red ball hits a stationary blue ball dead center. Does the red ball rebound and leave the blue ball where it was? No, the blue ball gets knocked away and the red ball loses most of its speed (all, in the perfect case). How about a cannonball and a golf ball, both moving at the same speed but in opposite directions, colliding head-on. Will they both bounce equally? No, the cannonball will continue, barely noticing the impact, but the golf ball will reverse direction and fly away faster than it came. To understand these collisions you have to understand momentum (and if you want collisions that aren't perfectly elastic, like when beanbags collide, you also have to understand energy). A basic physics textbook will cover this in an early chapter. If you just want to be able to simulate these things, use the center-of-mass frame: Vector2 CMVelocity = (basket.Mass*basket.Velocity + base.Mass*base.Velocity)/(basket.Mass + base.Mass); baseVelocity -= CMVelocity; baseVelocity = Vector2.Reflect(baseVelocity, baseNorm); baseVelocity += CMVelocity; basketVelocity -= CMVelocity; basketVelocity = Vector2.Reflect(basketVelocity, basketNorm); basketVelocity += CMVelocity;

A new approach for extracellular spin trapping of nitroglycerin-induced superoxide radicals both in vitro and in vivo. Anti-ischemic therapy with nitrates is complicated by the induction of tolerance that potentially results from an unwanted coproduction of superoxide radicals. Therefore, we analyzed the localization of in vitro and in vivo, glyceryl trinitrate (GTN)-induced formation of superoxide radicals and the effect of the antioxidant vitamin C and of superoxide dismutase (SOD). Sterically hindered hydroxylamines 1-hydroxy-3-carboxy-2,2,5,5-tetramethylpyrrolidine (CP-H) and 1-hydroxy-4-phosphonooxy-2,2,6,6-tetramethylpiperidin (PP-H) can be used for in vitro and in vivo quantification of superoxide radical formation. The penetration/incorporation of CP-H or PP-H and of their corresponding nitroxyl radicals was examined by fractionation of the blood and blood cells during a 1-h incubation. For monitoring in vivo, GTN-induced (130 microg/kg) O2*- formation CP-H or PP-H were continuously infused (actual concentration, 800 microM) for 90 to 120 min into rabbits. Formation of superoxide was determined by SOD- or vitamin C-inhibited contents of nitroxide radicals in the blood from A. carotis. The incubation of whole blood with CP-H, PP-H, or corresponding nitroxyl radicals clearly shows that during a 1-h incubation, as much as 8.3% of CP-H but only 0.9% of PP-H is incorporated in cytoplasm. Acute GTN treatment of whole blood and in vivo bolus infusion significantly increased superoxide radical formation as much as 4-fold. Pretreatment with 20 mg/kg vitamin C or 15,000 U/kg superoxide dismutase prevented GTN-induced nitroxide formation. The decrease of trapped radicals after treatment with extracellularly added superoxide dismutase or vitamin C leads to the conclusion that GTN increases the amount of extracellular superoxide radicals both in vitro and in vivo.

The world of esports has lately been gaining the attention of the world. The number of multiplayer games and players increase day by day. But that is not all as the prize pools for these competitive multiplayer gaming tournaments skyrocket as well. Derived from the words “electronic sports,” esports are multiplayer video games played by professional gamers competing with each other or one another. These tournaments are usually shown live for real-time viewing, and sometimes for real-money betting. The popularity of esports even got players into making it as a legitimate career path since the prize pools of each tournament can actually match, or even surpass, the earnings of professional athletes. But how are these massive prize pools accumulated? And how do each team earn from this growing industry? Massive Prize Pools Explained The reward given to players during tournaments is called a prize pool. It is collected money given to all the competing teams at the end of the tournament, with the winner getting the biggest share. Normally, the prize pool comes from the buy-ins of players’ entry fees. But that is not the case with esports. Companies get the prize pool from crowdfunding. They raise small amounts of money, usually via internet marketing, from a huge number of people or supporters of the game. For instance, this year, Dota 2 The International 2018’s prize pool was worth $25,532,177. But how did Valve, the developer of the game, come up with this massive prize pool? A blog post from The Esports Observer said that the prize pool was accumulated from Valve’s US$1.6 million contribution plus the sales of its in-game digital compendium called Battle Pass. The post said: “Valve takes 25% of Battle Pass sales and dumps them directly into the annual International prize pool, generating millions of dollars of additional reward for pro teams while amplifying the stakes for both players and viewers alike.” The Battle Pass can give players exclusive deals such as in-game treasure, cursor packs, new music, digital packs, and bonus game modes for US$9.99. Having a Battle Pass also gives players points that will help them level up. Players can buy a Battle Pass worth US$2.49 to increase five levels. To increase 11 levels, players need to buy a pass for US$4.99, and those planning to go up 24 levels should buy a pass for US$9.99. According to Statistica, there were over 739,000 players online at the same time on Steam on October 13, 2018. There was even a time in March 2016 that the number of players reached 1.29 million. With that huge number, it came as no surprise when Valve was able to accumulate a prize pool worth more than US$20 million, mostly out of the Battle Pass purchases. The more players and supporters an esport has, the higher the in-game purchases it can earn, and the bigger its prize pool can get. Biggest Prize Pools in Esports History As of this writing, the biggest esport in terms of prize pool is Fortnite. On May 21, 2018, The Fortnite Team released a statement through its Twitter account that its game developer, Epic Games, will provide US$100,000,000 for the 2018 to 2019 season tournament. The tweet read: “Grab your gear, drop in and start training. Epic Games will provide $100,000,000 to fund prize pools for Fortnite competitions.” Grab your gear, drop in and start training. Epic Games will provide $100,000,000 to fund prize pools for Fortnite competitions.https://t.co/ZcBe9fZD0S — Fortnite (@FortniteGame) May 21, 2018 After the announcement, esports analysts shared their thoughts on this massive prize pool. Bethany Lyons, SuperData esports analyst, said: “By dedicating $100MM to fund prize pools competitions, Epic has made Fortnite the biggest esports in the world in terms of prize money.” Before this announcement, the biggest prize pool in the history of esports was from Dota 2 The International 2018. The prize pool of that event was worth US$25,532,177. But that enormous prize was given not only to the winner. The money was divided among all the teams in the said tournament. When The International 2018 ended, the prize money was shared among 18 teams. According to Dota 2 Prize Pool Tracker, the following was the prize pool distribution: 18 th and 17 th place – US$63,830 and 17 place – US$63,830 16 th to 13 th place – US$127,661 to 13 place – US$127,661 9 TH to 12 th place – US$382,983 to 12 place – US$382,983 7 th to 8 th place – US$638,304 to 8 place – US$638,304 5 th to 6 th place – 1,148,948 to 6 place – 1,148,948 4 th place – US$1,787,252 place – US$1,787,252 3 rd place – US$2,680,879 place – US$2,680,879 2 nd place – US$4,085,148 place – US$4,085,148 1st place –US$11,234,158 In related manner, The International 2017 is the third in rank with its more than US$24 million prize pool. The fourth top esport tournament based on prize pool is The International 2016 with its prize money worth more than US$20 million. Confederations Cup, Other Sporting Events Outprized The increasing prize pools of esports tournaments are so massive that some even outprized the wins a player or a team could get from traditional sporting events. For one, last year’s Confederations Cup had only US$20 million prize pool. Indy 500 2018’s US$13 million also lost to The International 2018’s over US$25 million prize money. Other sports outprized were US Open 2018, The Masters 2017, Stanley Cup 2018, and ICC Championship Trophy. Even US Open 2017 Tennis’ US$50 million prize pool lost to Fortnite’s announced US$100 million. The International 2018 has a bigger prize pool than the US Open 2018. A report by ESPN said: “The 2018 U.S. Open announced a record $53 million prize pool for the tennis tournament later this month. When broken down by event, both the Men’s and Women’s singles tournaments have a total prize pool of approximately $20 million. The singles champion and runner up will receive $3.8 million and $1.85 million respectively.” Apparently, the Singles champion prize of US$1.85 million is still too far from The International 2018’s first-place prize of more than US$11 million. Esports in Online Betting Industry Esports have already reached too wide an audience that popular sportsbooks now add these gaming tournaments to the betting markets on offer. Not surprisingly, there are now online bookies that offer betting options for the biggest esports tournaments including CS:GO, Dota 2, StarCraft 2, Overwatch, and League of Legends, among others. Before, only basketball, baseball, football, soccer, and other popular physical sports were included in the list of betting options for online sportsbook punters. But today, esports now secured their own betting category in online betting portals. Among the popular sites that have esports betting are bet365, OneHash, Pinnacle, and Unibet. Who knows how much bigger the future esports prize pools can get? But given the immense support the industry has been getting, some people predict that there will come a time when the prize money offered by these gaming tournaments will reach $100,000,000.

Q: how to remap I want to use the following code to quickly go from insert mode -> command mode. The problem is it just doesnt seem to be registering my keys. I've tried with control (nothing happens) and i've also tried with command (D) and it just says spellcheck not activated. " Quick command mode from insert imap <C-;> <esc>: How do i go about doing this? is there an easier vim way that I'm not aware of? A: Certain Ctrl chords can't be mapped, including Ctrl-;. This is mentioned in this FAQ, see also the Vim FAQ: 20.4. I am not able to create a mapping for the <xxx> key. What is wrong? First make sure that the key is passed to Vim. In insert mode, press Ctrl-V followed by the desired key. You should see the keycode corresponding to the key. If you do see the keycode, then you can create a mapping for the key using the following command: :map <C-V><xxx> <your_command_to_be_mapped> For more information, read :help map-keys-fails :help :map-special-keys :help key-codes The tip about trying to print the character using Ctrl-V is good to remember if you run into this problem with another key combo.

Filed 8/12/14 P. v. Nava CA5 NOT TO BE PUBLISHED IN THE OFFICIAL REPORTS California Rules of Court, rule 8.1115(a), prohibits courts and parties from citing or relying on opinions not certified for publication or ordered published, except as specified by rule 8.1115(b). This opinion has not been certified for publication or ordered published for purposes of rule 8.1115. IN THE COURT OF APPEAL OF THE STATE OF CALIFORNIA FIFTH APPELLATE DISTRICT THE PEOPLE, F065082 Plaintiff and Respondent, (Super. Ct. No. 11CM2478) v. JOSE AMEZCUA NAVA, OPINION Defendant and Appellant. F068148 In re JOSE AMEZCUA NAVA, On Habeas Corpus. APPEAL from a judgment of the Superior Court of Kings County. Donna Tarter, Judge. ORIGINAL PROCEEDINGS; petition for writ of habeas corpus. Robert J. Beles and Manisha Daryani for Defendant and Appellant and for Petitioner.  Kamala D. Harris, Attorney General, Dane R. Gillette, Chief Assistant Attorney General, Michael P. Farrell, Assistant Attorney General, Catherine Chatman and Daniel B. Bernstein, Deputy Attorneys General, for Plaintiff and Respondent and for Respondent. -ooOoo- A jury convicted appellant Jose Amezcua Nava of one count of committing a lewd and lascivious act on a child under the age of 14 years, his 10-year-old niece. Nava contends reversal is required because of several evidentiary errors, prosecutorial misconduct, and ineffective assistance of counsel. We reject Nava’s contentions and affirm the judgment. FACTUAL AND PROCEDURAL SUMMARY Prosecution Evidence In July 2011, Nava was 59 years old. The victim would often come over to Nava’s house to play, swim in the pool, and get help with her homework from her 28- year-old cousin, Jesse. The charged offense occurred when the victim was staying at Nava’s house for a three-day visit from Friday, July 22, 2011, to Sunday, July 24. The victim testified that on Saturday morning Jesse, Nava’s 28-year-old daughter, asked her to get milk from a refrigerator located in a detached garage at the residence. The milk was for a baby that was staying at the house. The victim went alone to the garage. When she went inside, Nava was already there fixing something. The victim asked him if he could help her get milk from the refrigerator. Nava opened the refrigerator door and the victim started looking for the milk. While the victim was looking for the milk, Nava came up behind her, put his hand under her shirt, and started squeezing her breasts. Leaving one hand on her breasts, Nava put his other hand inside the victim’s shorts and underwear and started squeezing her vagina. It hurt and the victim told Nava to leave her alone. Nava stopped and warned the 2. victim that if she told her mom or dad, they would hit her, and if she told her aunt, her aunt would kick him out of the house. On Sunday the victim walked back to her own nearby house and told her mother, A.B., what had happened. The victim testified that after what happened in the garage, she was scared of Nava and did not walk past his house anymore. The victim acknowledged she sometimes laughed and smiled when she was nervous or scared. The victim consistently described the incident in the garage during interviews with Sheriff’s Deputy Rod Schulman and social service practitioner Delia Acosta-Perez. A.B. is Nava’s sister. She testified she went to Nava’s house after the victim told her about the incident in the garage. At Nava’s house, Jesse threatened that if A.B. called the police, the victim would be taken from her and A.B. would be deported to Mexico. A.B. had planned to call the police after she had had a chance to speak with her daughter thoroughly about what happened, but the police came to her house before she had a chance to call them. A.B. also testified she had always had a good relationship with Nava before he touched her daughter. Sheriff’s Deputy Lionel Alvarez assisted Schulman in interviewing the victim on July 27, 2011. Schulman testified that at one point during the interview, he asked Alvarez to go get a diagram. When Alvarez stepped out of the room, the victim became very upset and started to cry. The victim told Schulman she was crying because “she was afraid to be in the room with [him], as an older male, because of what her uncle had done to her.” Uncharged Sexual Offenses At trial, the parties and the court focused on one uncharged sexual offense involving allegations by another of Nava’s sisters, M.N., which allegedly occurred in 1969. The record, however, reflects considerable testimony by the victim, both live and via video recording, of three and possibly more uncharged sexual offenses committed by 3. Nava on the victim during the summer of 2011 where Nava touched her in a similar manner. M.N., who was 52 years old at the time of trial, testified that when she was around nine or 10 years old and living with Nava in Mexico, Nava had put his finger inside her vagina, causing it to hurt. She told their mother what happened and it never happened again. Since then, her relationship with Nava had been fine and she never spoke to him about what happened. She could not remember a lot about the incident but recalled it occurred in a bathroom. She previously had told Schulman, however, that the incident occurred in a bedroom. Acosta-Perez interviewed the victim on August 24, 2011. The video recording of the interview was played for the jury. There, the victim, verbally and with hand motions, described how Nava had touched her in the same manner as the charged offense on each of the three days she stayed with him over the weekend of July 22. Those occurred when Jesse would ask that she retrieve things from the garage refrigerator.1 She also testified that one night during the same weekend, Nava touched her in a similar manner while she was sitting on a couch in Nava’s house, after he locked the door to the room.2 Defense Evidence Nava testified and denied the victim’s accusation that he molested her on Saturday morning when she was getting milk from the refrigerator in the garage. According to Nava’s testimony, the only time he and the victim were in the garage together during her 1Thevictim also stated in the interview that “he touched me more, but like I don’t know to count .…” 2Nava did not object to the introduction of the video interview containing the evidence of the other instances of sexual touching in the garage or the testimony regarding the incident on the couch. 4. stay was on the night of Friday, July 22, 2011, when Nava went to get a cake out of the refrigerator. Nava explained that on Friday night he had more than 15 guests at his house for a church celebration. At some point Jesse asked him to bring the cake to her. Nava headed to the garage, followed by his wife. The victim ran up behind them and Nava handed her the cake when he took it out of the refrigerator. He did not touch the victim in any way at this time. Also, they never kept milk in the refrigerator in the garage but kept it in the refrigerator in the kitchen. Around 4:00 a.m. on Saturday morning, when it was still dark, Nava and his wife left the house to drive to Tehachapi to visit their son, who was in prison. Nava did not talk to the victim or say goodbye to anybody before they left. The drive to Tehachapi took approximately one hour 45 minutes. Nava did not see the victim until later that day when he met Jesse in front of a store to help her repair a flat tire. Nava saw the victim briefly on Sunday morning when they were getting ready for church, but they were never alone together that day. The next time he saw the victim was Monday, when he first found out about her accusation, which surprised him. Jesse called Nava while he was working and he came back to the house. When Nava arrived at the house, the victim’s mother was there and he asked her to bring the victim over to the house so the victim’s mother could see what the victim was saying was not true. Nava denied his sister’s allegations that he molested her approximately 43 years earlier. He testified the house in Mexico, where she alleged the molestation occurred, did not have a bathroom. Nava claimed he and his sister had had disagreements over the years. He explained that when his sister came to Los Angeles, she sold drugs and legal documents and they would argue about that. She would get mad and tell him to “butt [out of] her life.” Jesse also testified and described the Friday night church celebration at Nava’s house. During the celebration, Jesse sent Nava to get the cake from the refrigerator in the 5. garage. Jesse saw the victim run after Nava and his wife. The victim then returned carrying the cake, which Jesse took from the victim because it looked like she was going to drop it. The victim slept with Jesse in Jesse’s room on Friday night. On Saturday morning, Nava and his wife went to visit Jesse’s brother in prison, leaving the house around 4:00 a.m. Jesse did not see Nava again that Saturday until around 1:00 p.m. The victim was never alone with Nava on Saturday morning. Jesse was with the victim the whole time, except once when the victim went to change, but that was after Nava and Jesse’s mother had left the house. Jesse did not send the victim to get milk either for her or for a baby on Saturday morning. The only time the victim asked Jesse for milk was while Nava and his wife were away visiting Jesse’s brother in prison. Jesse and the victim were in the living room when the victim asked if she could get milk so she could eat cornflakes. Jesse said yes and the victim went into the kitchen to get it. Jesse learned about the victim’s accusations against Nava on Monday, when the victim’s mother called her. Jesse met the victim and the victim’s mother at Nava’s house. Jesse called Nava to come home. Jesse’s mother also was present. Jesse then asked the victim to show them what had happened to her. The victim brushed her hand down the middle of her chest and below her waist. While she did this, the victim was laughing and continued to laugh until her mother told her it was not a game. Jesse denied threatening to turn the victim’s mother in to immigration. After Nava was arrested, Jesse saw the victim walk past Nava’s house with other people on three separate occasions. As they passed by the house, the victim would turn and laugh. Stephanie Estrada testified she had known Nava for 24 to 25 years. She used to work at a school where she would see him two to three times a month and had had sufficient contacts with him to form an opinion as to his reputation for honesty. 6. According to Estrada, Nava was a very honest individual. He was a God-fearing and loving individual who cared about his community and the people with whom he was acquainted. He had a very excellent reputation for the way he treated women and children. Charges, Trial and Sentence On August 23, 2011, Nava was charged with one count of committing a lewd and lascivious act on a child under the age of 14 on or about July 23, 2011, in violation of Penal Code section 288, subdivision (a). Nava’s jury trial commenced on March 26, 2012, and concluded on March 29. After deliberating for just over an hour, the jury found Nava guilty as charged. On April 27, 2012, Nava was sentenced to six years in prison. DISCUSSION I. Admission of 1969 Uncharged Sexual Offense Nava contends the trial court abused its discretion in admitting the evidence of the 1969 uncharged sexual offense under Evidence Code section 1108. He also claims admission of the evidence violated his constitutional due process rights. Prior to the commencement of trial, the prosecution filed a motion seeking to admit evidence of Nava’s uncharged sexual offense against his sister pursuant to section 1108.3 In anticipation of the prosecution’s motion, the defense filed a separate motion asking the trial court to exclude the evidence under section 352. The trial court conducted a hearing pursuant to section 402, heard argument, and concluded the evidence was admissible. The trial court explained: “In this case, the amount of time between the incident with [Nava’s sister] and the new victim is 43 years. That is very remote in time. However, the victims were the same age, or approximately the same age when the alleged assault took place. They were family members; Mr. Nava had a position of 3All further statutory references are to the Evidence Code. 7.  either trust or authority over the two victims; the acts are substantially similar; the evidence that [Nava’s sister] … testified to is not particularly inflammatory; and there is very little possibility of confusion of the issues. The … testimony that was elicited from [Nava’s sister] took a very short period of time and, therefore, the amount of time involved in introducing and refuting the evidence is not lengthy. So the court—based on viewing the [section] 352 analysis, there is propensity evidence that is highly probative to show intent and far outweighs the prejudicial effect. Court is going to allow the [section] 1108 evidence.” Section 1108 was intended to sweep away the narrow categories of admissibility of other crimes evidence that had existed under section 1101. (People v. Britt (2002) 104 Cal.App.4th 500, 505.) Instead, such evidence is admissible whenever it may be helpful to the trier of fact, on a commonsense basis, for resolution of any issue in the case, including the probability or improbability that the defendant has been falsely accused. (See Britt, at p. 506.) A trial court, however, retains the discretion to admit or exclude evidence of another sexual offense under section 352. (People v. Rodriguez (1999) 20 Cal.4th 1, 9; People v. Callahan (1999) 74 Cal.App.4th 356, 367-368.) A trial court’s exercise of its discretion under section 352 is reviewed for abuse of discretion and will not be disturbed on appeal absent a showing the trial court exercised its discretion in an arbitrary, capricious or patently absurd manner. (Rodriguez, at pp. 9-10.) In People v. Harris (1998) 60 Cal.App.4th 727, the court set out five factors for evaluating the admissibility of prior offense evidence: (1) the inflammatory nature of the evidence; (2) the probability of confusion; (3) the remoteness in time of the prior incidents; (4) the consumption of time involved; and (5) the probative value of the prior offense evidence. (Id. at pp. 737-741.) Here, the trial court found these factors weighed in favor of admission of the evidence. Nava argues the evidence should have been excluded because it was too remote in time, particularly since he had led an unblemished life since the uncharged incident with his sister. But similarities between the charged and uncharged acts can balance out remoteness. (People v. Waples (2000) 79 Cal.App.4th 1389, 1395.) Here, the uncharged 8. and charged offenses were substantially similar. Both victims were young girls, were family members, and were staying in the same house as Nava at the time of the offenses. Both offenses included touching of the victims’ vaginas. The uncharged offense against Nava’s sister, which involved digital penetration of the vagina, was more inflammatory than the charged offense against the victim, but not greatly so. The sister’s testimony did not consume an undue amount of time. And the jury was instructed properly with CALCRIM No. 1191 that they could, but were not required to, consider this evidence for propensity purposes.4 No specific time limit is set forth in the statute and appellate courts have upheld admission of evidence of uncharged offenses that occurred 30 years before. (People v. Branch (2001) 91 Cal.App.4th 274, 285.) Remoteness is but one factor to be considered by the trial court. But 43 years is a long time, and if the sexual misconduct evidence was not similar, we likely would reach a different result. Because it is solely within the trial court’s discretion to determine whether sexual misconduct evidence is too remote, and where the record demonstrates the court wrestled with the issue and exercised its discretion, we will not disturb the court’s ruling. But, even assuming the trial court erred in allowing the introduction of the 43- year-old uncharged and unadjudicated sexual act, based on the evidence of the other three or more recent uncharged acts involving the victim, we conclude any such error not prejudicial, under either the Watson or Chapman standards. (People v. Watson (1956) 46 Cal.2d 818,836; Chapman v. California (1967) 386 U.S. 19, 24.) We can infer from the jury’s guilty verdict that they found the victim credible and not Nava. 4This instruction was not modified to identify the 1969 prior uncharged incident specifically, so the jury also could have considered the evidence of the recent uncharged sexual misconduct involving the victim for propensity purposes. 9.  Nava also claims the probative value of the uncharged sexual offense was “substantially diminished” because his sister’s testimony indicated he was between the ages of 16 and 18 and was, therefore, likely a juvenile at the time of the offense. He argues this is significant because of the language in recent United States Supreme Court cases recognizing that juvenile offenders have lessened culpability for crimes they commit and are less deserving of severe punishment than their adult counterparts. (See Graham v. Florida (2010) 560 U.S. 48; Roper v. Simmons (2005) 543 U.S. 551, 570.) These authorities do not help Nava, however, because they address the issue of culpability, not propensity. It is not unreasonable to conclude that a defendant with a juvenile history of committing sexual abuse would be more likely to commit sexual abuse as an adult than a defendant without such history. Evidence that Nava as a teenager sexually abused his young sister was highly probative on the issue of the likelihood he would as an adult sexually abuse his young niece. In short, there is no indication here that the trial court exercised its discretion in an arbitrary, capricious, or patently absurd manner. We thus conclude the trial court did not abuse its discretion in admitting evidence of Nava’s prior sexual offense under section 1108. Nava also has forfeited his challenge to admissibility of the evidence on constitutional due process grounds. “An appellate contention that the erroneous admission or exclusion of evidence violated a constitutional right is not preserved in the absence of an objection on that ground below. [Citations.]” (People v. Daniels (2009) 176 Cal.App.4th 304, 320, fn. 10.) No objection on the basis of a violation of constitutional rights was made by Nava. II. Exclusion of Proffered Testimony Nava contends the trial court erred in excluding as cumulative the proffered testimony of Luis Hernandez that he saw Nava and the victim getting a cake out of the refrigerator in the garage and nothing happened. 10.  The issue of Hernandez’s proposed testimony came up during a discussion of additional witnesses the defense planned to call: “THE COURT: [Hernandez is] going to testify about the cake, refrigerator incident? “[DEFENSE COUNSEL]: Yeah, about being in that area and seeing … him and [the victim], and that nothing happened. “THE COURT: I’m going to preclude you from calling Luis Hernandez as cumulative, undue consumption of time. I think it’s pretty— it’s been pretty well settled that both [the victim] and Mr. Nava got cake from the refrigerator and nothing happened. I’ll bet [the prosecutor] would even stipulate to that. “[THE PROSECUTOR]: I don’t know if I’d go quite that far, but— “[DEFENSE COUNSEL]: Well, your Honor, I think it’s the Government’s theory that that’s when the touching happened. “THE COURT: That is not. “[DEFENSE COUNSEL]: They’re going to stipulate to that? “THE COURT: It’s clear that the Government’s theory is when she went to get the milk is when the touching happened, not the cake. So it’s different. “[DEFENSE COUNSEL]: Well, can we have that stipulation on the record? “THE COURT: No, I’m just telling you that from the testimony, it is clear. You’ve had two witnesses now testify that [the victim] and the defendant went to the refrigerator to get the cake and there were people around and nothing happened. And actually, there’s three people that testified. Well, [the victim] testified that it was during the milk incident. “[THE PROSECUTOR]: And just for the record, Luis Hernandez’s testimony is going to be that he wasn’t—he was standing in the yard. He couldn’t even see the garage. So for the record, again, for appeal purposes, that’s what his testimony would be. That’s not relevant. 11.  “THE COURT: My ruling is that it would be cumulative, assuming that he would testify that he witnessed [the] getting of the cake from the refrigerator.” Section 352 gives “the court discretion to ‘exclude evidence if its probative value is substantially outweighed by the probability that its admission will (a) necessitate undue consumption of time or (b) create substantial danger of undue prejudice, of confusing the issues, or of misleading the jury.’ [Citation.]” (People v. Brown (2000) 77 Cal.App.4th 1324, 1337.) Here, the trial court excluded Hernandez’s proffered testimony as an undue consumption of time and cumulative of other evidence. This was a proper exercise of its discretion. The testimony had limited probative value. As the trial court observed in making its ruling, based on the evidence presented, the prosecution’s theory was that Nava molested the victim when she went into the garage in the morning to get milk, not when she followed Nava into the garage in the evening to get the cake. Hernandez’s proffered testimony tended to show Nava was telling the truth in testifying he did not molest the victim in the garage during the Friday evening church celebration when others were present at his house. It did not, however, tend to prove this was the only time Nava had gone into the garage with the victim or that the separate incident she described did not occur. We therefore disagree with Nava’s assertion that in excluding Hernandez’s testimony, the trial court prevented Nava from presenting “strong evidence” he was “innocent of the charged crime.” Contrary to Nava’s assertions, the prosecution did not suggest to the jury that it could find the charged offense occurred during the Friday evening celebration when Nava and the victim took the cake out of the refrigerator in the garage. Despite the prosecutor’s expressed reluctance to stipulate that nothing happened, she presented no evidence or argument suggesting Nava molested the victim during the cake incident or even disputing the incident occurred. Although, as Nava points out, the prosecutor did state in her opening argument that the charged offense occurred at night, she corrected 12. herself at the beginning of closing argument and explained to the jury she misspoke in opening argument. Nava also claims the prosecutor argued the alleged touching could have happened at any time over the weekend. Not so. On the pages of the reporter’s transcript Nava cites to support his assertion, the prosecutor argued that the alleged molestation could have “happened Friday morning, Saturday morning, or Sunday morning. You don’t have to agree on which day it happened, you just have to agree it happened on or about Saturday morning.” The trial court’s conclusion the proffered evidence was cumulative of other evidence also was reasonable. Both Nava and Jesse presented undisputed testimony that a number of guests attended the Friday evening church celebration at Nava’s house when he went to get the cake from the refrigerator in the garage. Jesse also testified she saw Nava’s wife go with him into the garage, followed by the victim, who reappeared shortly carrying the cake. Although Jesse did not testify she could see what was happening inside the garage after the victim followed Nava and his wife inside, Jesse’s testimony corroborated Nava’s testimony that his wife was present in the garage when he took the cake out of the refrigerator and gave it to the victim and was circumstantial evidence Nava did not molest the victim at that time. On the record before us, we cannot agree with Nava’s argument that Hernandez’s proffered testimony added significant details to the testimony already presented or was crucial to his defense. III. Exclusion of Evidence Relevant to the Victim’s Credibility Nava contends the trial court abused its discretion in excluding evidence he claims was relevant to the issue of the victim’s credibility. He also claims the exclusion of the evidence violated his constitutional rights to present evidence in his own defense and to confront and cross-examine witnesses. 13.  Specifically, Nava argues the trial court erred in excluding evidence of (1) four accusations of molestation “against other people by other family members” of the victim’s family; (2) Nava’s statement to Schulman that he maintained a certain distance from the victim because she gets angry and makes accusations, like accusing her neighbor, a young man, of “touch[ing] her butt,” and that Nava told the victim’s mother to watch over the victim closely; (3) “bad blood” between Nava and the victim’s mother; and (4) molestation of the victim’s older sister by Nava’s “brother, Silvano B.”5 We agree with the People that Nava forfeited most of his claims by failing to seek appropriate evidentiary rulings, and, to the extent his claims have been preserved for appellate review, he has failed to demonstrate an abuse of discretion by the trial court. Standard of Review A trial court’s rulings on the admissibility of evidence, including under section 352, are reviewed for abuse of discretion. (People v. Scott (2011) 52 Cal.4th 452, 490- 491 (Scott).) We will not set aside a judgment by reason of the erroneous exclusion of evidence unless the error resulted in a miscarriage of justice. (§ 354.) An error has resulted in a miscarriage of justice and warrants reversal “if in the absence of the error, the appealing party would have probably obtained a more favorable result. (See Cal. Const., art. VI, § 13; Code Civ. Proc., § 475; Evid. Code, § 354; [citations].)” (Greenspan v. LADT LLC (2010) 191 Cal.App.4th 486, 526-527.) “‘As a general matter, the “[a]pplication of the ordinary rules of evidence … does not impermissibly infringe on a defendant’s right to present a defense.” [Citations.]’ [Citation.]” (People v. McNeal (2009) 46 Cal.4th 1183, 1203.) Where the trial court 5Nava’s fourth claim of evidentiary error refers to evidence that Nava’s brother-in- law (not his brother) Silvano B. was charged in 1999 with molesting the victim’s older sister and the charges later were dropped. This is discussed in Nava’s petition for writ of habeas corpus filed approximately nine months after the appeal. (See part VII. of the DISCUSSION, post.) 14. does not preclude the defendant from presenting a defense but rejects some evidence concerning that defense, the error is reversible only if it has resulted in a miscarriage of justice. (Ibid.) With certain statutory exceptions, all relevant evidence is admissible. (§ 351.) “‘Relevant evidence’ means evidence, including evidence relevant to the credibility of a witness …, having any tendency in reason to prove or disprove any disputed fact that is of consequence to the determination of the action.” (§ 210.) The test is whether the evidence tends “‘“‘“‘“logically, naturally, and by reasonable inference” to establish material facts .… [Citations.]’ [Citation.]”’”’” (People v. Fields (2009) 175 Cal.App.4th 1001, 1016.) Circumstantial evidence is evidence from which a fact may be inferred. (People v. Nealy (1991) 228 Cal.App.3d 447, 451.) “An inference is a deduction of fact that may logically and reasonably be drawn from another fact or group of facts found or otherwise established in the action.” (§ 600, subd. (b).) The trial court has broad discretion to determine whether evidence, including circumstantial evidence, is relevant. (Scott, supra, 52 Cal.4th at p. 490; People v. Harris (2005) 37 Cal.4th 310, 337-338.) However, even relevant evidence may be excluded if its probative value is substantially outweighed by concerns of undue prejudice. (§ 352.) “Generally, in order to preserve any evidentiary point for appellate review, the proponent of the evidence must make an offer of proof regarding the anticipated testimony or ask questions which presage the expected response. [Citations.] Failure to make an adequate offer of proof precludes consideration of the alleged error on appeal. [Citation.]” (People v. Eid (1994) 31 Cal.App.4th 114, 126.) “An offer of proof should give the trial court an opportunity to change or clarify its ruling and in the event of appeal would provide the reviewing court with the means of determining error and assessing prejudice. [Citation.] To accomplish these purposes an offer of proof must be specific. It must set forth the actual evidence to be produced and not merely the facts or issues to be addressed and argued. [Citations.]” (People v. Schmies (1996) 44 Cal.App.4th 38, 53 15. (Schmies).) “An offer of proof must consist of material that is admissible, it must be specific in indicating the purpose of the testimony, the name of the witness and the content of the answer to be elicited.” (Semsch v. Henry Mayo Newhall Memorial Hospital (1985) 171 Cal.App.3d 162, 167.) Four Accusations of Molestation by Members of the Victim’s Family Nava’s failure to make a specific offer of proof forfeited his claim that the trial court erred in excluding evidence of four accusations of molestation by members of the victim’s family against other people. To counter the forfeiture argument, Nava correctly points out the prosecution preemptively sought to exclude such evidence in one of its in limine motions. When the prosecution asked the court to rule on the motion, however, defense counsel stated, “I think I can probably shortcut that because I recognize that I don’t have—although I’ve shared with the prosecutor the names of people and some other incidents, I don’t have any [of] these people available to testify.” When the issue arose during trial, defense counsel again indicated he was not planning to go into the area of the four prior accusations of molestation. Thereafter, the trial court indicated it agreed with the prosecutor’s argument that such evidence was irrelevant and unduly prejudicial as its only apparent purpose was to show that “this victim’s family is out to just accuse people.” Defense counsel did not object to this characterization of the evidence or attempt to make a showing that the evidence was relevant to any of the issues at trial. In the absence of a specific offer of proof, the trial court acted well within its discretion in excluding the evidence as irrelevant. Nava’s Statement to Schulman Nava has similarly forfeited his claim concerning the exclusion of his statement he made during his interview with Schulman regarding the victim’s claim that a young man in the neighborhood “touch[ed] her butt.” 16.  Nava argues the evidence was relevant because “it shows that [the victim] was familiar with sexual touching and had a basis and experience upon which to fabricate testimony.” (See People v. Daggett (1990) 225 Cal.App.3d 751, 757.) Nava, however, did not offer the evidence on this ground in the trial court. Rather, it was the prosecutor who brought the evidence of Nava’s statement to the court’s attention and pointed out she did not see how it was relevant.6 In response, defense counsel stated he thought the evidence had “some probative value” because Nava’s statement was given in response to the deputy’s question asking Nava why a 10-year-old girl would lie, which “tends to be a major question in a case like this.” The trial court deferred a ruling on the admissibility of the statement but indicated it disagreed with counsel’s relevancy argument, pointing out that for Nava to say “she’s accused other people, she may have been molested before” did not really answer the deputy’s question of why the victim would lie in this instance. It does not appear the court thereafter made a specific evidentiary ruling regarding Nava’s statement to Schulman. On the merits, assuming the trial court’s statements can be interpreted as a ruling to exclude Nava’s statement to Schulman, Nava has not demonstrated the trial court abused its discretion in excluding this evidence as irrelevant. Nava exaggerates his proffered statement by suggesting it established the victim had been sexually molested in 6The prosecutor read the relevant portion of the interview transcript as follows: “[Schulman:] ‘Why would a ten-year-old girl make an accusation against you? Who’s the family member if you did not do it?’ [¶] The defendant: ‘I’m going to tell you. When that little girl gets angry she says a lot of things from her house. She has said a lot of things and she has talked to—talked to about another young man who lives next to the house, that the young man would touch her—her butt area. And I told my sister, take care of that little girl because they want to touch her, and she did not listen. But for me to touch someone, no one.’” 17. the past. Nava merely reported that the victim claimed a young man touched her “butt area.” His statement disclosed nothing about the circumstances or the extent of the alleged touching. The minimal information presented by Nava’s statement, which may very well have been the extent of any possible testimony on the subject, was very different than the allegations the victim made against Nava of squeezing her breasts and vagina underneath her clothes. Because the jury would have to speculate to reach the conclusions asserted by Nava, the evidence was properly excluded. Bad Blood between Nava and the Victim’s Mother During Nava’s direct examination defense counsel asked, “[W]as one of your sisters angry at you before this accusation came out?” The prosecutor objected on relevancy grounds, and the trial court called for the jurors to take a break. Outside the presence of the jury, the trial court noted that it looked like defense counsel was “going to go into an area where—to establish some sort of bad blood between the victim’s mother and the defendant; is that correct?” Defense counsel responded, “I understood there was some bad blood between him and his sister … and she’d come in and said there was none.” The court asked what the relevance was and defense counsel stated: “Well, what if the mother has—is encouraging the daughter to lie about this? That’s the relevance.” The trial court replied: “There’s no evidence that if there was—assuming for the purpose of this discussion that there was bad blood between … the victim’s mother and the defendant, unless the victim knew about these problems, this conflict, then it’s not relevant. And there’s no evidence of any coaching or anything like that. So this area is not relevant. It’s—it is prejudicial and it’s undue consumption of time.” In response, defense counsel stated: “Well, to save time, I’m not going to disagree with you. I mean I disagree with you, but I’m not going to try to go into that area any further because what I’m really fishing for is why would [the victim] say this if it wasn’t true.” 18.  Nava argues that because the victim was only 10 years old at the time of the alleged molestation, we should infer that the victim’s mother likely was the person who instigated the criminal investigation. Therefore, evidence of the mother’s bias against Nava was relevant and should have been admitted. (See People v. Haxby (1962) 204 Cal.App.2d 791, 792, 795-796 [trial court erred in refusing to allow defense counsel to show hostile relationship existed between defendant and his next-door neighbors, who instigated investigation of defendant’s conduct that led to his conviction of committing lewd acts on his daughter and a neighbor girl].) But there was no evidence here the victim’s mother instigated the investigation into Nava’s conduct. Rather, there was evidence to the contrary in that the victim’s mother testified she was planning to call the police, but they came to her house before she had a chance to call them. We do not think the trial court abused its discretion in excluding the evidence. The evidence had minimal probative value. As the court noted, even assuming there was some kind of bad blood or hostility between Nava and the victim’s mother, it did not tend to prove the victim lied about Nava molesting her, absent evidence the victim was aware of the hostility or that her mother influenced her somehow in making her allegation against Nava. Victim’s Older Sister Was Molested Nava contends the trial court erred in excluding as irrelevant evidence that the victim’s older sister was molested by Nava’s brother because this was “direct evidence of a potential bias between not only the families, but also [the victim], if she had been informed of the conduct.” We disagree. Like the alleged bad blood between Nava and the victim’s mother, evidence the victim’s family was biased against Nava, based on his brother’s alleged molestation of the victim’s older sister, was irrelevant. There was no evidence the victim was aware of the molestation or that the victim’s family somehow influenced the victim in bringing her 19. accusation against Nava. Nor was there any evidence the victim’s family instigated the criminal investigation against Nava. The trial court did not err. Constitutional Challenge Nava forfeited his constitutional challenge by failing to raise in the trial court his claim that the exclusion of the evidence violated his constitutional rights to present a defense and confront witnesses. IV. Admission of Victim’s Hearsay Statement Nava claims the trial court erred when it admitted, over his hearsay objection, Schulman’s testimony that the victim said she was crying because “she was afraid to be in the room with [him], as an older male, because of what her uncle had done to her.” After holding a sidebar conference, the trial court allowed the evidence to be admitted under the exception to the hearsay rule contained in section 1360. Nava now contends section 1360 was inapplicable because the victim’s statement to Schulman was not “describing any act of child abuse … or describing any attempted act of child abuse .…” (§ 1360, subd. (a).) The rationale used by the trial court for admitting the evidence, however, is not a matter for this court’s review. (Davey v. Southern Pacific Co. (1897) 116 Cal. 325, 329.) It is judicial action and not judicial reasoning that is the proper subject of appellate review. (El Centro Grain Co. v. Bank of Italy, etc. (1932) 123 Cal.App. 564, 567.) As the victim’s statement was admissible under at least one exception to the hearsay rule, the trial court’s admission of the evidence must be upheld. (People v. Brown (2004) 33 Cal.4th 892, 901 [“If a judgment rests on admissible evidence it will not be reversed because the trial court admitted that evidence upon a different theory, a mistaken theory, or one not raised below”].) Section 1250 permits admission of “evidence of a statement of the declarant’s then existing state of mind, emotion, or physical sensation (including a statement of intent, plan, motive, design, mental feeling, pain, or bodily health) … when: [¶] (1) The 20. evidence is offered to prove the declarant’s state of mind, emotion, or physical sensation at that time or at any other time when it is itself an issue in the action; or [¶] (2) The evidence is offered to prove or explain acts or conduct of the declarant.” (Id., subd. (a).) “‘“[A] victim’s out-of-court statements of fear of an accused are admissible under section 1250 only when the victim’s conduct in conformity with that fear is in dispute. Absent such dispute, the statements are irrelevant. [Citations.]”’ [Citation.]” (People v. Jablonski (2006) 37 Cal.4th 774, 819 (Jablonski).) Nava placed the victim’s state of mind at issue. He asserted the victim fabricated her claims of molestation against him as a means of gaining attention. He also maintained her conduct in laughing and smiling when describing the molestation was inconsistent with the behavior one would expect to see in a sexual abuse victim who feared her abuser or was traumatized as a result of the abuse. (Jablonski, supra, 37 Cal.4th at p. 820 [victim’s fear of accused may be relevant where, according to defendant, victim behaved in manner inconsistent with that fear].)7 Evidence that the victim cried and said she feared being in the room with Schulman, as an older man, after what her uncle did to her, implicitly expressed the victim’s fear of Nava. It was relevant to dispute the defense theory that her behavior was 7For example, defense counsel argued: “Did you notice that [the victim] was depressed or quiet? Common sense tells us that’s what would happen if somebody’s traumatized or victimized like she said. But instead, you see the exact opposite. You see the laughing and smiling and in an effort to short that up, they try to bring up, well, she cried once during the interview and said she was afraid. I think the candid capture of this young girl on the video [of her interview with social services practitioner Acosta-Perez] is more revealing of what’s going on with her. She likes this moment. She’s finally getting what she wants. She’s getting more attention from her mom who has left her with this family. She’s getting the attention she finally wants. She’s getting attention from the lady on the video. She loves it. She basks in it.” A little later defense counsel argued: “So does it matter that her story’s consistent or does it matter that her behavior is inconsistent with the story? There are things that don’t fit.” 21. inconsistent with someone who had been sexually abused and indicated she fabricated her accusations because she wanted (and enjoyed) the resulting attention. As Schulman’s testimony was admissible under section 1250, the trial court did not err in allowing it. V. Prosecutorial Misconduct Nava argues the prosecutor committed misconduct by using, in cross-examination and closing argument, “reprehensible and deceptive means” to portray Nava “as an extraordinarily arrogant man with a family of criminals.” We disagree. During cross-examination, the prosecutor questioned Nava about his interview with Schulman, in relevant part, as follows: “Q. Okay. And then when you told the officer that your son, who you went to go visit in prison, was being falsely imprisoned, you told the officer that people lied there and that’s why he was being falsely imprisoned? “[DEFENSE COUNSEL]: Objection, irrelevant. “THE COURT: Sustained. “[DEFENSE COUNSEL]: Move to strike. “THE COURT: Stricken. “[DEFENSE COUNSEL]: The question and the answer. “[THE PROSECTUOR]: “Q. Is your son in prison? [¶] … [¶] “[DEFENDANT]: Yes. “[THE PROSECUTOR]: Is he being illegally incarcerated, in your opinion? “[DEFENSE COUNSEL]: Objection irrelevant. “THE COURT: Sustained. Sustained. You don’t answer that, sir. Don’t answer that. [¶] Go ahead. “[THE PROSECUTOR]: 22.  “Q. Is it always somebody else’s fault, Mr. Nava? “[DEFENSE COUNSEL]: Objection, argumentative. “THE COURT: Sustained.” In closing argument, the prosecutor stated: “Another one of his responses was asked about making mistakes. This is an arrogant man, that’s what the defendant is. And when asked if you ever have made a mistake, his response is, no, he’s never made a mistake. And talking the deputy, what about a traffic accident? [Sic.] Well, yeah, but it wasn’t my fault. It’s all somebody else’s fault with this defendant. It’s never his fault.” No objection was made to these statements. “‘“The applicable federal and state standards regarding prosecutorial misconduct are well established. ‘“A prosecutor’s … intemperate behavior violates the federal Constitution when it comprises a pattern of conduct ‘so egregious that it infects the trial with such unfairness as to make the conviction a denial of due process.’”’ [Citations.] Conduct by a prosecutor that does not render a criminal trial fundamentally unfair is prosecutorial misconduct under state law only if it involves ‘“‘the use of deceptive or reprehensible methods to attempt to persuade either the court or the jury.’”’ [Citation.]” [Citation.]’ [Citation.] [¶] Regarding the scope of permissible prosecutorial argument, ‘“‘a prosecutor is given wide latitude during argument. The argument may be vigorous as long as it amounts to fair comment on the evidence, which can include reasonable inferences, or deductions to be drawn therefrom. [Citations.] It is also clear that counsel during summation may state matters not in evidence, but which are common knowledge or are illustrations drawn from common experience, history or literature.’ [Citation.] ‘A prosecutor may “vigorously argue his case and is not limited to ‘Chesterfieldian politeness’” [citation], and he may “use appropriate epithets .…”’” [Citation.]’ [Citation.] [¶] Finally, ‘a defendant may not complain on appeal of prosecutorial misconduct unless in a timely fashion—and on the same ground—the defendant made an 23. assignment of misconduct and requested that the jury be admonished to disregard the impropriety. [Citation.]’ [Citation.]” (People v. Stanley (2006) 39 Cal.4th 913, 951- 952.) Any claim of prosecutorial misconduct has been forfeited because Nava did not object on that basis and did not request the jury be admonished to disregard any impropriety. Nava concedes no objection on the ground of prosecutorial misconduct or request for admonishment was made in the trial court but contends the issue is reviewable on appeal because an admonishment would not have cured the harm. He fails, however, to provide a meaningful explanation to support this contention. He simply asserts the challenged cross-examination questions were “so argumentative in nature and so aimed at developing an irrelevant character attack of [him] and his family that an admonition would not have cured the harm.” Merely asserting that the misconduct is “incurable” is insufficient. (People v. Foster (2010) 50 Cal.4th 1301, 1354 [California Supreme Court rejected defendant’s assertion he did not forfeit claim of prosecutorial misconduct by his failure to object and seek curative admonition where defendant did not explain why a curative admonition would not have cured any harm].) The circumstance of Nava’s son being in prison was already before the jury through defense evidence that Nava was visiting him there at the time the victim claimed she was molested. Therefore, the prosecutor’s referring to Nava’s son being in prison did not render the questions incurably inflammatory. We see no support in the record for Nava’s suggestion that the questions were aimed at showing he had “a family of criminals” or a son who was “a serious criminal.” Regardless, even assuming Nava preserved his claim of prosecutorial misconduct for appellate review, it fails on the merits for two reasons. First, there is no indication the jury construed or applied the prosecutor’s cross-examination questions in a manner contrary to the trial court’s instructions. The court sustained defense counsel’s objections to the complained-of questions on the grounds they were irrelevant and argumentative. 24. The court also granted defense counsel’s motion to strike the prosecutor’s question asking Nava whether he told Schulman his son was falsely imprisoned and that his son was in prison because people had lied. A jury is presumed to follow a court’s instruction in the absence of any indication that it was unwilling or unable to do so. (People v. Letner and Tobin (2010) 50 Cal.4th 99, 196.) The jurors were instructed: “During the trial, the attorneys may have objected to questions or moved to strike answers given by the witnesses. I ruled on the objections according to the law. If I sustained an objection, you must ignore the question. If the witness was not permitted to answer, do not guess what the answer might have been or why I ruled as I did. If I ordered testimony stricken from the record you must disregard it and must not consider that testimony for any purpose.” Here, there is no evidence in the record, and Nava has cited none, indicating the jury was unwilling or unable to follow these instructions. Second, it is clear the challenged remarks in closing argument were not objectionable. The prosecutor’s argument did not, as Nava suggests, refer back to the prosecutor’s objectionable questions during cross-examination. Rather, the prosecutor’s argument refers to questions preceding those. In cross-examination, Nava acknowledged he told Schulman he never made any mistakes, testifying, “No, I have never made mistakes.” Nava also acknowledged denying he was at fault when Schulman challenged his assertion about never making mistakes by observing Nava had hit someone in a car accident. Thus, Nava testified, “Yes, but it’s never been my fault. Never.” In light of Nava’s testimony professing never to make any mistakes, the prosecutor’s remark that Nava was an arrogant man constituted fair comment on the evidence and did not amount to misconduct. Even if we were to view the prosecutor’s remarks as improper, the jurors were instructed that comments from counsel were not evidence and they were to decide the case based upon the evidence, not bias, sympathy, 25. prejudice, or public opinion. Again, we see no evidence in the record that the jurors were unable or unwilling to follow these instructions. VI. Exclusion of Character Evidence Nava contends the trial court erred in excluding evidence of “specific instances of good conduct around young girls” to rebut the prosecution’s section 1108 evidence. We agree with the People that Nava’s failure to make a specific offer of proof regarding such evidence forfeited his claim on appeal. To avoid forfeiture Nava argues it would have been futile for him to make a specific offer of proof because the trial court made a “blanket ruling” excluding evidence of specific instances of good conduct. (§ 354, subd. (b); see Schmies, supra, 44 Cal.App.4th at p. 54, fn. 9.) Nava is correct that “[w]here an entire class of evidence has been declared inadmissible or the trial court has clearly intimated it will receive no evidence of a particular class or upon a particular issue, an offer of proof is not a prerequisite to raising the question on appeal, and an offer, if made, may be broad and general. [Citations.]” (Beneficial etc. Ins. Co. v. Kurt Hitke & Co. (1956) 46 Cal.2d 517, 522.) “An offer [or a more specific offer] under those circumstances would be an idle gesture.” (Caminetti v. Pacific Mut. Life Ins. Co. (1943) 23 Cal.2d 94, 100.) Such circumstances did not exist here. Nava failed to make an offer of proof before there was any indication by the trial court that a specific offer of proof would be futile. Instead, defense counsel expressly indicated he was not planning to introduce evidence of specific instances of good conduct and thus made no offer of proof before the trial court made its so-called blanket ruling to exclude such evidence. Moreover, throughout the trial, the trial court showed itself to be willing to revisit evidentiary issues addressed during the pretrial hearing based on evidence that had been presented. There is no indication the trial court would not have given meaningful 26. consideration to a specific offer of proof if counsel had come to view evidence of specific instances of good conduct as relevant to refute the prosecution’s section 1108 evidence. VII. Ineffective Assistance of Counsel Nava has filed a separate petition for writ of habeas corpus claiming his defense counsel rendered ineffective assistance at trial by (1) failing to obtain prison records corroborating his alibi defense; (2) neglecting to present good character evidence to rebut the prosecution’s section 1108 evidence; and (3) not making a specific relevance proffer about a prior accusation of molestation by the victim’s sister against his brother-in-law.8 Standard of Review The defendant has the burden of proving ineffective assistance of trial counsel. To prevail on a claim of ineffective assistance of trial counsel, the defendant must establish not only deficient performance, which is performance below an objective standard of reasonableness, but also prejudice. A court must indulge a strong presumption that counsel’s conduct falls within the wide range of reasonable professional assistance. Tactical errors generally are not deemed reversible. (People v. Maury (2003) 30 Cal.4th 342, 389 (Maury).) Counsel’s decisionmaking is evaluated in the context of the available facts. To the extent the record fails to disclose why counsel acted or failed to act in the manner challenged, appellate courts will affirm the judgment unless counsel was asked for an explanation and failed to provide one, or, unless there simply could be no satisfactory explanation. (Maury, supra, 30 Cal.4th at p. 389.) Prejudice must be affirmatively proved. The record must affirmatively demonstrate a reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different. (Maury, supra, 30 Cal.4th at p. 389.) 8By order filed June 5, 2014, we consolidated the appeal with the writ petition. 27. Attorneys are not expected to engage in tactics or to file motions that are futile. (Id. at p. 390; see also People v. Mendoza (2000) 24 Cal.4th 130, 166.) Analysis Nava’s contention fails for three reasons. First, he contends defense counsel rendered ineffective assistance by failing to obtain “easily obtainable” prison visitation records reflecting Nava and his wife visited their son at the prison in Tehachapi at 7:35 a.m. on Saturday, July 23, 2011, thereby corroborating Nava’s alibi defense. The record, however, discloses a reasonable explanation for counsel’s alleged omission. In a letter responding to questions posed by Nava’s retained appellate counsel, defense counsel explained he verified the prison visit with Nava, Nava’s wife, and Jesse. “The victim however testified that this happened early before he departed to visit the son.” Defense counsel further noted Nava’s wife “could have been a good witness but had an emotional meltdown suddenly just before I wanted to call her to the stand and had to be taken to the hospital. Later at sentencing she was back to normal.” On cross-examination, defense counsel elicited testimony from the victim that, to the best of her recollection, the molestation occurred early in the morning on Saturday, July 23, 2011, after Jesse woke her up to get milk from the refrigerator in the garage. Defense counsel’s clarification of the timing of the alleged incident demonstrates why presenting further corroborating evidence of Nava’s alibi would not have been particularly helpful to the defense and therefore counsel was not deficient for failing to develop it. The prosecution did not dispute that Nava and his wife visited their son in prison but posited the theory that if the molestation occurred Saturday morning as the victim recalled, then it must have occurred before Nava and his wife left to visit their son.9 9Forexample, the prosecutor argued: “We heard that the defendant left at 4:00 a.m. If Jesse’s got a baby in that house that early in the morning, doesn’t it make sense that she asked for milk? Doesn’t it make sense [the victim] was already up?... So if the 28.  These circumstances also demonstrate the absence of prejudice. As the prosecution did not dispute the prison trip occurred, Nava’s theory that the jury in this case would have expected him to present prison records to prove the trip and likely rejected his alibi defense based on his failure to do so fails. Second, Nava complains defense counsel was ineffective because “[d]espite being provided over 35 character letters, trial counsel did not request testimony from any number of young girls that were ready and willing to testify that either on family vacations, overnights, or dance rehearsals, despite being alone with [Nava], they were always treated with respect.” Again, the record reveals a satisfactory explanation for counsel’s alleged omission. In his letter to Nava’s appellate counsel, defense counsel explained that “[f]rom a tactical standpoint given the nature of the charges, where it happened, and his active involvement in cultural dancing with young girls I did not feel that would help.” In a case alleging the molestation of a young girl, it was not unreasonable for counsel to decide not to seek the admission of evidence calling the jury’s attention to Nava’s taking an interest in and participating in numerous activities bringing him in close proximity to young girls. Nava also claims defense counsel failed to call Arcelia Valdez as a character witness. She provided a declaration in which she described her long acquaintance and frequent contact with Nava. The record, however, does not disclose whether defense counsel even was aware of this witness. Finally, Nava contends defense counsel was ineffective for failing to investigate charges that were brought, and then later dropped, against Nava’s brother-in-law, Silvano B., for committing lewd acts on the victim’s sister in 1999. defendant’s up at 4:00 a.m. getting ready [to] go to Tehachapi to see his son in prison, then [the victim] was awake. So you can believe it happened Saturday morning because they were both there. There’s an opportunity there. Both there, they’re both awake, they’re moving around the house .…” 29.  Nava asserts that “evidence that there was a prior fabricated charge, or that there was an actual molestation, would both have been highly relevant” to his defense. This is so, he argues, because if the victim’s mother “encouraged her older daughter to lie, this would have significantly called into question her credibility.” On the other hand, if his brother-in-law molested the victim’s sister, this evidence would be relevant to show that the victim “was familiar with certain forms of sexual assault, since her older sister had experienced what, judging by the complaint, appears to be a similar type of sexual assault at a similar age.” This last claim of ineffective assistance of counsel fails because Nava cannot establish prejudice. His relevancy arguments are based entirely on speculation. They assume the victim’s mother either encouraged the victim to lie or the victim was familiar with the acts of sexual abuse allegedly inflicted on her sister over a decade before the alleged conduct in this case. Nava presents no evidence in his writ petition, and no evidence was presented at trial, suggesting the victim’s mother encouraged or coached the victim to fabricate her allegations against Nava, or that the victim was familiar with any prior allegations of sexual abuse against the victim’s sister. Thus, Nava has failed to establish that investigation into charges of sexual abuse by his brother-in-law against the victim’s sister would have yielded relevant evidence affecting the outcome of the proceedings. 30.  DISPOSITION The judgment is affirmed. The petition for a writ of habeas corpus is denied. _____________________ CORNELL, Acting P.J. WE CONCUR: _____________________ DETJEN, J. _____________________ FRANSON, J. 31. 

As one of the nation's leading academic research centers, the University of Pittsburgh has both an[unreadable] opportunity and an obligation to take the inherent risks associated with reengineering a successful research[unreadable] enterprise to undertake a transformative initiative that will result in the development and advancement of[unreadable] clinical and translational science as a distinct discipline in western Pennsylvania. The University is committed[unreadable] to transforming its culture, environment, and structure to achieve this goal by forming the Clinical and[unreadable] Translational Science Institute (CTSI). The CTSI will serve as the integrative academic home for clinical and[unreadable] translational scientists across the University's six health sciences schools; Carnegie Mellon University; the[unreadable] University of Pittsburgh Medical Center (UPMC), one of the nation's largest and most financially successful[unreadable] academic health care systems; and the region. The CTSI's primary focus is to develop, nurture, and support[unreadable] a cadre of clinical and translational scientists by building on the University's existing clinical research training[unreadable] programs (Roadmap K12, K30) to establish a comprehensive program with activities ranging from early[unreadable] research exposure for high school students to advanced doctoral programs. Through "integration and[unreadable] innovation," the CTSI will excel in the development of new biomedical knowledge and the translation of that[unreadable] knowledge from the basic and preclinical research settings to individuals, communities, and health practice.[unreadable] The Children's Hospital of Pittsburgh's General Clinical Research Center (GCRC) and the four sites of the[unreadable] University of Pittsburgh GCRC will be reengineered, integrated, and augmented by new CTSI communitybased[unreadable] and minority health focused centers to develop efficient, accessible, and widely used participant and[unreadable] clinical interaction resources. The CTSI Center for Clinical and Translational Informatics, which is developing[unreadable] translational research informatics tools for the NCI Cancer Biomedical Informatics Grid Initiative, will infuse[unreadable] informatics tools into the entire lifecycle of clinical research studies and develop an online collaborative[unreadable] research community. Innovative interdisciplinary research initiatives will be developed through the ten CTSI[unreadable] resource cores and translated to health practice via a novel CTSI community partnership program and[unreadable] through centralization of UPMC's extensive clinical networks. The resulting transformations in the institution,[unreadable] scientist, research, and health practice will improve health locally, regionally, and nationally.

Cytochrome P450 (CYP) enzymes are involved in numerous detoxification and synthetic processes in addition to generating potent lipid mediators from endogenous substrates. Even though many CYP isozymes can oxidize a spectrum of ω-6 and ω-3 polyunsaturated fatty acids such as retinoic acid, linoleic acid, eicosapentaenoic acid (EPA), and docosahexenoic acid (DHA), they are often referred to as the third pathway of arachidonic acid metabolism, mainly because the most is known about the biological actions of these products ([@bib28]). Angiogenesis is a tightly regulated and organized process, and although numerous studies have addressed the role of specific proteins at the different stages of vascular development ([@bib44]), the role of lipids is less clear. Although cyclooxygenases and prostaglandins have been the focus of some studies ([@bib48]), little is known about the role of CYP-derived lipid mediators. The first link between CYP enzymes and angiogenesis was obtained in co-cultures of astrocytes and endothelial cells in which arachidonic acid epoxides (epoxyeicosatrienoic acids \[EETs\]) released from astrocytes increased endothelial cell proliferation and elicited the formation of capillary-like structures ([@bib35]; [@bib65]). Also, the overexpression of the CYP2C9 epoxygenase in, and/or the application of 11,12- or 14,15-EET to monocultures of endothelial cells was associated with angiogenesis ([@bib32]; [@bib33]). In vivo data rapidly followed to support these in vitro findings, with EETs reported to induce angiogenesis in the chick chorioallantoic membrane ([@bib33]) and the vascularization of Matrigel plugs implanted into wild-type mice ([@bib32]; [@bib60]). Moreover, the overexpression of the human CYP2C11 and 2J2 enzymes in the ischemic rat hindlimb model was found to increase muscle capillary density ([@bib57]). However, such studies could not address the importance of endogenously generated CYP metabolites and were difficult to back up in knockout models, as there are major differences in CYP epoxygenase isoform expression between species. This is particularly true for the CYP2C family of proteins, which have been most frequently linked to angiogenesis ([@bib12]). The majority of studies have also concentrated on the signaling initiated by the EETs, despite the fact that CYP isozymes can oxidize other ω-6 and ω-3 polyunsaturated fatty acids. Intracellular levels of the lipid epoxides are carefully controlled and are determined by their rate of generation by the CYP enzymes, as well as by their metabolism to the corresponding diols by the soluble epoxide hydrolase (sEH; gene = Ephx2; [@bib28]; [@bib22]). The latter enzyme is highly conserved between species, and targeting its expression or activity in mice is an effective way of manipulating fatty acid epoxide and diol levels in vivo ([@bib49]). The aim of this study was to determine the consequences of sEH deletion/inhibition on retinal angiogenesis in mice and to use LC-MS/MS based lipid profiling to identify the lipid underlying the phenotype observed. Here, we report that an sEH-derived metabolite of the ω-3 fatty acid DHA, rather than arachidonic acid, regulates murine retinal angiogenesis, and that a DHA diol generated by Müller glia cells is a major determinant of postnatal retinal angiogenesis via its ability to inhibit the γ-secretase. RESULTS ======= Delayed angiogenesis in sEH^−/−^ retina --------------------------------------- sEH protein and activity were detected in retinas from adult and postnatal wild-type mice ([Fig. 1 A](#fig1){ref-type="fig"}), and although the highest sEH activity was detected in the liver, activity in the retina was greater than that detected in either the spleen or lung ([Fig. 1 B](#fig1){ref-type="fig"}). To study the importance of the sEH in sprouting angiogenesis in vivo, we focused on retinal angiogenesis during the first postnatal week. sEH^−/−^ mice displayed a significant delay in the radial extension of the vascular plexus from the optic nerve to the periphery at postnatal days 2 (P2) and 5, as well as fewer branch points ([Fig. 1 C](#fig1){ref-type="fig"}). The delay in retinal angiogenesis in sEH^−/−^ retinas at P5 was associated with a significant reduction in endothelial cell proliferation (BrdU incorporation) at the angiogenic front ([Fig. 1 D](#fig1){ref-type="fig"}) as well as with fewer tip cells and filopodia ([Fig. 1 E](#fig1){ref-type="fig"}). {#fig1} Because all of these processes are highly dependent on the Notch pathway ([@bib18]; [@bib53]), our findings suggested that Notch signaling may be activated in the sEH^−/−^ animals. We therefore studied the expression of the Notch receptor and ligands as well as the Notch downstream targets Hes1 and Hey1. Retinal Hey1 increased from P2 to P7 in sEH^−/−^ retina, whereas Hes1 levels were elevated in sEH^−/−^ retina on P2 and P5 but decreased by P7 ([Fig. 1 F](#fig1){ref-type="fig"}). Notch1 was up-regulated in sEH^−/−^ mice immediately after birth (P0) but RNA levels then decreased (after P2). A transient increase in retinal delta-like ligand (Dll) 1 was detected on P5 but there were no significant changes in Dll4, which is also a target of activated Notch receptors ([@bib42]). A more detailed analysis of purified (CD31 beads) sEH^−/−^ retinal endothelial cells confirmed the significant up-regulation of Dll4, Hes1, and Hey1 in 5-d-old sEH^−/−^ mice ([Fig. 1 G](#fig1){ref-type="fig"}). Although coverage of the retina by the primary vascular layer was almost complete at day 7 in both genotypes, delayed angiogenesis persisted so that the formation of the secondary capillary plexus was also attenuated at P9 in sEH^−/−^ mice ([Fig. 1 H](#fig1){ref-type="fig"}). Different approaches were taken to ensure that the phenotype observed was related to the lack of sEH activity during retinal development. First, floxed sEH mice (sEH^fl/fl^) mice were bred with CreERT2 mice to generate an inducible sEH knockout strain (sEH^iKO^). Then sEH^iKO^ pups were treated (on P1) with tamoxifen for 4 d, a procedure which decreased sEH expression by ∼80% from day 3 onwards ([Fig. 2 A](#fig2){ref-type="fig"}). The acute postnatal down-regulation of the sEH also attenuated retinal angiogenesis and increased retinal expression of Dll4, Hes1, and Hey1 ([Fig. 2, B--D](#fig2){ref-type="fig"}). The apparent difference in the sEH^−/−^ and sEH^iKO^ retinas can be most likely attributed to the fact that the sEH was still expressed at day 2 in the latter group of animals. The mammalian sEH protein is a homodimer, and each monomer consists of an N-terminal domain which displays lipid phosphatase activity and a larger C terminal which possesses classical α/β-hydrolase activity ([@bib9]; [@bib38]). Therefore, in a second approach to ensure that the defects observed in the sEH^−/−^ mice were due to the loss of epoxide hydrolase activity, newborn wild-type mice were treated with *trans*-4-\[4-(3-adamantan-1-ylureido)cyclohexyloxy\]-benzoic acid (*t*-AUCB), a specific sEH inhibitor which does not affect the activity of the lipid phosphatase domain ([@bib21]). Daily inhibitor treatment from P1 onwards significantly decreased retinal vascularization (assessed at P5) compared with vehicle-treated littermates, and this delay was accompanied by a significant increase in the expression of Dll4, Hes1, and Hey1 ([Fig. 2, E--G](#fig2){ref-type="fig"}). {#fig2} Expression of the sEH in Müller glia cells and astrocytes --------------------------------------------------------- In retinal cross sections, the sEH was detected in the ganglion cell, inner plexiform, and outer plexiform layers but did not colocalize with the Isolectin B4 staining of the vasculature. Rather, the sEH colocalized with glutamine synthase (GS; [Fig. 3 A](#fig3){ref-type="fig"}), as well as vimentin ([Fig. 3 B](#fig3){ref-type="fig"}), aquaporin 4 (AQP-4), and, to a certain extent, glial fibrillary acidic protein (GFAP; [Fig. 3 C](#fig3){ref-type="fig"}). In whole mount preparations, the sEH was detected in short AQP-4--positive processes surrounding the vasculature but not in the vessels themselves---i.e., consistent with expression in Müller cell processes ([Fig. 3 D](#fig3){ref-type="fig"}). Indeed, Müller cells freshly isolated from adult retinas displayed the typical morphology with multiple end-feet and expressed the sEH protein ([Fig. 3 E](#fig3){ref-type="fig"}). Although the morphology of the Müller cells in culture was markedly different from that of freshly isolated cells, they continued to express the sEH, as well as AQP-4 and the Müller cell marker cellular retinaldehyde binding-protein (CRALBP; [@bib6]; [Fig. 3 F](#fig3){ref-type="fig"}). {#fig3} To target sEH expression in Müller cells, sEH^fl/fl^ mice were crossed with platelet-derived growth factor receptor α polypeptide (Pdgfra)--cre deleter mice ([@bib47]). The resulting Müller cell--specific sEH knockout mice (sEH^ΔMC^) showed a greatly reduced expression of the sEH in the inner and outer plexiform layers of the retina ([Fig. 4 A](#fig4){ref-type="fig"}), decreased retinal epoxide hydrolase activity ([Fig. 4 B](#fig4){ref-type="fig"}), and reproduced the delay in angiogenesis ([Fig. 4, C--E](#fig4){ref-type="fig"}) as well as the altered tip cell morphology ([Fig. 4, F--H](#fig4){ref-type="fig"}) of the sEH^−/−^ mice. These phenomena were also linked to altered Notch signaling, as Hes1 and Hey1 expression were also increased in retinas from 5-d-old sEH^ΔMC^ mice ([Fig. 4 I](#fig4){ref-type="fig"}). {#fig4} Several studies have highlighted the role of the astrocyte framework as a template for the primary plexus outgrowth as well as a source of angiogenic growth factors ([@bib51]; [@bib15]). Given that the sEH is expressed in astrocytes ([@bib45]) and that the sEH was partially colocalized with GFAP, we generated mice with astrocyte-specific deletion of the sEH (sEH^fl/fl^ mice crossed with GFAP-cre deleter mice). Astrocyte-specific deletion of the sEH ([Fig. 5 A](#fig5){ref-type="fig"}) resulted in the development of a less dense upper capillary layer characterized by fewer branching points in central areas and at the leading edge ([Fig. 5 B](#fig5){ref-type="fig"}). The marked delay in radial extension observed in global sEH^−/−^ mice was, however, not observed. In addition, tip cell numbers and filopodia length were comparable in sEH^fl/fl^ and GFAP-sEH mice ([Fig. 5 C](#fig5){ref-type="fig"}), as were the expression of Hes1 and Hey1 ([Fig. 5 D](#fig5){ref-type="fig"}). These data indicated that the sEH expressed in Müller cells played a more prominent role in regulating angiogenesis than the sEH in astrocytes. {#fig5} Müller cell lipid profile ------------------------- Müller cells are a rich source of angiogenesis-promoting growth factors, in particular vascular endothelial cell growth factor (VEGF) A ([@bib43]; [@bib52]). Therefore, we compared the effects of conditioned medium (CM) from wild-type and sEH^−/−^ Müller cells on endothelial cell proliferation. CM from sEH-expressing Müller cells potentiated endothelial cell proliferation, whereas the CM from sEH^−/−^ Müller cells exhibited a weaker effect ([Fig. 6 A](#fig6){ref-type="fig"}). There were no differences in VEGFR1, VEGFR2, or VEGFR3 expression in retinas from wild-type and sEH^−/−^ mice over the first postnatal week (unpublished data). VEGF levels were also similar in whole retinas from wild-type and sEH^−/−^ mice up to day 9 ([Fig. 6 B](#fig6){ref-type="fig"}), but the generation of VEGF was reduced in Müller cells and in CM from sEH^−/−^ mice ([Fig. 6, C and D](#fig6){ref-type="fig"}). The expression of the anti-angiogenic pigment epithelium-derived factor ([@bib10]) was not different in CM from wild-type and sEH^−/−^ Müller cells. Despite the apparent difference in VEGF secretion, VEGF neutralizing antibodies elicited similar effects on endothelial cells treated with wild-type-CM or sEH^−/−^-CM ([Fig. 6 E](#fig6){ref-type="fig"}). {#fig6} However, a soluble factor was responsible for the observed phenotype, as when sEH^−/−^ Müller cells were co-cultured with, but kept physically separated from, endothelial cells (Transwell chambers), endothelial cell proliferation (assessed by BrdU incorporation) was attenuated. Moreover, the reexpression (adenoviral transduction) of the sEH in Müller cells normalized endothelial cell proliferation ([Fig. 6 F](#fig6){ref-type="fig"}). Similarly, the expression of Hes1 and Hey1 increased in endothelial cells cultured with sEH^−/−^ versus wild-type Müller cells ([Fig. 6 G](#fig6){ref-type="fig"}), and a rescue experiment, in which the sEH was reintroduced into sEH^−/−^ Müller cells, reversed these effects. To determine which of the sEH substrates or products could account for the delay in retinal angiogenesis, lipid profiles (LC-MS/MS) were generated from the retinas of 5-d-old wild-type and sEH^−/−^ mice. As expected, levels of 11,12- and 14,15-EET were increased in samples from sEH^−/−^ mice, whereas the levels of the corresponding diols were slightly decreased ([Fig. 7 A](#fig7){ref-type="fig"}). However, the most pronounced changes were in the levels of the DHA derivative 19,20-dihydroxydocosapentaenoic acid (DHDP). In CM from sEH^−/−^ Müller cells, 11,12-EET, 14,15-EET, and 19,20-epoxydocosapentaenoic acid (EDP) levels were significantly increased in CM and were restored to control by adenoviral transduction of the sEH ([Fig. 7 B](#fig7){ref-type="fig"}). The opposite was true for 19,20-DHDP, which was attenuated in CM from sEH^−/−^ Müller cells and restored by sEH overexpression. There was no significant change in 11,12-dihydroxyeicosatrienoic acid (DHET) levels, 14,15-DHET levels were below the limit of detection, and there were no apparent changes in the levels of the linoleic acid--derived fatty acids, 12,13-epoxyoctadecenoic acid (EpOME), or 12,13-dihydroxyoctadecenoic acid (DiHOME). {#fig7} 19,20-DHDP and Notch signaling ------------------------------ 19,20-DHDP is of particular interest, given that it was the most abundant lipid detected in the retinal lipid profile and that its precursor DPA has previously been reported to modulate the activity of the γ-secretase ([@bib11]; [@bib16]). When activated, the γ-secretase cleaves the Notch intracellular domain (NICD), which then either undergoes rapid proteasome degradation or translocates to the nucleus and binds to transcription factors ([@bib5]). We therefore compared the ability of 19,20-DHDP and the corresponding epoxide and sEH substrate 19,20-EDP to influence Dll4-stimulated Notch activity in murine endothelial cells. Although 19,20-EDP had no effect on the Dll4-induced increase in NICD cleavage, 19,20-DHDP decreased NICD levels by 54 ± 12% versus 89 ± 7% inhibition in the presence of the γ-secretase inhibitor *N*-\[*N*-(3,5-Difluorophenacetyl)-[l]{.smallcaps}-alanyl\]-S-phenylglycine t-butyl ester (DAPT; [Fig. 7 C](#fig7){ref-type="fig"}). Neither 19,20-EDP nor 19,20-DHDP affected VEFGR phosphorylation or downstream signaling in vascular endothelial cells (unpublished data). Treatment of cells with other sEH substrates, e.g., 11,12-EET, 14,15-EET, or 12,13-EpOME, or products, e.g., 11,12-DHET, 14,15-DHET, or 12,13-DiHOME, that were altered in the sEH^−/−^ retinas had no effect on the Dll4-induced cleavage of NICD ([Fig. 7, D--F](#fig7){ref-type="fig"}). When the nuclear translocation of the NICD was assessed by immunofluorescence ([Fig. 8 A](#fig8){ref-type="fig"}), 19,20-DHDP, but not 19,20-EDP, attenuated the Dll4-stimulated translocation. The fact that 19,20-DHDP also attenuated NICD levels in the presence of the proteasome inhibitor MG 132 ([Fig. 8 B](#fig8){ref-type="fig"}) suggested that the generation of the NICD, rather than its degradation, was targeted by the lipid. {#fig8} The binding of a Notch ligand to a Notch receptor initiates two steps of cleavage to generate the NICD, the first by ADAM (a disintegrin and metalloprotease) proteases and the second by the γ-secretase ([@bib29]). The effect of 19,20-DHDP on the γ-secretase was evaluated using cells expressing a Notch 1 mutant (NotchΔE) that lacks the extracellular domain of the protein and possesses only the γ-secretase cleavage site ([@bib17]). We found that 19,20-DHDP, but not 19,20-EDP, reduced the γ-secretase--dependent generation of NICD from NotchΔE ([Fig. 8 C](#fig8){ref-type="fig"}). The small effect observed at higher concentrations of 19,20-EDP can most likely be attributed to the endogenous expression of the sEH in the cells used ([@bib3]) and its conversion to 19,20-DHDP. The γ-secretase complex is made up of several proteins, including presenilin 1, nicastrin, anterior pharynx defective 1, and presenilin enhancer 2, and the active complex is localized to lipid rafts ([@bib56]; [@bib20]; [@bib55]). We therefore determined whether 19,20-DHDP could inhibit the γ-secretase by altering the localization of constituent proteins and concentrated on presenilin 1. After sucrose density gradient fractionation, presenilin 1 and the raft-associated protein flotillin-1/reggie-2 were detected in detergent-insoluble lipid raft fractions (fractions 5 and 6) isolated from solvent--treated cells. 19,20-DHDP caused the displacement of presenilin 1 from lipid rafts to non-raft fractions (fractions 10 to 12) without affecting the distribution of flotillin-1 ([Fig. 8 D](#fig8){ref-type="fig"}). The latter effect that was accompanied by a decrease in cholesterol levels in the raft fractions (fractions 5 and 6) and an increase in the non-raft fractions ([Fig. 8 E](#fig8){ref-type="fig"}). Neither 19,20-EDP (in the presence of the sEH inhibitor *t*-AUCB) nor DAPT affected the colocalization of presenilin 1 with flotillin-1 or the distribution of cholesterol between the different fractions. Finally, we determined whether or not 19,20-DHDP could rescue the retinal phenotype in sEH^−/−^ mice. Intravitreal injection of 19,20-DHDP was performed in 5-d-old wild-type and sEH^−/−^ pups and its effects compared with that of DAPT-treated mice. As before, the sEH^−/−^ retinas demonstrated a less dense vessel network as well as fewer tip cells and filopodia than their wild-type littermates, but this phenotype was largely reversed after the administration of 19,20-DHDP ([Fig. 9](#fig9){ref-type="fig"}). Similar effects were observed in retinas from animals receiving DAPT, whereas 19,20-EDP was without effect. In retinas from wild-type mice, 19,20-DHDP (but not 19,20-EDP) also induced a more complex vessel network, and increased tip cell and filopodia numbers ([Fig. 9](#fig9){ref-type="fig"}). {#fig9} The E3 ubiquitin ligase Fbxw7 is involved in the ubiquitination and degradation of the NICD ([@bib54]; [@bib61]), and endothelial cell-specific and inducible Fbxw7 mutant (Fbxw7^iECKO^) mice demonstrate retinal blood vessel growth impairment that is attributable to Notch activation ([@bib25]). To confirm the in vitro data suggesting that 19,20-DHDP inhibits γ-secretase activity, the lipid was given intravitrealy to Fbxw7^iECKO^ mice. Also in these animals, 19,20-DHDP increased sprouting and filopodia numbers ([Fig. 10](#fig10){ref-type="fig"}), confirming the inhibition of NICD generation by the lipid. Similar effects were observed in animals that received DAPT, whereas 19,20-EDP was without effect. {#fig10} DISCUSSION ========== The results of the present study show that the genetic deletion and inhibition of the sEH activates Notch signaling and attenuates physiological sprouting angiogenesis in the murine retina. Mechanistically, these effects could be attributed to the lack of the sEH product 19,20-DHDP, a fatty acid diol derived from DHA which inhibits the γ-secretase by inducing the redistribution of presenilin 1 from lipid rafts. Moreover, 19,20-DHDP, but not the parent epoxide, was able to rescue the defective angiogenesis in sEH^−/−^ mice as well as in animals lacking the Fbxw7 ubiquitin ligase which demonstrate strong basal activity of the Notch signaling cascade. Müller glia cells were identified as the major source of the sEH in the murine retina. These cells develop and maintain a close contact with both superficial vessels and deeper capillaries via their multiple end-feet ([@bib37]) and have been implicated in angiogenesis by virtue of their ability to produce angiogenic substances in response to hypoxia ([@bib43]; [@bib52]; [@bib46]). However, as the Müller cell--specific deletion of VEGF-A inhibited neovascularization in a mouse model of oxygen-induced retinopathy without affecting physiological vascularization or retinal morphology ([@bib2]), it was assumed that Müller cells play a greater role in proliferative retinopathy than physiological angiogenesis. In our study, the deletion of the sEH in Müller cells clearly affected endothelial cell proliferation as well as Notch signaling and tip cell filopodia formation in mice, indicating that Müller cells make a more important contribution to retinal angiogenesis than generally appreciated. Interestingly, several recent studies also support a role for VEGF-independent signaling by Müller cells in developmental angiogenesis. For example, Müller cells secrete Norrin, a retinal signaling molecule which binds to Frizzled-4 to activate canonical Wnt/β-catenin signaling, in the absence of which the development of the superficial vessels was attenuated and deeper intraretinal capillaries were not formed ([@bib63]; [@bib64]; [@bib39]). Also, the deletion of hypoxia inducible factor-1α in neuroretinal cells (includes Müller cells) results in impaired vascular development characterized by decreased tip cell filopodia and reduced vessel branching ([@bib36]), a phenotype very similar to that of sEH^−/−^ mice. The Müller cell--specific sEH knockout (sEH^ΔMC^) mice were generated using Pdgfra-cre mice ([@bib47]), which may also target astrocytes. Also, because astrocytes express the sEH ([@bib45]) and retinal angiogenesis is generally thought to be largely determined by astrocyte-derived VEGF ([@bib52]), we also focused on this interaction. However, although vessel branching was impaired in retinas from astrocyte-specific sEH knockout (GFAP-sEH) mice, the tip cell phenotype and Notch signaling (Hes1 and Hey1 expression) were unaltered, suggesting that another sEH-expressing cell type underlies the effects observed. Our findings do not completely rule out a role for astrocyte sEH in retinal angiogenesis but rather indicate that another source of the sEH plays a more dominant role. To date, the CYP--sEH axis has been linked with the promotion of angiogenesis, as the epoxides of arachidonic acid (particularly 11,12- and 14,15-EET) promote tube formation in vitro ([@bib35]; [@bib33]) and can promote the vascularization of matrix material and tumors in vivo ([@bib26]). For a while, sEH inhibitors looked like highly interesting compounds for the treatment of cardiovascular disorders ([@bib23]), but interest was dampened by speculation that the associated increase in EET levels could induce or promote cancer progression. Certainly, tumor growth and metastasis can be increased by sEH inhibition in transgenic mice with high vascular EET levels, i.e., animals that overexpressed either the human CYP2C8 or human CYP2J2 specifically in Tie-2--expressing cells or that were treated with high concentrations of 14,15-EET ([@bib40]). The latter models were somewhat artificial and focused on the products of arachidonic acid metabolism, largely ignoring the biological actions of other lipids that feed into the same CYP--sEH axis. It was partly to assess the role of the sEH in angiogenesis under more physiological conditions that we determined the effects of the global and induced deletion of the sEH, as well as its pharmacological inhibition in the postnatal murine retina. We found that sEH deletion and inhibition resulted in markedly decreased capillary branching and endothelial cell proliferation as well as reduced tip cell numbers and filopodia. If the epoxides of arachidonic acid possess angiogenic properties, why was angiogenesis attenuated in retinas from sEH^−/−^ mice which demonstrate higher than normal circulating and tissue levels of EETs? One logical explanation is that other lipids generated by the sequential activity of CYP epoxygenases and the sEH exert effects that dominate over, or antagonize those of, the EETs. The retinal phenotype observed could have been due to the accumulation of an anti-angiogenic epoxide or the loss of a pro-angiogenic diol, even though lipid diols are generally assumed to be either inactive or markedly less active than the parent epoxides ([@bib23]). To identify such lipids, we used an LC-MS/MS--based lipid profiling approach to screen for the lipid epoxide or diol most affected by the deletion of the sEH in the retina and found that 11,12-EET, 14,15-EET, and 12,13-EpOME were elevated, whereas 19,20-DHDP was attenuated in retinas from sEH^−/−^ mice. These findings could be confirmed in isolated Müller cells. Interestingly, DHA levels are known to be higher than those of arachidonic acid in the retina ([@bib1]), and we detected higher levels of the DHA-derived lipids 19,20-EDP and 19,20--DHDP than the epoxides and diols of arachidonic acid in retinas from 5-d-old mice. Which of the lipids affected by the deletion of the sEH could contribute to the development of the retinal phenotype in sEH^−/−^ mice? A significant delay in the radial extension of the vascular plexus, coupled with reduced endothelial cell proliferation as well as with fewer tip cells and filopodia, are indicative of Notch signaling activation ([@bib18]; [@bib53]). Therefore, we assessed the ability of all four candidate lipids to influence Dll4-stimulated Notch activity and the generation of the NICD. However, the sEH product 19,20,-DHDP and not the sEH substrates 19,20-EDP, 11,12-EET, 14,15-EET, or 12,13-EpOME attenuated NICD levels. Two distinct cleavage steps are required to generate the NICD; the first is dependent on ADAM proteases and the second on the activity of the γ-secretase ([@bib29]). To determine which of these proteases could be targeted by 19,20-DHDP, we assessed its ability to influence the Dll4-induced generation of the NICD from a so-called NotchΔE that possesses only the γ-secretase cleavage site ([@bib17]). The finding that 19,20-DHDP also affected NICD generation in this model indicated that its molecular target was the γ-secretase. To confirm this in vivo, we assessed the consequences of the intravitreal application of 19,20-EPD or 19,20-DHDP to 5-d-old mice with hyper-activated endothelial cell Notch signaling, i.e., animals lacking the E3 ubiquitin ligase involved in the ubiquitination and degradation of the NICD ([@bib25]). In the latter Fbxw7^iECKO^ animals, which demonstrate markedly impaired retinal blood vessel growth impairment that is attributable to Notch activation ([@bib25]), 19,20-DHDP increased vessel branching as well as tip cell and filopodia numbers. The γ-secretase inhibitor DAPT elicited similar effects, whereas 19,20-EDP was ineffective. Fatty acid epoxides such as the EETs are thought to act via a specific membrane receptor or the activation of peroxisome proliferator-activated receptors ([@bib50]; [@bib41]). How can fatty acid diols exert biological effects that could lead to the inhibition of the γ-secretase? A cell surface receptor may not be essential, as polyunsaturated fatty acids can incorporate into membrane phospholipids and potentially influence cell signaling and biological responses by modulating the lipid microenvironment therein ([@bib27]). Certainly, the fish oils EPA and DHA are readily incorporated into phospholipids and the resulting polyunsaturated phospholipids are able to infiltrate lipid rafts as well as form non-raft domains ([@bib62]). DHA, from which 19,20-DHDP is derived, is particularly interesting in this respect as it tends to incorporate more readily into lipid rafts than EPA ([@bib62]). Changes in the lipid composition of the plasma membrane can have marked consequences of the formation and stability of the γ-secretase complex ([@bib56]; [@bib55]). We found that 19,20-DHDP, but not the corresponding epoxide, altered cholesterol distribution in cellular membranes and displaced presenilin 1 from lipid rafts, where the processing of substrate proteins preferentially occurs, to non-raft fractions, thus disrupting the γ-secretase complex. Little is known about the incorporation of DHA-derived epoxides or diols into phospholipids, but this is highly likely given that the epoxides of arachidonic acid (the EETs) can be esterified to phosphatidylcholine, phosphatidylethanolamine, and phosphatidylinositols ([@bib7]). This is not the first time a link between DHA and the γ-secretase has been reported, and although DHA increased γ-secretase activity in vascular smooth muscle cells ([@bib11]), it was found to redistribute cholesterol out of lipid rafts ([@bib58], [@bib59]), thereby decreasing γ-secretase activity in a neuroblastoma cell line ([@bib16]). Our results agree with the latter study. Moreover, given that fatty acids such as DHA are rapidly metabolized by CYP and sEH enzymes and that the DHA epoxide 19,20-EDP had no effect on the γ-secretase, it is possible that the DHA diol 19,20-DHDP and the activity of the sEH may underlie the previously reported changes in lipid raft composition and biological responses rather than DHA on its own. These studies highlight the fact that consequences of sEH inhibition in vivo cannot be solely attributed to the accumulation of the epoxides or arachidonic acid. In fact, more and more evidence is being accumulated to demonstrate that other lipid products (epoxides and diols) of the CYP--sEH axis can affect different steps in vessel development and maturation. For example, the sEH products 12,13-DiHOME (generated from the linoleic acid epoxide 12,13-EpOME), and to a lesser extent 11,12-DHET (generated from 11,12-EET), attenuate angiogenesis and vascular repair via the activation of the canonical Wnt pathway ([@bib14]). Moreover, the combination of 19,20-EDP with an sEH inhibitor was reported to inhibit primary tumor growth and metastasis by preventing the VEGF-induced phosphorylation of VEGFR2 ([@bib66]). However, why the latter lipid failed to affect retinal angiogenesis is unclear at the moment but fits with our finding that the phenotype observed in the sEH^−/−^ mice was independent of VEGF signaling. Our study also highlights the role of the neuroretina and particularly the sEH in Müller cells in the regulation of retinal angiogenesis. It will be interesting to determine whether or not this mechanism can be exploited and whether the pathological retinal angiogenesis associated with Müller cell activation ([@bib8]), such as that observed in oxygen-induced retinopathy in preterm infants, can be linked to altered lipid signaling and modulated by sEH inhibition. MATERIALS AND METHODS ===================== ### Materials. The sEH inhibitor *t*-AUCB was provided by B. Hammock (University of California, Davis, Davis, CA). The proteasomal inhibitor MG-132 was obtained from EMD Millipore. Mouse and human Dll4 protein were obtained from R&D Systems. DAPT, tamoxifen, and all other chemicals were purchased from either Sigma-Aldrich or Merck and Roth. ### Animals. C57BL/6 mice (6--8 wk old) were purchased from Charles River. sEH^−/−^ mice ([@bib49]) were provided by F. Gonzalez (National Institutes of Health, Bethesda, MD) and cross-bred for ∼20 generations onto the C57BL/6 background in the animal house facility at Frankfurt University. Inducible and endothelial cell--specific Fbxw7^iECKO^ mutant mice, generated as described previously ([@bib25]), were treated with tamoxifen (1 mg/ml i.p., 2% ethanol in sunflower oil, 50 µl) once a day from days 1--4. Floxed sEH (sEH^fl/fl^) mice (Taconic) were generated using C57BL/6 embryonic stem cells for gene targeting. Positive selection cassettes flanked by FRT/F3 sites were deleted by crossing with a ubiquitously expressing FLP1 recombinase strain (Taconic). These animals were cross-bred with animals expressing Cre under the control of the GFAP promoter (FVB-TgGFAP-cre25Mes \[[@bib67]\] backcrossed in the C57BL/6 background) to generate mice lacking sEH in astrocytes. Inducible sEH knockout mice (iCreER-sEH) were generated by breeding sEH^fl/fl^ mice with the RosaCre-ER-T2 strain ([@bib24]; cross-bred to C57BL/6) and Müller cell--specific sEH-deficient mice (sEH^ΔMC^) were generated by crossing the sEH^fl/fl^ mice with C57BL/6-TgPdgfra-cre1Clc/J mice ([@bib47]; The Jackson Laboratory). Mice were housed in conditions that conform to the *Guide for the Care and Use of Laboratory Animals* published by the U.S. National Institutes of Health (publication no. 85--23). All experiments were approved by the governmental authorities (Regierungspräsidium Darmstadt: F28/06, F28/23, and F28/38; Münster: 84-02.04.2011.A257). Age-, gender-, and strain-matched animals (usually littermates) were used throughout. ### sEH inhibition and postnatal deletion studies. To inhibit epoxide hydrolase activity, C57BL/6 pups received *t*-AUCB (2 mg/kg i.p., twice per day) from days 1--4. To postnatally delete the sEH, iCreER-sEH transgenic mice were treated with tamoxifen (1 mg/ml i.p., 2% ethanol in sunflower oil, 50 µl) once a day from days 1--4 as previously described ([@bib4]). On day 5, retinas were isolated for Western blotting and immunohistochemistry. ### Intravitreal injection. Mice were anesthetized with 2% isoflurane. A drop of 0.5% proparacaine was administered as a topical local anesthesia. A fine glass micropipette connected to a 10 µl Hamilton glass syringe was inserted through the incision in the cornea and slid between the iris and the lens into the posterior chamber of the eye. Each eye received 0.5 µl 19,20-EDP, 100 µmol/liter 19,20-DHDP, or 200 µmol/liter DAPT or 0.5 µl of solvent (1% DMSO in PBS) as control. Injections were administered slowly over approximately one minute and monitored with a stereo microscope. ### Endothelial cell proliferation in vivo. Proliferating endothelial cells in P5 retinas were labeled by administering BrdU (50 mg/kg i.p.; BD) 4 h before sacrifice. After Isolectin B4 staining, retinas were post-fixed for 30 min in 4% paraformaldehyde (PFA), washed three times with PBS, incubated for 1 h in 1 M HCl, washed five times in PBS, blocked, and incubated overnight with anti-BrdU antibody (1:50; BD). Secondary detection was performed with Alexa Fluor--coupled secondary antibodies. ### Retina preparation and analysis. Retinas for whole-mount immunohistochemistry were fixed in 4% PFA for 2 h at room temperature, or overnight at 4°C. After fixation, retinas were blocked and permeabilized in 1% BSA and 0.5% Triton X-100 overnight at 4°C. Then retinas were washed three times in Pblec buffer (0.5% Triton X-100, 1 mM CaCl~2~, 1 mM MgCl~2~, and 1 mM MnCl~2~ in PBS, pH 6.8) and incubated overnight in Pblec containing Fitc-labeled Isolectin B4 (1:100; Sigma-Aldrich). The following primary antibodies were diluted in 1% BSA and 0.5% Triton X-100 and incubated overnight: sEH (1:250; provided by M. Arand, Institute of Pharmacology and Toxicology, Uinversity of Zurich, Zurich, Switzerland), GFAP (1:500; DAKO; or 1:1,000; EMD Millipore), NG-2 (1:1,000; EMD Millipore), AQP-4 (1:500; Santa Cruz Biotechnology, Inc.), GS (1:1,000; EMD Millipore), and vimentin (1:500; Nococastra). For secondary detection, Alexa Fluor--coupled secondary antibodies (1:200) were used. Cell nuclei were visualized with DAPI (1:200; Molecular Probes). After antibody staining, retinas were post-fixed in 4% PFA for 15 min before flat-mounting in mounting medium (Dako). 10-µm cryosections were cut after retinas were embedded within Tissue-Tek OCT Compound (Sakura). All quantification was done with high-resolution images taken using a meta laser scanning confocal microscope (LSM-510; Carl Zeiss) as previously described ([@bib30]). ### LC-MS/MS lipid profiles. Samples were extracted twice into ethyl acetate, evaporated under nitrogen, and resuspended in methanol/water (vol. 1:2). The eicosanoid profiles generated were determined with a mass spectrometer (API4000; AB Sciex) operating in multiple reaction monitoring (MRM) mode as previously described ([@bib34]; [@bib14]). Chromatography was performed on a Gemini C18 column (150 mm length, 2 mm inner diameter; particle size, 5 µm; Phenomenex). All fatty acid epoxides, diols, and deuterated analogues were obtained from Cayman Europe. ### Müller cell culture. Müller cells were isolated from 8--12-d-old mice, as previously described ([@bib19]), and cultured in DMEM/F12 containing 10% FCS (Biochrom), 100 U/ml penicillin, and 100 µg/ml streptomycin. After 6 d, the cultures were rinsed free of nonadherent tissue, and culture medium was replenished every 3--4 d. The purity of the Müller cell cultures was characterized after each passage by testing for the expression of vimentin, AQP-4, GS, and CRALBP ([@bib6]). Müller cell CM was generated by culturing confluent wild-type or sEH^−/−^ Müller cells for 48 h. ### Adenovirus generation and Müller cell transduction. The human sEH (Ephx2) cDNA sequence, with a C-terminal Myc/His tag, was excised via NotI--PmeI from *pcDNA3*.1myc/His as previously described ([@bib3]) and cloned into the NotI--EoRV site of the cloning plasmid pAdTrackCMV. The recombination into the adenoviral transfer plasmid and the generation of the viruses was performed as previously described ([@bib31]). For the sEH rescue experiments, sEH^−/−^ Müller cells were infected with sEH adenovirus for 6 h and then washed four times with Hank's solution. Experiments were performed after an additional 48 h and sEH expression was confirmed by immunoblotting. ### Müller cell and endothelial cell co-culture. 1 × 10^6^ endothelial cells isolated as previously described ([@bib13]) were seeded on 6-well plates coated with fibronectin. After 4 h, the nonadherent cells were removed and the medium was replaced with 1.1 ml DMEM/F12 containing 10% FCS. Then Transwell inserts (Millipore) were added to the wells and seeded with 6 × 10^5^ Müller cells. After 8 h, the insert was removed and RNA was extracted from the endothelial cells. ### Müller cell CM on endothelial cell proliferation. 5 × 10^4^ endothelial cells were seeded in 24-well plates and stimulated with M-CM from wild-type or sEH^−/−^ Müller cells (mixed with endothelial cell growth medium at 1:1 ratio), or solvent (DMEM/F12). Mouse VEGF at a final concentration of 50 ng/ml was used as a positive control. Viable cells were counted (CASY; Roche) after 24, 48, and 72 h of stimulation. For the BrdU incorporation, endothelial cells were seeded in 96-well plates (1 × 10^4^ cells/well) and stimulated with the same condition as described above for 48 h. BrdU was added to cells 4 h before the end of stimulation and then BrdU incorporation was assessed using BrdU ELISA kit (Roche) according to the manufacturer's instructions. ### Retinal endothelial cell isolation. Retinas were isolated from P5 mice. After digestion with 0.1% trypsin (Worthington Biochemical Corporation) and 70 U/ml collagenase type I (Biochrom) in DMEM/F12 medium at 37°C for 30 min, retinal endothelial cells were isolated with anti--mouse CD 31 antibody based on magnetic bead separation similar to that described above. CD-31--positive cells were used for gene expression analysis. ### RT-qPCR. Total RNA from retinas or cultured primary cells was extracted using an RNeasy kit (QIAGEN), and equal amounts (1 µg) of total RNA was reverse transcribed (Superscript III; Invitrogen). Gene expression levels were detected using SYBR Green (Absolute QPCR SYBR Green Mix; Thermo Fisher Scientific). The TaqMan probe used for the detection of the Dll4 RNA was from Applied Biosystems. The relative expression levels of the different genes studied was calculated using the 2^−ΔΔCt^ method with the 18S RNA as a reference. The primer sequences used were as follows: 18S forward, 5′-CTTTGGTCGCTCGCTCCTC-3′, reverse, 5′-CTGACCGGGTTGGTTTTGAT-3′; Dll1 forward, 5′-CATCCGATACCCAGGTTGTC-3′, reverse, 5′-ACGGCTTATGGTGAGTACAG-3′; Jagged1 forward, 5′-TCTCTGACCCCTGCCATAAC-3′, reverse, 5′-TTGAATCCATTCACCAGATCC-3′; Notch1 forward, 5′-CCTCAGATGGTGCTCTGATG-3′, reverse, 5′-CTCAGGTCAGGGAGAACTAC-3′, Hes1, forward, 5′-GAGGCGAAGGGCAAGAATAAA-3′, reverse, 5′-GTGGACAGGAAGCGGGTCA-3′; Hey1 forward, 5′-TGAGCTGAGAAGGCTGGTAC-3′, reverse, 5′-ACCCCAAACTCCGATAGTC-3′; and VEGF forward, 5′-GCACTGGACCCTGGCTTTACTGCTGTA-3′, and reverse, 5′-GAACTTGATCACTTCATGGGACTTCTGCTC-3′. ### Dll4-induced NICD production. Endothelial cells were seeded onto 6-well plates coated with or without 500 ng/ml recombinant mouse Dll4 protein (R&D Systems) in 0.1% BSA/PBS for 1 h at room temperature. Experiments were performed in the presence of solvent, 10 µmol/liter 19,20-EDP/DHDP, 10 µmol/liter 11,12-EET/DHET, 10 µmol/liter 14,15-EET/DHET, 3 µmol/liter 12,13-EpOME/DiHOME, 20 µmol/liter DAPT, or 10 µmol/liter MG 132. After an additional 4 h, cells were washed three times with PBS and lysed in a Triton X-100--containing buffer. Intracellular domain of Notch (NICD) levels were assessed by Western blotting with an S3 (γ-secretase) cleavage-specific antibody Val1744 (Cell Signaling Technology). For the nuclear translocation of NICD, cells were fixed with 4% PFA for 15 min at room temperature, followed by NICD and DAPI immunostaining. Co-localization of NICD with DAPI was quantified with ImageJ software (National Institutes of Health). ### γ-Secretase assay. HEK 293 cells were seeded in 10 cm culture dishes and transfected with 6 µg NotchΔE construct (provided by P. Greengard, Rockefeller University, New York, NY). Cells were then incubated with solvent, 19,20-EDP, 19,20-DHDP (0.1--10 µmol/liter), or DAPT (20 µmol/liter) for 24 h. The γ-secretase--dependent generation of NICD was detected as described above. ### Lipid rafts. HEK293 cells were incubated with solvent (0.1% DMSO), 10 µmol/liter 19,20-EDP, 10 µmol/liter 19,20-DHDP, or 20 µmol/liter DAPT in serum-free medium containing 10 µmol/liter of the sEH inhibitor *t*-AUCB for 2 h at 37°C. Cells were then washed twice with ice-cold PBS, harvested by scraping, recovered by centrifugation (300 g for 4 min), and resuspended in 2 ml MBS buffer (containing 1% CHAPSO in 25 mM MES, pH 6.5, 150 µmol/liter NaCl supplemented with a protease inhibitor mix, and 1 µmol/liter phenylmethylsulfonyl fluoride), and homogenized by using a glass homogenizer. Lipid rafts were isolated as previously described ([@bib56]). In brief, after 30 min at 4°C, equal amounts of protein were adjusted to a final sucrose concentration of 45% (final volume, 4 ml) and transferred to 12 ml ultracentrifuge tubes. A discontinuous sucrose gradient was then formed by sequentially overlaying 4 ml of 35% and 4 ml of 5% sucrose. Samples were subjected to ultracentrifugation (35,000 rpm, 4°C for 20 h) using a rotor (SW 41; Beckman). After centrifugation, twelve 1 ml fractions were collected using a capillary tube connected to a peristaltic pump, and equal volumes of each faction were analyzed by SDS-PAGE using antibodies against flotillin 1 (BD) and presenilin 1 (Santa Cruz Biotechnology, Inc.). Cholesterol levels in the different fractions were determined using the Amplex Red cholesterol assay (Invitrogen) according to the manufacturer's instructions. ### Statistical analysis. Data are expressed as mean ± SEM, and statistical evaluation was performed using Student's *t* test for unpaired data or one-way ANOVA, followed by a Bonferroni's *t* test where appropriate. Values of P \< 0.05 were considered statistically significant. The authors are indebted to Isabel Winter and Katharina Engel-Herbig for expert technical assistance. This work was supported by the Deutsche Forschungsgemeinschaft (SFB-TR 23/A6, GRK 880 and Exzellenzcluster 147 "Cardio-Pulmonary Systems"). J. Hu was supported by a research scholarship from China Scholarship Council (CSC). The authors declare no competing financial interests. Abbreviations used:AQP-4aquaporin 4CMconditioned mediumCRALBPcellular retinaldehyde binding-proteinCYPcytochrome P450DAPT*N*-\[*N*-(3,5-Difluorophenacetyl)-[l]{.smallcaps}-alanyl\]-S-phenylglycine t-butyl esterDHAdocosahexenoic acidDHDPdihydroxydocosapentaenoic acidDHETdihydroxyeicosatrienoic acidDiHOMEdihydroxyoctadecenoic acidDlldelta-like ligandEDPepoxydocosapentaenoic acidEETepoxyeicosatrienoic acidEPAeicosapentaenoic acidEpOMEepoxyoctadecenoic acidGFAPglial fibrillary acidic proteinGSglutamine synthaseNICDNotch intracellular domainPdgfraplatelet-derived growth factor receptor α polypeptidesEHsoluble epoxide hydrolase*t*-AUCB*trans*-4-\[4-(3-adamantan-1-ylureido)cyclohexyloxy\]-benzoic acidVEGFvascular endothelial cell growth factor

This website uses cookies We use cookies to analyse our traffic. We share information about your use of our site with our social media, advertising and analytics partners who may combine it with other information that you’ve provided to them. You consent to our cookies if you continue to use this website. See Cookie Policy here

Click here for the online Merit Badge ApplicationPrint on both sides of regular or blue paper or card stock and trim off the edges. This will print three blank forms per page. The back should line up with the front side. Ask your council what merit badge applications they will accept. Boy Scouts may work on merit badges from the time they join a Scout troop until they turn 18 years old. See Merit Badge for more details. The Merit Badge Application #34124 is known as a "blue card" due to its distinctive color. Though it has not been clearly stated in the past, units, districts, and local councils do not have the authority to implement a different system for merit badge approval and documentation. In any case, through the years, many councils have created new forms and approaches to the process, some including IT components. In an effort to gather and consider these potential best practices, councils are now asked to submit descriptions and copies of their blue card alternatives to the national Advancement Team. For very large events—such as the national Scout jamboree—the National Council may approve an alternative format and sizing for the blue card. This is done through the national Advancement Team.

Image copyright Getty Images Some 400 climate-change protesters have been arrested after several days of demonstrations in central London. Despite the arrests, the Extinction Rebellion activists say they will continue with their protests, which have caused traffic gridlock in places and disruption to parts of London's public transport network. Policing of the protests has been criticised by some, but what powers do officers have? And how have their tactics changed from previous protests? What powers are the police using? The majority of the arrests by the Metropolitan Police Service, have been for suspected breaches of a Section 14 Notice of the Public Order Act 1986. A "Section 14" is a "direction" that allows the police to impose conditions on a static protest - in other words, one where a group of people gather in one place and stay put - rather than marching somewhere. Media playback is unsupported on your device Media caption Who are Extinction Rebellion? In relation to the Extinction Rebellion protests, the Met has stipulated that activists should move to nearby Marble Arch if they wish to continue protesting. If individuals fail to follow the police instructions, they can be arrested and face the risk of criminal prosecution. The most likely sentence for breaching a Section 14 notice, especially if the convicted person was peaceful and has no recent convictions for similar offences, is a conditional discharge - where a court convicts but does not impose a sentence on the condition a further offence isn't committed. When are the police allowed to issue a Section 14 notice? Section 14 notices are quite rarely used, but they can be when police have advance notice of a big protest. Police officers can issue them if they "reasonably believe" that the protest "may result in serious public disorder, serious damage to property or serious disruption to the life of the community". They can also be be used if the police believe the purpose of the protest is to intimidate others. Image copyright Getty Images Image caption Met Commissioner Cressida Dick briefed the home secretary about the protests But protesters need to be aware that police have issued a Section 14 direction in respect of the protest before they can be arrested for a suspected breach. Police will often try to communicate that a Section 14 notice is in force through various means - for example, handing out leaflets, using social media or using a megaphone to address a crowd. But why isn't everyone being arrested? Officers have discretion over whether to make arrests. The Met said in a statement: "These are peaceful protesters; while disruptive their actions are not violent towards police, themselves or other members of the public. We are looking at other tactics such as tighter police cordons." In any protest, the police might take the view that attempting to carry out mass arrests could cause trouble - so by "waiting it out" they hope that the demonstrations might gradually wind down of their own accord. There is also a question of resources. Image copyright Getty Images Image caption One activist, Farhana Yamin, glued her own hands to the pavement outside Shell's headquarters One police officer at one of the protest spots told the BBC their tactic was to move people one-by-one. The officer added that while everyone was "technically arrestable", there were not enough officers to carry out mass arrests. Furthermore, there have been reports of some of those arrested being sent as far as Bromley, south London, due to a lack of police cells in central London. According to a 2017 Freedom of Information request, the Met has 683 cells across 27 custody suites. In response to concerns over capacity, the Met said it had "contingency plans" should cells become too crowded but refused to discuss its plans further for "operational reasons". How do the police tactics compare with previous protests? On May Day in 2000, the Metropolitan Police Service was accused of losing control of Westminster and the West End during a major anti-capitalism protest. There were running battles between officers and anarchist groups which led to significant damage. The following year, when some of the same groups returned, the Met responded in force. Thousands of officers used a tactic known as "kettling" to corral crowds until they were trapped in what amounted to a human pen at the very same Oxford Circus location now occupied by Extinction Rebellion. Some of the protesters turned violent - but many others were entirely peaceful and angry at being "kettled" and unable to leave. Police identified the "trouble-makers" and sent in snatch squads to arrest them. It took seven hours for the city to return to normal and the crowds to be dispersed. Kettling became the subject of an 11-year legal battle, which resulted in the European Court of Human Rights declaring that it was lawful. That landmark judgement meant the police could use this exceptional containment power if there was a genuine risk of violence. In 2005, more than 350 people were arrested at the Gleneagles G8 leaders' summit in Scotland following violent clashes between protesters and police. Extinction Rebellion is different because there is no threat of violence - just occupation and disruption, aided by DJs, mobile curry stalls and the occasional juggler. Image copyright Getty Images Image caption The 2019 Extinction Rebellion protest has focused on occupation and disruption So do the police have other tactics? In 2009, demonstrations against the UK meeting of G20 world leaders saw both peaceful and sporadically violent protests in London. Climate Camp, one of the peaceful demos, was similar to Extinction Rebellion. It was a sudden, no-notice occupation of a part of the City of London - home of the Bank of England and other financial institutions - which police couldn't stop. By early evening, the Met's commander on the ground was concerned there was going to be trouble - not from the Climate Campers, but from the dispersal of another nearby group which he feared could infiltrate the peaceful street occupation. His officers issued Section 14 orders telling the Climate Campers to depart by 19:00. When they failed to leave, he contained them while his officers dealt with the other group. What happened next deepened the controversy. Lines of police with riot shields and batons raised pushed through the Climate Camp to break up the occupation. Image copyright Getty Images Image caption The police "kettling" tactic was used during the 2009 G20 protests Eventually, after another mammoth legal battle, the courts backed the Met's decision-making that night, saying the commander was entitled to judge that violence would have been imminent had he not used the containment and dispersal strategy. It's worth pointing out that while the Met has won these two key legal cases over kettling and dispersal, a far more serious incident happened during the G20 protests. Ian Tomlinson, an entirely innocent newspaper vendor died after he was pushed over by a police officer as he tried to head home. An inquest jury found he had been unlawfully killed and the officer was later tried - but acquitted of manslaughter. No commander in charge of policing protests wants to go through that experience again. And that's why the right approach to containing and potentially dispersing a disruptive street demonstration on the scale of Extinction Rebellion comes down not only to the law and police resources, but the mood on the ground and some intelligence about the intentions of the demonstrators. Read more from Reality Check Send us your questions Follow us on Twitter

Experiences and problems in the disinfection of fibre endoscopes. The problem of the disinfection of fibre endoscopes is to achieve a good efficiency together with easy handling. There is not yet a disinfectant which shows a sufficient germicidal effect after a short time of contact with the fibre endoscope and which is harmless to the patient and medical staff. This paper deals with a two-stage disinfecting procedure in which the instrument undergoes a short intermediate disinfection using polyvidoneiodine (Betaisodona) between two investigations and a main disinfection at the end of a series of investigations using glutardialdehyde (Cidex). The measures lead to a sufficient reduction of the number of germs. The importance of the disinfection of all parts of the endoscopes, especially the channels and the accessories (water bottles etc.) is pointed out.

122 B.R. 527 (1990) In re GRAHAM SQUARE, INC. Bankruptcy No. 690-01182. United States Bankruptcy Court, N.D. Ohio. July 11, 1990. *528 John P. Van Abel, Amerman, Burt & Jones, Canton, Ohio, for debtor. Stephen D. Thompson, Black, McCuskey, Souers & Arbaugh, Canton, Ohio, for Union Nat. Bank of Pittsburgh. Russ Kendig, Krugliak, Wilkins, Griffiths & Dougherty, Canton, Ohio, for Central Trust Co. of Northeastern Ohio. MEMORANDUM OF DECISION JAMES H. WILLIAMS, Chief Judge. The court is presented with a motion for use of cash collateral filed by the debtor and debtor in possession, Graham Square, Inc. The debtor seeks to use the rental income secured to The Union National Bank of Pittsburgh (UNB) and The Central Trust Company of Northeastern Ohio (Central Trust) to operate its business as it attempts to reorganize. Both UNB and Central Trust strenuously contest the use of the rental income. A hearing was held pursuant to Bankruptcy Rule 4001(b), upon the conclusion of which the court took the matter under advisement. The court has jurisdiction in this matter by virtue of 28 U.S.C. § 1334(b) and General Order No. 84 entered in this district on July 16, 1984. This is a core proceeding under 28 U.S.C. § 157(b)(2)(M). This Memorandum of Decision constitutes the court's findings of fact and conclusions of law pursuant to Bankruptcy Rule 7052. FACTS The debtor obtained a construction loan from UNB in May, 1988, to build and operate a strip style shopping center located on Graham Road in the city of Stow, Ohio. The original amount of the loan was $8,550,000. At present, approximately $8,056,000 plus interest is due and owing UNB. As security for the loan, UNB possesses, among other things, a mortgage on the premises and an assignment of the rental income. In September, 1989, the debtor obtained an additional construction loan in the amount of $1,000,000 from Central Trust to construct several out-parcel buildings in front of the shopping center. Currently there is due Central Trust approximately $925,632 plus interest. Central Trust also possesses a mortgage on the real estate and an assignment of the rental income to secure the indebtedness. The debtor filed for relief under Chapter 11 of the Title 11 of the United States Code on June 6, 1990. It was stipulated at the hearing on the motion for the use of cash collateral that the debtor was in default of its loan obligations to UNB and Central Trust prior to its filing bankruptcy. At the hearing, the debtor presented income projections from rental income and common area maintenance fees for each of the months of July and August, 1990, of $68,826. This figure is based upon 89.8% occupancy of the shopping center. For these same months, the debtor estimates operating expenses of $11,649.72, leaving a balance of available cash each month of $57,176.28. The debtor has offered to pay this sum to UNB and Central Trust as adequate protection for the use of rent proceeds. The debt service to UNB and Central Trust at 11% interest, the original contract rate on each of the mortgages, is $82,332 per month. However, both loans provide that upon default the interest rate would increase to the prime rate plus 3%, or 14% interest.[1] At this level, the debt service is $99,587 per month. The debtor in its projections estimates that by December, 1990, it will be receiving $98,059.85 in monthly rental income and common area maintenance fees. Deducting projected monthly operating expenses of $14,709.14, there will remain a balance of $83,350.71. This increase in rental income from July to December reflects an expected increased occupancy rate of from 89.9% to 100%. However, to obtain the 100% occupancy status, an additional $404,000 *529 is needed to complete various portions of the project. The debtor anticipates obtaining an additional loan at 11.5% interest. Such an additional loan and the present obligations to UNB and Central Trust will create a monthly debt service load of approximately $86,202, assuming only an 11% interest rate. Both UNB and Central Trust argue that the debtor's offer of adequate protection, the payment of the debt service, is inadequate as the monthly payment falls short of the amount due per month, even at the 11% rate. Further, even if the debtor can achieve 100% occupancy, there remains a negative net cash flow when interest is paid on the three (3) loans that will, in the debtor's view, be necessary to accomplish its goals. DISCUSSION 11 U.S.C. § 363(c)(2) provides: (2) The trustee may not use, sell, or lease cash collateral under paragraph (1) of this subsection unless — (A) each entity that has an interest in such cash collateral consents; or (B) the court, after notice and a hearing, authorizes such use, sale, or lease in accordance with the provisions of this section. Section 363(e) further provides: Notwithstanding any other provision of this section, at any time, on request of an entity that has an interest in property used, sold, or leased, or proposed to be used, sold, or leased, by the trustee, the court, with or without a hearing, shall prohibit or condition such use, sale, or lease as is necessary to provide adequate protection of such interest. The threshold issue for the court to determine is whether the rental income that the debtor seeks to use is cash collateral within the meaning of Section 363(a).[2] Whether or not a mortgagee is entitled to rents and profits derived from the mortgaged property must be determined with reference to applicable state law. Butner v. United States, 440 U.S. 48, 99 S.Ct. 914, 59 L.Ed.2d 136 (1979); Saline State Bank v. Mahloch, 834 F.2d 690 (8th Cir.1987); In re Pfleiderer, Memorandum of Decision, Case No. 686-00198 (Bankr.N. D.Ohio, October 16, 1987). Ohio law has been summarized in 69 O.Jur.3d Mortgages, § 151 (1986): A mortgage of real property does not per se operate as a specific pledge of the rents and profits therefrom. To have that effect the mortgage must expressly include them. But even though a mortgage contains a pledge of rents and profits, and even though the condition is broken, the mortgagor is entitled to them so long as he retains possession. To be entitled to the rents and profits specifically pledged, the mortgagee must have taken possession of the premises, or must have taken some action, such as the appointment of a receiver, to reduce the rents and profits to possession. (footnotes omitted) (emphasis added). It appears to the court, from its review of the entire file, the papers relating to the motion for use of cash collateral and the statements made at the hearing, that neither UNB nor Central Trust took any affirmative steps prior to the debtor's bankruptcy filing to secure the rents in any fashion such as described above. Additionally, neither bank has filed any type of motion in this court to sequester the rents or a notice pursuant to Sections 546 or 552 of the Bankruptcy Code to exercise their rights to obtain possession of the rental income. Accordingly, the court must find that the rents and common area maintenance fees do not presently qualify as "cash collateral." As a result, the requirement of Section *530 363(e) that adequate protection be given by the debtor for the use of this rental income is not applicable. See, In the Matter of Hamlin's Landing Joint Venture, 77 B.R. 916 (Bkrtcy.M.D.Fla.1988); Saline State Bank v. Mahloch, supra. Graham Square, Inc., is entitled, therefore, for the present time, to use the rents and common area maintenance fees in the ordinary course of its business, subject of course, to the rules of this court and the provisions of the Bankruptcy Code. An order in accordance herewith shall issue. NOTES [1] The court assumes, from the testimony and statements made at the hearing that the default provisions in Central Trust's loan documents are the same as those of the UNB loan. The court, however, was never presented with any loan documents regarding Central Trust's loan to Graham Square. [2] 11 U.S.C. § 363(a) provides: In this section, "cash collateral" means cash, negotiable instruments, documents of title, securities, deposit accounts, or other cash equivalents whenever acquired in which the estate and an entity other than the estate have an interest and includes the proceeds, products, offspring, rents, or profits of property subject to a security interest as provided in section 552(b) of this title, whether existing before or after the commencement of a case under this title.

DoD News News Article Cohen Says U.S. Troops Are America's Ambassadors TOKYO, April 9, 1997 – One misdeed can destroy countless good works, U.S. Defense Secretary William S. Cohen told American troops stationed here. "Wherever you are stationed, you are carrying the honor of the United States," Cohen said. "Be always aware you are an ambassador as well as a soldier, sailor, airman or Marine." Cohen talked with airmen at Yokota Air Base and sailors at Naval Fleet Activities Yokosuka April 8 and with Marines at Marine Corps Air Station Iwakuni April 9. The defense secretary said the nation is indebted to its men and women in uniform for their service to the nation. "Many times you tend to forget how important the role is that each of you plays in our foreign and defense policies," he said. "You are a part of our strategy of engagement. By virtue of your presence, you are helping the United States carry out its role in stabilizing the entire region." The United States is a global superpower, Cohen said, but even the United States can't go it alone. "We have to have great allies like Japan," he said. Some in Japan, however, oppose continuing U.S. presence on their land, he said. Japanese support for U.S. forces in Japan took a severe blow last fall after two U.S. servicemen raped an Okinawan schoolgirl. The incident fueled the flames of other complaints against U.S. presence in Japan. "Any time you have an incident involving a member of the military in a country in which we are guests, it does create domestic concern," he said. "That's understandable. The difficulty is, there's so much good being done by our military day in and day out. Overwhelming effort is being devoted to community affairs, helping the local community with public works projects -- but it can all be overshadowed and undercut by one bad incident." The public tends to focus upon that one negative aspect, as opposed to all the good being done on day to day, he said. U.S. relations with Japan, however, have been and continue to be so strong they will not be unraveled by any one incident, Cohen said. While some object to the U.S. presence, Japanese Prime Minister Ryutaro Hashimoto has repeatedly assured U.S. officials he is committed to providing a stable, legal framework for continuing American forward-basing in Japan. The Okinawa case sparked U.S. efforts to reduce the intrusive footprint of U.S. military forces and activities on local communities. U.S. officials took steps to reduce noise and traffic, returned about 20 percent of the land formerly used for training and are working to relocate helicopter activities to an offshore base. Cohen asked U.S. troops to do their part in maintaining good relations by always being on their best behavior and acting responsibly. "Everything you do reflects upon our country wherever you're deployed," he said. "... On duty or off duty, you are having an impact and reflecting an image of who or what the United States is."

import React from 'react'; import { Link } from 'react-router-dom'; import theme from './Avatar.css'; import AvatarImage from '../image/avatar.gif'; function Avatar() { return ( <div className={theme.container}> <div className={theme['account-navs']}> <h2 className={theme['flex-1']}>头像</h2> <h2 className={theme['flex-1']}> <Link to="/account/information">信息</Link> </h2> </div> <img src={AvatarImage} alt="" width="100%" /> <div> <a href="/">首页</a> </div> </div> ); } export default Avatar;

History Killavullen GAA club is one of the oldest official clubs in North Cork, with the club being founded on the 15th of February 1888 in the parish under the auspices of the Gaelic Athletic Association, and with the exception of a North Cork Junior hurling championship that they came close to annexing in 1929; being narrowly defeated by Churchtown, in the final, the club toiled away without much initial success. However they had a proud tradition of promoting our Gaelic games, and the club is grateful to neighbouring parishes, particularly Shanballymore for inviting them to participate in hurling tournaments which sometimes brought success, but more importantly provided competitive games in preparation of championship fixtures. The prize for winning such tournaments was often the "Cut" of a suit, or gold watches, which in poorer times, were great incentives to do well. Persistence paid off and in 1962, the club contested the final of the Avondhu Novice hurling Championship, and won out this competition in 1968. In between, with the immense help of Johnny Beechinor, a native of Newmarket, who had come to settle in the parish Killavullen made a break through in Football in 1963, which was the clubs first football title, and went out to further glory in 1979, when they were crowned North Cork Junior (B) Hurling League Champions. Since the revival of the juvenile club in 1978, more success has been recorded in football rather than hurling, and with the growth in population of the parish, and trojan work been done within the juvenile structure of the club, the basis for the modern day success of the club has been laid. The club won North Cork Under-12 championships in 1980 and 1981 and won Avondhu championships at all age levels, including minor titles in 1986 and 1992. Indeed in 1992, we won North Cork Championships in Under 14 B hurling; minor football and we were league champions in minor hurling and junior football, which really was a historic year. In 1993 Killavullen won the North Cork Junior (B) Football Championship, won the divisional league title and contested, albeit unsuccessfully, the county Junior (B) Football Championship, which reflects the work done at the Juvenile level since 1978 by Johnny Beechinor, Owen O Neill, George Lane, Mattie Dorgan, Joe Taylor, Tim Nagle and other club members. The underage success of the club in the middle years on the 1980's came to fruition when the club contested the Junior Football ranks in North Cork, and won the league championship in 1997, before annexing their first North Cork Junior football championship in 1998. That same year, Killavullen were defeated in the county final by Newmarket. The team again won the North Cork junior Football Championship in 1999, before winning that same title again in the year 2000, and landing the County title, defeating Kiskeam in the final. This gained the club promotion to the intermediate grade. The Tadgh Crowley Cup was won in the year 2002, in our second year at the intermediate ranks, before the Tom Creedon cup was won in 2003, and again claimed in 2005. In the intervening year, Killavullen won the Munster Intermediate Football League in 2004. The club came within the width of a post of winning the Premier Intermediate Football title in 2007, being defeated by Mallow in a very closely fought final. Off the playing pitch, the club, to commemorate the centenary of the founding of the GAA, in 1984 the club were very grateful to the Browne Family from Killavullen from whom they purchased land near the village and since then have developed then into a magnificent playing field. Dressing Rooms were added in time, and recently, a second playing pitch and gym facilities have been added to the clubs portfolio. The playing pitch must surely be one of the most picturesque in all of Cork with it being on the banks of the Blackwater River, though it is subject to the vagaries of flooding. However careful sitting of the dressing rooms means that these do not flood. The club members are very grateful to the farmers who gave their fields down through the years, particularly the Lucey family, Ballymacmoy who provided the club with a playing pitch for over 30 years. Prior to that venue, games were played in the Inch Field of the property that is now part of the Nano Nagle Centre in Ballygriffin. Aside from these activities, Killavullen has been competing with success in Scor na n'Og and in Scor na bPaisti, and praise must be heaped upon the respective principals, both present and former, in the two schools in the parish, Killavullen National School, and Ballygown National School for all their help down through the years in this regard. Killavullen GAA Roll of Honour Year Titles Won 1963 North Cork Novice Football Championship 1968 North Cork Novice Hurling Championship 1974 North Cork Junior (B) Hurling Championship 1979 North Cork Junior (B) Hurling League Champions 1981 North Cork Junior (B) Hurling Championship 1986 North Cork Junior (B) Football Championship 1987 North Cork Under -21 (B) Hurling Championship 1992 North Cork Junior (B) Football League Champions 1993 North Cork Junior (B) Football Championship North Cork Junior (B) Football League Champions 1994 North Cork Under -21 (B) Football Championship 1995 North Cork Under -21 (B) Hurling Championship 1997 Winners of the North Cork Junior A Football "Avondhu" Cup1 North Cork Junior (A) Football League Champions - Fitzgerald Cup Winners of the North Cork Junior A Hurling "O'Leary" Cup2 1998 North Cork Junior (A) Football League Champions - Fitzgerald Cup North Cork Junior (A) Football Championship 1999 North Cork Junior (A) Football League Champions - Fitzgerald Cup North Cork Junior (A) Football Championship 2000 North Cork Junior (A) Football League Champions - Fitzgerald Cup North Cork Junior (A) Football Championship Cork County Junior (A) Football Championship 2001 Winners of the North Cork Junior A Hurling "O'Leary" Cup2 North Cork Under -21 (C) Football Championship

Battle of Cumberland Church The Battle of Cumberland Church was fought on April 7, 1865, between the Union Army's II Corps of the Army of the Potomac and the Confederate Army of Northern Virginia during the Appomattox Campaign of the American Civil War. After the Battle of Sailor's Creek on April 6, 1865, surviving Confederate troops of Lieutenant General Richard H. Anderson and Major General John B. Gordon headed for the High Bridge, a double-deck structure with a railroad bridge on top and a lower wagon road bridge over the Appomattox River to cross to the north side of the river and continue their retreat to the west. The Confederates intended to destroy the bridge, which they had fought to save the day before in the Battle of High Bridge, but through mistakes and delays did not start to do so until Union troops of Major General Andrew A. Humpreys's II Corps began to arrive at the bridges. After a second smaller Battle of High Bridge, the Union soldiers at the scene kept the railroad bridge from total destruction and saved the wagon bridge in shape for use. Humphreys's troops pursued the last division in the line of march, the division of Major General William Mahone to Cumberland Church about to the west and north of Farmville, Virginia where the Confederates began to fortify the high ground around the church. Soon, Lieutenant General James Longstreet with the entire remaining Confederate infantry moved up from Farmville, Virginia to join Mahone, burning the railroad and wagon bridges at Farmville after them. The Confederate cavalry and one infantry brigade that were left behind had to ford the river nearby. After some fighting in transit to Cumberland Church in which Brigadier General Thomas A. Smyth was mortally wounded, Humphreys ordered his two divisions, which reached the church soon after Mahone, to attack the Confederate line. Finding the Confederate position too strong to take, Humphreys recalled Brigadier General Francis Barlow's division, which had been heading directly to Farmville in pursuit of Gordon's corps, and sent a message to Army of the Potomac commander Major General George Meade that Lee's whole army was north of the river. Humphreys suggested that reinforcements be sent in order to engage the full Confederate Army. Neither Humphreys nor Meade knew until later that the Confederates had destroyed the bridges at Farmville and no reinforcements could reach Humphreys that afternoon. Mistaking a separate nearby cavalry engagement as the arrival of reinforcements engaging with Lee's infantry, Humphreys ordered another futile attack, which was repulsed by the Confederates. Colonel (Brevet Brigadier General) John Irvin Gregg was taken prisoner by the Confederates in the cavalry engagement. Since no reinforcements could quickly reach Humphreys and night was approaching, Humphreys ordered no further attacks. The II Corps had at least 571 casualties in the Battle at Cumberland Church and the cavalry had 74 in what is sometimes called the Battle of Farmville. Confederate casualties are unknown but have been estimated to be about half the number of the Union casualties. The National Park Service gives the number of Confederate casualties as 255. Realizing that the Union forces could close in on his men at Cumberland Church, General Robert E. Lee withdrew his army in another night march to the west at about 11:00 p.m. Although the Confederates held back the Union Army at Cumberland Church and had fewer casualties, they were delayed in their march, which helped other Union forces south of the Appomattox River to get past them to cut them off at the Battle of Appomattox Court House. Background Retreat from Sailor's Creek Because of the need to reduce steep grades in the vicinity of Farmville, the South Side Railroad crossed from the south side of the [Appomattox River to the north side over the High Bridge east of Farmville, about from Rice's Station. From High Bridge, the railroad ran parallel to the north bank of the river until it reached Farmville, Virginia, to the southwest. At Farmville, the railroad crossed back to the south side of the river and proceeded to the west. The railroad bridge at Farmville also had an adjacent wagon bridge. After the Battle of Sailor's Creek, Confederate Lieutenant General Richard H. Anderson and Major General John B. Gordon and the defeated survivors of their corps headed for the South Side Railroad's High Bridge and the wagon bridge below it to cross to the presumed relative safety of the north side of the Appomattox River. After fighting unsuccessfully to destroy the bridges after themselves, Gordon and his men followed the route of the railroad over the High Bridge, along the north bank of the river, and back to the south side of the river over the bridges at Farmville to obtain rations and then immediately return to the north side of the river from Farmville. Longstreet moves to, evacuates Farmville Lieutenant General James Longstreet's corps, which had reached Rice's Station on the morning of April 6, had moved past Sailor's Creek because they were at the head of the march and missed the battle there. After the minor Battle of Rice's Station on April 6, in a night march, Longstreet's corps slipped away from Major General John Gibbon's XXIV Corps of Major General Edward Ord's Army of the James, which had approached the Confederate lines at Rice's Station during afternoon. Longstreet's corps stayed south of the river and headed west for Farmville during the night. Confederate cavalry had been covering Longstreet's march but after crossing Bush River, Longstreet placed Alfred M. Scales's brigade (under the command of Colonel Joseph H. Hyman) of Major General Cadmus Wilcox's division as the rear guard. Longstreet's troops began to arrive at Farmville, under close pursuit by Union Major General George Crook's cavalry division, at about 9:00 a.m.on April 7. Gibbon's XXIV Corps of Ord's Army of the James, reinforced by Brigadier General William Birney's second division of the XXV Corps, followed closely behind Crook's cavalry. Rations were promptly issued to Longstreet's men at Farmville but the soldiers were told to march to the north side of the river to begin their meal preparations. When Confederate Commissary General Isaac M. St. John heard gunfire just outside town, he sent the ration trains west on the South Side Railroad to Pamplin's Depot or Pamplin's Station in an effort to prevent their capture. Sheridan's cavalry got to Pamplin's Station first and captured 3 engines and the rolling stock with the supplies. The Confederate provost guard and their prisoners were sent down the road without rations. Some rations were passed out to troops on the march out of Farmville and the empty wagons from which they were distributed were burned. Longstreet had to move his forces quickly from Farmville to the Cumberland Church vicinity to avoid Union forces closing in the town. Second Battle of High Bridge Major General Andrew A. Humphreys began to march the II Corps west from the Sailor's Creek Battlefield at 5:30 a.m. on April 7. When the leading unit of the Second Division of the II Corps under the command of Brigadier General Francis C. Barlow reached High Bridge at about 7:00 a.m., they found that the Confederates had blown up the redoubt at the bridgehead on the northern (and western) side of the bridge and were just starting to burn the railroad bridge and lower wagon and foot bridge. A mistaken or misunderstood order to Lieutenant General Anderson's guards not to allow anyone to pass kept Major General William Mahone's division from crossing until the order were clarified. Then more time was spent getting the required order to destroy the bridges to the Confederate engineers commanded by Colonel Thomas M. R. Talcott. This permitted the lead units of the II Corps to reach the bridge just as the Confederates were attempting to destroy it. The ensuing fight to save or destroy the bridges was the second Battle of High Bridge, following the battle the previous day when a Union raiding force tried to destroy the bridge to keep the Confederates on the south side of the Appomattox River and the Confederates successfully fought to save it, killing or capturing the Union force. On the morning of April 7, the Union troops led by the 19th Maine Volunteer Infantry Regiment under Colonel Isaac R. Starbird were able to prevent the total destruction of the railroad bridge, though it had been made unusable. They were able to preserve the wagon bridge for passage of the II Corps in their pursuit of Gordon, Anderson and Mahone. Colonel William A. Olmsted's First Brigade of Barlow's division and Brigadier General Thomas Alfred Smyth's Third Brigade, joined by Brigadier General Nelson Miles's division drove off a counterattack by Mahone's troops who were trying to complete the bridges' destruction. Mahone's division was in a defensive line with two redoubts on high ground on the north bank of the river when the II Corps began to arrive at the bridges but they moved to northwest when they could not destroy the bridges and prevent Humphreys's corps from crossing the river. Gordon's corps moved up river along the railroad bed toward Farmville. Humphreys, with Miles's and Brigadier General Regis de Trobriand's divisions, followed Mahone while Barlow's division followed Gordon. Pursuit; Smyth mortally wounded Gordon's corps and Barlow's division skirmished from High Bridge to the northern outskirts of Farmville. An advance Union party of 103 men from the 7th Michigan Infantry Regiment and the 59th New York Infantry Regiment were taken prisoner by the 9th (1st) North Carolina Battalion of Sharp Shooters of Brigadier General John A. Walker's division. Barlow's division caught up with part of Gordon's division and cut off and burned 135 wagons along the road. During this attack, Brigadier General Smyth was mortally wounded by a sniper shot as he rode with his skirmish line within of Gordon's rear guard near Farmville. Colonel Daniel Woodall of the 1st Delaware Infantry Regiment assumed command of the Third Brigade. Smyth's men were stunned by his fall and refused to move for several minutes as Gordon's men began to withdraw. Colonel William A. Olmsted's brigade moved to support Smyth's brigade but about 100 of his men were captured as Confederate rear guards rushed Smyth's hesitating troops. Barlow's men ceased the pursuit altogether as they reached the vicinity of the edge of Farmville north of the river. Union troops take Farmville; Confederates withdraw north While the VI Corps pursued Mahone's division and Gordon's corps north of the Appomattox River, Ord and Gibbons with the XXIV Corps and a brigade of cavalry from Crook's division engaged with Longstreet's rear guard, Scales's brigade and the cavalry, at Briery Creek, just short of Bush River. During this encounter, the other men of Cadmus Wilcox's division received rations at Farmville and moved to the north side of the river. The 1st Maine Cavalry Regiment, armed with 16-shot Henry repeating rifles, led the attack on the Confederate rear guard and the Union forces drove them toward Farmville. Wilcox led his division back across the river to protect the rear guard and the trains with rations just before they moved out toward Pamplin Station. Wilcox then took his men back to the north side of the river cross and Alexander's men started to burn the bridge. Since the Confederates burned the bridge before all their troops were across, Major General Thomas L. Rosser's and Colonel Thomas T. Munford's cavalry divisions and Brigadier General John Bratton's infantry brigade had to cross the river further upstream under fire from Crook's lead units. Crook's cavalry came into town and captured stragglers while the XXIV Corps was only away when the bridge was burned. By the time the main body of Crook's cavalry reached Farmville, most of the Confederates had crossed to the north side of the Appomattox River and burned the bridges leading from the main part of town to the north side of the river. The cavalry ford that Crook's men soon discovered was too deep for the infantry to cross. By 1:30 p.m., Union troops of the Army of the James and Major General Horatio Wright's VI Corps of the Army of the Potomac occupied Farmville while Crook's cavalry found a spot just northwest of town to ford the Appomattox River. Soon thereafter, Union Army General-in-Chief Lieutenant General Ulysses S. Grant and his staff arrived in Farmville, making the Randolph House (Prince Edward Hotel) his headquarters. When all of the Confederates crossed the river from Farmville, they formed a line of battle on Cumberland Heights overlooking the town. Artillery was set up to fire on the Union troops as they moved into Farmville. After a short period of time, the Confederates withdrew to the north to join Mahone's division around Cumberland Church. At this location, First Corps (Longstreet's) artillery chief Edward Porter Alexander told General Lee that the route of march to Appomattox Station was shorter on the south side of the river and they should have stayed there. About this time, Lee also learned that Union troops had followed the Confederates across High Bridge and he could not break away from them on the north side. Having burned the bridges at Farmville, Lee's troops could not cross back to the south side of the river there. Battle at Cumberland Church At Cumberland Church, about northwest of High Bridge and north of Farmville, Mahone's Confederates arrived first and controlled the high ground around the church. They entrenched on top of a long slope of open ground, covering the stage and plank roads to Lynchburg. Lieutenant General Anderson with little more than a few stragglers and a battalion of artillery headed west when Mahone told him that Humphreys was approaching on the same side of the river with them. Union II Corps commander Major General Humphreys, with Brigadier Generals Miles's and de Trobriand's divisions, arrived near the Lynchburg stage road at Cumberland Church about 1:00 p.m. and became engaged not just with Mahone's division, but with the entire remaining Confederate Army of Northern Virginia. As the Union force approached, Lieutenant Colonel William T. Poague's 16-gun artillery battalion opened fire on them. Soldiers from Miles's division temporarily captured the guns of Captain Arthur B. Williams's North Carolina Battery but troops of Major General Bryan Grimes's division recaptured them. Despite evident confusion and disorganization, especially among the survivors of commands hard hit at Sailor's Creek and Gordon's men who had not arrived at Farmville to get rations from the trains that had been there, the Confederates began to prepare defenses at Cumberland Church. As Humphreys's men approached, Gordon's men caught up with Mahone's and the entire effective Army of Northern Virginia was near Cumberland Church. Historian William Marvel stated that with artillery, about 12,000 effective Confederate men held the position at Cumberland Church. They immediately began to prepare a line of thin breastworks in a fish hook formation when they reached the hill. Mahone's division held the right of the line with Poague's battery on its right. Gordon's corps got in line to the right of Poague and Longstreet's corps moved into position next to Gordon. Some of Major General Fitzhugh Lee's cavalry moved to protect Mahone's left flank. Other Confederate cavalrymen were in the rear supported by Major General Henry Heth's division. Humphreys had about the same number of men, 12,000, but did not have the high ground, fortifications or interior lines of communication that the Confederate defenders had. A heavy Union skirmish line moved forward but Miles and de Trobriand found they were unable to make a successful flank attack against Mahone along the Jamestown Road. When Humphreys realized that the Confederates held a strong position in force, he recalled Barlow's division, which had been marching along the South Side Railroad following Gordon's corps toward Farmville and to guard against a Confederate move toward Danville, to rejoin the other II Corps divisions. At the same time, Humphreys sent a message to Army of the Potomac commander Major General George Meade telling him that Lee's entire army was at Cumberland Church and suggesting that a corps should come up from Farmville and attack the Confederates from the other side. Neither Humphreys nor Meade yet knew that the Confederates had destroyed the bridges at Farmville and infantry could not ford the swollen river there. After the initial Union foray was repulsed, Miles moved to his right and tried again to get his men around Mahone's left flank. Humphreys heard gunfire from the direction of Farmville and saw the Confederates shorten their right flank, so he thought the VI Corps had crossed the river and were starting their attack. He had Miles send Colonel George W. Scott's brigade to turn the Confederate left flank. Scott's men had to move uphill through some rolling hills, valleys and ravines to reach the Confederate fortifications. Despite heavy fire, including canister, and the movement of Brigadier General William Forney's Alabama brigade to the left to protect the flank, some of the Union force made it to the Confederate line. The 5th New Hampshire Infantry Regiment and their flag were captured when they got there. Reinforcements from Brigadier General George T. "Tige" Anderson's brigade of Georgians threw back other Union troops of Scott's brigade who were starting to have some success turning the Confederate flank. The counterattack forced Scott's men to withdraw since they were unsupported by the expected attack on the other side of the Confederate line. Mahone sent some of the men on his right to attack the Union skirmish line, but they were driven back twice. As night approached and no Union reinforcements had arrived to join the II Corps, Humphreys made no further attacks on the Confederate line. Barlow's division returned to the II Corps at Cumberland Church only in the evening. Barlow took Miles's place on the right flank of the Union line; Miles moved his division to the center of the line; de Trobriand held the left flank. Cavalry Battle of Farmville The gunfire heard by Humphreys was from Crook's cavalry, which had crossed the river and attacked a wagon train about up the Buckingham Plank Road from Farmville. Colonel (Brevet Brigadier General) J. Irvin Gregg's division led the attack, supported by Brigadier General Henry E. Davies's and Colonel (Brevet Brigadier General) Charles H. Smith's brigades and Lieutenant James H. Lord's Battery A of the 2nd United States Artillery Regiment. The wagon train was accompanied by the survivors of Anderson's (Johnson's) and Pickett's command, hundreds of stragglers, and several artillery batteries. As Gregg's brigade attacked the wagon train, Confederate Major General Thomas L. Rosser and Colonel Thomas T. Munford arrived and sent units of their cavalry divisions against Gregg's dispersed troopers, broke their attack and captured Gregg and some of his men as the others fell back. They were supported by Gordon's infantry having moved down from Cumberland Church. Colonel Samuel B.M. Young reformed the remainder of Gregg's brigade and with Davies's men and Lord's battery attacked the wagons again. By this time, the wagon train also had infantry and artillery support. Another Union attack was repulsed but Confederate Brigadier General William Gaston Lewis was severely wounded and captured. After Confederate troops prevented General Robert E. Lee from personally leading a counterattack by pledging to repulse the Union cavalrymen, they forced Crook's troopers to retreat across the Appomattox River to Farmville. At about this time, Sheridan had recalled Crook to Farmville and directed him to move his division to Prospect Station, or west of Farmville, where his division arrived about midnight. Casualties In the battle up to and at Cumberland Church, the II Corps sustained casualties of 40 killed, 210 wounded, 321 missing, total 571. Confederate casualties are unknown. In the cavalry fight, Crook's cavalry sustained casualties of 11 killed, 33 wounded and 30 missing (including General Gregg, taken prisoner), total 74. Confederate casualties, in what is sometimes separately called the Battle of Farmville, again are unknown. Confederate casualties have been estimated at about half the Union casualties. Aftermath Grant asks for surrender; Lee non-committal Grant wrote to Lee from Farmville at 5:00 p.m. on April 7: Grant's Adjutant-General, Brigadier General (Brevet Major General) Seth Williams, brought Grant's first message to Lee for delivery through the lines by General Humphreys's staff at about 8:30 p.m. Lee showed the message to Longstreet who said: "Not yet." Lee's reply to Grant stated that he disagreed with Grant about the hopelessness of his situation but he also was hoping to avoid "the useless effusion of blood" and asked Grant for the terms he would offer on condition of his army's surrender. Lee's answer came through the lines about one hour after it had been delivered and Williams returned with it to Farmville by a circuitous route over High Bridge. VI Corps cheers Grant, crosses river The VI Corps marched through Farmville on the night of April 7, lit bonfires and cheered General Grant as they passed his headquarters hotel. The corps then crossed the Appomattox River on a newly built footbridge, but waited for a pontoon bridge borrowed from the XXIV Corps for their artillery and wagons to cross after which they camped on the north bank. The XXIV Corps stayed south of the river and camped just to the west of Farmville. The second and third brigades of the XXV Corps were added to Gibbon's XXIV Corps as General Birney had been reassigned to the Petersburg area to command of the newly formed Division of Fort Powhatan, City Point and Wilson's Wharf. The first brigade from Birney's division did not catch up to the rest of the Army of the James until April 10. Cavalry, V Corps move west Brigadier General Ranald Mackenzie's small cavalry division rode from Burkeville (or Burke's Junction) to Prince Edward Court House (now Worsham, Virginia) to cut off the road to Farmville and the Richmond and Danville Railroad at Keysville, Virginia. They skirmished with some outpost troops and took 38 prisoners. By 3:00 p.m., units of Brigadier General (Brevet Major General) George Armstrong Custer's and Brigadier General Thomas Devins's divisions began to arrive at Prince Edward Court House from the Sailor's Creek Battlefield. After dinner, Union Cavalry commander Major General Philip Sheridan ordered them to move west and they eventually camped from Farmville and from Prospect Station. Brigadier General (Brevet Major General) Charles Griffin's V Corps arrived at Prince Edward Court House at about 7:30 p.m. and went into camp on the grounds of Hampden-Sydney College. Lee's night march west While Lee dealt with Grant's message, wagons continued west through roads that had become a quagmire from rain during the day and continued use, causing the death of some teams of mules and the burning of wagons that could not be moved. As the train moved on to the Richmond-Lynchburg Stage Road, conditions improved. Brigadier General Reuben Lindsay Walker had been moving the reserve artillery column independently of the main body of the Confederate army. He had moved through Cumberland Court House the previous afternoon. Having heard about the defeat at Sailor's Creek and the movement out of Farmville, Walker moved to the north, came to the Buckingham Plank Road and moved through Curdsville, Virginia about midday. There, the column moved toward New Store to the southwest and halted for the night, as did the wagon train and hospital train that had been moving from Farmville. Lee knew that his army again had to make a night march from Cumberland Church to get ahead of the Union forces, obtain rations at Appomattox Station and possibly still escape to Danville via Campbell Court House (now Rustburg, Virginia) and through Pittsylvania County, Virginia. Longstreet's corps, with General Lee, moved via the Piedmont Coal Mine Road, also known as the Buckingham Plank Road through Curdsville toward a back road to New Store, Virginia. They covered the retreat of the last wagon train before withdrawing from the Cumberland Church area. Gordon's corps, followed by Fitzhugh Lee's cavalry, moved via the Richmond-Lynchburg Stage Road, a shorter route, toward New Store. Humphreys sent intelligence to Meade at about 3:00 p.m. that Confederate prisoners had said that the Army of Northern Virginia was headed for Lynchburg so Meade and Sheridan had good intelligence about Lee's plan. That night at Cumberland Church, Humphreys's men slept while Lee's force withdrew. On the evening of April 7, Confederate Secretary of War John C. Breckinridge and the Confederacy's quartermaster general, commissary general, chief engineer and various War Department officials rode through Pamplin Station and saw the remainder of the supply trains from Farmville. The commissary general suggested the cars be moved because Union cavalry were not far from Pamplin but Breckinridge refused to do so without being able to notify Lee. On the next day, Crook's cavalry passed through Pamplin on their way to Appomattox Station and destroyed these cars. General Humphreys stated in his book that the delay forced upon the Confederate Army by the engagement at Cumberland Church lost Lee valuable time. It led to the loss of Confederate supplies on April 8 when Custer's division of Sheridan's cavalry arrived at Appomattox Station first and captured the supply trains there. Custer's cavalry captured Confederate supplies and about 25 cannons and took about 1,000 prisoners at the Battle of Appomattox Station. The delay also allowed the Union cavalry and the V Corps and XXIV Corp to reach Appomattox Court House on the morning of April 9 in order to prevent a Confederate breakthrough toward Lynchburg. Footnotes Notes References Calkins, Chris. The Appomattox Campaign, March 29 – April 9, 1865. Conshohocken, PA: Combined Books, 1997. . Humphreys, Andrew A., The Virginia Campaign of 1864 and 1865: The Army of the Potomac and the Army of the James. New York: Charles Scribners' Sons, 1883. . Retrieved March 5, 2015. Longacre, Edward G. The Cavalry at Appomattox: A Tactical Study of Mounted Operations During the Civil War's Climactic Campaign, March 27 – April 9, 1865. Mechanicsburg, PA: Stackpole Books, 2003. . Marvel, William. Lee's Last Retreat: The Flight to Appomattox. Chapel Hill: University of North Carolina Press, 2002. . Salmon, John S., The Official Virginia Civil War Battlefield Guide, Stackpole Books, 2001, . Starr, Steven. The Union Cavalry in the Civil War: The War in the East from Gettysburg to Appomattox, 1863–1865. Volume 2. Baton Rouge: Louisiana State University Press, 2007. Originally published 1981. . Trudeau, Noah Andre. Out of the Storm: The End of the Civil War, April–June 1865. Boston, New York: Little, Brown and Company, 1994. . Warner, Ezra J., Generals in Blue: Lives of the Union Commanders, Louisiana State University Press, 1964, . "CWSAC Battle Summary", National Park Service, Retrieved April 19, 2015. Further reading Calkins, Chris M. Thirty-Six Hours Before Appomattox: April 6 and 7, 1865: The Battles of Sayler's Creek, High Bridge, Farmville and Cumberland Church. Farmville, VA: Farmville Herald, 1980. . Category:Appomattox Campaign Category:Battles of the Eastern Theater of the American Civil War Category:Confederate victories of the American Civil War Category:Battles of the American Civil War in Virginia Category:Cumberland County, Virginia Category:1865 in the American Civil War Category:1865 in Virginia Category:April 1865 events

Q: SpecRun- Tests are not showing in Test Explorer I have a problem where my tests are not being shown in the Test Explorer, I have created a feature file and generated step definitions, I have the following packages installed - SpecFlow-3.1.97 - SpecFlow.Tools.MsBuild.Generation- 3.1.97 - SpecRun.SpecFlow.3-1-0 I have also tried to delete any related to specflow from the %TEMP% folder, It didn't help, also I have different project where I have set up specflow and everything is working fine. Note: without - SpecRun.SpecFlow.3-1-0 package, The test does get discorvered but i get the following error OneTimeSetUp: BoDi.ObjectContainerException : Interface cannot be resolved: TechTalk.SpecFlow.UnitTestProvider.IUnitTestRuntimeProvider('nunit') A: You get this error, because after you removed the SpecRun.SpecFlow.3-1-0 package, you don't have any package more, that configures the used unit test runner. One of the following packages has to be added to your specifications project (the one containing your tests) to select your unit test provider: SpecRun.Runner => for SpecFlow+ Runner SpecFlow.xUnit => for xUnit SpecFlow.MsTest => for MSTest SpecFlow.NUnit => for NUnit From https://specflow.org/2019/updating-to-specflow-3/ You get something discovered in Visual Studio because it has its own test discovery. But this has nothing to do when you execute them. We need stuff from the NuGet packages to execute the scenarios.

--- author: - Elena Kuchina title: | SOME ASPECTS OF\ DEEPLY VIRTUAL COMPTON SCATTERING --- =10000 I am grateful to many people for their help and support during all the years I have been a graduate student. I would particularly like to thank my thesis advisor Ian Balitsky for assigning me such a challenging and rewarding thesis project. I express my deep gratitude to Professor Anatoly Radyushkin for his continuing support during my graduate student career and for giving me many opportunities to learn good physics. It has been my privilege to collaborate with an amazing group of people, the Theory group at Jefferson Lab, who are equal parts teachers, scientists, and tireless workers. Igor Musatov, my colleague and friend, provided invaluable help in nearly every project I worked on. I would like to thank my family and friends for supporting me through what has been a difficult road. I thank my husband Sergey for his willingness to help and my mom for keeping my children safe and sound for me, for their patience, permanent spiritual support and encouragement. I would also like to extend my gratitude to the members of my committee, for taking time to read this manuscript and for their helpful comments and suggestions. Introduction ============ DVCS at small-$x$ ================= Generalized parton distributions ================================ Conclusion\[Summary\] ===================== [99]{}

Q: How to track multiple BackgroundworkerX.Runworkercompleted operations I am trying to use a single handler to cover the end of multiple backgroundworker activities and cannot find a way to get the information about the specific backgroundworker using the backgroundworkercompleted event. My code to catch the event is as below: Private Sub BGx_RunWorkerCompleted(ByVal sender As Object, ByVal e As System.ComponentModel.RunWorkerCompletedEventArgs) Handles BackgroundWorker1.RunWorkerCompleted, BackgroundWorker2.RunWorkerCompleted, BackgroundWorker3.RunWorkerCompleted, BackgroundWorker4.RunWorkerCompleted, BackgroundWorker5.RunWorkerCompleted, BackgroundWorker6.RunWorkerCompleted, BackgroundWorker7.RunWorkerCompleted, BackgroundWorker8.RunWorkerCompleted 'Do work here based on completed Backgroundworker For BG = 1 To 8 If Not DSWorkers(BG).IsBusy Then If DStatus(BG) = -2 Then : DStatus(BG) = -1 : End If End If Next Complete() End Sub There is nothing on the "Do Work Here" section because I do not know how to capture and have been unable to find details of the backgroundworkercompleted event id. Please - any pointers as to how I can identify the specific completed BackgroundWorker A: As with all event handlers, the sender parameter is a reference to the object that raised the event, so you can access the actual BackgroundWorker that has completed its work via that. If you need some data other than that, you assign it to the e.Result property in the DoWork event handler and get it back from the e.Result property in the RunWorkerCompleted event handler. e.Result works for getting data out of the DoWork event handler much as e.Argument works for getting data in. Check this out for some examples of using BackgroundWorker objects, including passing data out using e.Result. You might also like to checkout my own BackgroundMultiWorker class, which basically combines the functionality of multiple BackgroundWorker objects into a single BackgroundMultiWorker object. It identifies each task using a token. EDIT: Here's an example that may help with this issue and your task in general: Imports System.ComponentModel Imports System.Threading Public Class Form1 Private ReadOnly resultsByWorker As New Dictionary(Of BackgroundWorker, BackgroundWorkerResult) Private ReadOnly rng As New Random Private Sub Form1_Load(sender As Object, e As EventArgs) Handles MyBase.Load 'The NumericUpDown is used to select the index of a BackgroundWorker to cancel. With NumericUpDown1 .DecimalPlaces = 0 .Minimum = 0 .Maximum = 9 End With End Sub Private Sub Button1_Click(sender As Object, e As EventArgs) Handles Button1.Click 'Create 10 BackgroundWorkers and run them. For i = 1 To 10 Dim worker As New BackgroundWorker resultsByWorker.Add(worker, New BackgroundWorkerResult) AddHandler worker.DoWork, AddressOf workers_DoWork AddHandler worker.RunWorkerCompleted, AddressOf workers_RunWorkerCompleted worker.WorkerSupportsCancellation = True worker.RunWorkerAsync() Next End Sub Private Sub Button2_Click(sender As Object, e As EventArgs) Handles Button2.Click Dim index = Convert.ToInt32(NumericUpDown1.Value) Dim worker = resultsByWorker.Keys.ToArray()(index) If worker.IsBusy Then 'Cancel the BackgroundWorker at the specified index. worker.CancelAsync() End If End Sub Private Sub workers_DoWork(sender As Object, e As DoWorkEventArgs) Dim worker = DirectCast(sender, BackgroundWorker) 'Do work for a random number of seconds between 10 and 20. Dim period = rng.Next(10, 20 + 1) For i = 0 To period If worker.CancellationPending Then e.Cancel = True Return End If 'Simulate work. Thread.Sleep(1000) Next 'The work was completed without being cancelled. e.Result = period End Sub Private Sub workers_RunWorkerCompleted(sender As Object, e As RunWorkerCompletedEventArgs) Dim worker = DirectCast(sender, BackgroundWorker) Dim result = resultsByWorker(worker) If e.Cancelled Then result.WasCancelled = True Else result.Result = CInt(e.Result) End If Dim workers = resultsByWorker.Keys.ToArray() If Not workers.Any(Function(bgw) bgw.IsBusy) Then 'All work has completed so display the results. Dim results As New List(Of String) For i = 0 To workers.GetUpperBound(0) worker = workers(i) result = resultsByWorker(worker) results.Add($"Worker {i} {If(result.WasCancelled, "was cancelled", $"completed {result.Result} iterations")}.") Next MessageBox.Show(String.Join(Environment.NewLine, results)) End If End Sub End Class Public Class BackgroundWorkerResult Public Property WasCancelled As Boolean Public Property Result As Integer End Class Here is that example reworked to use a single instance of the BackgroundMultiWorker is linked to instead of multiple instances of the BackgroundWorker class. Imports System.Threading Public Class Form1 Private WithEvents worker As New BackgroundMultiWorker With {.WorkerSupportsCancellation = True} Private ReadOnly results(9) As BackgroundWorkerResult Private ReadOnly rng As New Random Private Sub Form1_Load(sender As Object, e As EventArgs) Handles MyBase.Load 'The NumericUpDown is used to select the index of a BackgroundWorker to cancel. With NumericUpDown1 .DecimalPlaces = 0 .Minimum = 0 .Maximum = results.GetUpperBound(0) End With End Sub Private Sub Button1_Click(sender As Object, e As EventArgs) Handles Button1.Click 'Create 10 BackgroundWorkers and run them. For i = 0 To results.GetUpperBound(0) results(i) = New BackgroundWorkerResult worker.RunWorkerAsync(i) Next End Sub Private Sub Button2_Click(sender As Object, e As EventArgs) Handles Button2.Click Dim index = Convert.ToInt32(NumericUpDown1.Value) If worker.IsBusy(index) Then 'Cancel the BackgroundWorker at the specified index. worker.CancelAsync(index) End If End Sub Private Sub worker_DoWork(sender As Object, e As DoWorkEventArgs) Handles worker.DoWork 'Do work for a random number of seconds between 10 and 20. Dim period = rng.Next(10, 20 + 1) For i = 0 To period If worker.IsCancellationPending(e.Token) Then e.Cancel = True Return End If 'Simulate work. Thread.Sleep(1000) Next 'The work was completed without being cancelled. e.Result = period End Sub Private Sub workers_RunWorkerCompleted(sender As Object, e As RunWorkerCompletedEventArgs) Handles worker.RunWorkerCompleted Dim result = results(CInt(e.Token)) If e.Cancelled Then result.WasCancelled = True Else result.Result = CInt(e.Result) End If If Not worker.IsBusy() Then 'All work has completed so display the results. Dim output As New List(Of String) For i = 0 To results.GetUpperBound(0) result = results(i) output.Add($"Task {i} {If(result.WasCancelled, "was cancelled", $"completed {result.Result} iterations")}.") Next MessageBox.Show(String.Join(Environment.NewLine, output)) End If End Sub End Class Public Class BackgroundWorkerResult Public Property WasCancelled As Boolean Public Property Result As Integer End Class

--- abstract: 'Recently, attention-based encoder-decoder (AED) models have shown state-of-the-art performance in automatic speech recognition (ASR). As the original AED models with global attentions are not capable of online inference, various online attention schemes have been developed to reduce ASR latency for better user experience. However, a common limitation of the conventional softmax-based online attention approaches is that they introduce an additional hyperparameter related to the length of the attention window, requiring multiple trials of model training for tuning the hyperparameter. In order to deal with this problem, we propose a novel softmax-free attention method and its modified formulation for online attention, which does not need any additional hyperparameter at the training phase. Through a number of ASR experiments, we demonstrate the tradeoff between the latency and performance of the proposed online attention technique can be controlled by merely adjusting a threshold at the test phase. Furthermore, the proposed methods showed competitive performance to the conventional global and online attentions in terms of word-error-rates (WERs).' author: - '[^1]' bibliography: - 'IEEEexample.bib' title: | Gated Recurrent Context: Softmax-free Attention\ for Online Encoder-Decoder Speech Recognition --- automatic speech recognition, online speech recognition, attention-based encoder-decoder model Introduction {#sec:introduction} ============ the last few years, the performance of deep learning-based end-to-end automatic speech recognition (ASR) systems has improved significantly through numerous studies mostly on the architecture designs and training schemes of neural networks (NNs). Among many end-to-end ASR systems, attention-based encoder-decoder (AED) models [@LAS; @Bahdanau] have shown better performance than the others, such as the connectionist temporal classification (CTC) [@CTC] and recurrent neural network transducer (RNN-T) [@RNN-T], and even outperformed the conventional DNN-hidden Markov model (HMM) hybrid systems in case a large training set of transcribed speech is available [@SOTA-LAS]. Such successful results of AED models come from the tightly integrated language modeling capability of the label-synchronous decoder, supported by the attention mechanism that provides proper acoustic information at each step [@garg2019improved]. A major drawback of the conventional AED models is that they cannot infer the ASR output in an online fashion, which degrades the user experience due to the large latency [@sainath2019two]. This problem is mainly caused by the following aspects of the AED models. Firstly, the encoders of most high-performance AED models make use of layers with global receptive fields, such as bidirectional long short-term memory (BiLSTM) or self-attention layer. More importantly, a conventional global attention mechanism (e.g., Bahdanau attention) considers the entire utterance to obtain the attention context vector at every step. The former issue can be solved by replacing the global-receptive encoder with an online encoder, where an encoded representation for a particular frame depends on only a limited number of future frames. The online encoder can be built straightforwardly by employing layers with finite future receptive field such as latency-controlled BiLSTM (LC-BiLSTM) [@zhang2016highway] and masked self-attention layers. However, reformulating the global attention methods for an online purpose is still a challenging problem. Conventional techniques for online attentionare usually two-step approaches where the window (i.e., chunk) for the current attention is determined first at each decoder step, then the attention weights are calculated using the softmax function defined over the window. As the softmax function is a common solution for representing discrete probability distributions (i.e., attention weights), existing online attentions mainly differ in how they determine the window. Neural transducers [@jaitly2015neural; @sainath2018improving] divide an encoded sequence into multiple chunks with a fixed length, and the attention-decoder produces an output sequence for each input chunk. In the windowed attention techniques [@hou2017gaussian; @tjandra2017local; @merboldt2019analysis], the position of each fixed-size window is decided by a position prediction algorithm. The window position is monotonically increasing in time, and some approaches employ a trainable position prediction model with a fixed Gaussian window. In MoChA-based approaches [@chiu2018monotonic; @miao2019online; @tsunoo2019towards], a fixed-size chunk is obtained using a monotonic endpoint prediction model, which is jointly trained considering all possible chunk endpoints. A common limitation of the aforementioned approaches is that the fixed-length of the window needs to be tuned according to the training data. Merely choosing a large window of a constant size causes a large latency while setting the window size too small results in degraded performance. Therefore multiple trials of the model training are required to find a proper value of the window length, consuming excessive computational resources. Moreover, the trained model does not guarantee to perform well on an unseen test set, since the window size is fixed for all datasets. Although a few variants of MoChA utilize an adaptive window length to remove the need for tuning the window size, such variants induce other problems. MAtChA [@chiu2018monotonic] regards the previous endpoint as the beginning of the current chunk. Occasionally, the window can be too short to contain enough speech content when two consecutive endpoints are too close, which may degrade the performance. AMoChA [@fan2018online] employs an auxiliary model that predicts the chunk size but also introduces an additional loss term for the prediction model. As the coefficient for the new loss needs tuning, AMoChA still requires repeated training sessions. Besides, several recent approaches [@moritz2019triggered; @dong2019cif] utilize strictly monotonic windows. But these methods have a limitation in that the decoder state is not used for determining the window, which means such algorithms might not fully exploit the advantage of AED models, i.e., the inherent capability of autoregressive language modeling. The aforementioned inefficiency in training the conventional online attentions is essentially caused by the fact that the softmax function needs a predetermined attention window to obtain the attention weights, which results in repetitive tuning process of the window-related hyperparameter. To overcome this drawback, we propose a novel softmax-free global attention method called gated recurrent context (GRC), inspired by the gate-based update in gated recurrent unit (GRU) [@cho2014learning]. Whereas conventional attentions are based on a kernel smoother (e.g. softmax function) [@wasserman2006all; @tsai2019transformer], GRC obtains an attention context vector by recursively aggregating the encoded vectors in a time-synchronous manner, using update gates. GRC can be reformulated for the purpose of online attention, which we refer to decreasing GRC (DecGRC), where the update gates are constrained to be decreasing over time. DecGRC is capable of deciding the attention-endpoint by thresholding the update gate values at the inference phase. DecGRC as well as GRC introduces no hyperparameter to be tuned at the training phase. The main contributions of this paper can be summarized as follows: - We propose a novel softmax-free attention method called **Gated Recurrent Context (GRC)**. To the best of our knowledge, GRC is the first attention method that represents discrete probability distributions without a kernel smoother. - We present a novel online attention method, **Decreasing GRC (DecGRC)**, a constrained variant of GRC. DecGRC does not need any new hyperparameter to be tuned at the training phase. At test time, the tradeoff between performance and latency can be adjusted using a simple thresholding technique. - We experimentally show that GRC and DecGRC perform competitive to the conventional global and online attention methods on the LibriSpeech test set. The remainder of this paper is organized as follows. In Section \[sec:backgrounds\], the general framework of attention-based encoder-decoder ASR is formally described, followed by conventional online attention methods and their common limitation. Section \[sec:proposed\] proposes formulations of both GRC and DecGRC and the algorithm for online inference. The experimental results with various attention methods are given in Section \[sec:experiments\]. Conclusions are presented in Section \[sec:conclusion\]. Backgrounds {#sec:backgrounds} =========== Attention-based Encoder-Decoder for ASR {#AED_based_ASR} --------------------------------------- An attention-based encoder-decoder model consists of two sub-modules $\mathrm{Encoder}(\cdot)$ and $\mathrm{AttentionDecoder}(\cdot)$, and it predicts the posterior probability of the output transcription given the input speech features as follows: $$\mathbf{h} = \mathrm{Encoder}(\mathbf{x}),$$ $$P(\mathbf{y}|\mathbf{x}) = \mathrm{AttentionDecoder}(\mathbf{h},\mathbf{y})%,$$ where $\mathbf{x}=[x_1, x_2, ..., x_{T_{in}} ]$ and $\mathbf{h}=[h_1, h_2, ..., h_{T} ]$ are sequences of input speech features and encoded vectors respectively, and $\mathbf{y}=[y_1, y_2, ..., y_U ]$ is a sequence of output text units. Either the start or end of the text is considered as one of the text units. In general, $\mathrm{Encoder}(\cdot)$ reduces its output length $T$ to be smaller than the input length $T_{in}$, cutting down the memory and computational footprint. A global $\mathrm{Encoder}(\cdot)$ is implemented with NN layers having powerful sequence modeling capacity, e.g., BiLSTM or self-attention layers with subsampling layers. On the other hand, an online $\mathrm{Encoder}(\cdot)$ must only consist of layers with finite future receptive field. $\mathrm{AttentionDecoder}(\cdot)$ operates at each output step recursively, emitting an estimated posterior probability over all possible text units given the outputs produced at the previous step. This procedure can be summarized as follows: $$s_u =\mathrm{RecurrentState}(s_{u-1}, y_{u-1}, c_{u-1}),$$ $$\label{AttentionContext} c_u = \mathrm{AttentionContext}(s_u, \mathbf{h}),$$ $$P(y_u | \mathbf{y}_{<u}, \mathbf{x}) = \mathrm{ReadOut}(s_u, y_{u-1}, c_u)%,$$ where $c_u$ denotes the $u$-th attention context vector and $s_u$ is the $u$-th decoder state. $\mathrm{RecurrentState}(\cdot)$ consists of unidirectional layers, e.g., unidirectional LSTM and masked self-attention layers. $\mathrm{ReadOut}(\cdot)$ usually contains a small NN followed by a softmax activation function. The most popular choice for $\mathrm{AttentionContext}(\cdot)$ is the global soft attention (GSA) [@Bahdanau; @luong2015effective] that includes the softmax function given as follows:$$\label{eq:GSA_context} c_u = \sum_{t=1}^{T}{\alpha_{u,t}} h_t ,$$ $$\label{GSA_weights} \alpha_{u,t} = \frac{\exp(e_{u,t})} {\sum_{j=1}^{T}{\exp(e_{u,j})}},$$ $$\label{GSA_score} e_{u,t} = \mathrm{Score}(s_u, h_t, \alpha_{<u,t})%,$$ in which $\alpha_{u,t}$ is an attention weight on the $t$-th encoded vector $h_t$ at the $u$-th decoder step, and $e_{u,t}$ is a score indicating the relevance of $h_t$ to the $u$-th decoder state. Many different choices are available for the similarity function $\mathrm{Score}(\cdot)$. The whole system is trained to maximize the log posterior probability on a training dataset $\mathbf{D}=\{(\mathbf{x}^{(n)},\mathbf{y}^{(n)}) \}_{n=1}^{N} $, $$\mathrm{max}_{\theta} \,\, \mathrm{E}_{ (\mathbf{x},\mathbf{y}) \sim \mathbf{D}} \Big[ %\sum_{u=1}^{U^{(n)}}{\log}P(y_u^{(n)}| \mathbf{y}^{(n)}_{<u}, \mathbf{x}^{(n)} ;\theta) \Big], %\sum_{u=1}^{U}{\log}P(y_u| \mathbf{y}_{<u}, \mathbf{x} ;\theta) \Big], \sum_{u=1}^{{|\mathbf{y}|}}{\log}P(y_u| \mathbf{y}_{<u}, \mathbf{x} ;\theta) \Big]%,$$ where $\theta$ denotes the set of all trainable parameters, and $|\mathbf{y}|$ is the text sequence length of the sampled data. Inference can be performed by searching the most likely text sequence: $$\mathbf{\hat{y}} = \mathrm{argmax}_{\theta} \,\, {\log}P(\mathbf{y} | \mathbf{x} ;\theta).$$ Online Attention {#sec:online_attention} ---------------- To achieve online attention, the context vector $c_u$ in Eq.  must have local dependency on the encoded vectors $\mathbf{h}$. Windowed attention and MoChA are widely-used online attention methods that show high performance for which only the $\mathrm{AttentionContext}(\cdot)$ function in Eq.  is modified in the general framework. ### Windowed attention Among various formulations of windowed attention, a simple heuristic using $\mathrm{argmax}$ for window boundary prediction [@merboldt2019analysis] has shown the best performance. This method can be described as follows: $$\label{Windowed_att_position} %p_1 = 0, \quad p_u = \mathrm{argmax}_{\tau}(\alpha_{u-1,\tau}), p_1 = 0, \quad p_u = \mathrm{argmax}_{t}(\alpha_{u-1, 1\leq t\leq T}),$$ $$\label{Windowed_att_context} c_u = \sum_{t = p_u}^{p_u+w-1}{\alpha_{u,t}} h_t ,$$ $$\label{Windowed_att_weights} \alpha_{u,t} = \frac{\exp(e_{u,t})} {\sum_{j=p_u}^{p_u+w-1}{\exp(e_{u,j})}}%,$$ where $p_u$ is the start point of the attention window at the $u$-th step, and $w$ is the window size. The windowed attention is online, as the attention context $c_u$ derived through Eqs. - does not depend on the entire encoded vector sequence $\mathbf{h}$. The tradeoff between performance and latency of windowed attention relies on the window length $w$. ### MoChA In MoChA [@chiu2018monotonic], an attention window endpoint is first decided, followed by attention weights calculation within a fixed-size window as follows: $$\label{MoChA_infer_context} c_{u} = \sum_{t=\tau_u-w+1}^{\tau_u}\beta_{u,t} h_t,$$ $$\label{MoChA_infer_beta} \beta_{u,t} = \frac{\exp(e_{u,t})} {\sum_{j=\tau_u-w+1}^{\tau_u}{\exp(e_{u,j})}},$$ $$\label{MoChA_infer_tau} \tau_{u} = \mathrm{MonotonicEndpoint}(\tilde{e}_{u,\geq\tau_{u-1}}) ,$$ $$\label{Monotonic_infer_chunkenergy} \tilde{e}_{u,t} = \mathrm{MonotonicScore}(s_u, h_t, \alpha_{<u,t}) + b%,$$ where $\mathrm{MonotonicScore(\cdot)}$ is a similarity function, $b$ is a trainable bias parameter, $\tilde{e}_{u,t}$ is the monotonic score, and $\mathrm{MonotonicEndpoint(\cdot)}$ is an window end-decision algorithm based on thresholding, and $\beta_{u,t}$ is an attention weight within the window. Note that Eqs. - are substitutes for Eqs. - in GSA. The performance and latency of MoChA are also known to depend on the chunk size $w$. Optimizing an AED model using these formulations is infeasible since they do not fit the backpropagation framework. To solve this problem, an expectation-based formulation is exploited for training [@chiu2018monotonic]: $$\label{MoChA_train_beta} %\beta_{u,t} = \sum_{k=t}^{t+w-1} \Big{(} \alpha_{u,t} \exp(e_{u,k}) \Big{/} \sum_{l=k-w+1}^{k}\exp(e_{u,l}) \Big{)}, \beta_{u,t} = \sum_{k=t}^{t+w-1} \frac{ \alpha_{u,t} \exp(e_{u,k})}{ \sum_{l=k-w+1}^{k}\exp(e_{u,l}) },$$ $$\label{MoChA_train_alpha} \alpha_{u,t} = p_{u,t} \Big( (1-p_{u,t-1})\frac{\alpha_{u,t-1}}{p_{u,t-1} } + \alpha_{u-1,t} \Big),$$ $$\label{MoChA_train_p} p_{u,t} = \sigma(e_{u,t})%,$$ where $p_{u,t}$ is a stopping probability at the $t$-th time step and $\alpha_{u,t}$ is an accumulated selection probability that the window endpoint is $t$. ### A limitation of the conventional methods As mentioned in Sec. \[sec:introduction\], the softmax function in the conventional online attentions (e.g., Eqs. -) requires a predetermined attention window, which induces a limitation in training efficiency since multiple trials of training are inevitable for tuning either the window length or the coefficient of an additional loss term. To overcome this limitation, in the next section, we propose a novel softmax-free global attention approach and its online version which is free from the tuning of hyperparameters in training. Proposed methods {#sec:proposed} ================ Gated Recurrent Context (GRC) ----------------------------- We propose a novel softmax-free global attention method called GRC, which recursively aggregates the information of the encoded sequence into an attention context vector in a time-synchronous manner. Specifically, the following formulas are employed in place of the Eqs. -: $$\label{GRC_context} c_{u} = d_{u,T},$$ $$\label{GRC_recursive} \quad d_{u,t} = (1-z_{u,t})d_{u,t-1} + z_{u,t}h_t,$$ $$\label{GRC_updategate} z_{u,1}=1, \quad z_{u,t} = \sigma(e_{u,t}) = \frac{1}{1+\exp(e_{u,t})},$$ $$\label{GRC_energy} e_{u,t} = \mathrm{Score}(s_u, h_t, \alpha_{<u,t}) + b%,$$ where $z_{u,t}$ and $d_{u,t}$ are the update gate and the intermediate attention context vector for the $t$-th time step at the $u$-th decoder step, respectively. GRC computes an intermediate value for the final context vector recursively in time, inspired by GRU [@cho2014learning]. Note that Eqs. - of GRC do not utilize the softmax function at all, unlike the conventional attentions. Nevertheless, GRC can be interpreted as a global attention method, since it calculates a weighted average of the encoded sequence over the whole time period, as explained in Sec. \[sec:relation\_to\_GSA\]. ### Relation to GSA {#sec:relation_to_GSA} The update gate $z_{u,t}$ of GRC in Eq.  and the attention weight $\alpha_{u,t}$ of GSA in Eq.  have one-to-one correspondence according to the following theorem: \[theorem\] \[thm1\] For arbitrary $n \in \mathbb{N}$, let $Z^n=\{x \in \mathbb{R}^{n} \,|\,x_1=1,\,\, 0 \leq x_j \leq 1 \,\, for \,\, j=2,3,\dots,n \}$ and $A^n=\{x \in \mathbb{R}^{n} \,|\,\sum_{j=1}^{T}x_j=1,\,\, 0 \leq x_j \leq 1 \,\, for \,\, j=1,2,\dots,n \}$. There exists a bijective function $\boldsymbol{\bar{\alpha}} :Z^T \rightarrow A^T$ s.t. for any $\mathbf{h}=[h_1, h_2, \dots, h_T]$ and $\mathbf{z}_{u}=[z_{u,1}, z_{u,2}, \dots, z_{u,T}]$, the following holds: $$\label{thm1_eq} d_{u,T} = \sum_{t=1}^{T} \boldsymbol{\bar{\alpha}}(\mathbf{z}_{u})_{t}h_t%,$$ where $\boldsymbol{\bar{\alpha}}(\mathbf{z}_{u})_{t}$ denotes the $t$-th element of $\boldsymbol{\bar{\alpha}}(\mathbf{z}_{u})$, and $d_{u,T}$ is obtained from $\mathbf{z}_{u}$ and $\mathbf{h}$ according to Eq. . Using the recursive Eq. , $$\label{thm1_proof_eq1} \begin{split} d_{u,T} = &(1-z_{u,T}) d_{u,T-1} + z_{u,T}h_T \\ = &(1-z_{u,T})(1-z_{u,T-1}) d_{u,T-2} \\ &+ (1-z_{u,T})z_{u,T-1} h_{T-1} + z_{u,T}h_T \\ = &\dots \\ = &\sum_{t=1}^T \Big(\prod_{j=t+1}^{T} (1-z_{u,j})\Big)z_{u,t}h_t . \end{split}$$ Therefore the function $\boldsymbol{\bar{\alpha}}$ that satisfies Eq. \[thm1\_eq\] is given by $$\label{GRC_GSA_dual_function} \boldsymbol{\bar{\alpha}}(\mathbf{z}_{u})_t:= z_{u,t}\prod_{j=t+1}^{T} (1-z_{u,j}) \quad \text{for $t=1, 2, \dots, T$.} % \begin{dcases} % z_{u,t}\prod_{j=t+1}^{T} (1-z_{u,j}) & \text{if $t \leq T$,} \\ % 0 & \text{otherwise,} \\ % \end{dcases} % \quad \text{for $t=1, 2, \dots, T$.}$$ Given that $\mathbf{z}_u \in Z^T$, the output $\boldsymbol{\bar{\alpha}}(\mathbf{z}_{u})$ is an element of $A^T$ because it is trivial to show that $0 \leq \boldsymbol{\bar{\alpha}}(\mathbf{z}_{u})_t \leq 1$ for $i = 1, 2, \dots, T$, and also $\sum_{t=1}^{T} \boldsymbol{\bar{\alpha}}(\mathbf{z}_{u})_t=1$ holds as follows: $$\begin{split} \sum_{t=1}^{T} \boldsymbol{\bar{\alpha}}(\mathbf{z}_{u})_t &= \sum_{t=1}^{T} z_{u,t}\prod_{j=t+1}^{T} (1-z_{u,j}) \\ &= \sum_{t=2}^{T} z_{u,t}\prod_{j=t+1}^{T} (1-z_{u,j}) + \prod_{j=2}^{T} (1-z_{u,j})\\ %&= \sum_{t=2}^{T} z_{u,t}\prod_{j=t+1}^{T} (1-z_{u,j}) + \prod_{j=3}^{T} (1-z_{u,j}) - z_{u,2}\prod_{j=3}^{T} (1-z_{u,j})\\ &= \sum_{t=3}^{T} z_{u,t}\prod_{j=t+1}^{T} (1-z_{u,j}) + \prod_{j=3}^{T} (1-z_{u,j})\\ &= \dots \\ &= z_{u,T} + (1-z_{u,T}) = 1. \end{split}$$ The $\boldsymbol{\bar{\alpha}}(\mathbf{z}_{u})$ is a bijective function since the inverse mapping of $\boldsymbol{\bar{\alpha}}$ exists as follows: $$\begin{split} z_{u,T} & = \boldsymbol{\bar{\alpha}}(\mathbf{z}_{u})_T, \\ z_{u,T-1} & = \begin{dcases} \frac{\boldsymbol{\bar{\alpha}}(\mathbf{z}_{u})_{T-1}}{\big(1-z_{u,T}\big)} = \frac{\boldsymbol{\bar{\alpha}}(\mathbf{z}_{u})_{T-1}}{1-\boldsymbol{\bar{\alpha}}(\mathbf{z}_{u})_T} & \text{if $\boldsymbol{\bar{\alpha}}(\mathbf{z}_{u})_T < 1$;} \\ 0 & \text{otherwise,} \\ \end{dcases} \\ z_{u,T-2} & = \begin{dcases} %\frac{\boldsymbol{\bar{\alpha}}(\mathbf{z}_{u})_{T-2}}{\big(1-z_{u,T}\big) \big(1-z_{u,T-1}\big)} = \frac{\boldsymbol{\bar{\alpha}}(\mathbf{z}_{u})_{T-2}}{1 - \boldsymbol{\bar{\alpha}}(\mathbf{z}_{u})_T - \boldsymbol{\bar{\alpha}}(\mathbf{z}_{u})_{T-1}} & \text{if %$\boldsymbol{\bar{\alpha}}(\mathbf{z}_{u})_{T-1} + \boldsymbol{\bar{\alpha}}(\mathbf{z}_{u})_{T} < 1$; $\sum_{j=T-1}^{T}\boldsymbol{\bar{\alpha}}(\mathbf{z}_{u})_{j}$; } \\ 0 & \text{otherwise,} \\ \end{dcases} \\ & \vdots\\ \therefore z_{u,t} & = \begin{dcases} \frac{\boldsymbol{\bar{\alpha}}(\mathbf{z}_{u})_{t}}{1 - \sum_{j=t+1}^{T} \boldsymbol{ \bar{\alpha}}(\mathbf{z}_{u})_j} & \text{if $\sum_{t=j+1}^{T} \boldsymbol{\bar{\alpha}}(\mathbf{z}_{u})_{T} < 1$;} \\ 0 & \text{otherwise,} \\ \end{dcases} %\quad \text{for $t= 1, 2, \dots, T$.} \\ \end{split}$$ for $t= 1, 2, \dots, T$. It is also trivial to show that $z_{u,1} = 1$ and $0 \leq z_{u,t}\leq 1$ for $i = 2, \dots, T$, given that $\boldsymbol{\bar{\alpha}}(\mathbf{z}_{u}) \in A^T$. Therefore, $\mathbf{z}_u \in Z^T$. +0.1in Note that $\boldsymbol{\bar{\alpha}}(\mathbf{z}_u)_t$ in Eq. corresponds to the attention weight $\alpha_{u,t}$ in Eq.  of GSA. By Thm. \[thm1\], the attention context vector $c_u$ of GRC is capable of expressing all possible weighted averages of the encoded representations over time, as in the GSA. Thus the range of $c_u$ in GRC or GSA is the same. Nonetheless, we empirically showed that GRC performs comparable to or even better than GSA, and the experimental results are given in Sec. \[sec:experiments\]. ### Relation to sMoChA {#relation_sMoChA} The sMoChA [@miao2019online] is a variant of MoChA where Eq.  is replaced by the following formula: $$\label{sMoChA} \alpha_{u,t} = p_{u,t} \prod_{j=1}^{t-1} (1-p_{u,j})%,$$ which enables the optimization process to be more stabilized. Eq.  is almost similar to the function $\boldsymbol{\bar{\alpha}}$ in Eq. , and implies evidence on the stability of GRC training. Despite this fact, sMoChA is an algorithm independent of GRC, as Eq.  is merely used as the selection probability component in the whole training formulas and not even used for inference. Decreasing GRC (DecGRC) ----------------------- The intermediate context $d_{u,t}$ of GRC is a weighted average of the encoded representations over time according to the following corollary: \[coroll1\_thm1\] Let $Z^n$ and $A^n$ be the sets defined in Thm. \[thm1\]. For any $\tau \in \{1, 2, \dots, T\}$, $\mathbf{z}_u \in Z^\tau$, and $\mathbf{h}=[h_1, h_2, \dots, h_T]$ , there exists a function $\boldsymbol{\bar{a}} :Z^\tau \rightarrow A^T$ that satisfies the following equation: $$%\label{thm2_eq} \label{coroll1_thm1_eq} d_{u,\tau} = \sum_{t=1}^{T} \boldsymbol{\bar{a}}(\mathbf{z}_u)_{t}h_{t}%,$$ where $\boldsymbol{\bar{a}}(\mathbf{z}_u)_{t}$ denotes the $t$-th element of $\boldsymbol{\bar{a}}(\mathbf{z}_u)$, and $d_{u,\tau}$ is obtained from $\mathbf{z}_{u}$ and $\mathbf{h}$ according to Eq. . By substituting every $T$ in the proof of Thm. \[thm1\] with $\tau$, there exists a bijective function $\boldsymbol{\bar{\alpha}}: Z^\tau \rightarrow A^\tau$ given by $$\boldsymbol{\bar{\alpha}}(\mathbf{z}_{u})_t:= z_{u,t}\prod_{j=t+1}^{\tau} (1-z_{u,j}) \quad \text{for $t=1, 2, \dots, \tau$,}%.}$$ such that $$%\label{thm1_coroll1_proof_eq1} d_{u,\tau} = \sum_{t=1}^\tau \boldsymbol{\bar{\alpha}}(\mathbf{z}_{u})_t h_t.$$ It is trivial to show that the following function $\boldsymbol{\bar{a}}: Z^\tau \rightarrow A^T$ satisfies the equation in Coroll. \[coroll1\_thm1\_eq\]: $$\boldsymbol{\bar{a}}(\mathbf{z}_u)_t = \begin{dcases} \boldsymbol{\bar{\alpha}}(\mathbf{z}_u)_t & \text{if $t \leq \tau$;} \\ 0 & \text{otherwise,} \\ \end{dcases} \quad \text{for $t= 1, 2, \dots, T$.} \\$$ Global attention methods including GSA and GRC cannot compute the attention weights without the entire sequence of the encoded vectors $\mathbf{h}$. However, considering that the attention techniques are methods for calculating the weighted average of the encoded vectors, Coroll. \[coroll1\_thm1\] enables us to treat an intermediate context $d_{u,t}$ as a substitute for the attention context vector $c_u$ in Eq.  of GRC even when the whole encoded sequence is not provided. Inspired by this, we further propose a novel online attention algorithm, namely DecGRC. DecGRC is a modified version of GRC, replacing Eq.  with $$\label{DecGRC} z_{u,1}=1, \quad z_{u,t} = \frac{1}{1+\sum_{j=1}^{t}\exp(e_{u,j})}.$$ Note that the update gate is constrained to be monotonically decreasing over time. Assume that there exists an *endpoint* index $t_{end}$ with which $z_{u,t_{end}}$ has a very small value (e.g. less than 0.001). Considering that $z_{u,t}<z_{u,t_{end}}$ holds for all $t>t_{end}$, the difference between $d_{u,t_{end}}$ and $d_{u,T}$ is small, as the numerical change $|d_{u,t}-d_{u,t-1}|$ for $t>t_{end}$ induced by the recursion rule in Eq.  is negligible if $z_{u,t}$ is small enough. Intuitively, intermediate context vectors roughly converge after the endpoint. DecGRC can operate as an online attention method if such an endpoint index $t_{end}$ exists at each decoder step and the index can be decided by the model. We experimentally observed that DecGRC models adequately learn the alignment between encoded vectors and text output units, and the intermediate context nearly converges after the aligned time index at each decoder step. Relevant experimental results are given in Sec. \[sec:attention\_analysis\] Accordingly, with an online encoder, online inference can be implemented via a well-trained DecGRC model. We describe the online inference technique in Alg. \[alg:example\], where the endpoint index is decided simply by thresholding the update gate values. encoded vectors $\mathbf{h}$ of length $T$, threshold $\nu$ $s_0=\vec{0}$, $u=1$, $y_0=\mathrm{StartOfSequence}$ $d_{u} = h_1$ $e_{u,t} = \mathrm{Score}(s_u, h_t,\alpha_{u,t})+b$ $z_{u,t} = 1 / (1+\sum_{j=1}^{t}\exp(e_{u,j}))$ $d_{u} = (1-z_{u,t}) d_{u} + z_{u,t} h_t$ **break** $s_u =\mathrm{RecurrentState}(s_{u-1}, y_{u-1}, d_{u})$ $\tilde{P}(y_u | \mathbf{y}_{<u}, \mathbf{h}) = \mathrm{ReadOut}(s_u, y_{u-1}, c_u)$ $y_u = \mathrm{Decide}\big(\tilde{P}(y_u | \mathbf{y}_{<u}, \mathbf{h})\big)$ $u=u+1$ Computational efficiency of proposed methods -------------------------------------------- GRC or DecGRC increases negligible amount of memory footprint, since only one trainable parameter $b$ in Eq.  is added to the standard GSA-based AED model. Both proposed methods have computational complexity of $O(TU)$, same as the conventional global attentions. However, in practice, a speech sequence is linearly aligned with the text sequence on average. As Alg. \[alg:example\] only regards to encoded vectors before endpoint indices, the total number of steps in the for loop is typically slightly larger than $TU/2$, if the threshold $\nu$ is set to an appropriate value. Therefore, DecGRC is computationally more efficient than the global attentions such as GRC and GSA. The recursive updating in Eq.  induces negligible amount of computation compared to the whole training or inference process. There still exists a room for faster computation by enabling parallel computation in time. The parallel computation can be implemented by utilizing Eq.  where $\alpha_{u,t}$ is replaced with $\bar{\alpha}(\mathbf{z}_u)$ in Eq. , instead of Eqs. -. ----- ---------------------- ----- -- -- -- ---------- ----------- ---------- ----------- E1 GSA (Bahdanau) - 15.15 E2 GRC - **14.59** E3 Windowed att. (w=11) E1 12.50 23.79 15.27 25.81 E4 Windowed att. (w=20) E1 5.78 14.82 5.71 15.90 E5 MoChA (w=2) - 6.49 17.11 6.17 18.18 E6 MoChA (w=8) - **4.74** 14.20 4.95 15.32 E7 - 4.91 14.85 5.10 15.85 E8 E2 4.97 **14.02** **4.83** **14.90** E9 - 5.54 15.49 5.51 16.91 E10 E1 **5.28** 15.44 **5.17** 16.40 E11 - 6.09 16.05 6.18 16.47 E12 E2 5.48 **15.14** 5.55 **15.88** E13 E10 12.82 24.10 15.14 26.94 E14 E10 5.62 15.86 5.56 16.96 E15 MoChA (w=2) E5 6.48 18.35 6.55 19.33 E16 MoChA (w=8) E6 **5.11** **15.10** **5.15** 16.45 E17 E8 5.77 16.24 5.87 17.04 E18 E12 5.79 15.67 6.04 **16.34** ----- ---------------------- ----- -- -- -- ---------- ----------- ---------- ----------- Most importantly, both proposed methods introduce no hyperparameter at the training phase. Thus the proposed methods do not need to repeat training to find a proper value of such a hyperparameter. Though the DecGRC inference in Alg. \[alg:example\] introduces a new hyperparameter (i.e., threshold $\nu$) at test phase, the threshold searching on development sets does not take a long time, because the size of the development sets are minor compared to the training set. Hence the total time spent to prepare an ASR system can be saved. Furthermore, the tradeoff between latency and performance can be adjusted by resetting the threshold value $\nu$ at inference phase, unlike the conventional online attention methods [@jaitly2015neural; @merboldt2019analysis; @chiu2018monotonic]. In these existing methods, the inference algorithms’ decision rules on the attention endpoints are determined at the training phase, and remains unchanged at the test stage. The experiments on DecGRC with different thresholds are demonstrated in Sec. \[sec:ablation\_study\]. Experiments {#sec:experiments} =========== Configurations -------------- All experiments were conducted on LibriSpeech dataset[^2], which contains 16 kHz read English speech with transcription. The dataset consists of 960 hours of a training set from 2,338 speakers, 10.8 hours of a dev set from 80 speakers, and 10.4 hours of a test set from 66 speakers, with no overlapping speakers between different sets. Both dev and test sets are split in half into clean and other subsets, depending upon the ASR difficulty of each speaker. We randomly chose 1,500 utterances from dev set as a validation set. All experiments [^3] shared the same network architecture and training scheme of a recipe of RETURNN toolkit [@zeyer2018returnn; @zeyer2018comprehensive], except the attention methods. Input features were 40-dimensional mel-frequency cepstral coefficients (MFCCs) extracted with Hanning window of 25 ms length and 10 ms hop size, followed by global mean-variance normalization. Output text units were 10,025 byte-pair encoding (BPE) units extracted from transcription of LibriSpeech training set. The $\mathrm{Encoder}(\cdot)$ consisted of 6 BiLSTM layers of 1,024 units for each direction, and max-pooling layers of stride 3 and 2 were applied after the first two BiLSTM layers respectively. For the online $\mathrm{Encoder}(\cdot)$, 6 LC-BiLSTM layers were employed in place of the BiLSTM layers, where the future context sizes were set to 36, 12, 6, 6, 6, and 6 for each layer from bottom to top and the chunk sizes were twice the future context sizes. Both $\mathrm{Score(\cdot)}$ and $\mathrm{MonotonicScore(\cdot)}$ functions were implemented using Bahdanau score with fertility-based weight feedback and 1,024-dimensional attention key, as in [@zeyer2018improved]. $\mathrm{RecurrentState}(\cdot)$ was implemented with an unidirectional LSTM layer with 1,000 units. $\mathrm{ReadOut(\cdot)}$ consisted of a max-out layer with 2$\times$500 units, followed by a softmax output layer with 10,025 units. Weight parameters were initialized with Glorot uniform method [@glorot2010understanding], and biases were initially set to zero. Optimization techniques were utilized during the training: teacher forcing, Adam optimizer, learning rate scheduling, curriculum learning, and the layer-wise pre-training scheme. Briefly, the models were trained for 13.5 epochs using a learning rate of 8$\times$10$^{-4}$ with a linear warm-up starting from 3$\times$10$^{-4}$ and the Newbob decay rule [@zeyer2017comprehensive]. Only the first two layers of $\mathrm{Encoder(\cdot)}$ with half-width were used at the beginning of a training, then the width and the number of layers gradually increased to the original size at 1.5 epoch. The CTC multi-task learning [@kim2017joint] with a lambda of 0.5 was employed to stabilize the learning, where CTC loss is measured with another 10,025-units softmax layer on the top of $\mathrm{Encoder(\cdot)}$. For the models which began the learning from parameters of a pre-trained model, the layer-wise pre-training was skipped. Every model was regularized by applying dropout rate 0.3 to $\mathrm{Encoder(\cdot)}$ layers and the softmax layer and employing label smoothing of 0.1. For each epoch of the training, both cross-entropy (CE) losses and output error rates were measured 20 times on the validation set with teacher forcing. During the inference phase, model with the lowest WER on the dev-other set among all checkpoints was selected as the final model, and performed beam search once on the dev and test sets with a beam size of 12. We trained MoChA models for 17.5 epochs with five times longer layer-wise pre-training to make them converge. A small learning rate of 1e-5 was used for training windowed attention models as in [@merboldt2019analysis]. Though the numbers of total epochs for different experiments were not the same, each model was optimized to converge and showed negligible improvements after that. Performance comparison between attentions ----------------------------------------- All experimental results are summarized in Tbl. \[LibriSpeechWER\]. For each experiment, we performed two trials of training with the same configuration and chose a model with the lowest word-error-rate (WER), a word-level Levenshtein distance divided by the number of ground-truth words, on dev-other set. In E1 to E2 and E9 to E12, GRC showed better performance than the other attention methods on test-other set, showing 3.7% and 3.2% relative error-reduction rate (RERR) compared to GSA when evaluated on BiLSTM and LC-BiLSTM encoder, respectively. In E3 to E6 and E13 to E16, performances of the conventional online attentions, i.e., windowed attention and MoChA, were shown to be highly dependent on a choice of window size hyperparameter $w$. On the other hand, DecGRC is trained without any additional hyperparameter and only involves a threshold $\nu$ at the inference phase. In E3 to E8 and E13 to E18, DecGRC outperformed the conventional online attention techniques on BiLSTM encoder. With LC-BiLSTM encoder, the performance of DecGRC on test-other set surpassed the conventional attentions including GSA, while the scores on test-clean set were worse than the competitors. The overall performance of GRC and DecGRC is degraded on LC-BiLSTM compared to their preferable performance on BiLSTM, which was conjectured to be caused by the following aspect of the proposed methods; $\boldsymbol{\bar{\alpha}}({\mathbf{z}_u})$ in Eq.  has a dependency on update-gate values of the future time-steps. Therefore using a short future receptive field of LC-BiLSTM may affected the degradation. Optimization speed ------------------ coordinates [(0.05,6.606168589465353)(0.1,6.0580554242266675)(0.15,5.821005939494658)(0.2,5.637596254754004)(0.25,5.476822565948908)(0.3,5.286328993667397)(0.35,4.451726806013469)(0.4,2.7567605693972244)(0.45,1.9853467248104093)(0.5,1.7591589752050938)(0.55,1.500116264659877)(0.6,1 .3590997123321222)(0.65,1.41116809282427)(0.7,1.182418052724496)(0.75,1.1360856823934176)(0.8,3.4904476188944025)(0.85,1.3652874665792123)(0.9,1.1532419330453445)(0.95,1.1711545478801069)(1.0,0.9821442753120624)(1.05,3.3585877305503358)(1.1,1.2729943696708703)(1.15,1.0319864420147544)(1.2,0.9444800880807288)(1.25,0.9032107740213039)(1.3,2.718888774161502)(1.35,1.0703718146236398)(1.4,0.9602280724499165)(1.45,0.9011093543550479)(1.5,0.8466466826631844)(1.55,3.0662677271494103)(1.6,1.0544228009472063)(1.65,0.9531792639494701)(1.7,0.9167340914163986)(1.75,0.9242554715334796)(1.8,1.1007853316288543)(1.85,0.9772334565055945)(1.9,0.9212078219821279)(1.95,0.9002388760792722)(2.0,0.8684908077417614)(2.05,0.7966433578341311)(2.1,0.8138547062660167)(2.15,0.9686228336853654)(2.2,0.848054439297636)(2.25,0.7525973736745635)(2.3,0.7193418306118422)(2.35,0.7230058440051947)(2.4,0.7244809093336416)(2.45,0.8127259755764654)(2.5,0.7995350016731883)(2.55,0.7673712483044441)(2.6,0.7352334490537166)(2.65,0.7098621039358713)(2.7,0.6989299670468696)(2.75,0.6906277038516802)(2.8,0.6848225992669935)(2.85,0.6898301332699912)(2.9,0.6823743090177716)(2.95,0.6752964629237224)(3.0,0.6562583264447094)(3.05,0.6009323914527542)(3.1,0.5968251925877095)(3.15,0.6057414576663485)(3.2,0.6002146034478392)(3.25,0.6013928208499276)(3.3,0.6035648586415112)(3.35,0.5857752168639716)(3.4,0.573500504946627)(3.45,0.5783581012975069)(3.5,0.5697252565252113)(3.55,0.5691341067663355)(3.6,0.5638701809333444)(3.65,0.5630846552952314)(3.7,0.5588877271502101)(3.75,0.5553489834615226)(3.8,0.5621746913316951)(3.85,0.5465522261592191)(3.9,0.539724238804123)(3.95,0.54364700396231)(4.0,0.5397546331026913)(4.05,0.46879056511342226)(4.1,0.4708146501009042)(4.15,0.47357014348986165)(4.2,0.46975204081223887)(4.25,0.4822745709916306)(4.3,0.46986516622355234)(4.35,0.4620966421226265)(4.4,0.4584159326990642)(4.45,0.46663637468509533)(4.5,0.45367360551003455)(4.55,0.45200244629266356)(4.6,0.45231778913175985)(4.65,0.4507563133897209)(4.7,0.4502040012731752)(4.75,0.45227423091012964)(4.8,0.4424708273393489)(4.85,0.4365653269214007)(4.9,0.43923449514887575)(4.95,0.43082952217114345)(5.0,0.4313594247712902)(5.05,0.3743281580270501)(5.1,0.3867014231521147)(5.15,0.3797073268530094)(5.2,0.37836197252365317)(5.25,0.37552276133567253)(5.3,0.37418228950481974)(5.35,0.37726457489225124)(5.4,0.37780144913095176)(5.45,0.3743523503195505)(5.5,0.36572160708143797)(5.55,0.3666218387245273)(5.6,0.3701146741882173)(5.65,0.36545146909227205)(5.7,0.36608765805583454)(5.75,0.3645264824504503)(5.8,0.36293230513627883)(5.85,0.3639392079081574)(5.9,0.36552864343038477)(5.95,0.35777063442970136)(6.0,0.36659982939399927)(6.05,0.31124351317196997)(6.1,0.31242810651222136)(6.15,0.3186086685776954)(6.2,0.32080803555222837)(6.25,0.32480878998867585)(6.3,0.32460357233071047)(6.35,0.31890445143795704)(6.4,0.32186127459909564)(6.45,0.31914318818563364)(6.5,0.31243005821649944)(6.55,0.30776411631693523)(6.6,0.31381766759961244)(6.65,0.3114361186457307)(6.7,0.309519418237535)(6.75,0.3093251208679713)(6.8,0.30628138550281053)(6.85,0.309011907913942)(6.9,0.3104019643717366)(6.95,0.3073381925159789)(7.0,0.3052899413145837)(7.05,0.26477615165429236)(7.1,0.2684837190523341)(7.15,0.2719009206283708)(7.2,0.274577799820348)(7.25,0.27710253846250466)(7.3,0.2818276908003217)(7.35,0.2800747598658199)(7.4,0.27184511825681174)(7.45,0.27280053542175314)(7.5,0.2731939191441712)(7.55,0.27329404831638865)(7.6,0.2660851376335041)(7.65,0.26538316704378245)(7.7,0.2681497796208545)(7.75,0.2689089944715839)(7.8,0.26426897558451695)(7.85,0.26042411114433794)(7.9,0.264454370714927)(7.95,0.26154378530154226)(8.0,0.2611000465647041)(8.05,0.22492122072174048)(8.1,0.22690474661665577)(8.15,0.22600653679921615)(8.2,0.22681658663076273)(8.25,0.22656074604154144)(8.3,0.2287841423788997)(8.35,0.22878097748789067)(8.4,0.2293920914724632)(8.45,0.22853387020851423)(8.5,0.22934122731150405)(8.55,0.22965601347525386)(8.6,0.22998600792878884)(8.65,0.22850762374219089)(8.7,0.23065205005796713)(8.75,0.22745802362185175)(8.8,0.22751058927163736)(8.85,0.22554271530097797)(8.9,0.2261955725728062)(8.95,0.229338392421983)(9.0,0.22512409993929894)(9.05,0.19694945238393155)(9.1,0.19825492929610528)(9.15,0.197828964004379)(9.2,0.19963102865643395)(9.25,0.2000879755659006)(9.3,0.19865253531064853)(9.35,0.20246316357129757)(9.4,0.20037957650633925)(9.45,0.200733286103027)(9.5,0.19823882042719917)(9.55,0.19859896665309257)(9.6,0.1998560283771629)(9.65,0.19844103213125527)(9.7,0.19801320266937641)(9.75,0.19882977845788)(9.8,0.19832143504638236)(9.85,0.19707337343524942)(9.9,0.19781922257409212)(9.95,0.1964261584705095)(10.0,0.19452738416415083)(10.05,0.17478246937088682)(10.1,0.17542804335044457)(10.15,0.17755608263692726)(10.2,0.17714353850790393)(10.25,0.177967753732663)(10.3,0.17832524877764588)(10.35,0.1767714930516409)(10.4,0.17727099583051165)(10.45,0.1758440983918837)(10.5,0.17827164655392344)(10.55,0.1785907619913131)(10.6,0.17554434987618564)(10.65,0.1771109903632465)(10.7,0.17576477054924017)(10.75,0.1769759962848957)(10.8,0.17597156420297988)(10.85,0.1749493226323683)(10.9,0.17585266111904652)(10.95,0.17539746970576892)(11.0,0.17597897950672667)(11.05,0.1598879189615477)(11.1,0.16118041075956221)(11.15,0.1616333857456364)(11.2,0.15945725714983028)(11.25,0.16168620092056862)(11.3,0.1607566406794696)(11.35,0.16117893164769784)(11.4,0.16199175620879666)(11.45,0.16140613487793445)(11.5,0.16370322798064055)(11.55,0.16259775496979556)(11.6,0.16028961175208786)(11.65,0.15951898525721522)(11.7,0.1605633468099001)(11.75,0.1622353714607244)(11.8,0.16291373613306223)(11.85,0.16207590757738308)(11.9,0.16117115581737512)(11.95,0.16024767817564328)(12.0,0.1621897264212411)(12.05,0.15046057562832157)(12.1,0.15125448261369867)(12.15,0.15130415467438948)(12.2,0.1507102064417709)(12.25,0.1503879458335411)(12.3,0.15334719679275002)(12.35,0.15070946357758006)(12.4,0.15074371016770718)(12.45,0.15250926979139776)(12.5,0.15141932285372672)(12.55,0.14974291031106593)(12.6,0.14889213098625434)(12.65,0.14926939871019007)(12.7,0.1480808880027816)(12.75,0.14875437380003678)(12.8,0.1481798429152866)(12.85,0.14958239266011275)(12.9,0.14931022051184406)(12.95,0.14969714434344628)(13.0,0.1492501203930928)(13.05,0.14717062809796225)(13.1,0.1454766470772957)(13.15,0.14686160694892855)(13.2,0.1479462422378774)(13.25,0.14704594232387996)(13.3,0.1461479603325077)(13.35,0.1473455082242145)(13.4,0.14725330796812172)(13.45,0.1478550625924793)(13.5,0.1477176245033056)(13.55,0.14866102488287447)(13.6,0.14924844335919618)(13.65,0.146920164451391)(13.7,0.14987474398642098)(13.75,0.1478449807764031)]{}; coordinates [(0.05,6.108530428731351)(0.1,5.798628193428268)(0.15,5.602965376011772)(0.2,5.432043708738972)(0.25,5.1942848683608)(0.3,4.865971920134973)(0.35,3.24153673790708)(0.4,2.1089283951569646)(0.45,1.768304751512065)(0.5,1.604418274513482)(0.55,1.3966846487104816)(0.6,1.3037 251803650298)(0.65,1.2513428691024155)(0.7,1.171425253698173)(0.75,1.1431799771991538)(0.8,1.4753423808508344)(0.85,1.3912839323666417)(0.9,1.0566592675682764)(0.95,1.0445984437756872)(1.0,0.9810305702061082)(1.05,1.2557128873155854)(1.1,1.1154110771006245)(1.15,1.0282446167423767)(1.2,0.9386105819672383)(1.25,0.9157266667703877)(1.3,1.1265616216913776)(1.35,1.0741187451027012)(1.4,0.9721355883633462)(1.45,0.8845744230150501)(1.5,0.8660752656811459)(1.55,1.1166370786144946)(1.6,0.9548611198092443)(1.65,0.9133830922273539)(1.7,0.8817091934452985)(1.75,0.8913139206382851)(1.8,0.9723702533007179)(1.85,0.8819918945877615)(1.9,0.8687062331582113)(1.95,0.8565648413588686)(2.0,0.831965448817698)(2.05,0.817491924008677)(2.1,0.8039357742361725)(2.15,0.9723779706140359)(2.2,0.7811884122234378)(2.25,0.73234789562014)(2.3,0.719365811619194)(2.35,0.7346941226116692)(2.4,0.7193388554484798)(2.45,0.7877814371936483)(2.5,0.8441326133733437)(2.55,0.7462738442541601)(2.6,0.7137098629304451)(2.65,0.7060258624490685)(2.7,0.6925830855448303)(2.75,0.6855186992648395)(2.8,0.6716065631796049)(2.85,0.674245229828022)(2.9,0.671142706596513)(2.95,0.662508990000192)(3.0,0.6535783978036558)(3.05,0.6563565049998877)(3.1,0.6509755680510012)(3.15,0.6552110649802705)(3.2,0.642515808721332)(3.25,0.6337120581748465)(3.3,0.6379133844781727)(3.35,0.6139194326109929)(3.4,0.6260990392992074)(3.45,0.6049562704681475)(3.5,0.5991632350354049)(3.55,0.6004793532775763)(3.6,0.588041204033388)(3.65,0.5925197433018097)(3.7,0.5826311636492739)(3.75,0.5842092504879609)(3.8,0.5740521083319607)(3.85,0.5713207576344345)(3.9,0.5775303670345432)(3.95,0.5615442543973489)(4.0,0.5664913038957843)(4.05,0.5598904923006413)(4.1,0.548830117468552)(4.15,0.5513764140634291)(4.2,0.5496109839781003)(4.25,0.5572148104590152)(4.3,0.5353561163797543)(4.35,0.5285849493960201)(4.4,0.5255682395253919)(4.45,0.5260014650266658)(4.5,0.51279458717868)(4.55,0.517553985017048)(4.6,0.5148995590473177)(4.65,0.5131302610017574)(4.7,0.5039697063585264)(4.75,0.5169108355685247)(4.8,0.497463660951999)(4.85,0.5032046470836894)(4.9,0.4984854327701161)(4.95,0.4888207882448862)(5.0,0.48855451209434264)(5.05,0.48656635541576465)(5.1,0.49554150149023735)(5.15,0.4823832478331194)(5.2,0.4907923738654579)(5.25,0.48332832518685775)(5.3,0.4779359011171109)(5.35,0.4767807435161065)(5.4,0.4845839157824354)(5.45,0.47012967634867275)(5.5,0.46137268648750057)(5.55,0.46208698314535035)(5.6,0.46302229221323193)(5.65,0.4560884378452904)(5.7,0.4525329571657733)(5.75,0.45899675132853696)(5.8,0.4506976558560747)(5.85,0.4558560211161927)(5.9,0.4417855279945312)(5.95,0.44357283729110747)(6.0,0.4491628831942999)(6.05,0.44619479146721763)(6.1,0.44286210250595665)(6.15,0.4435303185696367)(6.2,0.44611197199593977)(6.25,0.44086268943511414)(6.3,0.4435714240784577)(6.35,0.43781649840421855)(6.4,0.4368556500997015)(6.45,0.4372087374192674)(6.5,0.42553942981728143)(6.55,0.4254220280369811)(6.6,0.42371704651788056)(6.65,0.41994869731044476)(6.7,0.41693168600407043)(6.75,0.4140236627922832)(6.8,0.41229147669708505)(6.85,0.41171589956018784)(6.9,0.4091035620217041)(6.95,0.4063585988021993)(7.0,0.4028079222548878)(7.05,0.40747621369743886)(7.1,0.40834860828722375)(7.15,0.40666383459258637)(7.2,0.40927534829973283)(7.25,0.411030590508089)(7.3,0.41318280173084054)(7.35,0.41811504924556425)(7.4,0.40716029601038395)(7.45,0.40474831322974253)(7.5,0.40152907807782773)(7.55,0.39666965008957206)(7.6,0.3925158538769978)(7.65,0.39010543014250376)(7.7,0.392540058283581)(7.75,0.39629192179328737)(7.8,0.3886683829031403)(7.85,0.3820895737915793)(7.9,0.38228754892818734)(7.95,0.37967070776288697)(8.0,0.37562409800810176)(8.05,0.38739707717821853)(8.1,0.3827643165438976)(8.15,0.37872018340008917)(8.2,0.377981751182092)(8.25,0.3784089893693384)(8.3,0.3845845830479848)(8.35,0.3799854438962452)(8.4,0.37537667973366406)(8.45,0.37576520645386113)(8.5,0.37639544673048053)(8.55,0.37521822128829796)(8.6,0.3795229540079411)(8.65,0.37338995743336306)(8.7,0.3688532366467212)(8.75,0.36852763766221164)(8.8,0.3665050233424815)(8.85,0.36453162352528895)(8.9,0.3635191101572603)(8.95,0.36322010566708723)(9.0,0.36333968239002884)(9.05,0.36299909581553785)(9.1,0.36410247136060747)(9.15,0.3686233388598548)(9.2,0.365500792479795)(9.25,0.3678677676007554)(9.3,0.36728467434029377)(9.35,0.3627286411980336)(9.4,0.3616361429246525)(9.45,0.3638705865632626)(9.5,0.3574854215427613)(9.55,0.35738354416857065)(9.6,0.3581983836031482)(9.65,0.3579356239703881)(9.7,0.3603663218172988)(9.75,0.3544831326745545)(9.8,0.3530553601665312)(9.85,0.35331946974764866)(9.9,0.35105818716333687)(9.95,0.3576106573962134)(10.0,0.355018536055408)(10.05,0.3545310216506123)(10.1,0.3548636794881022)(10.15,0.35364523154467553)(10.2,0.35261300251116445)(10.25,0.3536336934867562)(10.3,0.35480634064095357)(10.35,0.35713153876775394)(10.4,0.35471844607092345)(10.45,0.35411120980745653)(10.5,0.3530343941009162)(10.55,0.354295728256924)(10.6,0.35239031545809796)(10.65,0.35102163552062354)(10.7,0.3492520123152074)(10.75,0.3499259363531532)(10.8,0.34761874156933237)(10.85,0.34716152355033697)(10.9,0.34812262863345267)(10.95,0.3471739675436202)(11.0,0.3475743328271256)(11.05,0.348163639245835)(11.1,0.3485102512790003)(11.15,0.348769103178757)(11.2,0.34857111873857655)(11.25,0.3477733853174192)(11.3,0.3482984842940556)(11.35,0.34799984054990285)(11.4,0.3473511024454301)(11.45,0.3463675345844872)(11.5,0.3462700270982597)(11.55,0.3469699053121706)(11.6,0.3467501116701251)(11.65,0.3463837079147728)(11.7,0.3451914886026975)(11.75,0.3450737735245961)(11.8,0.3458448275563186)(11.85,0.34509641818726405)(11.9,0.3448385350219851)(11.95,0.34462060162924385)(12.0,0.3442056604388365)(12.05,0.34446447489171417)(12.1,0.3445065323885825)(12.15,0.3439756510091011)(12.2,0.34446929460740355)(12.25,0.34445035625539633)(12.3,0.34395419348227324)(12.35,0.34385823597121007)(12.4,0.3433969055406181)(12.45,0.34313725911577725)(12.5,0.34410304110474577)(12.55,0.33984275919968476)(12.6,0.3395072304460018)(12.65,0.3393354336101496)(12.7,0.3395865181290289)(12.75,0.33900456773167603)(12.8,0.3387719659320767)(12.85,0.3385053121527092)(12.9,0.3384123763804576)(12.95,0.33864851572890203)(13.0,0.3387160706106534)(13.05,0.3385415790424849)(13.1,0.3378778002714276)(13.15,0.3394306487716451)(13.2,0.33859101078305764)(13.25,0.3381220872838744)(13.3,0.3385249556405827)(13.35,0.3392957447343889)(13.4,0.33825445581789243)(13.45,0.3383856007014322)(13.5,0.3388343169424843)(13.55,0.33855549148550784)(13.6,0.3382468442156311)(13.65,0.3378271737750454)(13.7,0.3387160822349455)(13.75,0.3385547411888548)]{}; coordinates [(0.05,8.406286575065327)(0.1,7.14201103086453)(0.15,6.992275509874851)(0.2,6.933837756473986)(0.25,6.943017567730235)(0.3,6.837342827206325)(0.35,6.481315562670667)(0.4,6.337117043572911)(0.45,6.208517103651352)(0.5,6.114876012041933)(0.55,5.9618679720766705)(0.6,5.813 193637286171)(0.65,5.723464365032351)(0.7,5.607619493022751)(0.75,5.544625222653081)(0.8,5.658852431214185)(0.85,5.5415536091454785)(0.9,5.453049196492626)(0.95,5.401160145158689)(1.0,5.3566702273970535)(1.05,5.4168955225520214)(1.1,5.302143530755093)(1.15,5.213423362787778)(1.2,5.169733735600509)(1.25,5.10866988372212)(1.3,5.335110839386759)(1.35,5.198955982383131)(1.4,5.124945934565066)(1.45,5.092638874351591)(1.5,5.036574844532098)(1.55,5.174902502366388)(1.6,5.044440848040173)(1.65,4.969778513124316)(1.7,4.928227070812125)(1.75,4.8761718387419455)(1.8,4.325697467110933)(1.85,4.269876114550827)(1.9,4.190417507600032)(1.95,2.8008753486232187)(2.0,1.5589738162369262)(2.05,1.1756951433491032)(2.1,1.0430874810841608)(2.15,0.9673295387842449)(2.2,0.9121061069921834)(2.25,0.8664988783426724)(2.3,0.8252328038641787)(2.35,0.797753014558661)(2.4,0.7786830748773113)(2.45,0.7456982387131387)(2.5,0.7229141365068738)(2.55,0.7076513972782222)(2.6,0.6955173702693008)(2.65,0.707928698576308)(2.7,0.6799344967233544)(2.75,0.6578781282030595)(2.8,0.641032088364839)(2.85,0.631911235666178)(2.9,1.2848543541585997)(2.95,0.867235254789501)(3.0,0.8164386092669731)(3.05,0.7303409710274935)(3.1,0.7350542337454917)(3.15,0.7156597998872914)(3.2,0.7047227049897613)(3.25,0.6805359169429543)(3.3,0.6936170788421739)(3.35,0.6761758079850427)(3.4,0.6588744997661325)(3.45,0.6422467522281037)(3.5,0.63521008230386)(3.55,0.6261960141066185)(3.6,0.6225536768752439)(3.65,0.6418963613406978)(3.7,0.6153179581155361)(3.75,0.5987750924859996)(3.8,0.5944626898952351)(3.85,0.5802826173321445)(3.9,0.5761882469168711)(3.95,0.5779525954021176)(4.0,0.5703865534470541)(4.05,0.513201011346366)(4.1,0.5088084661727337)(4.15,0.5004486572945304)(4.2,0.49748353898728576)(4.25,0.4873943927857756)(4.3,0.4878452151905451)(4.35,0.486404095471078)(4.4,0.4878929777784781)(4.45,0.48417691691036063)(4.5,0.4886556876880725)(4.55,0.47807034323677916)(4.6,0.48029944120179696)(4.65,0.4788868302610857)(4.7,0.47609721441178726)(4.75,0.47629158909702823)(4.8,0.46999736176746654)(4.85,0.4660251602684708)(4.9,0.466461208755589)(4.95,0.4730849472266876)(5.0,0.4661249334829371)(5.05,0.40622402209885927)(5.1,0.4151228541280056)(5.15,0.4136604437777902)(5.2,0.4147529183084921)(5.25,0.4147205993517703)(5.3,0.41278952628247184)(5.35,0.40689721278831237)(5.4,0.40745559932448805)(5.45,0.40486609043030447)(5.5,0.4160078037358194)(5.55,0.4103467194482884)(5.6,0.4063047589626231)(5.65,0.4010036449214145)(5.7,0.39986304711558796)(5.75,0.39590383310943417)(5.8,0.39812752210587915)(5.85,0.39738680508715424)(5.9,0.3938625883569091)(5.95,0.3951395755578779)(6.0,0.39077601374116877)(6.05,0.3433206854293202)(6.1,0.3594308875255259)(6.15,0.3479109321304907)(6.2,0.34506289199235507)(6.25,0.3471837025616942)(6.3,0.35153367722229945)(6.35,0.3516411217083576)(6.4,0.35334046502399974)(6.45,0.34903394002678556)(6.5,0.3453417595641817)(6.55,0.34789411476772947)(6.6,0.34526566503570655)(6.65,0.3414038393887537)(6.7,0.3442630670869288)(6.75,0.33903396509376016)(6.8,0.34311196150610357)(6.85,0.3446242925054519)(6.9,0.3440984068684178)(6.95,0.34204930443321635)(7.0,0.34537960931830153)(7.05,0.2874434954543034)(7.1,0.30012118451277353)(7.15,0.30289635124112296)(7.2,0.3041324702322161)(7.25,0.30344000608218225)(7.3,0.30299464831667766)(7.35,0.3015745778937562)(7.4,0.2964791866215405)(7.45,0.3025649009029621)(7.5,0.29430838061366854)(7.55,0.2946034599261612)(7.6,0.29640693226897014)(7.65,0.2886809727875214)(7.7,0.2866875008238243)(7.75,0.2881692904549133)(7.8,0.2860031451112796)(7.85,0.2852525910175307)(7.9,0.28134358654011604)(7.95,0.28032349657548034)(8.0,0.28163649505955046)(8.05,0.23378254204228827)(8.1,0.236511385073327)(8.15,0.23811296050753564)(8.2,0.24113675597874693)(8.25,0.2400814478458405)(8.3,0.2393943311248331)(8.35,0.23746223806656028)(8.4,0.23847221504428723)(8.45,0.24062003672898144)(8.5,0.2383445968845033)(8.55,0.23663272107267114)(8.6,0.23611413845403675)(8.65,0.2344122190800203)(8.7,0.23356606810185318)(8.75,0.23750367972935227)(8.8,0.2341025975714119)(8.85,0.23363992577624057)(8.9,0.23246115468904494)(8.95,0.22957148294193633)(9.0,0.23161705853301362)(9.05,0.19635913423884357)(9.1,0.19611084128898273)(9.15,0.19870920363338365)(9.2,0.2025615947440862)(9.25,0.19957932043265145)(9.3,0.19940810565240435)(9.35,0.19904914362071197)(9.4,0.19958866414135482)(9.45,0.19710792644606126)(9.5,0.19642878725634572)(9.55,0.19604335557118904)(9.6,0.19710953350575616)(9.65,0.19490696540243516)(9.7,0.1948197536593701)(9.75,0.196878825789012)(9.8,0.19500564041851437)(9.85,0.19706603154559668)(9.9,0.19598835221176855)(9.95,0.1927113793023761)(10.0,0.194791830145515)(10.05,0.17021746888148498)(10.1,0.17205125574284325)(10.15,0.173632729773847)(10.2,0.17348648741951853)(10.25,0.1732203518938148)(10.3,0.17218025998715653)(10.35,0.17356837270298137)(10.4,0.17428772577672696)(10.45,0.17202968474791963)(10.5,0.17357329520764742)(10.55,0.17196992733454064)(10.6,0.17241232974538745)(10.65,0.17201581472515615)(10.7,0.17213438743873977)(10.75,0.16936745687102273)(10.8,0.17061256815119408)(10.85,0.17067285885760847)(10.9,0.1699328735398011)(10.95,0.16804594354549487)(11.0,0.169070875282168)(11.05,0.15472957431926715)(11.1,0.15427443364998922)(11.15,0.15432553558255893)(11.2,0.154409503446434)(11.25,0.15425017166352809)(11.3,0.156108831550827)(11.35,0.1568358787239285)(11.4,0.15797202928437945)(11.45,0.15683259913998282)(11.5,0.1568705116687038)(11.55,0.15679193327516353)(11.6,0.15547979322001146)(11.65,0.1554303340363767)(11.7,0.15562839682113874)(11.75,0.15360715485253199)(11.8,0.15677349245362118)(11.85,0.15504375450055757)(11.9,0.15435311261558393)(11.95,0.15525422930927604)(12.0,0.15532267236533898)(12.05,0.14599388375472452)(12.1,0.14560385715678312)(12.15,0.14559735948855884)(12.2,0.14669212634946208)(12.25,0.14509142095033475)(12.3,0.14583206558586564)(12.35,0.14489649051808598)(12.4,0.14408240028384747)(12.45,0.14661059559816472)(12.5,0.14707425012138814)(12.55,0.14700707688452905)(12.6,0.14669286960019298)(12.65,0.14650220243420156)(12.7,0.14536765971945936)(12.75,0.1463698668578221)(12.8,0.1448846063524686)(12.85,0.14616061051090873)(12.9,0.14599058532260922)(12.95,0.14617440317694577)(13.0,0.14630588564095745)(13.05,0.13996310685882754)(13.1,0.1393131499281146)(13.15,0.14098154918026065)(13.2,0.14084821556087374)(13.25,0.14051643064736236)(13.3,0.14018420569131285)(13.35,0.14075791515815228)(13.4,0.14111708237144857)(13.45,0.14130654223026104)(13.5,0.14131520504332098)(13.55,0.14130077639509583)(13.6,0.14171326183980218)(13.65,0.1412639449231731)(13.7,0.14240242294844185)(13.75,0.14075396961030195)(13.8,0.14034719724107075)]{}; coordinates [(0.05,7.31327150769035)(0.1,7.143370456214577)(0.15,7.096564175939285)(0.2,7.044790967686032)(0.25,7.033707145522158)(0.3,6.513934798862032)(0.35,6.277778075447159)(0.4,6.135633996898611)(0.45,6.086771019595773)(0.5,6.036062638698562)(0.55,5.85841692046921)(0.6,5.76153 7151824908)(0.65,5.669668946907954)(0.7,5.5618716714121)(0.75,5.508595556627665)(0.8,5.521746317191602)(0.85,5.454713362478807)(0.9,5.409197263481727)(0.95,5.361511483943304)(1.0,5.32558175712414)(1.05,6.571283826927614)(1.1,5.941341542608717)(1.15,5.672227752156729)(1.2,5.367510257647911)(1.25,5.103052415357794)(1.3,5.216907683192228)(1.35,5.1634177364948854)(1.4,5.0262006089388)(1.45,5.004573093741681)(1.5,4.9369792326951565)(1.55,4.991686224822758)(1.6,4.904554363958364)(1.65,4.86230638387474)(1.7,4.800099891803548)(1.75,4.752969822088481)(1.8,4.195425298899062)(1.85,4.1325563101351275)(1.9,3.862215712479522)(1.95,2.3119604725150538)(2.0,2.1932521380085106)(2.05,1.8076599615519693)(2.1,1.893867885686234)(2.15,2.09217014270637)(2.2,1.6388031145521467)(2.25,1.183887861118845)(2.3,0.9393959953050375)(2.35,1.0360280224919796)(2.4,0.902715868733564)(2.45,0.8701728656182541)(2.5,0.8530981584232689)(2.55,0.8523652825310818)(2.6,0.7902490341819817)(2.65,0.7121082348417106)(2.7,0.7168022990420073)(2.75,0.6996332956050793)(2.8,0.6869204432631365)(2.85,0.7634602375183843)(2.9,0.9237150630390077)(2.95,0.8829274024419714)(3.0,1.0655912224841886)(3.05,1.1041935573617696)(3.1,0.9801086652706652)(3.15,1.2791274279716143)(3.2,1.0386245532617289)(3.25,0.9670872797777348)(3.3,0.8075987744569829)(3.35,0.7675510378478605)(3.4,0.745919101069802)(3.45,0.7281927557884748)(3.5,0.7146347701262296)(3.55,0.6643794133140036)(3.6,0.7718303451846508)(3.65,0.6820859987378733)(3.7,0.6560017439948129)(3.75,0.6749199501103071)(3.8,0.6317969715993006)(3.85,0.6472489402152113)(3.9,0.6356777870430634)(3.95,0.627025745401014)(4.0,0.6407769051786031)(4.05,2.2003999952446343)(4.1,0.9502262506669024)(4.15,0.7063525166757748)(4.2,0.6592243114547677)(4.25,0.8381062405607795)(4.3,0.6369567954285698)(4.35,0.683912570738726)(4.4,0.5994476770604962)(4.45,0.5785557998699196)(4.5,0.5435079788573322)(4.55,0.546081216941514)(4.6,0.549548441811794)(4.65,0.5396482143161417)(4.7,0.5355926283571498)(4.75,0.5374838501849637)(4.8,0.5281201176410372)(4.85,0.5345099572725356)(4.9,0.5446589585150555)(4.95,0.5313742647692483)(5.0,0.527801184113665)(5.05,0.5378506338169199)(5.1,0.524463981813958)(5.15,0.606183907300243)(5.2,0.8818107453491688)(5.25,0.6694209125357735)(5.3,0.5687644245495097)(5.35,0.5286447899618185)(5.4,0.510928972428597)(5.45,0.5027403169579016)(5.5,0.5068918687139804)(5.55,0.4858232888306256)(5.6,0.4876644352847015)(5.65,0.4812709507751928)(5.7,0.48244556438514924)(5.75,0.4809464291151954)(5.8,0.4913357412553016)(5.85,0.4808105946610805)(5.9,0.47413439011114045)(5.95,0.4778536207602722)(6.0,0.49984409207058184)(6.05,0.4772579906159885)(6.1,0.4814323262394445)(6.15,0.47168606304040345)(6.2,0.4832287902518289)(6.25,0.5849520292240281)(6.3,0.4813142523341482)(6.35,0.4703213355217343)(6.4,0.5429019091599019)(6.45,0.4524530469007377)(6.5,0.4507396644732726)(6.55,0.456799714238745)(6.6,0.4459158729454588)(6.65,0.4441275589206594)(6.7,0.44713937210399884)(6.75,0.4446689600240784)(6.8,0.44701157973056577)(6.85,0.44496381262071005)(6.9,0.4459094360653661)(6.95,0.44965607838231114)(7.0,0.43801533820990934)(7.05,0.43038436084960857)(7.1,0.4441346359737554)(7.15,0.4454563215903606)(7.2,0.43549588611302953)(7.25,0.4351369378625424)(7.3,0.44593548781255493)(7.35,0.43146567992163487)(7.4,0.42721885150152233)(7.45,0.435803855089571)(7.5,0.4173875668500288)(7.55,0.41532561903157866)(7.6,0.41656469299542936)(7.65,0.41159815681416123)(7.7,0.4059642884956082)(7.75,0.4013920536748351)(7.8,0.4027550675045386)(7.85,0.39939195981446646)(7.9,0.40014460858911244)(7.95,0.40150894079317484)(8.0,0.4022751446983911)(8.05,0.39814097422518663)(8.1,0.4103965829340957)(8.15,0.40178965313721526)(8.2,0.39496847632423704)(8.25,0.3956622581690208)(8.3,0.4078248068190982)(8.35,0.39229624503142996)(8.4,0.3961592954199562)(8.45,0.3887864903916739)(8.5,0.3893605724737341)(8.55,0.386937717828361)(8.6,0.3898630228629124)(8.65,0.3838018309676311)(8.7,0.3781523303942078)(8.75,0.38090790288521514)(8.8,0.37633312424969056)(8.85,0.3786578422837952)(8.9,0.37252778509835344)(8.95,0.37546474147635867)(9.0,0.37260336080817247)(9.05,0.376464175272047)(9.1,0.3732427030460814)(9.15,0.37597665496514776)(9.2,0.37411252941404216)(9.25,0.37528045744778576)(9.3,0.3702179617493421)(9.35,0.37101509241554664)(9.4,0.36766884448496745)(9.45,0.3673689652844533)(9.5,0.3682124560299719)(9.55,0.3665438636022175)(9.6,0.36391215878715577)(9.65,0.3649233906174687)(9.7,0.3692903728437124)(9.75,0.3610671598056718)(9.8,0.35954779080661403)(9.85,0.35764603105491216)(9.9,0.3585678543913464)(9.95,0.35872688749226317)(10.0,0.3588543761331305)(10.05,0.36179310366419787)(10.1,0.36271182257295836)(10.15,0.36388698866974817)(10.2,0.36266872432613395)(10.25,0.3611390608369752)(10.3,0.36189891377477534)(10.35,0.3620060118486594)(10.4,0.3597890859349412)(10.45,0.3599425072518072)(10.5,0.35759960130056284)(10.55,0.35873986264685104)(10.6,0.3572212268129103)(10.65,0.3575912750248173)(10.7,0.3542537929783172)(10.75,0.3523967013786011)(10.8,0.3535606025645394)(10.85,0.35067814004486564)(10.9,0.3517952718669877)(10.95,0.352815097726953)(11.0,0.35307456430218936)(11.05,0.353728625634799)(11.1,0.3530444324442432)(11.15,0.3558057073128377)(11.2,0.3565758488242482)(11.25,0.3529218321183765)(11.3,0.35897407442763396)(11.35,0.35807766856733897)(11.4,0.3573719706633924)(11.45,0.3551160487677389)(11.5,0.35531554537284515)(11.55,0.35376593280028806)(11.6,0.35515405551826323)(11.65,0.3536014429070223)(11.7,0.3531253858951194)(11.75,0.35328718189538116)(11.8,0.3528895016954366)(11.85,0.3511520821221687)(11.9,0.35110946236970314)(11.95,0.3509616462371662)(12.0,0.35095757922409443)(12.05,0.35225772039814723)(12.1,0.35195963765931954)(12.15,0.35213996358476524)(12.2,0.3515579459333905)(12.25,0.35141987786598017)(12.3,0.3515025855199725)(12.35,0.35194086774557776)(12.4,0.3516738889571824)(12.45,0.3514006317110884)(12.5,0.3534720984215102)(12.55,0.35180696732972644)(12.6,0.3553712569234378)(12.65,0.35194921793441175)(12.7,0.3520731781954397)(12.75,0.35095089626451126)(12.8,0.3496683235350435)(12.85,0.35011182036937827)(12.9,0.35015664114356215)(12.95,0.35004691543555394)(13.0,0.3524928217894321)(13.05,0.34993988338796045)(13.1,0.3502924827643352)(13.15,0.3495451940975693)(13.2,0.34943358041851774)(13.25,0.3498431567330918)(13.3,0.3505865914240574)(13.35,0.35010621460357594)(13.4,0.35014071448877215)(13.45,0.35021911662389166)(13.5,0.34958669858429126)(13.55,0.34916798778972846)(13.6,0.3497797769891338)(13.65,0.3499364186939817)(13.7,0.34958379556442354)(13.75,0.3499611355820442)(13.8,0.35006689107774425)]{}; coordinates [(0.05,8.36147686257847)(0.1,7.087779546539288)(0.15,6.956401353756784)(0.2,6.891344167225708)(0.25,6.913308596396137)(0.3,6.793939218289796)(0.35,6.7703217853644535)(0.4,6.789018834163316)(0.45,6.750823025863694)(0.5,6.666381235042926)(0.55,6.629496634665033)(0.6,6.659 824904168242)(0.65,6.641187011197276)(0.7,6.616349014858145)(0.75,6.601156350524258)(0.8,6.599295703053511)(0.85,6.569010914991907)(0.9,6.492166849498066)(0.95,6.4914782330058225)(1.0,6.565239047421843)(1.05,6.353977897864798)(1.1,6.337009267381585)(1.15,6.238737117706011)(1.2,6.279657682898293)(1.25,6.237854536854853)(1.3,6.507097514646458)(1.35,6.295325473655188)(1.4,6.133854321015416)(1.45,6.051639327036881)(1.5,5.949523819401636)(1.55,5.894726187674295)(1.6,5.812155398738772)(1.65,5.757765752187907)(1.7,5.690086124179352)(1.75,5.64514590765875)(1.8,5.6004439964759785)(1.85,5.5366010399336085)(1.9,5.5043285924464795)(1.95,5.474721879685134)(2.0,5.431273537406448)(2.05,5.299171616419968)(2.1,5.283055795859911)(2.15,5.299449534621267)(2.2,5.270986859100641)(2.25,5.322613355690559)(2.3,5.346455348146981)(2.35,5.322521157719037)(2.4,5.308348599556917)(2.45,5.28410244532708)(2.5,5.298549590325793)(2.55,5.22381157043027)(2.6,5.1934652748882675)(2.65,5.167579270871915)(2.7,5.13308044482291)(2.75,5.116886938808302)(2.8,5.08706000746674)(2.85,5.059523773598495)(2.9,5.047922548885686)(2.95,5.018213427418749)(3.0,5.011286271273447)(3.05,4.849764290090971)(3.1,4.869194597705853)(3.15,4.856437074408847)(3.2,4.83937898490505)(3.25,4.83520469924271)(3.3,4.850833029850868)(3.35,4.821819494936261)(3.4,4.799734119401401)(3.45,4.769805147465937)(3.5,4.7440523768564375)(3.55,4.715783926757411)(3.6,4.680396862015482)(3.65,4.519017401689703)(3.7,3.611077643083086)(3.75,2.7145109946414787)(3.8,5.338260129474434)(3.85,3.747336953997253)(3.9,2.4382760872776728)(3.95,2.017974135514813)(4.0,1.7866086412985984)(4.05,1.4537677646349585)(4.1,1.2844864776347258)(4.15,1.3205518489545662)(4.2,1.3177540717627558)(4.25,1.1000515005992362)(4.3,1.2565498409691)(4.35,1.0284601559710471)(4.4,1.6639142509352305)(4.45,2.277802932898003)(4.5,1.688669654132068)(4.55,1.5086385126294857)(4.6,1.3836458450329117)(4.65,1.4168502843562794)(4.7,1.3251740084531607)(4.75,1.2399547195479035)(4.8,1.2016826017420845)(4.85,1.193187566719414)(4.9,1.1052075041195877)(4.95,1.0916815887868054)(5.0,1.227481131099884)(5.05,3.731405240647728)(5.1,1.595921502929868)(5.15,1.6287866071549628)(5.2,1.1693819668034569)(5.25,1.0227411211163344)(5.3,1.060968547434411)(5.35,0.9373470100356998)(5.4,0.9143111108049742)(5.45,0.9324388877435196)(5.5,0.9017801859684789)(5.55,0.8467012063174592)(5.6,0.7561451494492134)(5.65,0.7261813718879124)(5.7,0.8365751177904587)(5.75,0.6870057041623114)(5.8,0.7040253617880796)(5.85,0.6486004003228969)(5.9,0.849315663936879)(5.95,0.6515473064944542)(6.0,0.6274015380690654)(6.05,0.5469594885452629)(6.1,0.5596752025823619)(6.15,0.5405562094658388)(6.2,0.6264088537408444)(6.25,0.5264877924744068)(6.3,3.1628946025259346)(6.35,0.9651022218598182)(6.4,0.8202483980598586)(6.45,0.811192612716552)(6.5,0.6930268979816092)(6.55,0.7044886241459954)(6.6,0.6446851514494661)(6.65,1.0216514529797331)(6.7,0.7047580556667882)(6.75,0.6145432928754362)(6.8,0.5809761073447133)(6.85,0.5931071380921246)(6.9,0.5938291366307634)(6.95,0.5550359052584737)(7.0,0.5873400272246482)(7.05,0.4953239204004775)(7.1,0.4933301015095071)(7.15,0.5015912780720567)(7.2,0.4984140595180192)(7.25,0.49899168713739706)(7.3,0.48141581444658965)(7.35,0.49986166084954936)(7.4,0.6164591976426622)(7.45,0.5408056237195585)(7.5,0.46855791129842983)(7.55,2.639946300882915)(7.6,0.9683470261350748)(7.65,0.7119010694240537)(7.7,0.6798161403838997)(7.75,0.6844185041160918)(7.8,0.5969284448945487)(7.85,0.5522090927063263)(7.9,0.5408058681871746)(7.95,0.541594937929843)(8.0,0.5361774353624027)(8.05,0.45934781597754193)(8.1,0.467124772800522)(8.15,0.48628648329856744)(8.2,0.4510879466644415)(8.25,0.43539085398862715)(8.3,0.42758256755488666)(8.35,0.45386695673113436)(8.4,0.41353631356818904)(8.45,0.40875285332821)(8.5,0.470714477619535)(8.55,0.4476859825723753)(8.6,0.41178975316062416)(8.65,0.41409153732420695)(8.7,0.42071576418231654)(8.75,0.41407303741308316)(8.8,0.6963166251481494)(8.85,0.6029217017991427)(8.9,0.5778652974725801)(8.95,0.5674910784490138)(9.0,0.5517435365414899)(9.05,0.48332372394751205)(9.1,0.48649555553153745)(9.15,0.47143376593769887)(9.2,0.47480104879452967)(9.25,0.4745042637308948)(9.3,0.46718284520077696)(9.35,0.4638348382167879)(9.4,0.4706005633812938)(9.45,0.4553220574669162)(9.5,0.44198531979681405)(9.55,0.4435088121125537)(9.6,0.45329555166916824)(9.65,0.44276287100060524)(9.7,0.442097561432701)(9.75,0.43851891546153265)(9.8,0.4397135505865975)(9.85,0.43071785599525597)(9.9,0.43472542192214353)(9.95,0.4502281979840952)(10.0,0.4604789775763778)(10.05,0.39327514386481816)(10.1,0.37353911626052605)(10.15,0.37186719543243624)(10.2,0.37617515663586243)(10.25,0.3880402005951219)(10.3,0.41530309089439826)(10.35,0.391809876798689)(10.4,0.4004076071679583)(10.45,0.4121755539722973)(10.5,0.3890708977515186)(10.55,0.371178139279341)(10.6,0.3731530032218216)(10.65,0.36566111838582543)(10.7,0.35972432709610924)(10.75,0.3555115887464266)(10.8,0.3667154761877675)(10.85,0.37813943682608536)(10.9,0.36235910568937607)(10.95,0.352796679071546)(11.0,0.35342561338927286)(11.05,0.3256709661046176)(11.1,0.31417925456030243)(11.15,0.32337298398554165)(11.2,0.32359892885639463)(11.25,0.3183816016862058)(11.3,0.3121702287313061)(11.35,0.3173291713377299)(11.4,0.31794701084153293)(11.45,0.3264974203883656)(11.5,0.3228822445431416)(11.55,0.32368360829388276)(11.6,0.3432115178602605)(11.65,0.3317823626175288)(11.7,0.31710633661916265)(11.75,0.3055470084266608)(11.8,0.3077038036169465)(11.85,0.305442429368438)(11.9,0.3048927533215612)(11.95,0.305844847528969)(12.0,0.31176753333394874)(12.05,0.275911255740873)(12.1,0.2686906274412183)(12.15,0.27019140716648266)(12.2,0.2729007527244132)(12.25,0.2706414367792938)(12.3,0.27151044232460897)(12.35,0.2692956842289182)(12.4,0.26803682424191033)(12.45,0.27199694821781845)(12.5,0.27440088703385834)(12.55,0.28052081988471406)(12.6,0.274642769137379)(12.65,0.27611998357501666)(12.7,0.27709938224490804)(12.75,0.27491979384734255)(12.8,0.2699032275278034)(12.85,0.28081778175483074)(12.9,0.27704958066025276)(12.95,0.27977529720969696)(13.0,0.2771637073227553)(13.05,0.24173486330758034)(13.1,0.24203370417946882)(13.15,0.24133969176060444)(13.2,0.26050911348714495)(13.25,0.2538886483673846)(13.3,0.2483952256085425)(13.35,0.24583142503425578)(13.4,0.2508403987525299)(13.45,0.24464024939487256)(13.5,0.24040872987382952)(13.55,0.23825432196919052)(13.6,0.24199111846377572)(13.65,0.23689362124625565)(13.7,0.23793798997131846)(13.75,0.23440082422589553)(13.8,0.23680546642906392)(13.85,0.23753477112935967)(13.9,0.23765892422434792)(13.95,0.23626642046337482)(14.0,0.24879851050981824)(14.05,0.2176316944862133)(14.1,0.2117841464111614)(14.15,0.21215681494182725)(14.2,0.21113570894378988)(14.25,0.21068890394109674)(14.3,0.21289732621508406)(14.35,0.21179316825487576)(14.4,0.21123644313390705)(14.45,0.21018433191409952)(14.5,0.20924031936901755)(14.55,0.2082008496321248)(14.6,0.211416357501672)(14.65,0.2091562850392054)(14.7,0.21357299770677554)(14.75,0.2109053960702659)(14.8,0.2084753061048062)(14.85,0.20780465707758483)(14.9,0.20768661993170678)(14.95,0.2067930956474127)(15.0,0.21129681248572815)(15.05,0.18808739188469126)(15.1,0.1870490797599572)(15.15,0.18651302326838193)(15.2,0.18713618820283837)(15.25,0.1873484381227392)(15.3,0.18824788385411068)(15.35,0.18803564783544882)(15.4,0.1881285302922155)(15.45,0.1872583576802411)(15.5,0.18829915668771852)(15.55,0.18486127865776317)(15.6,0.18551816568866517)(15.65,0.18556059434911346)(15.7,0.18646130294318358)(15.75,0.18493185894019223)(15.8,0.18510415715199763)(15.85,0.18447777026538206)(15.9,0.18558540611807472)(15.95,0.1862021922977735)(16.0,0.18661393719402283)(16.05,0.16671563116838153)(16.1,0.1691445470950074)(16.15,0.16979700732718092)(16.2,0.17076341345091414)(16.25,0.16971231999132202)(16.3,0.17188425318905187)(16.35,0.17104377109244637)(16.4,0.17020460801417683)(16.45,0.1706813540505914)(16.5,0.1714091420544386)(16.55,0.16985075946277747)(16.6,0.17028544845772942)(16.65,0.17121452815883834)(16.7,0.172104895521336)(16.75,0.17248935602044854)(16.8,0.17138726210456345)(16.85,0.17180803791728694)(16.9,0.1721384894406991)(16.95,0.17172266625626917)(17.0,0.17041520767391874)(17.05,0.15628760963598232)(17.1,0.15835498886124869)(17.15,0.15723237971481008)(17.2,0.15743890630045704)(17.25,0.1555523751144341)(17.3,0.157578688106103)(17.35,0.1585394015955068)(17.4,0.1571959198250993)(17.45,0.1583224354171047)(17.5,0.15859578447860764)]{}; coordinates [(0.05,7.284510047350189)(0.1,7.144500804237097)(0.15,7.0982345795054576)(0.2,7.10187558729314)(0.25,7.016213424972881)(0.3,7.038880541231043)(0.35,7.00836502852821)(0.4,7.028937194201535)(0.45,6.891526494591193)(0.5,7.023393263447894)(0.55,6.980129251104713)(0.6,7.0028 227770000555)(0.65,7.017146062761193)(0.7,6.941236645262024)(0.75,6.94838884916648)(0.8,6.862255359703437)(0.85,6.8909952266394585)(0.9,6.842717504967207)(0.95,6.802045118747749)(1.0,6.869646155248258)(1.05,6.908088649573546)(1.1,6.801852820760691)(1.15,6.7677091418390045)(1.2,6.856313934390435)(1.25,6.772818368382903)(1.3,6.2658351126086425)(1.35,6.153824857733203)(1.4,5.996293442013531)(1.45,5.931021309176699)(1.5,5.85416503932825)(1.55,5.755459285215447)(1.6,5.7073459198067304)(1.65,5.6346294943910324)(1.7,5.587628188243312)(1.75,5.505878474165089)(1.8,5.463588298570651)(1.85,5.42053907443275)(1.9,5.37867123048525)(1.95,5.337517491429558)(2.0,5.296924720361896)(2.05,5.275188172747828)(2.1,5.257762661362923)(2.15,5.225220184636543)(2.2,5.243202515605226)(2.25,5.354411627267228)(2.3,5.296315415325689)(2.35,5.301012431556546)(2.4,5.250359018731557)(2.45,5.232545472914656)(2.5,5.223382527077324)(2.55,5.229482378946343)(2.6,5.1666637490399046)(2.65,5.134815464680996)(2.7,5.103838247735424)(2.75,5.073562664719551)(2.8,5.05250512931135)(2.85,5.046093682715471)(2.9,5.06145022369034)(2.95,5.0023380315902415)(3.0,4.998690762112345)(3.05,5.039579570701205)(3.1,4.913840489680784)(3.15,4.9022219791257955)(3.2,4.871453107530218)(3.25,4.853035583114138)(3.3,4.842479156009748)(3.35,4.7871711324553194)(3.4,4.7783358287077835)(3.45,4.747992448572268)(3.5,4.701351230534924)(3.55,4.659714340786576)(3.6,4.603932783002197)(3.65,3.9705450756395746)(3.7,3.076037223624638)(3.75,2.619147462433169)(3.8,5.037940047035359)(3.85,2.791246845735491)(3.9,2.213846816947302)(3.95,2.009315409676204)(4.0,1.8106100356833499)(4.05,1.8136784822978511)(4.1,2.1825501563211356)(4.15,1.6581501476195422)(4.2,1.5924384094600297)(4.25,1.4898761296360703)(4.3,1.5344977415267065)(4.35,1.4115853478505285)(4.4,3.133771792838778)(4.45,2.013930442734659)(4.5,1.8997713034210002)(4.55,1.7710546231723105)(4.6,1.716222883510592)(4.65,1.772965642929499)(4.7,1.6315819512275844)(4.75,1.5940594414139035)(4.8,1.523322151983055)(4.85,1.5175653097379411)(4.9,1.4885988892127107)(4.95,1.5565348705026767)(5.0,1.5367038143522558)(5.05,1.8566733987569832)(5.1,1.2945207415152846)(5.15,1.3191412917684384)(5.2,1.1453576659956173)(5.25,1.0327809672569246)(5.3,0.9914895763129751)(5.35,0.9817495438541516)(5.4,0.9228256174388392)(5.45,0.9885034218630788)(5.5,1.1360146595499427)(5.55,0.8568421198319067)(5.6,0.8393934705866403)(5.65,0.8020001272157299)(5.7,0.8512917502463171)(5.75,0.7502818775989023)(5.8,0.7651483219938804)(5.85,0.716928231735727)(5.9,0.783451341348358)(5.95,0.728624750407169)(6.0,0.7290091213043749)(6.05,0.6972923254278589)(6.1,0.686521423823444)(6.15,0.8701164477587924)(6.2,0.6757946063711688)(6.25,0.6552704694435318)(6.3,1.1944516797758433)(6.35,1.1711068107309284)(6.4,0.8185461771290142)(6.45,0.8434568560022421)(6.5,0.7232312182748364)(6.55,0.7475858886688186)(6.6,0.7527021246686608)(6.65,0.9243103787731746)(6.7,0.7225538975943675)(6.75,0.7296733015396291)(6.8,0.6683246573236175)(6.85,0.6568243189285489)(6.9,0.695916281413891)(6.95,0.6786030600717039)(7.0,0.7277883241465226)(7.05,0.6493407366346015)(7.1,0.6239110223603523)(7.15,0.6399305631236993)(7.2,0.6329561983713631)(7.25,0.60529972810596)(7.3,0.6271005924241431)(7.35,0.644209932709939)(7.4,0.6445501027782616)(7.45,0.6104862094828641)(7.5,0.5879846645357041)(7.55,0.9113004301160925)(7.6,0.8184317401192656)(7.65,0.7811752536254223)(7.7,0.6924338799678675)(7.75,0.7259335766055859)(7.8,0.6461257827693523)(7.85,0.6090586995298057)(7.9,0.6407407602225218)(7.95,0.6454142452257102)(8.0,0.627494522307827)(8.05,0.6039336632296988)(8.1,0.7251494603799185)(8.15,0.610541027435588)(8.2,0.566768312997611)(8.25,0.5905330463499354)(8.3,0.5883448133409777)(8.35,0.5686404252566363)(8.4,0.5376648585685305)(8.45,0.553046735409684)(8.5,0.641126840002947)(8.55,0.5450667910522002)(8.6,0.5335885059410661)(8.65,0.5904337015631616)(8.7,0.5267513371755437)(8.75,0.5276587242773417)(8.8,0.6812036075406288)(8.85,0.6195049345887622)(8.9,0.6229725070529364)(8.95,0.6320181141356593)(9.0,0.593511653366391)(9.05,0.5756018579725616)(9.1,0.5713821682531962)(9.15,0.5630813718176203)(9.2,0.5565703175386992)(9.25,0.588746850821423)(9.3,0.5434969998455544)(9.35,0.560590078497035)(9.4,0.551265813210164)(9.45,0.5216770776745502)(9.5,0.5289840086201124)(9.55,0.5309651056938715)(9.6,0.5293153440173859)(9.65,0.5248371362240721)(9.7,0.5213396320944005)(9.75,0.5161634663210435)(9.8,0.5281758193958488)(9.85,0.5091731750208027)(9.9,0.5162380439113359)(9.95,0.5149516102802151)(10.0,0.5183014532663798)(10.05,0.5080327166404609)(10.1,0.49745893366932237)(10.15,0.49255865499871937)(10.2,0.48911887729135134)(10.25,0.49548642699264234)(10.3,0.5100619320653466)(10.35,0.4873111300164725)(10.4,0.5261531135561496)(10.45,0.5159834101958473)(10.5,0.4827410740608397)(10.55,0.4771361327506881)(10.6,0.47449085670615193)(10.65,0.46438217332846254)(10.7,0.45920441052062644)(10.75,0.4662707396522154)(10.8,0.4792271932294437)(10.85,0.47025402826522766)(10.9,0.46084472149584804)(10.95,0.4675677304019946)(11.0,0.47802228342070807)(11.05,0.4609966692164603)(11.1,0.4575407570121939)(11.15,0.45737454168005404)(11.2,0.4613332423074647)(11.25,0.4518459602980339)(11.3,0.454735194768275)(11.35,0.45144597276474646)(11.4,0.4513362628299166)(11.45,0.4512139497376541)(11.5,0.4494009152245356)(11.55,0.4482411059775608)(11.6,0.46642398248467953)(11.65,0.43920920007813574)(11.7,0.43184681422828786)(11.75,0.43015337061830805)(11.8,0.4315177774203358)(11.85,0.43014539602153884)(11.9,0.43190362833041995)(11.95,0.42454873807045473)(12.0,0.5115795696298305)(12.05,0.436119251261777)(12.1,0.42166598082628365)(12.15,0.42069585186902586)(12.2,0.4193301388622342)(12.25,0.4300122379696667)(12.3,0.41171160859171346)(12.35,0.41869579407727137)(12.4,0.4143992733546075)(12.45,0.4152817853770437)(12.5,0.42604889993025336)(12.55,0.41222874658748676)(12.6,0.41096242486573764)(12.65,0.41461933034825765)(12.7,0.40981096404092077)(12.75,0.40997121862837793)(12.8,0.40785289435821304)(12.85,0.4161285474153934)(12.9,0.40695881518735894)(12.95,0.42111899128687513)(13.0,0.41485737159626096)(13.05,0.417889313736227)(13.1,0.41599286276600217)(13.15,0.4054164737651971)(13.2,0.42191436543885236)(13.25,0.41073022797305475)(13.3,0.4131166115434915)(13.35,0.41086413419299045)(13.4,0.40705188460457653)(13.45,0.40649716583441736)(13.5,0.40133709581898086)(13.55,0.39922914290430456)(13.6,0.39990086626228216)(13.65,0.39804421574090326)(13.7,0.3919134963209619)(13.75,0.3975650416560536)(13.8,0.3945073775819782)(13.85,0.39761990254352003)(13.9,0.3918988319510305)(13.95,0.3894275454732716)(14.0,0.40125941824831596)(14.05,0.3937756955399429)(14.1,0.3932935309094595)(14.15,0.3926086526797338)(14.2,0.3899825458194731)(14.25,0.3884244816026361)(14.3,0.389191168909268)(14.35,0.3876010846281789)(14.4,0.38645054187115635)(14.45,0.38761973342268347)(14.5,0.3851358801124681)(14.55,0.3805860728021541)(14.6,0.38382657036772877)(14.65,0.3816649341960273)(14.7,0.3845446599140139)(14.75,0.3815780159997129)(14.8,0.3810814969685972)(14.85,0.3801526161139471)(14.9,0.37693230649879417)(14.95,0.37519131016125024)(15.0,0.3768956554024007)(15.05,0.38289748323089534)(15.1,0.38599630448018535)(15.15,0.3788731733164722)(15.2,0.38250278569470647)(15.25,0.38012905448135387)(15.3,0.37941365499498103)(15.35,0.3760880637010063)(15.4,0.3785673687448491)(15.45,0.3745466052286634)(15.5,0.37341473517625634)(15.55,0.375835577469835)(15.6,0.371022957864129)(15.65,0.3693070588665217)(15.7,0.37143321588212436)(15.75,0.36836973785765326)(15.8,0.3692878793940014)(15.85,0.3647549383679243)(15.9,0.3656485102535876)(15.95,0.3663353942887278)(16.0,0.36526953579945504)(16.05,0.3673265885135559)(16.1,0.3730798690717557)(16.15,0.37058217163486273)(16.2,0.3701459607432727)(16.25,0.3694499572423721)(16.3,0.37483035438973167)(16.35,0.3716950631516435)(16.4,0.3690407503496395)(16.45,0.36873262850892075)(16.5,0.36624646775804476)(16.55,0.3664013800450356)(16.6,0.3658243738902642)(16.65,0.3679525267678939)(16.7,0.3672182911719885)(16.75,0.36642103432867773)(16.8,0.3645050118980557)(16.85,0.3639822456001041)(16.9,0.3637398265602213)(16.95,0.36308813200098927)(17.0,0.36385913479526927)(17.05,0.36830717509812083)(17.1,0.3654394976620592)(17.15,0.3661582789922279)(17.2,0.36347609338069975)(17.25,0.3634172195604509)(17.3,0.3640110207289361)(17.35,0.36353285161593796)(17.4,0.36410270111814486)(17.45,0.36653427830435586)(17.5,0.3646317606607597)]{}; coordinates [(0.05,8.37727126625171)(0.1,7.147839296749115)(0.15,7.0058040208758605)(0.2,6.933355712243054)(0.25,6.942551868919154)(0.3,6.828752546300098)(0.35,6.470315988741071)(0.4,6.313455344380915)(0.45,6.180048552988477)(0.5,6.084835267635532)(0.55,5.9221377990159825)(0.6,5.76 2175892159592)(0.65,5.6706264542541795)(0.7,5.554662350273768)(0.75,5.491104288045348)(0.8,5.63132204420064)(0.85,5.508456489410134)(0.9,5.415265663871575)(0.95,5.361820925624268)(1.0,5.31457684128711)(1.05,5.38907664834944)(1.1,5.276866237033959)(1.15,5.18834371715672)(1.2,5.14488815600806)(1.25,5.0827023011947166)(1.3,5.240296265194806)(1.35,5.1282350430314)(1.4,5.047394499449821)(1.45,5.008562524456762)(1.5,4.953309347728223)(1.55,5.149804596308791)(1.6,5.172313438156097)(1.65,5.043898760213672)(1.7,5.001124076310532)(1.75,4.958519422388507)(1.8,4.388473387103625)(1.85,4.286311219147867)(1.9,4.216948968807611)(1.95,4.134518683588273)(2.0,4.0341782830686945)(2.05,3.8530841225075134)(2.1,3.7715436398307034)(2.15,3.680596984205215)(2.2,3.5840272602668923)(2.25,3.3776602348831344)(2.3,2.819014029029495)(2.35,1.9716637133570816)(2.4,1.7435220795177573)(2.45,1.386174376457276)(2.5,1.2498220798467674)(2.55,1.2924846994352412)(2.6,1.047453012094431)(2.65,1.304541164116448)(2.7,0.9983144555482816)(2.75,0.9845829568331153)(2.8,2.1886817921922406)(2.85,1.1311478549822997)(2.9,0.9737302621406014)(2.95,1.7040888168319728)(3.0,0.9905920102757524)(3.05,0.8029521441186549)(3.1,0.7678358998620268)(3.15,0.8450586813411698)(3.2,0.8979049185222719)(3.25,1.1123867101696552)(3.3,0.798554248092059)(3.35,0.8269363001416848)(3.4,0.8180170507683944)(3.45,0.714696590966373)(3.5,0.8342898129465923)(3.55,0.7396826993803787)(3.6,0.6964033802499358)(3.65,0.8700341773639886)(3.7,0.7249703533074198)(3.75,0.6722750181377676)(3.8,0.6512795827930983)(3.85,0.8861391787582299)(3.9,0.8473847848332654)(3.95,0.6689309149374855)(4.0,0.636854874358041)(4.05,0.5662308815338498)(4.1,0.6255771340628469)(4.15,0.982843582137996)(4.2,0.6095593631106377)(4.25,0.5756167876115524)(4.3,0.5750736228437927)(4.35,0.5854717564916435)(4.4,0.5803666987687668)(4.45,0.5648723952944276)(4.5,0.565937678222856)(4.55,0.5644513115485489)(4.6,0.5547750872261653)(4.65,0.6217968802738374)(4.7,0.5756904294877887)(4.75,0.6442067580404552)(4.8,0.5474826242405308)(4.85,0.5547844075655067)(4.9,0.5481657692773677)(4.95,0.5941025228555804)(5.0,0.5354819749851834)(5.05,0.46571767699361444)(5.1,0.4716934006279536)(5.15,0.4816419518115527)(5.2,0.4632571515824279)(5.25,0.43001132133502556)(5.3,0.40265429654271356)(5.35,0.39273500305467074)(5.4,0.3860483928091026)(5.45,0.3766884544896771)(5.5,0.3744676903340633)(5.55,0.3644089853790728)(5.6,0.36462465632395075)(5.65,0.35671335689533146)(5.7,0.3560469561577632)(5.75,0.350216717105434)(5.8,0.34849933831160884)(5.85,0.3445913156954386)(5.9,0.3399062452296746)(5.95,0.3410852009712044)(6.0,0.3394933363041897)(6.05,0.30335950197885553)(6.1,0.3026006277349023)(6.15,0.29951336030469683)(6.2,0.2998674540515362)(6.25,0.30033467610406256)(6.3,0.30048120475525875)(6.35,0.29666864585647734)(6.4,0.2961188431822177)(6.45,0.2971127960986731)(6.5,0.29537576341349375)(6.55,0.29635171396868226)(6.6,0.29493781959868826)(6.65,0.2924307521345077)(6.7,0.297250799216372)(6.75,0.29044310780115884)(6.8,0.2916717983556159)(6.85,0.2930762475529691)(6.9,0.2909220730953797)(6.95,0.2883574471610817)(7.0,0.2901049491367703)(7.05,0.2645641398927049)(7.1,0.26486892303033505)(7.15,0.2660009061850978)(7.2,0.2670578035866984)(7.25,0.2663456962023371)(7.3,0.26698386975033594)(7.35,0.2689380277847516)(7.4,0.265507015290703)(7.45,0.2686855760706562)(7.5,0.26483871155865707)(7.55,0.2689408515578891)(7.6,0.27235375305376014)(7.65,0.26515757929699757)(7.7,0.26731287846500723)(7.75,0.26711491854569147)(7.8,0.26895437161696367)(7.85,0.26802140721637974)(7.9,0.26366987299295996)(7.95,0.2627336593590726)(8.0,0.26483900426648005)(8.05,0.24057211405483625)(8.1,0.2433521805379187)(8.15,0.24595956832930072)(8.2,0.24423900371869944)(8.25,0.243054865266703)(8.3,0.24362228452352175)(8.35,0.2436642532315727)(8.4,0.2441017406977001)(8.45,0.24441686485184796)(8.5,0.24267684606383538)(8.55,0.2451696346610518)(8.6,0.24465703989912083)(8.65,0.24450411816877254)(8.7,0.24349980869325022)(8.75,0.24678850919663547)(8.8,0.24708386852781125)(8.85,0.24424006276547192)(8.9,0.24487293386858813)(8.95,0.24474495676620597)(9.0,0.24278386543405286)(9.05,0.22281424548559942)(9.1,0.22061170769175628)(9.15,0.2234080629929103)(9.2,0.2271503652089406)(9.25,0.22375251314566882)(9.3,0.2241393761101391)(9.35,0.22538049444681643)(9.4,0.2240981857590402)(9.45,0.2248583457819716)(9.5,0.22280743976571435)(9.55,0.2226283761703538)(9.6,0.22395239758383956)(9.65,0.22360257733551767)(9.7,0.22259402946044446)(9.75,0.22469266746495448)(9.8,0.22365172719633167)(9.85,0.2249130977352919)(9.9,0.2252883621767997)(9.95,0.22382185749039585)(10.0,0.22329502933971027)(10.05,0.20733601170250662)(10.1,0.2094615990287403)(10.15,0.21147635710387505)(10.2,0.21252156266623867)(10.25,0.21064742142424717)(10.3,0.20959092910846294)(10.35,0.2107714344369863)(10.4,0.21255682575978935)(10.45,0.21037311908162654)(10.5,0.2121155620164463)(10.55,0.21012304495819956)(10.6,0.2108499415916986)(10.65,0.21161681986687256)(10.7,0.21075747728735447)(10.75,0.2095466546365874)(10.8,0.2094765927810802)(10.85,0.20840894349810796)(10.9,0.21176543999106717)(10.95,0.20853450150326794)(11.0,0.20781677984637872)(11.05,0.20011574103903046)(11.1,0.20056428760271747)(11.15,0.20129911824274438)(11.2,0.1996751299060147)(11.25,0.2009118059883833)(11.3,0.19750838295055706)(11.35,0.19948313258074388)(11.4,0.20085690487936816)(11.45,0.20055780833736103)(11.5,0.20101260810833918)(11.55,0.20220536829271524)(11.6,0.20021237301960584)(11.65,0.20005134507898353)(11.7,0.20090138656317907)(11.75,0.1980169171609387)(11.8,0.20219691610020427)(11.85,0.20014976776251342)(11.9,0.20040670978049396)(11.95,0.2012409435944838)(12.0,0.2016615869063895)(12.05,0.1943017329894732)(12.1,0.19295711319373518)(12.15,0.19354277946064305)(12.2,0.19512023558525324)(12.25,0.1936165301859761)(12.3,0.19432242132721284)(12.35,0.19307519063128947)(12.4,0.19124721014149088)(12.45,0.19464820648308748)(12.5,0.19462831857551097)(12.55,0.19607104567012512)(12.6,0.1939415821004805)(12.65,0.19541776811236683)(12.7,0.19446713748958996)(12.75,0.19513715659305203)(12.8,0.19396890657242363)(12.85,0.19469563321364614)(12.9,0.1950821959238418)(12.95,0.1947026774512792)(13.0,0.19492150741119457)(13.05,0.19049936170411083)(13.1,0.18980614407409413)(13.15,0.19109689075442934)(13.2,0.19137267542628586)(13.25,0.1893494160376995)(13.3,0.19054429330649877)(13.35,0.1908683113861541)(13.4,0.1912866815401788)(13.45,0.19185154851778194)(13.5,0.19058270701517252)(13.55,0.19171591821417636)(13.6,0.1933786498368439)(13.65,0.19054290491593293)(13.7,0.19192038055243038)(13.75,0.1900627560768375)(13.8,0.1888835872123269)]{}; coordinates [(0.05,7.294876755408422)(0.1,7.143459007015522)(0.15,7.10697046820233)(0.2,7.039394155983152)(0.25,7.026308991460278)(0.3,6.510975320008498)(0.35,6.243115439072131)(0.4,6.105289175077003)(0.45,6.055312733639453)(0.5,5.935039702058106)(0.55,5.825901271325296)(0.6,5.6880 93231689677)(0.65,5.5867423886127545)(0.7,5.498111540657016)(0.75,5.44245493175869)(0.8,5.498157172914707)(0.85,5.417898859391104)(0.9,5.367460395266636)(0.95,5.326698343848651)(1.0,5.296824889730416)(1.05,6.0203644320612515)(1.1,5.951492591067398)(1.15,5.629974848372628)(1.2,5.291276689204172)(1.25,5.099740794831965)(1.3,5.146620468685676)(1.35,5.052022042315587)(1.4,4.964757934781348)(1.45,4.936920273710175)(1.5,4.877923799657347)(1.55,5.020638537303579)(1.6,5.070764465049533)(1.65,4.994068999133266)(1.7,4.957135528347418)(1.75,4.898700914689063)(1.8,4.241870363218415)(1.85,4.1448981785958745)(1.9,4.049506468529267)(1.95,3.891280157371325)(2.0,3.7353740890719047)(2.05,3.8909938631089274)(2.1,3.682976484285241)(2.15,3.5620111365644878)(2.2,3.4297623236734034)(2.25,2.9432068862274403)(2.3,2.1664755343266613)(2.35,1.8909225392502775)(2.4,1.8076962847788174)(2.45,1.4721344026921206)(2.5,1.4526095476885508)(2.55,1.4920326058543918)(2.6,1.1550142560841294)(2.65,1.1983065943279652)(2.7,1.062547475503181)(2.75,1.6694662455739397)(2.8,1.5796570336101317)(2.85,1.1158821138726907)(2.9,0.9473122962736089)(2.95,1.1670672326885256)(3.0,1.038697792354673)(3.05,0.9174663813843345)(3.1,0.9018216230285925)(3.15,2.477144567206786)(3.2,0.9206047669070757)(3.25,0.8815551069397457)(3.3,0.8126225182495597)(3.35,0.8611761393156186)(3.4,0.7926411851091579)(3.45,0.7371737478839958)(3.5,0.7802471912505505)(3.55,0.7463891014387589)(3.6,0.7377828179605033)(3.65,0.8078595782227387)(3.7,0.7645049814092778)(3.75,0.7060433718113253)(3.8,0.7086022748291205)(3.85,0.843256183545804)(3.9,0.7248512180854119)(3.95,0.6867248835765802)(4.0,0.6782928619958325)(4.05,0.6665662959556152)(4.1,0.9758636795784974)(4.15,0.7012710556147046)(4.2,0.6665186122871418)(4.25,0.6381702640233708)(4.3,0.6763554985828949)(4.35,0.6504905407969209)(4.4,0.6544181529326483)(4.45,0.6263901547541854)(4.5,0.6316339586266058)(4.55,0.637564212308961)(4.6,0.6344352711518662)(4.65,0.6679271155479036)(4.7,0.7276229010018751)(4.75,0.6137587334173562)(4.8,0.5884813674955226)(4.85,0.6184289425915641)(4.9,0.585997433317164)(4.95,0.5927840577692879)(5.0,0.5749813881331797)(5.05,0.5626287142203976)(5.1,0.5960067944376266)(5.15,0.6280227962545357)(5.2,0.5523787261104437)(5.25,0.4994123397467301)(5.3,0.4824490595768111)(5.35,0.4759646390416428)(5.4,0.4613345432206699)(5.45,0.45873671324804877)(5.5,0.4518391826114293)(5.55,0.4471541782596887)(5.6,0.44357929022644005)(5.65,0.441327487815143)(5.7,0.44071905383255)(5.75,0.4336441071456153)(5.8,0.4313850681301356)(5.85,0.4282499398131093)(5.9,0.42528441352329927)(5.95,0.4235418059036517)(6.0,0.4218797622510601)(6.05,0.42144524390440946)(6.1,0.4182964219182779)(6.15,0.4177564270227889)(6.2,0.4168608989632423)(6.25,0.4181030488622067)(6.3,0.41517383575375055)(6.35,0.41430180404182804)(6.4,0.41287789588964524)(6.45,0.4122367498839358)(6.5,0.4098295317314658)(6.55,0.4082834133005633)(6.6,0.40617848696683845)(6.65,0.40645270708496983)(6.7,0.40678591920431617)(6.75,0.4050203162677558)(6.8,0.40411272804472415)(6.85,0.4017664454118419)(6.9,0.3970757294997769)(6.95,0.40209725890991943)(7.0,0.3990218408330199)(7.05,0.4004870727267153)(7.1,0.3983834648565725)(7.15,0.39795929488173815)(7.2,0.3996291560368087)(7.25,0.397364789458027)(7.3,0.3983999388032953)(7.35,0.3964760807617469)(7.4,0.3951905317976588)(7.45,0.39385478594570544)(7.5,0.39630353650759226)(7.55,0.39404896760967456)(7.6,0.3936804652466766)(7.65,0.39301053505823386)(7.7,0.39032294505411497)(7.75,0.3897165330024663)(7.8,0.3879961563748509)(7.85,0.38700847288887186)(7.9,0.3854015569432422)(7.95,0.3893930856363974)(8.0,0.38726043797561865)(8.05,0.3876280491112605)(8.1,0.3884701823848394)(8.15,0.3889967017999761)(8.2,0.38794473816136543)(8.25,0.38644251454396655)(8.3,0.3853768613482451)(8.35,0.38589857067876904)(8.4,0.38444671785497087)(8.45,0.38424149542257746)(8.5,0.38636569394689535)(8.55,0.3825899865416747)(8.6,0.38301469158188234)(8.65,0.38363833953282983)(8.7,0.3813490537930746)(8.75,0.38275958097751656)(8.8,0.382397686399854)(8.85,0.38289347855671474)(8.9,0.379941885064304)(8.95,0.37839976225843186)(9.0,0.3763573228365485)(9.05,0.379682000181308)(9.1,0.37891822570704087)(9.15,0.3781726477256124)(9.2,0.37813665518493533)(9.25,0.3784049189209174)(9.3,0.378101687606327)(9.35,0.3759501809787691)(9.4,0.3759211645948709)(9.45,0.37714870422953106)(9.5,0.3751871466318765)(9.55,0.37387027916796856)(9.6,0.37412258479111926)(9.65,0.37423367557253595)(9.7,0.37377493356133096)(9.75,0.37237511334625417)(9.8,0.3735909059299931)(9.85,0.3732560063746375)(9.9,0.37048419557514617)(9.95,0.37135751696668934)(10.0,0.37228209478055085)(10.05,0.37254551484008835)(10.1,0.3733663894185437)(10.15,0.37288839520736494)(10.2,0.37237518815292286)(10.25,0.37266347062791016)(10.3,0.37262654564504)(10.35,0.3703907382280017)(10.4,0.37098151125759027)(10.45,0.3702662710652996)(10.5,0.37076660843106757)(10.55,0.3707777744255389)(10.6,0.3693601341232714)(10.65,0.37002725730613256)(10.7,0.3679216446172686)(10.75,0.3685420172064639)(10.8,0.36929761257578425)(10.85,0.36834781514067605)(10.9,0.36812417907050504)(10.95,0.36798616351862545)(11.0,0.36709971430004734)(11.05,0.36892332469982275)(11.1,0.36781952171623106)(11.15,0.3676762713423402)(11.2,0.36846260753951826)(11.25,0.3681166463394588)(11.3,0.36833902701771687)(11.35,0.3681854250140449)(11.4,0.36790469475463156)(11.45,0.3675104545958217)(11.5,0.36777130269770786)(11.55,0.36702723593751)(11.6,0.36737314311228825)(11.65,0.3670815004419397)(11.7,0.3671247603318658)(11.75,0.3668521483831425)(11.8,0.3663714211506519)(11.85,0.3667341726995936)(11.9,0.366396322535157)(11.95,0.36603379882234)(12.0,0.3657413971727037)(12.05,0.36562585265844144)(12.1,0.36588143709109355)(12.15,0.366762196416395)(12.2,0.3662263311969563)(12.25,0.3657294654299691)(12.3,0.3653326168116493)(12.35,0.36602320632457486)(12.4,0.3655373884429254)(12.45,0.36504828947436224)(12.5,0.36515653814009064)(12.55,0.36570600732389363)(12.6,0.36578777559923886)(12.65,0.36539111493069726)(12.7,0.3653343815847854)(12.75,0.3657834542409847)(12.8,0.3654466735746801)(12.85,0.36451183905484835)(12.9,0.3647384848835593)(12.95,0.36474604076829753)(13.0,0.3648705219117809)(13.05,0.3652829923134148)(13.1,0.3651208941700221)(13.15,0.36526465291295285)(13.2,0.3645191811791354)(13.25,0.3655175750822288)(13.3,0.365327268213914)(13.35,0.3650297481749715)(13.4,0.36500196738215135)(13.45,0.3653808434215469)(13.5,0.3647300113281774)(13.55,0.36522078479578257)(13.6,0.36516078930976226)(13.65,0.36494702547407815)(13.7,0.3648527068538758)(13.75,0.3650233707878496)(13.8,0.365511315045103)]{}; The cross-entropy loss curves on training and dev set in E1, E2, E6, and E7 are depicted in Fig. \[train\_dev\_loss\]. The model based on each attention method was trained from scratch until convergence, with a few spikes in its training loss curve. A spike in the loss curve occurs when the layer-wise pre-training algorithm inserts a new layer and units into the encoder, as the untrained new components make the performance worse right after the insertion. GRC converged later than GSA, and DecGRC converged slightly later than GRC. MoChA showed the slowest optimization speed, which was partly due to the 5 times longer layer-wise pre-training scheduling than the others. Such long pre-training was employed to stabilize the training of MoChA, whereas the both GRC and DecGRC successfully converged with the standard pre-training. Note that the longer pre-training of MoCha was adopted because it had failed to converge with a short pre-training in our initial experiments. The relatively stable learning of the proposed methods over MoChA can be explained in relation to sMoChA, as described in Sec. \[relation\_sMoChA\]; the sMoChA stabilized the training of MoChA by utilizing a modified selection probability formula, which is actually almost similar to the attention weight $\boldsymbol{\bar{\alpha}}({\mathbf{z}_u})$ of GRC in Eq. . Attention analysis {#sec:attention_analysis} ------------------ ![An input spectrogram, attention plots with the output BPE sequence of GSA (E1), GRC (E2), and DecGRC (E8), and the update gates of the DecGRC, from top to bottom. All results were obtained with BiLSTM encoder on an utterance 8254-84205-0009 in dev-other set. The update gates were obtained with teacher forcing, and the attention plots were results of the beam search w/ beam size 12. “\_\_" was inserted after a BPE unit end if it was not a word-end. []{data-label="attention_plot"}](att7.pdf){width="9.25cm"} GRC and DecGRC accurately learned alignments between encoded representations and output text units, as illustrated in Fig. \[attention\_plot\]. An interesting characteristic of GRC was observed that it tended to put much weight on the latter time indices of attention weights, compared to GSA. This can be regarded as an innate behavior of GRC, as the attention weight $\boldsymbol{\bar{\alpha}}({\mathbf{z}_u})$ in Thm. \[thm1\] is designed to weigh the latter indices when the update gates $z_{u,t}$ have similar value over several consecutive time-indices. The latter-time-weighing attribute could be especially effective for a long text unit (e.g., a BPE unit “swinging” in Fig. \[attention\_plot\]), as a long BPE unit often ends with a suffix that might be crucial to distinguish words (e.g., “-ing", “-n’t", or “-est" in English). A piece of statistical evidence is presented in Fig. \[median\_length\_of\_BPE\_units\_to\_WER\]; GRC outperformed GSA when the median length of BPE was larger than or equal to 6, while it showed similar performance for shorter median lengths. Attention weights of DecGRC tended to be much smoother (i.e., focused on longer time) than GRC and GSA. Such smoothness was hypothesized to be caused by the decreasing update gates, which made the model trained to be cautious for a sharp descent of update gate values, as it is irreversible in DecGRC. In addition, DecGRC did not attend on the first time index, unlike GSA and GRC. It is an intrinsic property of DecGRC, as the earliest update gates have values close to 1 and therefore difficult to carry information to later time. As the initial frames of an utterance usually contain helpful information such as background noise, this might cause DecGRC to be degraded compared to the global attentions. The last two plots in Fig. \[attention\_plot\] show that the update gate values of DecGRC mostly changed near the attention region. As the update gates rapidly decreased after the attention region, tight attention endpoints could easily be found by setting the threshold value approximately in a range of \[0.001, 0.2\]. For instance, with an inference threshold $\nu=0.01$ in Fig. \[attention\_plot\], the total number of steps in the for loop in Alg. \[alg:example\] was 459, which was approximately 54% of $TU=13\times65=845$. It implies that insignificant time indices were properly ignored during the inference. coordinates [(2,0.21849929211892402)(3,0.10279748360326596)(4,0.08669469039740035) (5,0.10634648370497427)(6,0.1729957805907173) (7,0.28125) (8,0)]{}; coordinates [(2,0.21755545068428503)(3,0.10001635955323547)(4,0.08326625525444951) (5,0.11492281303602059)(6,0.12658227848101267) (7,0.25) (8,0)]{}; Ablation study on DecGRC inference threshold {#sec:ablation_study} -------------------------------------------- We evaluated WERs of DecGRC models for different threshold values, and the results are summarized in Tbl. \[ablation\]. Setting the threshold to a value larger than 0.2 was found to be detrimental to the performance on the BiLSTM encoder (E8), with larger thresholds giving higher WERs. It means that some encoded vectors in the correct attention region were ignored due to the high threshold, as shown in the last two plots of Fig. \[attention\_plot\]. Impressively, the best performance was obtained with $\nu$ between 0.001 and 0.1, not $\nu=0$. This may be attributed to the fact that the thresholding not only reduced the latency, but also eliminated undesirable updates after the correct attention region. With thresholds higher than the best-performing threshold, the latency could be further reduced by taking the performance penalty, and vice versa. The best-performing threshold on the LC-BiLSTM encoder (E18) was $\nu=0.08$, which means that the optimal threshold for performance depends on the model architecture, thus threshold searching is required before test phase. Nevertheless, as the beam search inference on the dev set took less than 15 minutes using a single GPU, the time spent for the tuning process of the threshold was no more than 2.5 hours, which is much shorter than the model training time; a single epoch of training took 9 hours on average, and the total time for training a model from scratch was more than 5 days. ------- -- ---------- ----------- ---------- ----------- 0 5.03 14.08 4.89 14.83 0.001 4.96 **14.01** 4.86 **14.76** 0.01 4.97 14.02 **4.83** 14.90 0.05 **4.83** 14.02 4.84 14.93 0.2 5.09 14.34 5.10 15.10 0.25 5.85 14.79 5.98 15.64 0.4 10.59 18.35 11.58 19.24 0.6 39.67 39.08 40.42 40.70 0 6.07 15.80 6.30 16.51 0.001 6.08 15.77 6.32 16.50 0.01 6.01 15.75 6.22 16.45 0.05 5.83 **15.63** 6.06 16.35 0.08 5.79 15.67 6.04 **16.34** 0.1 **5.78** 15.66 **6.03** 16.36 0.2 6.41 16.20 6.77 16.93 0.25 7.27 16.78 7.84 17.77 0.4 15.53 21.93 16.76 23.94 0.6 70.54 66.38 71.82 70.20 ------- -- ---------- ----------- ---------- ----------- : Ablation study about the inference threshold of DecGRC on LibriSpeech dataset. []{data-label="ablation"} Conclusion {#sec:conclusion} ========== We proposed a novel softmax-free global attention method called GRC, and its variant for online attention, namely DecGRC. Unlike the conventional online attentions, DecGRC introduces no additional hyperparameter to be tuned at the training phase. Thus DecGRC does not require multiple trials of training, saving time for model preparation. Moreover at the inference of DecGRC, the tradeoff between ASR latency and performance can be controlled by adapting the scalar threshold which is related to the attention endpoint decision, whereas the conventional online attentions are not capable of changing the endpoint decision rule at test phase. Both GRC and DecGRC showed comparable ASR performance to the conventional global attentions. For further research, the proposed attention methods will be investigated in various applications which leverage AED models. We are particularly interested in applying DecGRC to simultaneous machine translation [@arivazhagan2019monotonic] and real-time scene text recognition [@liu2018squeezedtext], where the latency can be reduced by exploiting an online attention method. [^1]: The authors are with the Institute of New Media and Communications, Department of Electrical and Computer Engineering, Seoul National University, Seoul, Republic of Korea (e-mail: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]). This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible. [^2]: The LibriSpeech dataset can be downloaded from <http://www.openslr.org/12>. [^3]: The scripts for all experiments are available at <https://github.com/GRC-anonymous/GRC>.

Q: sql replication using remote server PI'm trying to do replicate two data bases. DB in LAN network (Publisher) DB in virtual dedicated network (Subscriber) According to my situation, replicate publisher implemented in my server in LAN network.But subscriber is implementing on a virtual dedicated server. i configured router port to my server machine in LAN network.using sql management studio on virtual server, i connected to the db in LAN network.But when i try to create subscriber using virtual server db i can't access to the publisher.(IN LAN network.) it gives errors as below. "SQL Server replication requires the actual server name to make a connection to the server. Connections through a server alias, IP address, or any other alternate name are not supported. Specify the actual server name, 'BESTLIFE\BESTLIFECROWN'. (Replication.Utilities)" Please help me to solve this . A: You didn't say whether you used the GUI or scripts, but somewhere, a call got made to one of the stored procedures (likely sp_addsubscription) with a server name that doesn't match the actual server name. For instance, the error message above says that the server is called BESTLIFE\BESTLIFECROWN. If that's not the name of the server that you put in as hosting the subscriber, it's not going to work. Whether you need to add a DNS alias or whatever, that's the only value that will work for this setup.

Who's Online Guest Users: 15 Saudi Cracks Down on Ethiopian Migrants Friday, November 15 2013 @ 08:30 AM MSK As garbage piled in the Saudi streets, 23,000 Ethiopians have surrendered to Saudi authorities after a clampdown on illegal migrant workers began in the oil-rich kingdom last week, stirring clashes after frustrated workers took to the streets. The clashes “occurred because the illegal workers were frustrated they did not have a way to surrender to the police,” Ambassador Mohammad Hassan Kabiera told the English-language Arab News daily, AFP reported on Thursday. The workers took to the streets to voice their concerns, prompting “clashes with some youths in the neighborhood”, he said. “We have been informed that so far about 23,000 Ethiopians have handed themselves in,” Kabiera said. The Ethiopian workers’ surrender followed clashes earlier this week between police and Ethiopian migrants on Saturday that led to the deaths of three people in the poor Manfuhah neighborhood of Riyadh. On Tuesday, Ethiopia said three of its citizens had died during clashes in the Gulf kingdom, without elaborating. "The act of killing innocent civilians is uncalled for, we condemn that," Foreign Ministry spokesman Dina Mufti told reporters, saying he had been informed of the death of three Ethiopian citizens. The troubles followed the kingdom’s move to round up thousands of illegal workers starting from November 4, following the expiry of a final amnesty for them to formalize their status. Among them were foreigners who overstayed their visas, pilgrims who had sought jobs, and migrants working under one sponsor trying to get jobs elsewhere. Having an official sponsor is a legal requirement in Saudi Arabia and most other Gulf states. Nearly a million migrants — Bangladeshis, Filipinos, Indians, Nepalis, Pakistanis and Yemenis among them — took advantage of the amnesty to leave. Another roughly four million were able to find employers to sponsor them. Expatriates account for a full nine million of the oil-rich kingdom’s population of 27 million. Defended Despite feeling the loss of the everyday work the foreign laborers provided, Saudis largely have cheered the campaign. Al Riyadh daily quoted Prince Khalid as defending the campaign, saying it “does not target a specific group but all illegal” workers and residents. “We will continue these campaigns until we ensure all residents in our country are staying legally,” he said. Residents have taken matters into their own hands on several occasions, despite police calling on the public not to make citizen arrests. Workers from neighboring Yemen also face harassment. Yemeni Nobel Peace Prize Laureate Tawakkol Karman posted a picture last week on her Facebook page of what appeared to be a Saudi man in his car grabbing hold of a Yemeni man for a police officer. Fresh violence erupted in Riyadh’s Manfouha district later on Wednesday, resulting in the death of a Sudanese resident. At least 17 expatriate workers were also reported injured as they clashed with local residents, police said. On Wednesday, Saudi media quoted Riyadh governor Prince Khalid Bin Bandar Bin Abdul Aziz as saying that “casualties and deaths have not surpassed three Saudis and two foreigners”, again without giving details. Saudi columnist Abdul-Rahman al-Rashed cautioned Saudis to remember that without “a strong state and oil revenues” they too may have emigrated in search of work. “Those deprived of the chance of a proper life can understand the feeling of those wanting to seek a better life,” he wrote in the Asharq al-Awsat newspaper. Don't take easy at any stage of your life during that process and work really hard to bring some positive results. That time i am working for one of the best educational network, where students can pay for college essays online and get their products online. Our network also provides money back guarantee on ever service which is really great for all students. A genuine write-up could possibly be rewarding. I really like this original write-up as well as whichever other people make a decision kid ought to be might permit you may be maintaining most current combined with the drastically beneficial, My family and i expected get pleasure from so as to thrust grows such a web sites we've got genuinely gotten to preserve i would authenticate speedily beneath of which likewise has a brand-new established that you have get to completing a unique complete. quick weight loss My wife or husband and we seriously appreciated examining which usually, you could possibly become a outstanding article article author. Prevent meaning that many of us you need to please take a notice with the web page all of which will return sometime rapidly. Access Control System

Q: Log4j and Jboss. How do I get my log.Debug("Debug") messages to display? I'm trying to start my jBoss server (v4.2) to run and display all of my log.debug messages. I need to achieve this by way of editing jboss configurations. Not programmatically through the application. my jboss-log4.xml <?xml version="1.0" encoding="UTF-8"?> <!DOCTYPE log4j:configuration SYSTEM "log4j.dtd"> <!-- ===================================================================== --> <!-- --> <!-- Log4j Configuration --> <!-- --> <!-- ===================================================================== --> <!-- $Id: jboss-log4j.xml 62403 2007-04-18 15:26:43Z [email protected] $ --> <!-- | For more configuration infromation and examples see the Jakarta Log4j | owebsite: http://jakarta.apache.org/log4j --> <log4j:configuration xmlns:log4j="http://jakarta.apache.org/log4j/" debug="false"> <!-- ================================= --> <!-- Preserve messages in a local file --> <!-- ================================= --> <!-- A time/date based rolling appender --> <appender name="FILE" class="org.jboss.logging.appender.DailyRollingFileAppender"> <errorHandler class="org.jboss.logging.util.OnlyOnceErrorHandler"/> <param name="File" value="${jboss.server.log.dir}/server.log"/> <param name="Append" value="false"/> <!-- Rollover at midnight each day --> <param name="DatePattern" value="'.'yyyy-MM-dd"/> <!-- Rollover at the top of each hour <param name="DatePattern" value="'.'yyyy-MM-dd-HH"/> --> <layout class="org.apache.log4j.PatternLayout"> <!-- The default pattern: Date Priority [Category] Message\n --> <param name="ConversionPattern" value="%d %-5p [%c] %m%n"/> <!-- The full pattern: Date MS Priority [Category] (Thread:NDC) Message\n <param name="ConversionPattern" value="%d %-5r %-5p [%c] (%t:%x) %m%n"/> --> </layout> </appender> <!-- A size based file rolling appender <appender name="FILE" class="org.jboss.logging.appender.RollingFileAppender"> <errorHandler class="org.jboss.logging.util.OnlyOnceErrorHandler"/> <param name="File" value="${jboss.server.log.dir}/server.log"/> <param name="Append" value="false"/> <param name="MaxFileSize" value="500KB"/> <param name="MaxBackupIndex" value="1"/> <layout class="org.apache.log4j.PatternLayout"> <param name="ConversionPattern" value="%d %-5p [%c] %m%n"/> </layout> </appender> --> <!-- ============================== --> <!-- Append messages to the console --> <!-- ============================== --> <appender name="CONSOLE" class="org.apache.log4j.ConsoleAppender"> <errorHandler class="org.jboss.logging.util.OnlyOnceErrorHandler"/> <param name="Target" value="System.out"/> <param name="Threshold" value="DEBUG"/> <layout class="org.apache.log4j.PatternLayout"> <!-- The default pattern: Date Priority [Category] Message\n --> <param name="ConversionPattern" value="%d{ABSOLUTE} %-5p [%c{1}] %m%n"/> </layout> </appender> <!-- ====================== --> <!-- More Appender examples --> <!-- ====================== --> <!-- Buffer events and log them asynchronously <appender name="ASYNC" class="org.apache.log4j.AsyncAppender"> <errorHandler class="org.jboss.logging.util.OnlyOnceErrorHandler"/> <appender-ref ref="FILE"/> <appender-ref ref="CONSOLE"/> <appender-ref ref="SMTP"/> </appender> --> <!-- EMail events to an administrator <appender name="SMTP" class="org.apache.log4j.net.SMTPAppender"> <errorHandler class="org.jboss.logging.util.OnlyOnceErrorHandler"/> <param name="Threshold" value="ERROR"/> <param name="To" value="[email protected]"/> <param name="From" value="[email protected]"/> <param name="Subject" value="JBoss Sever Errors"/> <param name="SMTPHost" value="localhost"/> <param name="BufferSize" value="10"/> <layout class="org.apache.log4j.PatternLayout"> <param name="ConversionPattern" value="[%d{ABSOLUTE},%c{1}] %m%n"/> </layout> </appender> --> <!-- Syslog events <appender name="SYSLOG" class="org.apache.log4j.net.SyslogAppender"> <errorHandler class="org.jboss.logging.util.OnlyOnceErrorHandler"/> <param name="Facility" value="LOCAL7"/> <param name="FacilityPrinting" value="true"/> <param name="SyslogHost" value="localhost"/> <layout class="org.apache.log4j.PatternLayout"> <param name="ConversionPattern" value="[%d{ABSOLUTE},%c{1}] %m%n"/> </layout> </appender> --> <!-- Log events to JMS (requires a topic to be created) <appender name="JMS" class="org.apache.log4j.net.JMSAppender"> <errorHandler class="org.jboss.logging.util.OnlyOnceErrorHandler"/> <param name="Threshold" value="ERROR"/> <param name="TopicConnectionFactoryBindingName" value="java:/ConnectionFactory"/> <param name="TopicBindingName" value="topic/MyErrorsTopic"/> </appender> --> <!-- Log events through SNMP <appender name="TRAP_LOG" class="org.apache.log4j.ext.SNMPTrapAppender"> <errorHandler class="org.jboss.logging.util.OnlyOnceErrorHandler"/> <param name="ImplementationClassName" value="org.apache.log4j.ext.JoeSNMPTrapSender"/> <param name="ManagementHost" value="127.0.0.1"/> <param name="ManagementHostTrapListenPort" value="162"/> <param name="EnterpriseOID" value="1.3.6.1.4.1.24.0"/> <param name="LocalIPAddress" value="127.0.0.1"/> <param name="LocalTrapSendPort" value="161"/> <param name="GenericTrapType" value="6"/> <param name="SpecificTrapType" value="12345678"/> <param name="CommunityString" value="public"/> <param name="ForwardStackTraceWithTrap" value="true"/> <param name="Threshold" value="DEBUG"/> <param name="ApplicationTrapOID" value="1.3.6.1.4.1.24.12.10.22.64"/> <layout class="org.apache.log4j.PatternLayout"> <param name="ConversionPattern" value="%d,%p,[%t],[%c],%m%n"/> </layout> </appender> --> <!-- Emit events as JMX notifications <appender name="JMX" class="org.jboss.monitor.services.JMXNotificationAppender"> <errorHandler class="org.jboss.logging.util.OnlyOnceErrorHandler"/> <param name="Threshold" value="WARN"/> <param name="ObjectName" value="jboss.system:service=Logging,type=JMXNotificationAppender"/> <layout class="org.apache.log4j.PatternLayout"> <param name="ConversionPattern" value="%d %-5p [%c] %m"/> </layout> </appender> --> <!-- ================ --> <!-- Limit categories --> <!-- ================ --> <!-- Limit the org.apache category to INFO as its DEBUG is verbose --> <category name="org.apache"> <priority value="DEBUG"/> </category> <!-- Limit the org.jboss.serial (jboss-serialization) to INFO as its DEBUG is verbose --> <category name="org.jboss.serial"> <priority value="DEBUG"/> </category> <!-- Limit the org.jgroups category to WARN as its INFO is verbose --> <category name="org.jgroups"> <priority value="DEBUG"/> </category> <!-- Limit the jacorb category to WARN as its INFO is verbose --> <category name="jacorb"> <priority value="DEBUG"/> </category> <!-- Limit JBoss categories --> <category name="org.jboss"> <priority value="DEBUG"/> </category> <!-- Limit the JSR77 categories --> <category name="org.jboss.management"> <priority value="DEBUG"/> </category> <!-- Enable JBossWS message tracing --> <category name="jbossws.SOAPMessage"> <priority value="TRACE"/> </category> <!-- Decrease the priority threshold for the org.jboss.varia category <category name="org.jboss.varia"> <priority value="DEBUG"/> </category> --> <!-- Show the evolution of the DataSource pool in the logs [inUse/Available/Max] <category name="org.jboss.resource.connectionmanager.JBossManagedConnectionPool"> <priority value="TRACE"/> </category> --> <!-- | An example of enabling the custom TRACE level priority that is used | by the JBoss internals to diagnose low level details. This example | turns on TRACE level msgs for the org.jboss.ejb.plugins package and its | subpackages. This will produce A LOT of logging output. | | Note: since jboss AS 4.2.x, the trace level is supported natively by | log4j, so although the custom org.jboss.logging.XLevel priority will | still work, there is no need to use it. The two examples that follow | will both enable trace logging. <category name="org.jboss.system"> <priority value="TRACE" class="org.jboss.logging.XLevel"/> </category> <category name="org.jboss.ejb.plugins"> <priority value="TRACE"/> </category> --> <!-- | Logs these events to SNMP: - server starts/stops - cluster evolution (node death/startup) - When an EJB archive is deployed (and associated verified messages) - When an EAR archive is deployed <category name="org.jboss.system.server.Server"> <priority value="INFO" /> <appender-ref ref="TRAP_LOG"/> </category> <category name="org.jboss.ha.framework.interfaces.HAPartition.lifecycle"> <priority value="INFO" /> <appender-ref ref="TRAP_LOG"/> </category> <category name="org.jboss.deployment.MainDeployer"> <priority value="ERROR" /> <appender-ref ref="TRAP_LOG"/> </category> <category name="org.jboss.ejb.EJBDeployer"> <priority value="INFO" /> <appender-ref ref="TRAP_LOG"/> </category> <category name="org.jboss.deployment.EARDeployer"> <priority value="INFO" /> <appender-ref ref="TRAP_LOG"/> </category> --> <!-- ======================= --> <!-- Setup the Root category --> <!-- ======================= --> <root> <appender-ref ref="CONSOLE"/> <appender-ref ref="FILE"/> </root> <!-- Clustering logging --> <!-- Uncomment the following to redirect the org.jgroups and org.jboss.ha categories to a cluster.log file. <appender name="CLUSTER" class="org.jboss.logging.appender.RollingFileAppender"> <errorHandler class="org.jboss.logging.util.OnlyOnceErrorHandler"/> <param name="File" value="${jboss.server.log.dir}/cluster.log"/> <param name="Append" value="false"/> <param name="MaxFileSize" value="500KB"/> <param name="MaxBackupIndex" value="1"/> <layout class="org.apache.log4j.PatternLayout"> <param name="ConversionPattern" value="%d %-5p [%c] %m%n"/> </layout> </appender> <category name="org.jgroups"> <priority value="DEBUG" /> <appender-ref ref="CLUSTER"/> </category> <category name="org.jboss.ha"> <priority value="DEBUG" /> <appender-ref ref="CLUSTER"/> </category> --> </log4j:configuration> Page I'm testing private static final Logger log = Logger.getLogger(ServerLogsController.class .getName()); @GET @Path("/testDebug") public String testDebug(@Context final ServletContext context) { log.error("This is an error which always shows!"); log.debug("This is a debug message"); return "Test Debug Page"; } A: There's two ways I can think of to get the debug messages to show here: 1) You can add a new entry in your xml file for ServerLogsController to display DEBUG messages. <category name="package.for.ServerLogsController"> <priority value="DEBUG" /> <appender-ref ref="CONSOLE"/> <appender-ref ref="FILE"/> </category> where you replace package.for.ServerLogsController with the correct package name. Note that you can also set that to just package.for and all files under that directory will log @ DEBUG level. This setting will override the root behavior mentioned in the second bullet (i.e. if root is set to DEBUG but package.for.ServerLogsController is set to ERROR, you will not see DEBUG messages from package.for.ServerLogsController. 2) Set debugging on the root package <root> <priority value="DEBUG"/> <appender-ref ref="CONSOLE"/> <appender-ref ref="FILE"/> </root> This will make all packages log at DEBUG level unless specifically overriden by a package.

djangoappengine - Django backends (DB, email, etc.) for App Engine ================================================================== Djangoappengine contains App Engine backends for Django-nonrel, e.g. the database and email backends. In addition we provide a testapp_ which contains minimal settings for running Django-nonrel on App Engine. Use it as a starting point if you want to use App Engine as your database for Django-nonrel. Take a look at the documentation below and subscribe to our `discussion group`_ for the latest updates. Contents -------- .. toctree:: :maxdepth: 2 :titlesonly: installation db services management admin mapreduce Contribute ------------------------------------------------------ If you want to help with implementing a missing feature or improving something please fork the source_ and send a pull request via Github or a patch to the `discussion group`_. .. _discussion group: http://groups.google.com/group/django-non-relational .. _testapp: https://github.com/django-nonrel/django-testapp .. _source: https://github.com/django-nonrel/djangoappengine

The Carolina Hurricanes will put their four-game winning streak on the line when they host the Philadelphia Flyers in a Metropolitan Division match-up on Thursday. The Canes will ice the same lineup seen in Chicago on Tuesday - and why mess with what's working? - including Petr Mrazek starting in net. GAME PREVIEW Here is the projected lineup for the Hurricanes.

IN THE COURT OF CRIMINAL APPEALS OF TEXAS PD-1374-14 WESLEY ALLEN DOTSON, Appellant v. THE STATE OF TEXAS ON APPELLANT’S PETITION FOR DISCRETIONARY REVIEW FROM THE FOURTH COURT OF APPEALS KARNES COUNTY Per curiam. Richardson, J., not participating. OPINION A jury convicted appellant of aggravated assault of a public servant, and sentenced him to fifty years’ confinement and assessed a $10,000 fine. On appeal, appellant claimed that the trial court erred by allowing the State to impeach a defense witness with prior felony convictions that were more than ten years old. The court of appeals upheld the trial court’s actions by applying the common law tacking doctrine to the remote convictions, and assessing their admissibility under Rule of Evidence 609(a)’s “outweigh” standard rather than Rule of Evidence 609(b)’s  WESLEY ALLEN DOTSON – 2 “substantially outweigh” standard. Dotson v. State, No. 04-13-00858-CR, slip op. at 3-4 (Tex. App.–San Antonio Sept. 10, 2015)(not designated for publication). Applying Rule 609(a), the court of appeals upheld the admission of the prior remote convictions. Id. at 6. Appellant has filed a petition for discretionary review of this decision. We recently addressed this issue in Meadows v. State, PD-0175-14, slip op. at 4 (Tex. Crim. App. Feb. 25, 2015), in which we held that the unambiguous plain language of Rule of Evidence 609 supplants the common-law tacking doctrine. Under Rule 609(b), evidence of a prior conviction is inadmissible to impeach a witness “if more than ten years has elapsed since the later of the date of conviction or release of the witness from confinement imposed for that conviction ‘unless the court determines, in the interests of justice, that the probative value of the conviction supported by specific facts and circumstances substantially outweighs its prejudicial effect.’” Id. at 6. The Court of Appeals in the instant case did not have the benefit of our opinion in Meadows. Accordingly, we grant ground (1) of appellant’s petition for discretionary review, vacate the judgment of the Court of Appeals, and remand this case to the Court of Appeals in light of our opinion in Meadows.1 DELIVERED April 15, 2015 DO NOT PUBLISH 1 Ground (2) of appellant’s petition for discretionary review is refused without prejudice. 

Q: py2neo graph.merge() behaves differently from Cypher MERGE? So, for an empty database MERGE (N1:A {name:"A"})-[:r]->(N2:B {name:"B"}) will create two nodes N1 and N2 with an edge r between them. The following python code however does not do that... but why? Should it not? from py2neo import Graph, authenticate, rel, Node graph = Graph() # set up authentication parameters authenticate("localhost:7474", <user>, <password>) # clear the data base graph.delete_all() graph.merge(rel(Node("A" , name="A"), "r", Node("B" , name="B"))) Running that script results in a still empty database. Why is that and how can I get the Cypher merge behaviour from py2neo without using graph.cypher.execute("MERGE ...")? A: In Py2neo graph.merge matches or creates a single node by label and (optionally) property, where you are wanting to MERGE on the entire pattern (node, relationship, other node). The pattern you are using for the Cypher MERGE statement does not appear to be supported in Py2neo outside of Cypher.

Q: How can I load a sound file into memory using NAudio and use it later? I'm playing the same sound files (randomly choosing between them) 5 times a second, and I'm always loading into memory, so the program uses a lot of memory. How could I load the sound file into memory, and start it from there? I'm using NAudio. Current code: var sound = "sounds/test.mp3"; using (var audioFile = new AudioFileReader(sound)) using (var outputDevice = new WaveOutEvent()) { outputDevice.Init(audioFile); outputDevice.Play(); while (outputDevice.PlaybackState == PlaybackState.Playing) { Thread.Sleep(1000); } threadStop(); } A: If you remove the using blocks then audioFile and outputDevice will not be disposed. You can then retain them in memory, and the same references will be used each time you play the audio. With using blocks, you are repeatedly instantiating NAudio objects whose memory may not be deallocated immediately. var sound = "sounds/test.mp3"; var audioFile = new AudioFileReader(sound); var outputDevice = new WaveOutEvent(); outputDevice.Init(audioFile); outputDevice.Play(); while (outputDevice.PlaybackState == PlaybackState.Playing) { Thread.Sleep(1000); } threadStop(); A: I fixed the whole issue by using the code found in this article. It uses the MixingSampleProvider. I load the sounds into a custom class called: CachedSound. And then I play them using another class called: AudioPlaybackEngine. Which handles the mixer, and I use the CachedSoundSampleProvider class to read the cached sound. The code looks like this: class AudioPlaybackEngine : IDisposable { private readonly IWavePlayer outputDevice; private readonly MixingSampleProvider mixer; public AudioPlaybackEngine(int sampleRate = 44100, int channelCount = 2) { outputDevice = new WaveOutEvent(); mixer = new MixingSampleProvider(WaveFormat.CreateIeeeFloatWaveFormat(sampleRate, channelCount)); mixer.ReadFully = true; outputDevice.Init(mixer); outputDevice.Play(); } public void PlaySound(string fileName) { var input = new AudioFileReader(fileName); AddMixerInput(new AutoDisposeFileReader(input)); } private ISampleProvider ConvertToRightChannelCount(ISampleProvider input) { if (input.WaveFormat.Channels == mixer.WaveFormat.Channels) { return input; } if (input.WaveFormat.Channels == 1 && mixer.WaveFormat.Channels == 2) { return new MonoToStereoSampleProvider(input); } throw new NotImplementedException("Not yet implemented this channel count conversion"); } public void PlaySound(CachedSound sound) { AddMixerInput(new CachedSoundSampleProvider(sound)); } private void AddMixerInput(ISampleProvider input) { mixer.AddMixerInput(ConvertToRightChannelCount(input)); } public void Dispose() { outputDevice.Dispose(); } public static readonly AudioPlaybackEngine Instance = new AudioPlaybackEngine(44100, 2); } class CachedSound { public float[] AudioData { get; private set; } public WaveFormat WaveFormat { get; private set; } public CachedSound(string audioFileName) { using (var audioFileReader = new AudioFileReader(audioFileName)) { // TODO: could add resampling in here if required WaveFormat = audioFileReader.WaveFormat; var wholeFile = new List<float>((int)(audioFileReader.Length / 4)); var readBuffer= new float[audioFileReader.WaveFormat.SampleRate * audioFileReader.WaveFormat.Channels]; int samplesRead; while((samplesRead = audioFileReader.Read(readBuffer,0,readBuffer.Length)) > 0) { wholeFile.AddRange(readBuffer.Take(samplesRead)); } AudioData = wholeFile.ToArray(); } } } class CachedSoundSampleProvider : ISampleProvider { private readonly CachedSound cachedSound; private long position; public CachedSoundSampleProvider(CachedSound cachedSound) { this.cachedSound = cachedSound; } public int Read(float[] buffer, int offset, int count) { var availableSamples = cachedSound.AudioData.Length - position; var samplesToCopy = Math.Min(availableSamples, count); Array.Copy(cachedSound.AudioData, position, buffer, offset, samplesToCopy); position += samplesToCopy; return (int)samplesToCopy; } public WaveFormat WaveFormat { get { return cachedSound.WaveFormat; } } } // This class automatically disposes the file reader that it contains. class AutoDisposeFileReader : ISampleProvider { private readonly AudioFileReader reader; private bool isDisposed; public AutoDisposeFileReader(AudioFileReader reader) { this.reader = reader; this.WaveFormat = reader.WaveFormat; } public int Read(float[] buffer, int offset, int count) { if (isDisposed) return 0; int read = reader.Read(buffer, offset, count); if (read == 0) { reader.Dispose(); isDisposed = true; } return read; } public WaveFormat WaveFormat { get; private set; } }

417 F.3d 715 UNITED STATES of America, Plaintiff-Appellee,v.John LASHAY, Defendant-Appellant. No. 04-3378. United States Court of Appeals, Seventh Circuit. Argued July 6, 2005. Decided August 3, 2005. Tracy M. Johnson (argued), Michelle L. Jacobs, Office of the United States Attorney, Milwaukee, WI, for Plaintiff-Appellee. Michael Holzman (argued), Rosen & Holzman, Waukesha, WI, for Defendant-Appellant. Before COFFEY, RIPPLE, and ROVNER, Circuit Judges. ILANA DIAMOND ROVNER, Circuit Judge. 1 A jury found John LaShay guilty of conspiracy to defraud the United States, 18 U.S.C. § 371, and tampering with a witness, id. § 1512(b)(1). At sentencing, the district court treated the sentencing guidelines as advisory and imposed concurrent terms of 24 months on the conspiracy count and 6 months on the witness-tampering count. LaShay now argues that there was insufficient evidence to support the witness-tampering conviction, and that the district court should have submitted sentencing issues to a jury.1 We affirm the convictions but vacate the sentences and remand for resentencing. 2 Beginning in June 2002, LaShay was involved in a scheme to help Pakistani nationals obtain permanent resident status in the United States by finding United States citizens for them to marry. LaShay's employer at a local gas station, Faryad Hussain, asked him if he knew anyone who would marry an acquaintance who needed a green card, and LaShay suggested that his daughter might do so. LaShay's daughter eventually married one of Hussain's acquaintances and sponsored his application for permanent residency. LaShay afterward approached several other women about marrying Pakistani nationals, including his daughter's mother and a former co-worker. According to Hussain, LaShay was promised $400 or $500 for his role in the conspiracy, but the payments were never made. 3 During the investigation of this scheme, a government agent discovered that LaShay had cashed a $2000 check written out to him by Hussain. This amount matched the sum LaShay's daughter had told authorities she was offered for her marriage. Because of this, and because $2000 was a much larger amount than Hussain typically gave employees for business purposes, the agent considered this check significant and followed up on it in a July 2003 interview with James Clark, LaShay's friend and fellow gas station employee. As Clark recounted that interview at trial, he told the agent that the station did keep cash on hand to cash paychecks for customers, but that $2000 was more than the customary amount. Hussain, however, testified at trial as a government witness that he indeed gave LaShay the check to use for petty cash at the gas station, and that it was not a payment for participation in marriage fraud. 4 According to Clark's trial testimony, LaShay mentioned the $2000 check to him several times in the three days immediately preceding his interview with the government agent. LaShay told Clark that Hussain had given him the check to provide funds for cashing checks for customers, but that he was "worried" about the check because immigration authorities had a copy of it. He asked if Clark remembered Hussain giving him the check, but, according to his testimony, Clark responded that there was no way he could have been present for the event because he and LaShay worked different shifts. Nonetheless, LaShay raised the subject daily for three days, stating that he wanted to make sure Clark remembered that the check had been for petty cash. Clark responded that he wasn't going to lie for anyone. On cross-examination, though, Clark conceded that he did not feel LaShay had been trying to threaten or intimidate him. When asked on redirect if he felt LaShay had been asking him to lie, he responded yes, although on re-cross he also contradicted that statement by agreeing that defense counsel was "correct" in saying that LaShay hadn't really been asking him to lie. 5 After the guilty verdicts, the district court proceeded to sentencing. It stated that, in light of Blakely v. Washington, 542 U.S. 296, 124 S.Ct. 2531, 159 L.Ed.2d 403 (2004), and United States v. Booker, 375 F.3d 508 (7th Cir.2004), it did not believe the sentencing guidelines bound its sentencing determination. Rather, the district court decided to proceed with "[t]he Guidelines used as just that, a guide." Although LaShay had filed objections to the presentence report based on upward adjustments made to his offense level for committing the offense while on pretrial release and having a leadership role in the conspiracy, the district court did not resolve them. Instead, taking into account matters of punishment, deterrence, retribution, rehabilitation, and the safety of the public, the district court accepted the government's recommendation of 24 months' imprisonment. 6 The witness-tampering count alleges that LaShay "did knowingly attempt to corruptly persuade James Clark with the intent to influence his testimony" at LaShay's trial, in violation of 18 U.S.C. § 1512(b)(1). LaShay argues that there was insufficient evidence to support the jury's guilty verdict on this charge. He argues that there was no evidence that he threatened, intimidated, or harassed Clark, or in any way encouraged him to lie. 7 In reviewing for the sufficiency of the evidence, we consider the evidence in the light most favorable to the government and ask whether any rational jury could have found the elements of the offense beyond a reasonable doubt. United States v. Henningsen, 387 F.3d 585, 589 (7th Cir.2004). We neither reweigh the evidence nor substitute our judgment of the facts for that of the factfinder. United States v. Masten, 170 F.3d 790, 794 (7th Cir.1999). 8 In relevant part, § 1512(b)(1) subjects to imprisonment anyone who "knowingly uses intimidation, threatens, or corruptly persuades another, or attempts to do so," with the intent to "influence, delay, or prevent the testimony of any person in an official proceeding." To have convicted LaShay of this offense as alleged in the indictment, the government was required to prove that: 1) Clark was a witness or prospective witness; 2) LaShay attempted to persuade Clark to provide false testimony; and 3) LaShay acted knowingly and with the intent to influence Clark's testimony. United States v. Arocho, 305 F.3d 627, 639 (7th Cir.2002), superseded by statute on other grounds as stated in United States v. Rodriguez-Cardenas, 362 F.3d 958 (7th Cir.2004); United States v. Johnson, 903 F.2d 1084, 1087 (7th Cir.1990). 9 The jury's verdict is supported by sufficient evidence. LaShay focuses on the fact that he never threatened Clark, but a defendant need not use physical force or intimidation to be guilty of witness tampering, so long as he attempts to "corruptly persuade" a witness to testify falsely. United States v. LaFontaine, 210 F.3d 125, 133 (2d Cir.2000); United States v. Pennington, 168 F.3d 1060, 1066 (8th Cir. 1999); United States v. Gabriel, 125 F.3d 89, 102 (2d Cir.1997); see also Arocho, 305 F.3d at 640 (holding that conviction under § 1512(b)(1) was supported by sufficient evidence where defendants never threatened witness but repeatedly urged him to change his story and wrote false statement that witness signed). For instance, "corrupt" persuasion occurs "`where a defendant tells a potential witness a false story as if the story were true, intending that the witness believe the story and testify to it.'" Gabriel, 125 F.3d at 102 (quoting United States v. Rodolitz, 786 F.2d 77, 82 (2d Cir.1986)). In Gabriel, the defendant faxed a witness a false account of a meeting relevant to the government's investigation of whether the defendant made misrepresentations in pursuit of government contracts. Gabriel, 125 F.3d at 93-94. He urged the witness to "think this through" before answering any questions about his memory of that meeting. Id. at 94. The Second Circuit held that this was sufficient to support a conviction for witness tampering. Id. at 105. Here, LaShay similarly placed a story before Clark in the hopes that he would adopt it if interviewed. 10 Similarly, the Second Circuit upheld a district court's finding of witness tampering where the defendant tried to persuade a witness to give a false account that tracked the defendant's position. LaFontaine, 210 F.3d at 132 (affirming revocation of bail based on "practical probability" of witness tampering). The defendant had done nothing more than "remind" the witness that her mother had undergone a hernia operation, when in fact the mother had received cosmetic procedures which the defendant's clinic disguised as a hernia operation when billing the insurance company. Id. at 128. This case presents a similar situation, with LaShay repeatedly urging Clark to "remember" that he saw Hussain hand over the $2000 check for petty cash, when Clark knew he had not actually seen the check change hands. A jury could properly view LaShay's remarks as an unstated invitation to lie. Id. This is true even though the evidence at trial never established that LaShay's account of the check was false; LaShay was suggesting that Clark claim personal knowledge of the transaction when in fact he had none. See id. at 132-33 (practical probability of witness tampering even though defendant had some evidence to suggest that the version of events she proposed to witness could be true). 11 In his opening brief, which he filed before United States v. Booker, ___ U.S. ___, 125 S.Ct. 738, 160 L.Ed.2d 621 (2005), LaShay argues that the district court's use of the sentencing guidelines as a reference violated the Sixth Amendment. He contends that the court should have submitted the question whether he was an organizer or manager in the conspiracy to a sentencing jury. But this position is no longer tenable; Booker holds that application of the sentencing guidelines does not violate the Sixth Amendment so long as the district court treats them as advisory rather than mandatory. 125 S.Ct. at 750. The district court here made clear that it did not consider itself bound by the guidelines, but still found them to be a useful reference. Under those circumstances, LaShay had "no right to a jury determination of the facts that the judge deems relevant." Id. 12 In his reply brief, however, LaShay for the first time argues that the district court erred when it failed to resolve the dispute over the applicable guideline range before sentencing him. Typically, arguments first raised in a reply brief are considered waived. Kelso v. Bayer Corp., 398 F.3d 640, 643 (7th Cir.2005). But LaShay's argument here is based on Booker, which was not decided until after his opening brief had already been filed, meaning he raised the argument as soon as it was reasonably available to him. Accordingly, we will review the district court's sentence for harmless error, since LaShay raised a Blakely objection in the district court. United States v. Schlifer, 403 F.3d 849, 854 (7th Cir.2005). 13 LaShay is correct that, though the guidelines are no longer mandatory, district courts still must consult them and take them into account when sentencing. United States v. Baretz, 411 F.3d 867 (7th Cir.2005). The district court should have calculated the guideline range accurately and then explained any deviation it chose to make in LaShay's case. United States v. George, 403 F.3d 470, 473 (7th Cir.2005). Here, the district court did not bother to calculate a range at all, but simply accepted the government's sentencing recommendation. This was error. Id. 14 Nor can this error can be considered harmless. Although the government conceded at sentencing that LaShay had not committed the offense while on pretrial release, leaving only his purported leadership role in the conspiracy in dispute, this dispute might seriously have impacted LaShay's term of imprisonment. The government contended that the applicable guideline range was 21 to 27 months, but without this three-level increase in the offense level, LaShay's range would have been 12 to 18 months. LaShay's counsel argued for 14 months at sentencing. Under these circumstances, the government cannot demonstrate that the district court's failure to calculate the guideline range did not impact LaShay's sentence. 15 Accordingly, although we AFFIRM LaShay's convictions, we REMAND this case to the district court for resentencing. Notes: 1 Although counsel reported at argument that LaShay has now been released from imprisonment, his appeal is not moot because he is still serving his term of supervised releaseUnited States v. Trotter, 270 F.3d 1150, 1152 (7th Cir. 2001). On remand, the district court could still alter LaShay's overall sentence. Id.

1. Field of the Invention The present invention relates to a lens barrel which holds a zoom lens consisting of multiple lens groups, to a photographic apparatus which takes photographs by capturing light from a subject entering through the zoom lens held by the lens barrel, and to an optical apparatus which has a lens consisting of multiple lens groups. 2. Description of the Related Art Recently, digital cameras have been spreading rapidly, and increasingly higher image quality is demanded together with smaller size and thinner profiles. A thin card-size digital camera equipped with a lens barrel which holds a zoom lens consisting of multiple lens groups has been proposed (see Non-patent Document 1: Internet URL—http://www.business-ijP/sentan/jusyou/2003/pentax/) and introduced commercially as one of digital cameras which meet the above needs. This camera is capable of high image quality zooming using optical zoom whereas earlier thin digital cameras use an electronic zoom function for magnification. One of the trends in user needs is to have higher-powered optical zoom capabilities while achieving smaller size and thinner profiles. Patent Document 1 (Japanese Patent Laid-Open No. 2003-295031) proposes a technique for collapsing a lens barrel equipped with a zoom lens in a thin camera body using an ingenious method for forming cam grooves. An internal configuration of the lens barrel disclosed in Patent Document 1 is described below. FIGS. 1, 2, and 3 are sectional views taken along the optical axis of the lens barrel mounted in a digital camera. Of these, FIGS. 1 and 2 show the lens as it is extended. According to Patent Document 1, FIG. 1 is a diagram showing a telephoto end while FIG. 2 is a diagram showing a wide-angle end. FIG. 3 is a diagram showing the lens barrel as it is collapsed. FIG. 4 is a developed view illustrating cam grooves used to extend and collapse the lens barrel from/into the camera body. A configuration of a lens barrel 100 is described with reference to FIGS. 1 to 4. The lens barrel 100 holds a four-group zoom lens composed of a first lens group 210, second lens group 220, third lens group 230, and fourth lens group 240. Of the four lens groups, the second lens group 220 is moved along the optical axis for adjustment of focal distance while the fourth lens group 240 serving as a focus lens is moved along the optical axis for focus adjustment. The first lens group 210 is held in an inner tube 110. The inner tube 110 is equipped with cam pins 111, which are engaged with cam grooves 121 (see FIG. 4) formed in the inner surface of an outer tube 120. Three cam pins 111 are installed on the outer wall of the inner tube 110 at unequal intervals and three cam grooves 121 are formed on the inner wall of the outer tube 120 at unequal intervals (see FIG. 4) to engage with the cam pins 111. Consequently, when rotation of a zoom motor 270 is transmitted to a gear 124 installed on the inner wall of the outer tube 120 via a coupling gear 271 (see FIG. 2), rotating the outer tube 120, the first lens group 210 extends together with the inner tube 110 following the shape of the cam grooves 121. The second lens group 220 is held by a lens group holding frame 221, on whose circumference three cam pins 222 are installed at unequal intervals. The cam pins 222 are engaged with respective cam grooves 122 formed in the inner wall of the outer tube 120 (see FIG. 4). A guide rod 1132 is passed through a through-hole 221a made in the lens group holding frame 221 of the second lens group 220. Along with rotation of the outer tube 120, the second lens group 220 moves along the optical axis, being guided by the guide rod 1132. The guide rod 1132 is supported by a tip support 1132a while a guide rod 1133 which guides the third lens group 230 is supported by another tip support 1133a. The tip supports 1132a and 1133a support the respective guide rods 1132 and 1133 and are also used as members which support the inner tube 110. The tip supports 1132a and 1133a are equipped with an intermediate frame 1101 and a retainer ring 1102. The intermediate frame 1101 is inserted slidably along the inner wall of the inner tube 110 and the retainer ring 1102 is installed at the rear end of the inner tube 110 to prevent the intermediate frame 1101 from moving backward. The intermediate frame 1101 and retainer ring 1102 are equipped with respective spring pegs 1101a and 1102b. A spring 1103 is bridged between the spring pegs to restrict the movement of the intermediate frame 1101 by urging the intermediate frame 1101 forward along the sliding surface so that the intermediate frame 1101 will not move backward when it is extended together with the inner tube 110. Furthermore, a cam groove 123 (see FIG. 4) is formed between the cam grooves 121 and 122 in the outer tube 120 to engage with a cam pin (not shown) installed on a lens group holding frame 130 of the third lens group 230. Consequently, as the outer tube 120 rotates by receiving the driving force of the zoom motor 270 via the coupling gear 271 and gear 124 (see FIG. 2), the third lens group 230 moves along the optical axis following the shape of the cam groove 123. Incidentally, a shutter unit 131 is linked to the lens group holding frame 130 which holds the third lens group 230. A through-hole 131a is made in the lens group holding frame 130 as in the case of the lens group holding frame 221. The guide rod 1132 is passed through the through-hole 131a. Furthermore, the guide rod 1132 is also passed through a through-hole 141a made in a lens group holding frame 140 which holds the fourth lens group 240 described later. In this way, this example employs a configuration in which the second lens group 220, third lens group 230, and fourth lens group 240 are guided by the common guide rod 1132 to avoid misalignment of optical axes. Extension operation of the lens barrel 100 with this configuration is described in detail with reference to FIG. 4. When the outer tube 120 is rotated by the zoom motor 270, the inner tube 110 extends from a collapsed state to an extended position (B-side end of the area indicated by symbol A) following the shape of the cam grooves 121 (area indicated by symbol A) and held at the extended position (area indicated by symbol B). Until the outer tube 120 is held at the extended position, the second lens group 220 moves along the area indicated by symbol C following the shape of the cam grooves 122 and reaches the end of the area indicated by symbol C when the inner tube 110 is extended to the extended position. As a zoom switch (not shown) is operated at this time, the second lens group 220 enters the area indicated by symbol D, and moves to the end of the area indicated by symbol D if the zoom switch continues to be operated. On the other hand, as the inner tube 110 rotates, the third lens group 230 leaves a collapsed position, moves along the cam groove 123 through an extension area (area indicated by symbol E) and reaches an extended position (intersection of areas indicated by symbols E and F). It remains held at the extended position (area indicated by symbol F) even if the zoom switch is operated. In this way, by arranging cam grooves ingeniously in the inner wall of the outer tube 120, it is possible to move the first lens group, second lens group 220, and third lens group 230 among the four lens groups along the optical axis following the shape of the cam grooves by the rotation of the single tube 120 and move the second lens group along the optical axis by the operation of the zoom switch, and thereby do zooming. In the lens barrel 100 shown in FIGS. 1 to 3, the fourth lens group 240 at the tail of the four lens groups composing the zoom lens is used as a focus lens. The zoom lens held in the lens barrel 100 has a high zoom ratio, and consequently the fourth lens group 240 acting as the focus lens must have a relatively long travel distance. Thus, in this example, a column screw 1131 (see FIG. 1) as long as the long travel distance is installed along the optical axis, a nut 141b is fastened to the lens group holding frame 140 which holds the fourth lens group 240, and the column screw 1131 is screwed into the nut 141b for accurate focus adjustment. Focus is adjusted as the column screw 1131 is rotated by rotational driving force of a focus motor (not shown) transmitted via a gear train (not shown) and a lens group holding frame 141 moves along the optical axis by the distance equivalent to the rotation of the column screw 1131, being guided by the column screw 1131 and guide rods 1132 and 1133. When an image taking lens starts to catch a subject, focus is adjusted by generating image data by means of an image pickup device 280, detecting a focus position based on the image data, and moving the fourth lens group 240 acting as the focus lens to the focus position through rotation of the column screw. After focus adjustment, when a shutter button (not shown) is pressed, the shutter unit 131 provided in the lens group holding frame 130 which holds the third lens group 230 is operated in, synchronization with full depression of the shutter button to take a photograph. Consequently, light from the subject passes through the first lens group 210, second lens group 220, third lens group 230, and fourth lens group 240 (focus lens) and forms an image on a light-receiving surface of the image pickup device 280, which generates an image signal which represents the subject image formed on the light-receiving surface. Even if attempts are made to further reduce the length of the lens barrel, since the lens barrel shown in FIGS. 1 to 4 consists of two tubes, it is not possible to reduce the length sufficiently. It is conceivable to reduce the length of the lens barrel by increasing the number of tubes of the lens barrel to three or four. In that case, however, to arrange lens groups of the collapsed lens barrel along the optical axis, it is necessary to build a safety factor into the length of the lens barrel to allow adjacent lens groups to stay clear of each other. Besides, for high-powered zooming, it is also necessary to increase the length of the extended lens barrel to allow for a longer adjustable distance between lens groups composing the zoom lens. Thus, due to the increase in the length of the extended lens barrel, the length of the collapsed lens barrel cannot be decreased as it ought to be.

[Meckel's diverticulum and obstruction of the small intestine in adult patients: report of two cases]. Meckel's Diverticulum (MD) is the most frequent congenital anomaly of the ileum. However, its clinical manifestations in adult patients are rare. In this paper, we report cases with two different mechanisms of small intestine obstruction due to MD and we discuss diagnosis and management of these symptomatic lesions in adult patients.

Q: Python 3, for a in list and a in word: How would I write the following? a='somelongword' for r in a_list and for r in a: if r==x: new_list.append(r) else: return 1 Obviously, the above for statement isn't correct. How can I rewrite it? I'm sorry, I tried to make it easier to read! I want to check, from a list of letters, say a_list, if any of those letters are also in the word a. At the moment, my code checks this: for r in reveal_word(a): if r==guess: new_list.append(r) else: new_list.append('*') So, for some letter in the word a, if r is equal to some guessed letter, then append that letter to a list. Otherwise, when that letter is not equal to that guessed letter, append asterisks to the new list. It's for a piece of code for hangman. So, I want it to cycle through a list of guessed letters, check if the 'secret' word contains any of those letters and if it does, add them to a new list and where it doesn't contain those letters, add asterisks. I'll then add this to a dictionary object of key 'guessed_word' or something... A: Assuming this is what you want >>> word ='somelongword' >>> guesses = ['a', 'd', 'm', 'e', 's'] >>> [c if c in guesses else '*' for c in word] ['s', '*', 'm', 'e', '*', '*', '*', '*', '*', '*', '*', 'd'] Printed as a string >>> print(''.join([c if c in guesses else '*' for c in word])) s*me*******d That looks like what you would want for hangman.

var oauthModule = require('./oauth') , OAuth = require('oauth').OAuth; var yahoo = module.exports = oauthModule.submodule('yahoo') .definit( function () { var oauth = this.oauth = new OAuth( this.oauthHost() + this.requestTokenPath() , this.oauthHost() + this.accessTokenPath() , this.consumerKey() , this.consumerSecret() , '1.0', null, 'HMAC-SHA1'); }) .apiHost('http://social.yahooapis.com/v1') .oauthHost('https://api.login.yahoo.com/oauth/v2') .requestTokenPath('/get_request_token') .accessTokenPath('/get_token') .authorizePath('/request_auth') .entryPath('/auth/yahoo') .callbackPath('/auth/yahoo/callback') .fetchOAuthUser( function (accessToken, accessTokenSecret, params) { var promise = this.Promise(); this.oauth.get(this.apiHost() + '/user/' + params.xoauth_yahoo_guid + '/profile?format=json', accessToken, accessTokenSecret, function (err, data) { if (err) return promise.fail(err); var oauthUser = JSON.parse(data).profile; promise.fulfill(oauthUser); }); return promise; });

Q: Reorder not working in ggplot with my current data frame I'm currently trying to make my own graphical timeline like the one at the bottom of this page. I scraped the table from that link using the rvest package and cleaned it up. Here is my code: library(tidyverse) library(rvest) library(ggthemes) library(lubridate) URL <- "https://en.wikipedia.org/wiki/List_of_Justices_of_the_Supreme_Court_of_the_United_States" justices <- URL %>% read_html %>% html_node("table.wikitable") %>% html_table(fill = TRUE) %>% data.frame() # Removes weird row at bottom of the table n <- nrow(justices) justices <- justices[1:(n - 1), ] # Separating the information I want justices <- justices %>% separate(Justice.2, into = c("name","year"), sep = "\\(") %>% separate(Tenure, into = c("start", "end"), sep = "\n–") %>% separate(end, into = c("end", "reason"), sep = "\\(") %>% select(name, start, end) # Removes wikipedia tags in start column justices$start <- gsub('\\[e\\]$|\\[m\\]|\\[j\\]$$','', justices$start) justices$start <- mdy(justices$start) # This will replace incumbencies with NA justices$end <- mdy(justices$end) # Incumbent judges are still around! justices[is.na(justices)] <- today() justices$start = as.Date(justices$start, format = "%m/%d%/Y") justices$end = as.Date(justices$end, format = "%m/%d%/Y") justices %>% ggplot(aes(reorder(x = name, X = start))) + geom_segment(aes(xend = name, yend = start, y = end)) + coord_flip() + scale_y_date(date_breaks = "20 years", date_labels = "%Y") + theme(axis.title = element_blank()) + theme_fivethirtyeight() + NULL This is the output from ggplot (I'm not worried about aesthetics yet I know it looks terrible!): The goal for this plot is to order the judges chronologically from their start date, so the judge with the oldest start date should be at the bottom while the judge with the most recent should be at the top. As you can see, There are multiple instances where this rule is broken. Instead of sorting chronologically, it simply lists the judges as the order they appear in the data frame, which is also the order Wikipedia has it in. Therefore, a line segment above another segment should always start further right than the one below it My understanding of reorder is that it will take the X = start from geom_segment and sort that and list the names in that order. The only help I could find to this problem is to factor the dates and then order them that way, however I get the error Error: Invalid input: date_trans works with objects of class Date only. Thank you for your help! A: You can make the name column a factor and use forcats::fct_reorder to reorder names based on start date. fct_reorder can take a function that's used for ordering start; you can use min() to order by the earliest start date for each justice. That way, judges with multiple start dates will be sorted according to the earliest one. Only a two line change: add a mutate at the beginning of the pipe, and remove the reorder inside aes. justices %>% mutate(name = as.factor(name) %>% fct_reorder(start, min)) %>% ggplot(aes(x = name)) + geom_segment(aes(xend = name, yend = start, y = end)) + coord_flip() + scale_y_date(date_breaks = "20 years", date_labels = "%Y") + theme(axis.title = element_blank()) + theme_fivethirtyeight() Created on 2018-06-29 by the reprex package (v0.2.0).

Q: US extradition or domestic prosecution of US based foreign cyber crime? I’m a cybersecurity professional familiar with the extradition of foreign cyber criminals to the US. But what if the situation were flipped? Given the following scenario what would likely happen? US citizen on US soil Perpetrates cyber crimes with effects in foreign countries which do not have extradition treaties with the US Commits no acts of “cybercrime or intellectual property crime against the interests of the United States or the citizens of the United States” (6 U.S. Code § 1531) What is the likely legal recourse against such a person? Perhaps: Do nothing? Would the US view such an action as a prosecutable offense? Can the US (or does the US) domestically prosecute individuals for crimes against foreigners committed on US soil? Extradite, with or without a request from that country? A: The Computer Fraud and Abuse Act applies to "protected computers", defined amongst other things as: [...] used in or affecting interstate or foreign commerce or communication, including a computer located outside the United States that is used in a manner that affects interstate or foreign commerce or communication of the United States [...] The interpretation of "affecting" is very broad. AIUI if a computer sends packets over the Internet, and those packets enter the USA, then the computer has affected foreign commerce or communication of the United States. They don't need to be sent to a computer inside the USA; merely passing through a router on USA territory is sufficient, and it might even be interpreted to include a router owned by a USA company on foreign territory, as that would affect commerce. The size of the effect may be minuscule, but as long as it is not zero the computer that sent the packets is "protected". In your scenario the foreign computer must have sent packets to the criminal, so it must have affected US communications. Hence the law enforcement authorities would have a case under the CFA. Whether they would pursue that case in practice depends on a bunch of factors, including the amount of harm done, the likely costs of conviction, and probably political factors to do with international legal cooperation.

Great Cumbrian Run The Great Cumbrian Run is an annual half marathon road running event held in Carlisle, Cumbria, United Kingdom. Dave Cannon finished first in the inaugural Cumbrian Run in 1982 completing the course in 1:05:06 while Francis Bowen of Kenya holds the race record of 1:03:35 achieved in 2004. Course Recent winners References External links Category:Half marathons in the United Kingdom Category:Recurring sporting events established in 1982 Category:Sport in Cumbria Category:1982 establishments in England

Q: How to scale a texture in webgl? I have a texture of size 800x600. How do I scale it on a webgl <canvas> at another size and keep the original aspect ratio? Assuming that the drawing buffer and the canvas have the same dimensions. A: Given the WebGL only cares about clipsapce coordinates you can just draw a 2 unit quad (-1 to +1) and scale it by the aspect of the canvas vs the aspect of the image. In other words const canvasAspect = canvas.clientWidth / canvas.clientHeight; const imageAspect = image.width / image.height; let scaleY = 1; let scaleX = imageAspect / canvasAspect; Note that you need to decide how you want to fit the image. scaleY= 1 means the image will always fit vertically and horizontally will just be whatever it comes out to. If you want it to fit horizontally then you need to make scaleX = 1 let scaleX = 1; let scaleY = canvasAspect / imageAspect; If you want it to contain then let scaleY = 1; let scaleX = imageAspect / canvasAspect; if (scaleX > 1) { scaleY = 1 / scaleX; scaleX = 1; } If you want it to cover then let scaleY = 1; let scaleX = imageAspect / canvasAspect; if (scaleX < 1) { scaleY = 1 / scaleX; scaleX = 1; } let scaleMode = 'fitV'; const gl = document.querySelector("canvas").getContext('webgl'); const vs = ` attribute vec4 position; uniform mat4 u_matrix; varying vec2 v_texcoord; void main() { gl_Position = u_matrix * position; v_texcoord = position.xy * .5 + .5; // because we know we're using a -1 + 1 quad } `; const fs = ` precision mediump float; varying vec2 v_texcoord; uniform sampler2D u_tex; void main() { gl_FragColor = texture2D(u_tex, v_texcoord); } `; let image = { width: 1, height: 1 }; // dummy until loaded const tex = twgl.createTexture(gl, { src: 'https://i.imgur.com/TSiyiJv.jpg', crossOrigin: 'anonymous', }, (err, tex, img) => { // called after image as loaded image = img; render(); }); const programInfo = twgl.createProgramInfo(gl, [vs, fs]); const bufferInfo = twgl.createBufferInfoFromArrays(gl, { position: { numComponents: 2, data: [ -1, -1, // tri 1 1, -1, -1, 1, -1, 1, // tri 2 1, -1, 1, 1, ], } }); function render() { // this line is not needed if you don't // care that the canvas drawing buffer size // matches the canvas display size twgl.resizeCanvasToDisplaySize(gl.canvas); gl.viewport(0, 0, gl.canvas.width, gl.canvas.height); gl.useProgram(programInfo.program); twgl.setBuffersAndAttributes(gl, programInfo, bufferInfo); const canvasAspect = gl.canvas.clientWidth / gl.canvas.clientHeight; const imageAspect = image.width / image.height; let scaleX; let scaleY; switch (scaleMode) { case 'fitV': scaleY = 1; scaleX = imageAspect / canvasAspect; break; case 'fitH': scaleX = 1; scaleY = canvasAspect / imageAspect; break; case 'contain': scaleY = 1; scaleX = imageAspect / canvasAspect; if (scaleX > 1) { scaleY = 1 / scaleX; scaleX = 1; } break; case 'cover': scaleY = 1; scaleX = imageAspect / canvasAspect; if (scaleX < 1) { scaleY = 1 / scaleX; scaleX = 1; } break; } twgl.setUniforms(programInfo, { u_matrix: [ scaleX, 0, 0, 0, 0, -scaleY, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, ], }); gl.drawArrays(gl.TRIANGLES, 0, 6); } render(); window.addEventListener('resize', render); document.querySelectorAll('button').forEach((elem) => { elem.addEventListener('click', setScaleMode); }); function setScaleMode(e) { scaleMode = e.target.id; render(); } html, body { margin: 0; height: 100%; } canvas { width: 100%; height: 100%; display: block; } .ui { position: absolute; left: 0; top: 0; } <script src="https://twgljs.org/dist/4.x/twgl-full.min.js"></script> <canvas></canvas> <div class="ui"> <button id="fitV">fit vertical</button> <button id="fitH">fit horizontal</button> <button id="contain">contain</button> <button id="cover">cover</button> </div> The code above uses a 4x4 matrix to apply the scale gl_Position = u_matrix * position; It could just as easily pass in the scale directly uniform vec2 scale; ... gl_Position = vec4(scale * position.xy, 0, 1);

Konstantin Zhuravlyov Konstantin Zhuravlyov (born 13 February 1976) is a Uzbekistani sprinter. He competed in the men's 4 × 100 metres relay at the 2000 Summer Olympics. References Category:1976 births Category:Living people Category:Athletes (track and field) at the 2000 Summer Olympics Category:Uzbekistani male sprinters Category:Olympic athletes of Uzbekistan Category:Place of birth missing (living people)

I’m different now, Montgomery said. We [he and Gimenez] joked about it the other day. He didn’t have to block as many balls in the dirt this time. I used to spike a lot of heaters. Gimenez admitted that his shins and knees took a beating that minor league season, but now the lefty throws […]

Q: User Current Location Prompt i am deleting and re-installing the application, it still not prompt the user "Use Current Location Prompt" when they select the "Use current Location" .button A: The iPhone caches the location services authorization status. Even if you reinstall the app, it won't prompt you about this. You can reset this (for all apps!) under Settings.app > General > Reset > Reset Location Warnings

Q: My symbols are rendered as ff. Have I ran out of memory? I have been trying to solve a 4x4 matrix and have done a hell a lot of writing. Now I am near the end and I am not getting the requested output out of the LaTeX compiler. When I write for example: \begin{minipage}[t]{\textwidth} \begin{equation}\label{e50} R^{-1} = \frac{1}{\left|R\right|} \rm{adj}(R) = \alpha \end{equation} \end{minipage} the last symbol isn't shown as "alpha" but rather as a very weird symbol ff like I show in the picture below: I get the same weird symbol for any symbol that I put in place of \alpha or even after it. From this point on all of my symbols are typeset as ff. If I put same source code in my other new document all of my symbols again render as they were supposed to. Q1: Have I ran out of memory? Q2: How can I fix this? A: You shouldn't be getting ff instead of \alpha. But your input has a bad error anyway: \rm is not a command with an argument, but rather a declaration that extends its influence until the group in which it's been issued ends, in this case up to \end{equation}. The command \rm must not be used. Forget it, together with \it, \bf and \tt. They are obsolete and, as in your case, can be misunderstood by users. In place of them use \textrm, \textit, \texttt and the others that you find in any user guide. In math there are \mathrm and some similar commands, but for operator names there's a better way: say \usepackage{amsmath} \DeclareMathOperator{\adj}{adj} in your preamble and then your formula can be input as \begin{equation}\label{e50} R^{-1} = \frac{1}{\lvert R\rvert} \adj(R) = \alpha \end{equation} The minipage you use is not needed and probably makes vertical spacing awkward. Note also \lvert and \rvert instead of \left| and `\right|.

Any UAB employee who handles or offers for transport an Infectious Substance, Category A substance must complete training every two years or if regulations change. This is a Federal requirement per the Hazardous Materials Regulation. A copy of the completed training certificate must be maintained in the laboratory files and presented to the proper authorities upon request. Once you pass the assessment for the course, you may go to the My Transcript tab and located the course link. Once you click on the link, you will see another link that brings up your certificate for the course. Choose File, Print to print it.Below you will find reference material and job aids to assist you in shipping Infectious Substances, Category A. After completing the course, feel free to return here and print any of the material you find useful.

1. Field of the Invention The present invention relates to a motor-vehicle-mounted radar apparatus mounted on a vehicle and used to explore obstacles such as other vehicles traveling around the vehicle, for example, ahead of the vehicle, as targets thus securing safety of driving. 2. Description of the Related Art Conventionally, motor-vehicle-mounted radar apparatus that explores obstacles of a vehicle in the direction of traveling has been developed. As motor-vehicle-mounted radar apparatus, the FM-CW system is employed in which an exploration wave as a continuous sending wave frequency-modulated using a triangular wave as a modulating wave and a beat component caused by a reflected wave from a target are extracted as a beat signal, and the relative velocity and relative range to the target are obtained based on the frequency of the beat signal. Related arts concerning the FM-CW system radar apparatus are disclosed, for example, in the Japanese Patent Laid-Open No. 111395/1977, the Japanese Patent Laid-Open No. 120549/1995, the Japanese Patent Laid-Open No. 80184/1997 and the Japanese Patent Laid-Open No. 145824/1997. Especially in the Japanese Patent Laid-Open No. 120549/1995, a configuration is disclosed whereby the radiation direction of a beam-shaped radio wave transmitted from motor-vehicle-mounted radar apparatus can be changed in order to correctly explore a target such as a vehicle traveling diagonally ahead for example in curvilinear traveling. FIG. 16 shows a schematic configuration of conventional motor-vehicle-mounted radar apparatus of the FM-CW system 1. The motor-vehicle-mounted radar apparatus 1 explores a target 2 and transmits a radio wave for exploration from an antenna 3 in order to calculate the range to the target 2 and the relative velocity to the target 2. The antenna 3 receives the reflected radio wave reflected off the target 2. The antenna 3 is formed in a beam shape having a sharp range whose gain is high. Thus it is possible to execute scanning with beam direction changed via a scanning mechanism 4 and to detect the direction of the target 2 from the direction of beam in receiving a reflected signal from the target 2. With the range to and direction of the target 2 obtained, the position of the target 2 can be relatively obtained based on the position of the vehicle. In the exploration according to FM-CW system, a transmitter circuit 5 is used to give an exploration signal frequency-modulated by a triangular wave to the antenna 3 for transmission, a reflected signal received by the antenna 3 is amplifier and frequency-converted by a receiver circuit 6 and converted to a digital signal by an analog-to-digital (hereinafter referred to as A/D) converter circuit 7, then converted to a frequency component by a fast Fourier Transform (hereinafter referred to as FFT) circuit 8. An object detection circuit 9, based on the range R and the relative velocity V to the target 2 based on the frequency component from the FFT circuit 8. Each of FIGS. 17A and 17B show a principle in which the object detection circuit shown in FIG. 16 explores the target 2 and calculates the range R and the relative velocity V in accordance with the FM-CW system. From the antenna 3 in FIG. 16, an exploration signal 10 for frequency-modulated continuous wave (CW) is transmitted so that frequencies maybe continuously varied on the triangular wave at a constant variation velocity. An exploration 10 wave reflects off the target 2 and a resulting reflected signal 11 that is received by the antenna 3 is delayed as long as the period corresponding to the range R from the exploration signal 10. This causes difference in frequencies for the exploration signal 10 whose frequency is in variation. The relative velocity V is generated to the target 2. Thus the Doppler shift effect is generated on the reflected signal 11, causing difference in frequency from the exploration signal 10. In the FM-CW system, as shown in FIG. 17B, variation in frequency caused by the Doppler shift effect is reflected differently on an upbeat signal 12 as a beat signal in the frequency rise section where the frequency shift amount of frequency modulation is increasing, and on a downbeat signal 13 as a beat signal in the frequency drop section where the frequency shift amount of frequency modulation is decreasing. Thus the frequency fub of the upbeat signal 12 and the frequency fdb of the downbeat signal 13 can be represented as the following expressions 1 and 2 using the range frequency fr and the Doppler shit frequency fd that are standard beat signal frequencies. fub=frxe2x88x92fdxe2x80x83xe2x80x83(1) fdb=fr+fbxe2x80x83xe2x80x83(2) Here, the range frequency fr is in proportion to the range R to the target 2 and can be represented by the following expression 3 assuming the frequency shift amplitude of the exploration signal of the FM-CW system 10 as a triangular wave as xcex94f, modulating frequency as a triangular wave fm, and the velocity of light C. The Doppler shift frequency fd can be represented by the following expression 4 assuming the relative velocity to the target 2 as V and the wavelength of the exploration signal as xcex. It is also possible to calculate the range R and the relative velocity V respectively from the range frequency fr and the Doppler shift frequency fd by using the expressions 3 and 4. xe2x80x83fr=4xc3x97xcex94fxc3x97fmxc3x97R/Cxe2x80x83xe2x80x83(3) fd=2V/xcexxe2x80x83xe2x80x83(4) As shown in FIG. 16, related arts concerning the motor-vehicle-mounted radar apparatus that scans the beam direction of the antenna 3 are disclosed, for example, in the Japanese Patent Laid-Open No. 64499/1999, the Japanese Patent Laid-Open No. 72651/1999, the Japanese Patent Laid-Open No. 84001/1999 and the Japanese Patent Laid-Open No. 121053/1999. In the Japanese Patent Laid-Open No. 82673/1996, a configuration whereby the short range and long range are switched over for exploring a target. In the Japanese Patent Laid-Open No. 282220/1998, a related technology is disclosed whereby part of data obtained in radar exploration is used to identify a target for a radar for an airframe. In the Japanese Patent Laid-Open No. 38141/1999, a related art is disclosed whereby a mobile-vehicle-mounted radar is used to recognize an obstacle in a three-dimensional image. Of the related arts that scan the beam direction of an antenna, for example in the Japanese Patent Laid-Open No. 84001/1999 and the Japanese Patent Laid-Open No. 231053/1999, a philosophy is described that a plurality of explorations are carried out in a single scan period and the target direction is estimated from the peak of the reflected signal level obtained according to the variation in beam direction. In order to upgrade the exploration accuracy of the exploration in the target direction using such a philosophy in the related arts, it is necessary to explore a target in more directions and to increase of the frequency of exploration. Such a method suffers from high load of operation processing so that special hardware for high-speed signal processing is required to process data at a high speed. High-speed signal processing has a problem of heat as well as costs. Smooth operation requires a corresponding circuit scale thus upsizing the system configuration. A method is also available whereby the limits of angle of exploration in a specific section where a target is present is narrowed. This approach requires a complicated hardware configuration and has few merits in terms of costs. In recognizing a target, it is necessary to prepare a complicated logic in order to extract a true target in case reflected signals from a guard rail, tunnel, or sound-proof wall is received. In the case of a guard rail, the relative velocity calculated after paring processing in which the frequency in the frequency rise section and the frequency in the frequency drop section according to the FM-CW system are combined does to equal the actual velocity of the vehicle. The target thus appears as a moving object, not a stationary object. The travel amount does not coincide with the value obtained from the relative velocity so that the target is determined as a guard rail based on such information. However, as the frequency of exploration increases, the data update rate is accelerated and travel amount is decreased. Thus the relative velocity obtained from the travel amount is less accurate and makes difficult the comparison of relative velocity. Further, in the FM-CW system, while a combination of the frequency rise section and the frequency drop section is used to calculate the range and the relative velocity, the Doppler shift effect is large for an object approaching at a high relative velocity so that the difference in frequency is large between the frequency rise section and the frequency drop section. In the case of an approach to a target, the beat signal in the frequency rise section deviates to the lower frequency band and processing is difficult for an extremely low-frequency beat signal thus increasing the minimum range in which a target can be detected, compared with the stationary state. As a result, it is impossible to trace the approaching target and get information on whether the target is approaching or deviating in another direction. In case targets detected during driving are determined as stationary objects, they include objects that the vehicle can clear or pass through. Conventionally, these objects are not under specific criterion but subject to alarms or deceleration control. While a stationary objects may not be determined as a target but excluded from the target of control for these objects, objects that cannot be cleared or passed through cannot be excluded from the target of control. An object of the invention is to provide motor-vehicle-mounted radar apparatus that can correctly recognize the position and nature of the target. In the invention, there is provided a motor-vehicle-mounted radar apparatus that is mounted on a vehicle for exploring targets around the vehicle, characterized in that the apparatus comprises: an antenna formed to have a high gain in a predetermined beam direction, for transmitting an exploration signal in the beam direction and receiving a reflected signal from a target of the exploration signal, scanning means for performing a scan that varies the beam direction of the antenna within predetermined limits, direction detecting means for detecting the beam direction varied by the scanning means, exploration control means for making control to scan repeatedly in the beam direction of the antenna via the scanning means within the predetermined limits and to explore targets in a plurality of beam directions detected by the direction detecting means, each direction making an angle with each other every time exploration is made, and target recognizing means for calculating the range to a target based on the reflected signal from the target received by the antenna and the exploration signal transmitted from the antenna and for recognizing the target based on the range and the beam direction of the antenna detected by the direction detecting means. According to the invention, the motor-vehicle-mounted radar apparatus that is mounted on a vehicle for exploring targets around the vehicle transmits an exploration signal in a predetermined beam direction from an antenna and receives a reflected signal from a target. The beam direction of the antenna is varied within predetermined limits by scanning means in a scan and detected by direction detecting means. Exploration control means makes control to scan repeatedly in the beam direction of the antenna via the scanning means within the predetermined limits and to explore targets in a plurality of beam directions detected by the direction detecting means, each direction making an angle with each other every time exploration is made. Target recognizing means calculates the range to a target based on the reflected signal from the target received by the antenna and the exploration signal transmitted from the antenna. Target recognition is made based on the range and the beam direction of the antenna detected by the direction detecting means. Since a plurality of beam directions that are different from antenna scan to antenna scan is obtained by exploration control means, a combination of exploration results obtained from a plurality of explorations assures as high accuracy in exploration as the results obtained from explorations made in beam directions making a smaller angle with each other. Since an angle between beam directions in explorations during a single scan may be larger than a final angle, the target direction can be determined at an accuracy similar to that in a high-speed processing. The invention is characterized in that the target recognizing means recognizes a target based on the results of explorations in a plurality of beam directions per scan and recognizes the target based on a combination of exploration results obtained from a plurality of prespecified scans when the number of beam directions of a reflected signal is smaller than a prespecified reference value. According to the invention, recognition of a target can be made based on exploration results obtained from a plurality of scans in case a sufficient number of exploration results cannot be obtained via a single scan. This can upgrade the target recognition accuracy. The invention is characterized in that the target recognizing means calculates a Doppler shift frequency from frequency shift amount that accompanies traveling of a target based on exploration results obtained from the plurality of scans and varies reflected signals to be combined for target recognition depending on the calculation results. According to the invention, it is possible to combine candidate exploration results considering time delay between a plurality of scans in order to perform high-accuracy recognition. The invention is characterized in that the target recognizing means recognizes reflected signals having frequencies excluded from the combination depending on the exploration results as reflected signals from unwanted reflecting objects, not as reflected signals from a target. According to the invention, reflected signals having frequencies that cannot be combined among a plurality of exploration results are recognized as reflected signals from unwanted reflecting objects, not as reflected signals from a target. Thus it is possible to avoid burdening the system with load of processing on a target that need not pay attention to and perform processing focused on a target that need to pay attention to. The invention is characterized in that the target exploration is performed via the FM-CW system and that the target recognizing means subtracts the Doppler shift component of the velocity of the vehicle from the peak data obtained from reflected signals having frequencies excluded from the combination depending on the exploration results in the frequency rise section and the frequency drop section of the FM-CW system in order to calculate the range and the direction. According to the invention, the Doppler shift component of the velocity of the vehicle is subtracted in the frequency rise section and the frequency drop section of the FM-CW system in order to calculate the range and the direction even if the exploration results are excluded from the combination for a specific target. This allows the position of a target to be explored efficiently. The invention is characterized in that the target recognizing means recognizes the calculation results of the range and direction as data on unwanted reflecting objects and obtains the position where unwanted reflecting objects assemble from the range and the direction, and determines the position as a shoulder. According to the invention, the calculation results based on exploration results that could not be combined are recognized as data on unwanted reflecting objects and the position where unwanted reflecting objects assemble is determined as a shoulder Thus, an object that is on or beyond the shoulder can be excluded from attention in order to reduce processing load. In the invention, there is provided a motor-vehicle-mounted radar apparatus that is mounted on a vehicle for exploring targets around the vehicle, characterized in that the apparatus comprises: an antenna formed to have a high gain in a predetermined beam direction, for transmitting an exploration signal in the beam direction and receiving a reflected signal from a target of the exploration signal, and target recognizing means for calculating the range to a target based on the reflected signal from the target received by the antenna and the exploration signal transmitted from the antenna, for recognizing the target based on the range and the beam direction of the antenna detected by the direction detecting means, and for determining the height of the stationary target depending on whether or not a multipath error has effects on the variation in the reflected signal level within a specific frequency limits caused by the range from the target. According to the invention, it is determined whether or not the target can be cleared based on the variation in the level of the reflected signal from the target caused by the range from the target. The reflected signal from the target is either a signal directly received by the antenna or a signal reflected off a road surface and received by the antenna. These signals generate a phase difference based on the path difference of the reflected signals thus varying the level of a signal received by the antenna. When the target is low, variation in the level of a reflected signal is small and the possibility of clearing the target is high. Since it is determined whether or not the target can be cleared based on a variation in the level of a reflected signal caused by the range from the target, a secure decision is possible via a simple configuration. A reflected signal from a target on the road surface on which the vehicle is traveling passes through a plurality of paths, i.e., a path directly reaching the antenna and a path once reflected off the road surface and reaching the antenna, and thus received in a state where a phase difference based on the path difference is generated. Phase difference occurs between reflected signals received in such a multipath error. Thus a phase difference close to 180 degrees results in a large signal attenuation. Since such effects caused by a multipath error appear depending on the height of a target, it is possible to properly determine the height of a stationary target based on the effects of multipath error. The invention is characterized in that the target recognizing means determines that the stationary target can be cleared in case the level of the reflected signal from the target suddenly drops in an approach to the target. According to the invention, it is determined that the target can be cleared in case the level of the reflected signal from the target suddenly drops in an approach to the target. Since the target that can be cleared has a small height from the road surface and goes beyond the beam direction limits of the antenna when the range to the antenna becomes smaller, the level of the reflected signal suddenly drops. Thus, in case the level of the reflected signal from the target suddenly drops in an approach to the target, it is highly possible that the target is at least lower than the antenna position and thus can be cleared. The invention is characterized in that the target recognizing means has data indicating the variation in the threshold value of the reflected signal level for the range in advance and determines that the stationary target can be cleared in case the level of the reflected signal from the target has dropped below the threshold value within a predetermined limits of range. According to the invention, a map is formed indicating the variation in the threshold value of the reflected signal level for the range in advance. Thus, in case the level of the reflected signal in an approach to the target has dropped below the threshold value, it can be determined that the stationary target is low and can be cleared. In the invention, there is provided a motor-vehicle-mounted radar apparatus that is mounted on a vehicle for exploring targets around the vehicle, characterized in that the apparatus comprises: an antenna formed to have a high gain in a predetermined beam direction, for transmitting an exploration signal in the beam direction and receiving a reflected signal from a target of the exploration signal, and target recognizing means for calculating the range to a target based on the reflected signal from the target received by the antenna and the exploration signal transmitted from the antenna, for recognizing the target based on the range and the beam direction of the antenna detected by the direction detecting means, the target recognizing means having data indicating the variation in the threshold value of the reflected signal level for the range in advance and determines that the height of the target based on the range where the reflected signal from the stationary target has dropped below the threshold value. According to the invention, the range where the reflected signal level drops below the threshold value while the user""s car is approaching the stationary target corresponds to the height of the stationary target. Thus it can be determined that the target is low in case the reflected signal level drops at relatively long range and gets higher as the level drops at relatively short range. The invention is characterized in that the target recognizing means determines whether or not the stationary target can be cleared based on the state in which the reflected signal level drops as the range gets shorter. According to the invention, it is determined that the target can be cleared in case the reflected signal level has dropped above a certain extent in an approach to the target from the level at long range. This evades the effects of a variation in the reflected signal level caused by difference of target material. In the invention, there is provided a motor-vehicle-mounted radar apparatus that is mounted on a vehicle for exploring targets around the vehicle, characterized in that the apparatus comprises: an antenna formed to have a high gain in a predetermined beam direction, for transmitting an exploration signal in the beam direction and receiving a reflected signal from a target of the exploration signal, and target recognizing means for calculating the range to a target based on the reflected signal from the target received by the antenna and the exploration signal transmitted from the antenna, for recognizing the target based on the range and the beam direction of the antenna detectedby the direction detecting means, and for determining the height of the stationary target based on the range where the reflected signal level suddenly drops in an approach to the target. According to the invention, the height of a stationary target is estimated depending on the range where the reflected signal level has dropped below a certain level in an approach to the stationary target compared with the level at long range. When the stationary target is relatively low, the target goes beyond the beam limits from the antenna at relatively long range thus avoiding a drop in the reflected signal level. In case the target is relatively high, the drop in the reflected signal level becomes large at short range. Since the reference level used to evaluate the drop in the reflected signal level is the reflected signal level from the target at long range, effects such as the target material can be reduced to estimate the target height at a high accuracy. In the invention, there is provided a motor-vehicle-mounted radar apparatus that is mounted on a vehicle for exploring targets around the vehicle, characterized in that the apparatus comprises: an antenna formed to have a high gain in a predetermined beam direction, for transmitting an exploration signal in the beam direction and receiving a reflected signal from a target of the exploration signal, and target recognizing means for calculating the range to a target based on the reflected signal from the target received by the antenna and the exploration signal transmitted from the antenna and for recognizing the target based on the range and the beam direction of the antenna detected by the direction detecting means, the target recognizing means recognizing that the target is not a target having a height to be alerted for a traveling vehicle in case the reflected signal level that is recognized exceeds the prespecified reference while the target recognizing means is recognizing the stationary target at longer range than the predetermined range and the reflected signal level drops considerably as the vehicle approaches the target from the vehicle position at the time of recognition. According to the invention, it is recognized that the target is not a target having a height to be alerted for a traveling vehicle in case the reflected signal level that is recognized exceeds the prespecified reference while the stationary target at longer range than the predetermined range is recognized and the reflected signal level drops considerably as the vehicle approaches the target from the vehicle position at the time of recognition. For example, in case a reflected signal is received from an object, such as a billboard, a sign, and a two-level crossing, above the road surface on which the vehicle is traveling, the target is within the antenna beam direction limits at a distance but goes beyond the antenna beam direction limits as the vehicle approaches the target and causing the reflected signal level to drop considerably. When such an object as will not interfere with the traveling of the vehicle is detected, the object is recognized as a target not to be alerted. This eliminates control such as unnecessary alert or braking. The invention is characterized in that the target height is the height of the highest section of the target. According to the invention, it is possible to determine the possibility of clearing the target from the height of the highest section of the target. The invention is characterized in that the target recognizing means determines whether or not the target can be cleared based on the determined height. According to the invention, it is possible to determine whether or not the target can be cleared from the height of the highest section of the target. This assures a correct decision. The invention is characterized in that the target height is the height of the lowest section of the target. According to the invention, it is possible to determine the possibility of passing through the target from the height of the lowest section of the target. The invention is characterized in that the target recognizing means determines whether or not the target can be passed through based on the determined height. According to the invention, it is possible to determine whether or not the target can be passed through from the height of the lowest section of the target. This assures a correct decision. The invention is characterized in that the target recognizing means determines the target height in a plurality of sections depending on the range from the target and derives a signal for predetermined alarm and/or braking based on the decision results for each section. According to the invention, the target height is determined based on the presence/absence of the effects of a multipath error in a plurality of sections in an approach to a stationary target. Thus it is possible to make a plurality of decisions in an approach to the target and to avoid unnecessary alarm or braking in case it is determined that the target can be cleared in an early period. The invention is characterized in that the target exploration is performed via the FM-CW system and that the target recognizing means uses only data in the frequency drop section to estimate the range and the relative velocity to the target when the target recognizing means determines that the relative velocity to the target is larger than the reference velocity. According to the invention, it is possible to perform an highly accurate detection of a target that is rapidly approaching.

Alimak takes your personal data seriously. We use cookies to optimize our websites for your needs. By using this website you consent to our personal data policy. If you want to find out more or disable cookies, please click here. ALIMAK SC 65 The all-new fully modular ALIMAK SC 65 is based on the proven design of the ALIMAK SCANDO 650, which has been very successful in the market. Alimak has taken construction hoist technology to unprecedented levels with the introduction of the all-new Alimak SC 65; a modular system that can adapt to different applications using relatively few modules and components. Available in single or twin car configuration, the Alimak SC 65 offers payloads of 1,500 to 3,200 kg/car and has a standard maximum lifting height of 250 m, which can be increased on request. The free-standing mast height is 22.5 m when equipped with an optional pipe support. Two or three FC frequency control or two DOL motors drive the system through a high efficiency gearbox. The Alimak SC 65 operates at speeds up to 65 m/min and provides more hoistpower, while using as much as 40% less power than previous models. In a 'time is money' world, we ensure customers get materials and personnel transported higher, faster and safer. Payload capacity 1,500–3,200 kg Speed 38–65 m/min. Max. lifting height 250 m (Increased lifting height on request.) Car width (internal) 1.5 m Car length (internal) 3.2–5.0 m Car height (internal) 2.3 m Motor control DOL/FC No. of motors 2–3 Safety device type GFD-II Power supply range 380–500 V, 50 or 60 Hz, 3 phase Type of mast A-50, Tubular steel with integrated rack Length mast section 1.508 m Weight mast section with 1 rack 118 kg Rack module 5 Request for quotation Construction Hoists If you have an inquiry, please complete the form below and we will return with a quote.

The best thing I can pass along is not to make the common mistake of selling your services below market value. New business owners often get flustered by not getting the amount of business they want...

Of course now that it’s 2010 they’re no longer called expat wives, they’re “trailing spouses,” yep, thanks for that, I feel so much better now. I love the visual of me trailing behind G, hunched over and waiting for direction. Maybe we’ll forget about the title. So, who and what is she? In my experience she’s like any group of women, she’s a nurse, a doctor, a dentist, a hairdresser, a chef, a banker. The one thing she usually has in common with her expat friends, is that at some stage she sat down with her partner and had to make a practical choice on whether they were going to take “the job” overseas. In our case, I was 8 weeks pregnant when that conversation came. We did the math and it seemed impractical to turn the job down, the salary G was offered was the nearly the same as our two salaries in Australia, our worries of affordable child care and negotiating maternity leave arrangements would be non existent, it just seemed to make sense to go. G was an expat child, he was incredibly excited about hitting the road again, there was a piece of family nostalgia there for him and he was happy with the idea of showing a child the expat life, me, not so much. The plan was 2 years in Indonesia, save some money, enjoy the experience and come home. I didn’t resign from work, I took a leave of absence, 11 years later and I still haven’t been able to formally resign from that role. What do you think Freud would say about that? When we arrived in Jakarta and G went off to his first day at the office, I sat in our hotel room looking out over the grey city skyline, all logic and practicality disappeared from my mind. I quickly forgot our agreement. I wondered what on earth had possessed me to give up my career, friends and family to take on the role where my whole existence appeared to be being Mrs G. In fact, that’s what the staff at the hotel called me, Mrs G! As I wandered around the city I felt incredibly lonely. If I wasn’t working then who was I? I kept looking in the mirror at my 5 month pregnant body not really knowing who she was either. After a couple of very quiet days the phone began to ring, British, American and Australian accents at the end of the line. “My husband mentioned there was a new Australian at the office and his wife was pregnant, do you have a doctor? I had a baby last year” a woman with a thick Scottish accent said. Someone invited me on a museum tour, someone else for a coffee “have you heard about ANZA?”. None of these women were the same, they were all from different parts of the world, all different ages but they had all been the woman in the hotel room, they had a pretty good idea on what was going through my mind. When I started to spend time with them I realized that it doesn’t matter if you’re a hippy, or a conservative, at any age, the story from the very well dressed dignified woman in the corner about how she had to poo in her handbag while stuck in traffic in Mumbai with a serious case of Delhi belly is hysterical to everyone. They laughed about their language disasters, rats in their dryer pipes, no electricity or phone for days, cold showers, doctors who diagnosed them with terrible non existent diseases and the tragic haircut where “just cut a little bit off” translated to “just leave a little bit there” (it took me two years to grow that haircut out). An expat wife acquires the skill of looking across the room and thinking (as my friend Jen later told me) “I’ll have her, she’s mine” as they see something in someone that looks familiar. A lifelong friendship can be made in a moment, over the death of a family member or a terrifying health scare for a child. You’ll find yourself sharing intimate stories with a friend you’ve only known for a few weeks, the terrible ex boyfriend, the miscarriage and the fight you had with your sister when you were 8, because you need to share, if you’re going to be good friends she needs to know the details. That’s why when you phone her the next day to say the car won’t start and your husbands in China, she’ll be there. An expat wife will nervously walk in to a room full of strangers biting the side of her cheek, armed with a list of questions Is the milk okay to drink? Do you have a good doctor, mechanic, dentist or physio? Can you draw me a map to the school? Where do I buy a decent bra? What sort of cab should I get in to? Do they have Napisan here? Why is there a sign “this meat does not contain traces of mad cow disease” in the supermarket? Why can’t I find tampons? Where can I find a math tutor? It will be more than likely that she will leave the room with the answers, a list of phone numbers and an invitation for tomorrow. She may not have met one person she can see herself being friends with but that fear of never meeting anyone will be gone. She’ll feel indestructible, it will be better than the best performance review she’s ever had. That weekend you’ll see her, leading the way with her trailing spouse behind her, she’ll be showing him how the city works and what she’s learnt during the week, because in reality we all know who the real trailing spouse is. Share this: Thanks so much for that perspective Great post. I am a bit of an ex-pat here, but have moved to my husband’s home town/country, so it wasn’t a work thing. We both had to leave our jobs in Oz to get here. I am lucky to have found a group of ex-pat English speakers, all of whom are lovely and have filled me in on all the important things you mentioned re. kids and baby stuff… but I sometimes get envious of the beautiful houses, and ‘work-paid-for-our-move-and-shipped-our-stuff’ side of things, (we are living in a cheap, tumbledown house as we came over here pennyless) BUT, we have family here. Cousins, Aunts, Grandparents to babysit and to spend weekends with etc. And that is why we are here and that support is invaluable. So I will try not to be tooo envious of my new ex-pat friends… And I will hope that they stick around here as long as me!! http://livinglifeasme.wordpress.com/ livinglifeasme Great insight into your world as an ex pat, trailing spouse (WTF)! Sounds like, despite the initial isolation and unknown, the kinship women share is universal. And that is wonderful. xx http://www.blogger.com/profile/08404421856986720832 Kerri Sackville Brilliant post. FASCINATING. May I suggest you write a book? ‘Mrs G – Tales of a Trailing Spouse’. I’d buy it for SURE.xxxxxx http://www.blogger.com/profile/07529172123543772763 Linda T Fabulous post, and I love the name Shamozal, it’s one of my favourite words x I agree with Kerri, a book is a must!! http://www.blogger.com/profile/05615149112130152767 Christy I have to admit, I’m a little jealous – I know it must be hard too, but it sounds awesome! Great post; I really enjoyed reading it! A nice peek into a totally different lifestyle than my own. http://anjwritesabout.com/ anjwritesabout.com I can identify with a lot of what you write, except of course, I bounce into overseas moves excitedly! Agree wholeheartedly with Kerri as well…write these stories down (had an idea about that myself some time ago) as they will certainly sell!! http://davesag.tumblr.com/ Dave Sag Great post. I was an expat for years and now I live in Canberra I still feel like an expat. http://www.blogger.com/profile/02743336097657087832 JANE Really fascinating, Kirsty. Brilliant insights. It took me back to when I was 18 and an exchange student in Germany for the year. First time away from home – what a culture shock! The expats were such a supportive bunch and made life so much easier. J x http://www.blogger.com/profile/04729571991002856725 gilly troo. great post. thanks! katherine Oh, Kristy, thank you so much for this.I am tangentially familiar with the expat life — my mom has been one for 40+ years and I know what her community of expat friends has done for her; and I, myself, have lived overseas for half my life (though most of that has been as a student. ) Nevertheless, I am a very reluctant to become that “trailing wife” — my husband is already in Doha, and my not-yet-two year old daughter and I are still behind in NY, just because, if I stay here a little longer, I will be given the chance to keep my job and work from afar — from Doha. Mind you, I don’t even really like my job. I don’t like being away from my husband. I certainly don’t like my daughter being away from her father. And I so very much wish I could spend more time with my daughter, rather than running on this mad treadmill, trying to balance work, family, and now single parenthood…. Yet… Yet. I am terrified of becoming “Mrs K” http://www.blogger.com/profile/16561240096806868742 Toni “Trailing spouse”??? what the ….??? that made me want to punch someone.I’m in awe of ex-pat wives.My hubs is currently working overseas and until last week, it was looking like we would be joining him.While I was kinda excited, I also had so many questions — which he couldn’t answer, despite having been in this job for over 6 months. Things like, what are the supermarkets like? Would I be able to tell what was in cans and packets? What would happen if the kids got sick? etc etc. That was kind of one of the reasons why I started following your blog, cos I knew I would need to know people who could help out while I found my feet. (plus it’s just so INTERESTING!)I know a few ex-pat wives, in Switzerland, Sakhalin, Indonesia, Japan. Their husbands would be the first ones to admit they couldn’t do their jobs as well, without their wives making a home for them to come to every night. http://juststopspeaking.wordpress.com/ juststopspeaking Fantastic post – and can I just say – WOW you’re brave. We did the travel thing-but from one side of Oz to the otherand whilst it may have been in the same country but, can I tell you, worlds apart….. and unfortunately the locals Not.So.Welcoming. Anyhoo – we got through and came out the other side – what does not kill us makes us stronger Write a book – I’d buy it for sure. http://www.blogger.com/profile/09677312773827236567 Kath Lockett Great post, Shamozal. Just in case you were wondering, I never thought of the expat wife as ‘gin swilling’ – I always assumed that they had it harder than their working spouse – ‘things’ such as education, house, bills paid, outings arranged and familiarisation don’t happen by themselves….. Ex Dhaka Great post – brings back wonderful memories of my time as the … non working part of the couple. First six weeks (in a hotel) were awful – particularly as husband went to Dublin for 2 weeks 3 days after kids n I arrived, and all the expats went on 2 months summer leave a week after that. But things improved once we were settled. I’ve always described it as a shallow but fulfilling life filled with coffee, shopping, and some incredible friends who I will never forget. http://www.blogger.com/profile/17798190669591053390 Expat mum A penny just dropped for me – I consider myself an expat, even though I’ve been in the States for 20 years. The reason I still feel like an expat and not an immigrant I’ve just realised, is because most other people here grew up here so I’ve had very few people to bond with. Aaaah… Darinaw Hi Kirsty,What a gift !!! I have been an Expat for almost 30 years, 10 different countries, and three kids. You have just articulated beautifully, what it’s all about. Trailing spouse…. I don’t think so !!! Well done. xx http://www.blogger.com/profile/05388656296518535181 Kristin :) Hi! I’m a fellow expat, and I loved reading your post!! You couldn’t have said it better! Expat wives couldn’t be any farther from a trailing spouse! http://www.blogger.com/profile/07695203425270299420 Very Bored in Catalunya I love this post! Being an ex-pat wife really does mean you have to break out of your comfort zone, but the rewards in befriending people who you would normally come across are brilliant. http://www.21stcenturymummy.com/ 21st century mummy Great post. I am about to become an expat. We are moving to Singapore and I am pregnant. It’s exciting but scary too. http://www.blogger.com/profile/15757789955664659689 Ingrid I love your post! I can totally relate to you wrote, I have recently (one month today !!) moved to Doha and am finding things a bit tough, meeting other women in the same boat has been my life saver. Thanks for brightening up my day. http://www.blogger.com/profile/18254275544017629129 bigwords is… Great to know women throughout the world reach out to others when you need it most. Sounds like a perfect first chapter xx http://www.blogger.com/profile/02777700715209427380 SDSJMcManus Great post Kirsty, I recently met a woman who before leaving her home country went for a mammogram and was clear. 3 Months into expat life she was showering and found the lump we all dread, she went into overdrive and needed to find a doctor quickly. She had just met my friend who lives in the country she had just expatriated to and confided in this stranger who had happened into her life…. Within days she had a doctor, a best friend and a group of other expat women that supported her through this unforeseen challenge. Her special group of friends, who looked after children when needed, made meals, held her hand and became her best friends. 6 months later she is in remission, but without that knowledge and special connection that she made it most certainly would have been much more difficult for her to survive this. AND she made a best friend for life! http://www.blogger.com/profile/08680909356053872951 Lynn MacDonald Wow…that was an incredibly interesting post. I just visited my cousin who has lived in London for close to 20 years. I met soooo many people from NZ and Austrailia, South Africa, Switzerland….It was so interesting that I almost got jealous of the it multinational existence. Sometimes, I wish I had been more adventurous. I came out to join a French boyfriend who I then married and then later divorced so I’m an expat but not a very mobile one because I’m still here! I’m also integrated into the French system as I work and pay my taxes and am a resident although I couldn’t ever face taking French nationality! So I recognise a lot of what you say – having expat friends is a real plus in a strange land because coping with strangeness 24/7 is very exhausting and it’s good to giggle with people who know who you are, if you see what I mean. Of course, there are some totally weird and bizarre people living as expats who I would avoid with a bargepole wherever I was. http://www.lifeintheexpatlane.com/ Miss Footloose Hi Kristy, glad you found me! Yes, I certainly identify with (almost) everything in your post having lived in a number of foreign countries. I was married in Kenya, had my first baby in Ghana, and so on and so forth. Been the trailing spouse all right. Hailing from the Netherlands, I have a travel/exploring gene and I always wanted to see the world, so I got lucky. Lucky also because I have my own portable career as a writer, which means I don’t have to “sacrifice” myself for the good of my husband’s career. http://www.blogger.com/profile/08635723714090524880 Lorna Great Post! – if you ever need help in Houston, TX contact Lorna – [email protected] for 2 years and now been here 8 years – sound familiar! http://www.kylieladd.com.au/ Kylie L I was a trailing spouse for 5 years- first in Scotland, then in Canada. The experience was tough, lonely, frightening (the latter two especially in Montreal, where it seemed no-one spoke English and I had two tiny children), brilliant, character-forming, mundane, fascinating and overwhelming. I was actually termed a “dependent spouse” in the Uk, where I was refused a work permit unless I had the same surname as my husband, necessitating a lot of paperwork & a new signature before we left (and this was in 1999, not 1959!). Though tough at times, I loved our half decade overseas. More to the point, they gave me my writing… I began writing out of loneliness in Scotland after we first arrived, then really got into it seriously (and started the novel that would become After The Fall, published last year) in Montreal, where I couldn’t work at my ‘real’ job (as a psychologist) due to language restrictions. Ten years later, we’ve come full circle… I’m a trailing spouse again, though this time following my husband’s mid life crisis to Broome, WA- and with my writing as my only occupation. Thank you for a lovely post prompting all these memories! http://www.blogger.com/profile/06972149809628579689 Alexis Jacobs As a “trailing spouse” myself, all I have to say is AMEN! You hit the nail on the head. Great writing! Pamela May I re-post your blog on my Facebook page? It will give my friends back in the US a wonderful perspective on what life is like overseas! http://globalcoachcenter.wordpress.com/ globalcoachcenter Love the post and, especially, the “we all know who the real trailing spouse is”! Funny and very true at the same time! I’ve loved all my expat assignments and, in fact, I find myself itching for the next one when we live back home (like now). I help others love it too so a dream job/lifestyle combination. http://www.blogger.com/profile/15425166687421078790 homekeepsmoving Hi, I have just stumbled upon your blog – I love it! Thought you might be interested in mine…I recently wrote a book about growing up as a Third Culture Kid and it got published this summer…Heidi http://homekeepsmoving.blogspot.com/ My husband is an expat kid too – and here we are in the UK. Love the post http://www.blogger.com/profile/03924035710478459520 EmmaK Brilliant! I suppose it was the opposite for me because I moved from UK to USA where there is much more choice of products. Like my head was spinning in the supermarket thinking: why are there 56 different versions of mayonnaise and whicih one shall I choose??? http://www.blogger.com/profile/15999271127735997770 Naturally Carol When I was eighteen I moved to Australia from New Zealand. I didn’t expect so many differences in culture. It was lonely in Canberra at that time…then I met my husband. Five grown up kids later and now in Queensland…I’m still here. I love your post, what misconceptions are out there. You are brave women! Elissa Sarich Great post Kirsty (as always)…keep them coming! http://www.blogger.com/profile/08647596319711811125 cjtato Really loved reading this. It reminds me that some of our best friends are found in places we least expect. I have found two extremely great friends through meeting mothers at school (something I swore I’d never do for fear of only having children as a common bond but then realised that this is what starts the friendship but very rarely continues it). On a weirder note, the expat lifestyle appeals to me but then I’ve always been a traveller so could explain it. Hubby, however, has his feet firmly planted on his home turf. LOL http://swrightboucher.wordpress.com/ swrightboucher Kirsty, this is your best yet. Killer closing line. You nailed it! Thank you for letting us travel with you. Anonymous Brilliant! Every word you wrote is so true! After 7 International moves I am known as ‘Tagga’ – short for tag-a-long! It has a nicer ring than ‘trailing spouse’ but only just. http://www.blogger.com/profile/09890862541890241422 Raine and Sage My husband and I met travelling, and have consequently lived and worked in a few different cities, and now renovating our 2nd home. I wonder if we got addicted to moving?! I’ve not heard of trailing spouses before. It’s a funny term, and condescending a bit. We were considering if we’d move to Qatar about 18mths ago… Great post. I love how you write! It gave me insight into expat wives as I do think they get a bad rap. At least they have each other though. I am the one who is working and there doesn’t appear to be an “expat husbands” club. And the expat wives won’t have much to do with me…mostly because they form their friends when I can’t. Company parties aren’t a lot of fun! Working wives want girlfriends too Brett Great post, we’re from Australia also but now in Chennai (or Madras as it was once called) before that Singapore. You’re right my wonderful wife Helen does all the hard work and makes everything happen. She putting out 4 year old to bed after a mad hospital dash across horrendous wet season traffic. Some sort of weird fever that seems to have settled down now. I can’t wait to show her your post as it will be like looking in a mirror for her. I spent a long time expatting alone and then ended up back in sydney for a few years where we met and married. After a one year of parenthood in Sydney I persuaded Helen to hit the road and while it’s been crazy we are having fun and are closer in way that I don’t think we would have if we’d stayed in surburbia. (Plus I would have died of boredom)Once you’ve survived a fight with a monkey in your kitchen your world is irreversibly different.keep up the blog, you have a gift that should be shared!CheersBrett http://www.blogger.com/profile/13700717302832203631 London City Mum Fabulous post. Where do I join? Can swap you for Kevin. LCM x Single Working Girl Abroad Ladies, suggest you purchase a book called “Diplomatic Incidents: The Memoirs of an Undiplomatic Wife” by Cherry Denman, available on amazon.co.uk – so, so funny, I found myself laughing out loud with no-one there to hear me. Might chivvy Mrs G into writing her own book. http://www.blogger.com/profile/04925824005829442967 Expatriate Chef Nice post, am sending the link to a friend who is living this life, too! Thanks for stopping by my kitchen. Thank you so much for the comments. I don’t have the technological capabilities to do single replies here in the comments so where I can I’m coming to find you! Either on your blogs or via your emails. To those who are thinking of living in Doha or are currently living in Doha please drop me an email at [email protected] I would REALLY like to talk to you. The book is on the way. I was contacted by a publisher a few weeks ago and fingers crossed 4 kids, 20 suitcases and a beagle will get written and published in a year or (depending on how demanding the beagle is) two. Once again, thanks for the comments, I loved reading them, (particularly the monkey fighting Australian), new post coming tomorrow, I hope you can find time to come back again. Kirsty http://www.blogger.com/profile/00621170366257326241 MultipleMum It certainly sounds like an interesting way to spend a life! I can’t say it would be for everyone, but as a fellow traveller, I reckon I could get used to it too. Make the most of any situation you find yourself in because it is unique and interesting and yours http://www.blogger.com/profile/14796758557993482943 Desiree Loved the post and your writing. I was an expat for most of the past fifteen years. Met great people and had great experiences, some of the best friends ever. Am having a hard time with our most recent move back to the US. Tough to not have the expat identity and the willingness of people to get to know you and have a coffee! My goodness, how brave are you?! I get stressed just going on holiday never mind living in numerous different countries! http://www.blogger.com/profile/09763064258020825441 Elizabeth Abbot great post (a friend forwarded it to me). I see you have the Expat Women website listed — check out the December article on “The Pyramid of Expat Needs”. It’s a model I often use in trainings and presentations that seems to hit the spot. Best wishes! Elizabeth http://www.blogger.com/profile/10652566920769043373 Olga Thank you for this post. It brings so many memories. http://www.blogger.com/profile/12951250749027163763 Victoria Hi! Thanks for stopping by my blog.. hopefully I’ll be posting more regularly now that I’m no longer pregnant!I’m a grad school wife, and it sounds a bit like being an ex-pat wife.. except I assure you no one ever thinks we’re going to be glamorous! But you DO have to get in good with women quickly, be sharp & have a talent to stand out. Whew! I’ll visit your blog again soon! http://www.blogger.com/profile/16232358885177196136 katherine It was like reading a short story about my life as an ‘expat wife’. I found myself smiling from ear to ear and laughing out loud. SO TRUE! I lived in Abu Dhabi for 13 years and came to the States for 2 years to get my teens settled into college. Now, it’s back to UAE and the hubby – and I’m more than ready to end that long-distance commute thing. Every expat woman will tell you it started out as a two year contract, but they all end up staying. When we meet a new girl in the group and she talks about that 2-year contract schpiel, we all hold in our giggles. The truth is – once you become an expat, you can never UNbecome one. You will always relate most easily with other expats once you’ve lived in this way. I have missed that camaraderie so much. Even as I contemplate leaving Alabama and moving back in January to Abu Dhabi, we are planning his retirement in South America. And so it goes! Anonymous Absolutely loved this post! We were much older when we began our expat adventures, and we are coming up on our 1-year anniversary. We are the only ones from my husband’s company in the country, and I truly love all the women I’ve met. However, I’m still wondering if I’ll find that one BFF. Every time I find someone that I really like and think we could be best of friends, I find out she is leaving in the next month or so. Maybe lots of friends will have to do instead of one BFF? Thank you for making my day with your blog! Anonymous and just imagine what’s about being a “trailing husband”because if the majority is female gender, some are male… http://www.blogger.com/profile/03482513767849843084 My New Normal Great post, and I can totally relate as an expat living in London. The expat community over here has been my family away from home and they have helped me through some very difficult times. http://www.blogger.com/profile/10244439931661355439 qatar foodie This is a great post and just for that I voted for you at babble and you are 16 now… the book idea suggested by some in the comments is worth a thought… keep blogging wow, i’ve been also the woman in the hotel room already in brussels, warsaw and now in a small village close to Geneva (with a 2 months old baby girl) and next year there ll be a new destination…I started my blog yesterday and I ended up here browsing the lists of other mothers expat blogs.. you really hit the point!! http://www.blogger.com/profile/17165208811776097332 Heather Excellent post lady! Just excellent! Nicola Adair I’m the blonde Scottish woman you met at the end of last night’s QPWN event. Really enjoyed chatting to you and so couldnt resist looking up your blog….I love it! Your comments re Trailing Spouse made me laugh and cry – your description of feeling empowered after walking into a room full of unknown women is exactly how I felt last night……and today I have an extra spring in my step due to my ‘bravery’! http://www.blogger.com/profile/04503327503533034631 Rhanda This post is great! I am an American, now living as an expat in Oman (we are practically neighbors). We have been here 2.5 years! I can so relate with what you have written. My second week here, I found myself in tears at the local market unsuccesfully trying to find pesto. Now, I see it all over the place. Oh to have a WalMart (or Target) here! A girl can dream can’t she? Wonderful post! Thanks for your lovely comment on my blog, I am so glad we have found each others blogs – as me and our two young girls are soon moving to Hong Kong (hubby is already there), so I guess I am soon to be an expat wife (I will not call myself the other version, it is awful)I could be stopping by your blog to pick your brain when I am wondering what the flippin heck I have done!Nice to meet you Cat http://www.preludefinancial.com.au/ Expats To be honest, the days of gold-rimmed expat packages are disappearing fast, whoops … they are gone, and therefor many expats now work under reasonably modest (local) terms, but no matter how you look at it, your life is often more luxurious. http://www.blogger.com/profile/08499944412217904302 vegemitevix I can’t believe I haven’t read this brilliant post before Kirsty. It really does say it all. I wrote about trailing spouses on my blog too – Who Moves, and also about the loss of identity when you move – Who am I? Can’t wait for your book to come out. Vxx Wonderful post! Thanks for your lovely comment on my blog, I am so glad we have found each others blogs – as me and our two young girls are soon moving to Hong Kong (hubby is already there), so I guess I am soon to be an expat wife (I will not call myself the other version, it is awful)I could be stopping by your blog to pick your brain when I am wondering what the flippin heck I have done!Nice to meet you Cat This post is great! I am an American, now living as an expat in Oman (we are practically neighbors). We have been here 2.5 years! I can so relate with what you have written. My second week here, I found myself in tears at the local market unsuccesfully trying to find pesto. Now, I see it all over the place. Oh to have a WalMart (or Target) here! A girl can dream can’t she? http://www.blogger.com/profile/09763064258020825441 Elizabeth Abbot great post (a friend forwarded it to me). I see you have the Expat Women website listed — check out the December article on “The Pyramid of Expat Needs”. It’s a model I often use in trainings and presentations that seems to hit the spot. Best wishes! Elizabeth http://www.blogger.com/profile/04925824005829442967 Expatriate Chef Nice post, am sending the link to a friend who is living this life, too! Thanks for stopping by my kitchen. Hi, I have just stumbled upon your blog – I love it! Thought you might be interested in mine…I recently wrote a book about growing up as a Third Culture Kid and it got published this summer…Heidi http://homekeepsmoving.blogspot.com/ http://globalcoachcenter.wordpress.com/ globalcoachcenter Love the post and, especially, the “we all know who the real trailing spouse is”! Funny and very true at the same time! I’ve loved all my expat assignments and, in fact, I find myself itching for the next one when we live back home (like now). I help others love it too so a dream job/lifestyle combination. http://www.lifeintheexpatlane.com/ Miss Footloose Hi Kristy, glad you found me! Yes, I certainly identify with (almost) everything in your post having lived in a number of foreign countries. I was married in Kenya, had my first baby in Ghana, and so on and so forth. Been the trailing spouse all right. Hailing from the Netherlands, I have a travel/exploring gene and I always wanted to see the world, so I got lucky. Lucky also because I have my own portable career as a writer, which means I don’t have to “sacrifice” myself for the good of my husband’s career. Darinaw Hi Kirsty,What a gift !!! I have been an Expat for almost 30 years, 10 different countries, and three kids. You have just articulated beautifully, what it’s all about. Trailing spouse…. I don’t think so !!! Well done. xx http://juststopspeaking.wordpress.com/ juststopspeaking Fantastic post – and can I just say – WOW you’re brave. We did the travel thing-but from one side of Oz to the otherand whilst it may have been in the same country but, can I tell you, worlds apart….. and unfortunately the locals Not.So.Welcoming. Anyhoo – we got through and came out the other side – what does not kill us makes us stronger Write a book – I’d buy it for sure. Ex Dhaka Great post – brings back wonderful memories of my time as the … non working part of the couple. First six weeks (in a hotel) were awful – particularly as husband went to Dublin for 2 weeks 3 days after kids n I arrived, and all the expats went on 2 months summer leave a week after that. But things improved once we were settled. I’ve always described it as a shallow but fulfilling life filled with coffee, shopping, and some incredible friends who I will never forget. http://davesag.tumblr.com/ Dave Sag Great post. I was an expat for years and now I live in Canberra I still feel like an expat. http://strongsoutherly.blogspot.com/ Andrea So so so right – I wasn’t the expat spouse, just the expat, but I certainly experienced the cameraderie and openness, and the willingness to make people feel welcome. It is something you really miss when you come back to Aus. Everybody has their lives, their routine, their set ‘support crew’, so when you move somewhere it doesn’t take a number of weeks to find people, it takes years. Even now, 5 years on my friendship group pales in comparison to my expat one. Lynn There’s no life like it! Robandmerinplus Everything you say about the life of an expat wife is just so true…..but now being back “home” after a 2 year expat experience I think I’d like to go again!!! Life is never, ever boring or mundane as an expat wife!! Rachel I adore this post. You have summed up the strength it takes to get out there, and the sheer bloody mindedness you need to make a success of it – all while pinning a smile on your face to camouflage the butterflies in your stomach. Can’t wait to read the rest of your posts! LK Great post! So true on every count! Was an expat girlfriend for a year, and now an expat wife for 10…. Your post just reminded me of my early days! Nice one!!! Fam Wanten could I use this article to publish in our bimonthly newsletter of the British Womens Association? That would be a great story to share .. thanks in advance Anonymous I love this!!! Everything is so true. After 3 different countries in a 12 year span we are now living back in the USA. At the beginning I didn’t want this kind of life, but after our third move ( where 2 of my 3) children were born I actually began to love the expat life. We saved a great amount of money, we learned another language, we had a ton of help , and some of my best friends have come from these adventures. I love being settled in one place again, but I will never forget the experiences our family shared. Never say never.. http://irunyourun.wordpress.com/ Carla I wish I had the same experience as you did, arriving in a foreign country and having other spouses reach out. I instead was alone in that room (not a hotel, but isolated temporary quarters nonetheless) for a month. (To be fair one spouse did reach out and I had a glorious second day in town, but she was pregnant and on her way out, so I was left alone after that, since she had urgent matters to take care of, but the people who were here? Nothing. I vowed to never be that person, and I have met friends by reaching out instead as 5 months later, I’m no longer “new”) Amwaters You got the essence of “trailing” spouse expat wife. There is a fantastic play here in singapore called Expat Wife and it was also so close to the truth. Melinda Mccabe That is thee truth. Living in India was all that you listed but the friends I made seem closer then friends I have in the states… http://twitter.com/TorreDoro Dorotea Torretta Hi, I’m @TorreDoro . I read this blog last week and I totally agree with the content. When I started my expat-wife life, 16 years ago, I arrived in a little town in Central Mexico, 2 hours by car from Mexico City. My daughter was 6 and my son was 1 year old. We came from a rich, well served and safe North of Italy, Milan. In that “pueblito” there wasn’t anything. Even expats, apart a very small Italian colony, my husband’s colleagues, but as was summer the few wives leaving there were on vacation abroad.The second day there, I called a pediatrician (suggested by hubby’s colleague) for my son. This very generous man the same day sent his wife to my place to help me. Since that afternoon we became inseparable. She was my sister, mother, friend, conforter.I don’t see her in person since 14 years, but we are continuously in touch – by e-mails, phone, FB, skype: our affection still lasts, no matter the distance. All of us can tell stories like this. How is painful grow away from these persons that are part of our life, that our children love as the true family. And how is extraordinary being aware that this kind of ralationships can be built wherever. Cheers,Dorotea Rissa Hi Kirsty, in a few months I will become a trailing spouse/expat wife and this blog made me feel so much better! We haven’t even started to pack yet and I am already feeling that I am losing my identity. The thought that after we move, it will be me who has the answers to the ‘where do we go?’ ‘where can I get?’ is reassuring. It’s me who has those answers here and I like the fact that this at least, won’t change! Keep writing, your words are thoroughly entertaining!!! Em Preparing for “my” first expat wife assignment (China) I came accross your blog. Thanks for sharing your experiences. Look forward to new making friends but will miss the “old” ones very much… Jeanne @ Collage of Life I am not sure if I comment on this before…even if I did, I have to say, once again, this is brilliant and so true! Well done! Patricia LOve your blog! Would you like to follow each other? My blog is on spanish designers I promote , style and my experience as a western woman and expat wife in Qatar! So true as also the last paragraph !If it is the partner to follow he’ll receive the nice title of “prince consort”… but mens are certainly not so organized as we are and certainly are not sharing ideas or comfort or help in welcoming groups. They could really struggle to adapt in a new environment that is not work ! Sandra Great article! As an expat wife, I reconglize myself on that story. I remember our first move to Vietnam, what a shock!! first time in Asia for us, with a little one, and getting pregned in the first month there. I was not happy at all. But everything changes when you start meeting people. Making friends make your life so much better werever you are. We still moving and moving, I like this life. Meeting new people is a little bit of a chalenge and more when your kids start getting older, but knowing a new place with exiting things to know and learn, is great!Even tho, the move is one of the most stressful things in a person life, we start been a little bit of an expert, still very stress out, but well…. it came w the territorie, right? I think now a days I couldn t live anymore in a one place, may be onde day we move to a place were we ll feel like home and we stay there forever. For now….forever is a world we do not use. There is nothing forever in life, so we ll enjoy while we can. Melina Brilliant Post!!! Loved reading it…and looking back through our blog! Im a newbie Expat, having just done a year in KL but on our way to Doha now…perhaps we can meet up once we are there! Thanks for a brilliant blog! http://www.blogger.com/profile/05681227038111050561 Melody This is such a fabulous post. You wrote everything just the way it is. And you always think you are alone until you start talking to other westerners around you. http://www.blogger.com/profile/07285709209953730580 Fraudster Again, fabulous writing. Cheers for the insight on a Sunday morning in same old Melbourne. http://www.blogger.com/profile/10122834410061492424 bloggingher lovely post…i totally relate…is that book out yet? u really should write one… http://www.blogger.com/profile/01095913641167115662 Chuzai Living I’m ‘training spouse’ in Jakarta and sometimes I feel that we get a better deal. What you wrote is so true and I am nodding as I read. Great writing! http://www.blogger.com/profile/01023787854786251105 aussiemama Another expat wife/trailing spouse/mum here. We moved from Sydney to Dublin (2 years) then Yokohama (3 years) adn now in the UK. I can totally relate to all that you wrote in this blog – tho rather than ‘she’s mine’ i call it ‘girl dating’ as you meet someone for the first time, get to know them, chat a bit, after a few meetings you exchange phone numbers, then wait to see who rings who first, then offer/accept an invitation to coffee ….. and before you know it the deepest of friendships is formed, or not. And you learn that’s ok too. Then you add husbands and kids into the mix and again, it either works or it doesn’t. it’s hard work moving country every few years, closing one chapter and opening another. Long gone are the days of tennis, lunches, endless shopping, expat clubs etc (tho i hear it also depends on where you are gigging and the industry the employed partner is in). The move to the UK’s actually been very difficult cos we’re not living in an expat community. Like a lot of expat wives/ trailing spouses i also started blogging just before we left Japan as friends said ‘write a book’ . Love to have your company and comments at ms-havachat.blogspot.com meanwhile, onwards and upwards http://www.blogger.com/profile/10441442451871326442 Chris Wow, thanks. I appreciate you wrote this post 2 years ago, but it’s a great inspiration. I moved to Ghana with my husband and three kids from Australia at the beginning of this year. As we live an hour out of the second biggest city, and we don’t have a expat pool of people close by, and it’s been tough. Your post has been a kick up the butt to get out there a bit and enjoy some of the upsides to expat life! And no more feeling guilty about morning tennis lessons…. Anonymous After 2 years still loving this article!!! Anonymous Thank you for the good writeup. It in fact was a amusement account it. Look advanced to far added agreeable from you! Terrific post! I sometimes think it’s easier to make friends living abroad than it was back in an American city – at least among other expats. You share the same struggles and can empathize with newcomers’ challenges. It’s a small, generally supportive, manageable community. I actually think I’ll miss that easy camaraderie when we go back to the States. Loved this post! I immigrated from the UK to the USA to be with my husband, and it’s been an interesting journey of constant adaptation and unpredicted culture shock to say the least! Got to find the humor in it though don’t you!? I love this – completely spot on. I lived in Qatar last year, Russia and Libya before that and we’re off to Angola; every new start begins with at least one day in the hotel/apartment figuring out of it was the right move. I’ve got a bunch of lovely friends around the world now though, all of whom have helped me out when needed and who I will do the same for in a heartbeat. The bit about being able to show my husband (aka the office dweller) around the city when we’ve just arrived is so true. http://www.blogger.com/profile/11555869161874166113 Megan Absolutely right! Thanks for articulating it all. In the past 10 years of doing some “trailing” of my own I have encountered my share of “ladies who lunch”, but mostly I have also met the most important women in my life who have continued to be what I assume will be lifelong friendships, and who have gone out of their way to be there for me and either help me feel at home in a new home, or have joined forces in figuring it all out. People may often be nervous about moving to a new country, especially if they don’t know what they will be doing there, but I have been much more enthusiastic about moving to a new place where I knew I would feel more “at home” with others who know what it’s like to be the new kid on the block than in a new town in my home country. Plus we have friends all over the world whose couches we look forward to crashing on for many years to come! http://www.blogger.com/profile/01401601942849055034 Jane Lee I’m impressed with your ideas and the way you express them in this article. I share many of your views and enjoyed reading this article. It’s a pleasure to see well-written original content these days. What about trailing husbands? Think they are a new trend without the wives support network to help them… And often ‘trailing’ wives who are not earning as much as your “expat wives'” spouses (e.g. Many teachers in my experience)… Where do we fit in? http://shamozal.blogspot.com Kirsty Rice 4kids20suitcases Thanks Esta, I’ve written about trailing husbands before, but this post was focussed on women. Regarding your comment on teachers, the post was about ALL women e.g.. “In my experience she’s like any group of women, she’s a nurse, a doctor, a dentist, a hairdresser, a chef, a banker.” Cheers and thanks for your comment. Esta Styles Ah, this was posted by someone and popped up on my FB feed. Haven’t seen your other posts I was definitely nodding my head when you said “That weekend you’ll see her, leading the way with her trailing spouse behind her, she’ll be showing him how the city works and what she’s learnt during the week” that totally describes me. Thanks so much for writing this post and sharing the true essence of what an expat wife is and is not. Thank you so much for your article, you made me laugh and cry at the same time … and you gave me hope! I am a new expat wife and mum to a 6 month old in Bishkek, Kyrgyzstan (your will probably need to google map that one), what I mean by new is I am 11 days in and I am slowly getting to that stage of feeling a little bit alone when walking around the city. But I am determined to find a friend with the added challenge of my husband working as 1 of 2 expats in the company (the other expat doesn’t have a partner over here) So I have a question… what advise would you give to new expat wife in a random country wanting to find new friends? http://shamozal.blogspot.com Kirsty Rice 4kids20suitcases Hi Marta, are you on the 4 kids, 20 suitcases Facebook page. I might put a note out there asking if anyone else is there at the moment. I’ve definitely had comments from Kyrgyzstan before. Let me know when you “like” the page and I’ll put a note out to readers. xxx Marta Diemar Hello Kirsty, Thank you so much for your reply and doing that for me! Meeting fello expat wives would be really nice right about now. I ‘liked’ your page yesterday you will be able to find me with the same name. x AnExpatWife LOVE this blog… I’ve been an expat wife for nearly 10 years now and am in my 4th and hopefully final country !! I wouldn’t change any of it if I had the choice again. I’ve shed many tears and met lots of ‘ladies who lunch’ but I’ve also met some amazing people, travelled to some great places and even got the chance to study for 2 bachelors degrees…. Your post describes many things perfectly !! It really takes another expat to truly understand the frustrations. almost_bedtime hi kirsty – i am a british expat living in canada who is moving to australia in a couple of months time – one of my blog readers suggest that i check out your blog and i’m so glad i did – i lived in the states for 10 years and had a brilliant expat community but i haven’t found one at all in vancouver and i’ve found it very hard going – just reading this piece has given me hope that when we move to Aus i have a good chance of finding a new community again – many of the women i met here had never left vancouver and although i tried really hard to fit in sometimes i just needed to chat to someone who wasn’t from here – i wish i had discovered your blog sooner but i love pieces like this – they make me feel normal – thanks

// stdafx.cpp : source file that includes just the standard includes // xrLBuilder.pch will be the pre-compiled header // stdafx.obj will contain the pre-compiled type information #include "stdafx.h" // TODO: reference any additional headers you need in STDAFX.H // and not in this file

Ben, DC is still treating me well. 10 years here and still going. My wife and I moved into our new home in June. So I've been getting settled in the new place all summer, real busy but it's been fun. Just had the 1 year anniversary, can't believe it's been that long. Flather Hall, Kitty's, and 'box' seem like yesterday... Marlon's wedding sounds like fun, too bad it's Homecoming weekend. Would be nice to see everybody. We need to make some plans and all get together soon. The marathon sounds great, good luck, the only running I have time for these days is from Home Depot and back to the house. But I manage to go to the gym a few days a week to maintain my 250 lbs slim figure. I still hang out with Penney and his new little girl, never thought he would be the first to have a baby. Saw Ken and Theo a few weeks ago in DC. Dolan and Sofie are now in San Fran and I talk to him from time to time. Hard to make plans with a big bunch of people. Sometimes you just gotta make the plans with a few and throw it out there for the rest of the group. Let me know if you have any ideas or possible locations. Maybe just plan a weekend somewhere, DC or otherwise and try to get as many people together. Maybe a big party or something like that. We'll hire Vince O'Neill to organize it. An un-official homecoming party. See ya Christopher J. Lukawski Project Manager HITT Contracting Inc. http://www.hitt-gc.com <http://www.hitt-gc.com> 703-444-9453 voice 703-444-7096 fax 703-814-1235 pager -----Original Message----- From: [email protected] [SMTP:[email protected]] Sent: Thursday, September 14, 2000 4:41 PM To: [email protected] Subject: Luke: How are ya? How is the DC area these days? Houston is still fucking hot and Enron is still doing well and the price is still screaming so things are good. In response to your inquiry regarding Homecoming. Megan and I are actually, if you can believe it, going to Marlon and Marissa's wedding the first weekend in October in NJ. Also, OB, Chobu, Mike, Corpina and Danny, Patty D, Sammy, Amy, etc. are going as well - should be interesting. I haven't seen some of those people for quite awhile. I was trying to set something up, but that came up as well as I'm running in this year's NYC Marathon, which is the first Sunday in Novemebr, so trying to arrange something became even harder. I still want to get together with you guys, but trying to figure out when and where. Suggestions are welcomed. Hope all is going well and talk to you soon. Ben (713) 853-7998 Office Ph. ********************************************************************** This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please notify the system manager. **********************************************************************

Filter Product Type + Size + Flavor Type + Flavor + Brand + Nic Level + Collection The Drip Co. Coffee Shop Edition The Drip Company brings us a unique vaping experience with their Coffee Shop Edition (formerly Stars-N-Bucks) line of premium eJuices. Offering flavors of your favorite coffee shop in very unique packaging! Newsletter NOT FOR SALE TO MINORS | CALIFORNIA PROPOSITION 65 - Warning: This product contains nicotine, a chemical known to the state of California to cause birth defects or other reproductive harm. eJuice Direct products are not smoking cessation products and have not been evaluated by the Food and Drug Administration, nor are they intended to treat, prevent or cure any disease or condition. Products may contain Diacetyl and/or Acetyl Propionyl, which have been linked to Bronchiolitis Obliterans. KEEP OUT OF REACH OF CHILDREN AND PETS All product names, trademarks and images are the property of their respective owners, which are in no way associated or affiliated with eJuice Direct. Product names and images are used solely for the purpose of identifying the specific products. Use of these names does not imply any co-operation or endorsement.

/* Copyright 2019 The Kubernetes Authors. Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at http://www.apache.org/licenses/LICENSE-2.0 Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License. */ // +k8s:deepcopy-gen=package // +k8s:protobuf-gen=package // +k8s:openapi-gen=true // +k8s:prerelease-lifecycle-gen=true // +groupName=discovery.k8s.io package v1beta1 // import "k8s.io/api/discovery/v1beta1"

Q: Why does this home experiment show no trace of a magnetic field? The photos below show what I did. Where are the fields? The first photo shows welding cables over paper with iron filings. 55 amps DC and 70 amps DC produced only faint hints of a field. The second photo shows 60 Hz over paper with iron filings (notice I separated the conductors so they wouldn't cancel each other). I then ran a small motor with the ac and no noticeable field was created in the filings. Then I remembered that the magnetic field is perpendicular to the electric field so I ran the wire perpendicular to the paper and still no field. The very last photo shows the hint of a field with the welding cables. In summary here's what I did: I tried DC with a lot of current. I tried AC with a motor as the draw. I tried perpendicular and parallel, and basically got nothing. Besides rolling his eyes at me, what would Maxwell say? A: The magnetic effect of a direct current is very weak. At 1cm from a wire carrying 70 A the magnetic field is about 14 gauss. This is significantly stronger than the Earth's magnetic field (0.5 gauss) but still weaker than a fridge magnet (about 100 gauss). Does a fridge magnet produce a pattern in the iron filings? Getting a pattern with a strong magnet is difficult enough. You will not see any effect in the setup of figs 1, 2 & 4 because the magnetic field is vertical at the paper. It will not push iron filings either towards or away from the wire, nor clump them in lines parallel to the wire. You would only see an effect in fig 3. But all that will happen in this case is that the filings will each turn parallel to the field - you have to look very closely to see the effect. The filings should be longer in one direction, otherwise they will not turn and you will not notice they have turned. If the particles are round they will not turn in the magnetic field. If the filings are close enough together a ring of filings might join up end to end around the wire. Your filings are much too far apart for this too happen. They will not move closer to the wire. The magnetic force may be weaker than the static friction force, keeping the filings from moving at all. You will not see any effect with AC, because the magnetic field changes direction 60 times per second. The iron filings do not have time to react before the direction of the force on them reverses.

Q: How do I show and hide forms in Visual C++? Hey guys, I'm brand new to Visual C++, but not C++. I'm having issues trying to figure out how to show/hide forms. Let's say I have a form Form1 and another form TestForm. In a button click function in Form1.h I have the code: Form1::Hide(); TestForm^ form = gcnew TestForm(); form->Show(); And it works fine. I click the button, and Form1 disappears and TestForm appears. But if I do the same thing in TestForm.h (just changing which forms are set to appear/disappear) I get a plethora of compiler errors in both Form1.h (which used to work) and TestForm.h Form1.cpp c:\users\alex\documents\visual studio 2010\projects\test\test\TestForm.h(86): error C2065: 'Form1' : undeclared identifier c:\users\alex\documents\visual studio 2010\projects\test\test\TestForm.h(86): error C2065: 'form' : undeclared identifier c:\users\alex\documents\visual studio 2010\projects\test\test\TestForm.h(86): error C2061: syntax error : identifier 'Form1' c:\users\alex\documents\visual studio 2010\projects\test\test\TestForm.h(87): error C2065: 'form' : undeclared identifier c:\users\alex\documents\visual studio 2010\projects\test\test\TestForm.h(87): error C2227: left of '->Show' must point to class/struct/union/generic type type is ''unknown-type'' TestForm.cpp c:\users\alex\documents\visual studio 2010\projects\test\test\Form1.h(103): error C2065: 'TestForm' : undeclared identifier c:\users\alex\documents\visual studio 2010\projects\test\test\Form1.h(103): error C2065: 'form' : undeclared identifier c:\users\alex\documents\visual studio 2010\projects\test\test\Form1.h(103): error C2061: syntax error : identifier 'TestForm' c:\users\alex\documents\visual studio 2010\projects\test\test\Form1.h(104): error C2065: 'form' : undeclared identifier c:\users\alex\documents\visual studio 2010\projects\test\test\Form1.h(104): error C2227: left of '->Show' must point to class/struct/union/generic type type is ''unknown-type'' A: The problem is most likely due to the order in which you're including the headers into the .cpp files. In the original case, when you were working in Form1.cpp, "TestForm" was a known type before Form1.h was included. As soon as you switch the header files will your method calls, this isn't the case anymore. I recommend moving your event handlers (the button click functions) to your .cpp files. Your .cpp files can both include both headers, and the compiler will find the form definitions, with their methods, appropriately, no matter what order you include the headers.

Plastic passion How are corporates dealing with the increasing backlash against plastic and the expectations that they play a larger role in cleanup? And how does this affect how these brands are perceivedDelshad Irani&Amit Bapna | ETBrandEquity | Updated: July 18, 2018, 09:40 IST When travelers traipsing about the Indian countryside, especially at higher altitudes, stop for a snack at a dhaba, Maggi noodles is always a top choice. It’s an association so strong that even Nestle’s popular instant noodle brand created a commercial with snowy mountains in the background, weary tourists bundled in sweaters served Maggi by friendly locals. It’s a mouthwatering sight. Until, of course, every strand of Maggi is consumed and the city-slickers return home. Last year, India’s largest trekking community, Indiahikes conducted its first-ever Himalayan Waste Audit, as part of its long running efforts to clean up popular trekking destinations. For the past few years, the organization has been providing all members with an Eco Bag, which can be easily buckled around the waist, so trekkers can pick up litter along the way. The audit headed by avid trekker and environmentalist Ori Gutin was the latest of these efforts and the first that puts manufacturers at the center of the mess. It covered two of the most popular Himalayan treks, Roopkund and Hampta Pass, which are 765 kilometers apart, and in two different states. The results weren’t shocking if you consider our love for Maggi in the mountains. 13 companies, including Parle, Mondelez, Perfetti, Britannia, Mars, PepsiCo and Haldirams, among others, accounted for 77.5% of the total waste collected. But 20.2% of all waste collected from both locations was from Nestle, “meaning that nearly 1 out of every 5 pieces of identifiable trash in the Himalayas is from Nestle”, as per the report. At No 2 is Parle, contributing 10.8% of waste. The report concludes: “Companies have no control over how consumers use their products. It is not the fault of Nestle, Parle, Mondelez International (makers of Cadbury) or any other company that their products are ending up as litter in the most beautiful mountain range in the world, slowly killing wildlife and vegetation, damaging waterways, and polluting the atmosphere when burned. However, that does not mean they cannot do something about it.” So what can they do? The final report lists some measures: Develop waste management systems in rural Himalayan villages. Facilitate programs to educate locals not to litter. Hire mule drivers to collect waste from shops out in the middle of the mountain treks. Hire people to remove the mass amounts of waste already strewn about the Himalayas. However, these companies, despite immense resources at their disposal, haven’t exactly stepped up to do any of the above, the report concludes. When Brand Equity spoke to Lakshmi Selvakumaran, who heads the Green Trails Initiative at IndiaHikes, she told us one of the easier on-ground solutions is to encourage hungry trekkers and travellers to go for, say, egg bhurjee over Maggi. Eventually, one supposes, dhaba owners will have no reason to keep Maggi on the menu. As people across the state of Maharashtra scrambled to secure and protect their cloth bags following the plastic bag ban, another notification by the state government could bring major FMCG companies under the Extended Producers Responsibility (EPR) model. In short, companies have to clean up after their consumers. This is what Nestle said in an email response to a query sent by BE: “Nestlé India shares the ambition that no plastic waste should end up discarded in the environment and believe that with the right approach it can be collected or recycled without a detrimental impact.” One of the approaches includes street plays. Nestle is collaborating with industry bodies and organisations like Indian Pollution Control Association (IPCA), NEPRA and Saahas Waste Management to conduct education and awareness programmes, workshops and street plays for waste pickers, waste dealers, traders and aggregator communities. The country’s biggest FMCG company HUL had this to say; “We completely support the government initiative of significantly improving the state of plastic waste management in the country… At HUL, we have reduced one-third of our plastic packaging since 2010. Further, we are committed to using packaging which will be reusable, recyclable or compostable by 2025.” ITC’s solid waste management initiatives today extend to 10 states across the country, covering cities, towns, villages and even temples. Areas around religious sites often bear the brunt of environmental abuse at the hands of humans and religious events/festivals are known to leave in their wake a severely damaged ecosystem. For now though, optimizing packaging design and packaging weight reduction are the key areas of focus for manufacturers. Says a Mondelez India spokesperson, “design simplification has been undertaken over the years to make our packaging single layer foil from three layer packaging.” PepsiCo India is also reducing the quantity of plastic . According to a company spokesperson, it has resized packaging for snacks master brands Lay’s and Kurkure. Pepsi is working with partners to develop sustainable, environmentally friendly packaging solutions and will pilot its first-ever 100% compostable, plant-based packaging for Lay's and Kurkure in the fourth quarter of 2018. Savvy marketers have long used environmental planks to bolster their brands’ green credentials. However, national and state level legislations are quickly putting plastic waste management and reduction out of their reach. Says Ravi Agrawal, director, Toxics Link, “Many companies bring out a sustainability report. It’s something they can do voluntarily. But if it becomes a regulatory requirement then it’s no longer about doing good. It goes out of the ambit of ‘oh, I’m doing so many good things so why should I do this?’ It has a much more serious implication because it’s not voluntary. You need to set up the systems, redesign, rethink etc. It hits at the very heart of the product you sell.” People like Selvakumaran think it’s about time companies clean up after their consumers and she’s keen to see what effect this will have on-ground. A 2017 post by Gutin asked in the headline ‘Maggi or the Mountains, What Will You Choose?’ The answer is not a 2-minute one. However, in the meantime, might we suggest an Eco Bag for every employee this Diwali?

NVIDIA Corporation Corporate Office Founded in 1993, NVIDIA Corporation is a global technology company which is best known for manufacturing GPUs or graphics processing units. Aside from GPUs, however, NVIDIA also offers several other products including PC platform chipsets, wireless communications processors, and digital media player software. Some of the most notable product families in NVIDIA\'s product portfolio include GeForce, Quadro, Tegra. Tesla, and nForce.

#ifndef __ARCH_ARM_MACH_OMAP2_CM_REGBITS_34XX_H #define __ARCH_ARM_MACH_OMAP2_CM_REGBITS_34XX_H /* * OMAP3430 Clock Management register bits * * Copyright (C) 2007-2008 Texas Instruments, Inc. * Copyright (C) 2007-2008 Nokia Corporation * * Written by Paul Walmsley * * This program is free software; you can redistribute it and/or modify * it under the terms of the GNU General Public License version 2 as * published by the Free Software Foundation. */ /* Bits shared between registers */ /* CM_FCLKEN1_CORE and CM_ICLKEN1_CORE shared bits */ #define OMAP3430ES2_EN_MMC3_MASK (1 << 30) #define OMAP3430ES2_EN_MMC3_SHIFT 30 #define OMAP3430_EN_MSPRO_MASK (1 << 23) #define OMAP3430_EN_MSPRO_SHIFT 23 #define OMAP3430_EN_HDQ_MASK (1 << 22) #define OMAP3430_EN_HDQ_SHIFT 22 #define OMAP3430ES1_EN_FSHOSTUSB_MASK (1 << 5) #define OMAP3430ES1_EN_FSHOSTUSB_SHIFT 5 #define OMAP3430ES1_EN_D2D_MASK (1 << 3) #define OMAP3430ES1_EN_D2D_SHIFT 3 #define OMAP3430_EN_SSI_MASK (1 << 0) #define OMAP3430_EN_SSI_SHIFT 0 /* CM_FCLKEN3_CORE and CM_ICLKEN3_CORE shared bits */ #define OMAP3430ES2_EN_USBTLL_SHIFT 2 #define OMAP3430ES2_EN_USBTLL_MASK (1 << 2) /* CM_FCLKEN_WKUP and CM_ICLKEN_WKUP shared bits */ #define OMAP3430_EN_WDT2_MASK (1 << 5) #define OMAP3430_EN_WDT2_SHIFT 5 /* CM_ICLKEN_CAM, CM_FCLKEN_CAM shared bits */ #define OMAP3430_EN_CAM_MASK (1 << 0) #define OMAP3430_EN_CAM_SHIFT 0 /* CM_FCLKEN_PER, CM_ICLKEN_PER shared bits */ #define OMAP3430_EN_WDT3_MASK (1 << 12) #define OMAP3430_EN_WDT3_SHIFT 12 /* CM_CLKSEL2_EMU, CM_CLKSEL3_EMU shared bits */ #define OMAP3430_OVERRIDE_ENABLE_MASK (1 << 19) /* Bits specific to each register */ /* CM_FCLKEN_IVA2 */ #define OMAP3430_CM_FCLKEN_IVA2_EN_IVA2_MASK (1 << 0) #define OMAP3430_CM_FCLKEN_IVA2_EN_IVA2_SHIFT 0 /* CM_CLKEN_PLL_IVA2 */ #define OMAP3430_IVA2_DPLL_RAMPTIME_SHIFT 8 #define OMAP3430_IVA2_DPLL_RAMPTIME_MASK (0x3 << 8) #define OMAP3430_IVA2_DPLL_FREQSEL_SHIFT 4 #define OMAP3430_IVA2_DPLL_FREQSEL_MASK (0xf << 4) #define OMAP3430_EN_IVA2_DPLL_DRIFTGUARD_SHIFT 3 #define OMAP3430_EN_IVA2_DPLL_DRIFTGUARD_MASK (1 << 3) #define OMAP3430_EN_IVA2_DPLL_SHIFT 0 #define OMAP3430_EN_IVA2_DPLL_MASK (0x7 << 0) /* CM_IDLEST_IVA2 */ #define OMAP3430_ST_IVA2_MASK (1 << 0) /* CM_IDLEST_PLL_IVA2 */ #define OMAP3430_ST_IVA2_CLK_SHIFT 0 #define OMAP3430_ST_IVA2_CLK_MASK (1 << 0) /* CM_AUTOIDLE_PLL_IVA2 */ #define OMAP3430_AUTO_IVA2_DPLL_SHIFT 0 #define OMAP3430_AUTO_IVA2_DPLL_MASK (0x7 << 0) /* CM_CLKSEL1_PLL_IVA2 */ #define OMAP3430_IVA2_CLK_SRC_SHIFT 19 #define OMAP3430_IVA2_CLK_SRC_MASK (0x3 << 19) #define OMAP3430_IVA2_DPLL_MULT_SHIFT 8 #define OMAP3430_IVA2_DPLL_MULT_MASK (0x7ff << 8) #define OMAP3430_IVA2_DPLL_DIV_SHIFT 0 #define OMAP3430_IVA2_DPLL_DIV_MASK (0x7f << 0) /* CM_CLKSEL2_PLL_IVA2 */ #define OMAP3430_IVA2_DPLL_CLKOUT_DIV_SHIFT 0 #define OMAP3430_IVA2_DPLL_CLKOUT_DIV_MASK (0x1f << 0) /* CM_CLKSTCTRL_IVA2 */ #define OMAP3430_CLKTRCTRL_IVA2_SHIFT 0 #define OMAP3430_CLKTRCTRL_IVA2_MASK (0x3 << 0) /* CM_CLKSTST_IVA2 */ #define OMAP3430_CLKACTIVITY_IVA2_SHIFT 0 #define OMAP3430_CLKACTIVITY_IVA2_MASK (1 << 0) /* CM_REVISION specific bits */ /* CM_SYSCONFIG specific bits */ /* CM_CLKEN_PLL_MPU */ #define OMAP3430_MPU_DPLL_RAMPTIME_SHIFT 8 #define OMAP3430_MPU_DPLL_RAMPTIME_MASK (0x3 << 8) #define OMAP3430_MPU_DPLL_FREQSEL_SHIFT 4 #define OMAP3430_MPU_DPLL_FREQSEL_MASK (0xf << 4) #define OMAP3430_EN_MPU_DPLL_DRIFTGUARD_SHIFT 3 #define OMAP3430_EN_MPU_DPLL_DRIFTGUARD_MASK (1 << 3) #define OMAP3430_EN_MPU_DPLL_SHIFT 0 #define OMAP3430_EN_MPU_DPLL_MASK (0x7 << 0) /* CM_IDLEST_MPU */ #define OMAP3430_ST_MPU_MASK (1 << 0) /* CM_IDLEST_PLL_MPU */ #define OMAP3430_ST_MPU_CLK_SHIFT 0 #define OMAP3430_ST_MPU_CLK_MASK (1 << 0) /* CM_AUTOIDLE_PLL_MPU */ #define OMAP3430_AUTO_MPU_DPLL_SHIFT 0 #define OMAP3430_AUTO_MPU_DPLL_MASK (0x7 << 0) /* CM_CLKSEL1_PLL_MPU */ #define OMAP3430_MPU_CLK_SRC_SHIFT 19 #define OMAP3430_MPU_CLK_SRC_MASK (0x3 << 19) #define OMAP3430_MPU_DPLL_MULT_SHIFT 8 #define OMAP3430_MPU_DPLL_MULT_MASK (0x7ff << 8) #define OMAP3430_MPU_DPLL_DIV_SHIFT 0 #define OMAP3430_MPU_DPLL_DIV_MASK (0x7f << 0) /* CM_CLKSEL2_PLL_MPU */ #define OMAP3430_MPU_DPLL_CLKOUT_DIV_SHIFT 0 #define OMAP3430_MPU_DPLL_CLKOUT_DIV_MASK (0x1f << 0) /* CM_CLKSTCTRL_MPU */ #define OMAP3430_CLKTRCTRL_MPU_SHIFT 0 #define OMAP3430_CLKTRCTRL_MPU_MASK (0x3 << 0) /* CM_CLKSTST_MPU */ #define OMAP3430_CLKACTIVITY_MPU_SHIFT 0 #define OMAP3430_CLKACTIVITY_MPU_MASK (1 << 0) /* CM_FCLKEN1_CORE specific bits */ #define OMAP3430_EN_MODEM_MASK (1 << 31) #define OMAP3430_EN_MODEM_SHIFT 31 /* CM_ICLKEN1_CORE specific bits */ #define OMAP3430_EN_ICR_MASK (1 << 29) #define OMAP3430_EN_ICR_SHIFT 29 #define OMAP3430_EN_AES2_MASK (1 << 28) #define OMAP3430_EN_AES2_SHIFT 28 #define OMAP3430_EN_SHA12_MASK (1 << 27) #define OMAP3430_EN_SHA12_SHIFT 27 #define OMAP3430_EN_DES2_MASK (1 << 26) #define OMAP3430_EN_DES2_SHIFT 26 #define OMAP3430ES1_EN_FAC_MASK (1 << 8) #define OMAP3430ES1_EN_FAC_SHIFT 8 #define OMAP3430_EN_MAILBOXES_MASK (1 << 7) #define OMAP3430_EN_MAILBOXES_SHIFT 7 #define OMAP3430_EN_OMAPCTRL_MASK (1 << 6) #define OMAP3430_EN_OMAPCTRL_SHIFT 6 #define OMAP3430_EN_SAD2D_MASK (1 << 3) #define OMAP3430_EN_SAD2D_SHIFT 3 #define OMAP3430_EN_SDRC_MASK (1 << 1) #define OMAP3430_EN_SDRC_SHIFT 1 /* AM35XX specific CM_ICLKEN1_CORE bits */ #define AM35XX_EN_IPSS_MASK (1 << 4) #define AM35XX_EN_IPSS_SHIFT 4 #define AM35XX_EN_UART4_MASK (1 << 23) #define AM35XX_EN_UART4_SHIFT 23 /* CM_ICLKEN2_CORE */ #define OMAP3430_EN_PKA_MASK (1 << 4) #define OMAP3430_EN_PKA_SHIFT 4 #define OMAP3430_EN_AES1_MASK (1 << 3) #define OMAP3430_EN_AES1_SHIFT 3 #define OMAP3430_EN_RNG_MASK (1 << 2) #define OMAP3430_EN_RNG_SHIFT 2 #define OMAP3430_EN_SHA11_MASK (1 << 1) #define OMAP3430_EN_SHA11_SHIFT 1 #define OMAP3430_EN_DES1_MASK (1 << 0) #define OMAP3430_EN_DES1_SHIFT 0 /* CM_ICLKEN3_CORE */ #define OMAP3430_EN_MAD2D_SHIFT 3 #define OMAP3430_EN_MAD2D_MASK (1 << 3) /* CM_FCLKEN3_CORE specific bits */ #define OMAP3430ES2_EN_TS_SHIFT 1 #define OMAP3430ES2_EN_TS_MASK (1 << 1) #define OMAP3430ES2_EN_CPEFUSE_SHIFT 0 #define OMAP3430ES2_EN_CPEFUSE_MASK (1 << 0) /* CM_IDLEST1_CORE specific bits */ #define OMAP3430ES2_ST_MMC3_SHIFT 30 #define OMAP3430ES2_ST_MMC3_MASK (1 << 30) #define OMAP3430_ST_ICR_SHIFT 29 #define OMAP3430_ST_ICR_MASK (1 << 29) #define OMAP3430_ST_AES2_SHIFT 28 #define OMAP3430_ST_AES2_MASK (1 << 28) #define OMAP3430_ST_SHA12_SHIFT 27 #define OMAP3430_ST_SHA12_MASK (1 << 27) #define OMAP3430_ST_DES2_SHIFT 26 #define OMAP3430_ST_DES2_MASK (1 << 26) #define OMAP3430_ST_MSPRO_SHIFT 23 #define OMAP3430_ST_MSPRO_MASK (1 << 23) #define OMAP3430_ST_HDQ_SHIFT 22 #define OMAP3430_ST_HDQ_MASK (1 << 22) #define OMAP3430ES1_ST_FAC_SHIFT 8 #define OMAP3430ES1_ST_FAC_MASK (1 << 8) #define OMAP3430ES2_ST_SSI_IDLE_SHIFT 8 #define OMAP3430ES2_ST_SSI_IDLE_MASK (1 << 8) #define OMAP3430_ST_MAILBOXES_SHIFT 7 #define OMAP3430_ST_MAILBOXES_MASK (1 << 7) #define OMAP3430_ST_OMAPCTRL_SHIFT 6 #define OMAP3430_ST_OMAPCTRL_MASK (1 << 6) #define OMAP3430_ST_SDMA_SHIFT 2 #define OMAP3430_ST_SDMA_MASK (1 << 2) #define OMAP3430_ST_SDRC_SHIFT 1 #define OMAP3430_ST_SDRC_MASK (1 << 1) #define OMAP3430_ST_SSI_STDBY_SHIFT 0 #define OMAP3430_ST_SSI_STDBY_MASK (1 << 0) /* AM35xx specific CM_IDLEST1_CORE bits */ #define AM35XX_ST_IPSS_SHIFT 5 #define AM35XX_ST_IPSS_MASK (1 << 5) /* CM_IDLEST2_CORE */ #define OMAP3430_ST_PKA_SHIFT 4 #define OMAP3430_ST_PKA_MASK (1 << 4) #define OMAP3430_ST_AES1_SHIFT 3 #define OMAP3430_ST_AES1_MASK (1 << 3) #define OMAP3430_ST_RNG_SHIFT 2 #define OMAP3430_ST_RNG_MASK (1 << 2) #define OMAP3430_ST_SHA11_SHIFT 1 #define OMAP3430_ST_SHA11_MASK (1 << 1) #define OMAP3430_ST_DES1_SHIFT 0 #define OMAP3430_ST_DES1_MASK (1 << 0) /* CM_IDLEST3_CORE */ #define OMAP3430ES2_ST_USBTLL_SHIFT 2 #define OMAP3430ES2_ST_USBTLL_MASK (1 << 2) #define OMAP3430ES2_ST_CPEFUSE_SHIFT 0 #define OMAP3430ES2_ST_CPEFUSE_MASK (1 << 0) /* CM_AUTOIDLE1_CORE */ #define OMAP3430_AUTO_MODEM_MASK (1 << 31) #define OMAP3430_AUTO_MODEM_SHIFT 31 #define OMAP3430ES2_AUTO_MMC3_MASK (1 << 30) #define OMAP3430ES2_AUTO_MMC3_SHIFT 30 #define OMAP3430ES2_AUTO_ICR_MASK (1 << 29) #define OMAP3430ES2_AUTO_ICR_SHIFT 29 #define OMAP3430_AUTO_AES2_MASK (1 << 28) #define OMAP3430_AUTO_AES2_SHIFT 28 #define OMAP3430_AUTO_SHA12_MASK (1 << 27) #define OMAP3430_AUTO_SHA12_SHIFT 27 #define OMAP3430_AUTO_DES2_MASK (1 << 26) #define OMAP3430_AUTO_DES2_SHIFT 26 #define OMAP3430_AUTO_MMC2_MASK (1 << 25) #define OMAP3430_AUTO_MMC2_SHIFT 25 #define OMAP3430_AUTO_MMC1_MASK (1 << 24) #define OMAP3430_AUTO_MMC1_SHIFT 24 #define OMAP3430_AUTO_MSPRO_MASK (1 << 23) #define OMAP3430_AUTO_MSPRO_SHIFT 23 #define OMAP3430_AUTO_HDQ_MASK (1 << 22) #define OMAP3430_AUTO_HDQ_SHIFT 22 #define OMAP3430_AUTO_MCSPI4_MASK (1 << 21) #define OMAP3430_AUTO_MCSPI4_SHIFT 21 #define OMAP3430_AUTO_MCSPI3_MASK (1 << 20) #define OMAP3430_AUTO_MCSPI3_SHIFT 20 #define OMAP3430_AUTO_MCSPI2_MASK (1 << 19) #define OMAP3430_AUTO_MCSPI2_SHIFT 19 #define OMAP3430_AUTO_MCSPI1_MASK (1 << 18) #define OMAP3430_AUTO_MCSPI1_SHIFT 18 #define OMAP3430_AUTO_I2C3_MASK (1 << 17) #define OMAP3430_AUTO_I2C3_SHIFT 17 #define OMAP3430_AUTO_I2C2_MASK (1 << 16) #define OMAP3430_AUTO_I2C2_SHIFT 16 #define OMAP3430_AUTO_I2C1_MASK (1 << 15) #define OMAP3430_AUTO_I2C1_SHIFT 15 #define OMAP3430_AUTO_UART2_MASK (1 << 14) #define OMAP3430_AUTO_UART2_SHIFT 14 #define OMAP3430_AUTO_UART1_MASK (1 << 13) #define OMAP3430_AUTO_UART1_SHIFT 13 #define OMAP3430_AUTO_GPT11_MASK (1 << 12) #define OMAP3430_AUTO_GPT11_SHIFT 12 #define OMAP3430_AUTO_GPT10_MASK (1 << 11) #define OMAP3430_AUTO_GPT10_SHIFT 11 #define OMAP3430_AUTO_MCBSP5_MASK (1 << 10) #define OMAP3430_AUTO_MCBSP5_SHIFT 10 #define OMAP3430_AUTO_MCBSP1_MASK (1 << 9) #define OMAP3430_AUTO_MCBSP1_SHIFT 9 #define OMAP3430ES1_AUTO_FAC_MASK (1 << 8) #define OMAP3430ES1_AUTO_FAC_SHIFT 8 #define OMAP3430_AUTO_MAILBOXES_MASK (1 << 7) #define OMAP3430_AUTO_MAILBOXES_SHIFT 7 #define OMAP3430_AUTO_OMAPCTRL_MASK (1 << 6) #define OMAP3430_AUTO_OMAPCTRL_SHIFT 6 #define OMAP3430ES1_AUTO_FSHOSTUSB_MASK (1 << 5) #define OMAP3430ES1_AUTO_FSHOSTUSB_SHIFT 5 #define OMAP3430_AUTO_HSOTGUSB_MASK (1 << 4) #define OMAP3430_AUTO_HSOTGUSB_SHIFT 4 #define OMAP3430ES1_AUTO_D2D_MASK (1 << 3) #define OMAP3430ES1_AUTO_D2D_SHIFT 3 #define OMAP3430_AUTO_SAD2D_MASK (1 << 3) #define OMAP3430_AUTO_SAD2D_SHIFT 3 #define OMAP3430_AUTO_SSI_MASK (1 << 0) #define OMAP3430_AUTO_SSI_SHIFT 0 /* CM_AUTOIDLE2_CORE */ #define OMAP3430_AUTO_PKA_MASK (1 << 4) #define OMAP3430_AUTO_PKA_SHIFT 4 #define OMAP3430_AUTO_AES1_MASK (1 << 3) #define OMAP3430_AUTO_AES1_SHIFT 3 #define OMAP3430_AUTO_RNG_MASK (1 << 2) #define OMAP3430_AUTO_RNG_SHIFT 2 #define OMAP3430_AUTO_SHA11_MASK (1 << 1) #define OMAP3430_AUTO_SHA11_SHIFT 1 #define OMAP3430_AUTO_DES1_MASK (1 << 0) #define OMAP3430_AUTO_DES1_SHIFT 0 /* CM_AUTOIDLE3_CORE */ #define OMAP3430ES2_AUTO_USBHOST (1 << 0) #define OMAP3430ES2_AUTO_USBHOST_SHIFT 0 #define OMAP3430ES2_AUTO_USBTLL (1 << 2) #define OMAP3430ES2_AUTO_USBTLL_SHIFT 2 #define OMAP3430ES2_AUTO_USBTLL_MASK (1 << 2) #define OMAP3430_AUTO_MAD2D_SHIFT 3 #define OMAP3430_AUTO_MAD2D_MASK (1 << 3) /* CM_CLKSEL_CORE */ #define OMAP3430_CLKSEL_SSI_SHIFT 8 #define OMAP3430_CLKSEL_SSI_MASK (0xf << 8) #define OMAP3430_CLKSEL_GPT11_MASK (1 << 7) #define OMAP3430_CLKSEL_GPT11_SHIFT 7 #define OMAP3430_CLKSEL_GPT10_MASK (1 << 6) #define OMAP3430_CLKSEL_GPT10_SHIFT 6 #define OMAP3430ES1_CLKSEL_FSHOSTUSB_SHIFT 4 #define OMAP3430ES1_CLKSEL_FSHOSTUSB_MASK (0x3 << 4) #define OMAP3430_CLKSEL_L4_SHIFT 2 #define OMAP3430_CLKSEL_L4_MASK (0x3 << 2) #define OMAP3430_CLKSEL_L3_SHIFT 0 #define OMAP3430_CLKSEL_L3_MASK (0x3 << 0) #define OMAP3630_CLKSEL_96M_SHIFT 12 #define OMAP3630_CLKSEL_96M_MASK (0x3 << 12) /* CM_CLKSTCTRL_CORE */ #define OMAP3430ES1_CLKTRCTRL_D2D_SHIFT 4 #define OMAP3430ES1_CLKTRCTRL_D2D_MASK (0x3 << 4) #define OMAP3430_CLKTRCTRL_L4_SHIFT 2 #define OMAP3430_CLKTRCTRL_L4_MASK (0x3 << 2) #define OMAP3430_CLKTRCTRL_L3_SHIFT 0 #define OMAP3430_CLKTRCTRL_L3_MASK (0x3 << 0) /* CM_CLKSTST_CORE */ #define OMAP3430ES1_CLKACTIVITY_D2D_SHIFT 2 #define OMAP3430ES1_CLKACTIVITY_D2D_MASK (1 << 2) #define OMAP3430_CLKACTIVITY_L4_SHIFT 1 #define OMAP3430_CLKACTIVITY_L4_MASK (1 << 1) #define OMAP3430_CLKACTIVITY_L3_SHIFT 0 #define OMAP3430_CLKACTIVITY_L3_MASK (1 << 0) /* CM_FCLKEN_GFX */ #define OMAP3430ES1_EN_3D_MASK (1 << 2) #define OMAP3430ES1_EN_3D_SHIFT 2 #define OMAP3430ES1_EN_2D_MASK (1 << 1) #define OMAP3430ES1_EN_2D_SHIFT 1 /* CM_ICLKEN_GFX specific bits */ /* CM_IDLEST_GFX specific bits */ /* CM_CLKSEL_GFX specific bits */ /* CM_SLEEPDEP_GFX specific bits */ /* CM_CLKSTCTRL_GFX */ #define OMAP3430ES1_CLKTRCTRL_GFX_SHIFT 0 #define OMAP3430ES1_CLKTRCTRL_GFX_MASK (0x3 << 0) /* CM_CLKSTST_GFX */ #define OMAP3430ES1_CLKACTIVITY_GFX_SHIFT 0 #define OMAP3430ES1_CLKACTIVITY_GFX_MASK (1 << 0) /* CM_FCLKEN_SGX */ #define OMAP3430ES2_CM_FCLKEN_SGX_EN_SGX_SHIFT 1 #define OMAP3430ES2_CM_FCLKEN_SGX_EN_SGX_MASK (1 << 1) /* CM_IDLEST_SGX */ #define OMAP3430ES2_ST_SGX_SHIFT 1 #define OMAP3430ES2_ST_SGX_MASK (1 << 1) /* CM_ICLKEN_SGX */ #define OMAP3430ES2_CM_ICLKEN_SGX_EN_SGX_SHIFT 0 #define OMAP3430ES2_CM_ICLKEN_SGX_EN_SGX_MASK (1 << 0) /* CM_CLKSEL_SGX */ #define OMAP3430ES2_CLKSEL_SGX_SHIFT 0 #define OMAP3430ES2_CLKSEL_SGX_MASK (0x7 << 0) /* CM_CLKSTCTRL_SGX */ #define OMAP3430ES2_CLKTRCTRL_SGX_SHIFT 0 #define OMAP3430ES2_CLKTRCTRL_SGX_MASK (0x3 << 0) /* CM_CLKSTST_SGX */ #define OMAP3430ES2_CLKACTIVITY_SGX_SHIFT 0 #define OMAP3430ES2_CLKACTIVITY_SGX_MASK (1 << 0) /* CM_FCLKEN_WKUP specific bits */ #define OMAP3430ES2_EN_USIMOCP_SHIFT 9 #define OMAP3430ES2_EN_USIMOCP_MASK (1 << 9) /* CM_ICLKEN_WKUP specific bits */ #define OMAP3430_EN_WDT1_MASK (1 << 4) #define OMAP3430_EN_WDT1_SHIFT 4 #define OMAP3430_EN_32KSYNC_MASK (1 << 2) #define OMAP3430_EN_32KSYNC_SHIFT 2 /* CM_IDLEST_WKUP specific bits */ #define OMAP3430ES2_ST_USIMOCP_SHIFT 9 #define OMAP3430ES2_ST_USIMOCP_MASK (1 << 9) #define OMAP3430_ST_WDT2_SHIFT 5 #define OMAP3430_ST_WDT2_MASK (1 << 5) #define OMAP3430_ST_WDT1_SHIFT 4 #define OMAP3430_ST_WDT1_MASK (1 << 4) #define OMAP3430_ST_32KSYNC_SHIFT 2 #define OMAP3430_ST_32KSYNC_MASK (1 << 2) /* CM_AUTOIDLE_WKUP */ #define OMAP3430ES2_AUTO_USIMOCP_MASK (1 << 9) #define OMAP3430ES2_AUTO_USIMOCP_SHIFT 9 #define OMAP3430_AUTO_WDT2_MASK (1 << 5) #define OMAP3430_AUTO_WDT2_SHIFT 5 #define OMAP3430_AUTO_WDT1_MASK (1 << 4) #define OMAP3430_AUTO_WDT1_SHIFT 4 #define OMAP3430_AUTO_GPIO1_MASK (1 << 3) #define OMAP3430_AUTO_GPIO1_SHIFT 3 #define OMAP3430_AUTO_32KSYNC_MASK (1 << 2) #define OMAP3430_AUTO_32KSYNC_SHIFT 2 #define OMAP3430_AUTO_GPT12_MASK (1 << 1) #define OMAP3430_AUTO_GPT12_SHIFT 1 #define OMAP3430_AUTO_GPT1_MASK (1 << 0) #define OMAP3430_AUTO_GPT1_SHIFT 0 /* CM_CLKSEL_WKUP */ #define OMAP3430ES2_CLKSEL_USIMOCP_MASK (0xf << 3) #define OMAP3430_CLKSEL_RM_SHIFT 1 #define OMAP3430_CLKSEL_RM_MASK (0x3 << 1) #define OMAP3430_CLKSEL_GPT1_SHIFT 0 #define OMAP3430_CLKSEL_GPT1_MASK (1 << 0) /* CM_CLKEN_PLL */ #define OMAP3430_PWRDN_EMU_PERIPH_SHIFT 31 #define OMAP3430_PWRDN_CAM_SHIFT 30 #define OMAP3430_PWRDN_DSS1_SHIFT 29 #define OMAP3430_PWRDN_TV_SHIFT 28 #define OMAP3430_PWRDN_96M_SHIFT 27 #define OMAP3430_PERIPH_DPLL_RAMPTIME_SHIFT 24 #define OMAP3430_PERIPH_DPLL_RAMPTIME_MASK (0x3 << 24) #define OMAP3430_PERIPH_DPLL_FREQSEL_SHIFT 20 #define OMAP3430_PERIPH_DPLL_FREQSEL_MASK (0xf << 20) #define OMAP3430_EN_PERIPH_DPLL_DRIFTGUARD_SHIFT 19 #define OMAP3430_EN_PERIPH_DPLL_DRIFTGUARD_MASK (1 << 19) #define OMAP3430_EN_PERIPH_DPLL_SHIFT 16 #define OMAP3430_EN_PERIPH_DPLL_MASK (0x7 << 16) #define OMAP3430_PWRDN_EMU_CORE_SHIFT 12 #define OMAP3430_CORE_DPLL_RAMPTIME_SHIFT 8 #define OMAP3430_CORE_DPLL_RAMPTIME_MASK (0x3 << 8) #define OMAP3430_CORE_DPLL_FREQSEL_SHIFT 4 #define OMAP3430_CORE_DPLL_FREQSEL_MASK (0xf << 4) #define OMAP3430_EN_CORE_DPLL_DRIFTGUARD_SHIFT 3 #define OMAP3430_EN_CORE_DPLL_DRIFTGUARD_MASK (1 << 3) #define OMAP3430_EN_CORE_DPLL_SHIFT 0 #define OMAP3430_EN_CORE_DPLL_MASK (0x7 << 0) /* CM_CLKEN2_PLL */ #define OMAP3430ES2_EN_PERIPH2_DPLL_LPMODE_SHIFT 10 #define OMAP3430ES2_PERIPH2_DPLL_RAMPTIME_MASK (0x3 << 8) #define OMAP3430ES2_PERIPH2_DPLL_FREQSEL_SHIFT 4 #define OMAP3430ES2_PERIPH2_DPLL_FREQSEL_MASK (0xf << 4) #define OMAP3430ES2_EN_PERIPH2_DPLL_DRIFTGUARD_SHIFT 3 #define OMAP3430ES2_EN_PERIPH2_DPLL_SHIFT 0 #define OMAP3430ES2_EN_PERIPH2_DPLL_MASK (0x7 << 0) /* CM_IDLEST_CKGEN */ #define OMAP3430_ST_54M_CLK_MASK (1 << 5) #define OMAP3430_ST_12M_CLK_MASK (1 << 4) #define OMAP3430_ST_48M_CLK_MASK (1 << 3) #define OMAP3430_ST_96M_CLK_MASK (1 << 2) #define OMAP3430_ST_PERIPH_CLK_SHIFT 1 #define OMAP3430_ST_PERIPH_CLK_MASK (1 << 1) #define OMAP3430_ST_CORE_CLK_SHIFT 0 #define OMAP3430_ST_CORE_CLK_MASK (1 << 0) /* CM_IDLEST2_CKGEN */ #define OMAP3430ES2_ST_USIM_CLK_SHIFT 2 #define OMAP3430ES2_ST_USIM_CLK_MASK (1 << 2) #define OMAP3430ES2_ST_120M_CLK_SHIFT 1 #define OMAP3430ES2_ST_120M_CLK_MASK (1 << 1) #define OMAP3430ES2_ST_PERIPH2_CLK_SHIFT 0 #define OMAP3430ES2_ST_PERIPH2_CLK_MASK (1 << 0) /* CM_AUTOIDLE_PLL */ #define OMAP3430_AUTO_PERIPH_DPLL_SHIFT 3 #define OMAP3430_AUTO_PERIPH_DPLL_MASK (0x7 << 3) #define OMAP3430_AUTO_CORE_DPLL_SHIFT 0 #define OMAP3430_AUTO_CORE_DPLL_MASK (0x7 << 0) /* CM_AUTOIDLE2_PLL */ #define OMAP3430ES2_AUTO_PERIPH2_DPLL_SHIFT 0 #define OMAP3430ES2_AUTO_PERIPH2_DPLL_MASK (0x7 << 0) /* CM_CLKSEL1_PLL */ /* Note that OMAP3430_CORE_DPLL_CLKOUT_DIV_MASK was (0x3 << 27) on 3430ES1 */ #define OMAP3430_CORE_DPLL_CLKOUT_DIV_SHIFT 27 #define OMAP3430_CORE_DPLL_CLKOUT_DIV_MASK (0x1f << 27) #define OMAP3430_CORE_DPLL_MULT_SHIFT 16 #define OMAP3430_CORE_DPLL_MULT_MASK (0x7ff << 16) #define OMAP3430_CORE_DPLL_DIV_SHIFT 8 #define OMAP3430_CORE_DPLL_DIV_MASK (0x7f << 8) #define OMAP3430_SOURCE_96M_SHIFT 6 #define OMAP3430_SOURCE_96M_MASK (1 << 6) #define OMAP3430_SOURCE_54M_SHIFT 5 #define OMAP3430_SOURCE_54M_MASK (1 << 5) #define OMAP3430_SOURCE_48M_SHIFT 3 #define OMAP3430_SOURCE_48M_MASK (1 << 3) /* CM_CLKSEL2_PLL */ #define OMAP3430_PERIPH_DPLL_MULT_SHIFT 8 #define OMAP3430_PERIPH_DPLL_MULT_MASK (0x7ff << 8) #define OMAP3630_PERIPH_DPLL_MULT_MASK (0xfff << 8) #define OMAP3430_PERIPH_DPLL_DIV_SHIFT 0 #define OMAP3430_PERIPH_DPLL_DIV_MASK (0x7f << 0) #define OMAP3630_PERIPH_DPLL_DCO_SEL_SHIFT 21 #define OMAP3630_PERIPH_DPLL_DCO_SEL_MASK (0x7 << 21) #define OMAP3630_PERIPH_DPLL_SD_DIV_SHIFT 24 #define OMAP3630_PERIPH_DPLL_SD_DIV_MASK (0xff << 24) /* CM_CLKSEL3_PLL */ #define OMAP3430_DIV_96M_SHIFT 0 #define OMAP3430_DIV_96M_MASK (0x1f << 0) #define OMAP3630_DIV_96M_MASK (0x3f << 0) /* CM_CLKSEL4_PLL */ #define OMAP3430ES2_PERIPH2_DPLL_MULT_SHIFT 8 #define OMAP3430ES2_PERIPH2_DPLL_MULT_MASK (0x7ff << 8) #define OMAP3430ES2_PERIPH2_DPLL_DIV_SHIFT 0 #define OMAP3430ES2_PERIPH2_DPLL_DIV_MASK (0x7f << 0) /* CM_CLKSEL5_PLL */ #define OMAP3430ES2_DIV_120M_SHIFT 0 #define OMAP3430ES2_DIV_120M_MASK (0x1f << 0) /* CM_CLKOUT_CTRL */ #define OMAP3430_CLKOUT2_EN_SHIFT 7 #define OMAP3430_CLKOUT2_EN_MASK (1 << 7) #define OMAP3430_CLKOUT2_DIV_SHIFT 3 #define OMAP3430_CLKOUT2_DIV_MASK (0x7 << 3) #define OMAP3430_CLKOUT2SOURCE_SHIFT 0 #define OMAP3430_CLKOUT2SOURCE_MASK (0x3 << 0) /* CM_FCLKEN_DSS */ #define OMAP3430_EN_TV_MASK (1 << 2) #define OMAP3430_EN_TV_SHIFT 2 #define OMAP3430_EN_DSS2_MASK (1 << 1) #define OMAP3430_EN_DSS2_SHIFT 1 #define OMAP3430_EN_DSS1_MASK (1 << 0) #define OMAP3430_EN_DSS1_SHIFT 0 /* CM_ICLKEN_DSS */ #define OMAP3430_CM_ICLKEN_DSS_EN_DSS_MASK (1 << 0) #define OMAP3430_CM_ICLKEN_DSS_EN_DSS_SHIFT 0 /* CM_IDLEST_DSS */ #define OMAP3430ES2_ST_DSS_IDLE_SHIFT 1 #define OMAP3430ES2_ST_DSS_IDLE_MASK (1 << 1) #define OMAP3430ES2_ST_DSS_STDBY_SHIFT 0 #define OMAP3430ES2_ST_DSS_STDBY_MASK (1 << 0) #define OMAP3430ES1_ST_DSS_SHIFT 0 #define OMAP3430ES1_ST_DSS_MASK (1 << 0) /* CM_AUTOIDLE_DSS */ #define OMAP3430_AUTO_DSS_MASK (1 << 0) #define OMAP3430_AUTO_DSS_SHIFT 0 /* CM_CLKSEL_DSS */ #define OMAP3430_CLKSEL_TV_SHIFT 8 #define OMAP3430_CLKSEL_TV_MASK (0x1f << 8) #define OMAP3630_CLKSEL_TV_MASK (0x3f << 8) #define OMAP3430_CLKSEL_DSS1_SHIFT 0 #define OMAP3430_CLKSEL_DSS1_MASK (0x1f << 0) #define OMAP3630_CLKSEL_DSS1_MASK (0x3f << 0) /* CM_SLEEPDEP_DSS specific bits */ /* CM_CLKSTCTRL_DSS */ #define OMAP3430_CLKTRCTRL_DSS_SHIFT 0 #define OMAP3430_CLKTRCTRL_DSS_MASK (0x3 << 0) /* CM_CLKSTST_DSS */ #define OMAP3430_CLKACTIVITY_DSS_SHIFT 0 #define OMAP3430_CLKACTIVITY_DSS_MASK (1 << 0) /* CM_FCLKEN_CAM specific bits */ #define OMAP3430_EN_CSI2_MASK (1 << 1) #define OMAP3430_EN_CSI2_SHIFT 1 /* CM_ICLKEN_CAM specific bits */ /* CM_IDLEST_CAM */ #define OMAP3430_ST_CAM_MASK (1 << 0) /* CM_AUTOIDLE_CAM */ #define OMAP3430_AUTO_CAM_MASK (1 << 0) #define OMAP3430_AUTO_CAM_SHIFT 0 /* CM_CLKSEL_CAM */ #define OMAP3430_CLKSEL_CAM_SHIFT 0 #define OMAP3430_CLKSEL_CAM_MASK (0x1f << 0) #define OMAP3630_CLKSEL_CAM_MASK (0x3f << 0) /* CM_SLEEPDEP_CAM specific bits */ /* CM_CLKSTCTRL_CAM */ #define OMAP3430_CLKTRCTRL_CAM_SHIFT 0 #define OMAP3430_CLKTRCTRL_CAM_MASK (0x3 << 0) /* CM_CLKSTST_CAM */ #define OMAP3430_CLKACTIVITY_CAM_SHIFT 0 #define OMAP3430_CLKACTIVITY_CAM_MASK (1 << 0) /* CM_FCLKEN_PER specific bits */ /* CM_ICLKEN_PER specific bits */ /* CM_IDLEST_PER */ #define OMAP3430_ST_WDT3_SHIFT 12 #define OMAP3430_ST_WDT3_MASK (1 << 12) #define OMAP3430_ST_MCBSP4_SHIFT 2 #define OMAP3430_ST_MCBSP4_MASK (1 << 2) #define OMAP3430_ST_MCBSP3_SHIFT 1 #define OMAP3430_ST_MCBSP3_MASK (1 << 1) #define OMAP3430_ST_MCBSP2_SHIFT 0 #define OMAP3430_ST_MCBSP2_MASK (1 << 0) /* CM_AUTOIDLE_PER */ #define OMAP3630_AUTO_UART4_MASK (1 << 18) #define OMAP3630_AUTO_UART4_SHIFT 18 #define OMAP3430_AUTO_GPIO6_MASK (1 << 17) #define OMAP3430_AUTO_GPIO6_SHIFT 17 #define OMAP3430_AUTO_GPIO5_MASK (1 << 16) #define OMAP3430_AUTO_GPIO5_SHIFT 16 #define OMAP3430_AUTO_GPIO4_MASK (1 << 15) #define OMAP3430_AUTO_GPIO4_SHIFT 15 #define OMAP3430_AUTO_GPIO3_MASK (1 << 14) #define OMAP3430_AUTO_GPIO3_SHIFT 14 #define OMAP3430_AUTO_GPIO2_MASK (1 << 13) #define OMAP3430_AUTO_GPIO2_SHIFT 13 #define OMAP3430_AUTO_WDT3_MASK (1 << 12) #define OMAP3430_AUTO_WDT3_SHIFT 12 #define OMAP3430_AUTO_UART3_MASK (1 << 11) #define OMAP3430_AUTO_UART3_SHIFT 11 #define OMAP3430_AUTO_GPT9_MASK (1 << 10) #define OMAP3430_AUTO_GPT9_SHIFT 10 #define OMAP3430_AUTO_GPT8_MASK (1 << 9) #define OMAP3430_AUTO_GPT8_SHIFT 9 #define OMAP3430_AUTO_GPT7_MASK (1 << 8) #define OMAP3430_AUTO_GPT7_SHIFT 8 #define OMAP3430_AUTO_GPT6_MASK (1 << 7) #define OMAP3430_AUTO_GPT6_SHIFT 7 #define OMAP3430_AUTO_GPT5_MASK (1 << 6) #define OMAP3430_AUTO_GPT5_SHIFT 6 #define OMAP3430_AUTO_GPT4_MASK (1 << 5) #define OMAP3430_AUTO_GPT4_SHIFT 5 #define OMAP3430_AUTO_GPT3_MASK (1 << 4) #define OMAP3430_AUTO_GPT3_SHIFT 4 #define OMAP3430_AUTO_GPT2_MASK (1 << 3) #define OMAP3430_AUTO_GPT2_SHIFT 3 #define OMAP3430_AUTO_MCBSP4_MASK (1 << 2) #define OMAP3430_AUTO_MCBSP4_SHIFT 2 #define OMAP3430_AUTO_MCBSP3_MASK (1 << 1) #define OMAP3430_AUTO_MCBSP3_SHIFT 1 #define OMAP3430_AUTO_MCBSP2_MASK (1 << 0) #define OMAP3430_AUTO_MCBSP2_SHIFT 0 /* CM_CLKSEL_PER */ #define OMAP3430_CLKSEL_GPT9_MASK (1 << 7) #define OMAP3430_CLKSEL_GPT9_SHIFT 7 #define OMAP3430_CLKSEL_GPT8_MASK (1 << 6) #define OMAP3430_CLKSEL_GPT8_SHIFT 6 #define OMAP3430_CLKSEL_GPT7_MASK (1 << 5) #define OMAP3430_CLKSEL_GPT7_SHIFT 5 #define OMAP3430_CLKSEL_GPT6_MASK (1 << 4) #define OMAP3430_CLKSEL_GPT6_SHIFT 4 #define OMAP3430_CLKSEL_GPT5_MASK (1 << 3) #define OMAP3430_CLKSEL_GPT5_SHIFT 3 #define OMAP3430_CLKSEL_GPT4_MASK (1 << 2) #define OMAP3430_CLKSEL_GPT4_SHIFT 2 #define OMAP3430_CLKSEL_GPT3_MASK (1 << 1) #define OMAP3430_CLKSEL_GPT3_SHIFT 1 #define OMAP3430_CLKSEL_GPT2_MASK (1 << 0) #define OMAP3430_CLKSEL_GPT2_SHIFT 0 /* CM_SLEEPDEP_PER specific bits */ #define OMAP3430_CM_SLEEPDEP_PER_EN_IVA2_MASK (1 << 2) /* CM_CLKSTCTRL_PER */ #define OMAP3430_CLKTRCTRL_PER_SHIFT 0 #define OMAP3430_CLKTRCTRL_PER_MASK (0x3 << 0) /* CM_CLKSTST_PER */ #define OMAP3430_CLKACTIVITY_PER_SHIFT 0 #define OMAP3430_CLKACTIVITY_PER_MASK (1 << 0) /* CM_CLKSEL1_EMU */ #define OMAP3430_DIV_DPLL4_SHIFT 24 #define OMAP3430_DIV_DPLL4_MASK (0x1f << 24) #define OMAP3630_DIV_DPLL4_MASK (0x3f << 24) #define OMAP3430_DIV_DPLL3_SHIFT 16 #define OMAP3430_DIV_DPLL3_MASK (0x1f << 16) #define OMAP3430_CLKSEL_TRACECLK_SHIFT 11 #define OMAP3430_CLKSEL_TRACECLK_MASK (0x7 << 11) #define OMAP3430_CLKSEL_PCLK_SHIFT 8 #define OMAP3430_CLKSEL_PCLK_MASK (0x7 << 8) #define OMAP3430_CLKSEL_PCLKX2_SHIFT 6 #define OMAP3430_CLKSEL_PCLKX2_MASK (0x3 << 6) #define OMAP3430_CLKSEL_ATCLK_SHIFT 4 #define OMAP3430_CLKSEL_ATCLK_MASK (0x3 << 4) #define OMAP3430_TRACE_MUX_CTRL_SHIFT 2 #define OMAP3430_TRACE_MUX_CTRL_MASK (0x3 << 2) #define OMAP3430_MUX_CTRL_SHIFT 0 #define OMAP3430_MUX_CTRL_MASK (0x3 << 0) /* CM_CLKSTCTRL_EMU */ #define OMAP3430_CLKTRCTRL_EMU_SHIFT 0 #define OMAP3430_CLKTRCTRL_EMU_MASK (0x3 << 0) /* CM_CLKSTST_EMU */ #define OMAP3430_CLKACTIVITY_EMU_SHIFT 0 #define OMAP3430_CLKACTIVITY_EMU_MASK (1 << 0) /* CM_CLKSEL2_EMU specific bits */ #define OMAP3430_CORE_DPLL_EMU_MULT_SHIFT 8 #define OMAP3430_CORE_DPLL_EMU_MULT_MASK (0x7ff << 8) #define OMAP3430_CORE_DPLL_EMU_DIV_SHIFT 0 #define OMAP3430_CORE_DPLL_EMU_DIV_MASK (0x7f << 0) /* CM_CLKSEL3_EMU specific bits */ #define OMAP3430_PERIPH_DPLL_EMU_MULT_SHIFT 8 #define OMAP3430_PERIPH_DPLL_EMU_MULT_MASK (0x7ff << 8) #define OMAP3430_PERIPH_DPLL_EMU_DIV_SHIFT 0 #define OMAP3430_PERIPH_DPLL_EMU_DIV_MASK (0x7f << 0) /* CM_POLCTRL */ #define OMAP3430_CLKOUT2_POL_MASK (1 << 0) /* CM_IDLEST_NEON */ #define OMAP3430_ST_NEON_MASK (1 << 0) /* CM_CLKSTCTRL_NEON */ #define OMAP3430_CLKTRCTRL_NEON_SHIFT 0 #define OMAP3430_CLKTRCTRL_NEON_MASK (0x3 << 0) /* CM_FCLKEN_USBHOST */ #define OMAP3430ES2_EN_USBHOST2_SHIFT 1 #define OMAP3430ES2_EN_USBHOST2_MASK (1 << 1) #define OMAP3430ES2_EN_USBHOST1_SHIFT 0 #define OMAP3430ES2_EN_USBHOST1_MASK (1 << 0) /* CM_ICLKEN_USBHOST */ #define OMAP3430ES2_EN_USBHOST_SHIFT 0 #define OMAP3430ES2_EN_USBHOST_MASK (1 << 0) /* CM_IDLEST_USBHOST */ #define OMAP3430ES2_ST_USBHOST_IDLE_SHIFT 1 #define OMAP3430ES2_ST_USBHOST_IDLE_MASK (1 << 1) #define OMAP3430ES2_ST_USBHOST_STDBY_SHIFT 0 #define OMAP3430ES2_ST_USBHOST_STDBY_MASK (1 << 0) /* CM_AUTOIDLE_USBHOST */ #define OMAP3430ES2_AUTO_USBHOST_SHIFT 0 #define OMAP3430ES2_AUTO_USBHOST_MASK (1 << 0) /* CM_SLEEPDEP_USBHOST */ #define OMAP3430ES2_EN_MPU_SHIFT 1 #define OMAP3430ES2_EN_MPU_MASK (1 << 1) #define OMAP3430ES2_EN_IVA2_SHIFT 2 #define OMAP3430ES2_EN_IVA2_MASK (1 << 2) /* CM_CLKSTCTRL_USBHOST */ #define OMAP3430ES2_CLKTRCTRL_USBHOST_SHIFT 0 #define OMAP3430ES2_CLKTRCTRL_USBHOST_MASK (3 << 0) /* CM_CLKSTST_USBHOST */ #define OMAP3430ES2_CLKACTIVITY_USBHOST_SHIFT 0 #define OMAP3430ES2_CLKACTIVITY_USBHOST_MASK (1 << 0) /* * */ /* OMAP3XXX CM_CLKSTCTRL_*.CLKTRCTRL_* register bit values */ #define OMAP34XX_CLKSTCTRL_DISABLE_AUTO 0x0 #define OMAP34XX_CLKSTCTRL_FORCE_SLEEP 0x1 #define OMAP34XX_CLKSTCTRL_FORCE_WAKEUP 0x2 #define OMAP34XX_CLKSTCTRL_ENABLE_AUTO 0x3 #endif

--- abstract: 'Let $(X,\mu)$ be a probability space, $G$ a countable amenable group and $(F_n)_n$ a left Følner sequence in $G$. This paper analyzes the non-conventional ergodic averages $$\frac{1}{|F_n|}\sum_{g \in F_n}\prod_{i=1}^d (f_i\circ T_1^g\cdots T_i^g)$$ associated to a commuting tuple of $\mu$-preserving actions $T_1$, …, $T_d:G{\curvearrowright}X$ and $f_1$, …, $f_d \in L^\infty(\mu)$. We prove that these averages always converge in $\|\cdot\|_2$, and that they witness a multiple recurrence phenomenon when $f_1 = \ldots = f_d = 1_A$ for a non-negligible set $A\subseteq X$. This proves a conjecture of Bergelson, McCutcheon and Zhang. The proof relies on an adaptation from earlier works of the machinery of sated extensions.' author: - | <span style="font-variant:small-caps;">Tim Austin</span>[^1]\ \ \ \ bibliography: - 'bibfile.bib' --- Introduction ============ Let $(X,\mu)$ be a probability space, $G$ a countable amenable group, and $T_1$, …, $T_d:G{\curvearrowright}(X,\mu)$ a tuple of $\mu$-preserving actions of $G$ which commute, meaning that $$i\neq j \quad \Longrightarrow \quad T_i^gT_j^h = T_j^hT_i^g \quad \forall g,h \in G.$$ Also, let $(F_n)_n$ be a left Følner sequence of subsets of $G$; this will be fixed for the rest of the paper. In this context, Bergelson, McCutcheon and Zhang have proposed in [@BerMcCZha97] the study of the non-conventional ergodic averages $$\begin{aligned} \label{eq:Lambda} {\Lambda}_n(f_1,\ldots,f_d) := \frac{1}{|F_n|}\sum_{g \in F_n}\prod_{i=1}^d(f_i\circ T_1^g\cdots T_i^g)\end{aligned}$$ for functions $f_1,\ldots,f_d \in L^\infty(\mu)$. These are an analog for commuting $G$-actions of the non-conventional averages for a commuting tuple of transformations, as introduced by Furstenberg and Katznelson [@FurKat78] for their proof of the multi-dimensional generalization of Szemerédi’s Theorem. Other analogs are possible, but the averages above seem to show the most promise for building a theory: this is discussed in [@BerMcCZha97] and, for topological dynamics, in [@BerHin92], where some relevant counterexamples are presented. The main results of [@BerMcCZha97] are that these averages converge, and that one has an associated multiple recurrence phenomenon, when $d=2$. The first of these conclusions can be extended to arbitrary $d$ along the lines of Walsh’s recent proof of convergence for polynomial nilpotent non-conventional averages ([@Walsh12]). **Theorem A.***In the setting above, the functional averages ${\Lambda}_n(f_1,\ldots,f_d)$ converge in the norm of $L^2(\mu)$ for all $f_1$, …, $f_d \in L^\infty(\mu)$.* Zorin-Kranich has made the necessary extensions to Walsh’s argument in [@Zor13]. However, that proof gives essentially no information about the limiting function, and in particular does not seem to enable a proof of multiple recurrence. The present paper gives both a new proof of Theorem A, and a proof of the following. **Theorem B.***If $\mu(A) > 0$, then $$\begin{gathered} \lim_{n{\longrightarrow}\infty}\int_X {\Lambda}_n(1_A,\ldots,1_A)\,{\mathrm{d}}\mu\\ = \lim_{n{\longrightarrow}\infty}\frac{1}{|F_n|}\sum_{g \in F_n}\mu\big(T_1^{g^{-1}}A\cap \cdots \cap (T_1^{g^{-1}}\cdots T_d^{g^{-1}})A\big) > 0.\end{gathered}$$ In particular, the set $$\big\{g \in G\,\big|\ \mu\big(T_1^{g^{-1}}A\cap \cdots \cap (T_1^{g^{-1}}\cdots T_d^{g^{-1}})A\big) > 0\big\}$$ has positive upper Banach density relative to $(F_n)_{n\geq 1}$.* As in the classical case of [@FurKat78], this implies the following Szemerédi-type result for amenable groups. If $E \subseteq G^d$ has positive upper Banach density relative to $(F_n^d)_{n\geq 1}$, then the set $$\begin{gathered} \big\{g \in G\,\big|\ \exists (x_1,\ldots,x_d) \in G^d\\ \hbox{s.t.}\ \{(g^{-1}x_1,x_2,\ldots,x_d),\ldots,(g^{-1}x_1,\ldots,g^{-1}x_d)\} \subseteq E\big\}\end{gathered}$$ has positive upper Banach density relative to $(F_n)_{n\geq 1}$. This deduction is quite standard, and can be found in [@BerMcCZha97]. Our proofs of Theorems A and B are descended from some work for commuting tuples of transformations: the proof of non-conventional-average convergence in [@Aus--nonconv], and that of multiple recurrence in [@Aus--newmultiSzem]. Both of those papers offered alternatives to earlier proofs, using new machinery for extending an initially-given probability-preserving action to another action under which the averages behave more simply. The present paper will adapt to commuting tuples of $G$-actions the notion of a ‘sated extension’, which forms the heart of the streamlined presentation of that machinery in [@Aus--thesis]. Further discussion of this method may be found in that reference. The generalization of the notion of satedness is nontrivial, but fairly straightforward: see Section \[sec:func\] below. However, more serious difficulties appear in how it is applied. Heuristically, if a given system satisfies a satedness assumption, then, in any extension of that system, this constrains how some canonical ${\sigma}$-subalgebra ‘sits’ relative to the ${\sigma}$-algebra lifted from the original system. An appeal to satedness always relies on constructing a particular extension for which this constraint implies some other desired consequence. The specific constructions of system extensions used in [@Aus--nonconv; @Aus--newmultiSzem; @Aus--thesis] do not generalize to commuting actions of a non-Abelian group $G$. This is because they rely on the commutativity of the diagonal actions $T_i\times \cdots \times T_i$ of $G$ on $X^d$ with the ‘off-diagonal’ action generated by $$T_1^g\times \cdots \times (T_1^g\cdots T^g_d), \quad g \in G.$$ Thus, a key part of this paper is a new method of extending probability-preserving $G^d$-systems. It is based on a version of the Host-Kra self-joinings from [@HosKra05] and [@Hos09]. It also relies on a quite general result about probability-preserving systems, which may be of independent interest: Theorem \[thm:recoverG\] asserts that, given a probability-preserving action of a countable group and an extension of that action restricted to a subgroup, a compatible further extension may be found for the action of the whole group. Developing ideas from [@HosKra05], we find that the asymptotic behaviour of our non-conventional averages can be estimated by certain integrals over these Host-Kra-like extensions (Theorem \[thm:ineq\]). On the other hand, a suitable satedness assumption on a system gives extra information on the structure of those extensions, and combining these facts then implies simplified behaviour for the non-conventional averages for that system. Finally, the existence of sated extensions for all systems (Theorem \[thm:sateds-exist\]) then enables proofs of convergence and multiple recurrence similar to those in [@Aus--nonconv] and [@Aus--newmultiSzem], respectively. An interesting direction for further research is suggested by the work of Bergelson and McCutcheon in [@BerMcC07]. They study multiple recurrence phenomena similar to Theorem B when $d=3$, but without assuming that the group $G$ is amenable. As output, they prove that the set $$\big\{g \in G\,\big|\ \mu\big(T_1^{g^{-1}}A\cap (T_1^{g^{-1}}T_2^{g^{-1}} )A \cap (T_1^{g^{-1}}T_2^{g^{-1}} T_3^{g^{-1}})A\big) > 0\big\}$$ is ‘large’ in a sense adapted to non-amenable groups, in terms of certain special ultrafilters in the Stone-Čech compactification of $G$. In particular, their result implies that this set is syndetic in $G$. Could their methods be combined with those below to extend this result to larger values of $d$? Generalities on actions and extensions ====================================== Preliminaries ------------- If $d \in {\mathbb{N}}$ then $[d] := \{1,2,\ldots,d\}$, and more generally if $a,b \in {\mathbb{Z}}$ with $a \leq b$ then $$(a;b] = [a+1;b] = [a+1;b+1) = (a;b+1):= \{a+1,\ldots,b\}.$$ The power set of $[d]$ will be denoted ${\mathcal{P}}[d]$, and we let $\binom{[d]}{\geq p}:= \{e \in {\mathcal{P}}[d]\,|\ |e|\geq p\}$. Next, if ${\mathcal{A}} \subseteq {\mathcal{P}}[d]$, then it is an **up-set** if $$a,b \in {\mathcal{A}} \quad \Longrightarrow \quad a\cup b \in{\mathcal{A}}.$$ The set $$\langle e \rangle := \{a \subseteq [d]\,|\ a \supseteq e\}$$ is an up-set for every $e \subseteq [d]$, and every up-set is a union of such examples. On the other hand, ${\mathcal{B}} \subseteq {\mathcal{P}}[d]$ is an **antichain** if $$a,b\in {\mathcal{B}} \quad \hbox{and} \quad a \subseteq b \quad \Longrightarrow \quad a = b.$$ Any up-set contains a unique anti-chain of minimal elements. Standard notions from probability theory will be assumed throughout this paper. If $(X,\mu)$ is a probability space with ${\sigma}$-algebra ${\Sigma}$, and if $\Phi,{\Sigma}_1,{\Sigma}_2 \subseteq {\Sigma}$ are ${\sigma}$-subalgebras with $\Phi \subseteq {\Sigma}_1\cap {\Sigma}_2$, then ${\Sigma}_1$ and ${\Sigma}_2$ are **relatively independent** over $\Phi$ under $\mu$ if $$\int_X fg\,{\mathrm{d}}\mu = \int_X{\mathsf{E}}_\mu(f\,|\,\Phi){\mathsf{E}}_\mu(g\,|\,\Phi)\,{\mathrm{d}}\mu$$ whenever $f,g \in L^\infty(\mu)$ are ${\Sigma}_1$- and ${\Sigma}_2$-measurable, respectively. Relatedly, if $(X,\mu)$ is standard Borel, then on $X^2$ we may form the **relative product** measure $\mu\otimes_\Phi \mu$ over $\Phi$ by letting $x\mapsto \mu_x$ be a disintegration of $\mu$ over the ${\sigma}$-subalgebra $\Phi$ and then setting $$\mu\otimes_\Phi\mu := \int_X \mu_x\otimes \mu_x\,\mu({\mathrm{d}}x).$$ If $G$ is a countable group, then a **$G$-space** is a triple $(X,\mu,T)$ consisting of a probability space $(X,\mu)$ and an action $T:G{\curvearrowright}X$ by measurable, $\mu$-preserving transformations. Passing to an isomorphic model if necessary, we will henceforth assume that $(X,\mu)$ is standard Borel. Often, a $G$-space will also be denoted by a boldface letter such as ${\mathbf{X}}$. If ${\mathbf{X}}= (X,\mu,T)$ is a $G$-space, then ${\Sigma}_X$ or ${\Sigma}_{\mathbf{X}}$ will denote its ${\sigma}$-algebra of $\mu$-measurable sets. A **factor** of such a $G$-space is a ${\sigma}$-subalgebra $\Phi \leq {\Sigma}_{\mathbf{X}}$ which is globally $T$-invariant, meaning that $$A \in \Phi \quad \Longleftrightarrow \quad T^g(A) \in \Phi \quad \forall g \in G.$$ Relatedly, a **factor map** from one $G$-space ${\mathbf{X}}= (X,\mu,T)$ to another ${\mathbf{Y}}= (Y,\nu,S)$ is a measurable map $\pi:X{\longrightarrow}Y$ such that $\pi_\ast\mu = \nu$ and $S^g\circ \pi = \pi \circ T^g$ for all $g \in G$, $\mu$-a.e. In this case, $\pi^{-1}({\Sigma}_{\mathbf{Y}})$ is a factor of ${\mathbf{X}}$. Such a factor map is also referred to as a **$G$-extension**, and ${\mathbf{X}}$ may be referred to as an **extension** of ${\mathbf{Y}}$. On the other hand, if ${\mathbf{X}}= (X,\mu,T)$ is a $G$-space and $H \leq G$, then the **$H$-subaction** of ${\mathbf{X}}$, denoted ${\mathbf{X}}^{{\!\upharpoonright}H} = (X,\mu,T^{{\!\upharpoonright}H})$, is the $H$-space with probability space $(X,\mu)$ and action given by the transformations $(T^h)_{h \in H}$. The associated ${\sigma}$-algebra of $H$-almost-invariant sets, $$\{A \in {\Sigma}_X\,|\ \mu(T^h(A)\triangle A) = 0\ \forall h \in H\},$$ will be denoted by either ${\Sigma}_{\mathbf{X}}^H$ or ${\Sigma}_{\mathbf{X}}^{T^{{\!\upharpoonright}H}}$, as seems appropriate. In the sequel, we will often consider a space $(X,\mu)$ endowed with a commuting tuple $T_1$, …, $T_d$ of $G$-actions. Slightly abusively, we shall simply refer to this as a ‘$G^d$-action’ or ‘$G^d$-space’ (leaving the distinguished $G$-subactions to the reader’s understanding), and denote it by $(X,\mu,T_1,\ldots,T_d)$. Also, if $(X,\mu,T_1,\ldots,T_d)$ is a $G^d$-space and $a,b \in [d]$ with $a \leq b$, then we shall frequently use the notation $$T_{[a;b]}^g = T_{(a-1;b]}^g = T_{[a,;b+1)}^g := T_a^gT_{a+1}^g \cdots T_b^g \quad \forall g \in G.$$ Because the actions $T_i$ commute, this defines another $G$-action for each $a,b$. Actions of groups and their subgroups ------------------------------------- Our approach to Theorems A and B is descended from the notions of ‘pleasant’ and ‘isotropized’ extensions. These were introduced in [@Aus--nonconv] and [@Aus--newmultiSzem] respectively, where they were used to give new proofs of the analogs of Theorems A and B for commuting tuples of single transformations. Subsequently, the more general notion of ‘sated’ extensions was introduced in [@Aus--thesis]. It simplifies and clarifies those earlier ideas as special cases. In this paper we shall show how ‘sated’ extensions can be adapted to the non-Abelian setting of Theorems A and B. An important new difficulty is that we will need to consider certain natural ${\sigma}$-subalgebras of a probability-preserving $G$-spaces which need not be factors in case $G$ is not Abelian. This subsection focuses on a key tool for handling this situation, which seems to be of interest in its own right. Given $H \leq G$, it enables one to turn an extension of an $H$-subaction into an extension of a whole $G$-action. Satedness will then be introduced in the next subsection. \[thm:recoverG\] Suppose $H \leq G$ is an inclusion of countable groups, that ${\mathbf{X}}= (X,\mu,T)$ is a $G$-space and that $${\mathbf{Y}}= (Y,\nu,S) \stackrel{{\beta}}{{\longrightarrow}} {\mathbf{X}}^{{\!\upharpoonright}H}$$ is an extension of $H$-spaces. Then there is an extension of $G$-spaces ${\widetilde{{\mathbf{X}}}}\stackrel{\pi}{{\longrightarrow}} {\mathbf{X}}$ which admits a commutative diagram of $H$-spaces $\phantom{i}$ This theorem was proved for Abelian $G$ and $H$ in [@Aus--lindeppleasant1 Subsection 3.2]. The non-Abelian case is fairly similar. We shall construct the new $G$-space ${\widetilde{{\mathbf{X}}}}$ by a kind of ‘relativized’ co-induction of ${\mathbf{Y}}$ over ${\mathbf{X}}$, and then show that it has the necessary properties. The construction of a suitable standard Borel dynamical system $({\widetilde{X}},{\widetilde{T}})$, deferring the construction of the measure, is easy. Let $${\widetilde{X}} := \{(y_g)_g \in Y^G\,|\ y_{gh} = S^{h^{-1}}y_g\ \hbox{and}\ {\beta}(y_g) = T^{g^{-1}}{\beta}(y_e)\ \forall g\in G, h \in H\},$$ and let ${\widetilde{T}}:G{\curvearrowright}{\widetilde{X}}$ be the restriction to ${\widetilde{X}}$ of the left-regular representation: $${\widetilde{T}}^k((y_g)_{g \in G}) = (y_{k^{-1}g})_{g\in G}$$ (it is easily seen that this preserves ${\widetilde{X}} \subseteq Y^G$). Also, let $${\alpha}:{\widetilde{X}}{\longrightarrow}Y: (y_g)_g\mapsto y_e$$ and $$\pi:= {\beta}\circ {\alpha}:{\widetilde{X}}{\longrightarrow}X: (y_g)_g \mapsto {\beta}(y_e).$$ These maps fit into a commutative diagram of the desired shape by construction. It remains to specify a suitable measure ${\widetilde{\mu}}$ on ${\widetilde{X}}$. It will be constructed as a measure on $Y^G$ for which ${\widetilde{\mu}}({\widetilde{X}}) = 1$. Let $X{\longrightarrow}\Pr\,Y:x\mapsto\nu_x$ be a disintegration of $\nu$ over the map ${\beta}:Y{\longrightarrow}X$. Using this, define new probability measures for each $x \in X$ as follows. First, for each $g \in G$, define ${\widetilde{\nu}}_{g,x}$ on $Y^{gH}$ by $${\widetilde{\nu}}_{g,x} := \int_Y \delta_{(S^{h^{-1}}y)_{gh \in gH}}\,\nu_x({\mathrm{d}}y).$$ Now let $C \subseteq G$ be a cross-section for the space $G/H$ of left-cosets, identify $Y^G = \prod_{c \in C}Y^{cH}$, and on this product define $${\widetilde{\nu}}_x := \bigotimes_{c \in C}{\widetilde{\nu}}_{c,T^{c^{-1}}x}.$$ One may easily write down the finite-dimensional marginals of ${\widetilde{\nu}}_x$ directly. If $c_1,\ldots,c_m \in C$, and $h_{i,1}$, …, $h_{i,n_i} \in H$ for each $i \leq m$, and also $A_{i,j} \in {\Sigma}_Y$ for all $i \leq m$ and $j\leq n_i$, then $$\begin{aligned} \label{eq:margs} &&{\widetilde{\nu}}_x\big\{(y_g)_g\,\big|\ y_{c_ih_{i,j}} \in A_{i,j}\ \forall i\leq m,\,j\leq n_i\big\} \nonumber\\ &&= \prod_{i=1}^m{\widetilde{\nu}}_{c_i,T^{c_i^{-1}}x}\big\{(y_{c_ih})_{h \in H}\,\big|\ y_{c_ih_{i,j}} \in A_{i,j}\ \forall j \leq n_i\big\} \nonumber\\ &&= \prod_{i=1}^m\nu_{T^{c_i^{-1}}x}\big(S^{h_{i,1}}(A_{i,1})\cap \cdots \cap S^{h_{i,n_i}}(A_{i,n_i})\big).\end{aligned}$$ The following basic properties of ${\widetilde{\nu}}_x$ are now easily checked: - If $g_1H = g_2H$, say with $g_1 = g_2h_1$, and $x \in X$, then $$\begin{aligned} {\widetilde{\nu}}_{g_1,T^{g_1^{-1}}x} &=& \int_Y \delta_{(S^{h^{-1}}y)_{g_1h \in g_1H}}\ \nu_{T^{g_1^{-1}}x}({\mathrm{d}}y)\\ &=& \int_Y \delta_{(S^{h^{-1}}y)_{g_2h_1h \in g_2H}}\ \nu_{T^{h_1^{-1}}T^{g_2^{-1}}x}({\mathrm{d}}y)\\ &=& \int_Y \delta_{(S^{h^{-1}}y)_{g_2h_1h \in g_2H}}\ (S^{h_1^{-1}}_\ast\nu_{T^{g_2^{-1}}x})({\mathrm{d}}y)\\ &=& \int_Y \delta_{(S^{h^{-1}}S^{h_1^{-1}}y)_{g_2h_1h \in g_2H}}\ \nu_{T^{g_2^{-1}}x}({\mathrm{d}}y)\\ &=& {\widetilde{\nu}}_{g_2,T^{g_2^{-1}}x}.\end{aligned}$$ It follows that ${\widetilde{\nu}}_x$ does not depend on the choice of cross-section $C$, and the formula (\[eq:margs\]) holds with any choice of $C$. - For each $g \in G$, say $g = ch \in cH$, the marginal of ${\widetilde{\nu}}_x$ on coordinate $g$ is $$S^{h^{-1}}_\ast\nu_{T^{c^{-1}}x} = \nu_{T^{h^{-1}}T^{c^{-1}}x} = \nu_{T^{g^{-1}}x}.$$ - If $(y_g)_g$ is sampled at random from ${\widetilde{\nu}}_x$ and $g \in cH$, then $y_c$ a.s. determines the whole tuple $(y_{ch})_{ch \in cH}$: specifically, $$y_{ch} = S^{h^{-1}}y_c \quad \hbox{a.s.}$$ It also holds that if $g_1$, …, $g_m$ lie in distinct left-cosets of $H$ and $(y_g)_g \sim {\widetilde{\nu}}_x$, then the coordinates $y_{g_1}$, …, $y_{g_m}$ are independent, but we will not need this fact. Finally, let $${\widetilde{\mu}} := \int_X {\widetilde{\nu}}_x\,\mu({\mathrm{d}}x).$$ Recalling the definition of ${\widetilde{X}}$, properties (ii) and (iii) above imply that ${\widetilde{\nu}}_x({\widetilde{X}}) = 1$ for all $x$, and hence also ${\widetilde{\mu}}({\widetilde{X}}) = 1$. We have seen that the left-regular representation defines an action of $G$ on ${\widetilde{X}}$, and the required triangular diagram commutes by the definition of $\pi$, so it remains to check the following. - (The new $G$-space $({\widetilde{X}},{\widetilde{\mu}},{\widetilde{T}})$ is probability-preserving.) Suppose that $k \in G$ and $x \in X$, that $c_1,\ldots,c_m \in C$, that $h_{i,1}$, …, $h_{i,n_i} \in H$ for each $i \leq m$, and that $A_{i,j} \in {\Sigma}_Y$ for all $i \leq m$ and $j\leq n_i$. Then one has $$\begin{aligned} &&{\widetilde{T}}^k_\ast{\widetilde{\nu}}_x\big\{(y_g)_g\,\big|\ y_{c_ih_{i,j}} \in A_{i,j}\ \forall i\leq m,\,j\leq n_i\big\} \\ &&= {\widetilde{\nu}}_x\big\{{\widetilde{T}}^{k^{-1}}(y_g)_g\,\big|\ y_{c_ih_{i,j}} \in A_{i,j}\ \forall i\leq m,\,j\leq n_i\big\} \\ &&= {\widetilde{\nu}}_x\big\{(y_g)_g\,\big|\ y_{k^{-1}c_ih_{i,j}} \in A_{i,j}\ \forall i\leq m,\,j\leq n_i\big\}\end{aligned}$$ Since $C$ is a cross-section of $G/H$, so is $k^{-1}C$. We may therefore apply (\[eq:margs\]) with the cross-section $k^{-1}C$ to deduce that the above is equal to $$\prod_{i=1}^m\nu_{T^{c_i^{-1}k}x}\big(S^{h_{i,1}}(A_{i,1})\cap \cdots \cap S^{h_{i,n_i}}(A_{i,n_i})\big).$$ On the other hand, equation (\[eq:margs\]) applied with the cross-section $C$ gives that this is equal to $${\widetilde{\nu}}_{T^kx}\big\{(y_g)_g\,\big|\ y_{c_ih_{i,j}} \in A_{i,j}\ \forall i\leq m,\,j\leq n_i\big\}.$$ Therefore ${\widetilde{T}}^k_\ast{\widetilde{\nu}}_x = {\widetilde{\nu}}_{T^kx}$, and integrating this over $x$ gives ${\widetilde{T}}^k_\ast{\widetilde{\mu}} = {\widetilde{\mu}}$. - (The map ${\alpha}$ defines a factor map of $H$-spaces.) If $h \in H$ and $(y_g)_g \in {\widetilde{X}}$, then $${\alpha}({\widetilde{T}}^h((y_g)_g)) = {\alpha}((y_{h^{-1}g})_g) = y_{h^{-1}} = S^hy_e = S^h{\alpha}((y_g)_g),$$ where the penultimate equality is given by property (iii) above. Also, property (ii) above gives that $${\alpha}_\ast{\widetilde{\mu}} = \int_X{\alpha}_\ast{\widetilde{\nu}}_x\,\mu({\mathrm{d}}x) = \int_X \nu_x\,\mu({\mathrm{d}}x) = \nu.$$ - (The map $\pi$ defines a factor map of $G$-spaces.) If $k \in G$ and $(y_g)_g \in {\widetilde{X}}$, then $$\pi({\widetilde{T}}^k((y_g)_g)) = {\beta}({\alpha}((y_{k^{-1}g})_g)) = {\beta}(y_{k^{-1}}).$$ If $x\in X$ and $(y_g)_g \sim {\widetilde{\nu}}_x$, then property (ii) above gives that $y_{k^{-1}} \sim \nu_{T^kx}$, and hence ${\beta}(y_{k^{-1}}) = T^kx = T^k{\beta}(y_e)$ a.s. Since this holds for every $x$, integrating over $x$ gives $$\pi({\widetilde{T}}^k((y_g)_g)) = T^k\pi((y_g)_g) \quad \hbox{a.s.}$$ Also, another appeal to property (ii) above gives $$\pi_\ast{\widetilde{\mu}} = \int_X \pi_\ast{\widetilde{\nu}}_x\,\mu({\mathrm{d}}x) = \int_X {\beta}_\ast\nu_x\,\mu({\mathrm{d}}x) = \int_X \delta_x\,\mu({\mathrm{d}}x) = \mu.$$ Functorial ${\sigma}$-subalgebras and subspaces, and satedness {#sec:func} ============================================================== Given $G$, a **functorial ${\sigma}$-subalgebra of $G$-spaces** is a map ${\mathsf{F}}$ which to any $G$-space ${\mathbf{X}}= (X,\mu,T)$ assigns a $\mu$-complete ${\sigma}$-subalgebra $${\Sigma}^{\mathsf{F}}_{\mathbf{X}}\subseteq {\Sigma}_{\mathbf{X}},$$ and such that for any $G$-extension $\pi:{\mathbf{X}}{\longrightarrow}{\mathbf{Y}}$ one has $${\Sigma}^{\mathsf{F}}_{\mathbf{X}}\supseteq \pi^{-1}({\Sigma}^{\mathsf{F}}_{\mathbf{Y}}).$$ Similarly, a **functorial $L^2$-subspace of $G$-spaces** is a map ${\mathsf{V}}$ which to each $G$-space ${\mathbf{X}}= (X,\mu,T)$ assigns a closed subspace $${\mathsf{V}}_{\mathbf{X}}\leq L^2(\mu),$$ and such that for any $G$-extension $\pi:{\mathbf{X}}{\longrightarrow}{\mathbf{Y}}$ one has $${\mathsf{V}}_{\mathbf{X}}\geq {\mathsf{V}}_{\mathbf{Y}}\circ \pi := \{f\circ \pi\,|\ f \in {\mathsf{V}}_{\mathbf{Y}}\}.$$ In this setting, $P^{\mathsf{V}}_{\mathbf{X}}: L^2(\mu) {\longrightarrow}{\mathsf{V}}_{\mathbf{X}}$ will denote the orthogonal projection onto ${\mathsf{V}}_{\mathbf{X}}$. The above behaviour relative to factors is called the **functoriality** of ${\mathsf{F}}$ or ${\mathsf{V}}$. Its first consequence is that ${\mathsf{F}}$ and ${\mathsf{V}}$ respect isomorphisms of $G$-spaces: if ${\alpha}:{\mathbf{X}}\stackrel{\cong}{{\longrightarrow}} {\mathbf{Y}}$, then $${\Sigma}^{\mathsf{F}}_{\mathbf{X}}= {\alpha}^{-1}({\Sigma}^{\mathsf{F}}_{\mathbf{Y}})$$ (where strict equality holds owing to the assumption that these ${\sigma}$-algebras are both $\mu$-complete) and $${\mathsf{V}}_{\mathbf{X}}= {\mathsf{V}}_{\mathbf{Y}}\circ {\alpha}.$$ If $H\leq G$ is any subgroup, then the map ${\mathbf{X}}\mapsto {\Sigma}^H_{\mathbf{X}}$ (the ${\sigma}$-subalgebra of $H$-almost-invariant sets) defines a functorial ${\sigma}$-subalgebra of $G$-spaces. In case $H \unlhd G$, this actually defines a factor of ${\mathbf{X}}$, but otherwise it may not: in general, one has $$T^g({\Sigma}^H_{\mathbf{X}}) = {\Sigma}^{gHg^{-1}}_{\mathbf{X}}.$$ This class of examples will provide the building blocks for all of the other functorial ${\sigma}$-subglebras that we meet later. [$\lhd$]{} If ${\mathsf{F}}$ is a functorial ${\sigma}$-subalgebra of $G$-spaces, then setting $${\mathsf{V}}_{\mathbf{X}}:= L^2(\mu|{\Sigma}^{\mathsf{F}}_{\mathbf{X}})$$ defines a functorial $L^2$-subspace of $G$-spaces, where this denotes the subspace of $L^2(\mu)$ generated by the ${\Sigma}^{\mathsf{F}}_{\mathbf{X}}$-measurable functions. In this case, $P^{\mathsf{V}}_{\mathbf{X}}$ is the operator of conditional expectation onto ${\Sigma}^{\mathsf{F}}_{\mathbf{X}}$. However, not all functorial $L^2$-subspaces arise in this way. For instance, given any two functorial $L^2$-subspaces ${\mathsf{V}}_1$, ${\mathsf{V}}_2$ of $G$-spaces, a new functorial $L^2$-subspace may be defined by $${\mathsf{V}}_{\mathbf{X}}:= {\overline{{\mathsf{V}}_{1,{\mathbf{X}}} + {\mathsf{V}}_{2,{\mathbf{X}}}}}.$$ If $H_1$, $H_2 \leq G$, then this gives rise to the example $${\mathsf{V}}_{\mathbf{X}}:= {\overline{L^2(\mu|{\Sigma}^{H_1}_{\mathbf{X}}) + L^2(\mu|{\Sigma}^{H_2}_{\mathbf{X}})}}.$$ The elements of this subspace generate the functorial ${\sigma}$-algebra ${\Sigma}^{H_1}_{\mathbf{X}}\vee {\Sigma}^{H_2}_{{\mathbf{X}}}$, but in general one may have $${\overline{L^2(\mu|{\Sigma}^{H_1}_{\mathbf{X}}) + L^2(\mu|{\Sigma}^{H_2}_{\mathbf{X}})}} \lneqq L^2(\mu|{\Sigma}^{H_1}_{\mathbf{X}}\vee {\Sigma}^{H_2}_{\mathbf{X}}).$$ In fact, the functorial $L^2$-subspaces that appear later in this work will all correspond to functorial ${\sigma}$-subalgebras. However, the theory of satedness depends only on the subspace structure, so it seems appropriate to develop it in that generality. To prepare for the next definition, recall that if ${\mathfrak{K}}_1,{\mathfrak{K}}_2 \leq {\mathfrak{H}}$ are two closed subspaces of a real Hilbert space, and ${\mathfrak{L}}\leq {\mathfrak{K}}_1 \cap {\mathfrak{K}}_2$ is a common further closed subspace, then ${\mathfrak{K}}_1$ and ${\mathfrak{K}}_2$ are **relatively orthogonal** over ${\mathfrak{L}}$ if $$\langle u,v\rangle = \langle P_{{\mathfrak{L}}}u,P_{{\mathfrak{L}}}v\rangle \quad \forall u \in {\mathfrak{K}}_1,\ v \in{\mathfrak{K}}_2,$$ where $P_{{\mathfrak{L}}}$ is the orthogonal projection onto ${\mathfrak{L}}$. This requires that in fact ${\mathfrak{L}}= {\mathfrak{K}}_1\cap {\mathfrak{K}}_2$, and is equivalent to asserting that $P_{{\mathfrak{K}}_2}u = P_{\mathfrak{L}}u$ for all $u \in {\mathfrak{K}}_1$, and vice-versa. Clearly it suffices to verify this for elements drawn from any dense subsets of ${\mathfrak{K}}_1$ and ${\mathfrak{K}}_2$. Let ${\mathsf{V}}$ be a functorial $L^2$-subspace of $G$-spaces. A $G$-space ${\mathbf{X}}= (X,\mu,T)$ is **${\mathsf{V}}$-sated** if the following holds: for any $G$-extension ${\mathbf{Y}}= (Y,\nu,S)\stackrel{\xi}{{\longrightarrow}} (X,\mu,T)$, the subspaces $L^2(\mu)\circ \xi$ and ${\mathsf{V}}_{\mathbf{Y}}$ are relatively orthogonal over their common further subspace ${\mathsf{V}}_{\mathbf{X}}\circ \xi$. More generally, a $G$-extension ${\widetilde{{\mathbf{X}}}}\stackrel{\pi}{{\longrightarrow}}{\mathbf{X}}$ is **relatively ${\mathsf{V}}$-sated** if the following holds: for any further $G$-extension ${\mathbf{Y}}\stackrel{\xi}{{\longrightarrow}}{\widetilde{{\mathbf{X}}}}$, the subspaces $L^2(\mu)\circ (\pi \circ \xi)$ and ${\mathsf{V}}_{\mathbf{Y}}$ are relatively orthogonal over ${\mathsf{V}}_{{\widetilde{{\mathbf{X}}}}}\circ \pi$. Clearly a $G$-space ${\mathbf{X}}$ is ${\mathsf{V}}$-sated if and only if ${\mathbf{X}}\stackrel{{\mathrm{id}}}{{\longrightarrow}} {\mathbf{X}}$ is relatively ${\mathsf{V}}$-sated. In case ${\mathsf{V}}_{\mathbf{X}}= L^2(\mu|{\Sigma}^{\mathsf{F}}_{\mathbf{X}})$ for some functorial ${\sigma}$-algebra ${\mathsf{F}}$, one may write that a $G$-space or $G$-extension is **${\mathsf{F}}$-sated**, rather than ${\mathsf{V}}$-sated. For a $G$-space ${\mathbf{X}}= (X,\mu,T)$, this asserts that for any $G$-extension $\xi:{\mathbf{Y}}= (Y,\nu,S){\longrightarrow}{\mathbf{X}}$, the ${\sigma}$-subalgebras $$\xi^{-1}({\Sigma}_X) \quad \hbox{and} \quad {\Sigma}^{\mathsf{F}}_{\mathbf{Y}}$$ are relatively independent over $\xi^{-1}({\Sigma}^{\mathsf{F}}_{\mathbf{X}})$. The key feature of satedness is that all $G$-spaces have sated extensions. This generalizes the corresponding result for satedness relative to idempotent classes ([@Aus--thesis Theorem 2.3.2]). The proof here will be a near-verbatim copy of that one, once we have the following auxiliary lemmas. \[lem:bigger-extn-still-rel-sated\] If ${\widetilde{{\mathbf{X}}}}\stackrel{\pi}{{\longrightarrow}} {\mathbf{X}}$ is a relatively ${\mathsf{V}}$-sated $G$-extension, and ${\mathbf{Z}}= (Z,\theta,R)\stackrel{{\alpha}}{{\longrightarrow}} {\widetilde{{\mathbf{X}}}}$ is a further $G$-extension, then ${\mathbf{Z}}\stackrel{{\alpha}\circ \pi}{{\longrightarrow}} {\mathbf{X}}$ is also relatively ${\mathsf{V}}$-sated. Suppose that ${\mathbf{Y}}= (Y,\nu,S)\stackrel{\xi}{{\longrightarrow}} {\mathbf{Z}}$ is another $G$-extension, and that $f \in L^2(\mu)$ and $g \in {\mathsf{V}}_{\mathbf{Y}}$. Then applying the definition of relative satedness to the composed extension ${\mathbf{Y}}\stackrel{{\alpha}\circ \xi}{{\longrightarrow}} {\widetilde{{\mathbf{X}}}}$ gives $$\int_Y(f\circ \pi\circ {\alpha}\circ \xi)\cdot g\,{\mathrm{d}}\nu = \int_Y (P^{\mathsf{V}}_{{\widetilde{{\mathbf{X}}}}}(f\circ\pi)\circ {\alpha}\circ \xi)\cdot g\,{\mathrm{d}}\nu.$$ This will turn into the required equality of inner products if we show that $$(P^{\mathsf{V}}_{{\widetilde{{\mathbf{X}}}}}(f\circ \pi))\circ {\alpha}= P^{\mathsf{V}}_{\mathbf{Z}}(f\circ \pi\circ {\alpha}).$$ However, in light of the inclusion ${\mathsf{V}}_{{\widetilde{{\mathbf{X}}}}}\circ {\alpha}\subseteq {\mathsf{V}}_{\mathbf{Z}}$ and standard properties of orthogonal projection, this is equivalent to the equality $$\int_Z(P^{\mathsf{V}}_{{\widetilde{{\mathbf{X}}}}}(f\circ \pi)\circ {\alpha})\cdot h\,{\mathrm{d}}\theta = \int_Z(f\circ \pi\circ {\alpha})\cdot h\,{\mathrm{d}}\theta \quad \forall h \in {\mathsf{V}}_{\mathbf{Z}},$$ and this is precisely the relative ${\mathsf{V}}$-satedness of $\pi$ applied to ${\alpha}$. \[lem:inv-lim-sated\] If $$\cdots \stackrel{\pi_2}{{\longrightarrow}} {\mathbf{X}}_2 \stackrel{\pi_1}{{\longrightarrow}}{\mathbf{X}}_1 \stackrel{\pi_0}{{\longrightarrow}} {\mathbf{X}}_0$$ is an inverse sequence of $G$-spaces in which each $\pi_i$ is relatively ${\mathsf{V}}$-sated, and if ${\mathbf{X}}_\infty$, $(\psi_m)_m$ is the inverse limit of this sequence, then ${\mathbf{X}}_\infty$ is ${\mathsf{V}}$-sated. First, all the resulting $G$-extensions ${\mathbf{X}}_\infty \stackrel{\psi_m}{{\longrightarrow}} {\mathbf{X}}_m$ are relatively ${\mathsf{V}}$-sated, because we may factorize $\psi_m = \pi_m\circ\psi_{m+1}$ and then apply Lemma \[lem:bigger-extn-still-rel-sated\]. However, this now implies that for any further $G$-extension ${\mathbf{Y}}\stackrel{\xi}{{\longrightarrow}} {\mathbf{X}}_\infty$ and for $\pi := {\mathrm{id}}_{{\widetilde{X}}}$, we have $$\int_Y (f\circ \xi)\cdot g\,{\mathrm{d}}\nu = \int_Y ((P^{\mathsf{V}}_{{\mathbf{X}}_\infty}f)\circ \xi)\cdot g\,{\mathrm{d}}\nu$$ for all $g \in {\mathsf{V}}_{\mathbf{Y}}$ and all $f \in \bigcup_{m\geq 1}(L^2(\mu_m)\circ \psi_m)$. Since this last union is dense in $L^2(\mu_\infty)$, this completes the proof. \[thm:sateds-exist\] If ${\mathsf{V}}$ is a functorial $L^2$-subspace of $G$-spaces, then every $G$-space has a ${\mathsf{V}}$-sated extension. Let ${\mathbf{X}}= (X,\mu,T)$ be a $G$-space. *Step 1*We first show that ${\mathbf{X}}$ has a relatively ${\mathsf{V}}$-sated extension. This uses the same ‘energy increment’ argument as in [@Aus--thesis]. Let $\{f_r\,|\ r\geq 1\}$ be a countable dense subset of the unit ball of $L^2(\mu)$, and let $(r_i)_{i\geq 1}$ be a member of ${\mathbb{N}}^{\mathbb{N}}$ in which every non-negative integer appears infinitely often. We will now construct an inverse sequence $({\mathbf{X}}_m)_{m\geq 0}$, $(\psi^m_k)_{m\geq k \geq 0}$ by the following recursion. First let ${\mathbf{X}}_0 := {\mathbf{X}}$. Then, supposing that for some $m_1 \geq 0$ we have already obtained $({\mathbf{X}}_m)_{m=0}^{m_1}$, $(\psi^m_k)_{m_1 \geq m\geq k\geq 0}$, let $\psi^{m_1+1}_{m_1}:{\mathbf{X}}_{m_1+1}{\longrightarrow}{\mathbf{X}}_{m_1}$ be an extension such that the difference $$\|P^{\mathsf{V}}_{{\mathbf{X}}_{m_1+1}}(f_{r_{m_1}}\circ \psi^{m_1+1}_0)\|_2 - \|P^{\mathsf{V}}_{{\mathbf{X}}_{m_1}}(f_{r_{m_1}}\circ \psi^{m_1}_0\,)\|_2$$ is at least half its supremal possible value over all extensions of ${\mathbf{X}}_{m_1}$, where of course we let $\psi^{m_1+1}_0 := \psi^{m_1}_0\circ \psi^{m_1+1}_{m_1}$. Let ${\mathbf{X}}_\infty$, $(\psi_m)_{m \geq 0}$ be the inverse limit of this sequence. We will show that ${\mathbf{X}}_\infty\stackrel{\psi_0}{{\longrightarrow}}{\mathbf{X}}$ is relatively ${\mathsf{V}}$-sated. Letting $\pi:{\mathbf{Y}}{\longrightarrow}{\mathbf{X}}_\infty$ be an arbitrary further extension, this is equivalent to showing that $$P^{\mathsf{V}}_{{\mathbf{Y}}}(f\circ\psi_0\circ \pi) = P^{\mathsf{V}}_{{\mathbf{X}}_\infty}(f\circ \psi_0)\circ\pi \quad \forall f\in L^2(\mu).$$ It suffices to prove this for every $f_r$ in our previously-chosen dense subset. Also, since ${\mathsf{V}}_{{\mathbf{Y}}} \supseteq {\mathsf{V}}_{{\mathbf{X}}_\infty}\circ \pi$, the result will follow if we only show that $$\|P^{\mathsf{V}}_{{\mathbf{Y}}}(f_r\circ\psi_0\circ\pi)\|_2 = \|P^{\mathsf{V}}_{{\mathbf{X}}_\infty}(f_r\circ\psi_0)\|_2.$$ Suppose, for the sake of contradiction, that the left-hand norm here were strictly larger. The sequence of norms $$\|P^{\mathsf{V}}_{{\mathbf{X}}_m}(f_r\circ\psi^m_0)\|_2$$ is non-decreasing as $m{\longrightarrow}\infty$, and bounded above by $\|f_r\|_2$. Therefore it would follow that for some sufficiently large $m$ we would have $r_m = r$ (since each integer appears infinitely often as some $r_m$) but also $$\begin{gathered} \|P^{\mathsf{V}}_{{\mathbf{X}}_{m+1}}(f_r\circ\psi^{m+1}_0)\|_2 - \|P^{\mathsf{V}}_{{\mathbf{X}}_m}(f_r\circ\psi^m_0)\|_2\\ < \frac{1}{2}\Big(\|P^{\mathsf{V}}_{{\mathbf{Y}}}(f_r\circ\psi_0\circ\pi)\|_2 - \|P^{\mathsf{V}}_{{\mathbf{X}}_\infty}(f\circ\psi_0)\|_2\Big)\\ \leq \frac{1}{2}\Big(\|P^{\mathsf{V}}_{{\mathbf{Y}}}(f_r\circ\psi_0\circ\pi)\|_2 - \|P^{\mathsf{V}}_{{\mathbf{X}}_m}(f\circ\psi^m_0)\|_2\Big).\end{gathered}$$ This would contradict the choice of ${\mathbf{X}}_{m+1}{\longrightarrow}{\mathbf{X}}_m$ in our construction above, so we must actually have the equality of $L^2$-norms required. *Step 2*Iterating the construction of Step 1, we may let $$\cdots \stackrel{\pi_2}{{\longrightarrow}} {\mathbf{X}}_2 \stackrel{\pi_1}{{\longrightarrow}} {\mathbf{X}}_1 \stackrel{\pi_0}{{\longrightarrow}} {\mathbf{X}}$$ be an inverse seqeuence in which each extension $\pi_i$ is relatively ${\mathsf{V}}$-sated. Letting ${\mathbf{X}}_\infty$, $(\pi_m)_{m\geq 0}$ be its inverse limit, Lemma \[lem:inv-lim-sated\] completes the proof. \[cor:mult-sateds-exist\] Let ${\mathsf{V}}_1$, ${\mathsf{V}}_2$, …be any countable family of functorial $L^2$-subspaces of $G$-spaces. Then every $G$-space has an extension which is simultaneously ${\mathsf{V}}_r$-sated for every $r$. Let $(r_i)_i$ be an element of ${\mathbb{N}}^{\mathbb{N}}$ in which every positive integer appears infinitely often. By repeatedly implementing Theorem \[thm:sateds-exist\], let $$\cdots \stackrel{\pi_2}{{\longrightarrow}} {\mathbf{X}}_2 \stackrel{\pi_1}{{\longrightarrow}} {\mathbf{X}}_1 \stackrel{\pi_0}{{\longrightarrow}} {\mathbf{X}}$$ be an inverse sequence in which each ${\mathbf{X}}_i$ is ${\mathsf{V}}_{r_i}$-sated. Also, let $\pi^n_m := \pi_m\circ \cdots \circ \pi_{n-1}$ whenever $m < n$. Finally, let ${\mathbf{X}}_\infty$ be the inverse limit of this sequence. Then for each $r \geq 1$, there is an infinite subsequence $i_1(r) < i_2(r) < \ldots $ in ${\mathbb{N}}$ such that $r_{i_1(r)} = r_{i_2(r)} = \cdots = r$, and ${\mathbf{X}}_\infty$ may be identified with the inverse limit of the thinned-out inverse sequence $$\cdots \stackrel{\pi^{i_3(r)}_{i_2(r)}}{{\longrightarrow}} {\mathbf{X}}_{i_2(r)} \stackrel{\pi^{i_2(r)}_{i_1(r)}}{{\longrightarrow}} {\mathbf{X}}_{i_1(r)} \stackrel{\pi^{i_1(r)}_0}{{\longrightarrow}} {\mathbf{X}}.$$ Given this, Lemma \[lem:inv-lim-sated\] implies that ${\mathbf{X}}_\infty$ is ${\mathsf{V}}_r$-sated. Since $r$ was arbitrary, this completes the proof. Characteristic subspaces and proof of convergence ================================================= Subgroups associated to commuting tuples of actions --------------------------------------------------- We now begin to work with commuting tuples of $G$-actions. We will need to call on several different subgroups of $G^d$ in the sequel, so the next step is to set up some bespoke notation for handling them. We will sometimes use a boldface ${\mathbf{g}}$ to denote a tuple $(g_i)_{i=1}^d$ in $G^d$, and will denote the identity element of $G$ by $1_G$. Fix $G$ and $d$, and let $e = \{i_1 < \ldots < i_r\} \subseteq [d]$ with $r\geq 2$ and also $\{i < j\} \subseteq [d]$. Then we define $$H_e := \{{\mathbf{g}}\in G^d\,|\ g_{i_s+1} = g_{i_s+2} = \ldots = g_{i_{s+1}}\ \hbox{for each}\ s=1,\ldots,r-1\},$$ $$K_{\{i,j\}} := \{{\mathbf{g}} \in H_{\{i,j\}}\,|\ g_\ell = 1_G \ \forall \ell \in (i;j]\},$$ and $$L_e := \{{\mathbf{g}} \in H_e\,|\ g_i = 1_G \ \forall i \in [d]\setminus (i_1;i_r]\}.$$ Routine calculations give the following basic properties. \[lem:subgp-props\] 1. For $e = \{i_1 < \ldots < i_r\}$ as above, the subgroups $L_e$ and $K_{\{i_1,i_r\}}$ commute and generate $H_e$. 2. If $a \subseteq e \subseteq [k]$ with $|a| \geq 2$, then $L_a \leq L_e$. 3. If $a \subseteq e \subseteq [k]$ with $|a| \geq 2$, and also $$e \cap [\min a;\max a] = a,$$ then $L_a \unlhd H_e$. In particular, $L_e \unlhd H_e$. Part (3) of this lemma has the following immediate consequence. \[cor:some-invce\] If $a \subseteq e \subseteq [k]$ with $|a| \geq 2$, and also $$e \cap [\min a;\max a] = a,$$ then ${\Sigma}^{L_a}_{\mathbf{X}}$ is globally $H_e$-invariant. The Host-Kra inequality ----------------------- In order to show that a suitably-sated $G$-space has some other desirable property, one must find an extension of it for which the relative independence given by satedness implies that other property. The key to such a proof is usually constructing the right extension. Where satedness was used in the previous works [@Aus--nonconv] and [@Aus--newmultiSzem], that extension could be constructed directly from the Furstenberg self-joining arising from some non-conventional averages. However, this seems to be more problematic in the present setting, and we will take a different approach. The construction below is a close analog of the construction by Host and Kra of certain ‘cubical’ extensions of a ${\mathbb{Z}}$-space in [@HosKra05]. That machinery has also been extended by Host to commuting tuples of ${\mathbb{Z}}$-actions in [@Hos09]. Fix now a $G^d$-space ${\mathbf{X}}= (X,\mu,T_1,\ldots,T_d)$, and let ${\mathbf{Y}}^{(0)} := {\mathbf{X}}$. Our next step is to construct recursively a height-$(d+1)$ tower of new probability-preserving $G^d$-spaces, which we shall denote by $$\begin{aligned} \label{eq:tower} {\mathbf{Y}}^{(d)} \stackrel{\xi^{(d)}}{{\longrightarrow}} {\mathbf{Y}}^{(d-1)} \stackrel{\xi^{(d-1)}}{{\longrightarrow}} \cdots \stackrel{\xi^{(2)}}{{\longrightarrow}} {\mathbf{Y}}^{(1)} \stackrel{\xi^{(1)}}{{\longrightarrow}} {\mathbf{Y}}^{(0)} = {\mathbf{X}}.\end{aligned}$$ The construction will also give some other auxiliary $G^d$-spaces ${\mathbf{Z}}^{(j)}$, and they too will be used later. Supposing the tower has already been constructed up to some level $j \leq d-1$, the next extension is constructed in the following steps. - From ${\mathbf{Y}}^{(j)} = (Y^{(j)},\nu^{(j)},S^{(j)})$, define a new $H_{\{d-j-1,d\}}$-action ${\widetilde{S}}^{(j)}$ on the same space by setting $$\begin{aligned} \label{eq:comm1} ({\widetilde{S}}^{(j)}_i)^g := (S^{(j)}_i)^g \quad \forall g \in G,\ i < d-j-1,\end{aligned}$$ $$\begin{aligned} \label{eq:comm2} ({\widetilde{S}}^{(j)}_{d-j-1})^g := (S^{(j)}_{[d-j-1;d]})^g \quad \forall g \in G,\end{aligned}$$ and $$\begin{aligned} \label{eq:comm3} ({\widetilde{S}}^{(j)}_{(d-j-1;d]})^g := {\mathrm{id}} \quad \forall g \in G\end{aligned}$$ (with the understanding that (\[eq:comm1\]) and (\[eq:comm2\]) are vacuous in case $j=d-1$). - Now consider the $H_{\{d-j-1,d\}}$-space $$\begin{aligned} {\mathbf{Z}}^{(j+1)} &=& (Z^{(j+1)},\theta^{(j+1)},R^{(j+1)})\\ &:=& \big(Y^{(j)}\times Y^{(j)},\ \nu^{(j)}\otimes_{{\Sigma}^{L_{\{d-j-1,d\}}}_{{\mathbf{Y}}^{(j)}}} \nu^{(j)},\ (S^{(j)})^{{\!\upharpoonright}H_{\{d-j-1,d\}}}\times {\widetilde{S}}^{(j)}\big).\end{aligned}$$ Let $\xi^{(j+1)}_0,\xi^{(j+1)}_1:Z^{(j+1)}{\longrightarrow}Y^{(j)}$ be the two coordinate projections. They are both factor maps of $H_{\{d-j-1,d\}}$-spaces. Notice that $\theta^{(j+1)}$ is $R^{(j+1)}$-invariant because both of the actions $(S^{(j)})^{{\!\upharpoonright}H_{\{d-j-1,d\}}}$ and ${\widetilde{S}}^{(j)}$ preserve the ${\sigma}$-subalgebra ${\Sigma}^{L_{\{d-j-1,d\}}}_{{\mathbf{Y}}^{(j)}}$, by Corollary \[cor:some-invce\]. - Finally, let ${\mathbf{Y}}^{(j+1)}\stackrel{\xi^{(j+1)}}{{\longrightarrow}} {\mathbf{Y}}^{(j)}$ be an extension of $G^d$-spaces for which there is a commutative diagram $\phantom{i}$ as provided by Theorem \[thm:recoverG\]. Having made this construction, for each $j\in \{1,2,\ldots,d\}$ we also define a family of maps $$\pi^{(j)}_\eta:Y^{(j)} {\longrightarrow}X$$ indexed by $\eta \in \{0,1\}^j$, by setting $$\pi^{(j)}_{(\eta_1,\ldots,\eta_j)}:= \xi^{(1)}_{\eta_1}\circ{\alpha}^{(1)}\circ \xi^{(2)}_{\eta_2}\circ {\alpha}^{(2)} \circ \cdots\circ \xi^{(j)}_{\eta_j}\circ {\alpha}^{(j)}.$$ Clearly $(\pi^{(j)}_\eta)_\ast \nu^{(j)} = \mu$ for every $\eta$. Also, $$\pi^{(j)}_{0^j} = \xi^{(1)}\circ\cdots\circ \xi^{(j)}:{\mathbf{Y}}^{(j)}{\longrightarrow}{\mathbf{X}}$$ is a factor map of $G^d$-spaces, where $0^j:= (0,0,\ldots,0) \in \{0,1\}^j$. \[lem:intertwine\] Let $r \in [d]$, let $\eta \in \{0,1\}^r\setminus\{0^r\}$, and let $\ell \in [r]$ be maximal such that $\eta_\ell = 1$. Then $\pi^{(r)}_\eta$ satisfies the following intertwining relations $$\begin{aligned} \label{eq:comm4} \pi^{(r)}_\eta\circ S^{(r)}_i = T_i\circ \pi^{(r)}_\eta \quad \forall i < d-\ell,\end{aligned}$$ $$\begin{aligned} \label{eq:comm5} \pi^{(r)}_\eta\circ S^{(r)}_{d-\ell} = T_{[d-\ell;d]}\circ \pi^{(r)}_\eta\end{aligned}$$ and $$\begin{aligned} \label{eq:comm6} \pi^{(r)}_\eta\circ S^{(r)}_{(d-\ell;d]} = \pi^{(r)}_\eta.\end{aligned}$$ There are no such simple relations for the compositions $\pi^{(r)}_\eta\circ S^{(r)}_i$ when $i \geq d-\ell+1$, but we will not need these. [$\lhd$]{} By the definition of $\ell$, for this $\eta$ we may write $\pi^{(r)}_\eta = \pi'\circ\pi''$, where $$\begin{gathered} \label{eq:dfn-pieta} \pi' := \pi^{(\ell)}_{(\eta_1,\ldots,\eta_\ell)} = \xi^{(1)}_{\eta_1}\circ{\alpha}^{(1)}\circ \xi^{(2)}_{\eta_2}\circ {\alpha}^{(2)}\circ \cdots\circ \xi^{(\ell)}_1\circ {\alpha}^{(\ell)}\\ \hbox{and}\quad \pi'' := \xi^{(\ell+1)}\circ \cdots\circ \xi^{(r)}.\end{gathered}$$ All three of the desired relations concern the actions of subgroups of $H_{\{d-\ell,d\}}$, and all the maps in the compositions in (\[eq:dfn-pieta\]) are factor maps of $H_{\{d-\ell,d\}}$-spaces. We will read off the desired results from the simpler relations (\[eq:comm1\]), (\[eq:comm2\]) and (\[eq:comm3\]). In the first place, each $\xi^{(j)}$ appearing in the definition of $\pi''$ actually intertwines the whole $G^d$-actions by construction, so $$\pi''\circ S_i^{(r)} = S_i^{(\ell)}\circ \pi'' \quad \forall i \in [d].$$ It therefore suffices to prove that $\pi' \circ S^{(\ell)}_i = T_i\circ \pi'$ for all $i < d - \ell$, and similarly for the other two desired relations. *Step 1.*Suppose that $i \leq d-\ell$ and $j \leq \ell$. Then the definitions of ${\alpha}^{(j)}$, $\xi_0^{(j)}$ and $\xi_1^{(j)}$ give $$\xi^{(j)}_\eta\circ{\alpha}^{(j)}\circ S_i^{(j)} = \xi^{(j)}_\eta\circ R^{(j)}_i \circ {\alpha}^{(j)}= \left\{\begin{array}{ll} S^{(j-1)}_i\circ \xi^{(j)}_\eta\circ{\alpha}^{(j)} & \quad \hbox{if}\ \eta = 0\\ {\widetilde{S}}^{(j-1)}_i \circ \xi^{(j)}_\eta\circ{\alpha}^{(j)} & \quad \hbox{if}\ \eta = 1.\end{array}\right.$$ In case $i < d - \ell \leq d - j$, this is equal to $S^{(j)}_i\circ \xi^{(j)}_\eta\circ{\alpha}^{(j)}$ for either value of $\eta$, by (\[eq:comm1\]). Applying this repeatedly for $j=\ell,\ell-1,\ldots,1$ in the composition that defines $\pi'$, we obtain $$\pi'\circ S_i^{(\ell)} = T_i\circ \pi'.$$ As explained above, this proves (\[eq:comm4\]). *Step 2.*The same calculation as above gives $$\xi^{(\ell)}_1\circ {\alpha}^{(\ell)}\circ S^{(\ell)}_{d-\ell} = \xi^{(\ell)}_1\circ R^{(\ell)}_{d-\ell} \circ {\alpha}^{(\ell)}= {\widetilde{S}}^{(\ell-1)}_{d-\ell}\circ \xi^{(\ell)}_1\circ {\alpha}^{(\ell)},$$ and now this is equal to $S^{(\ell-1)}_{[d-\ell,d]}\circ \xi^{(\ell)}_1\circ {\alpha}^{(\ell)}$, by (\[eq:comm2\]). On the other hand, if $j \leq \ell-1$, then another call to the definitions of ${\alpha}^{(j)}$, $\xi_0^{(j)}$ and $\xi_1^{(j)}$ gives $$\xi^{(j)}_\eta\circ{\alpha}^{(j)}\circ S_{[d-\ell,d]}^{(j)} = \xi^{(j)}_\eta\circ R^{(j)}_{[d-\ell,d]}\circ {\alpha}^{(j)} = \left\{\begin{array}{ll} S^{(j-1)}_{[d-\ell;d]}\circ \xi^{(j)}_\eta\circ{\alpha}^{(j)} & \quad \hbox{if}\ \eta = 0\\ {\widetilde{S}}^{(j-1)}_{[d-\ell;d]}\circ \xi^{(j)}_\eta\circ{\alpha}^{(j)} & \quad \hbox{if}\ \eta = 1.\end{array}\right.$$ This time, since $j \leq \ell-1$, one has $$\begin{aligned} {\widetilde{S}}^{(j-1)}_{[d-\ell;d]} &=& {\widetilde{S}}^{(j-1)}_{d-\ell}\circ {\widetilde{S}}^{(j-1)}_{d-\ell+1}\circ \cdots \circ {\widetilde{S}}^{(j-1)}_{d-j}\circ {\widetilde{S}}^{(j-1)}_{(d-j;d]}\\ &=& S^{(j-1)}_{d-\ell}\circ S^{(j-1)}_{d-\ell+1}\circ \cdots \circ S^{(j-1)}_{[d-j;d]}\circ {\mathrm{id}} = S^{(j-1)}_{[d-\ell;d]},\end{aligned}$$ using (\[eq:comm1\]) and (\[eq:comm2\]). So we obtain $$\xi^{(j)}_\eta\circ{\alpha}^{(j)}\circ S_{[d-\ell,d]}^{(j)} = S^{(j-1)}_{[d-\ell,d]} \circ \xi^{(j)}_\eta\circ{\alpha}^{(j)}$$ for all $j \leq \ell-1$ and either value of $\eta$. Combining these two calculations gives $$\pi'\circ S^{(\ell)}_{d-\ell} = (\xi^{(1)}_{\eta_1}\circ {\alpha}^{(1)}\circ \cdots \circ \xi^{(\ell-1)}_{\eta_{\ell-1}}\circ {\alpha}^{(\ell-1)})\circ S^{(\ell-1)}_{[d-\ell;d]} = T_{[d-\ell;d]}\circ \pi',$$ and hence (\[eq:comm5\]). *Step 3.*Finally, (\[eq:comm3\]) gives $$\xi^{(\ell)}_1\circ {\alpha}^{(1)}\circ S^{(\ell)}_{(d-\ell;d]} = {\widetilde{S}}^{(\ell-1)}_{(d-\ell;d]}\circ \xi^{(\ell)}_1\circ {\alpha}^{(1)} = \xi^{(\ell)}_1\circ {\alpha}^{(1)},$$ from which (\[eq:comm6\]) follows immediately. \[cor:intertwine\] If $r \in [d]$, $\eta \in \{0,1\}^r$, and if $j \in [r]$ is such that $\eta_i = 0$ for all $i \geq j+1$, then $\pi^{(r)}_\eta$ satisfies the following intertwining relations $$\begin{aligned} \label{eq:comm7} \pi^{(r)}_\eta\circ S^{(r)}_{[d-j;d]} = T_{[d-j;d]}\circ \pi^{(r)}_\eta.\end{aligned}$$ We next prove an estimate relating the multi-linear forms ${\Lambda}_n$ in (\[eq:Lambda\]) to certain integrals over these new $G^d$-spaces ${\mathbf{Y}}^{(j)}$. This will be the key estimate that enables an appeal to satedness. The following theorem relies on an iterated application of the van der Corput estimate, and follows essentially the same lines as Theorem 12.1 in [@HosKra05]. \[thm:ineq\] Let ${\mathbf{X}}= (X,\mu,T_1,\ldots,T_d)$ be a $G^d$-space, let $1 \leq j \leq d$, and let the tower (\[eq:tower\]) and the maps $\pi^{(j)}_\eta:Y^{(j)}{\longrightarrow}X$ for $\eta\in \{0,1\}^j$ be constructed as above. For $f_{d-j+1}$, …, $f_d \in L^\infty(\mu)$, let $${\Lambda}_n^{(j)}(f_{d-j+1},\ldots,f_d) := \frac{1}{|F_n|}\sum_{g \in F_n}\prod_{i=d-j+1}^d(f_i\circ T_{[d-j+1;i]}^g).$$ If $f_{d-j+1}$, …, $f_d$ are all uniformly bounded by $1$, then $$\limsup_{n{\longrightarrow}\infty}\|{\Lambda}_n^{(j)}(f_{d-j+1},\ldots,f_d)\|_2 \leq \Big(\int_{Y^{(j)}} \prod_{\eta \in \{0,1\}^j} ({\mathcal{C}}^{|\eta|}f_d\circ \pi^{(j)}_\eta)\,{\mathrm{d}}\nu^{(j)}\Big)^{2^{-j}},$$ where $|\eta| := \sum_i\eta_i \mod 1$ and ${\mathcal{C}}$ is the operator of complex conjugation. Note that ${\Lambda}_n^{(d)} = {\Lambda}_n$, the averages in (\[eq:Lambda\]). The integral appearing on the right-hand side of the last inequality actually defines a seminorm of the function $f_d$: these are the adaptations of the Host-Kra seminorms to the present setting. However, our approach will not emphasize the seminorm axioms. This is proved by induction on $j$. *Step 1: Base case.*When $j=1$ the Norm Ergodic Theorem for amenable groups gives $${\Lambda}_n^{(1)}(f_d) {\longrightarrow}{\mathsf{E}}_\mu(f_d\,|\,{\Sigma}_{\mathbf{X}}^{T_d}) = {\mathsf{E}}_\mu(f_d\,|\,{\Sigma}_{\mathbf{X}}^{L_{\{d-1,d\}}}) \quad \hbox{in}\ \|\cdot\|_2,$$ and the square of the norm of this limit is equal to $$\begin{gathered} \int_{X\times X} (f_d\otimes {\overline{f_d}})\,{\mathrm{d}}\big(\mu\otimes_{{\Sigma}_{\mathbf{X}}^{L_{\{d-1,d\}}}}\mu\big) = \int_{Z^{(1)}}f_d\circ \xi_0^{(1)}\cdot{\overline{f_d\circ \xi_1^{(1)}}}\,{\mathrm{d}}\theta^{(1)}\\ = \int_{Y^{(1)}}f_d\circ \pi_0^{(1)}\cdot{\overline{f_d\circ \pi_1^{(1)}}}\,{\mathrm{d}}\nu^{(1)},\end{gathered}$$ by the definition of $Y^{(1)}$ and $\nu^{(1)}$. *Step 2: Van der Corput estimate.*Now suppose the result is known up to some $j-1 \in \{1,2,\ldots,d-1\}$. By the amenable-groups version of the van der Corput estimate ([@BerMcCZha97 Lemma 4.2]), one has $$\begin{gathered} \label{eq:vdC} \limsup_{n{\longrightarrow}\infty}\|{\Lambda}_n^{(j)}(f_{d-j+1},\ldots,f_d)\|^2_2\\ \leq \limsup_{m{\longrightarrow}\infty}\frac{1}{|F_m|^2}\sum_{h,k \in F_m}\limsup_{n{\longrightarrow}\infty}\Big|\frac{1}{|F_n|}\sum_{g \in F_n}\int_X \prod_{i=d-j+1}^d(f_i\circ T_{[d-j+1;i]}^{hg}){\overline{(f_i\circ T_{[d-j+1;i]}^{kg})}}\,{\mathrm{d}}\mu\Big|\end{gathered}$$ For fixed $h$ and $k$, we may use the $T^g_{d-j+1}$-invariance of $\mu$ to re-arrange the above integral as follows: $$\begin{aligned} &&\Big|\frac{1}{|F_n|}\sum_{g \in F_n}\int_X \prod_{i=d-j+1}^d(f_i\circ T_{[d-j+1;i]}^{hg}){\overline{(f_i\circ T_{[d-j+1;i]}^{kg})}}\,{\mathrm{d}}\mu\Big|\\ &&= \Big|\frac{1}{|F_n|}\sum_{g \in F_n}\int_X (f_{d-j+1}\circ T_{d-j+1}^h)\cdot {\overline{(f_{d-j+1}\circ T_{d-j+1}^k)}}\\ &&\quad\quad\quad\quad\quad\quad\quad\quad \cdot\Big(\prod_{i=d-j+2}^d((f_i\circ T_{[d-j+1;i]}^h)\cdot {\overline{(f_i\circ T_{[d-j+1;i]}^k)}})\circ T_{(d-j+1;i]}^g\Big)\,{\mathrm{d}}\mu\Big|.\end{aligned}$$ (At this point we have made crucial use of the commutativity of the different actions $T_i$.) By the Cauchy-Bunyakowski-Schwartz Inequality, this, in turn, is bounded above by $$\begin{aligned} &&\|(f_{d-j+1}\circ T_{d-j+1}^h)\cdot {\overline{(f_{d-j+1}\circ T_{d-j+1}^k)}}\|_2\\ &&\quad\quad\quad\quad \cdot\Big\|\frac{1}{|F_n|}\sum_{g \in F_n}\prod_{i=d-j+2}^d((f_i\circ T_{[d-j+1;i]}^h)\cdot {\overline{(f_i\circ T_{[d-j+1;i]}^k)}})\circ T_{(d-j+1;i]}^g\Big\|_2\\ &&\leq \big\|{\Lambda}_n^{(j-1)}\big((f_{d-j+2}\circ T_{[d-j+1;d-j+2]}^h)\cdot {\overline{(f_{d-j+2}\circ T_{[d-j+1;d-j+2]}^k)}},\\ &&\quad\quad\quad\quad\quad\quad\quad\quad\quad\quad\quad\quad\quad\quad\quad\quad \ldots,(f_d\circ T_{[d-j+1;d]}^h)\cdot {\overline{(f_d\circ T_{[d-j+1;d]}^k)}}\big)\big\|_2\end{aligned}$$ (since $\|f_{d-j+1}\|_\infty \leq 1$). *Step 3: Use of inductive hypothesis.*Combining these inequalities and using the inductive hypothesis, this gives $$\begin{gathered} \label{eq:use-of-ind-hyp} \limsup_{n{\longrightarrow}\infty}\Big|\frac{1}{|F_n|}\sum_{g \in F_n}\int_X \prod_{i=d-j+1}^d(f_i\circ T_{[d-j+1;i]}^{hg}){\overline{(f_i\circ T_{[d-j+1;i]}^{kg})}}\,{\mathrm{d}}\mu\Big|\\ \leq \Big(\int_{Y^{(j-1)}}\prod_{\eta \in \{0,1\}^{j-1}}\big({\mathcal{C}}^{|\eta|}((f_d\circ T_{[d-j+1;d]}^h)\cdot {\overline{(f_d\circ T_{[d-j+1;d]}^k)}})\circ \pi^{(j-1)}_\eta\big)\,{\mathrm{d}}\nu^{(j-1)}\Big)^{2^{-(j-1)}}\end{gathered}$$ for each $h$ and $k$. To lighten notation, now let $$F := \prod_{\eta \in \{0,1\}^{j-1}}({\mathcal{C}}^{|\eta|}f_d\circ \pi_\eta^{(j-1)}).$$ In terms of this function, Corollary \[cor:intertwine\] with $r := j-1$ allows us to write $$\begin{gathered} \prod_{\eta \in \{0,1\}^{j-1}}({\mathcal{C}}^{|\eta|}(f_d\circ T^h_{[d-j+1;d]})\circ \pi_\eta^{(j-1)})\\ = \prod_{\eta \in \{0,1\}^{j-1}}({\mathcal{C}}^{|\eta|}f_d\circ \pi_\eta^{(j-1)}\circ (S^{(j-1)}_{[d-j+1;d]})^h) = F\circ (S^{(j+1)}_{[d-j+1;d]})^h,\end{gathered}$$ and similarly $$\prod_{\eta \in \{0,1\}^{j-1}}({\mathcal{C}}^{|\eta|}{\overline{(f_d\circ T^k_{[d-j+1;d]})}}\circ \pi_\eta^{(j-1)}) = {\overline{F}}\circ (S^{(j-1)}_{[d-j+1;d]})^k.$$ *Step 4: Finish.*Substituting into the right-hand side of (\[eq:use-of-ind-hyp\]), one obtains $$\begin{gathered} \limsup_{n{\longrightarrow}\infty}\Big|\frac{1}{|F_n|}\sum_{g \in F_n}\int_X \prod_{i=d-j+1}^d(f_i\circ T_{[d-j+1;i]}^{hg}){\overline{(f_i\circ T_{[d-j+1;i]}^{kg})}}\,{\mathrm{d}}\mu\Big|\\ \leq \Big(\int_{Y^{(j-1)}}(F\circ (S^{(j+1)}_{[d-j+1;d]})^h)\cdot ({\overline{F}}\circ (S^{(j-1)}_{[d-j+1;d]})^k)\,{\mathrm{d}}\nu^{(j-1)}\Big)^{2^{-(j-1)}}.\end{gathered}$$ Inserting this back into (\[eq:vdC\]) and using Hölder’s Inequality for the average over $(h,k)$, one obtains $$\begin{aligned} &&\limsup_{n{\longrightarrow}\infty}\|{\Lambda}_n^{(j)}(f_{d-j+1},\ldots,f_d)\|^2_2\\ &&\leq \limsup_{m{\longrightarrow}\infty}\frac{1}{|F_m|^2}\sum_{h,k \in F_m}\Big(\int_{Y^{(j-1)}}(F\circ (S^{(j+1)}_{[d-j+1;d]})^h)\cdot ({\overline{F}}\circ (S^{(j-1)}_{[d-j+1;d]})^k)\,{\mathrm{d}}\nu^{(j-1)}\Big)^{2^{-(j-1)}}\\ &&\leq \limsup_{m{\longrightarrow}\infty}\Big(\frac{1}{|F_m|^2}\sum_{h,k \in F_m}\int_{Y^{(j-1)}}(F\circ (S^{(j+1)}_{[d-j+1;d]})^h)\cdot ({\overline{F}}\circ (S^{(j-1)}_{[d-j+1;d]})^k)\,{\mathrm{d}}\nu^{(j-1)}\Big)^{2^{-(j-1)}}.\end{aligned}$$ Finally, the averages on the last line here actually converge as $m{\longrightarrow}\infty$, by the Norm Ergodic Theorem for amenable groups, giving $$\begin{aligned} &&\limsup_{n{\longrightarrow}\infty}\|{\Lambda}_n^{(j)}(f_{d-j+1},\ldots,f_d)\|^2_2\\ &&\leq \Big(\int_{Y^{(j-1)}} {\mathsf{E}}_{\nu^{(j-1)}}\big(F\,\big|\,{\Sigma}_{{\mathbf{Y}}^{(j-1)}}^{L_{\{d-j,d\}}}\big)\cdot {\mathsf{E}}_{\nu^{(j-1)}}\big({\overline{F}}\,\big|\,{\Sigma}_{{\mathbf{Y}}^{(j-1)}}^{L_{\{d-j,d\}}}\big)\,{\mathrm{d}}\nu^{(j-1)}\Big)^{2^{-(j-1)}}\\ &&= \Big(\int_{Z^{(j)}} F\circ \xi^{(j)}_0\cdot {\overline{F\circ \xi^{(j)}_1}}\,{\mathrm{d}}\theta^{(j)}\Big)^{2^{-(j-1)}}\\ &&= \Big(\int_{Y^{(j)}} \Big(\prod_{\eta \in \{0,1\}^j}{\mathcal{C}}^{|\eta|}f_d\circ \pi^{(j)}_\eta\Big)\,{\mathrm{d}}\nu^{(j)}\Big)^{2^{-(j-1)}},\end{aligned}$$ where ${\mathbf{Z}}^{(j)}$ is the auxiliary $H_{\{d-j,d\}}$-space constructed along with ${\mathbf{Y}}^{(j)}$. Taking square-roots, this continues the induction. Partially characteristic subspaces and the proof of convergence {#subs:part-char} --------------------------------------------------------------- Consider a probability space $(X,\mu)$, and a sequence $\Xi_n$ of multi-linear forms on $L^\infty(\mu)$ which are separately continuous for the norm $\|\cdot\|_2$ in each entry. A closed subspace $V \leq L^2(\mu)$ is **partially characteristic in position $i$** for the sequence $\Xi_n$ if $$\big\|\Xi_n(f_1,\ldots,f_d) - \Xi_n(f_1,\ldots,f_{i-1},P^Vf_i,f_{i+1},\ldots,f_d)\big\|_2 {\longrightarrow}0$$ as $n {\longrightarrow}\infty$ for all $f_1$, …, $f_d \in L^\infty(\mu)$, where $P^V$ is the orthogonal projection onto $V$. The following proposition will quickly lead to a proof of Theorem A. In fact, it gives rather more than one needs for the proof of Theorem A, but that extra strength will be used during the proof of Theorem B. \[prop:pleasant\] For $1 \leq i \leq j \leq d$, let ${\mathsf{F}}_{i,j}$ be the functorial ${\sigma}$-algebra $${\Sigma}^{{\mathsf{F}}_{i,j}}_{\mathbf{X}}:= \bigvee_{\ell = 0}^{i-1}{\Sigma}_{\mathbf{X}}^{T_{(\ell;i]}}\vee \bigvee_{\ell = i+1}^j{\Sigma}_{\mathbf{X}}^{T_{(i;\ell]}},$$ and let ${\mathsf{V}}_{i,j,{\mathbf{X}}} := L^2(\mu|{\Sigma}^{{\mathsf{F}}_{i,j}}_{\mathbf{X}})$ be the associated functorial $L^2$-subspace. Also, let $${\widehat{{\Lambda}}}_n^{(j)}(f_1,\ldots,f_j) := \frac{1}{|F_n|}\sum_{g \in F_n}\prod_{i=1}^{j}(f_i\circ T_{[1;i]}^g).$$ If ${\mathbf{X}}$ is ${\mathsf{V}}_{i,j}$-sated whenever $1 \leq i\leq j \leq d$, then, for each $j\in [d]$, the subspaces $${\mathsf{V}}_{1,j,{\mathbf{X}}},\ \ldots,\ {\mathsf{V}}_{j,j,{\mathbf{X}}}$$ are partially characteristic in positions $1$, …, $j$ for the averages ${\widehat{{\Lambda}}}_n^{(j)}$. Notice that we still have ${\widehat{{\Lambda}}}_n^{(d)} = {\Lambda}_n$ (the averages in (\[eq:Lambda\])), but otherwise these averages differ from the averages ${\Lambda}_n^{(j)}$ considered in Theorem \[thm:ineq\]. This is proved by induction on $j$. When $j=1$, hence also $i=1$, we have ${\Sigma}^{{\mathsf{F}}_{i,j}}_{\mathbf{X}}= {\Sigma}^{T_1}_{\mathbf{X}}$, and this is always partially characteristic because the Norm Ergodic Theorem gives $${\widehat{\Lambda}}^{(1)}_n(f_1) {\longrightarrow}{\mathsf{E}}_\mu(f_1\,|\,{\Sigma}^{T_1}_{\mathbf{X}}) \quad \hbox{in}\ \|\cdot\|_2$$ for any $G$-space. So now we focus on the recursion clause. For this it clearly suffices to assume $j=d-1$, and prove the result for the averages ${\widehat{\Lambda}}^{(d)}_n$, which will lighten the notation. *Step 1.*We first show that ${\mathsf{V}}_{d,d}$ is partially characteristic in position $d$. Let $f_d \in L^\infty(\mu)$. By decomposing it as $$P^{{\mathsf{V}}_{d,d}}_{\mathbf{X}}f_d + (f_d - P^{{\mathsf{V}}_{d,d}}_{\mathbf{X}}f_d),$$ and using the multi-linearity of ${\widehat{{\Lambda}}}^{(d)}_n$, it suffices to show that $$P^{{\mathsf{V}}_{d,d}}_{\mathbf{X}}f_d = 0 \quad \Longrightarrow \quad \|{\widehat{{\Lambda}}}^{(d)}_n(f_1,\ldots,f_d)\|_2 {\longrightarrow}0 \quad \forall f_1,\ldots,f_{d-1} \in L^\infty(\mu).$$ We will prove this in its contrapositive, so suppose that $$\limsup_{n{\longrightarrow}\infty}\|{\widehat{{\Lambda}}}^{(d)}_n(f_1,\ldots,f_d)\|_2 > 0 \quad \hbox{for some}\ f_1,\ldots,f_{d-1}.$$ Then Theorem \[thm:ineq\] gives $$\int_{Y^{(d)}}\Big(\prod_{\eta \in \{0,1\}^d}{\mathcal{C}}^{|\eta|}f_d\circ \pi_\eta^{(d)}\Big)\,{\mathrm{d}}\nu^{(d)} \neq 0.$$ However, recalling relation (\[eq:comm6\]) from Lemma \[lem:intertwine\], we see that if $\eta \in \{0,1\}^d\setminus \{0^d\}$ and $\ell \in [d]$ is maximal such that $\eta_\ell \neq 0$, then $${\mathcal{C}}^{|\eta|}f_d\circ \pi_\eta^{(d)}\circ S_{(d-\ell;d]}^g = {\mathcal{C}}^{|\eta|}f_d\circ \pi_\eta^{(d)} \quad \forall g \in G,$$ and so the function $$\prod_{\eta \in \{0,1\}^d\setminus \{0^d\}}{\mathcal{C}}^{|\eta|}f_d\circ \pi^{(d)}_\eta$$ is measurable with respect to $$\bigvee_{\ell=1}^d {\Sigma}_{{\mathbf{Y}}^{(d)}}^{S^{(d)}_{(d-\ell;d]}} = {\Sigma}_{{\mathbf{Y}}^{(d)}}^{{\mathsf{F}}_{d,d}}.$$ Therefore the non-vanishing of the above integral implies that $$P^{{\mathsf{V}}_{d,d}}_{{\mathbf{Y}}^{(d)}} (f_d\circ \pi_{0^d}^{(d)}) \neq 0.$$ Since ${\mathbf{X}}$ is ${\mathsf{V}}_{d,d}$-sated, this implies that also $P^{{\mathsf{V}}_{d,d}}_{\mathbf{X}}f_d \neq 0$, as required. This proves that the required subspace is partially characteristic for $i=j=d$. *Step 2.*By Step 1, for any $f_1$, …, $f_d \in L^\infty(\mu)$, we have $${\widehat{{\Lambda}}}_n^{(d)}(f_1,\ldots,f_d) - {\widehat{{\Lambda}}}_n^{(d)}(f_1,\ldots,f_{d-1},P^{{\mathbf{V}}_{d,d}}_{\mathbf{X}}f_d) {\longrightarrow}0.$$ Also, $P^{{\mathbf{V}}_{d,d}}_{\mathbf{X}}f_d$ still lies in $L^\infty(\mu)$, because $P^{{\mathbf{V}}_{d,d}}_{\mathbf{X}}$ is actually a conditional expectation operator. It therefore suffices to check that the required factors are partially characteristic in the other positions under the additional assumption that $f_d$ is ${\Sigma}^{{\mathsf{F}}_{d,d}}_{\mathbf{X}}$-measurable. This assumption implies that $f_d$ may be approximated in $\|\cdot\|_2$ by a finite sum of products of the form $$\begin{aligned} \label{eq:product-decomp} h_0\cdot \cdots \cdot h_{d-1}\end{aligned}$$ where $h_i$ is ${\Sigma}_{\mathbf{X}}^{T_{(i;d]}}$-measurable for each $i$. By multi-linearity, it therefore suffices to prove that the required factors are partially characteristic in the other positions when $f_d$ is just one such product function. However, at this point a simple re-arrangement gives $$\begin{aligned} {\widehat{{\Lambda}}}_n^{(d)}(f_1,\ldots,f_{d-1},h_0\cdots h_{d-1}) &=& \frac{1}{|F_n|}\sum_{g \in F_n}\Big(\prod_{i=1}^{d-1}(f_i\circ T_{[1;i]}^g)\Big)\cdot ((h_0\cdots h_{d-1})\circ T_{[1;d]}^g)\\ &=& h_0\cdot\frac{1}{|F_n|}\sum_{g \in F_n}\prod_{i=1}^{d-1}((f_ih_i)\circ T_{[1;i]}^g)\\ &=& h_0\cdot {\widehat{{\Lambda}}}_n^{(d-1)}(f_1h_1,\ldots,f_{d-1}h_{d-1}),\end{aligned}$$ using the partial invariances of each of the $h_i$s. Therefore, by the inductive hypothesis for $j = d-1$, if these averages do not vanish as $n {\longrightarrow}\infty$ then $$P^{{\mathsf{V}}_{i,d-1}}_{\mathbf{X}}(f_ih_i) = {\mathsf{E}}_\mu\big(f_ih_i\,\big|\,{\Sigma}^{{\mathsf{F}}_{i,d-1}}_{\mathbf{X}}\big)\neq 0$$ for each $i \in [d-1]$. Since $h_i$ is ${\Sigma}_{\mathbf{X}}^{T_{(i;d]}}$-measurable, this implies that $${\mathsf{E}}_\mu\big(f_i\,\big|\,{\Sigma}^{{\mathsf{F}}_{i,d-1}}_{\mathbf{X}}\vee {\Sigma}^{T_{(i;d]}}_{\mathbf{X}}\big) = {\mathsf{E}}_\mu(f_i\,|\,{\Sigma}^{{\mathsf{F}}_{i,d}}_{\mathbf{X}}) \neq 0.$$ Arguing as at the start of Step 1, this implies that for the averages ${\widehat{{\Lambda}}}_n^{(d)}$, the subspace ${\mathsf{V}}_{i,d,{\mathbf{X}}}$ is partially characteristic in position $i$ for each $i\leq d-1$. Therefore the induction continues. The proof is by induction on $d$, and uses only the partially characteristic factor in position $d$. When $d=1$, convergence is given by the Norm Ergodic Theorem for amenable groups, so suppose $d\geq 2$, and that convergence is known for all commuting tuples of fewer than $d$ actions. By Theorem \[thm:sateds-exist\], we may ascend from ${\mathbf{X}}$ to an extension which is ${\mathsf{V}}_{d,d}$-sated, and so simply assume that ${\mathbf{X}}$ is itself ${\mathsf{V}}_{d,d}$-sated. This implies that $$\big\|{\Lambda}_n(f_1,\ldots,f_d) - {\Lambda}_n(f_1,\ldots,f_{d-1},P^{{\mathsf{V}}_{d,d}}_{\mathbf{X}}f_d)\big\|_2 {\longrightarrow}0\quad \hbox{as}\ n{\longrightarrow}\infty,$$ as in the proof of the previous proposition. It therefore suffices to prove convergence for the right-hand averages inside these norms. However, again as in the proof of the previous proposition, for these we may approximate ${\mathsf{E}}_\mu(f_d\,|\,{\Sigma}_{\mathbf{X}}^{{\mathsf{F}}_{d,d}})$ by a finite sum of products of the form in (\[eq:product-decomp\]), and then re-arrange the resulting averages into the form $$h_0\cdot\frac{1}{|F_n|}\sum_{g \in F_n}\prod_{i=1}^{d-1}((f_ih_i)\circ T_{[1;i]}^g).$$ Ignoring the factor $h_0$, which is uniformly bounded and does not depend on $n$, this is now a system of non-conventional averages for a commuting $(d-1)$-tuple of $G$-actions, so convergence follows by the inductive hypothesis. Proof of multiple recurrence ============================ Given a $G^d$-space, the convergence proved in Theorem A implies that the sequence of measures $${\lambda}_n := \frac{1}{|F_n|}\sum_{g \in F_n} \delta_{(T_1^gx,T_{[1;2]}^gx,\ldots,T_{[1;d]}^gx)}$$ converges in the usual topology on the convex set of $d$-fold couplings of $\mu$. (This is the same as the topology on joinings when one gives all spaces the action of the trivial group; the joining topology is explained, for instance, in [@Gla03 Section 6.1].) Letting ${\lambda}:= \lim_{n{\longrightarrow}\infty}{\lambda}_n$, it follows that for any measurable $A \subseteq X$ one has $$\frac{1}{|F_n|}\sum_{g \in F_n}\mu(T_1^gA\cap \cdots \cap T_{[1;d]}^gA) {\longrightarrow}{\lambda}(A^d).$$ To complete the proof of Theorem B, we will show that $$\begin{aligned} \label{eq:multirec} {\lambda}(A_1\times \cdots \times A_d) = 0 \quad \Longrightarrow \quad \mu(A_1\cap \cdots \cap A_d) = 0\end{aligned}$$ for any $G^d$-space ${\mathbf{X}}= (X,\mu,T)$ and measurable subsets $A_1$, …, $A_d \subseteq X$. This gives the desired conclusion by setting $A_1 := \ldots := A_d := A$. First, by replacing ${\mathbf{X}}$ with a suitable extension and lifting each $A_i$ to that extension, we may reduce this task to the case in which ${\mathbf{X}}$ is sated with respect to any chosen family of functional ${\sigma}$-subalgebras. After doing this, the result will be proved by making contact with a modification of Tao’s Infinitary Removal Lemma from [@Tao06--hyperreg]. This is the same strategy as in [@Aus--newmultiSzem]. The modification of the Removal Lemma is essentially as in [@Aus--newmultiSzem], but we shall use its more explicit formulation from [@Aus--thesis]: \[prop:infremove\] Let $(X,\mu)$ be a standard Borel probability space with ${\sigma}$-algebra ${\Sigma}$. Let $\pi_i:X^d {\longrightarrow}X$ be the coordinate projection for each $i\leq d$. Let $\theta$ be a $d$-fold coupling of $\mu$ on $X^d$. Finally, suppose that $(\Psi_e)_e$ is a collection of ${\sigma}$-subalgebras of ${\Sigma}$, indexed by $e \in \binom{[d]}{\geq 2}$, with the following properties: - if $a \subseteq e$ then $\Psi_a \supseteq \Psi_e$; - if $i,j \in e$ and $A \in \Psi_e$ then $\theta(\pi_i^{-1}(A)\triangle \pi_j^{-1}(A)) = 0$, so that we may let ${\widehat{\Psi}}_e$ denote the common $\theta$-completion of the lifted ${\sigma}$-algebras $\pi_i^{-1}(\Psi_e)$ for $i \in e$; - if we define ${\widehat{\Psi}}_{\mathcal{I}}:= \bigvee_{e \in {\mathcal{I}}}{\widehat{\Psi}}_e$ for each up-set ${\mathcal{I}}\subseteq \binom{[d]}{\geq 2}$, then the ${\sigma}$-algebras ${\widehat{\Psi}}_{\mathcal{I}}$ and ${\widehat{\Psi}}_{{\mathcal{J}}}$ are relatively independent under $\theta$ over ${\widehat{\Psi}}_{{\mathcal{I}}\cap {\mathcal{J}}}$. In addition, suppose that ${\mathcal{I}}_{i,j}$ for $i=1,2,\ldots,d$ and $j = 1,2,\ldots,k_i$ are up-sets in $\binom{[d]}{\geq 2}$ such that $[d] \in {\mathcal{I}}_{i,j}\subseteq \langle i\rangle$ for each $i,j$, and that $A_{i,j} \in \bigvee_{e \in {\mathcal{I}}_{i,j}}\Psi_e$ for each $i,j$. Then $$\theta\Big(\prod_{i=1}^d\Big(\bigcap_{j=1}^{k_i}A_{i,j}\Big)\Big) = 0 \quad \Longrightarrow \quad \mu\Big(\bigcap_{i=1}^d\bigcap_{j=1}^{k_i}A_{i,j}\Big) = 0.$$ This result is proved by a rather lengthy induction on the up-sets ${\mathcal{I}}_{i,j}$, which requires the full generality above: see [@Aus--thesis Subsection 4.3] for a proof and additional discussion. However, as in [@Aus--newmultiSzem], we will apply it only for $k_i = 1$ and ${\mathcal{I}}_{i,1} = \langle i \rangle$ for each $i$, in which case it asserts that, if $A_i \in \Psi_{\langle i\rangle}$ for each $i$, then $$\begin{aligned} \label{eq:spec-case} \theta(A_1\times\cdots \times A_d) = 0 \quad \Longrightarrow \quad \mu(A_1\cap \cdots \cap A_d) = 0.\end{aligned}$$ We will apply Proposition \[prop:infremove\] with $\theta$ equal to the limit coupling ${\lambda}$, and with the following family of ${\sigma}$-subalgebras. Suppose that $e = \{i_1 < \ldots < i_\ell\}\subseteq [d]$ is non-empty, and define a new functorial ${\sigma}$-subalgebra of $G^d$-spaces ${\mathbf{X}}$ by $$\begin{gathered} \Phi^e_{\mathbf{X}}:= {\Sigma}_{\mathbf{X}}^{L_e} = \big\{A \in {\Sigma}_{\mathbf{X}}\,\big|\ \mu(T^g_{(i_1;i_2]}A \triangle A) = \mu(T^g_{(i_2;i_3]}A\triangle A) =\\ \ldots = \mu(T^g_{(i_{\ell-1};i_\ell]}A\triangle A) = 0\ \forall g\in G\big\},\end{gathered}$$ where we interpret this as ${\Sigma}_{\mathbf{X}}$ in case $|e| = 1$. Observe that if $a \subseteq e$, then $L_a \leq L_e$, and so $\Phi^a_{\mathbf{X}}\supseteq \Phi^e_{\mathbf{X}}$. For any up-set ${\mathcal{I}}\subseteq \binom{[d]}{\geq 2}$, let $$\Phi^{\mathcal{I}}_{\mathbf{X}}:= \bigvee_{e \in {\mathcal{I}}}\Phi^e_{\mathbf{X}}.$$ Most of our remaining work will go into checking properties (i)–(iii) above for the joint distribution of the lifted ${\sigma}$-algebras $\pi_i^{-1}(\Phi_{\mathbf{X}}^e)$, subject to a certain satedness assumption on ${\mathbf{X}}$. The $G^d$-space ${\mathbf{X}}= (X,\mu,T)$ is **fully sated** if it is sated for every functorial ${\sigma}$-subalgebra of the form $$\bigvee_{s=1}^r\Phi^{e_s}_{\mathbf{X}}$$ for some $e_1$, …, $e_r \in \binom{[d]}{\geq 2}$. \[thm:struct-of-coupling\] Let ${\mathbf{X}}$ be fully sated, and let ${\lambda}$ be its limit coupling as above. 1. The coordinates factors $\pi_i:X^d{\longrightarrow}X$ are relatively independent under ${\lambda}$ over the further ${\sigma}$-subalgebras $\pi_i^{-1}(\Phi^{\langle i\rangle}_{\mathbf{X}})$, $i=1,2,\ldots,d$. 2. The collection $(\Phi^e_{\mathbf{X}})_e$ satisfies properties (i)–(iii) of Proposition \[prop:infremove\]. By passing to an extension as given by Corollary \[cor:mult-sateds-exist\], it suffices to prove Theorem B for fully sated $G^d$-spaces. Specifically, we will prove the implication (\[eq:multirec\]). Let $A_1$, …, $A_d \subseteq X$ be measurable. Part (1) of Theorem \[thm:struct-of-coupling\] gives $${\lambda}(A_1\times \cdots \times A_d) = \int_X f_1\otimes \cdots \otimes f_d \,{\mathrm{d}}{\lambda},$$ where $f_i := {\mathsf{E}}_\mu(1_{A_i}\,|\,\Phi_{\mathbf{X}}^{\langle i\rangle})$. Letting $B_i := \{f_i > 0\}$ for each $i$, it follows that $$\begin{gathered} {\lambda}(A_1\times \cdots \times A_d) = 0 \\ \Longrightarrow \quad {\lambda}\{f_1\otimes \cdots \otimes f_d > 0\} = {\lambda}(B_1\times \cdots \times B_d) = 0.\end{gathered}$$ On the other hand, each $B_i$ is $\Phi^{\langle i\rangle}_{\mathbf{X}}$-measurable. By Part (2) of Theorem \[thm:struct-of-coupling\], we may therefore apply Proposition \[prop:infremove\] in the form of the implication (\[eq:spec-case\]), to conclude $${\lambda}(B_1\times \cdots \times B_d) = 0 \quad \Longrightarrow \quad \mu(B_1\cap \cdots \cap B_d) = 0.$$ Since $$\mu(A_i\setminus B_i) = \int 1_{A_i}\cdot 1_{X\setminus B_i}\,{\mathrm{d}}\mu = \int f_i\cdot 1_{X\setminus B_i}\,{\mathrm{d}}\mu = 0$$ for each $i$, this completes the proof. The rest of this section will go into proving Theorem \[thm:struct-of-coupling\]. Subsection \[subs:jointdist1\] will establish various necessary joint-distribution properties of the ${\sigma}$-algebras $\Phi^e_{\mathbf{X}}$ in $(X,\mu)$ itself, and then Subsection \[subs:jointdist2\] will deduce the required properties of the lifts $\pi_i^{-1}(\Phi^e_{\mathbf{X}})$ from these. Joint distribution of some ${\sigma}$-algebras of invariant sets {#subs:jointdist1} ---------------------------------------------------------------- The next proposition will be the second major outing for satedness in this paper. It shows that if a $G^d$-space ${\mathbf{X}}$ is sated relative to a suitable family of functorial ${\sigma}$-subalgebras constructed out of the collection $\Phi^e_{{\mathbf{X}}}$, $e\subseteq [k]$, then this forces some relative independence among those ${\sigma}$-subalgebras. \[prop:first-rel-ind\] Suppose that $\{i < j\} \subseteq [d]$, suppose that $e_1$, …, $e_r \in \binom{[d]}{\geq 2}$, and let ${\mathbf{X}}= (X,\mu,T_1,\ldots,T_d)$ be a $G^d$-space. 1. If $e_s\cap [i;j) = \{i\}$ for every $s \leq r$, and if ${\mathbf{X}}$ is ${\mathsf{F}}$-sated for $${\Sigma}^{\mathsf{F}}_{\mathbf{X}}:= \bigvee_{s=1}^r\Phi_{{\mathbf{X}}}^{e_s\cup \{j\}},$$ then $$\Phi_{\mathbf{X}}^{\{i,j\}} \quad \hbox{and} \quad \bigvee_{s=1}^r \Phi^{e_s}_{\mathbf{X}}\quad \hbox{are rel. indep. over} \quad {\Sigma}^{\mathsf{F}}_{\mathbf{X}}.$$ 2. If $e_s\cap (i;j] = \{j\}$ for every $s \leq r$, and if ${\mathbf{X}}$ is ${\mathsf{G}}$-sated for $${\Sigma}^{\mathsf{G}}_{{\mathbf{X}}} := \bigvee_{s=1}^r\Phi_{{\mathbf{X}}}^{e_s\cup \{i\}},$$ then $$\Phi_{\mathbf{X}}^{\{i,j\}} \quad \hbox{and} \quad \bigvee_{s=1}^r \Phi^{e_s}_{\mathbf{X}}\quad \hbox{are rel. indep. over} \quad {\Sigma}^{\mathsf{G}}_{\mathbf{X}}.$$ As always, the appeal to satedness will depend on constructing the right extension. Let $e := e_1 \cup \ldots \cup e_r \cup \{j\}$, so that our assumptions give $e\cap [i;j] = \{i,j\}$. *Step 1.*We first construct a suitable extension of the $L_e$-subaction ${\mathbf{X}}^{{\!\upharpoonright}L_e}$. Since $e \cap [i;j] = \{i,j\}$, Corollary \[cor:some-invce\] promises that $\Phi_{\mathbf{X}}^{\{i,j\}}$ is globally $L_e$-invariant, so the relative product measure $\nu:=\mu\otimes_{\Phi_{\mathbf{X}}^{\{i,j\}}}\mu$ is invariant under the diagonal action of $L_e$ on $Y := X^2$. We will make use of this by constructing a *non*-diagonal action of $L_e$ on $Y$ which still preserves $\nu$. Let $e$ be enumerated as $\{i_1 < \ldots < i_m\}$. Our assumptions imply that $\{i,j\} = \{i_{\ell_0},i_{\ell_0+1}\}$ for some $\ell_0 \leq m-1$. In these terms, $L_e$ is generated by its $m-1$ commuting subgroups $$L_{\{i_\ell,i_{\ell+1}\}} = {\varphi}_\ell(G),\quad \ell = 1,\ldots,m-1,$$ where ${\varphi}_\ell:G{\longrightarrow}G^d$ is the injective homomorphism defined by $$\big({\varphi}_\ell(g)\big)_i := \left\{\begin{array}{ll}g &\quad \hbox{if}\ i \in (i_\ell;i_{\ell+1}] \\ e &\quad \hbox{else}.\end{array}\right.$$ Specifying an action of $L_e$ is equivalent to specifying commuting actions of its subgroups ${\varphi}_\ell(G)$. We define our new, non-diagonal action $S:L_e{\curvearrowright}(Y,\nu)$ as follows: $$S^{{\varphi}_\ell(g)} := \left\{\begin{array}{ll}T_{(i_\ell;i_{\ell+1}]}^g\times T_{(i_\ell,i_{\ell+1}]}^g \quad \hbox{(i.e., diagonal)} &\quad \hbox{if}\ \ell \not\in \{\ell_0,\ell_0+1\}\\ T_{(i_{\ell_0};i_{\ell_0+1}]}^g\times {\mathrm{id}} &\quad \hbox{if}\ \ell = \ell_0\\ T_{(i_{\ell_0+1};i_{\ell_0+2}]}^g\times T_{(i_{\ell_0},i_{\ell_0+2}]}^g &\quad \hbox{if}\ \ell = \ell_0+1\end{array}\right.$$ (where the last option here is vacuous in case $\ell_0 = m-1$). Each of these transformations leaves $\nu$ invariant: we have already remarked this for the diagonal transformations, and for the last two possibilities we need only observe that $\nu$ is a relative product over a ${\sigma}$-algebra on which $T^g_{(i_{\ell_0};i_{\ell_0+1}]} = T^g_{(i;j]}$ acts trivially for all $g \in G$. Now let ${\beta}_1,{\beta}_2:Y{\longrightarrow}X$ be the two coordinate projections. The above definition gives ${\beta}_1\circ S^{{\varphi}_\ell(g)} = T^{{\varphi}_\ell(g)}\circ {\beta}_1$ for every $\ell$ and $g$, so letting ${\mathbf{Y}}$ denote the $L_e$-space given by the above transformations $S$ on $(Y,\nu)$, we have a factor map ${\mathbf{Y}}\stackrel{{\beta}_1}{{\longrightarrow}} {\mathbf{X}}^{{\!\upharpoonright}L_e}$. On the other hand, recalling that $\{i_{\ell_0},i_{\ell_0+1}\} = \{i,j\}$, the above definitions give that under $S$, the subgroup $L_{\{i,j\}} \leq L_e$ acts trivially on the second coordinate in $Y$. *Step 2.*Next, applying Theorem \[thm:recoverG\] enlarges this to an extension ${\mathbf{X}}_1\stackrel{\pi}{{\longrightarrow}} {\mathbf{X}}$ of $G^d$-spaces which factorizes through some $L_e$-extension ${\mathbf{X}}_1^{{\!\upharpoonright}L_e}\stackrel{{\alpha}}{{\longrightarrow}} {\mathbf{Y}}$. *Step 3.*Now suppose that $f \in L^\infty(\mu|_{\Phi_{\mathbf{X}}^{\{i,j\}}})$ and that $g_s \in L^\infty(\mu|_{\Phi^{e_s}_{\mathbf{X}}})$ for each $s \leq r$. From the definition of $\nu$ and the fact that $f$ is $\Phi^{\{i,j\}}_{\mathbf{X}}$-measurable, we have $$\begin{gathered} \label{eq:move-int} \int_X f\cdot \prod_{s \leq r}g_s\,{\mathrm{d}}\mu = \int_Y (f\circ {\beta}_2)\cdot \Big(\prod_{s \leq r}g_s\circ {\beta}_2\Big)\,{\mathrm{d}}\nu\\ = \int_Y (f\circ{\beta}_1)\cdot \Big(\prod_{s\leq r}g_s\circ {\beta}_2\Big)\,{\mathrm{d}}\nu.\end{gathered}$$ Consider the function $g_s\circ {\beta}_2$ for some $s\leq r$. Our next step is to show that it is invariant under the whole of $S^{{\!\upharpoonright}L_{e_s\cup \{j\}}}$. For a given $s$, if we enumerate $e_s \cup \{j\} =: \{p_1 < \ldots < p_n\}$, then it suffices to prove invariance under each subgroup $L_{\{p_k,p_{k+1}\}}$ for $k \in \{1,2,\ldots,n-1\}$. There are three cases to consider. Firstly, if $p_k \not\in \{i,j\}$, then, since $e\cap [i;j) = \{i\}$, it follows that $(p_k;p_{k+1}]$ is disjoint from $(i_{\ell_0};i_{\ell_0+2}]$ (using again the notation from Step 1). In this case the definition of $S$ gives ${\beta}_2 \circ S_{(p_k;p_{k+1}]} = T_{(p_k;p_{k+1}]}\circ {\beta}_2$, so the required invariance follows from the fact that $g_s$ itself is $L_{e_s}$-invariant. Second, if $p_k = i$, then the assumption $e_s \cap [i;j) = \{i\}$ implies that $p_{k+1} = j$, and hence the definition of $S$ implies that ${\beta}_2 \circ S_{(p_k;p_{k+1}]} = {\beta}_2$, from which the $S_{(p_k;p_{k+1}]}$-invariance of $g_s\circ {\beta}_2$ is obvious. Finally, if $p_k = j$, then the definition of $S$ gives $${\beta}_2\circ S_{(p_k;p_{k+1}]} = T_{(i;p_{k+1}]}\circ {\beta}_2.$$ Since $\{i,p_{k+1}\} \subseteq e_s$ (even if $j\not\in e_s$), once again the $L_{e_s}$-invariance of $g_s$ gives $$g_s\circ {\beta}_2\circ S^g_{(p_k;p_{k+1}]} = g_s\circ T^g_{(i;p_{k+1}]}\circ {\beta}_2 = g_s\circ {\beta}_2,$$ as required. *Step 4.*In light of Step 3, the function $\prod_{s \leq r}(g_s\circ {\beta}_2\circ {\alpha})$ is measurable with respect to $$\bigvee_{s\leq r}\Phi_{{\mathbf{X}}_1}^{e_s\cup \{j\}} = {\Sigma}^{\mathsf{F}}_{{\mathbf{X}}_1}.$$ By the assumed ${\mathsf{F}}$-satedness, it follows that the right-hand integral in (\[eq:move-int\]) is equal to $$\int_Y ({\mathsf{E}}_\mu(f\,|\ {\Sigma}_{\mathbf{X}}^{\mathsf{F}})\circ{\beta}_1)\cdot \Big(\prod_{s\leq r}g_s\circ {\beta}_2\Big)\,{\mathrm{d}}\nu,$$ and by the same reasoning that gave (\[eq:move-int\]) itself this is equal to $$\int_X {\mathsf{E}}_\mu(f\,|\ {\Sigma}_{\mathbf{X}}^{\mathsf{F}})\cdot \prod_{s \leq r}g_s\,{\mathrm{d}}\mu = \int_X {\mathsf{E}}_\mu(f\,|\ {\Sigma}_{\mathbf{X}}^{\mathsf{F}})\cdot {\mathsf{E}}_\mu\Big(\prod_{s \leq r}g_s\,\Big|\,{\Sigma}^{\mathsf{F}}_{\mathbf{X}}\Big)\,{\mathrm{d}}\mu.$$ Since $f$ and each $g_s$ were arbitrary subject to their measurability assumptions, this implies that $\Phi_{\mathbf{X}}^{\{i,j\}}$ and $\bigvee_{s=1}^r\Phi_{\mathbf{X}}^{e_s}$ are relatively independent over ${\Sigma}^{\mathsf{F}}_{\mathbf{X}}$. This follows exactly the same steps as Part 1, except that now the new $L_e$-action $S$ on $$(Y,\nu) := (X^2,\mu\otimes_{\Phi^{\{i,j\}}_{\mathbf{X}}}\mu)$$ is defined as follows: $$S^{{\varphi}_\ell(g)} := \left\{\begin{array}{ll}T_{(i_\ell;i_{\ell+1}]}^g\times T_{(i_\ell,i_{\ell+1}]}^g \quad \hbox{(i.e., diagonal)} &\quad \hbox{if}\ \ell \not\in \{\ell_0-1,\ell_0\}\\ T_{(i_{\ell_0};i_{\ell_0+1}]}^g\times {\mathrm{id}} &\quad \hbox{if}\ \ell = \ell_0\\ T_{(i_{\ell_0-1};i_{\ell_0}]}^g\times T_{(i_{\ell_0-1},i_{\ell_0+1}]}^g &\quad \hbox{if}\ \ell = \ell_0-1,\end{array}\right.$$ where now $e := e_1\cup \cdots \cup e_r\cup \{i\} = \{i_1 < \ldots < i_m\}$, and $\ell_0$ is such that $\{i,j\} = \{i_{\ell_0},i_{\ell_0 + 1}\}$, as before. The next two propositions contain the consequences of full satedness that we need. The first modifies the conclusion of Proposition \[prop:pleasant\] in the case of integrated averages, rather than functional averages. \[prop:char-factors-again\] Suppose that ${\mathbf{X}}$ is fully sated, that $\{i_1 < \ldots < i_k\} \subseteq [d]$ and that $f_1$, …, $f_k \in L^\infty(\mu)$. Then $$\begin{gathered} \lim_{n{\longrightarrow}\infty}\int_X \frac{1}{|F_n|}\sum_{g \in F_n}\prod_{j=1}^k (f_j\circ T_{[1;i_j]}^g)\,{\mathrm{d}}\mu\\ = \lim_{n{\longrightarrow}\infty}\int_X \frac{1}{|F_n|}\sum_{g \in F_n}\prod_{j=1}^k ({\mathsf{E}}_\mu(f_j\,|\,\Delta_j)\circ T_{[1;i_j]}^g)\,{\mathrm{d}}\mu,\end{gathered}$$ where $$\Delta_j := \bigvee_{\ell = 1}^{j-1}{\Sigma}^{T_{(i_\ell;i_j]}}_{\mathbf{X}}\vee \bigvee_{\ell = j+1}^k{\Sigma}^{T_{(i_j;i_\ell]}}_{\mathbf{X}}.$$ This is proved by induction on $k$. When $k=1$ it is trivial, by the $T_{[1;i_1]}$-invariance of $\mu$, so suppose $k\geq 2$. Because $\mu$ is $T_{[1;i_1]}$-invariant, the desired conclusion is equivalent to $$\begin{gathered} \int_X f_1\cdot \Big(\lim_{n{\longrightarrow}\infty}\frac{1}{|F_n|}\sum_{g \in F_n}\prod_{j=2}^k (f_j\circ T_{(i_1;i_j]}^g)\Big)\,{\mathrm{d}}\mu\\ = \int_X {\mathsf{E}}_\mu(f_1\,|\,\Delta_1)\cdot \Big(\lim_{n{\longrightarrow}\infty}\frac{1}{|F_n|}\sum_{g \in F_n}\prod_{j=2}^k ({\mathsf{E}}_\mu(f_j\,|\,\Delta_j)\circ T_{(i_1;i_j]}^g)\Big)\,{\mathrm{d}}\mu.\end{gathered}$$ However, ${\mathbf{X}}$ being fully sated implies that the $G^{k-1}$-space ${\mathbf{X}}'$ defined by $$T'_j := T_{(i_j,i_{j+1}]} \quad \hbox{for}\ j=1,2,\ldots,k-1$$ is also fully sated: otherwise, we could turn a $G^{k-1}$-extension witnessing the failure of satedness for ${\mathbf{X}}'$ back into a $G^d$-extension of ${\mathbf{X}}$ using Theorem \[thm:recoverG\]. Therefore Proposition \[prop:pleasant\] applied to this $G^d$-space gives $$\begin{gathered} \int_X f_1\cdot \Big(\lim_{n{\longrightarrow}\infty}\frac{1}{|F_n|}\sum_{g \in F_n}\prod_{j=2}^k (f_j\circ T_{(i_1;i_j]}^g)\Big)\,{\mathrm{d}}\mu\\ = \int_X f_1\cdot \Big(\lim_{n{\longrightarrow}\infty}\frac{1}{|F_n|}\sum_{g \in F_n}\prod_{j=2}^k ({\mathsf{E}}_\mu(f_j\,|\,\Delta_j)\circ T_{(i_1;i_j]}^g)\Big)\,{\mathrm{d}}\mu,\end{gathered}$$ since the ${\sigma}$-algebras $\Delta_j$ for $j\geq 2$ are those that arise by applying the functorial ${\sigma}$-subalgebras ${\mathsf{F}}_{\bullet,\bullet}$ of that proposition to the $G^{k-1}$-space ${\mathbf{X}}'$. This almost completes the proof. To finish, observe that, as in the proof of Theorem A, we may now approximate $f_k$ (say) by a finite sum of finite products of functions measurable with respect to ${\Sigma}_{\mathbf{X}}^{T_{(i_\ell,i_k]}}$ for $\ell=1,\ldots,k-1$, and having done so we may re-arrange the above into an analogous system of averages with only $k-1$ transformations. At that point, the inductive hypothesis allows us to replace $f_1$ with ${\mathsf{E}}_\mu(f_1\,|\,\Delta_1)$, completing the proof. \[prop:next-rel-ind\] Suppose that ${\mathbf{X}}$ is fully sated, and that $e_0$, $e_1$, …, $e_r \in \binom{[d]}{\geq 2}$ are sets such that $\bigcap_{0 \leq s\leq r}e_s \neq \emptyset$. Let $g$ be $\Phi^{e_0}_{\mathbf{X}}$-measurable and $f_s$ be $\Phi^{e_s}_{\mathbf{X}}$-measurable for $1 \leq s \leq r$, and suppose $g$ and every $f_s$ are bounded. Then $$\int_X g\cdot f_1\cdot\cdots \cdot f_r\,{\mathrm{d}}\mu = \int_X{\mathsf{E}}_\mu(g\,|\,\Psi)\cdot f_1\cdot \cdots \cdot f_r\,{\mathrm{d}}\mu,$$ where $$\Psi:= \bigvee_{s=1}^r \Phi^{e_0\cup e_s}_{\mathbf{X}}.$$ As is standard, this is equivalent to the assertion that $${\mathsf{E}}_\mu\Big(g\,\Big|\,\bigvee_{s=1}^r\Phi^{e_s}_{\mathbf{X}}\Big) = {\mathsf{E}}_\mu(g\,|\,\Psi).$$ If $e_0 = e_1 = e_2 = \cdots = e_r$, then $\Phi^{e_s}_{\mathbf{X}}= \Psi$ for all $s$, so the result is trivial. The general case is proved by an outer induction on $r$, and an inner induction on the cardinality of the up-set $${\mathcal{A}}_{e_0,e_1,\ldots,e_r} := \langle e_0\rangle \cup \cdots \cup \langle e_r\rangle.$$ The base clause corresponds to $r=1$ and $|{\mathcal{A}}_{e_0,e_1}| = 1$, hence $e_0 = e_1 = [k]$: this is among the trivial cases described above. For the recursion clause, we may assume that there are at least two distinct sets among the $e_s$ for $0\leq s \leq r$, so $\bigcap_{0 \leq s \leq r}e_s \neq \bigcup_{0 \leq s \leq r}e_s$ (else we would be in the trivial case treated above). We have also assumed that $\bigcap_{0 \leq s \leq r}e_s \neq \emptyset$, so there must be $\{i < j\} \subseteq [k]$ such that $e_s \cap [i+1;j) = \emptyset$ for all $0 \leq s\leq r$, and - either $i \in e_s$ for all $0 \leq s \leq r$, but $j$ lies in some $e_s$ but not all; - or $j \in e_s$ for all $0 \leq s \leq r$, but $i$ lies in some $e_s$ but not all. We will complete the induction in the first of these cases, the second being exactly analogous. In this first case we have $e_s\cap [i;j) = \{i\}$ for all $s$. It now breaks into two further sub-cases. *Case 1.*First assume that $j \in e_0$. Then, since $\Phi^{e_0}_{\mathbf{X}}\subseteq \Phi^{\{i,j\}}_{\mathbf{X}}$, Part 1 of Proposition \[prop:first-rel-ind\] gives that $$\int_X g\cdot (f_1\cdot \cdots \cdot f_r)\,{\mathrm{d}}\mu = \int_X{\mathsf{E}}_\mu\Big(g\,\Big|\,\bigvee_{s=1}^r\Phi^{e_s\cup \{j\}}_{\mathbf{X}}\Big)\cdot (f_1\cdot \cdots \cdot f_r)\,{\mathrm{d}}\mu.$$ Now observe that $${\mathcal{A}}_{e_0,e_1\cup\{j\},\ldots,e_r\cup \{j\}} \subseteq {\mathcal{A}}_{e_0,\ldots,e_r},$$ and this inclusion is strict, because the right-hand family contains some set that does not contain $j$, whereas every element of the left-hand family does contain $j$. Therefore we may apply the inductive hypothesis to $e_0$ and $e_1\cup \{j\}$, $e_2\cup \{j\}$, …, $e_r\cup\{j\}$, to conclude that $${\mathsf{E}}_\mu\Big(g\,\Big|\,\bigvee_{s=1}^r\Phi^{e_s\cup \{j\}}_{\mathbf{X}}\Big) = {\mathsf{E}}_\mu\Big(g\,\Big|\,\bigvee_{s=1}^r\Phi^{e_0 \cup e_s}_{\mathbf{X}}\Big) = {\mathsf{E}}_\mu(g\,|\,\Psi).$$ *Case 2.*Now assume that $j \not\in e_0$. By re-labeling the other sets if necessary, we may assume $j \in e_1$. Then the argument used in Case 1 gives $$\int_X g\cdot f_1\cdot\cdots \cdot f_r\,{\mathrm{d}}\mu = \int_X g\cdot {\mathsf{E}}_\mu(f_1\,|\,\Psi_1)\cdot f_2 \cdot \cdots \cdot f_r\,{\mathrm{d}}\mu,$$ where $$\Psi_1:= \bigvee_{s\in\{0\}\cup [r]\setminus \{1\}}\Phi^{e_1\cup e_s}_{\mathbf{X}},$$ and similarly $$\int_X {\mathsf{E}}_\mu(g\,|\,\Psi)\cdot f_1\cdot\cdots \cdot f_r\,{\mathrm{d}}\mu = \int_X {\mathsf{E}}_\mu(g\,|\,\Psi)\cdot {\mathsf{E}}_\mu(f_1\,|\,\Psi_1)\cdot f_2\cdot \cdots \cdot f_r\,{\mathrm{d}}\mu.$$ It therefore suffices to prove the desired equality when $f_1$ is $\Psi_1$-measurable. Since such an $f_1$ can be approximated in $\|\cdot\|_2$ by finite sums of products of $\Phi^{e_1\cup e_s}_{\mathbf{X}}$-measurable functions for $s \in \{0\}\cup [r]\setminus \{1\}$, it suffices furthermore to assume that $$f_1 = f_{10}\cdot f_{12}\cdot \cdots \cdot f_{1r}$$ is one such product. However, now the integral of interest may be written as $$\int_X (g f_{10})\cdot (f_2f_{12})\cdot \cdots \cdot (f_rf_{1r})\,{\mathrm{d}}\mu,$$ and by the inductive hypothesis on $r$ this is equal to $$\int_X {\mathsf{E}}_\mu(g f_{10}\,|\,\Psi_2)\cdot (f_2f_{12})\cdot \cdots \cdot (f_rf_{1r})\,{\mathrm{d}}\mu$$ with $$\Psi_2:= \bigvee_{s=2}^r\Phi^{e_0\cup e_s}_{\mathbf{X}}.$$ Since $\Psi \geq \Psi_2$ and $f_{10}$ is $\Psi$-measurable, the Law of Iterated Conditional Expectation gives $${\mathsf{E}}_\mu(g f_{10}\,|\,\Psi_2) = {\mathsf{E}}_\mu({\mathsf{E}}_\mu(g\,|\,\Psi) f_{10}\,|\,\Psi_2).$$ Substituting this back into the integral completes the desired equality. \[cor:second-rel-ind\] Suppose that ${\mathbf{X}}$ is fully sated and that $e_1$, …, $e_r \in \binom{[d]}{\geq 2}$ are such that $\bigcap_{s\leq r}e_s \neq \emptyset$. Let $$\Psi_s:= \bigvee_{t \in [r]\setminus s} \Phi^{e_t\cup e_s}_{\mathbf{X}}\quad \hbox{for}\ s=1,2,\ldots,r,$$ and let $f_s$ be $\Phi^{e_s}_{\mathbf{X}}$-measurable for each $s$. Then $$\int_X \prod_{s=1}^r f_s\ {\mathrm{d}}\mu = \int_X\prod_{s=1}^r{\mathsf{E}}_\mu(f_s\,|\,\Psi_s)\ {\mathrm{d}}\mu.$$ Simply apply Proposition \[prop:next-rel-ind\] to each factor of the integrand in turn. Structure of the limit couplings {#subs:jointdist2} -------------------------------- We now turn to the structure of the limit coupling ${\lambda}$, as discussed at the beginning of this section. The following lemma and definition are again essentially copied from [@Aus--newmultiSzem]. \[lem:diag\] If $i,j \in e \in \binom{[d]}{\geq 2}$ and $A \in \Phi^e_{\mathbf{X}}$, then $${\lambda}(\pi_i^{-1}(A)\triangle \pi_j^{-1}(A)) = 0.$$ In particular, $\pi_i^{-1}(\Phi^e_{\mathbf{X}})$ and $\pi_j^{-1}(\Phi_{\mathbf{X}}^e)$ agree up to ${\lambda}$-negligible sets. If $i = j$ then this is trivial, so assume without loss of generality that $i < j$. By definition, $$\begin{aligned} {\lambda}(\pi_i^{-1}(A)\cap \pi_j^{-1}(A)) &=& \lim_{n{\longrightarrow}\infty}\frac{1}{|F_n|}\sum_{g \in F_n} \mu(T_{[1;i]}^{g^{-1}}A \cap T_{[1;j]}^{g^{-1}}A)\\ &=& \lim_{n{\longrightarrow}\infty}\frac{1}{|F_n|}\sum_{g \in F_n} \mu(T_{[1;i]}^{g^{-1}}A \cap T_{[1;i]}^{g^{-1}}T_{(i;j]}^{g^{-1}}A)\\ &=& \lim_{n{\longrightarrow}\infty}\frac{1}{|F_n|}\sum_{g \in F_n} \mu(T_{[1;i]}^{g^{-1}}A \cap T_{[1;i]}^{g^{-1}}A) = \mu(A),\end{aligned}$$ because $\Phi^e_{\mathbf{X}}\leq \Phi^{\{i,j\}}_{\mathbf{X}}$, so $A$ is $T_{(i;j]}$-invariant. Therefore $${\lambda}(\pi_i^{-1}(A)\cap \pi_j^{-1}(A)) = {\lambda}(\pi_i^{-1}(A)) = {\lambda}(\pi_j^{-1}(A)),$$ and so ${\lambda}(\pi_i^{-1}(A)\triangle \pi_j^{-1}(A)) = 0$. In the setting above, the common ${\lambda}$-completion of the lifted ${\sigma}$-algebras $$\pi_i^{-1}(\Phi^e_{\mathbf{X}}) \quad \hbox{for}\ i \in e$$ is called the **oblique copy** of $\Phi^e_{\mathbf{X}}$. It will be denoted by ${\widehat{\Phi}}^e_{\mathbf{X}}$. If ${\mathcal{I}}$ is a non-empty up-set in $\binom{[d]}{\geq 2}$ and ${\mathbf{X}}$ is a $G^d$-space, then the ${\mathcal{I}}$-oblique ${\sigma}$-algebra is $${\widehat{\Phi}}^{{\mathcal{I}}}_{\mathbf{X}}:= \bigvee_{e \in {\mathcal{I}}}{\widehat{\Phi}}^e_{\mathbf{X}}.$$ The remainder of the proof of Theorem B follows almost exactly the same lines as [@Aus--thesis Subsection 4.2]; we include the following lemma again for completeness. \[lem:rel-ind\] If ${\mathcal{I}}$ and ${\mathcal{J}}$ are non-empty up-sets in $\binom{[d]}{\geq 2}$ and ${\mathbf{X}}$ is a fully sated $G^d$-space, then ${\widehat{\Phi}}^{{\mathcal{I}}}_{\mathbf{X}}$ and ${\widehat{\Phi}}^{{\mathcal{J}}}_{\mathbf{X}}$ are relatively independent over ${\widehat{\Phi}}^{{\mathcal{I}}\cap {\mathcal{J}}}_{\mathbf{X}}$ under ${\lambda}$. *Step 1.* Suppose first that ${\mathcal{J}}= \langle e\rangle$ where $e$ is a maximal member of $\binom{[d]}{\geq 2}\setminus {\mathcal{I}}$. Let $\{a_1,a_2,\ldots,a_m\}$ be the antichain of minimal elements of ${\mathcal{I}}$, so that ${\widehat{\Phi}}^{\mathcal{I}}_{\mathbf{X}}= \bigvee_{k\leq m}{\widehat{\Phi}}^{a_k}_{\mathbf{X}}$. The maximality assumption on $e$ implies that $e\cup \{j\}$ contains some $a_k$ for every $j \in [d]\setminus e$, and so ${\mathcal{I}}\cap {\mathcal{J}}$ is precisely the up-set generated by these sets $e\cup \{j\}$ for $j\in[d]\setminus e$. We must therefore show that ${\widehat{\Phi}}^e_{\mathbf{X}}$ is relatively independent from $\bigvee_{k\leq m}{\widehat{\Phi}}^{a_k}_{\mathbf{X}}$ under ${\lambda}$ over the ${\sigma}$-subalgebra $\bigvee_{j\in [d]\setminus e}{\widehat{\Phi}}^{e\cup\{j\}}_{\mathbf{X}}$. Since $e \not\in {\mathcal{I}}$ we can find some $j_k \in a_k\setminus e$ for each $k\leq m$. Moreover, each $j\in [d]\setminus e$ must appear as some $j_k$ in this list, since it appears for any $k$ for which $a_k \subseteq e \cup \{j\}$. Now Lemma \[lem:diag\] implies that ${\widehat{\Phi}}^{a_k}_{\mathbf{X}}$ agrees with $\pi_{j_k}^{-1}(\Phi^{a_k}_{\mathbf{X}})$ up to ${\lambda}$-negligible sets. On the other hand, we clearly have $\pi_{j_k}^{-1}(\Phi^{a_k}_{\mathbf{X}})\leq \pi_{j_k}^{-1}({\Sigma}_{\mathbf{X}})$, and so in fact it will suffice to show that ${\widehat{\Phi}}^e_{\mathbf{X}}$ is relatively independent from $\bigvee_{j\in [d]\setminus e}\pi_j^{-1}({\Sigma}_{\mathbf{X}})$ over $\bigvee_{j\in [d]\setminus e}{\widehat{\Phi}}^{e\cup\{j\}}_{\mathbf{X}}$. Picking $i \in e$, Lemma \[lem:diag\] also gives that ${\widehat{\Phi}}^e_{\mathbf{X}}$ agrees with $\pi_i^{-1}(\Phi^e_{\mathbf{X}})$ up to ${\lambda}$-negligible sets. On the other hand, for any sets $$A_j \in \pi_j^{-1}({\Sigma}_{\mathbf{X}})\ \hbox{for}\ j \in [d]\setminus e\ \hbox{and}\ B \in \Phi^e_{\mathbf{X}},$$ the definition of ${\lambda}$ gives $${\lambda}\Big(\Big(\bigcap_{j \in [d]\setminus e}\pi_j^{-1}(A_j)\Big)\cap \pi_i^{-1}(B)\Big) = \lim_{n{\longrightarrow}\infty}\int_X {\Lambda}_n(f_1,\ldots,f_d)\,{\mathrm{d}}\mu$$ with $$f_\ell := \left\{\begin{array}{ll}1_{A_\ell} & \quad \hbox{if}\ \ell \in [d]\setminus e\\ 1_B & \quad \hbox{if}\ \ell = i\\ 1 & \quad\hbox{else.} \end{array}\right.$$ By Proposition \[prop:char-factors-again\], one obtains the same limit from $$\lim_{n{\longrightarrow}\infty}\int_X {\Lambda}_n(f_1',\ldots,f_d')\,{\mathrm{d}}\mu$$ with $$f'_\ell := \left\{\begin{array}{ll}1_{A_\ell} & \quad \hbox{if}\ \ell \in [d]\setminus e \\ {\mathsf{E}}_\mu(1_B\,|\,\Delta) & \quad \hbox{if}\ \ell = i\\ 1 & \quad\hbox{else,} \end{array}\right.$$ and where in turn $$\Delta := \bigvee_{\ell \in [d]\setminus e,\,\ell < i}{\Sigma}^{T_{(\ell;i]}}_{\mathbf{X}}\vee \bigvee_{\ell \in [d]\setminus e,\, \ell > i}{\Sigma}^{T_{(i;\ell]}}_{\mathbf{X}}= \bigvee_{j \in [d]\setminus e}\Phi^{\{i,j\}}_{\mathbf{X}}.$$ This implies that $\pi_i^{-1}(\Phi^e_{\mathbf{X}})$ is relatively independent from $\bigvee_{j\in [d]\setminus e}\pi_j^{-1}({\Sigma}_{\mathbf{X}})$ over $\pi_i^{-1}(\Delta)$ under ${\lambda}$. On the other hand, Corollary \[cor:second-rel-ind\] gives that $\Phi^e_{\mathbf{X}}$ is relatively independent from $\Delta$ over $\bigvee_{j\in [d]\setminus e}\Phi^{e\cup\{j\}}_{\mathbf{X}}$. Combining these conclusions completes the proof in this case. *Step 2.*The general case can now be treated for fixed ${\mathcal{I}}$ by induction on ${\mathcal{J}}$. If ${\mathcal{J}}\subseteq {\mathcal{I}}$ then the result is clear, so now let $e$ be a minimal member of ${\mathcal{J}}\setminus{\mathcal{I}}$ of maximal size, and let ${\mathcal{K}} := {\mathcal{J}}\setminus\{e\}$. It will suffice to prove that if $F \in L^\infty({\lambda})$ is ${\widehat{\Phi}}^{\mathcal{J}}_{\mathbf{X}}$-measurable, then $${\mathsf{E}}_{\lambda}(F\,|\,{\widehat{\Phi}}^{\mathcal{I}}_{\mathbf{X}}) = {\mathsf{E}}_{\lambda}(F\,|\,{\widehat{\Phi}}^{{\mathcal{I}}\cap{\mathcal{J}}}_{\mathbf{X}}).$$ Furthermore, by an approximation in $\|\cdot\|_2$ by finite sums of products, it suffices to prove this only for $F$ that are of the form $F_1\cdot F_2$ with $F_1$ and $F_2$ being bounded and respectively ${\widehat{\Phi}}^{\langle e \rangle}_{\mathbf{X}}$- and ${\widehat{\Phi}}^{{\mathcal{K}}}_{\mathbf{X}}$-measurable. However, for such a product we can write $${\mathsf{E}}_{\lambda}(F\,|\,{\widehat{\Phi}}^{{\mathcal{I}}}_{\mathbf{X}}) = {\mathsf{E}}_{\lambda}\big({\mathsf{E}}_{\lambda}(F\,|\,{\widehat{\Phi}}^{{\mathcal{I}}\cup{\mathcal{K}}}_{\mathbf{X}})\,\big|\,{\widehat{\Phi}}^{{\mathcal{I}}}_{\mathbf{X}}\big) = {\mathsf{E}}_{\lambda}\big({\mathsf{E}}_{\lambda}(F_1\,|\,{\widehat{\Phi}}^{{\mathcal{I}}\cup{\mathcal{K}}}_{\mathbf{X}})\cdot F_2\,\big|\,{\widehat{\Phi}}^{{\mathcal{I}}}_{\mathbf{X}}\big).$$ By Step 1 we have $${\mathsf{E}}_{\lambda}(F_1\,|\,{\widehat{\Phi}}^{{\mathcal{I}}\cup{\mathcal{K}}}_{\mathbf{X}}) = {\mathsf{E}}_{\lambda}(F_1\,|\,{\widehat{\Phi}}^{({\mathcal{I}}\cup{\mathcal{K}})\cap\langle e\rangle}_{\mathbf{X}}),$$ and on the other hand $({\mathcal{I}}\cup{\mathcal{K}})\cap\langle e\rangle \subseteq {\mathcal{K}}$ (because ${\mathcal{K}}$ contains every subset of $[d]$ that strictly includes $e$, since ${\mathcal{J}}$ is an up-set). Therefore $({\mathcal{I}}\cup{\mathcal{K}})\cap\langle e\rangle = {\mathcal{K}}\cap \langle e\rangle$ and therefore another appeal to Step 1 gives $${\mathsf{E}}_{\lambda}(F_1\,|\,{\widehat{\Phi}}^{({\mathcal{I}}\cup{\mathcal{K}})\cap\langle e\rangle}_{\mathbf{X}}) = {\mathsf{E}}_{\lambda}(F_1\,|\,{\widehat{\Phi}}^{{\mathcal{K}}}_{\mathbf{X}}).$$ Therefore the above expression for ${\mathsf{E}}_{\lambda}(F_1F_2\,|\,{\widehat{\Phi}}^{{\mathcal{I}}}_{\mathbf{X}})$ simplifies to $$\begin{gathered} {\mathsf{E}}_{\lambda}\big({\mathsf{E}}_{\lambda}(F_1\,|\,{\widehat{\Phi}}^{{\mathcal{K}}}_{\mathbf{X}})\cdot F_2\,\big|\,{\widehat{\Phi}}^{{\mathcal{I}}}_{\mathbf{X}}\big) = {\mathsf{E}}_{\lambda}\big({\mathsf{E}}_{\lambda}(F_1\cdot F_2\,|\,{\widehat{\Phi}}^{{\mathcal{K}}}_{\mathbf{X}})\,\big|\,{\widehat{\Phi}}^{{\mathcal{I}}}_{\mathbf{X}}\big)\\ = {\mathsf{E}}_{\lambda}\big({\mathsf{E}}_{\lambda}(F\,|\,{\widehat{\Phi}}^{{\mathcal{K}}}_{\mathbf{X}})\,\big|\,{\widehat{\Phi}}^{{\mathcal{I}}}_{\mathbf{X}}\big) = {\mathsf{E}}_{\lambda}(F\,|\,{\widehat{\Phi}}^{{\mathcal{I}}\cap {\mathcal{K}}}_{\mathbf{X}}) = {\mathsf{E}}_{\lambda}(F\,|\,{\widehat{\Phi}}^{{\mathcal{I}}\cap{\mathcal{J}}}_{\mathbf{X}}),\end{gathered}$$ where the third equality follows by the inductive hypothesis applied to ${\mathcal{K}}$ and ${\mathcal{I}}$. *Part (1).*Since ${\mathbf{X}}$ is fully sated, in particular it is sated with respect to the functorial ${\sigma}$-subalgebra $$\bigvee_{\ell=1}^{i-1}{\Sigma}^{T_{(\ell;i]}}_{\mathbf{X}}\vee \bigvee_{\ell=i+1}^d{\Sigma}^{T_{(i;\ell]}}_{\mathbf{X}}= \bigvee_{\ell=1}^{i-1}\Phi^{\{\ell,i\}}_{\mathbf{X}}\vee \bigvee_{\ell=i+1}^d\Phi^{\{i,\ell\}}_{\mathbf{X}}= \Phi_{\mathbf{X}}^{\langle i\rangle}$$ for each $1 \leq i \leq d$. Therefore Proposition \[prop:char-factors-again\] gives $$\begin{aligned} \int_{X^d} f_1\otimes \cdots \otimes f_d\,{\mathrm{d}}{\lambda}&=& \lim_{n{\longrightarrow}\infty} \frac{1}{|F_n|}\sum_{g \in F_n} \int_X \prod_{i=1}^d (f_i\circ T_{[1;i]}^g)\,{\mathrm{d}}\mu\\ &=& \lim_{n{\longrightarrow}\infty} \frac{1}{|F_n|}\sum_{g \in F_n} \int_X \prod_{i=1}^d({\mathsf{E}}_\mu(f_i\,|\,\Phi^{\langle i\rangle}_{\mathbf{X}})\circ T_{[1;i]}^g)\,{\mathrm{d}}\mu\\ &=& \int_{X^d} {\mathsf{E}}_\mu(f_1\,|\,\Phi^{\langle 1\rangle}_{\mathbf{X}})\otimes \cdots \otimes {\mathsf{E}}_\mu(f_d\,|\,\Phi^{\langle d\rangle}_{\mathbf{X}}) \,{\mathrm{d}}{\lambda}\end{aligned}$$ for any $f_1$, …, $f_d \in L^\infty(\mu)$. *Part (2).*Property (i) of Proposition \[prop:infremove\] holds by construction of the ${\sigma}$-algebras $\Phi^e_{\mathbf{X}}$; property (ii) is given by Lemma \[lem:diag\]; and property (iii) is given by Lemma \[lem:rel-ind\]. [^1]: Research supported by a fellowship from the Clay Mathematics Institute

Alcohol-mediated calcium signals dysregulate pro-survival Snai2/PUMA/Bcl2 networks to promote p53-mediated apoptosis in avian neural crest progenitors. Prenatal alcohol exposure causes distinctive craniofacial anomalies that arise, in part, from the apoptotic elimination of neural crest (NC) progenitors that form the face. This vulnerability of NC to alcohol is puzzling as they normally express the transcriptional repressor Snail1/2 (in chick Snai2), which suppresses apoptosis and promotes their migration. Here, we investigate alcohol's impact upon Snai2 function. Chick cranial NC cells were treated with acute alcohol (52 mM, 2 hr). We evaluated NC migration, gene expression, proliferation, and apoptosis thereafter. Transient alcohol exposure induced Snai2 (191% ± 23%; p = .003) and stimulated NC migration (p = .0092). An alcohol-induced calcium transient mediated this Snai2 induction, and BAPTA-AM blocked whereas ionomycin mimicked these pro-migratory effects. Alcohol suppressed CyclinD1 protein content (59.1 ± 12%, p = .007) and NC proliferation (19.7 ± 5.8%, p < .001), but these Snai2-enriched cells still apoptosed in response to alcohol. This was explained because alcohol induced p53 (198 ± 29%, p = .023), and the p53 antagonist pifithrin-α prevented their apoptosis. Moreover, alcohol counteracted Snai2's pro-survival signals, and Bcl2 was repressed (68.5 ± 6.0% of controls, p = .016) and PUMA was not induced, while ATM (1.32-fold, p = .01) and PTEN (1.30-fold, p = .028) were elevated. Alcohol's calcium transient uncouples the Snai2/p53 regulatory loop that normally prevents apoptosis during EMT. This represents a novel pathway in alcohol's neurotoxicity, and complements demonstrations that alcohol suppresses PUMA in mouse NC. We propose that the NCs migratory behavior, and their requirement for Snai2/p53 co-expression, makes them vulnerable to stressors that dysregulate Snai2/p53 interactions, such as alcohol.

Introduction {#s1} ============ Sleep is a conserved behavioral state of fundamental significance. Nevertheless, the nature of its relevance to brain function is still a matter of active research and debate ([@bib33]). Most studies of sleep function focus on consequences of sleep loss for performance, in particular cognitive ability, or for overall physiology. Little is known about the impact of sleep on basic cell physiology or the extent to which cellular perturbations affect sleep. Amid the existing hypotheses for the purpose of sleep, multiple avenues exist for glial contribution to both regulation and function ([@bib17]), which have been minimally explored. Glial function has been linked to sleep in some contexts ([@bib6]; [@bib7]; [@bib22]; [@bib15]; [@bib49]). Sleep is well-associated with improved learning and memory, and appears to support synaptic remodeling ([@bib33]; [@bib52]), which may be accomplished in part by astrocyte and microglial activation ([@bib4]). Clearance of brain interstitial fluid has been shown to require the glymphatic system that involves astrocytes and whose function is enhanced during sleep ([@bib58]). Glial signaling may also contribute to sleep need ([@bib6]; [@bib22]), potentially as a consequence of astrocytic sensing of metabolic or energetic conditions ([@bib8]). In *Drosophila* ([@bib49]) and mice ([@bib22]), astrocytes reportedly influence sleep amount or quality following sleep deprivation. Disruption of vesicular trafficking in mouse astrocytes by dnSNARE expression prevented homeostatic rebound sleep following deprivation, suggesting that glia release sleep-promoting factors ([@bib22]). Likewise, *Notch* signaling acts in fly astrocytes to modulate sleep and learning in response to sleep loss ([@bib49]). Endocytic/exocytic trafficking is also important in fly glia for the regulation of circadian behavior ([@bib38]; [@bib37]), but effects on sleep have not been explored. Given the genetic tools and accessibility of defined glial populations in *Drosophila* ([@bib14]), the model is well positioned to discover new aspects of glial influence on sleep. Results {#s2} ======= Blocking vesicular trafficking in glia increases sleep {#s2-1} ------------------------------------------------------ To address the role of glia in homeostatic sleep in *Drosophila*, specifically rebound sleep following deprivation, we blocked vesicular trafficking during sleep deprivation using *Shibire^ts1^* (or *Shi^1^*), a temperature-sensitive dominant negative allele of *dynamin*, a GTPase involved in membrane scission ([@bib53]). Flies expressing Shibire^ts1^ (UAS-*Shi^1^*) by a pan-glial driver (*Repo*-GAL4) were raised at permissive temperature (18°C) until adulthood and subjected to mechanical sleep deprivation while being monitored for sleep/activity levels. Concurrently with the start of mechanical deprivation at lights off (Zeitgeber Time (ZT) 12), *Shi^1^* was induced by raising the temperature to a restrictive 30°C. Mechanical deprivation was applied for 12 hr, concluding at the beginning of the following day (ZT 0), when the temperature was returned to permissive 18°C. While mechanical sleep deprivation is not always total, and acclimation occurs over longer timespans, Repo-GAL4 \> UAS-Shi^1^ flies exhibited a surprisingly large amount of sleep above controls throughout the deprivation period, precluding any analysis of subsequent rebound. Experimental flies slept over 300 min in the course of 12 hr of mechanical deprivation, significantly greater than both parental controls ([Figure 1A](#fig1){ref-type="fig"}). This effect was present in both males and females and was confirmed with an additional line containing an independent single *Shi^1^* insertion (UAS-*20xShi^1^*) ([Figure 1A](#fig1){ref-type="fig"}). {#fig1} In principle, *Shi^1^* is an allele for a temperature-sensitive dynamin whose vesicle-forming membrane scission function declines with heightened temperature to act as a dominant negative. However, quite surprisingly, we found that increased sleep evident during mechanical deprivation was not dependent on a shift from permissive to restrictive temperature, as *Repo \> Shi^1^* flies deprived at 18°C exhibited greater sleep as compared to controls, commensurate to experimental flies at 30°C ([Figure 1B](#fig1){ref-type="fig"}). As discussed later, the temperature benchmarks established for dynamin inactivation may not be appropriate for all cell populations or phenotypes ([@bib31]). Importantly, the locomotor activity during wake (activity index) of *Repo* \> *Shi^1^* flies at permissive temperature does not significantly differ from that of controls ([Figure 2](#fig2){ref-type="fig"}), suggesting that the effect is particular to sleep and not simply locomotor impairment which may otherwise be observed with glial manipulation ([@bib62]). Additionally, to rule out that *Repo* \> *Shi^1^* flies may be less sensitive to the mechanical deprivation stimulus, we stimulated flies between ZT11 and ZT12, a time when most flies are awake. We found that stimulation of awake flies increased beam crossing, and that the average locomotor activity for experimental flies during this time did not differ significantly from that of controls ([Figure 1---figure supplement 1](#fig1s1){ref-type="fig"}). Therefore, it is unlikely that the increased sleep of *Repo* \> *Shi^1^* flies during mechanical deprivation is a result of diminished sensitivity to the stimulus. {#fig2} Based on the resistance to sleep deprivation displayed by *Repo* \> *Shi^1^* flies, which may represent greater sleep pressure, we closely examined daily, unperturbed sleep. Flies expressing *Shi^1^* in all glia had significantly greater total sleep than controls ([Figure 2A](#fig2){ref-type="fig"}) at 18°C, which was consistent across independent *Shi^1^* lines ([Figure 2B](#fig2){ref-type="fig"}) in females as well as males ([Figure 2---figure supplement 1A--B](#fig2s1){ref-type="fig"}). Expression of *Shi^1^* with *Repo*-GAL4 at permissive temperature resulted in an equivalent number of sleep bouts across the day, but significantly longer average bout lengths in almost all lines and sexes (see exception in [Figure 2---figure supplement 1A](#fig2s1){ref-type="fig"}), suggesting that sleep was also better consolidated ([Figure 2](#fig2){ref-type="fig"} and [Figure 2---figure supplement 1A--B](#fig2s1){ref-type="fig"}). While we did not observe an appreciable difference in resistance to sleep deprivation between 18° C and 30°C, adult *Repo*-\>*Shi^1^* flies raised at 18°C were also shifted to 30°C for 2 days to examine if the baseline sleep phenotype could be exaggerated. The pattern of *Drosophila* sleep across the day is known to alter at temperatures exceeding those preferred by flies (\~25°C)-- with greater sleep occurring during the daytime at the expense of nighttime sleep ([@bib40]). Appropriately, when kept at 30°C, sleep loss occurred during the nighttime in all flies, but an appreciably greater amount of nighttime sleep was seen in *Repo* \> *Shi^1^* flies ([Figure 2---figure supplement 2A](#fig2s2){ref-type="fig"}). Total sleep at 30°C in experimental female flies was equivalent to that at 18 degrees ([Figure 2---figure supplement 2B](#fig2s2){ref-type="fig"}, and even diminished in males [Figure 2---figure supplement 2C](#fig2s2){ref-type="fig"}), suggesting that the effect of glial *Shi^1^* on daily sleep is substantial even at allegedly permissive temperature. A nighttime phenotype is not seen at 18°C most likely due to ceiling amounts of sleep in controls but is clearly present at 30°C, where the temperature-dependent change in sleep pattern shifts control sleep to daytime hours. This suggests that *Shi*^1^ expression can increase sleep throughout the day. To determine whether increased sleep is the result of a partially dominant-negative *Shi^1^* at lower temperature or the overexpression of *Shi^1^*, we compared the effects of WT dynamin (*Shi^WT^*) with *Shi^K44A^*, a constitutive dominant negative allele of *dynamin*. While *Shi^K44A^* is also deficient for GTP hydrolysis, it is a distinct mutation from *Shi^1^*(G273D) ([@bib53]). Expression of *Shi^K44A^* in all glia resulted in significantly elevated sleep, in particular during the day, while expression of *Shi^WT^* had no effect ([Figure 2C](#fig2){ref-type="fig"}, [Figure 2---figure supplement 3](#fig2s3){ref-type="fig"} for total sleep). Given that a constitutively dominant negative dynamin phenocopies the effect of *Shi^1^* expression at 18°C, the sleep phenotype likely arises from inhibition of dynamin by *Shi^1^* even at the reportedly permissive temperature. Glia of the blood brain barrier link sleep to vesicular trafficking {#s2-2} ------------------------------------------------------------------- As noted above, astrocytes have previously been implicated in sleep ([@bib6]; [@bib22]; [@bib49]). In the fly, various glial classes have been described with corresponding GAL4 drivers generated for these classes ([@bib50]). Apart from astrocytes ([@bib11]), *Drosophila* glia include the ensheathing/wrapping glia ([@bib11]; [@bib28]), cortex glia ([@bib3]), and the two layers of the surface or blood-brain barrier glia ([@bib3]; [@bib46]). To determine whether any specific glial class is sufficient for the *Shi^1^* phenotype, we expressed *Shi^1^* at the permissive temperature using a panel of previously published GAL4 drivers targeting different glial populations ([Figure 3A](#fig3){ref-type="fig"}), with preference given to drivers which would isolate individual glial classes. Surprisingly, we did not observe a significant effect on sleep as compared to controls when using drivers expressed in astrocyte-like glia. Instead, expression of *Shi^1^* in the subperineurial (SPG) glia of the blood brain barrier, by *moody*-GAL4, produced a significant increase in total ([Figure 3A](#fig3){ref-type="fig"}) and daytime sleep [Figure 3---figure supplement 1](#fig3s1){ref-type="fig"}. The same was true, although to a lesser extent, with the NP6293-GAL4 driver, which expresses in the perineurial glia (PG), the second layer of the fly BBB (example expression [Figure 3---figure supplement 2](#fig3s2){ref-type="fig"}. The cumulative sleep increase produced by the two drivers is somewhat less than with Repo-G4, which likely reflects discrepancy in driver strength, but could also suggest a role for other populations -- nevertheless only the surface glial drivers were sufficient in isolation. Surface glia were also sufficient for the resistance to sleep deprivation phenotype, although only through the PG population ([Figure 3---figure supplement 3](#fig3s3){ref-type="fig"}Fig., S3.3). Together these data indicate an involvement of BBB glia in *Shi^1^*-mediated effects on fly sleep. {#fig3} Inducing block of vesicular trafficking in adult blood brain barrier glia increases sleep {#s2-3} ----------------------------------------------------------------------------------------- As an additional glial sub-type GAL4 driver, we observed that *Rab9*-GAL4 exhibits prominent expression in the BBB glia. To identify the BBB glial population labeled by *Rab9*-G4 we drove expression of a nuclear-localized GFP, which can differentiate the surface glial populations based on nuclear morphology and abundance ([@bib9]). The nuclei of perineurial glial are numerous and relatively small, while those of subperineurial glia are distinctively large and limited in number, consistent with the size and distribution of this glial class. The pattern of expression of nuclear GFP under the control of *Rab9*-G4 was typical of the subperineurial glia ([Figure 3---figure supplement 2A](#fig3s2){ref-type="fig"}), marking Rab9-G4 as an SPG driver. In addition, *Rab9*-G4 also shows some, albeit sparse, neuronal expression in the brain ([Figure 3---figure supplement 2B](#fig3s2){ref-type="fig"} for *Repo*-Gal80 images). While constitutive expression of *Shi^1^*, even at permissive temperature, with *Rab9*-G4 was lethal, we observed prominently enhanced sleep when *Shi^1^* was conditionally expressed in adulthood using the TARGET system ([@bib35]) ([Figure 3B](#fig3){ref-type="fig"}). Further, conditional expression of *Shi^1^* with *Rab9*-G4 in the presence of *Repo*-GAL80, an inhibitor of GAL4 localized to all glial cells, prevented the increase in sleep otherwise seen with this driver ([Figure 3C](#fig3){ref-type="fig"}), concomitant with elimination of barrier expression ([Figure 3---figure supplement 2B](#fig3s2){ref-type="fig"}). These data exclude a contribution of neuronal expression of this driver and confirm the sufficiency of the subperineurial glia for the *Shi^1^*-mediated increase in sleep. We confirmed adult-specificity of *Shi^1^* effects by also using the TARGET system to induce *Repo* \> *Shi^1^* expression conditionally during adulthood. A temperature-shift increases sleep in these flies, and upon return to restrictive temperature, the sleep of experimental flies was no longer significantly different from that of controls ([Figure 3---figure supplement 1B](#fig3s1){ref-type="fig"}), demonstrating that the phenotype is not likely the result of permanent detrimental effects. Together these data establish that sleep increase upon *Shi* expression is inducible and reversible in adult animals, and is not solely a result of developmental changes. *Shibire^1^* expression alters ultrastructure but not general integrity of the barrier {#s2-4} -------------------------------------------------------------------------------------- Behavioral, electrophysiological or morphological phenotypes have previously been noted from *Shi^1^* expression at permissive temperature ([@bib21]; [@bib31]). Most prominently, overexpression of *Shi^1^* at 19°C within photoreceptor cells produces a modified ERG profile, accompanied by a decrement in endocytic vesicles and accumulation of microtubule bundles, amounting to gross morphological alteration at the ultrastructural level ([@bib21]). To determine whether the morphology of fly glia is affected by *Shi^1^* expression, we performed transmission electron microscopy (TEM) on *Repo* \> *Shi^1^* flies raised at permissive temperature. When viewed as a horizontal section through the fly brain, the BBB glia form the two most superficial cell layers which continuously envelope the entire circumference of the brain. The thicker and seemingly less electron-dense PG layer faces the luminal side, and the underlying SPG, considered to be the tight barrier by virtue of septate junctions, appear as a predominantly thinner layer with periodic expanded involutions and protrusions. Examination of the superficial layers reveals a dense and distinct cytoplasmic make-up in *Repo* \> *Shi^1^* brains. Most prominently, amassed and repeated ring-like structures, resembling the microtubule bundles previously reported ([@bib21]), appear in the experimental animals but are not seen in controls ([Figure 4A](#fig4){ref-type="fig"}). Additionally, the PG layer appears relatively thinner as compared to controls. This finding shows that in glia, as in neurons, *Shi^1^* expression even at permissive temperature results in a morphological alteration. {#fig4} The principle role of the hemolymph-brain barrier is to preserve a selective barrier preventing free exchange of hemolymph and brain interstitial fluid. To assess whether expression of *Shi^1^* in glia compromises the general integrity of the barrier, we injected *Repo* \>*Shi^1^* flies with 10kD dextran conjugated to Alexa647, and dissected brains the next day for fixation and analysis by confocal microscopy. In animals with intact barriers, 10kd dextran is restricted from the brain, and forms a layer external to the surface glia ([@bib42]). In both *Repo* \>*Shi^1^* flies as well as parental controls, the fluorescent dextran was seen at the periphery of the brain ([Figure 4B](#fig4){ref-type="fig"}) indicating that *Shi^1^* expression does not create a leaky barrier. Taken together with the electron micrographs, these images support the finding that with *Shi^1^* expression, general barrier integrity is intact, but intracellular morphology is altered, although the link of the altered morphology to endocytosis is unclear. Alterations in specific Rab proteins affect sleep {#s2-5} ------------------------------------------------- To better define the relevant pathway, we sought a relatively unbiased method for perturbing trafficking. The Rab proteins are a family of membrane-bound GTPases that demarcate trafficking compartments (e.g. early, recycling, late endosomes, etc) and are known to regulate diverse vesicle movement within the cell, including exo- and endocytotic pathways. We asked which trafficking pathways are important in glia for sleep regulation by screening an existing collection of UAS-Rab lines ([@bib60]), which include both constitutively active (CA) and dominant negative (DN) constructs for the currently described *Drosophila* Rabs. For most Rabs in this collection, two separate insertions exist for each construct (DN and CA), and in the initial screen all available lines were crossed to a pan-glial driver, with hits defined as Rabs that showed opposing effects on sleep of all available CA as compared to all DN lines. In order to avoid potential lethality from overexpression of dominant negative constructs, we used a newly created hormone-inducible pan-glial driver, *Repo*-GeneSwitch (RepoGS). While lethality was not seen for any Rab using this driver, with or without RU486 feeding, we observed that expression was leaky ([Figure 4---figure supplement 1A](#fig4s1){ref-type="fig"}), that is not totally dependent on RU486, which has also been reported for other GeneSwitch lines ([@bib47]). Unfortunately, leakiness of RepoGS seemed particularly prominent in the surface glia, and so we examined effects of DN and CA lines on sleep in the presence and absence of RU486. When UAS-Rabs were expressed by RepoGS in the absence of RU, significant differences in sleep between DN and CA lines were found for Rab3, Rab9 and Rab11 ([Figure 4C](#fig4){ref-type="fig"}). In the presence of RU486, Rab1, Rab5, Rab27 and Rab30 were found significant ([Figure 4---figure supplement 1B](#fig4s1){ref-type="fig"}), albeit with only one line available for each DN and CA construct in the case of Rab30. As preliminary screening compared only experimental flies (RepoGS \> UAS Rab) with or without RU, the Rabs of interest were confirmed with *Repo*-GAL4, including the full complement of UAS and GAL4 controls. For Rab3, Rab5 and Rab9, we did not observe sleep increases by this driver ([Figure 4---figure supplement 1C](#fig4s1){ref-type="fig"}), and while some Rab27 and Rab30 constructs did exhibit significant increases, we did not find consistent phenotypes when these lines were expressed by surface-glial drivers (data not shown). Rab1 CA also increased sleep relative to parental controls, but given that it affects ER/Golgi transport ([@bib32]), we considered it is less likely to be mechanistically similar to *Shi^1^*. In contrast, expression of recycling-endosome associated *Rab11* CA resulted in a robust sleep increase when driven in all glia ([Figure 4D](#fig4){ref-type="fig"}) and was significant for both CA lines. Unlike with RepoGS, Rab11 DN driven by *Repo*-GAL4 yielded lethality ([Figure 4---figure supplement 1C](#fig4s1){ref-type="fig"}), with no adult flies detected, which may be indicative of *Repo*-GAL4 as a stronger driver. Likewise, expression of Rab11 DN in either BBB population resulted in lethality (data not shown) while expression of *Rab11* CA in either surface glial population was sufficient to produce increases in total sleep ([Figure 4D](#fig4){ref-type="fig"}). Unlike with *Shi^1^*, we did not see resistance to sleep deprivation with Rab11 driven in glia ([Figure 4---figure supplement 2B](#fig4s2){ref-type="fig"}), which could indicate more widespread alteration in trafficking by *Shi^1^* than by Rab11 alone. Rab11 is associated with the recycling endosome, and through multiple effectors, is also necessary for endocytic transport, particularly in polarized cells ([@bib29]). Together these results identify the importance of Rab11 in the surface glia, and support the role of endocytic/endosomal trafficking in the sleep function of the surface glia. Endocytosis occurs during sleep and is influenced by prior wakefulness {#s2-6} ---------------------------------------------------------------------- Considering that inhibition of endocytic trafficking in the fly hemolymph-brain barrier increases sleep, could rates of endocytosis at the barrier depend on sleep-wake state? Recent work in mice suggests that sleep serves to promote clearance of interstitial fluid, and consequently harmful metabolites which may have accumulated as a result of sustained neuronal activity during wake, through the glymphatic system ([@bib58]). Whether such clearance during sleep is a conserved feature in other organisms is currently unknown, but given that the barrier glia divide brain and body hemolymph in the fly, they are a prime candidate for regulating transport from brain interstitial fluid to the periphery as a function of sleep-wake state. In order to evaluate effects of sleep-wake state on endocytosis, we collected flies from an early night time point (ZT14 - 2 hr after lights off) and an early day time point (ZT2 - 2 hr after lights on) ([Figure 5A](#fig5){ref-type="fig"}). At ZT14, sleep pressure is typically high following daytime arousal (flies are diurnal) and a majority of flies experience consolidated sleep, while ZT2 is associated with high locomotor activity. To quantify the amount of endocytosis in the surface glia, we labeled both surface glial layers with UAS-CD8::GFP using the 9--137-GAL4 driver and dissociated the brains to allow access of fluorescence-conjugated 10 kD dextran to the basolateral surface of the barrier (as even a smaller dextran did not penetrate the brain from the apical surface of the BBB [Figure 5---figure supplement 1](#fig5s1){ref-type="fig"}). Dissociated brain cells were incubated for 30 min at room temperature, prior to analysis by flow cytometry (Materials and methods) ([Figure 5---figure supplement 2](#fig5s2){ref-type="fig"}, for gating strategy). Barrier cells from flies at ZT14 contained significantly higher dextran, suggesting that endocytosis is favored by high sleep or sleep need ([Figure 5A](#fig5){ref-type="fig"}). To verify that incorporation of dextran into GFP+ surface glial cells reflects endocytosis, samples were also pre-incubated with dynasore ([@bib34]), an inhibitor of GTP hydrolysis in dynamin family proteins. At all time points, the AF-dextran signal was significantly diminished by the presence of dynasore ([Figure 5A](#fig5){ref-type="fig"}), indicating that a substantial portion of the signal is due to endocytosis. {#fig5} Given that a difference between ZT14 and ZT2 may be indicative of a circadian effect, independent of sleep-wake state, we evaluated flies at an identical time-point (ZT2), but following sleep deprivation (SD). Flies in the SD condition were mechanically sleep deprived for 12 hr beginning at ZT12 and ending at ZT0, and allowed 2 hr to recover, before also being dissected at ZT2. Hence, circadian time is equivalent between SD and non-SD animals, but flies recovering from SD should have greater sleep pressure. We find that flies taken from a state of recovery sleep following deprivation show substantially higher rates of endocytosed AF-dextran than their un-deprived controls at ZT2 ([Figure 5A](#fig5){ref-type="fig"}) -- a level of endocytosis commensurate to that from flies at ZT14. Brains collected at 6 hr intervals over the course of the 24 hr day showed a rhythmic pattern, which was not significant as a circadian cycle by JTK, but indicated lowest and highest levels of endocytosis at ZT2 and ZT14, respectively ([Figure 5---figure supplement 3B](#fig5s3){ref-type="fig"}). To determine when endocytosis declines after the peak at ZT14, we assessed rates over the course of sleep to find that endocytosis is higher at ZT14, as compared to ZT18, ZT22 and ZT2, which are not statistically distinguishable ([Figure 5B](#fig5){ref-type="fig"}). This would suggest that endocytosis at the barrier is maximal and largely accomplished within the first half of the night. Since enhanced endocytosis at ZT2 following sleep deprivation could be a result of stress rather than enhanced sleep, we fed the sleep-inducing drug, Gaboxadol (GBX) (otherwise known as THIP) ([@bib5]; [@bib10]; [@bib55]). While exposure to Gaboxadol during the day did not alter the high endocytosis observed at ZT14 ([Figure 5---figure supplement 3A](#fig5s3){ref-type="fig"}), Gaboxadol-treated flies showed increased endocytosis at ZT2 ([Figure 5C](#fig5){ref-type="fig"}). Together these findings demonstrate that endocytosis at the hemolymph-brain barrier occurs preferentially during sleep and reflects sleep need. It is normally high during the early night when sleep predominates but can be increased in the early day if flies undergo recovery sleep following deprivation the prior night or if they are induced to sleep by drug treatment. Discussion {#s3} ========== Glial cells are an important constituent of blood or hemolymph-brain barriers, a conserved feature of complex organisms that protects the central nervous system from unregulated exchange with humoral fluids. In this study, we show that manipulations of endocytosis and vesicular trafficking in the surface glia, particularly the subperineurial glia of the fly BBB, are sufficient to elevate baseline sleep amounts, and that at the BBB, endocytosis itself is influenced by sleep-wake state. Our findings expand the scope of glial involvement in sleep and suggest that trafficking through hemolymph- or blood-brain barriers is an underexplored determinant of sleep-wake behavior. Expression of *Shibire^1^* has been a popular method to probe circuit-level influence on various fly behavior, with the prevailing interpretation being that exposure to non-permissive temperature abrogates synaptic transmission via an eventual depletion of the synaptic vesicle pool \-- a consequence of inhibited endocytosis. Ultrastructural work in neurons has revealed, however, that expression of *Shibire^1^* at the lower limit of permissive temperature markedly alters electrophysiological properties and intracellular structures ([@bib21]), which we also now find in glia. The microtubule defects we report in glia support cellular effects of *Shibire^1^* at permissive temperatures, although they are not necessarily related to behavioral phenotypes of *Shibire^1^* expression in glia (this work) or neurons ([@bib31]) under these conditions. Evidence for dynamin interaction with microtubules has primarily been from early in vitro experiments ([@bib16]), with in vivo work showing that only certain dynamin mutations alter microtubules ([@bib51]), underscoring a potentially separable impact of dynamin on endocytosis and microtubules. *Shibire^1^* expression in *Drosophila* astrocytes was previously shown to disrupt circadian rhythms ([@bib37]), but those experiments were designed to prevent expression of *Shibire^1^* at permissive temperatures, leaving open the question of whether other glial phenotypes are independent of temperature. Interestingly, this study showed a decrement in rhythmicity upon expression of *Shibire^1^* at non-permissive temperature in either of the surface glia, although the total percentage of rhythmic animals was unaffected ([@bib37]). While sleep was not directly analyzed, a loss of robustness might be explained by increased sleep, present in the context of otherwise intact daily activity patterns. In our experiments, conditional expression of *Shibire^1^* in the surface glia or all glia of adult flies was sufficient to increase sleep, demonstrating that while a phenotype is evident at permissive temperature, it is likewise inducible in adulthood and therefore cannot be accounted for by developmental effects. Dynamin (*Shibire*) is known to be fundamental to endocytosis and vesicle formation. Nevertheless, the stages of vesicular trafficking are intimately linked, making it difficult to exclusively manipulate, even at the level of molecular machinery, the processes of endocytosis and exocytosis ([@bib57]). Indeed, SNARE family members such as synaptobrevin, considered to be specific to exocytosis, have also been found to affect endocytosis ([@bib59]; [@bib61]). Therefore, we cannot exclude a contribution of exocytotic processes to the influence of the barrier on sleep, although our data clearly implicate endocytosis in barrier cells. Rab GTPase family members orchestrate vesicular and membrane traffic within cells including glia ([@bib13]; [@bib36]) and the BBB, although knowledge of their roles in these populations has been limited. Knockdown of the microtubular motor protein, kinesin heavy chain (Khc), results in locomotor deficits attributable to developmental alterations in subperineurial glia ([@bib44]), and also seen with pan-glial Rab21 knockdown. While dynamin also could interact with microtubules, we did not observe locomotor deficits in flies expressing *Shibire^1^*. Through screening of available CA and DN Rab constructs, we identified Rab11 as important for sleep regulation in the barrier glia. We find that Rab11 CA expression, like *Shibire^1^*, produces increases in total sleep when driven in the barrier populations. Constitutive expression of a dominant negative Rab11 is lethal when present in all glia or even just in surface glia, underscoring the importance of this Rab in barrier cells. Rab11 is considered a marker of the recycling endosome compartment, regulating recycling of plasma membrane constituent proteins, although evidence also links Rab11 to more direct participation in early endocytosis. For example, Rab11 functions in *Drosophila* astrocytes to regulate endocytosis of the GABA transporter (GAT) ([@bib62]) and Rab11 knockdown has been shown to inhibit endocytosis and transcytosis of LDL in human iPSC-derived neurons ([@bib56]). Considering that DN and CA mutations of Rab11 can alter trafficking in the same direction in some conditions ([@bib30]), it is possible that Rab11CA inhibits endocytosis at the BBB. Alternatively, Rab11CA could produce a similar phenotype to *Shi^1^* by a different, but related mechanism. For example, while Rab11 DN inhibits recycling of plasma membrane components ([@bib32]), Rab11 CA could enhance recycling. If the mechanism underlying sleep increase at the BBB involves trafficking a particular receptor then according to this model, both inhibiting endocytosis and enhancing recycling would keep this receptor at the membrane. While we demonstrate that interference with vesicle trafficking and endocytosis at the BBB can alter sleep, and endocytosis in these populations is affected by sleep-wake state, the nature of the sleep-relevant substrates or molecules remains to be determined. The hemolymph-brain barrier, like its mammalian counterpart, contains numerous transporters and receptors ([@bib9]) serving to selectively move nutrients and metabolites in, and deleterious compounds such as xenobiotics out. Mutants for the xenobiotic transporter, *Mdr65*, show greater total sleep, but effects have not been directly mapped to the barrier ([@bib24]). We demonstrated recently that efflux transporters in the *Drosophila* BBB are under circadian regulation such that they preferentially pump out xenobiotics during the day. At night, a decrease in transporter activity increases permeability of xenobiotics into the brain ([@bib63]). Based on our current findings, we suggest that activities of the BBB are temporally compartmentalized, such that efflux occurs during the day and endocytosis at night. However, endocytosis is dependent on sleep and sleep history rather than the circadian clock as it will occur during daytime recovery sleep following sleep deprivation at night. Neurotransmitter/modulator levels are known to influence, and vary according to, sleep-wake state ([@bib48]) and are regulated by glial populations, which could include the BBB glia. To determine if this was the case, we assessed biogenic amine and amino acid content in the brains of *Repo* \> *Shi^1^* flies, but found no consistent differences relative to controls ([Figure 5---figure supplement 4](#fig5s4){ref-type="fig"}). In mammals, a glymphatic system, involving aquaporin channels in astrocytes, promotes flow of interstitial fluid along the brain vasculature and mixing with cerebrospinal fluid to allow exchange of brain fluids and clearance of waste products ([@bib27]; [@bib58]). Interestingly, this flow is enhanced during the sleep state ([@bib58]). *Drosophila*, as other invertebrates, possess an open circulatory system and thereby lack vasculature within the brain. Nevertheless, separation between the interstitial fluid of the brain and the hemolymph at large is maintained by the hemolymph-brain barrier. Our finding of a role for barrier glia in fly sleep, and the influence of sleep state on endocytosis at the barrier, supports the possibility that clearance or interstitial exchange is a conserved function of sleep. As we find that peripherally injected 10kd dextran does not enter the brain, we speculate that sleep-dependent endocytosis in dissociated BBB cells reflects trafficking from, rather than to, the brain. We suggest that BBB endocytosis occurs during sleep to resolve products of wakefulness and thereby restore metabolic/neural homeostasis, which would account for the increased sleep need that results from a block in endocytosis. In other words, endocytosis would be highest during sleep, and decrease during the night as sleep pressure is resolved. Flies with inhibited endocytosis would be unable to perform this function and resolve sleep pressure, and would in consequence exhibit constantly elevated sleep. By this model, administration of Gaboxadol increases sleep pressure and so is associated with high endocytosis at ZT2 and ZT14, regardless of the baseline levels of endocytosis at these time points. Sleep was previously linked to permeability of the BBB in rodents, where sleep loss, and even just REM restriction, was shown to increase BBB permeability ([@bib20]; [@bib23]; [@bib26]). These effects were attributed to adenosine ([@bib26]), which is also considered a somnogenic molecule ([@bib43]). In fact, adenosine may be relevant to the interaction between astrocytes and sleep ([@bib18]; [@bib22]; [@bib45]). Given that astrocytes contribute to the mammalian BBB ([@bib1]), it is possible that sleep effects produced by inhibiting astrocyte signaling ([@bib22]) involve secondary effects on the BBB. *Drosophila* sleep studies have focused on neuronal circuits controlling sleep-wake ([@bib2]; [@bib12]), with comparatively few studies investigating the contribution of glia to this circuitry. We identify the barrier glia as regulators of sleep and propose that endocytic trafficking through the barrier is an important function of the sleep state. The specific differences between mammalian and invertebrate barriers notwithstanding, the identification of the glial/endothelial barrier as a population that can mediate changes in daily sleep aligns with emerging ideas of BBB involvement in behavior ([@bib25]; [@bib41]) and sleep function ([@bib39]; [@bib54]), and provides a readily identifiable cellular target to interrogate in other model organisms. Materials and methods {#s4} ===================== Fly lines {#s4-1} --------- Stocks present in the lab collection include---; ;*Repo*-GAL4/TM3,Sb and UAS-*Shi.ts1*; UAS-*Shi.ts1* (referred to as UAS-*Shi^1^* in the text) and UAS-CD8::GFP and UAS*-*nGFP. For control genotypes, GAL4 and UAS lines were crossed to iso31.; ;UAS*-20xShi.ts1* (referred to as UAS*-20xShi^1^*) was shared by Gerald Rubin. UAS*-Shi.WT* and UAS-*Shi.K44A* were shared by Konrad Zinsmaier. NP2222-GAL4, NP6293-GAL4, *Alrm*-GAL4, MZ0709-GAL4, MZ0097-GAL4 were shared by Marc Freeman. *Moody-*GAL4 and 9--137 GAL4 was shared by Roland Bainton. UAS-Rab CA and DN lines were acquired from the Bloomington Drosophila Stock Center. *Rab9*-GAL4 (\#51587) was also acquired from Bloomington. To generate the; ;*Repo*-GeneSwitch 2301 line a 4.2 kb genomic fragment (between coordinates 18231884 and 18236068, FlyBase release 6.19) was amplified by PCR and cloned into a PUAST-AttB-Sfi-GeneSwitch vector. Transgenesis for this construct was performed by BestGene at landing site AttP154 (97D2). Behavior {#s4-2} -------- Flies were raised in bottles on standard food at either room, or non-permissive temperature (18--19°C) for Shibire and TARGET system ([@bib35]) experiments. For sleep recording, flies were loaded into glass locomotor tubes containing 5% sucrose in 2% agar, and additionally 0.5 mM RU-486 (mifepristone) in the case of GeneSwitch experiments. Locomotor data was collected using the *Drosophila* Activity Monitoring (DAM) System (Trikinetics, Waltham, MA) and processed using PySolo([@bib19]). Data are processed as 1 min bins, with sleep defined as 5 min without activity. Activity index refers to the average number of beam crossings within an active bout. Sleep deprivation was performed by mechanical disruption using a vortexer triggered to shake randomly for 2 s out of every 20, for the span of 12 hr beginning at lights-OFF, ZT12. Flies were kept in incubators under 12 hr light: 12 hr dark (LD12:12) cycles and constant temperature (25°C or 18°C or 30[°C]{.ul}). In the case of the initial temperature shift experiments with *Shibire^1^*, flies were kept at 18°C, and shifted to 30°C for the duration of the night, before being returned to 18°C at lights-ON, ZT0. For behavioral experiments, adult flies of at least 6-days post-eclosion were used, except in the case of tubGal80.ts experiments, where younger flies were loaded to grant additional time to compensate for the loss of strength due to incomplete de-repression of tubGal80.ts. Confocal imaging {#s4-3} ---------------- Fly brains were fixed in 4% PFA in PBS with 0.1% TritonX-100 for 15--20 min, and washed in PBST for 30 min at room temperature, before mounting in VectaShield H-1000 (Vector Laboratories). Confocal imaging was performed on a Leica TCS SP5. Electron microscopy {#s4-4} ------------------- Heads from female *Repo* \> UAS-*20xShi^1^* flies approximately two weeks post-eclosion were dissected in a cold room in PBS and fixed with 2.5% glutaraldehyde, 2.0% paraformaldehyde in 0.1M sodium cacodylate buffer, pH7.4, overnight at 4°C. Samples were post-fixed in 2.0% osmium tetroxide for 1 hour at room temperature, and rinsed in DH~2~O prior to *en bloc* staining with 2% uranyl acetate. After dehydration through a graded ethanol series, the tissue was infiltrated and embedded in EMbed-812 (Electron Microscopy Sciences, Fort Washington, PA). Flies were not strictly circadian entrained or taken at an exact time of day, but dissections were performed together in the afternoon. Reported images represent sections from single flies from each genotype, although the experimental was compared to both parental controls. The described microtubule structures were not seen in the controls. Embedding, staining and sectioning was performed by the Electron Microscopy Resource Laboratory at Penn. Thin sections were stained with uranyl acetate and lead citrate and examined with a JEOL 1010 electron microscope fitted with a Hamamatsu digital camera and AMT Advantage image capture software. Hemolymph-brain barrier permeability {#s4-5} ------------------------------------ Integrity of the barrier was assessed as previously described ([@bib42]). Female flies were injected with Alexa fluor 647-conjugated 10 kd dextran (ThermoFisher D22914) the day before dissection and kept in standard food vials. Heads were removed and fixed in 4% PFA in PBS for 10--15 min before brains were further dissected out and cleaned. Brains were additionally washed in PBS for 30 min before being mounted in VectaShield H-1000 (Vector Laboratories) and imaged by confocal microscopy. HPLC/LC-MS {#s4-6} ---------- The brains of female flies ranging from 10 to 20 days post-eclosion were dissected in cold PBS, and collected to 20 brains per vial. Excess PBS was pipetted off after spinning the samples down by micro-centrifuge. The samples were frozen on dry ice and submitted to the Children's Hospital of Pennsylvania Metabolomic Core for HPLC analysis. Endocytosis assay and flow cytometry {#s4-7} ------------------------------------ Flies expressing membrane bound GFP in the surface glia (9--137 GAL4;UAS-CD8::GFP) were dissected in ice-cold AHL (108 mM NaCl, 5 mM KCl, 2 mM CaCl~2~, 8.2 mM MgCl~2~, 4 mM NaHCO~3~, 1 mM NaH~2~PO~4~-H~2~O, 5 mM trehalose, 10 mM sucrose, 5 mM HEPES; pH 7.5), and 20 brains were collected per condition. To each sample, 60 μL of Collagenase IV (25 mg/mL) and 10 μL of DNase I (1 mg/mL) were added. Following shaking at 37°C for 15 min, brains were broken up by 3 rounds of gently pipetting. Using a 70-μm cell strainer, the brain samples were then transferred to FACS tubes, to which 3 mL of AHL was added and spun down for 5 min (2500 RPM). After removing the supernatant, the pellet was resuspended in AHL and brought to a volume of 200 μL, at which point the sample was split into two new FACS tubes of 100 μL each. One tube received an additional 50 μL of vehicle solution while 50 μL of dynasore hydrate solution (10 μM final concentration) was added to the other. The samples were mixed well and incubated at room temperature for 10 min. Following incubation, 50 μL of Alexa fluor 647-10kd dextran (50 μg/ml final concentration) was added to each tube, mixed, and incubated at room temperature for 30 min. 4 mL of FACS buffer (0.5% BSA w/v + 0.1% w/v sodium azide in PBS) were added to each tube, spun down and the supernatant removed. Samples were immediately analyzed using a FACS Canto II (BD Biosciences). Gaboxadol feeding {#s4-8} ----------------- Gaboxadol hydrochloride (Cayman Chemical Company) was added to standard 5% sucrose in 2% agar food in locomotor tubes at 0.1 mg/mL. For the ZT14 timepoint, flies were flipped to Gaboxadol food at lights ON (ZT0) while for the ZT2 timepoint they were flipped onto drug just before lights OFF (ZT12). In both cases, they remained on this food for \~14 hr until dissection. Live imaging {#s4-9} ------------ 6-cm plates were coated with poly-L-lysine (0.01%) for greater than 10 min, removed, and let dry. Brains were dissected in AHL, placed on the coated plates, and incubated with 50 μL droplet of 500 μL/mL 3kD FITC-conjugated dextran and either imaged continuously for 10 mins or dye was washed off and brains were imaged in AHL for 5 min. Confocol imaging was performed with 20x submergible water objective (with the lens touching the dye droplet) at excitation/emission wavelengths of 488/520. For a permeable dye, brains were dissected in AHL and incubated with 50 μL droplet of 125 μL/mL Rhodamine B and imaged continuously for 10 min. Imaging was performed with 20x submergible water objective at excitation/emission wavelengths of 543/555. Statistical analysis {#s4-10} -------------------- Graphs and statistical tests were completed using Excel and GraphPad Prism. FACS analysis was performed in FlowJo. For measures of sleep, Controls (GAL4 alone and UAS alone) were compared to Experimental (GAL4 \> UAS) animals by one-way ANOVA with Holm-Sidak post-hoc correction. For comparison of endocytosis between ZT2 and ZT14 or ZT2 and ZT2 SD or between each time point control and inhibitor, Students' T test was performed. Additional details regarding tests and significance values are provided in the figure legends. Funding Information =================== This paper was supported by the following grants: - http://dx.doi.org/10.13039/100000002National Institutes of Health T32-HL07953 to Gregory Artiushin. - http://dx.doi.org/10.13039/100000002National Institutes of Health R37NS048471 to Amita Sehgal. - http://dx.doi.org/10.13039/100000863Ellison Medical Foundation AG-SS-2939-12 to Amita Sehgal. - http://dx.doi.org/10.13039/100000011Howard Hughes Medical Institute to Amita Sehgal. We thank Zhifeng Yue and Kiet Luu for technical support and members of the Sehgal lab for discussion and reagents. We would like to also thank the Gerald Rubin, Konrad Zinsmaier, Marc Freeman, and Roland Bainton labs for sharing reagents, and the Electron Microscopy Resource Laboratory at Penn and the Children's Hospital of Pennsylvania Metabolomic Core. Additional information {#s5} ====================== No competing interests declared. Conceptualization, Formal analysis, Investigation, Visualization, Methodology, Writing---original draft, Writing---review and editing. Formal analysis; Investigation; Visualization; Methodology; Writing---review and editing. Resources; Writing---review and editing. Conceptualization; Supervision; Funding acquisition; Writing---review and editing. Additional files {#s6} ================ 10.7554/eLife.43326.020 Data availability {#s7} ----------------- All data generated or analysed during this study are included in the manuscript and supporting files. 10.7554/eLife.43326.023 Decision letter Ramaswami Mani Reviewing Editor Trinity College Dublin Ireland In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included. \[Editors' note: a previous version of this study was rejected after peer review, but the authors submitted for reconsideration. The first decision letter after peer review is shown below.\] Thank you for submitting your work entitled \"Endocytosis at the blood-brain barrier as a function for sleep\" for consideration by *eLife*. Your article has been reviewed by a Senior Editor, a Reviewing Editor, and three reviewers. The following individuals involved in review of your submission have agreed to reveal their identity: Paul Taghert (Reviewer \#3). Our decision has been reached after consultation between the reviewers. Based on these discussions and the individual reviews below, we regret to inform you that your work will not be considered further for publication in *eLife*. However, we do feel the work could be made acceptable but only after the essential changes noted below are handled experimentally. Should you choose to submit a new version of this work, we will endeavor to have it reviewed by the same Board member and referees. Summary: This paper explores the functional contributions of the blood brain barrier (BBB) to sleep regulation in *Drosophila*. It uses a full range of genetic cellular and imaging methods to consider the hypothesis that endocytosis/exocytosis in BBB glia changes in ways that supports the role of these cells in helping resolve levels of metabolites that increase in relation to sleep-need. Further it provides evidence for the involvement of glial Rab11 in the cellular mechanism. Experimentally analyzing glial function in sleep, the work touches an aspect of brain physiology that has just begun to be explored but that has been proposed as one of the crucial functions of sleep. Thus, this work is a potentially significant contribution of interest to a general audience. Though work is clearly described, thorough in most aspects and also several caveats are clearly acknowledged, there are issues that need to be clarified and conclusions that need to be strengthened in order to better establish the key conclusion of the paper that inhibition of endocytosis in glia increases sleep. Essential revisions: 1\) The authors nicely describe various difficulties experienced with genetic tools to perturb endocytosis in targeted glia (e.g. temperature-independent and potentially endocytosis-independent effects of shi-ts expression, as well as leaky GsGal4 expression). In addition, they report similar effects on sleep following expression of constitutively active and dominant negative forms of Rab11 as well as perturbation of Rabs that may not participate in endocytosis (Figure 4 and Figure 4---figure supplement 1). Thus, it is important to better establish sleep phenotypes arise from changes in endocytosis rather than non-specific glial malfunction. Some suggestions of additional experiments to support the hypothesised role of endocytosis are proposed below. 1a) To better characterise sleep-associated endocytosis, it would be useful to measure rates of endocytosis at 4-6 time points to define a daily cycle, not just two time points. 1b) Given the possibility that even adult-specific expression of mutant dynamin or Rab RNAi can have long-lasting changes in glial physiology or alterations in BBB permeability, it would be useful to test if these flies revert to normal sleep behaviour if allowed to recover for an appropriate period of time. 1c) To show that a different amount of sleep deprivation (more or less) produces a proportionate change in endocytosis (i.e., two or more points on the concentration-effect curve). 1d) To show bidirectionality of regulation and to control for potential contribution of stress rather than sleep drive, the authors should test whether endocytosis following more sleep is reduced (this may require orthogonal genetics: e.g. to drive sleep-inducing FB neurons). 2\) Different Gal4 lines are used to target subsets of glia. The specificity of these lines needs to be clearly documented. The logic for using one or other for different experiments needs to be clarified. And at least one case, an additional Gal4 may need to be used. 2a) The description of Gal4 expression lines (Figure 3---figure supplement 2) should be expanded Critically, to demonstrate that each one the Gal4s used is indeed expressed preferentially in glia, the figure should include counter stains with glia or neuron-specific antibodies (Repo and Elav, respectively). Glial-specific expression would be also more convincing if the authors could test how well the expression pattern of these Gal4s (Rab9-Gal4 in particular) is blocked by Repo\>Gal80. (More minor, it might help if an outline of the brain were drawn). 2b) The tubGal80.ts experiments utilize Rab9-Gal4. To more convincingly implicate the glial cells presumed to be targeted, these experiments should also be repeated with Moody-Gal4 (and possibly Repo-Gal4 as well). 3\) Figure 1 shows that of Shi-ts expression in glia induces resistance to mechanical stimulation during sleep. Can a control experiment be done to rule out the possibility of a sensory defect that makes these flies unresponsive to mechanical stimulation? E.g. To test the responsiveness of awake Repo\>Shi flies to mechanical stimulation? On this note, are the PG and SPG drivers expressed outside the brain in sensory structures? Other concerns and suggestions: 4\) Is it necessary to dissociate cells before adding Alexa647-dextran? Why not measure before dissociation as most cellular processes including endocytosis could be more robust in an intact brain. If it is simply to allow access to Alexa647-dextran, then could an alternative smaller endocytic tracer be tried? 5\) Is resistance to mechanical stimulation following SD specific only to Shi, or also seen following endocytic Rab manipulation? 6\) The shi-ts expression results suggest that sleep during SD happens at both temperatures, but nighttime sleep seems to elevated sooner in both males and females at 18 degrees, than in controls. Is this impression correct? and if so, is there an explanation? This difference (if reproducible) involves latencies and time of day differences. The fly sleep literature appears to point to an increasing fragmentation of cellular mechanisms between those that affect daytime versus nighttime sleep. Is there any basis to think or expect that the cell biological mechanisms (exo- and endocytosis) should be geared to daytime versus nighttime? 7\) Figure 2C suggests that the effects of Shi-ts on sleep are not due to merely overexpression of Dynamin since the expression of Shi-wt does not have a significant effect. However, for Shi-wt, only total sleep is shown, whereas the main effect for Shi-ts at 18C is on day sleep. To exclude the possibility that a Shi-wt effect on sleep is masked when only total sleep is shown the figure should be revised to include data showing the effect of Shi-wt on daytime sleep. 8\) Indeed, it would useful to consistently show data for total sleep and daytime sleep for all genotypes in the paper (these data must be available). 9\) In Figure 3A, to exclude possible effects due to expression of Shi in neurons (in addition to expression in the BBB glia), the authors could show that these phenotypes remain when the expression in all glia is blocked using Repo\>Gal80. 10\) It would be valuable to test, with existing tools, to what extent surface glia are responsible for the phenotypes observed with the pan-glial driver, Repo. Does Shi expression in surface glia also cause resistance to mechanical sleep deprivation? Is the phenotype of either NP6293 or moody \>Shi weaker than that of Repo\>Shi and does combining NP6293 and moody Gal4s generate a phenotype comparable to that of Repo-Gal4? Why is the 9-137-Gal4 driver (SPG & PG) that is later used in Figure 5 (subsection "Endocytosis occurs during sleep and is influenced by prior wakefulness") not used here? 11\) Figure 3 localizing effect to surface glia: In Figure 3A -- which UAS-shi was used? The effect does not appear to match that of repo-Gal4. The difference is no more than 100 min; whereas. in Figure 2A, the difference is \>200 min. Is this lack of quantitative matching significant? Maybe it reflects the strength of the Gal4s but alternatively it could mean that other cells contribute. This should be discussed. 12\) In the same figure -- the lack of effect of astrocyte expression seems to be correlated with an unusually high value for the UAS control. Is there a reason to be concerned here? 13\) Regarding EM images in Figure 4A -- how abundant and frequent are these ultra-structural changes and how do they correlate with sleep phenotypes? The observations themselves (amassed ring-like structures resembling microtubules, and a thinner PG layer) are only described to the level of positioning of two arrows in one of the micrographs. There should be some definition of the phenotypes and indication of prevalence, severity or quantification (either within or between animals). Were the suspected MTs coated or bridged as originally described in the paper cited? There no details provided in the methods regarding gender, time of day, age, or \"n\" value. In addition, it would be useful to know if these are also seen following induction of wild-type dynamin, because WT-dynamin expression does not cause sleep phenotypes and there is the potential to examine how relevant these ultrastructural defects might be to the sleep effects observed. (Perhaps the cited EM paper has addressed this issue in neurons?) Are there any changes in membranes or vesicular structures that could potentially be connected to a block in endocytosis? What is the predicted EM phenotype associated with Rab11 CA expression? 14\) Is the overall structural integrity of BBB, as shown in Figure 4B, also preserved when Rab11CA is expressed in glia? 15\) In Figure 4C, the significance is determined based on a difference between DN and CA allele of each Rab protein. Is this the best criterion? Wouldn\'t it be better to compare each manipulation to the parental controls? In any event, it should be made clear why only Rab11 and not others is deemed to have crossed a significance threshold. 16\) The overall logic that explains the link between the manipulations, endocytosis and sleep is confusing. Glial expression of Rab11CA increases sleep, similarly to Shi-ts. Do the authors propose that Rab11CA inhibits endocytosis, similarly to Shi-ts? 17\) On similar lines, the data shown in Figure 4C seem to contradict the data in Figure 4D. In Figure 4D, Repo\>Rab11CA causes increased sleep, compared to parental controls. Figure 4C, shows that using the RepoGS driver, Rab11 DN induces even more sleep than Rab11CA. Thus, both of these opposing manipulations cause increased sleep. Could it be that both manipulations simply make glial cells sick, and therefore, in both cases, sleep is increased due to malfunction of glia, and not due to specific changes in endocytosis? The possibility for a non-specific effect is supported by similar data in Figure 4---figure supplement 1C, where expression of either DN or CA versions of Rab27 and Rab30 produces exactly the same phenotype of increased sleep, unless the authors believe that these also work by affecting endocytosis. These issues should be explained and addressed in a revised manuscript. 18\) In Figure 4---figure supplement 1B -- are the effects of Rab3, Rab9 and Rab11 that were significant in Figure 4A (without RU486) still significant in the Figure 4---figure supplement 1B (with RU486)? The significance stars in Figure 4---figure supplement 1B are shown only for Rab1, Rab5, Rab27 and Rab30. 19\) The apparent contradiction between Figure 1, Figure 2, Figure 3, Figure 4 and Figure 5 should be more clearly explained. Figure 5 shows that SD (which increases sleep pressure) causes increased endocytosis, but, Figure 1, Figure 2, Figure 3 and Figure 4, show that inhibiting endocytosis actually increases sleep, suggesting that endocytosis in BBB promotes wakefulness. An explanation that connects the two observations should be included. 20\) Are the changes in endocytosis that correlate with sleep state specific only to surface glia? Are there changes in endocytosis in other populations of glial cells, as function of sleep? *Reviewer \#1*: In this manuscript, Artiushin et al., aim to explore the role of glia endocytosis in sleep regulation focussing on a *Drosophila* model of the blood-brain barrier. The work is very interesting in principle as it touches an aspect of brain physiology that has been hardly explored so far but that has been proposed as one of the crucial functions of sleep. The manuscript is well written in general and the experiments are well reported. The main weakness of the work is that in almost all experiments the genetic tools adopted did not behave as expected. The temperature sensitive form of shibire, (which is instrumental for experiments in Figure 1, Figure 2 and Figure 3) does not appear to provide any specific insight, possibly due to a temperature insensitive activation. The repoGENESwitch tool (instrumental for experiments in Figure 4) does not seem to be properly responsive when it comes to controlling activation. All in all, the fact that the genetic tools employs are not behaving appropriately is a major weakness for the study as it is basically impossible to understand the exact specificity of the effect in terms of timing. For instance, it is impossible to exclude that this a developmental effect. One possibility to reinforce the findings could be to adopt different genetic tools with greater reported specificity, such as the ones described in https://pubs.acs.org/doi/10.1021/acssynbio.7b00302 Presentation of data could be improved. For instance: \# There is no reason to limit labelling of panels within figures. Please use more letters, ideally one per panel. It facilitates discussion. \# Avoid using barplots. Barplots are uninformative and should almost always be discouraged. Use plots that are more informative, giving more insights on the distribution (such as boxplot or violin plots) \# The experiment in Figure 3---figure supplement 2 needs some kind of co-staining confirmation \# Please quantify the observations collected using electron microscopy? *Reviewer \#2*: Most of the manuscript (Figure 1, Figure 2, Figure 3 and Figure 4) describes how manipulating glia affects sleep. Overall, the behavioral phenotypes are strong. However, there are multiple issues with specificity of the methods, with data representation and with the interpretation of some of the results. Another question is -- is it really so surprising and interesting that manipulating large numbers of glial cells has an effect on brain function, and sleep in particular? The second part (Figure 5) is more interesting and promising, as it provides evidence for correlation between sleep/wake states and the rate of endocytosis in BBB glia. However, this part needs further development. Also, the link between this part and the rest of the manuscript seems a bit artificial, as they do not really support or complement each other well. Specific comments: 1\) The authors show in Figure 1 that expression of the temperature sensitive allele of dynamin, Shi-ts in glia induces resistance to mechanical sleep deprivation. The authors show that the basal wake activity of these flies is not affected. However, another interpretation of the data should be considered. These flies may be unresponsive to mechanical stimulation. The authors should test the responsiveness of awake Repo\>Shi flies to mechanical stimulation. The authors should also check if Repo\>Shi provides resistance to sleep deprivation induced by alternative methods (thermo-genetic). Finally, the authors test the effects on SD only with Shi, but do not do it with the rest of manipulations, presented later in the paper, such as expressing mutant Rab proteins. Are the effects on resistance to SD specific only to Shi, but not to other manipulations of endocytosis? 2\) The authors observe strong effects for the temperature sensitive Shi allele even at the permissive temperature of 18C. This raises the question whether Shi is a good tool in this specific system. The main concern is that the observed phenotypes are mostly due to non-specific effects that make the glial cells sick. The authors should, at least, test if knocking down the endogenous Shi using RNAi produces a phenotype similar to expressing dominant negative Shi. 3\) The authors claim in Figure 2C, that the effects of Shi-ts on sleep are not due to merely overexpression of Dynamin since the expression of Shi-wt does not have a significant effect. However, for Shi-wt, only total sleep is shown, whereas the main effect for Shi-ts at 18C is on day sleep. Could it be that the effect of Shi-wt on sleep is masked when only total sleep is shown? The authors should demonstrate the effect of Shi-wt on day sleep. 4\) In Figure 3A, to exclude possible effects due to expression of Shi in neurons (in addition to expression in the BBB glia), the authors should show that these phenotypes remain when the expression in all glia is blocked, using Repo\>Gal80. 5\) To what extent are surface glia responsible for the phenotypes observed with the pan-glial driver, Repo? For instance, does Shi expression in surface glia also cause resistance to mechanical sleep deprivation? Is the phenotype of either NP6293 or moody \>Shi weaker than that of Repo\>Shi? Will combining both NP6293 and moody Gal4s (SPG & PG) generate a phenotype comparable to that of Repo-Gal4? Why don\'t the authors use here the 9-137-Gal4 driver (SPG & PG) that is later used in Figure 5 (subsection "Endocytosis occurs during sleep and is influenced by prior wakefulness")? 6\) The images in Figure 3---figure supplement 2 should be of better quality. The outline of the brain should be drawn. The authors also should demonstrate that each one the Gal4s used in this Figure is indeed expressed preferentially in glia by co-staining with glia or neuron-specific antibodies (Repo and Elav, respectively). The glia-specific expression would be also more convincing if the authors could show the expression pattern of these Gal4s (Rab9-Gal4 in particular) when the expression of the UAS-nGFP reporter is blocked in glia, using Repo\>Gal80. 7\) The EM images in Figure 4A -- how abundant and frequent are these ultra-structural changes? Are these changes also observed upon the expression of Rab11CA? Also, could the authors detect any accumulation of vesicles that cannot be secreted due to expression of mutant Shi? 8\) Is the overall structural integrity of BBB, as shown in Figure 4B, also preserved when Rab11CA is expressed in glia? 9\) In Figure 4C, the significance is determined based on a difference between DN and CA allele of each Rab protein. Is this the best criterion? Wouldn\'t it be better to compare each manipulation to the parental controls? Also, what would be the effect of overexpressing WT versions of these Rab proteins? 10\) The overall logic that explains the link between the manipulations, endocytosis and sleep is confusing. Glial expression of Rab11CA increases sleep, similarly to Shi-ts. Do the authors propose that Rab11CA inhibits endocytosis, similarly to Shi-ts? Do they suggest that expression of constitutively active Rab11 is equivalent to the inhibition of vesicular trafficking? If so, what would be the effect of Rab11 RNAi? Would it produce an opposite phenotype (less sleep)? Repo\>Rab11 DN could be useful, but it was lethal. The authors should try using tub-gal80ts, to achieve inducible expression of Rab11 DN. Also, does Repo\>Rab11CA cause resistance to mechanical sleep deprivation, just like Shi-ts? 11\) One of the major claims of this paper is that inhibition of endocytosis in glia (Shi-ts, Rab11CA) increases sleep. It is important to show that an opposite manipulation of increasing endocytosis actually reduces sleep. The data shown in Figure 4C is confusing and seems to contradict the data in Figure 4D. In Figure 4D, Repo\>Rab11CA causes increased sleep, compared to parental controls. The interpretation is that this happens due to inhibition of endocytosis. However, in Figure 4C, using RepoGS driver, Rab11 DN induces even more sleep than Rab11CA. It seems that both of these opposing manipulations cause increased sleep, which is counter-intuitive. Could it be that both manipulations simply make glial cells sick, and therefore, in both cases, sleep is increased due to malfunction of glia, and not due to specific changes in endocytosis? The possibility for a non-specific effect is supported by similar data in Figure 4---figure supplement 1C, where expression of either DN or CA versions of Rab27 and Rab30 produces exactly the same phenotype of increased sleep. 12\) In Figure 4---figure supplement 1B -- are the effects of Rab3, Rab9 and Rab11 that were significant in Figure 4A (without RU486) still significant in the Figure 4---figure supplement 1B (with RU486)? The significance stars in Figure 4---figure supplement 1B are shown only for Rab1, Rab5, Rab27 and Rab30. 13\) How do the authors explain the phenotypes caused by the expression of other Rabs, besides Rab11? Do they think that they all work by affecting endocytosis? 14\) In Figure 5, the authors should demonstrate that their genetic manipulations (Shi-ts and Rab11CA) actually can inhibit endocytosis in the surface glia using the 10kd dextran absorption assay. Also, could it be that the changes they see after SD are in fact caused by physical damage to the BBB? They may test if the effects of SD on endocytosis are resolved once the flies are allowed few hours of recovery after SD. They can also test how thermo-genetic SD affects their assay. They should also do EM on BBB after SD. 15\) There is an apparent contradiction between Figure 1, Figure 2, Figure 3Figure 4 and Figure 5. According to Figure 5, SD causes increased endocytosis. SD of course increases sleep pressure. But, in the Figure 1, Figure 2, Figure 3 and Figure 4Figures, the authors claim that inhibiting endocytosis actually increases sleep, suggesting that endocytosis in BBB promotes wakefulness. A possible solution to this paradox is that during normal sleep, there is an increase in endocytosis. This is important to support the function of sleep because increased endocytosis eventually helps to reduce sleep pressure. This is why inhibition of endocytosis promotes sleep -- because in the absence of proper endocytosis, sleep quality is compromised, and sleep is then less effective in reducing the sleep pressure. 16\) Are the changes in endocytosis that correlate with sleep state specific only to surface glia? The authors should have additional control in Figure 5 -- test if there are any changes in endocytosis in other populations of glial cells, as function of sleep. *Reviewer \#3*: This paper explores the functional contributions of the blood brain barrier (BBB) to sleep regulation in the model system *Drosophila*. It is a careful study that uses a full range of genetic cellular and imaging methods to consider the hypothesis that endocytosis/exocytosis in BBB glia changes in ways that supports the role of these cells in helping resolve levels of metabolites that increase in relation to sleep-need. Further it makes a strong case for the involvement of Rab11 in the cellular mechanism. The emerging involvement of glial compartments in the fundamentals of brain physiology (here sleep regulation) makes this work a significant contribution of interest to a general audience. The work is clearly described, thorough in most aspects of the analysis and discussed in a scholarly fashion. I have a few questions and concerns that may help improve the paper. Shibere mis-expression and effects on sleep. The paper wisely pays careful attention to the effects (and lack of effects) of temperature with a classic ts allele of dynamin. In addition, I also commend the use of different shi isoforms to help focus on the biologically (not technically) important results. I had one question about this dataset that involved, not amount of sleep, but its temporal patterning. The results suggest that sleep during SD happens at both temperatures but I notice that nighttime sleep is elevated sooner in both males and females at 18 degrees, than in controls. Is this impression correct? and if so, is there an explanation? This difference (if reproducible) involves latencies and time of day differences. In my reading of the fly sleep literature, there is an increasing fragmentation of cellular mechanisms between those that affect daytime versus nighttime sleep. Is there any basis to think or expect that the cell biological mechanisms (exo- and endocytosis) should be geared to daytime versus nighttime? Dynamin and MTs The ultrastructural work is an effective supporting body of evidence to evaluate the effects of UAS-shi at permissive temperatures. However, the results -- amassed ring-like structures resembling microtubules, and a thinner PG layer -- were not well-described, beyond the positioning of two arrows in one of the micrographs. There was no clear definition of the phenotypes nor any indication of prevalence, severity or quantification (either within or between animals). Were the suspected MTs coated or bridged as originally described in the paper cited? There no details provided in the methods regarding gender, time of day, age, or \"n\" value. On a tangential (but relevant) note, was ultrastructure observed at only a single phase point? This issue is clearly beyond the scope of the current study. Cell Specificity Figure 3 localizing effect to surface glia; 3A -- which UAS-shi was used? \- Male or female flies? \- The effect does not appear to match that of repo-Gal4. Difference is no more than 100 m; whereas. In Figure 2A, the difference is \>200 min. Is this lack of quantitative matching significant? Maybe it reflects the strength of the Gal4s but alternatively it could mean that other cells contribute. I did not find any discussion of this point. Rab involvement The screens were careful, unbiased, quantitative and very large in scope. I commend the authors for determining that the Gs-Gal4 line was leaky but was somewhat confused by the overall definitions of hits. There was a set of hits with no RU486 (Rab 3, 5 and 9) and a different one with RU486 (Rab1, 5, 27 and 30). Yet at the end, only Rab11 is judged to have experimental support -- I was not clear why (1) only Rab11 and not others crossed threshold. Also (2) whether Rab11 is a positive or negative regulator of sleep. "As preliminary screening compared only experimental flies (RepoGS\>UAS-Rab) with or without RU, the Rabs of interest were confirmed with Repo-GAL4, including the full complement of UAS and GAL4 controls." 1\) But only Rab 11 is shown, not Rab 3 or Rab9 -- follow-up on Rabs 3 and 9 were neither shown nor mentioned, "Expression of recycling-endosome associated Rab11 CA resulted in a robust sleep increase when driven in all glia (Figure 4D) and was significant for both CA lines." 2\) Paradoxically, Rab11 activity appears to increase sleep in the follow-up but in the preliminary large scale screen the DN isoforms of Rab11 produced significantly more sleep than did the CA isoforms -- by my reading that means Rab11 in glia is a negative regulator. So, I\'m not sure how to interpret the combined dataset. Endocytosis 1\) Why dissociate cells before adding Alexa647-dextran? Why not measure before dissociation? The results are fortunately very clear, but I would think cellular processes like endocytosis would be more robust in an intact brain. 2\) Also it would be a stronger case if a different amount of SD produces a proportionate change in endocytosis (i.e., two points on the concentration-effect curve). Also, would it would strengthen the case to measure endocytosis following more sleep (to control for potential contribution of stress to change); This would require orthogonal genetics to drive sleep-inducing FB neurons, so I suspect the genetics would be too complicated for a two-month period of additional study. 10.7554/eLife.43326.024 Author response \[Editors' note: the author responses to the first round of peer review follow.\] > This paper explores the functional contributions of the blood brain barrier (BBB) to sleep regulation in Drosophila. It uses a full range of genetic cellular and imaging methods to consider the hypothesis that endocytosis/exocytosis in BBB glia changes in ways that supports the role of these cells in helping resolve levels of metabolites that increase in relation to sleep-need. Further it provides evidence for the involvement of glial Rab11 in the cellular mechanism. Experimentally analyzing glial function in sleep, the work touches an aspect of brain physiology that has just begun to be explored but that has been proposed as one of the crucial functions of sleep. Thus, this work is a potentially significant contribution of interest to a general audience. > > Though work is clearly described, thorough in most aspects and also several caveats are clearly acknowledged, there are issues that need to be clarified and conclusions that need to be strengthened in order to better establish the key conclusion of the paper that inhibition of endocytosis in glia increases sleep. > > Essential revisions: > > 1\) The authors nicely describe various difficulties experienced with genetic tools to perturb endocytosis in targeted glia (e.g. temperature-independent and potentially endocytosis-independent effects of shi-ts expression, as well as leaky GsGal4 expression). In addition, they report similar effects on sleep following expression of constitutively active and dominant negative forms of Rab11 as well as perturbation of Rabs that may not participate in endocytosis (Figure 4 and Figure 4---figure supplement 1). Thus, it is important to better establish sleep phenotypes arise from changes in endocytosis rather than non-specific glial malfunction. Some suggestions of additional experiments to support the hypothesised role of endocytosis are proposed below. > > 1a) To better characterise sleep-associated endocytosis, it would be useful to measure rates of endocytosis at 4-6 time points to define a daily cycle, not just two time points. We have expanded the nighttime time points and included these findings in (Figure 5B). We also show time points around the clock (Figure 5---figure supplement 2B). > 1b) Given the possibility that even adult-specific expression of mutant dynamin or Rab RNAi can have long-lasting changes in glial physiology or alterations in BBB permeability, it would be useful to test if these flies revert to normal sleep behaviour if allowed to recover for an appropriate period of time. To examine whether adult-induced expression of Shi produces irreversible alterations in sleep, which could signify permanent changes in function of the barrier, we expressed Shi in glia under the control of tubGal80.ts. In these experiments, temperature-dependent expression of tubGal80.ts suppresses Shi expression during development but allows expression in adults. We find that sleep is increased upon adult expression of Shibire and returns to levels which are not significantly different from controls when expression is blocked (in each case with a temperature shift) (Figure 3---figure supplement 1B). > 1c) To show that a different amount of sleep deprivation (more or less) produces a proportionate change in endocytosis (i.e., two or more points on the concentration-effect curve). Although this is a good point and we would certainly like to resolve this question, there are a number of scientific and technical difficulties that prevent us from doing so. First, more or less deprivation does not necessarily correlate with a linearly proportionate response, whether in behavioral rebound or cellular measures (for instance, 12 hours of deprivation does not necessarily produce a larger rebound than 6 hours). Second, due to the degree of variability in the endocytosis assay, we do not have the resolution to detect such a difference. If we expand our time points after a period of deprivation, we still may not be positioned to detect a difference. We have examined later night time points (Figure 5B) and found that while endocytosis is high during early night sleep and low by the end of the night/early day, it is difficult to determine a difference in the time points in between, meaning that endocytosis may be largely complete within the early hours of sleep or we simply do not have the resolution with this technique. > 1d) To show bidirectionality of regulation and to control for potential contribution of stress rather than sleep drive, the authors should test whether endocytosis following more sleep is reduced (this may require orthogonal genetics: e.g. to drive sleep-inducing FB neurons). To examine potential bidirectionality and assess the contribution of stress, we have elected to use Gaboxadol to induce sleep rather than introducing the added complications of orthogonal genetics and the requisite temperature-induction. We found that enhancing sleep throughout the day did not decrease endocytosis at ZT14, but endocytosis was increased at ZT2 if flies were treated with Gaboxadol prior to this point (Figure 5C). We suggest that Gaboxadol increases sleep pressure and so promotes endocytosis at all times, thereby accounting for lack of a decrease at ZT14. Importantly, these data indicate that it is not the stress of sleep deprivation which is responsible for elevating endocytosis, as two methods of inducing sleep at a time when wake predominates both increase endocytosis. > 2\) Different Gal4 lines are used to target subsets of glia. The specificity of these lines needs to be clearly documented. The logic for using one or other for different experiments needs to be clarified. And at least one case, an additional Gal4 may need to be used. In addition to the discussion of sub-points below, we have added a few remarks throughout the text clarifying why the given Gal4 lines were used for particular experiments. > 2a) The description of Gal4 expression lines (Figure 3---figure supplement 2) should be expanded Critically, to demonstrate that each one the Gal4s used is indeed expressed preferentially in glia, the figure should include counter stains with glia or neuron-specific antibodies (Repo and Elav, respectively). Glial-specific expression would be also more convincing if the authors could test how well the expression pattern of these Gal4s (Rab9-Gal4 in particular) is blocked by Repo\>Gal80. (More minor, it might help if an outline of the brain were drawn). The Gal4 lines used in this study, other than Rab9-G4, have all been previously described and used for glial expression. We have emphasized this point and ensured that appropriate references for all drivers are included in the text. The Gal4 lines we used (Figure 3---figure supplement 2) have limited expression, if any, outside the BBB; given that, it would be difficult to co-localize expression with a pan-neuronal or pan-glial stain. In lieu of a counter-stain, we have provided Gal4\>GFP expression in the presence and absence of *repo*-Gal80, which by comparison makes clear what expression is neuronal and what glial (Figure 3---figure supplement 2B). This is particularly evident for the drivers relevant for behavioral differences, such as Rab9-Gal4, as the surface glia are easily distinguishable from all other cell types. As requested by the reviewer, we provide an outline of the brain in Figure 3---figure supplement 2A. > 2b) The tubGal80.ts experiments utilize Rab9-Gal4. To more convincingly implicate the glial cells presumed to be targeted, these experiments should also be repeated with Moody-Gal4 (and possibly Repo-Gal4 as well). The SPG glial cells are implicated by two drivers, Rab9-Gal4 and Moody-Gal4. To confirm the adult-specific effect of Shi, we now provide the tubGal80.ts experiment with a second driver, Repo-Gal4 (Figure 3---figure supplement 1). The data confirm that conditional adult expression is sufficient for the glial sleep phenotype. > 3\) Figure 1 shows that of Shi-ts expression in glia induces resistance to mechanical stimulation during sleep. Can a control experiment be done to rule out the possibility of a sensory defect that makes these flies unresponsive to mechanical stimulation? E.g. To test the responsiveness of awake Repo\>Shi flies to mechanical stimulation? On this note, are the PG and SPG drivers expressed outside the brain in sensory structures? To rule out the possibility that a sensory defect in Repo\>Shi flies makes them unresponsive to the mechanical stimulation of sleep deprivation, we stimulated the experimental and control flies at ZT11 -- 12, a time when flies are predominantly awake, as suggested. We find no difference in the resultant beam crossings induced by stimulation between Repo\>Shi flies and controls (Figure 1---figure supplement 1), suggesting that resistance to sleep deprivation cannot be accounted for by impaired sensitivity to the stimulus. Also, we are not aware of sensory structure expression of PG and SPG. > Other concerns and suggestions: > > 4\) Is it necessary to dissociate cells before adding Alexa647-dextran? Why not measure before dissociation as most cellular processes including endocytosis could be more robust in an intact brain. If it is simply to allow access to Alexa647-dextran, then could an alternative smaller endocytic tracer be tried? While we agree that a minimally perturbed sample might better reflect cellular processes occurring in vivo, in our hands we have found that dissociation is necessary to allow access of AF647-Dextran to both sides of the barrier. Even with a smaller (3KD) dextran we did not observe any appreciable penetration into an intact brain (Figure 5---figure supplement 3). > 5\) Is resistance to mechanical stimulation following SD specific only to Shi, or also seen following endocytic Rab manipulation? We have added this experiment (Figure 4---figure supplement 2). Resistance to mechanical stimulation appears to be specific to Shi, with surface glial being sufficient. It is plausible that Shi provides a more general perturbation of trafficking than Rab11, hence capturing both baseline and resistance phenotypes. These data also demonstrate that these effects can be separated as Rab11 only affects baseline sleep. > 6\) The shi-ts expression results suggest that sleep during SD happens at both temperatures, but nighttime sleep seems to elevated sooner in both males and females at 18 degrees, than in controls. Is this impression correct? and if so, is there an explanation? This difference (if reproducible) involves latencies and time of day differences. The fly sleep literature appears to point to an increasing fragmentation of cellular mechanisms between those that affect daytime versus nighttime sleep. Is there any basis to think or expect that the cell biological mechanisms (exo- and endocytosis) should be geared to daytime versus nighttime? We are not sure that this difference in latencies that the reviewer mentions is a consistent finding in our data. The difference in latency between 18 and 30 degrees could also be a consequence of the shift in sleep patterns related to temperature, as touched on elsewhere in the text. With respect to time-of-day differences, it is true that some publications demonstrate effects predominantly on daytime or nighttime sleep when given neuronal populations are manipulated, but in our view, there is not yet a clear delineation between circuits and mechanisms which govern daytime vs. nighttime sleep. Likewise, there is little evidence to suggest that daytime sleep and nighttime sleep serve different functions in the fly. > 7\) Figure 2C suggests that the effects of Shi-ts on sleep are not due to merely overexpression of Dynamin since the expression of Shi-wt does not have a significant effect. However, for Shi-wt, only total sleep is shown, whereas the main effect for Shi-ts at 18C is on day sleep. To exclude the possibility that a Shi-wt effect on sleep is masked when only total sleep is shown the figure should be revised to include data showing the effect of Shi-wt on daytime sleep. The figure has been updated to include daytime sleep. The result is still the same -- Shi-wt is not significantly different from both controls, in terms of total or day sleep. > 8\) Indeed, it would useful to consistently show data for total sleep and daytime sleep for all genotypes in the paper (these data must be available). Daytime and total sleep figures have now been included, either in the main text or supplements in order to not crowd the figures. > 9\) In Figure 3A, to exclude possible effects due to expression of Shi in neurons (in addition to expression in the BBB glia), the authors could show that these phenotypes remain when the expression in all glia is blocked using Repo\>Gal80. We showed that the increased sleep phenotype is present with Repo-G4 (only glial expression) as well as Rab9-G4 (BBB expression), but not when glial expression of Rab9-Gal4 is blocked. Comparing expression of surface glial drivers (moody, Rab9, NP6293) in the presence and absence of *repo*-Gal80 (Figure 3---figure supplement 2B) shows little or, in the case of moody-G4, no expression in neurons, therefore making it quite unlikely that the phenotype is due to non-glial expression. > 10\) It would be valuable to test, with existing tools, to what extent surface glia are responsible for the phenotypes observed with the pan-glial driver, Repo. Does Shi expression in surface glia also cause resistance to mechanical sleep deprivation? Is the phenotype of either NP6293 or moody \>Shi weaker than that of Repo\>Shi and does combining NP6293 and moody Gal4s generate a phenotype comparable to that of Repo-Gal4? Why is the 9-137-Gal4 driver (SPG & PG) that is later used in Figure 5 (subsection "Endocytosis occurs during sleep and is influenced by prior wakefulness") not used here? We conducted additional experiments to address this question and find that the resistance to sleep deprivation phenotype indeed tracks to surface glia, with NP6293 G4 being sufficient (Figure 3---figure supplement 3). In the screen of glial sub-type drivers, we chose to use the individual surface glial drivers instead of 9-137 G4 in order to establish the effects of each population in isolation, just as the other glial drivers were chosen for specificity to individual glial sub-types. Nevertheless, Shi expression with 9-137 G4 also produces an increase in sleep, albeit smaller than would be expected if combining the effects of the surface glial lines individually (and smaller than that produced by repo-Gal4). This may be a consequence of differences in driver strength, which are not readily assessable or comparable. Given that NP6293 G4 and moody G4 are on the same chromosome, the effort required to perform these experiments may not be worth the potential result because there still may be a discrepancy in driver strength between the surface glial drivers and repo. > 11\) Figure 3 localizing effect to surface glia: In Figure 3A -- which UAS-shi was used? The effect does not appear to match that of repo-Gal4. The difference is no more than 100 min; whereas. in Figure 2A, the difference is \>200 min. Is this lack of quantitative matching significant? Maybe it reflects the strength of the Gal4s but alternatively it could mean that other cells contribute. This should be discussed. A line has been added regarding this point. As the reviewers point out, it may very well be a difference in driver strength, and/or that repo-G4 encompasses both surface glial populations while the lines in Figure 3A do not. > 12\) In the same figure -- the lack of effect of astrocyte expression seems to be correlated with an unusually high value for the UAS control. Is there a reason to be concerned here? We do not believe so as the UAS control does not achieve ceiling levels of sleep, so there is still room for an increase. Also, as sleep varies some from experiment to experiment, we are careful to always compare to controls within the same experiment. Importantly, a different astrocyte driver also does not increase sleep significantly above the controls. As these data used the other Shibire line, rather than the 20xShibire used in much of the manuscript and in the rest of Figure 3A, we did not include them in the manuscript, but provide them for the reviewer ([Author response image 1](#respfig1){ref-type="fig"}). {#respfig1} > 13\) Regarding EM images in Figure 4A -- how abundant and frequent are these ultra-structural changes and how do they correlate with sleep phenotypes? > > The observations themselves (amassed ring-like structures resembling microtubules, and a thinner PG layer) are only described to the level of positioning of two arrows in one of the micrographs. There should be some definition of the phenotypes and indication of prevalence, severity or quantification (either within or between animals). Were the suspected MTs coated or bridged as originally described in the paper cited? There no details provided in the methods regarding gender, time of day, age, or \"n\" value. The details requested regarding the animals used have now been provided in the methods and/or figure caption. Briefly, our intention is to highlight the presence of unusual MT and intracellular structure in glial cells expressing Shi at 18 degrees. This simply demonstrates that by EM, expression of Shi at permissive temperature is not without consequence in glia, as has also been shown for neurons in the paper cited. While perhaps we could quantify something like the number of MTs over a length of surface glial area, it is not clear that this or any other quantification will be a useful measurement, because the phenotype is binary -- the structures are present in Repo\>Shi tissue, and not found in controls. > In addition, it would be useful to know if these are also seen following induction of wild-type dynamin, because WT-dynamin expression does not cause sleep phenotypes and there is the potential to examine how relevant these ultrastructural defects might be to the sleep effects observed. (Perhaps the cited EM paper has addressed this issue in neurons?) Are there any changes in membranes or vesicular structures that could potentially be connected to a block in endocytosis? What is the predicted EM phenotype associated with Rab11CA expression? Our hope was to identify membrane or vesicular structures in the control surface glial images to have a target for comparison with the experimental images. Unfortunately, perhaps due to resolution or preparation, we were unable to pinpoint obvious endocytic or vesicular trafficking events in the control images, thereby making it hard to compare to Shi flies and to say, beyond the gross changes described, where the alteration connected to blocking endocytosis is expressed. Determining whether WT-dynamin causes ultrastructural defects could be interesting, but even if so, we could not exclude a contribution of such defects to the sleep phenotype produced by Shibire as these defects may be necessary but not sufficient. > 14\) Is the overall structural integrity of BBB, as shown in Figure 4B, also preserved when Rab11CA is expressed in glia? Yes, the barrier does not appear permissive to dextran when Rab11CA is expressed (Figure 4---figure supplement 2). > 15\) In Figure 4C, the significance is determined based on a difference between DN and CA allele of each Rab protein. Is this the best criterion? Wouldn\'t it be better to compare each manipulation to the parental controls? In any event, it should be made clear why only Rab11 and not others is deemed to have crossed a significance threshold. We agree that the ideal comparison would be between each manipulation and its respective parental controls. But since our initial purpose was to screen as many Rabs as possible (which included multiple CA and DN lines per Rab in most cases), we chose to compare CA and DN effects of the experimental animals only. As noted below, we expected that CA and DN would have opposing effects, although this is not always the case. We followed up Rabs that were promising in the initial screen by using Repo-G4 and including all of the parental controls. Rab11 was a positive in the DN/CA screen and, in subsequent assays, Rab11 CA consistently increased sleep relative to controls, regardless of whether it was expressed with Repo or BBB drivers. This is now clarified in the text. > 16\) The overall logic that explains the link between the manipulations, endocytosis and sleep is confusing. Glial expression of Rab11CA increases sleep, similarly to Shi-ts. Do the authors propose that Rab11CA inhibits endocytosis, similarly to Shi-ts? Rab11 can affect endocytosis, as knockdown of Rab11 has been shown to inhibit endocytosis and transcytosis, for example, of LDL (Woodruff et al., 2016). Assuming that CA and DN have opposing effects, Rab11CA would not be expected to block endocytosis; however, DN and CA versions can sometimes go in the same direction, with WT or CA mutants blocking the pathway at a specific point due to limiting amounts of binding partners. In addition, for the Rab collection we used, the Rab CA and DNs are called so based on the best prediction that the sites mutated will be GTPase-defective and GTP-binding defective, rather than functional confirmation of their CA and DN activity; in fact, Rab3 is mislocalized with CA and DN mutations (Zhang et al., 2007).In the case of Rab11, in particular, both the DN and CA mutations we used alter trafficking in one direction in certain contexts (Khvotchev et al., 2003). Therefore, it is possible that Rab11 CA inhibits endocytosis. Alternatively, if the mechanism underlying the sleep phenotype involves, for example, a certain receptor, it is also possible that Rab11 CA and Shi lead to the same behavioral phenotype by keeping the receptor on the plasma membrane -- Rab11CA by promoting its recycling, and Shi by inhibiting its endocytosis. > 17\) On similar lines, the data shown in Figure 4C seem to contradict the data in Figure 4D. In Figure 4D, Repo\>Rab11CA causes increased sleep, compared to parental controls. Figure 4C, shows that using the RepoGS driver, Rab11 DN induces even more sleep than Rab11CA. Thus, both of these opposing manipulations cause increased sleep. Could it be that both manipulations simply make glial cells sick, and therefore, in both cases, sleep is increased due to malfunction of glia, and not due to specific changes in endocytosis? The possibility for a non-specific effect is supported by similar data in Figure 4---figure supplement 1C, where expression of either DN or CA versions of Rab27 and Rab30 produces exactly the same phenotype of increased sleep, unless the authors believe that these also work by affecting endocytosis. These issues should be explained and addressed in a revised manuscript. We have driven Rab27 and Rab30 with the surface glial drivers, but have not found consistent phenotypes when limiting expression to these glial populations. Thus, Rab11 is specific in terms of increasing sleep through this population, as sleep-promoting effects of other Rabs likely map to other glia. The similarity of increased sleep may therefore suggest sleep functions of Rabs in other glial populations. Concerning the discrepancy between RepoG4 and RepoGS sleep effects of Rab11CA and DN, it is difficult to compare as Rab11 DN is lethal with constitutively expressed Gal4s, while CA is not. As we have stated above, it is possible that CA and DN mutations do not really have opposing effects at the cellular level. In this case one might argue that CA is simply a weaker blocking mutation than DN. Nevertheless, Repo\>Rab11CA flies do not appear to be sick and show intact general integrity of the BBB (Figure 4---figure supplement 2). > 18\) In Figure 4---figure supplement 1B- are the effects of Rab3, Rab9 and Rab11 that were significant in Figure 4A (without RU486) still significant in the Figure 4---figure supplement 1B (with RU486)? The significance stars in Figure 4---figure supplement 1B are shown only for Rab1, Rab5, Rab27 and Rab30. Our standard for selection, as further clarified in the manuscript, is that all available DNs be significantly different from all available CAs. By this measure Rab3, 9 and 11 are different in the no-RU condition, and 1, 5, 27 and 30 are such in the RU condition, hence the difference in stars. Nevertheless, if we look at individual comparisons for the aforementioned Rabs in the RU+ condition, we find that the lone DN line of Rab3 is significantly different from one (of two) CA lines, a single DN line for Rab9 is significantly different from both CAs, and a single DN line for Rab11 is significantly different from both CAs. > 19\) The apparent contradiction between Figure 1, Figure 2, Figure 3, Figure 4 and Figure 5 should be more clearly explained. Figure 5 shows that SD (which increases sleep pressure) causes increased endocytosis, but, Figure 1, Figure 2, Figure 3 and Figure 4, show that inhibiting endocytosis actually increases sleep, suggesting that endocytosis in BBB promotes wakefulness. An explanation that connects the two observations should be included. To address this point we have built upon a relevant line in the Discussion section. In short, we propose that endocytosis in the BBB is a function of sleep and serves a homeostatic need. As shown in Figure 5, endocytosis is enhanced during times of baseline sleep, or during rebound sleep induced by sleep deprivation. If endocytosis during sleep fulfills a homeostatic role, it would follow that endocytosis would decline as sleep continues (see Figure 5B), just as sleep pressure does. In the experiments where we inhibit endocytosis at the surface glia, we believe we have created a situation where the potential homeostatic function of endocytosis cannot be accomplished, hence these flies may not be able to dissipate sleep pressure, and therefore have perpetually increased sleep. > 20\) Are the changes in endocytosis that correlate with sleep state specific only to surface glia? Are there changes in endocytosis in other populations of glial cells, as function of sleep? Our interest in endocytosis in the surface glia stems from our findings that Shi expression in these populations produces changes in sleep behavior. This does not exclude the possibility of sleep-dependent changes in endocytosis in other glial populations (or even neuronal populations). For example, phagocytic engulfment may be altered in astrocytes by sleep deprivation (Bellisi et al., 2017).

Introduction {#Sec1} ============ The diagnostic approach is statistical\[[@CR1],[@CR2]\]. Age is important: before 5 years old, malignant tumour is almost always metastatic neuroblastoma; between 5 and 15 years old, it is often osteosarcoma or Ewing's sarcoma; after 40 years old, it tends to be metastasis or myeloma. Clinical symptoms vary little and are most often pain and swelling. Although fever suggests infection, it may also be found in Ewing's sarcoma. The first step in the evaluation of a tumour is to determine its aggressiveness by conventional radiology. Important parameters include the tumour size, type of matrix, and periosteal reaction. Certain tumours are more common in particular bones. Adamantinoma, usually found in the adult, selectively involves the tibia and fibula. The most common epiphyseal tumour in childhood is the chondroblastoma. In addition, the aggressiveness of some tumours may relate to their location in the axial or appendicular skeleton: in the hand, cartilaginous tumours are almost always benign, whereas in the pelvis, they are often malignant. If necessary, multiple lesions may be estimated with bone scanning. Multiple lesions are seen in chondromas, osteochondromas, histiocytosis X and metastases. The first necessary step is to definitively diagnose benign lesions based on clinical and radiologic signs, and for which biopsy is not necessary. These lesions are fibrous cortical defect, non-ossifying fibroma, periosteal desmoid, fibrous dysplasia, osteochondroma or exostoses, chondroma, simple bone cyst, vertebral angiomas and myositis ossificans. Diagnosis may be difficult. In these cases, the next step is CT. Problems can result from bone locations which are difficult to evaluate on conventional X-ray (short, flat bones especially the pelvis, sacrum, sternum, and vertebrae). Sometimes, the study of the tumour matrix can provide features necessary for the diagnosis. It can be fluid density, small calcifications allowing diagnosis of cartilaginous tumour, osteoid matrix, or fat\[[@CR3]\]. CT is the examination of choice in the diagnosis of the nidus of osteoid osteoma in dense bone\[[@CR4]\]. Small lytic lesions of the cortex, localized involvement of the soft tissues and thin peripheral periosteal reaction can be seen; lesions with slow evolution which displace and expand the cortex peripherally can be distinguished from more aggressive lesions which cross the cortex. CT can show the tumour on both sides of the cortex before it is destroyed. This is the case for Ewing sarcomas and osteosarcomas\[[@CR5]\]. CT allows measurement of the thickness of a non-calcified cuff of a cartilaginous tumour: the cuff is thin in benign lesions and thick (more than 3 cm) in chondrosarcomas\[[@CR6],[@CR7]\]. In other cases, evidence favours a malignant lesion. Examination for metastases, and MRI examination before biopsy, must be performed. Staging examination {#Sec2} =================== Local {#Sec3} ----- Focal extent and staging is based on magnetic resonance imaging (MRI)\[[@CR8]--[@CR10]\]. The main advantages are high contrast and the possibility of choosing the plane of examination without moving the patient. On the other hand, MRI is rarely useful for diagnosis. T1 and T2 measurements are not reliable and reproducible because most large tumours are heterogeneous with variable signal intensity. Therefore, it is impossible to characterize such tumours and to distinguish benign from malignant lesions\[[@CR11]\]. MRI is rarely useful in the diagnosis of fluid levels in blood-filled cavities, especially anevrismal bone cysts. Intramedullary extension {#Sec4} ------------------------ In the medullary cavity, diaphyseal and metaphyseal extension in both adult and child, and extension across the growth plate in the child, are best demonstrated on MR images, due to their excellent contrast and ability to image in the longitudinal plane. Although CT can also evaluate extension, it is limited to the axial plane. Both CT and MRI have other limitations: the inability to detect very small lesions and the overestimation of the tumour volume on T2-weighted sequences because of peritumoral edema\[[@CR12]\]. With accurate evaluation of tumour extent, the surgeon can determine the level of bone resection and the size of the prosthesis. The growth plate in children and the joints in adults can sometimes be preserved when uninvolved. In periosteal tumors, MRI demonstrates the periostal location, its extension into cortex, and the medullary cavity\[[@CR13]\]. CT can also define extent of diaphyseal periosteal lesions, but not of metaphyseal periosteal tumours. Intraarticular extension is detected with better sensitivity than with any other imaging technique because of direct visualization of joint cartilage. Skip lesions (small metastases separated from the main tumourby healthy tissue)are easily detected on MR scans parallel to the long axis of the bone. MRI also shows excellent demonstration of vessels and their relationship to the tumour without injecting contrast media. Both CT and MRI can show skin and subcutaneous extension. In summary, MRI should be used as the principal test forevaluating extension of malignant tumours. Examination of distant spread {#Sec5} ----------------------------- Bone metastases are best detected on radionuclide bone scans. Pulmonary metastases are evaluated on conventional chest radiographs and chest CT\[[@CR14]\]. Conclusion {#Sec6} ========== Imaging should begin with the plain radiograph, which defines a lesion as benign and requires no treatment. When the type of matrix needs to be clarified, CT should be performed. MRI however, should be the primary study for staging the tumour extent. This paper is available online at <http://www.cancerimaging.org>. In the event of a change in the URL address, please use the DOI provided to locate the paper.

Greater than normal prevalence of seropositivity for Helicobacter pylori among patients who have suffered myocardial infarction. There is evidence to suggest that inflammation plays a role in the development of atherosclerosis. Chronic infections may activate an inflammatory response in the walls of blood vessels. To investigate the possibility of there being an association between infection with Helicobacter pylori (H. pylori) and coronary heart disease. We examined 100 consecutive patients documented to have recently suffered acute myocardial infarction and 100 control subjects from the same geographical area for whom there was no evidence of coronary heart disease, carefully matched both for age and sex. Blood samples were tested for the presence of immunoglobulin G antibodies against H. pylori with a serological test. In comparison with controls, patients were more commonly smokers (26 versus 12%/0, P < 0.05) and had more commonly been treated for hypertension (37 versus 20%, P< 0.01). There was a significant association between seropositivity for H. pylori and having previously suffered acute myocardial infarction (68 versus 53%, odds ratio 1.36 with 95% confidence interval 1.02-1.82, P=0.034). These findings remained valid in a multivariate analysis including possible confounding factors (age, sex, smoking and hypertension; odds ratio 1.35 with 95% confidence interval 1.01-1.83, P=0.046). The positive association between seropositivity for H. pylori and having previously suffered acute myocardial infarction found in this study provides further support for the hypothesis that there is a causal association between chronic infection with H. pylori and the development of coronary heart disease.

# Amahi Home Server # Copyright (C) 2007-2013 Amahi # # This program is free software; you can redistribute it and/or # modify it under the terms of the GNU General Public License v3 # (29 June 2007), as published in the COPYING file. # # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # file COPYING for more details. # # You should have received a copy of the GNU General Public # License along with this program; if not, write to the Amahi # team at http://www.amahi.org/ under "Contact Us." require 'socket' # Sockets are in standard library class DiskUtils # return information on hdd temperature - requires hddtemp service running! class << self def stats host = 'localhost' port = 7634 begin s = TCPSocket.open(host, port) rescue return [] end res = '' while line = s.gets # Read lines from the socket res += line.chop # And print with platform line terminator end s.close # Close the socket when done disks = res.split '||' res = [] disks.each do |disk| # split the info and do cleanup i = disk.gsub(/^\||\|$/, '').split('|') model = i[1].gsub(/[^A-Za-z0-9\-_\s\.]/, '') rescue "Unkown" next if model == '???' t = i[2].to_i rescue 0 tempcolor = "cool" celsius = "-" farenheight = "-" if t > 0 celsius = t.to_s tempcolor = "warm" if t > 39 tempcolor = "hot" if t > 49 farenheight = (t * 1.8 + 32).to_i.to_s end d = Hash.new d[:device] = i[0] d[:model] = model d[:temp_c] = celsius d[:temp_f] = farenheight d[:tempcolor] = tempcolor res.push(d) end res end def mounts s = `df -BK`.split( /\r?\n/ )[1..-1] || ["","Incorrect data returned"] mount = [] res = [] s.each do |line| word = line.split(/\s+/) mount.push(word) end mount.each do |key| d = Hash.new d[:filesystem] = key[0] d[:bytes] = key[1].to_i * 1024 d[:used] = key[2].to_i * 1024 d[:available] = key[3].to_i * 1024 d[:use_percent] = key[4] d[:mount] = key[5] res.push(d) unless ['tmpfs', 'devtmpfs', 'none'].include? d[:filesystem] end res.sort { |x,y| x[:filesystem] <=> y[:filesystem] } end end end

This invention relates to the field of large video display systems of the type appropriate for installation at a stadium. Such displays are usually formed by a large matrix of variable intensity display devices as, for example, incandescent bulbs, which are driven by a display system usually computer controlled. The display system receives a video input, such as the line feed from a network broadcast or a video tape recording and digitizes the video information into a complete frame of digital data. In prior systems the digitized data was stored in a computer memory and then at an appropriate point transferred from memory to the display device. Computers utilized for such a system include the Digital Equipment Corporation PDP Series 8. Although such mini computers are relatively powerful devices, their data transfer rate, as compared in random access memories, is low. As a result the computer represents a limiting element in the system with respect to the rate at which data can be digitized and transferred to the display device thereby limiting the versatility of the system with respect to other desirable features, such as maintaining statistics on participants, displaying caricatures, cartoons, or still photographs of the players. It is therefore a desirable objective to retain computer control of the display system but to remove the computer from the data path to the display board. The PDP computer referred to employs data break cycles to transmit information to the display board. It does not have time to do the other tasks as mentioned, such as disk storage input and output, statistical updating and message inputting in addition to refreshing the display. Furthermore, when the computer is included in the data path it is necessary to synchronize the computer to the master clock. This slows down processor time still further. It is accordingly an object of the present invention to provide a display system which is capable of higher speed than prior devices by virtue of removing the computer from the data path. By so doing the computer is able to perform a variety of other tasks as indicated. More importantly, it is possible to obtain a heretofore unavailable display in which a portion of the video picture being displayed can be enlarged to permit better viewing thereof. Often this is analogized to "zooming" in the manner permitted by an adjustable lens in photography. To obtain an enlarged or zoom picture it is necessary to utilize greater data rates than present systems can handle. The present invention, by removing the computer from the data path, is capable of operating at these higher data rates. Exemplary of prior systems for displaying data on large display devices are U.S. Pat. Nos. 4,009,335, 3,941,926, and 3,961,365 assigned to the assignee of the present invention.

Thursday, February 14, 2019 EPISODE 291 This week we find 2 truly awesome hidden gems, and one bona fide classic. First we talk about a straight to video action flick that Kyle and John neglected to watch in Lady Avenger. Then we talk about a bonkers 80's ghost/supernatural/haunted house/ ghost flick in SUPERSTITION. And then John shows Kyle BAY OF BLOOD for the first time. Also, we talk about beautiful Italian girls, Bernie Mac, Pittsburgh Geography, The geneva convention and it's connection to war crimes, and lousy Queen movies. So Download this episode or Bohemian Rhapsody will not win the Oscar.RIGHT CLICK TO DOWNLOAD

Fed Credit Score Study Bogs Down A Federal Trade Commission report on the insurance industry's controversial use of consumer credit records to rate them as risks will not be ready ... By Matt Brady|March 11, 2005 at 07:00 PM X Share with Email sending now... Thank you for sharing! Your article was successfully shared with the contacts you provided. A Federal Trade Commission report on the insurance industry’s controversial use of consumer credit records to rate them as risks will not be ready by its federally mandated December deadline The report was called for in the Fair and Accurate Credit Transactions, or FACT Act of 2003, which extended provisions of the Fair Credit Reporting Act and implemented a number of reforms designed to help consumers track and protect their credit information. Added to the legislation through an amendment by Rep. Luis Gutierrez, D-Ill., the report is officially due on Dec. 4. The insurance industry describes credit scoring as a useful tool, predictive of risk that helps keep rates down. Opponents say it impacts hardest on low income and minority consumers and fails to account for onetime events that can impact credit. A spokesman for Rep. Barney Frank, D-Mass., said the report was delayed because the FTC “needs more time to pull together information” from insurance companies. Rep. Frank is the ranking member of the House Financial Services Committee, where the FACT Act, and the amendment mandating the study, originated. However, David Snyder, vice president and assistant general counsel for the American Insurance Association, said the insurance industry is not the cause of the delay. The AIA had heard “some speculation” the report might not be ready on time, he said, but did not know of any specific reason for the delay. “One reason it cannot be is for any lack of cooperation on our part,” he said, adding that the AIA’s policy has been to cooperate fully with the FTC and provide whatever information it is asked for. The credit scoring issue has generally been handled at the state level with many states adopting a model regulation crafted by the National Conference of Insurance Legislators, or something similar to it. Efforts to ban the use of credit scoring in Arkansas, Colorado, Montana, Texas and Washington were defeated in 2005. Montana instead adopted a measure based on the NCOIL model, and New Mexico passed its own legislation that was supported by the insurance industry. In Michigan, Gov. Jennifer Granholm and Democratic members of the state House and Senate said this month they will include a provision prohibiting insurers from using a consumer’s credit history or credit score for determining their insurance rates in a legislative insurance reform package. A prior attempt to bar the use of credit scores through a state Office of Financial and Insurance Services regulation was overturned by a state circuit court judge earlier this year.

2014 American documentary, filmed and narrated by wartime journalist's father and son, Mike Boettcher and Carlos Boettcher. Summary: In real life there are no re-spawns, this is no call of duty. A fantastic real life account of the ever present war on terror being fought in Afghanistan. A film that takes you deep into 'The hornets nest', deep into situations you wouldn't even dream about, situations where all hope for humanity is lost, where words cannot describe the horrors that are present, where all faith is questioned. We follow father and son journalist's Mike and Carlos Boettcher as they spend a year with the real heroes the 101st airborne in Afghanistan. This is not based on a true story, this is a true story. A sight we always see, a report we always hear in our every day lives, but do we ignore it? or is it seemingly not real that there is a war going on? That is just one of many questioned posed in this real life insight to the ever present war on terrorism. Although stated as 'directed by' David Salzberg and Christian Tureaud "The Hornets nest" is filmed entirely by father and son Mike and Carlos Boettcher, and is an eye witnessed account of their experience living, breathing and doing everything these soldiers of the 101st airborne do day in and day out in Afghanistan. Mike Boettcher is a journalist who has spend the best part of 37 years reporting from some of the worst war zones in history, he has spend the majority of his reporting life in the deep end bringing people at home the truth of war. As well as being an account of the war in Afghan, we follow Mike's quest to build a lost relationship between him and his son Carlos, who in attempt to rebuild this relationship embarks on this journey with his father. The journey is narrated by Mike throughout, as he gives a step by step as to what he and his son were witnessing and feeling. This however, proved to be a bit misleading, as the film suffers a lack of time line, because we forget that they were based in Afghan for over 12 months, and the jump between story is rather sporadic. However, this is understandable as there is clearly a , countless amount of footage, but clarity mixed with overtly dramatic music throughout proves to be a bit confusing. Although this does not detract from the fact that what we as an audience are viewing is real, and actually has actually happened. The reality of what is occurring in Afghanistan is emphasised early on, when we are with Carlos interviewing a couple of soldiers regarding a road block, it is at this point soldiers are fired upon. The film is predominantly based in the notorious 'Korangal Valley' known by the locals as 'Death Valley', a fitting and obvious title. Mike and Carlos both follow a unit who are sent out to clear the pass of snipers as this pass is used to transport non lethal supplies to the soldiers. Almost instantly the soldiers are fired upon, and also both Mike and Carlos receive heavy fire and must continue to film whilst taking cover. It is at this point that Carlos must move to better cover up at the top of the hill as they cannot ascertain where fire is coming from, we follow him running up the hill, only to fall down whilst fire is still coming over head, and to our horror he stops moving. There are moments throughout this documentary that the relationship between father and son (Mike and Carlos) is really drilled home, and the notion of parent and child, which allows us as an audience to understand how these soldiers are someone's friend, father, husband/partner. The climax of this story is when Carlos leaves Afghan ahead of his father, and Mike stays on to follow the 101st airborne into "The Hornets Nest", the Hornets nest is a term used to describe an ambush where you are surrounded, and the centre of an ambush. For 9 days the 101st stand and fight back against Muslim militia. Although as a film there are many cinematic mishaps, and areas where the continuity and narrative arch are complicated, it is the reality of war that makes this documentary award winning. Finally, it is the end of the documentary which draws all glamour, fiction and Hollywood of war away from this documentary. And with one final bullet wound to the viewer, Mike Boettcher shows us the real loss of war, as we witness the funeral of the fallen men who died during this nine day battle of 'The Hornets Nest', a truly moving sequence. Clint Eastwood : I want the troops from Great Britain and the U.S. to be successful, but by the same token, Afghanistan has always been a screw-up.

Estrogen regulates the growth of a proportion of breast cancers through an estrogen receptor (ER) mediated mechanism. Loss of the ER is associated with a loss of estrogen (E2) regulation and a resistance to antiestrogen therapy. Since no effective therapeutic intervention is available to treat the ER negative patient new initiatives are essential to develop a new treatment strategy. One approach would be either to reactivate the ER gene or to develop a targeted "gene therapy" to re-introduce constitutive production of ER. Our strategic goal is to determine whether the ER could reassert control over the growth of receptor negative breast cancer cells. We are the first group to transfect a breast cancer cell line (MDA-MB-231) with sense and antisense cDNA for both mutant (valine replacing glycine at AA400) and wild type ER. The 8 different cloned cell lines, that constitutively produce ER, were selected by polycistronic production of mRNA for aminoglycoside phosphotransferase which confers resistance in media containing G418. The ER is functional and can activate vit ERETK-CAT however the wild type transfectant is approximately 20 times more sensitive that the mutant transfectant. The ER levels, determined by whole cell uptake of (3 HE-2) are in the range (300-520 fmole/mg protein) of MCF-7 breast cancer cells (500 fmole/mg protein). Interestingly enough E2 inhibits growth of all the transfectants in a concentration related manner. The inhibition is blocked by the pure antiestrogen ICI 164,384. In contrast an analog of the triphenylethylene 4 hydroxytamoxifen (40HT) is a partial agonist/antagonist in the wild type transfectants but an estrogen agonist in the mutant transfectants. The aims of the proposal are to characterize the cell cycle kinetics and cell biology of our current and additional transfectants and to determine the effect of increasing the receptor number in other breast cancer cell lines. Colon cancer cell line HT-29 will be transfected with ER to broaden our finding to other target organs not previously regulated by estrogen. Additionally we have been intrigued by the alteration in pharmacology of 40HT with the mutant receptor; we will establish a model for antiestrogen drug resistance based on alterations in the ligand, receptor and hormone response element. There is evidence that the inhibitory effect of E2 on ER transfected receptor negative cells is universal so these studies will provide the first solid mechanistic evidence to pursue a new therapeutic approach to breast cancer. In the future the pharmacologist may not only alter the ligand requirements but also be able to target or reactive quiescent receptors as a new general therapeutic strategy for cancer.

Check out our new site Makeup Addiction add your own caption add your own caption add your own caption add your own caption add your own caption add your own caption I got 99 problems and chromosomes

1. Field of the Invention The present invention relates to a fixing device and an image forming apparatus and, more particularly, to a fixing device that includes a cleaning unit that presses an elongated cleaning web running from a supply core to a take-up core against a fixing roller of a fixing device that includes the fixing roller provided with a heat source and a pressure-applying roller that is in pressure contact with the fixing roller, thereby cleaning the fixing roller. 2. Description of the Related Art An image forming apparatus such as a copier, a printer, or a facsimile typically employs a fixing device of a heating-roller type that fixes a not-yet-fixed toner image on a recording medium, such as transfer paper, with heat and pressure by bringing a fixing roller (i.e., a heating roller) and a pressure-applying roller into pressure contact with each other and causing the recording medium to pass through between the rollers. Known examples of this conventional type of fixing device include a fixing device disclosed in Japanese Patent Application Laid-open No. 2003-255745. This fixing device includes a web-type cleaning unit that cleans a surface of a fixing roller by bringing the fixing roller into sliding contact with a cleaning web that is being taken up from a supply core, on which the cleaning web is wound, onto a take-up roller to prevent toner from being peeled off from transfer paper and adhering to the fixing roller (toner offset). The web-type cleaning unit uses heat-resistant fiber such as nonwoven textile impregnated with a releasing agent such as silicone oil as the cleaning web, thereby not only wiping off toner adhering to the surface of the fixing toner with the cleaning web but also supplying the releasing agent from the cleaning web to the surface of the fixing roller. The web-type cleaning has cleaning performance superior to other type of cleaning such as roller-type cleaning that brings a cleaning member, a roller, into contact with a surface of a fixing roller or felt-type cleaning that brings a cleaning member made of felt into sliding contact with a fixing roller. The cleaning unit of the conventional fixing device is configured to maintain cleaning performance by feeding only a minuscule length (1 mm, for example) of the cleaning web each time a predetermined number of recording sheets are subjected to fixing operation so that a new portion, which is impregnated with silicone oil, of the cleaning web is regularly brought into sliding contact the fixing roller. However, the web-type cleaning unit of the fixing device disclosed in Japanese Patent Application Laid-open No. 2003-255745 has a disadvantage. That is, because oil seeps from an outer-radius portion of the rolled-up cleaning web and moves toward the take-up core which is a core of the cleaning web, an amount of oil absorbed and retained in the cleaning web wound on the take-up core in a roll form is large in an inner-radius portion but small in the outer-radius portion. For example, as illustrated in FIGS. 5A and 5B, when a cleaning web has an overall length of 32 m, an amount of oil absorbed and retained in a portion of the cleaning web, which corresponds to an outer-radius portion of the cleaning web wound on a take-up core, within 0.5 m from a leading end of the cleaning web is 10% to 20% smaller than a lower threshold of a target range, while an amount of oil absorbed and retained in a portion of the cleaning web, which corresponds to an inner-radius portion of the cleaning web wound on the take-up core, within 0.5 m from a trailing end of the cleaning web is 10% to 20% larger than an upper threshold of the target range. There is a problem in the portion of the cleaning web within 0.5 m from the leading end where the amount of oil is small. That is, because oil on the surface of the fixing roller becomes insufficient and cleaning performance is delivered insufficiently, toner offset onto the surface of the fixing roller occurs, which undesirably results in contamination of an image. Furthermore, when toner offset onto the surface of the fixing roller occurs, the moved toner can be accumulated on a thermistor that contacts the fixing roller and undesirably damage a surface layer of the fixing roller. There is also a problem in the portion of the cleaning web within 0.5 m from the trailing end where the amount of oil is large. That is, because oil on the surface of the fixing roller becomes excessive, toner escaping occurs, which undesirably results in contamination of an image. Put another way, there is a problem that because oil absorbed and retained in the cleaning web is deficient in the portion near the leading end and excessive in the portion near the trailing end, the surface of the fixing roller becomes nonuniform in anti-adhesive properties. Therefore, there is a need for a fixing device and an image forming apparatus capable of equalizing anti-adhesion properties across a surface of a fixing roller whichever portion of a cleaning web is fed to clean the surface.

MANILA, Philippines — Malacañang is encouraging the public to express their views on state policies following the release of a poll indicating that more than half of Filipinos think printing or broadcasting anything critical of the administration is dangerous. A Social Weather Stations nationwide survey conducted from June 22 to 26 found that 59% of Filipino adults agree that they can say anything they want openly and without fear, even if it is against the administration. Only 18% disagreed with the statement for a "very strong" net agreement score of +41. However, the poll also found that 51% of Filipinos agreed that it is "dangerous" to print or broadcast anything critical of the administration, even if it is the truth. Only 29% of the 1,200 respondents did not agree with the statement for a net agreement of +31. READ: Reports critical of Duterte administration being pulled down — report Presidential spokesman Salvador Panelo said Malacañang was "curious" as to why 51% of the respondents thought it was dangerous to publicize anything critical of the administration. He said there is no prior restraint or punishment for those who practice the freedom to dissent in the Duterte presidency. Advertising Scroll to continue "The president respects criticisms as long as the same is not baseless, unfounded or false. He even urges the people, including writers and reporters, to freely express whatever sentiments they have," Panelo said in a statement issued Saturday night. "As long as the speech or expression is within the ambit of the constitutional guarantees, it will not face any government interference," he added. READ: SWS: More Filipinos get news from Facebook than from radio, newspapers Panelo said dissent against the administration and its policies is "loud and not curtailed," citing anti-Duterte opinions in the media and social networking sites. "We reiterate our call to all Filipinos not to hesitate in articulating their thoughts on government policies, critical or otherwise," he said. "Ours is a president who not only respects everyone's right to free speech but listens to our plight as he ensures that the fibers of our country's democracy are preserved and enhanced." READ: Filipino journalists critical of gov't continue to face harassment — US State Dept Panelo claimed the survey indicating that 59% of Filipinos think they can say anything they want openly and without fear was a repudiation of the opposition's claim that freedom of expression is being curtailed by the administration. "The survey rating means we have a vibrant and robust exercise of those freedoms," the presidential spokesman said. Palace: Respect Duterte's appointment power The results of the SWS poll were released two days after Duterte hurled a string of insults at Sen. Richard Gordon for criticizing his appointment of former military officials to civilian posts. Last Thursday, Duterte said Gordon is a "smart a**" who walks like a penguin. He also claimed the senator's stomach is "a fart away from disaster." READ: Duterte said Gordon ran for vice president. He didn't. Panelo insisted that the president's appointment power should be respected. He also defended Duterte's preference for former military officials, saying they are known for their discipline. "Appointment is both a privilege and a right of a president. We have to respect it... If a soldier retires, he becomes a civilian," he said. "You should be in favor of them because they are known for their discipline. Civilians do not have discipline. They argue too much." READ: Gordon swallows Duterte's comments about stomach Panelo claimed Gordon's experience won't deter lawmakers from expressing views critical of the administration. He said opposition Sen. Leila de Lima still criticizes the president even if she is in jail. "Filipinos are not afraid to criticize. Look at social media. The criticisms of those who do not like him (Duterte) are relentless. That's our nature," Panelo said. READ: Filipino journalists identify poor wages, cyber attacks as key threats — report

199 N.J. Super. 329 (1985) 489 A.2d 715 STATE OF NEW JERSEY, PLAINTIFF-APPELLANT, v. SAMUEL VERDUCCI, DEFENDANT-RESPONDENT. Superior Court of New Jersey, Appellate Division. Argued January 29, 1985. Decided February 20, 1985. *330 Before Judges MICHELS, PETRELLA and BAIME. *331 Janet Berberian, Assistant Prosecutor, argued the cause for appellant (George L. Schneider, Essex County Prosecutor, attorney; Hilary L. Brunell, Assistant Prosecutor, on the brief). Thomas C. Brown argued the cause for respondent. PER CURIAM. This is an appeal from an order entered by the Superior Court, Law Division reducing defendant's custodial sentence because of illness and releasing him pursuant to R. 3:21-10(b)(2). The trial judge's decision was based upon expert psychiatric testimony which disclosed a substantial likelihood that defendant would attempt to commit suicide were he to remain incarcerated. The State contends that defendant poses a substantial potential threat to the public and that the trial judge mistakenly exercised his discretion. We agree and reverse. The salient facts are not in serious dispute. Defendant has an extensive record of delinquency both as a juvenile and as an adult. On July 21, 1978, defendant was sentenced to concurrent 12 month terms for atrocious assault and battery and obstruction of justice. Defendant subsequently escaped from the Essex County Corrections Center and was ultimately apprehended in Ohio. Upon his return to New Jersey, defendant entered pleas of guilty to assault with an offensive weapon and escape. For those offenses, defendant was sentenced to an aggregate term of not less than three nor more than five years in New Jersey State Prison. The sentences were to run consecutively to the custodial term remaining on the prior charges. Defendant's motion for a reduction of sentence was subsequently granted and he was placed on probation. The offenses which are the subject of this appeal were committed while defendant was under probationary supervision. On April 3, 1980, defendant forcibly entered Makar's Jewelry and Gift Shop and robbed the owner at gunpoint. Although the record pertaining to this offense is somewhat sketchy, it is undisputed that defendant fired a shot at the victim narrowly *332 missing him during the course of the robbery. The record reflects that defendant threatened to kill the victim and made off with approximately $30,000 worth of gold jewelry and $1000 in cash. Defendant ultimately entered pleas of guilty to first degree robbery (N.J.S.A. 2C:15-1) and unlawful possession of a handgun (N.J.S.A. 2C:39-5b). Defendant also pleaded guilty to a separate indictment charging him with theft by receiving stolen property (N.J.S.A. 2C:20-7) and unlawful possession of a firearm (N.J.S.A. 2C:39-5b). While awaiting disposition of charges pertaining to his violation of probation, defendant escaped from the holding cell adjacent to the courtroom and fled to Florida where he was ultimately apprehended after committing the crime of grand theft. On March 26, 1982, defendant was sentenced to an aggregate custodial term of 15 years. Under the sentences imposed, defendant was to serve a term of four years before being eligible for parole. Unfortunately, the record discloses that defendant has abjectly refused to comply with prison rules from the very commencement of his incarceration. His encounters with prison officials need not be recounted at length. Suffice it to say, defendant's derelictions have been substantial and include possession of a weapon, abusive conduct, fighting and attempted escape. It is undisputed that defendant's misbehavior has resulted in his being placed in administrative segregation for the greater part of his incarceration. The record clearly reveals defendant's apparent inability to cope with the realities of prison life. Uncontradicted evidence was presented that defendant made a serious attempt at suicide and came perilously close to succeeding. Apparently, this problem is not of recent vintage. The record reflects that defendant made a similar attempt several years ago when he was incarcerated for another offense. Whatever its etiology, it is beyond dispute that the risk of suicide is real. *333 On December 10, 1984, defendant moved for a reduction of sentence pursuant to R. 3:21-10(b)(2). Defendant contended that he suffered from a mental disease or emotional illness which rendered him incapable of conforming to prison life. It was alleged that defendant presented a major suicidal risk. In support of that claim, Dr. Thomas Latimer, a psychiatrist, testified that defendant was a "prime candidate" for suicide because of his "severe" psychiatric problems. While acknowledging that defendant was not "actively psychotic" and was neither "delusional" nor "paranoid," the witness ascribed his past misbehavior to a "basic immaturity." According to Dr. Latimer, defendant's illness precluded him from being able to "conform his conduct [to prison rules], to take life seriously [and] to obey the law [and] respect others." Continued administrative segregation over a lengthy period of time had seriously impaired defendant's will to live within the confines of the prison environment. The doctor noted that treatment within the prison system was insufficient and that defendant would benefit from intensive psychotherapy. According to Dr. Latimer, releasing defendant from administrative segregation and placing him in the general prison population might prove beneficial, but it could "increase his suicidal tendencies." Dr. Steven S. Simring, a psychiatrist, testified on behalf of the State. According to his testimony, "defendant did not suffer from a significant psychiatric illness." Specifically, defendant did not "show psychotic symptomology" in that he did not harbor hallucinations or delusions. Nevertheless, the doctor noted that defendant had "serious psychiatric difficulties" and that he suffered from an "antisocial personality disorder." The witness testified that defendant presented a danger to society because of his "gross immaturity." Dr. Simring stated that defendant could be treated at facilities within the prison system. However, he essentially agreed with Dr. Latimer's assessment that continued incarceration posed a substantial risk of suicidal behavior. *334 Understandably concerned with the potential danger of suicide, the trial judge placed defendant on probation for five years. Initially, the court directed that defendant be placed in a suitable psychiatric institution as a condition of probation. However, it subsequently developed that for a variety of reasons no hospital would accept defendant as a patient. Nevertheless, Dr. Latimer ultimately agreed to treat defendant on an out-patient basis. The trial judge, thus, directed that defendant receive psychotherapy from Dr. Latimer as a condition of probation. An order releasing defendant from custody was subsequently entered. This appeal followed. Pursuant to the State's application, we stayed the trial judge's decision and accelerated this appeal. Preliminarily, we note that R. 3:21-10(b)(2) does not distinguish between mental and physical illness. Although the issue is one of first impression, we are entirely satisfied that under appropriate circumstances a custodial sentence may be amended to permit the release of a defendant because of a mental infirmity. Moreover, the provisions of N.J.S.A. 30:4-82 which authorize the transfer of a prison inmate to a psychiatric facility do not necessarily preclude amendment of a custodial sentence to permit out-patient treatment or conditional release. See State v. Carter, 64 N.J. 382 (1974). The statute only authorizes confinement of persons who are a hazard to themselves or others. Aponte v. State, 30 N.J. 441, 455 (1959). It is not a blanket authorization to commit inmates suffering from any condition of mental illness or impairment. State v. Caralluzzo, 49 N.J. 152, 156 (1967). In any event, our Supreme Court has rejected the argument that the "beneficent" provisions of R. 3:21-10(b)(2) should never be indulged when the disease upon which the application is predicated is amenable to treatment within the prison setting. State v. Tumminello, 70 N.J. 187, 193 (1976). Clearly, the exercise of judicial discretion in this area requires painstaking inquiry and analysis. The humane objective *335 of providing appropriate care and treatment for the mentally impaired must be balanced against the demands of public security. In seeking to accommodate these often countervailing policies, the trial judge should carefully scrutinize the nature of the inmate's mental or emotional illness. Where the pathology from which the defendant suffers poses a serious risk to the public safety, the scale necessarily tips in favor of continued incarceration despite the possibility of adverse psychological effects that might well ensue. In that regard, we emphasize that primary among the hierarchy of governmental objectives is the obligation to protect the citizen against criminal attack. The point to be stressed is that the aim of the law is to guard the innocent from injury by the "sick as well as the bad." State v. Maik, 60 N.J. 203, 213 (1972). Hence, public security must be the paramount goal. We do not mean to suggest that a mentally ill inmate must be forever warehoused in a prison without exit. Whatever the crime committed, the prisoner is entitled to fair and humane treatment. Where the inmate presents a risk of harm to himself by virtue of his mental infirmity, appropriate action must be taken to protect and treat him. We recognize that this may well be easier said than done. Nevertheless, it is an imperative governmental obligation. Against this backdrop, we are thoroughly convinced that the trial judge erred in granting defendant's application for amendment of his custodial sentence. While it is true that defendant has already served a substantial portion of the parole ineligibility term imposed by the sentencing court, this circumstance does not by itself warrant immediate release. We note in that regard that parole eligibility does not mandate automatic freedom from custodial restraint. See In re Trantino Parole Application, 89 N.J. 347, 366-369 (1982). In any event, defendant's initial parole eligibility date has been delayed by virtue of his misconduct while in prison. In our view, consideration of all the factors, especially the serious nature of the *336 crimes and the circumstances attendent to their commission, compels the conclusion that the trial judge mistakenly exercised his discretion. State v. Sanducci, 167 N.J. Super. 503, 510 (App.Div. 1979), certif. den. 82 N.J. 263 (1979). Although the court was required to consider the circumstances pertaining to defendant's illness and the conditions under which he is presently incarcerated, the crime committed and the danger to the public attendent to his release should have been given paramount consideration. See State v. Hodge, 95 N.J. 369 (1984); State v. Roth, 95 N.J. 334 (1984). See also State v. Stanley, 149 N.J. Super. 326, 328 (App.Div. 1977) certif. den. 75 N.J. 21 (1977). We are, therefore, constrained to reverse the order amending defendant's custodial sentences. The sentences imposed initially by the trial judge are hereby reinstated.

1. Field of the Invention The present invention relates to street sweepers or the like and more specifically relates to apparatus for maintaining a substantially constant pressure and pick-up broom pattern on a surface being swept for the useful life of the broom, which broom varies in size, weight and stiffness during its life due to wear. 2. Description of the Prior Art Street sweepers and controls for operating the same are known and disclosed in the above referred to applications. The sweeper includes a main or pick-up broom which engages streets or the like to be swept. In the Kassai application a hydraulic cylinder is disclosed for raising and lowering the sides of the pick-up broom, and a pressure gauge located in the cab is connected to the cylinder thereby enabling the operator to maintain a light, medium, or heavy pick-up broom pressure on the surface being swept. In the above type of sweeper, whether a three wheel or a four wheel sweeper, it is desirable to maintain a substantially constant pick-up broom sweeping pressure and pattern against the surface being swept during the life of the broom. It is well known that a new pick-up broom of about 36 inches in diameter is heavy and thus requires a lifting force acting thereon to provide the desired sweeping pressure and pattern. When the pick-up broom is worn to its minimum acceptable diameter of about 16 inches, the broom is much lighter and requires a downward force to provide the desired sweeping pattern. The pick-up broom is journaled on the ends of a pair of laterally spaced arms that are provided to the chassis of the sweeper and is pivotally moved between a raised transport position and a lowered working position against the surface to be cleaned by a hydraulic cylinder.

Senior Ted Trzaska has been a solid contributor to the success of Minster's boys cross country team this season. Trzaska and the rest of the Wildcats will compete at the state cross country meet on Saturday at National Trail Raceway in Hebron.

Castles in Great Britain and Ireland Castles have played an important military, economic and social role in Great Britain and Ireland since their introduction following the Norman invasion of England in 1066. Although a small number of castles had been built in England in the 1050s, the Normans began to build motte and bailey and ring-work castles in large numbers to control their newly occupied territories in England and the Welsh Marches. During the 12th century the Normans began to build more castles in stone – with characteristic square keep – that played both military and political roles. Royal castles were used to control key towns and the economically important forests, while baronial castles were used by the Norman lords to control their widespread estates. David I invited Anglo-Norman lords into Scotland in the early 12th century to help him colonise and control areas of his kingdom such as Galloway; the new lords brought castle technologies with them and wooden castles began to be established over the south of the kingdom. Following the Norman invasion of Ireland in the 1170s, under Henry II, castles were established there too. Castles continued to grow in military sophistication and comfort during the 12th century, leading to a sharp increase in the complexity and length of sieges in England. While in Ireland and Wales castle architecture continued to follow that of England, after the death of Alexander III the trend in Scotland moved away from the construction of larger castles towards the use of smaller tower houses. The tower house style would also be adopted in the north of England and Ireland in later years. In North Wales Edward I built a sequence of militarily powerful castles after the destruction of the last Welsh polities in the 1270s. By the 14th century castles were combining defences with luxurious, sophisticated living arrangements and heavily landscaped gardens and parks. Many royal and baronial castles were left to decline, so that by the 15th century only a few were maintained for defensive purposes. A small number of castles in England and Scotland were developed into Renaissance Era palaces that hosted lavish feasts and celebrations amid their elaborate architecture. Such structures were, however, beyond the means of all but royalty and the richest of the late-medieval barons. Although gunpowder weapons were used to defend castles from the late 14th century onwards it became clear during the 16th century that, provided artillery could be transported and brought to bear on a besieged castle, gunpowder weapons could also play an important attack role. The defences of coastal castles around the British Isles were improved to deal with this threat, but investment in their upkeep once again declined at the end of the 16th century. Nevertheless, in the widespread civil and religious conflicts across the British Isles during the 1640s and 1650s, castles played a key role in England. Modern defences were quickly built alongside existing medieval fortifications and, in many cases, castles successfully withstood more than one siege. In Ireland the introduction of heavy siege artillery by Oliver Cromwell in 1649 brought a rapid end to the utility of castles in the war, while in Scotland the popular tower houses proved unsuitable for defending against civil war artillery – although major castles such as Edinburgh put up strong resistance. At the end of the war many castles were slighted to prevent future use. Military use of castles rapidly decreased over subsequent years, although some were adapted for use by garrisons in Scotland and key border locations for many years to come, including during the Second World War. Other castles were used as county gaols, until parliamentary legislation in the 19th closed most of them down. For a period in the early 18th century, castles were shunned in favour of Palladian architecture, until they re-emerged as an important cultural and social feature of England, Wales and Scotland and were frequently "improved" during the 18th and 19th centuries. Such renovations raised concerns over their protection so that today castles across the British Isles are safeguarded by legislation. Primarily used as tourist attractions, castles form a key part of the national heritage industry. Historians and archaeologists continue to develop our understanding of British castles, while vigorous academic debates in recent years have questioned the interpretation of physical and documentary material surrounding their original construction and use. Norman Invasion Anglo-Saxon fortifications The English word "castle" derives from the Latin word castellum and is used to refer to the private fortified residence of a lord or noble. The presence of castles in Britain and Ireland dates primarily from the Norman invasion of 1066. Before the arrival of the Normans the Anglo-Saxons had built burhs, fortified structures with their origins in 9th-century Wessex. Most of these, especially in urban areas, were large enough to be best described as fortified townships rather than private dwellings and are therefore not usually classed as castles. Rural burhs were smaller and usually consisted of a wooden hall with a wall enclosing various domestic buildings along with an entrance tower called a burh-geat, which was apparently used for ceremonial purposes. Although rural burhs were relatively secure their role was primarily ceremonial and they too are not normally classed as castles. There were, however, a small number of castles which were built in England during the 1050s, probably by Norman knights in the service of Edward the Confessor. These include Hereford, Clavering, Richard's Castle and possibly Ewyas Harold Castle and Dover. Invasion William, Duke of Normandy, invaded England in 1066 and one of his first actions after landing was to build Hastings Castle to protect his supply routes. Following their victory at the battle of Hastings the Normans began three phases of castle building. The first of these was the establishment, by the new king, of a number of royal castles in key strategic locations. This royal castle programme focused on controlling the towns and cities of England and the associated lines of communication, including Cambridge, Huntingdon, Lincoln, Norwich, Nottingham, Wallingford, Warwick and York. Of the castles built by William the Conqueror two-thirds were built in towns and cities, often those with the former Anglo-Saxon mints. These urban castles could make use of the existing town's walls and fortifications, but typically required the demolition of local houses to make space for them. This could cause extensive damage, and records suggest that in Lincoln 166 houses were destroyed, with 113 in Norwich and 27 in Cambridge. Some of these castles were deliberately built on top of important local buildings, such as the burhs or halls of local nobles, and might be constructed so as to imitate aspects of the previous buildings – such as the gatehouse at Rougemont Castle in Exeter, which closely resembled the previous Anglo-Saxon burh tower – this was probably done to demonstrate to the local population that they now answered to their new Norman rulers. The second and third waves of castle building were led by the major magnates, and then by the more junior knights on their new estates. The apportionment of the conquered lands by the king influenced where these castles were built. In a few key locations the king gave his followers compact groups of estates including the six rapes of Sussex and the three earldoms of Chester, Shrewsbury and Hereford; intended to protect the line of communication with Normandy and the Welsh border respectively. In these areas a baron's castles were clustered relatively tightly together, but in most of England the nobles' estates, and therefore their castles, were more widely dispersed. As the Normans pushed on into South Wales they advanced up the valleys building castles as they went and often using the larger castles of the neighbouring earldoms as a base. As a result, castle building by the Norman nobility across England and the Marches lacked a grand strategic plan, reflecting local circumstances such as military factors and the layout of existing estates and church lands. Castles were often situated along the old Roman roads that still formed the backbone for travel across the country, both to control the lines of communication and to ensure easy movement between different estates. Many castles were built close to inland river ports and those built on the coast were usually located at the mouths of rivers or in ports, Pevensey and Portchester being rare exceptions. Some groups of castles were located so as to be mutually reinforcing – for example the castles of Littledean Camp, Glasshouse Woods and Howle Hill Camp were intended to act as an integrated defence for the area around Gloucester and Gloucester Castle for Gloucester city itself, while Windsor was one of a ring of castles built around London, each approximately a day's march apart. Some regional patterns in castle building can also be seen – relatively few castles were built in East Anglia compared to the west of England or the Marches; this was probably due to the relatively settled and prosperous nature of the east of England and reflected a shortage of available serfs, or unfree labour. Not all of the castles were occupied simultaneously. Some were built during the invasions and then abandoned while other new castles were constructed elsewhere, especially along the western borders. Recent estimates suggest that between 500 and 600 castles were occupied at any one time in the post-conquest period. Architecture There was a large degree of variation in the size and exact shape of the castles built in England and Wales after the invasion. One popular form was the motte and bailey, in which earth would be piled up into a mound (called a motte) to support a wooden tower, and a wider enclosed area built alongside it (called a bailey); Stafford Castle is a typical example of a post-invasion motte castle. Another widespread design was the ring work in which earth would be built up in a circular or oval shape and topped with a wooden rampart; Folkestone Castle is a good example of a Norman ring work, in this case built on top of a hill although most post-invasion castles were usually sited on lower ground. Around 80 per cent of Norman castles in this period followed the motte-and-bailey pattern, but ring works were particularly popular in certain areas, such as south-west England and south Wales. One theory put forward to explain this variation is that ringworks were easier to build in these shallow-soil areas than the larger mottes. The White Tower in London and the keep of Colchester Castle were the only stone castles to be built in England immediately after the conquest, both with the characteristic square Norman keep. Both these castles were built in the Romanesque style and were intended to impress as well as provide military protection. In Wales the first wave of the Norman castles were again made of wood, in a mixture of motte-and-bailey and ringwork designs, with the exception of the stone built Chepstow Castle. Chepstow too was heavily influenced by Romanesque design, reusing numerous materials from the nearby Venta Silurum to produce what historian Robert Liddiard has termed "a play upon images from Antiquity". The size of these castles varied depending on the geography of the site, the decisions of the builder and the available resources. Analysis of the size of mottes has shown some distinctive regional variation; East Anglia, for example, saw much larger mottes being built than the Midlands or London. While motte-and-bailey and ring-work castles took great effort to build, they required relatively few skilled craftsmen allowing them to be raised using forced labour from the local estates; this, in addition to the speed with which they could be built – a single season, made them particularly attractive immediately after the conquest. The larger earthworks, particularly mottes, required an exponentially greater quantity of manpower than their smaller equivalents and consequently tended to be either royal, or belong to the most powerful barons who could muster the required construction effort. Despite motte-and-bailey and ringworks being common designs amongst Norman castles, each fortification was slightly different – some castles were designed with two baileys attached to a single motte, and some ring works were built with additional towers added on; yet other castles were built as ringworks and later converted to motte-and-bailey structures. 12th century Developments in castle design From the early 12th century onwards the Normans began to build new castles in stone and convert existing timber designs. This was initially a slow process, picking up speed towards the second half of the century. Traditionally this transition was believed to have been driven by the more crude nature of wooden fortifications, the limited life of timber in wooden castles and its vulnerability to fire; recent archaeological studies have however shown that many wooden castles were as robust and as complex as their stone equivalents. Some wooden castles were not converted into stone for many years and instead expanded in wood, such as at Hen Domen. Several early stone keeps had been built after the conquest, with somewhere between ten and fifteen in existence by 1100, and more followed in the 12th century until around 100 had been built by 1216. Typically these were four sided designs with the corners reinforced by pilaster buttresses. Keeps were up to four storeys high, with the entrance on the first storey to prevent the door from being easily broken down. The strength of the design typically came from the thickness of the walls: usually made of rag-stone, as in the case of Dover Castle, these walls could be up to thick. The larger keeps were subdivided by an internal wall while the smaller versions, such as that at Goodrich, had a single, slightly cramped chamber on each floor. Stone keeps required skilled craftsmen to build them; unlike unfree labour or serfs, these men had to be paid and stone keeps were therefore expensive. They were also relatively slow to erect – a keep's walls could usually only be raised by a maximum of a year, the keep at Scarborough was typical in taking ten years to build. Norman stone keeps played both a military and a political role. Most of the keeps were physically extremely robust and, while they were not designed as an intended location for the final defence of a castle, they were often placed near weak points in the walls to provide supporting fire. Many keeps made compromises to purely military utility: Norwich Castle included elaborate blind arcading on the outside of the building, in a Roman style, and appears to had a ceremonial entrance route; The interior of the keep at Hedingham could have hosted impressive ceremonies and events, but contained numerous flaws from a military perspective. Similarly there has been extensive debate over the role of Orford Castle whose expensive, three-cornered design most closely echoes imperial Byzantine palaces and may have been intended by Henry II to be more symbolic than military in nature. Another improvement from the 12th century onwards was the creation of shell keeps, involving replacing the wooden keep on the motte with a circular stone wall. Buildings could be built around the inside of the shell, producing a small inner courtyard. Restormel Castle is a classic example of this development with a perfectly circular wall and a square entrance tower while the later Launceston Castle, although more ovoid than circular, is another good example of the design and one of the most formidable castles of the period. Round castles were unusually popular throughout Cornwall and Devon. Although the circular design held military advantages, these only really mattered in the 13th century onwards; the origins of 12th-century circular design were the circular design of the mottes; indeed, some designs were less than circular in order to accommodate irregular mottes, such as that at Windsor Castle. Economy and society English castles during the period were divided into those royal castles owned by the king, and baronial castles controlled by the Anglo-Norman lords. According to chronicler William of Newburgh royal castles formed the "bones of the kingdom". A number of royal castles were also designated as shrieval castles, forming the administrative hub for a particular county – for example Winchester Castle served as the centre of Hampshire. These castles formed a base for the royal sheriff, responsible for enforcing royal justice in the relevant shire; the role of the sheriff became stronger and clearer as the century progressed. A number of royal castles were linked to forests and other key resources. Royal forests in the early medieval period were subject to special royal jurisdiction; forest law was, as historian Robert Huscroft describes it, "harsh and arbitrary, a matter purely for the King's will" and forests were expected to supply the king with hunting grounds, raw materials, goods and money. Forests were typically tied to castles, both to assist with the enforcement of the law and to store the goods being extracted from the local economy: Peveril Castle was linked to the Peak Forest and the local lead mining there; St Briavels was tied to the Forest of Dean; and Knaresborough, Rockingham and Pickering to their eponymous forests respectively. In the south-west, where the Crown oversaw the lead mining industry, castles such as Restormel played an important role running the local stannery courts. Baronial castles were of varying size and sophistication; some were classed as a caput, or the key stronghold of a given lord, and were usually larger and better fortified than the norm and usually held the local baronial honorial courts. The king continued to exercise the right to occupy and use any castle in the kingdom in response to external threats, in those cases he would staff the occupied castles with his own men; the king also retained the right to authorise the construction of new castles through the issuing of licenses to crenellate. It was possible for bishops to build or control castles, such as the important Devizes Castle linked to the Bishop of Salisbury, although this practice was challenged on occasion. In the 12th century the practice of castle-guards emerged in England and Wales, under which lands were assigned to local lords on condition that the recipient provided a certain number of knights or sergeants for the defence of a named castle. In some cases, such as at Dover, this arrangement became quite sophisticated with particular castle towers being named after particular families owing castle-guard duty. The links between castles and the surrounding lands and estates was particularly important during this period. Many castles, both royal and baronial, had deer parks or chases attached to them for the purposes of hunting. These usually stretched away from the village or borough associated with the castle, but occasionally a castle was placed in the centre of a park, such as at Sandal. The Anarchy Civil war broke out in England and raged between 1139 and 1153, forming a turbulent period in which the rival factions of King Stephen and the Empress Matilda struggled for power. Open battles were relatively rare during the war, with campaigns instead centred on a sequence of raids and sieges as commanders attempted to gain control over the vital castles that controlled the territory in the rival regions. Siege technology during the Anarchy centred on basic stone-throwing machines such as ballistae and mangonels, supported by siege towers and mining, combined with blockade and, occasionally, direct assault. The phase of the conflict known as "the Castle War" saw both sides attempting to defeat each other through sieges, such as Stephen's attempts to take Wallingford, the most easterly fortress in Matilda's push towards London, or Geoffrey de Mandeville's attempts to seize East Anglia by taking Cambridge Castle. Both sides responded to the challenge of the conflict by building many new castles, sometimes as sets of strategic fortifications. In the south-west Matilda's supporters built a range of castles to protect the territory, usually motte and bailey designs such as those at Winchcombe, Upper Slaughter, or Bampton. Similarly, Stephen built a new chain of fen-edge castles at Burwell, Lidgate, Rampton, Caxton, and Swavesey – all about six to nine miles (10–15 km) apart – in order to protect his lands around Cambridge. Many of these castles were termed "adulterine" (unauthorised), because no formal permission was given for their construction. Contemporary chroniclers saw this as a matter of concern; Robert of Torigny suggested that as many as 1,115 such castles had been built during the conflict, although this was probably an exaggeration as elsewhere he suggests an alternative figure of 126. Another feature of the war was the creation of many "counter-castles". These had been used in English conflicts for several years before the civil war and involved building a basic castle during a siege, alongside the main target of attack. Typically these would be built in either a ringwork or a motte-and-bailey design between 200 and 300 yards (180 and 270 metres) away from the target, just beyond the range of a bow. Counter-castles could be used to either act as firing platforms for siege weaponry, or as bases for controlling the region in their own right. Most counter-castles were destroyed after their use but in some cases the earthworks survived, such as the counter-castles called Jew's Mount and Mount Pelham built by Stephen in 1141 outside Oxford Castle. Matilda's son Henry II assumed the throne at the end of the war and immediately announced his intention to eliminate the adulterine castles that had sprung up during the war, but it is unclear how successful this effort was. Robert of Torigny recorded that 375 were destroyed, without giving the details behind the figure; recent studies of selected regions have suggested that fewer castles were probably destroyed than once thought and that many may simply have been abandoned at the end of the conflict. Certainly many of the new castles were transitory in nature: Archaeologist Oliver Creighton observes that 56 per cent of those castles known to have been built during Stephen's reign have "entirely vanished". The spread of castles in Scotland, Wales and Ireland Castles in Scotland emerged as a consequence of the centralising of royal authority in the 12th century. Prior to the 1120s there is very little evidence of castles having existed in Scotland, which had remained less politically centralised than in England with the north still ruled by the kings of Norway. David I of Scotland spent time at the court of Henry I in the south, until he became the Earl of Huntingdon, and returned to Scotland with the intention of extending royal power across the country and modernising Scotland's military technology, including the introduction of castles. The Scottish king encouraged Norman and French nobles to settle in Scotland, introducing a feudal mode of landholding and the use of castles as a way of controlling the contested lowlands. The quasi-independent polity of Galloway, which had resisted the rule of David and his predecessors, was a particular focus for this colonisation. The size of these Scottish castles, primarily wooden motte-and-bailey constructions, varied considerably from larger designs, such as the Bass of Inverurie, to smaller castles like Balmaclellan. As historian Lise Hull has suggested, the creation of castles in Scotland was "less to do with conquest" and more to do with "establishing a governing system". The Norman expansion into Wales slowed in the 12th century, but remained an ongoing threat to the remaining native rulers. In response the Welsh princes and lords began to build their own castles, usually in wood. There are indications that this may have begun from 1111 onwards under Prince Cadwgan ap Bleddyn with the first documentary evidence of a native Welsh castle being at Cymmer in 1116. These timber castles, including Tomen y Rhodwydd, Tomen y Faerdre and Gaer Penrhôs, were of equivalent quality to the Norman fortifications in the area and it can prove difficult to distinguish the builders of some sites from the archaeological evidence alone. At the end of the 12th century the Welsh rulers began to build castles in stone, primarily in the principality of North Wales. Ireland remained ruled by native kings into the 12th century, largely without the use of castles. There was a history of Irish fortifications called ráths, a type of ringfort, some of which were very heavily defended but which are not usually considered to be castles in the usual sense of the word. The kings of Connacht constructed fortifications from 1124 which they called caistel or caislen, from the Latin and French for castle, and there has been considerable academic debate over how far these resembled European castles. The Norman invasion of Ireland began between 1166 and 1171, under first Richard de Clare and then Henry II of England, with the occupation of southern and eastern Ireland by a number of Anglo-Norman barons. The rapid Norman success depended on key economic and military advantages, with castles enabling them to control the newly conquered territories. The new lords rapidly built castles to protect their possessions, many of these were motte-and-bailey constructions; in Louth at least 23 of these were built. It remains uncertain how many ringwork castles were built in Ireland by the Anglo-Normans. Other castles, such as Trim and Carrickfergus, were built in stone as the caput centres for major barons. Analysis of these stone castles suggests that building in stone was not simply a military decision; indeed, several of the castles contain serious defensive flaws. Instead the designs, including their focus on large stone keeps, were intended both to increase the prestige of the baronial owners and to provide adequate space for the administrative apparatus of the new territories. Unlike in Wales the indigenous Irish lords do not appear to have constructed their own castles in any significant number during the period. 13th–14th centuries Military developments Castle design in Britain continued to change towards the end of the 12th century. After Henry II mottes ceased to be built in most of England, although they continued to be erected in Wales and along the Marches. Square keeps remained common across much of England in contrast to the circular keeps increasingly prevailing in France; in the Marches, however, circular keep designs became more popular. Castles began to take on a more regular, enclosed shape, ideally quadrilateral or at least polygonal in design, especially in the more prosperous south. Flanking towers, initially square and latterly curved, were introduced along the walls and gatehouses began to grow in size and complexity, with portcullises being introduced for the first time. Castles such as Dover and the Tower of London were expanded in a concentric design in what Cathcart King has labelled the early development of "scientific fortification". The developments spread to Anglo-Norman possessions in Ireland where this English style of castles dominated throughout the 13th century, although the deteriorating Irish economy of the 14th century brought this wave of building to an end. In Scotland Alexander II and Alexander III undertook a number of castle building projects in the modern style, although Alexander III's early death sparked conflict in Scotland and English intervention under Edward I in 1296. In the ensuing wars of Scottish Independence castle building in Scotland altered path, turning away from building larger, more conventional castles with curtain walls. The Scots instead adopted the policy of slighting, or deliberately destroying, castles captured in Scotland from the English to prevent their re-use in subsequent invasions – most of the new Scottish castles built by nobles were of the tower house design; the few larger castles built in Scotland were typically royal castles, built by the Scottish kings. Some of these changes were driven by developments in military technology. Before 1190 mining was used rarely and the siege engines of the time were largely incapable of damaging the thicker castle walls. The introduction of the trebuchet began to change this situation; it was able to throw much heavier balls, with remarkable accuracy, and reconstructed devices have been shown to be able to knock holes in walls. Trebuchets were first recorded in England in 1217, and were probably used the year before as well. Richard I used them in his sieges during the Third Crusade and appears to have started to alter his castle designs to accommodate the new technology on his return to Europe. The trebuchet seems to have encouraged the shift towards round and polygonal towers and curved walls. In addition to having fewer or no dead zones, and being easier to defend against mining, these castle designs were also much less easy to attack with trebuchets as the curved surfaces could deflect some of the force of the shot. Castles saw an increasing use of arrowslits by the 13th century, especially in England, almost certainly linked to the introduction of crossbows. These arrow slits were combined with firing positions from the tops of the towers, initially protected by wooden hoarding until stone machicolations were introduced in England in the late 13th century. The crossbow was an important military advance on the older short bow and was the favoured weapon by the time of Richard I; many crossbows and vast numbers of quarrels were needed to supply royal forces, in turn requiring larger scale iron production. In England, crossbows were primarily made at the Tower of London but St Briavels Castle, with the local Forest of Dean available to provide raw materials, became the national centre for quarrel manufacture. In Scotland, Edinburgh Castle became the centre for the production of bows, crossbows and siege engines for the king. One result of this was that English castle sieges grew in complexity and scale. During the First Barons' War from 1215 to 1217, the prominent sieges of Dover and Windsor Castle showed the ability of more modern designs to withstand attack; King John's successful siege of Rochester required an elaborate and sophisticated assault, reportedly costing around 60,000 marks, or £40,000. The siege of Bedford Castle in 1224 required Henry III to bring siege engines, engineers, crossbow bolts, equipment and labourers from across all of England. The Siege of Kenilworth Castle in 1266, during the Second Barons' War, was larger and longer still. Extensive water defences withstood the attack of the future Edward I, despite the prince targeting the weaker parts of the castle walls, employing huge siege towers and attempting a night attack using barges brought from Chester. The costs of the siege exhausted the revenues of ten English counties. Sieges in Scotland were initially smaller in scale, with the first recorded such event being the 1230 siege of Rothesay Castle where the besieging Norwegians were able to break down the relatively weak stone walls with axes after only three days. When Edward I invaded Scotland he brought with him the siege capabilities which had evolved south of the border: Edinburgh Castle fell within three days, and Roxburgh, Jedburgh, Dunbar, Stirling, Lanark and Dumbarton castles surrendered to the king. Subsequent English sieges, such as the attacks on Bothwell and Stirling, again used considerable resources including giant siege engines and extensive teams of miners and masons. Economy and society A number of royal castles, from the 12th century onwards, formed an essential network of royal storehouses in the 13th century for a wide range of goods including food, drink, weapons, armour and raw materials. Castles such as Southampton, Winchester, Bristol and the Tower of London were used to import, store and distribute royal wines. The English royal castles also became used as gaols – the Assize of Clarendon in 1166 insisted that royal sheriffs establish their own gaols and, in the coming years, county gaols were placed in all the shrieval royal castles. Conditions in these gaols were poor and claims of poor treatment and starvation were common; Northampton Castle appears to have seen some of the worst abuses. The development of the baronial castles in England were affected by the economic changes during the period. During the 13th and 14th centuries the average incomes of the English barons increased but wealth became concentrated in the hands of a smaller number of individuals, with a greater discrepancy in incomes. At the same time the costs of maintaining and staffing a modern castle were increasing. The result was that although there were around 400 castles in England in 1216, the number of castles continued to diminish over the coming years; even the wealthier barons were inclined to let some castles slide into disuse and to focus their resources on the remaining stock. The castle-guard system faded into abeyance in England, being replaced by financial rents, although it continued in the Welsh Marches well into the 13th century and saw some limited use during Edward I's occupation of Scotland in the early 14th century. The remaining English castles became increasingly comfortable. Their interiors were often painted and decorated with tapestries, which would be transported from castle to castle as nobles travelled around the country. There were an increasing number of garderobes built inside castles, while in the wealthier castles the floors could be tiled and the windows furnished with Sussex Weald glass, allowing the introduction of window seats for reading. Food could be transported to castles across relatively long distances; fish was brought to Okehampton Castle from the sea some away, for example. Venison remained the most heavily consumed food in most castles, particularly those surrounded by extensive parks or forests such as Barnard Castle, while prime cuts of venison were imported to those castles that lacked hunting grounds, such as Launceston. By the late 13th century some castles were built within carefully "designed landscapes", sometimes drawing a distinction between an inner core of a herber, a small enclosed garden complete with orchards and small ponds, and an outer region with larger ponds and high status buildings such as "religious buildings, rabbit warrens, mills and settlements", potentially set within a park. A gloriette, or a suite of small rooms, might be built within the castle to allow the result to be properly appreciated, or a viewing point constructed outside. At Leeds Castle the redesigned castle of the 1280s was placed within a large water garden, while at Ravensworth at the end of the 14th century an artificial lake was enclosed by a park to produce an aesthetically and symbolically pleasing entrance to the fortification. The wider parklands and forests were increasingly managed and the proportion of the smaller fallow deer consumed by castle inhabitants in England increased as a result. Welsh castles During the 13th century the native Welsh princes built a number of stone castles. The size of these varied considerably from smaller fortifications, such as Dinas Emrys in Snowdonia, to more substantial castles like Deganwy Castle and the largest, Castell y Bere. Native Welsh castles typically maximised the defensive benefits of high, mountainous sites, often being built in an irregular shape to fit a rocky peak. Most had deep ditches cut out of the rock to protect the main castle. The Welsh castles were usually built with a relatively short keep, used as living accommodation for princes and nobility, and with distinctive rectangular watch-towers along the walls. In comparison to Norman castles the gatehouses were much weaker in design, with almost no use of portcullises or spiral staircases, and the stonework of the outer walls was also generally inferior to Norman built castles. The very last native Welsh castles, built in the 1260s, more closely resemble Norman designs; in the case of Dinas Brân including a round keep and Norman gatehouse defences. Edward I's castles in Wales In 1277 Edward I launched a final invasion of the remaining native Welsh strongholds in North Wales, intending to establish his rule over the region on a permanent basis. As part of this occupation he instructed his leading nobles to construct eight new castles across the region; Aberystwyth and Builth in mid-Wales and Beaumaris, Conwy, Caernarfon, Flint, Harlech and Rhuddlan Castle in North Wales. Historian R. Allen Brown has described these as "amongst the finest achievements of medieval military architecture [in England and Wales]". The castles varied in design but were typically characterised by powerful mural towers along the castle walls, with multiple, over-lapping firing points and large and extremely well defended barbicans. The castles were intended to be used by the king when in the region and included extensive high-status accommodation. Edward also established various new English towns, and in several cases the new castles were designed to be used alongside the fortified town walls as part of an integrated defence. Historian Richard Morris has suggested that "the impression is firmly given of an elite group of men-of-war, long-standing comrades in arms of the king, indulging in an orgy of military architectural expression on an almost unlimited budget". James of Saint George, a famous architect and engineer from Savoy, was probably responsible for the bulk of the construction work across the region. The castles were extremely costly to build and required labourers, masons, carpenters, diggers, and building resources to be gathered by local sheriffs from across England, mustered at Chester and Bristol, before being sent on to North Wales in the spring, returning home each winter. The number of workers involved placed a significant drain on the country's national labour force. The total financial cost cannot be calculated with certainty, but estimates suggest that Edward's castle building programme cost at least £80,000 – four times the total royal expenditure on castles between 1154 and 1189. The Edwardian castles also made strong symbolic statements about the nature of the new occupation. For example, Caernarvon was decorated with carved eagles, equipped with polygonal towers and expensive banded masonry, all designed to imitate the Theodosian Walls of Constantinople, then the idealised image of imperial power. The actual site of the castle may also have been important as it was positioned close to the former Roman fort of Segontium. The elaborate gatehouse, with an excessive five sets of doors and six portcullises, also appears to have been designed to impress visitors and to invoke an image of an Arthurian castle, then believed to have been Byzantine in character. Palace-fortresses In the middle of the 13th century Henry III began to redesign his favourite castles, including Winchester and Windsor, building larger halls, grander chapels, installing glass windows and decorating the palaces with painted walls and furniture. This marked the beginning of a trend towards the development of grand castles designed for elaborate, elite living. Life in earlier keeps had been focused around a single great hall, with privacy for the owner's family provided by using an upper floor for their own living accommodation. By the 14th century nobles were travelling less, bringing much larger households with them when they did travel and entertaining visitors with equally large retinues. Castles such as Goodrich were redesigned in the 1320s to provide greater residential privacy and comfort for the ruling family, while retaining strong defensive features and a capacity to hold over 130 residents at the castle. The design influenced subsequent conversions at Berkeley and by the time that Bolton Castle was being built, in the 1380s, it was designed to hold up to eight different noble households, each with their own facilities. Royal castles such as Beaumaris, although designed with defence in mind, were designed to hold up to eleven different households at any one time. Kings and the most wealthy lords could afford to redesign castles to produce palace-fortresses. Edward III spent £51,000 on renovating Windsor Castle; this was over one and a half times Edward's typical annual income. In the words of Steven Brindle the result was a "great and apparently architecturally unified palace... uniform in all sorts of ways, as to roof line, window heights, cornice line, floor and ceiling heights", echoing older designs but without any real defensive value. The wealthy John of Gaunt redesigned the heart of Kenilworth Castle, like Windsor the work emphasised a unifying, rectangular design and the separation of ground floor service areas from the upper stories and a contrast of austere exteriors with lavish interiors, especially on the 1st floor of the inner bailey buildings. By the end of the 14th century a distinctive English perpendicular style had emerged. In the south of England private castles were being built by newly emerging, wealthy families; like the work at Windsor, these castles drew on the architectural themes of earlier martial designs, but were not intended to form a serious defence against attack. These new castles were heavily influenced by French designs, involving a rectangular or semi-rectangular castle with corner towers, gatehouses and moat; the walls effectively enclosing a comfortable courtyard plan not dissimilar to that of an unfortified manor. Bodiam Castle built in the 1380s possessed a moat, towers and gunports but, rather than being a genuine military fortification, the castle was primarily intended to be admired by visitors and used as a luxurious dwelling – the chivalric architecture implicitly invoking comparisons with Edward I's great castle at Beaumaris. In the north of England improvements in the security of the Scottish border, and the rise of major noble families such as the Percies and the Nevilles, encouraged a surge in castle building at the end of the 14th century. Palace-fortresses such as Raby, Bolton and Warkworth Castle took the quadrangular castle styles of the south and combined them with exceptionally large key towers or keeps to form a distinctive northern style. Built by major noble houses these castles were typically even more opulent than those built by the nouveau riche of the south. They marked what historian Anthony Emery has described as a "second peak of castle building in England and Wales", after the Edwardian designs at the end of the 14th century. Introduction of gunpowder Early gunpowder weapons were introduced to England from the 1320s onwards and began to appear in Scotland by the 1330s. By the 1340s the English Crown was regularly spending money on them and the new technology began to be installed in English castles by the 1360s and 1370s, and in Scottish castles by the 1380s. Cannons were made in various sizes, from smaller hand cannons to larger guns firing stone balls of up to . Medium-sized weapons weighing around 20 kg each were more useful for the defence of castles, although Richard II eventually established 600 pound (272 kilo) guns at the Tower of London and the 15,366 pound (6,970 kilo) heavy Mons Meg bombard was installed at Edinburgh Castle. Early cannons had only a limited range and were unreliable; in addition early stone cannonballs were relatively ineffective when fired at stone castle walls. As a result, early cannon proved most useful for defence, particularly against infantry assaults or to fire at the crews of enemy trebuchets. Indeed, early cannons could be quite dangerous to their own soldiers; James II of Scotland was killed besieging Roxburgh Castle in 1460 when one of his cannons, called "Lion", exploded next to him. The expense of early cannons meant that they were primarily a weapon deployed by royalty rather than the nobility. Cannons in English castles were initially deployed along the south coast where the Channel ports, essential for English trade and military operations in Europe, were increasingly threatened by French raids. Carisbrooke, Corfe, Dover, Portchester, Saltwood and Southampton Castle received cannon during the late 14th century, small circular "keyhole" gunports being built in the walls to accommodate the new weapons. Carisbrooke Castle was subject to an unsuccessful French siege in 1377, the Crown reacting by equipping the castle with cannon and a mill for producing gunpowder in 1379. Some further English castles along the Welsh borders and Scotland were similarly equipped, with the Tower of London and Pontefract Castle acting as supply depots for the new weapons. In Scotland the first cannon for a castle appears to have been bought for Edinburgh in 1384, which also became an arsenal for the new devices. 15th–16th centuries Decline of English castles By the 15th century very few castles were well maintained by their owners. Many royal castles were receiving insufficient investment to allow them to be maintained – roofs leaked, stone work crumbled, lead or wood was stolen. The Crown was increasingly selective about which royal castles it maintained, with others left to decay. By the 15th century only Windsor, Leeds, Rockingham and Moor End were kept up as comfortable accommodation; Nottingham and York formed the backbone for royal authority in the north, and Chester, Gloucester and Bristol forming the equivalents in the west. Even major fortifications such as the castles of North Wales and the border castles of Carlisle, Bamburgh and Newcastle upon Tyne saw funding and maintenance reduced. Many royal castles continued to have a role as the county gaol, with the gatehouse frequently being used as the principal facility. The ranks of the baronage continued to reduce in the 15th century, producing a smaller elite of wealthier lords but reducing the comparative wealth of the majority. and many baronial castles fell into similar decline. John Leland's 16th-century accounts of English castles are replete with descriptions of castles being "sore decayed", their defences "in ruine" or, where the walls might still be in good repair, the "logginges within" were "decayed". English castles did not play a decisive role during the Wars of the Roses, fought between 1455 and 1485, which were primarily in the form of pitched battles between the rival factions of the Lancastrians and the Yorkists. Renaissance palaces The 15th and 16th centuries saw a small number of British castles develop into still grander structures, often drawing on the Renaissance views on architecture that were increasing in popularity on the continent. Tower keeps, large solid keeps used for private accommodation, probably inspired by those in France had started to appear in the 14th century at Dudley and Warkworth. In the 15th century the fashion spread with the creation of very expensive, French-influenced palatial castles featuring complex tower keeps at Wardour, Tattershall and Raglan Castle. In central and eastern England castles began to be built in brick, with Caister, Kirby Muxloe and Tattershall forming examples of this new style. North of the border the construction of Holyrood Great Tower between 1528 and 1532 picked up on this English tradition, but incorporated additional French influences to produce a highly secure but comfortable castle, guarded by a gun park. Royal builders in Scotland led the way in adopting further European Renaissance styles in castle design. James IV and James V used exceptional one-off revenues, such as the forfeiture of key lands, to establish their power across their kingdom in various ways including constructing grander castles such as Linlithgow, almost invariably by extending and modifying existing fortifications. These Scottish castle palaces drew on Italian Renaissance designs, in particular the fashionable design of a quadrangular court with stair-turrets on each corner, using harling to giving them a clean, Italian appearance. Later the castles drew on Renaissance designs in France, such as the work at Falkland and Stirling Castle. The shift in architectural focus reflected changing political alliances, as James V had formed a close alliance with France during his reign. In the words of architectural historian John Dunbar the results were the "earliest examples of coherent Renaissance design in Britain". These changes also included shifts in social and cultural beliefs. The period saw the disintegration of the older feudal order, the destruction of the monasteries and widespread economic changes, altering the links between castles and the surrounding estates. Within castles, the Renaissance saw the introduction of the idea of public and private spaces, placing new value on castles having private spaces for the lord or his guests away from public view. Although the elite in Britain and Ireland continued to maintain and build castles in the style of the late medieval period there was a growing understanding through the Renaissance, absent in the 14th century, that domestic castles were fundamentally different from the military fortifications being built to deal with the spread of gunpowder artillery. Castles continued to be built and reworked in what cultural historian Matthew Johnson has described as a "conscious attempt to invoke values seen as being under threat". The results, as at Kenilworth Castle for example, could include huge castles deliberately redesigned to appear old and sporting chivalric features, but complete with private chambers, Italian loggias and modern luxury accommodation. Although the size of noble households shrank slightly during the 16th century, the number of guests at the largest castle events continued to grow. 2,000 came to a feast at Cawood Castle in 1466, while the Duke of Buckingham routinely entertained up to 519 people at Thornbury Castle at the start of the 16th century. When Elizabeth I visited Kenilworth in 1575 she brought an entourage of 31 barons and 400 staff for a visit that lasted an exceptional 19 days; Leicester, the castle's owner, entertained the Queen and much of the neighbouring region with pageants, fireworks, bear baiting, mystery plays, hunting and lavish banquets. With this scale of living and entertainment the need to find more space in older castles became a major issue in both England and Scotland. Tower houses Tower houses were a common feature of British and Irish castle building in the late medieval period: over 3,000 were constructed in Ireland, around 800 in Scotland and over 250 in England. A tower house would typically be a tall, square, stone-built, crenelated building; Scottish and Ulster tower houses were often also surrounded by a barmkyn or bawn, a walled courtyard designed to hold valuable animals securely, but not necessarily intended for serious defence. Many of the gateways in these buildings were guarded with yetts, grill-like doors made out of metal bars. Smaller versions of tower houses in northern England and southern Scotland were known as Peel towers, or pele houses, and were built along both sides of the border regions. In Scotland a number were built in Scottish towns. It was originally argued that Irish tower houses were based on the Scottish design, but the pattern of development of such castles in Ireland does not support this hypothesis. The defences of tower houses were primarily aimed to provide protection against smaller raiding parties and were not intended to put up significant opposition to an organised military assault, leading historian Stuart Reid to characterise them as "defensible rather than defensive". Gunports for heavier guns were built into some Scottish tower houses by the 16th century but it was more common to use lighter gunpowder weapons, such as muskets, to defend Scottish tower houses. Unlike Scotland, Irish tower houses were only defended with relatively light handguns and frequently reused older arrowloops, rather than more modern designs, to save money. Analysis of the construction of tower houses has focused on two key driving forces. The first is that the construction of these castles appears to have been linked to periods of instability and insecurity in the areas concerned. In Scotland James IV's forfeiture of the Lordship of the Isles in 1494 led to an immediate burst of castle building across the region and, over the longer term, an increased degree of clan warfare, while the subsequent wars with England in the 1540s added to the level of insecurity over the rest of the century. Irish tower houses were built from the end of the 14th century onward as the countryside disintegrated into the unstable control of a large number of small lordships and Henry VI promoted their construction with financial rewards in a bid to improve security. English tower houses were built along the frontier with Scotland in a dangerous and insecure period. Secondly, and paradoxically, appears to have been the periods of relative prosperity. Contemporary historian William Camden observed of the northern English and the Scots, "there is not a man amongst them of a better sort that hath not his little tower or pile", and many tower houses seem to have been built as much as status symbols as defensive structures. Along the English-Scottish borders the construction pattern follows the relative prosperity of the different side: the English lords built tower houses primarily in the early 15th century, when northern England was particularly prosperous, while their Scottish equivalents built them in late 15th and early 16th centuries, boom periods in the economy of Scotland. In Ireland the growth of tower houses during the 15th century mirrors the rise of cattle herding and the resulting wealth that this brought to many of the lesser lords in Ireland. Further development of gunpowder artillery Cannons continued to be improved during the 15th and 16th centuries. Castle loopholes were adapted to allow cannons and other firearms to be used in a defensive role, but offensively gunpowder weapons still remained relatively unreliable. England had lagged behind Europe in adapting to this new form of warfare; Dartmouth and Kingswear Castles, built in the 1490s to defend the River Dart, and Bayard's Cover, designed in 1510 to defend Dartmouth harbour itself, were amongst the few English castles designed in the continental style during the period, and even these lagged behind the cutting edge of European design. Scottish castles were more advanced in this regard, partially as a result of the stronger French architectural influences. Ravenscraig Castle in Scotland, for example, was an early attempt in the 1460s to deploy a combination of "letter box" gun-ports and low-curved stone towers for artillery weapons. These letter box gun-ports, common in mainland Europe, rapidly spread across Scotland but were rarely used in England during the 15th century. Scotland also led the way in adopting the new caponier design for castle ditches, as constructed at Craignethan Castle. Henry VIII became concerned with the threat of French invasion during 1539 and was familiar with the more modern continental designs. He responded to the threat by building a famous sequence of forts, called the Device Forts or Henrician Castles, along the south coast of England specifically designed to be equipped with, and to defend against, gunpowder artillery. These forts still lacked some of the more modern continental features, such as angled bastions. Each fort had a slightly different design, but as a group they shared common features, with the fortification formed around a number of compact lobes, often in a quatrefoil or trefoil shape, designed to give the guns a 360-degree angle of fire. The forts were usually tiered to allow the guns to fire over one another and had features such as vents to disperse the gunpowder smoke. It is probable that many of the forts were also originally protected by earth bulwarks, although these have not survived. The resulting forts have been described by historian Christopher Duffy as having "an air at once sturdy and festive, rather like a squashed wedding cake". These coastal defences marked a shift away from castles, which were both military fortifications and domestic buildings, towards forts, which were garrisoned but not domestic; often the 1540s are chosen as a transition date for the study of castles as a consequence. The subsequent years also marked almost the end of indigenous English fortification design – by the 1580s English castle improvements were almost entirely dominated by imported European experts. The superiority of Scottish castle design also diminished; the Half Moon battery built at Edinburgh Castle in 1574, for example, was already badly dated in continental terms by the time it was built. The limited number of modern fortifications built in Ireland, such as those with the first gunports retrofitted to Carrickfergus Castle in the 1560s and at Corkbeg in Cork Harbour and built in the 1570s in fear of an invasion, were equally unexceptional by European standards. Nonetheless, improved gunpowder artillery played a part in the reconquest of Ireland in the 1530s, where the successful English siege of Maynooth Castle in 1530 demonstrated the power of the new siege guns. There were still relatively few guns in Ireland however and, during the Nine Years' War at the end of the century, the Irish were proved relatively unskilled in siege warfare with artillery used mainly by the English. In both Ireland and Scotland the challenge was how to transport artillery pieces to castle sieges; the poor state of Scottish roads required expensive trains of pack horses, which only the king could afford, and in Ireland the river network had to be frequently used to transport the weapons inland. In these circumstances older castles could frequently remain viable defensive features, although the siege of Cahir Castle in 1599 and the attack on Dunyvaig Castle on Islay in 1614 proved that if artillery could be brought to bear, previously impregnable castle walls might fall relatively quickly. 17th century Wars of the Three Kingdoms In 1603 James VI of Scotland inherited the crown of England, bringing a period of peace between the two countries. The royal court left for London and, as a result – with the exceptions of occasional visits, building work on royal castles north of the border largely ceased. Investment in English castles, especially royal castles, declined dramatically. James sold off many royal castles in England to property developers, including York and Southampton Castle. A royal inspection in 1609 highlighted that the Edwardian castles of North Wales, including Conwy, Beaumaris and Caernarfon were "[u]tterlie decayed".; a subsequent inspection of various English counties in 1635 found a similar picture: Lincoln, Kendal, York, Nottingham, Bristol, Queenborough, Southampton and Rochester were amongst those in a state of dilapidation. In 1642 one pamphlet described many English castles as "muche decayed" and as requiring "much provision" for "warlike defence". Those maintained as private homes; such as Arundel, Berkeley, Carlisle and Winchester were in much better condition, but not necessarily defendable in a conflict; while some such as Bolsover were redesigned as more modern dwellings in a Palladian style. A handful of coastal forts and castles, amongst them Dover Castle, remained in good military condition with adequate defences. In 1642 the English Civil War broke out, initially between supporters of Parliament and the Royalist supporters of Charles I. The war expanded to include Ireland and Scotland, and dragged on into three separate conflicts in England itself. The war was the first prolonged conflict in Britain to involve the use of artillery and gunpowder. English castles were used for various purposes during the conflict. York Castle formed a key part of the city defences, with a military governor; rural castles such as Goodrich could be used a bases for raiding and for control of the surrounding countryside; larger castles, such as Windsor, became used for holding prisoners of war or as military headquarters. During the war castles were frequently brought back into fresh use: existing defences would be renovated, while walls would be "countermured", or backed by earth, in order to protect from cannons. Towers and keeps were filled with earth to make gun platforms, such as at Carlisle and Oxford Castle. New earth bastions could be added to existing designs, such as at Cambridge and Carew Castle and at the otherwise unfortified Basing House the surrounding Norman ringwork was brought back into commission. The costs could be considerable, with the work at Skipton Castle coming to over £1000. Sieges became a prominent part of the war with over 300 occurring during the period, many of them involving castles. Indeed, as Robert Liddiard suggests, the "military role of some castles in the seventeenth century is out of all proportion to their medieval histories". Artillery formed an essential part of these sieges, with the "characteristic military action" according to military historian Stephen Bull, being "an attack on a fortified strongpoint" supported by artillery. The ratio of artillery pieces to defenders varied considerably in sieges, but in all cases there were more guns than in previous conflicts; up to one artillery piece for every nine defenders was not unknown in extreme cases, such as near Pendennis Castle. The growth in the number and size of siege artillery favoured those who had the resources to purchase and deploy these weapons. Artillery had improved by the 1640s but was still not always decisive, as the lighter cannon of the period found it hard to penetrate earth and timber bulwarks and defences – demonstrated in the siege of Corfe. Mortars, able to lob fire over the taller walls, proved particularly effective against castles – in particular those more compact ones with smaller courtyards and open areas, such as at Stirling Castle. The heavy artillery introduced in England eventually spread to the rest of the British Isles. Although up to a thousand Irish soldiers who had served in Europe returned during the war, bringing with them experience of siege warfare from the Thirty Years' War in Europe, it was the arrival of Oliver Cromwell's train of siege guns in 1649 that transformed the conflict, and the fate of local castles. None of the Irish castles could withstand these Parliamentary weapons and most quickly surrendered. In 1650 Cromwell invaded Scotland and again his heavily artillery proved decisive. The Restoration The English Civil War resulted in Parliament issuing orders to slight or damage many castles, particularly in prominent royal regions. This was particularly in the period of 1646 to 1651, with a peak in 1647. Around 150 fortifications were slighted in this period, including 38 town walls and a great many castles. Slighting was quite expensive and took some considerable effort to carry out, so damage was usually done in the most cost-effective fashion with only selected walls being destroyed. In some cases the damage was almost total, such as Wallingford Castle or Pontefract Castle which had been involved in three major sieges and in this case at the request of the townsfolk who wished to avoid further conflict. By the time that Charles II was restored to the throne in 1660, the major palace-fortresses in England that had survived slighting were typically in a poor state. As historian Simon Thurley has described, the shifting "functional requirements, patterns of movement, modes of transport, aesthetic taste and standards of comfort" amongst royal circles were also changing the qualities being sought in a successful castle. Palladian architecture was growing in popularity, which sat awkwardly with the typical design of a medieval castle. Furthermore, the fashionable French court etiquette at the time required a substantial number of enfiladed rooms, in order to satisfy court protocol, and it was impractical to fit these rooms into many older buildings. A shortage of funds curtailed Charles II's attempts to remodel his remaining castles and the redesign of Windsor was the only one to be fully completed in the Restoration years. Many castles still retained a defensive role. Castles in England, such as Chepstow and York Castle, were repaired and garrisoned by the king. As military technologies progressed the costs of upgrading older castles could be prohibitive – the estimated £30,000 required for the potential conversion of York in 1682, approximately £4,050,000 in 2009 terms, gives a scale of the potential costs. Castles played a minimal role in the Glorious Revolution of 1688, although some fortifications such as Dover Castle were attacked by mobs unhappy with the religious beliefs of their Catholic governors, and the sieges of King John's Castle in Limerick formed part of the endgame to the war in Ireland. In the north of Britain security problems persisted in Scotland. Cromwellian forces had built a number of new modern forts and barracks, but the royal castles of Edinburgh, Dumbarton and Stirling, along with Dunstaffnage, Dunollie and Ruthven Castle, also continued in use as practical fortifications. Tower houses were being built until the 1640s; after the Restoration the fortified tower house fell out of fashion, but the weak state of the Scottish economy was such that while many larger properties were simply abandoned, the more modest castles continued to be used and adapted as houses, rather than rebuilt. In Ireland tower houses and castles remained in use until after the Glorious Revolution, when events led to a dramatic shift in land ownership and a boom in the building of Palladian country houses; in many cases using timbers stripped from the older, abandoned generation of castles and tower houses. 18th century Military and governmental use Some castles in Britain and Ireland continued to have modest military utility into the 18th century. Until 1745 a sequence of Jacobite risings threatened the Crown in Scotland, culminating in the rebellion in 1745. Various royal castles were maintained during the period either as part of the English border defences, like Carlisle, or forming part of the internal security measures in Scotland itself, like Stirling Castle. Stirling was able to withstand the Jacobite attack in 1745, although Carlisle was taken; the siege of Blair Castle, at the end of the rebellion in 1746, was the final castle siege to occur in the British Isles. In the aftermath of the conflict Corgaff and many others castles were used as barracks for the forces sent to garrison the Highlands. Some castles, such as Portchester, were used for holding prisoners of war during the Napoleonic Wars at the end of the century and were re-equipped in case of a popular uprising during this revolutionary period. In Ireland Dublin Castle was rebuilt following a fire and reaffirmed as the centre of British administrative and military power. Many castles remained in use as county gaols, run by gaolers as effectively private businesses; frequently this involved the gatehouse being maintained as the main prison building, as at Cambridge, Bridgnorth, Lancaster, Newcastle and St Briavels. During the 1770s the prison reformer John Howard conducted his famous survey of prisons and gaols, culminating in his 1777 work The State of the Prisons. This documented the poor quality of these castle facilities; prisoners in Norwich Castle lived in a dungeon, with the floor frequently covered by an inch of water; Oxford was "close and offensive"; Worcester was so subject to jail fever that the castle surgeon would not enter the prison; Gloucester was "wretched in the extreme". Howard's work caused a shift in public opinion against the use of these older castle facilities as gaols. Social and cultural use By the middle of the century medieval ruined castles had become fashionable once again. They were considered an interesting counterpoint to the now conventional Palladian classical architecture, and a way of giving a degree of medieval allure to their new owners. Historian Oliver Creighton suggests that the ideal image of a castle by the 1750s included "broken, soft silhouettes and [a] decayed, rough appearance". In some cases the countryside surrounding existing castles was remodelled to highlight the ruins, as at Henderskelfe Castle, or at "Capability" Brown's reworking of Wardour Castle. Alternatively, ruins might be repaired and reinforced to present a more suitable appearance, as at Harewood Castle. In other cases mottes, such as that at Groby Castle, were reused as the bases for dramatic follies, or alternatively entirely new castle follies could be created; either from scratch or by reusing original stonework, as occurred during the building of Conygar Tower for which various parts of Dunster Castle were cannibalised. At the same time castles were becoming tourist attractions for the first time. By the 1740s Windsor Castle had become an early tourist attraction; wealthier visitors who could afford to pay the castle keeper could enter, see curiosities such as the castle's narwhal horn, and by the 1750s buy the first guidebooks. The first guidebook to Kenilworth Castle followed in 1777 with many later editions following in the coming decades. By the 1780s and 1790s visitors were beginning to progress as far as Chepstow, where an attractive female guide escorted tourists around the ruins as part of the popular Wye Tour. In Scotland Blair Castle became a popular attraction on account of its landscaped gardens, as did Stirling Castle with its romantic connections. Caernarfon in North Wales appealed to many visitors, especially artists. Irish castles proved less popular, partially because contemporary tourists regarded the country as being somewhat backward and the ruins therefore failed to provide the necessary romantic contrast with modern life. The appreciation of castles developed as the century progressed. During the 1770s and 1780s the concept of the picturesque ruin was popularised by the English clergyman William Gilpin. Gilpin published several works on his journeys through Britain, expounding the concept of the "correctly picturesque" landscape. Such a landscape, Gilpin argued, usually required a building such as a castle or other ruin to add "consequence" to the natural picture. Paintings in this style usually portrayed castles as indistinct, faintly coloured objects in the distance; in writing, the picturesque account eschewed detail in favour of bold first impressions on the sense. The ruins of Goodrich particularly appealed to Gilpin and his followers; Conwy was, however, too well preserved and uninteresting. By contrast the artistic work of antiquarians James Bentham and James Essex at the end of the century, while stopping short of being genuine archaeology, was detailed and precise enough to provide a substantial base of architectural fine detail on medieval castle features and enabled the work of architects such as Wyatt. 19th century Military and governmental use The military utility of the remaining castles in Britain and Ireland continued to diminish. Some castles became regimental depots, including Carlisle Castle and Chester Castle. Carrickfergus Castle was re-equipped with gunports in order to provide coastal defences at the end of the Napoleonic period. Political instability was a major issue during the early 19th century and the popularity of the Chartist movement led to proposals to refortify the Tower of London in the event of civil unrest. In Ireland Dublin Castle played an increasing role in Ireland as Fenian pressures for independence grew during the century. The operation of local prisons in locations such as castles had been criticised, since John Howard's work in the 1770s, and pressure for reform continued to grow in the 1850s and 1860s. Reform of the legislation surrounding bankruptcy and debt in 1869 largely removed the threat of imprisonment for unpaid debts, and in the process eliminated the purpose of the debtor's prisons in castles such as St Briavels. Efforts were made to regularise conditions in local prisons but without much success, and these failures led to prison reform in 1877 which nationalised British prisons, including prisons at castles like York. Compensation was paid to the former owners, although in cases such as York where the facilities were considered so poor as to require complete reconstruction, this payment was denied. In the short term this led to a 39 per cent reduction in the number of prisons in England, including some famous castle prisons such as Norwich; over the coming years, centralisation and changes in prison design led to the closure of most remaining castle prisons. Social and cultural use Many castles saw increased visitors by tourists, helped by better transport links and the growth of the railways. The armouries at the Tower of London opened for tourists in 1828 with 40,000 visitors in their first year; by 1858 the numbers had grown to over 100,000 a year. Attractions such as Warwick Castle received 6,000 visitors during 1825 to 1826, many of them travelling from the growing industrial towns in the nearby Midlands, while Victorian tourists recorded being charged six-pence to wander around the ruins of Goodrich Castle. The spread of the railway system across Wales and the Marches strongly influenced the flow of tourists to the region's castles. In Scotland tourist tours became increasingly popular during the 19th century, usually starting at Edinburgh complete with Edinburgh Castle, and then spending up to two weeks further north, taking advantage of the expanding rail and steamer network. Blair Castle remained popular, but additional castles joined the circuit – Cawdor Castle became popular once the railway line reached north to Fort William. Purchasing and reading guidebooks became an increasingly important part of visiting castles; by the 1820s visitors could buy an early guidebook at Goodrich outlining the castle's history, the first guidebook to the Tower of London was published in 1841 and Scottish castle guidebooks became well known for providing long historical accounts of their sites, often drawing on the plots of Romantic novels for the details. Indeed, Sir Walter Scott's historical novels Ivanhoe and Kenilworth helped to establish the popular Victorian image of a Gothic medieval castle. Scott's novels set in Scotland also popularised several northern castles, including Tantallon which was featured in Marmion. Histories of Ireland began to stress the role of castles in the rise of Protestantism and "British values" in Ireland, although tourism remained limited. One response to this popularity was in commissioning the construction of replica castles. These were particularly popular at beginning of the 19th century, and again later in the Victorian period. Design manuals were published offering details of how to recreate the appearance of an original Gothic castles in a new build, leading to a flurry of work, such as Eastnor in 1815, the fake Norman castle of Penrhyn between 1827 and 1837 and the imitation Edwardian castle of Goodrich Court in 1828. The later Victorians built the Welsh Castell Coch in the 1880s as a fantasy Gothic construction and the last such replica, Castle Drogo, was built as late as 1911. Another response was to improve existing castles, bringing their often chaotic historic features into line with a more integrated architectural aesthetic in a style often termed Gothic Revivalism. There were numerous attempts to restore or rebuild castles so as to produce a consistently Gothic style, informed by genuine medieval details, a movement in which the architect Anthony Salvin was particularly prominent – as illustrated by his reworking of Alnwick and much of Windsor Castle. A similar trend can be seen at Rothesay where William Burges renovated the older castle to produce a more "authentic" design, heavily influenced by the work of the French architect Eugène Viollet-le-Duc. North of the border this resulted in the distinctive style of Scots Baronial Style architecture, which took French and traditional medieval Scottish features and reinvented them in a baroque style. The style also proved popular in Ireland with George Jones' Oliver Castle in the 1850s, for example, forming a good example of the fashion. As with Gothic Revivalism, Scots Baronial architects frequently "improved" existing castles: Floors Castle was transformed in 1838 by William Playfair who added grand turrets and cupolas. In a similar way the 16th-century tower house of Lauriston Castle was turned into the Victorian ideal of a "rambling medieval house". The style spread south and the famous architect Edward Blore added a Scots Baronial touch to his work at Windsor. With this pace of change concerns had begun to grow by the middle of the century about the threat to medieval buildings in Britain, and in 1877 William Morris established the Society for the Protection of Ancient Buildings. One result of public pressure was the passing of the Ancient Monuments Protection Act 1882, but the provisions of the act focused on unoccupied prehistoric structures and medieval buildings such as castles were exempted from it leaving no legal protection. 20th–21st century 1900–1945 During the first half of the century several castles were maintained, or brought back into military use. During the Irish War of Independence Dublin Castle remained the centre of the British administration, military and intelligence operations in Ireland until the transfer of power and the castle to the Irish Free State in 1922. During the Second World War the Tower of London was used to hold and execute suspected spies, and was used to briefly detain Rudolf Hess, Adolf Hitler's deputy, in 1941. Edinburgh Castle was used as a prisoner of war facility, while Windsor Castle was stripped of more delicate royal treasures and used to guard the British royal family from the dangers of the Blitz. Some coastal castles were used to support naval operations: Dover Castle's medieval fortifications used as basis for defences across the Dover Strait; Pitreavie Castle in Scotland was used to support the Royal Navy; and Carrickfergus Castle in Ireland was used as a coastal defence base. Some castles, such as Cambridge and Pevensey, were brought into local defence plans in case of a German invasion. A handful of these castles retained a military role after the war; Dover was used as a nuclear war command centre into the 1950s, while Pitreavie was used by NATO until the turn of the 21st century. The strong cultural interest in British castles persisted in the 20th century. In some cases this had destructive consequences as wealthy collectors bought and removed architectural features and other historical artefacts from castles for their own collections, a practice that produced significant official concern. Some of the more significant cases included St Donat's Castle, bought by William Randolph Hearst in 1925 and then decorated with numerous medieval buildings removed from their original sites around Britain, and the case of Hornby, where many parts of the castle were sold off and sent to buyers in the United States. Partially as a result of these events, increasing legal powers were introduced to protect castles – acts of parliament in 1900 and 1910 widened the terms of the earlier legislation on national monuments to allow the inclusion of castles. An act of Parliament in 1913 introduced preservation orders for the first time and these powers were extended in 1931. Similarly, after the end of the Irish Civil War, the new Irish state took early action to extend and strengthen the previous British legislation to protect Irish national monuments. Around the beginning of the century there were a number of major restoration projects on British castles. Before the outbreak of the First World War work was undertaken at Chepstow, Bodiam, Caernarfon and Tattershal; after the end of the war various major state funded restoration projects occurred in the 1920s with Pembroke, Caerphilly and Goodrich amongst the largest of these. This work typically centred on cutting back the vegetation encroaching on castle ruins, especially ivy, and removing damaged or unstable stonework; castles such as Beaumaris saw their moats cleaned and reflooded. Some castles such as Eilean Donan in Scotland were substantially rebuilt in the inter-war years. The early UK film industry took an interest in castles as potential sets, starting with Ivanhoe filmed at Chepstow Castle in 1913 and starring US leading actor King Baggot. 1945–21st century After the Second World War picturesque ruins of castles became unfashionable. The conservation preference was to restore castles so as to produce what Oliver Creighton and Robert Higham have described as a "meticulously cared for fabric, neat lawns and [a] highly regulated, visitor-friendly environment", although the reconstruction or reproduction of the original appearance of castles was discouraged. As a result, the stonework and walls of today's castles, used as tourist attractions, are usually in much better condition than would have been the case in the medieval period. Preserving the broader landscapes of the past also rose in importance, reflected in the decision by the UNESCO World Heritage Site programme to internationally recognise several British castles including Beaumaris, Caernarfon, Conwy, Harlech, Durham and the Tower of London as deserving of special international cultural significance in the 1980s. The single largest group of English castles are now those owned by English Heritage, created out of the former Ministry of Works in 1983. The National Trust increasingly acquired castle properties in England in the 1950s, and is the second largest single owner, followed by the various English local authorities and finally a small number of private owners. Royal castles such as the Tower of London and Windsor are owned by the Occupied Royal Palaces Estate on behalf of the nation. Similar organisations exist in Scotland, where the National Trust for Scotland was established 1931, and in Ireland, where An Taisce was created in 1948 to working alongside the Irish Ministry of Works to maintain castles and other sites. Some new organisations have emerged in recent years to manage castles, such as the Landmark Trust and the Irish Landmark Trust, which have restored a number of castles in Britain and Ireland over the last few decades. Castles remain highly popular attractions: in 2018 nearly 2.9 million people visited the Tower of London, 2.1 million visited Edinburgh Castle, 466,000 visited Leeds Castle and 365,000 visited Dover Castle. Ireland, which for many years had not exploited the tourist potential of its castle heritage, began to encourage more tourists in the 1960s and 1970s and Irish castles are now a core part of the Irish tourist industry. British and Irish castles are today also closely linked to the international film industry, with tourist visits to castles now often involving not simply a visit to a historic site, but also a visit to the location of a popular film. The management and handling of Britain's historic castles has at times been contentious. Castles in the late 20th and early 21st century are usually considered part of the heritage industry, in which historic sites and events are commercially presented as visitor attractions. Some academics, such as David Lowenthal, have critiqued the way in which these histories are constantly culturally and socially reconstructed and condemned the "commercial debasement" of sites such as the Tower of London. The challenge of how to manage these historic properties has often required very practical decisions. At one end of the spectrum owners and architects have had to deal with the practical challenges of repairing smaller decaying castles used as private houses, such as that at Picton Castle where damp proved a considerable problem. At the other end of the scale the fire at Windsor Castle in 1992 opened up a national debate about how the burnt-out castle wing should be replaced, the degree to which modern designs should be introduced and who should pay the £37 million costs (£50.2 million in 2009 terms). At Kenilworth the speculative and commercial reconstruction of the castle gardens in an Elizabethan style led to a vigorous academic debate over the interpretation of archaeological and historical evidence. Trends in conservation have altered and, in contrast to the prevailing post-war approach to conservation, recent work at castles such as Wigmore, acquired by English Heritage in 1995, has attempted to minimise the degree of intervention to the site. Historiography The earliest histories of British and Irish castles were recorded, albeit in a somewhat fragmented fashion, by John Leland in the 16th century and, by the 19th century, historical analysis of castles had become popular. Victorian historians such as George Clark and John Parker concluded that British castles had been built for the purposes of military defence, but believed that their history was pre-Conquest – concluding that the mottes across the countryside had been built by either the Romans or Celts. The study of castles by historians and archaeologists developed considerably during the 20th century. The early-20th-century historian and archaeologist Ella Armitage published a ground-breaking book in 1912, arguing convincingly that British castles were in fact a Norman introduction, while historian Alexander Thompson also published in the same year, charting the course of the military development of English castles through the Middle Ages. The Victoria County History of England began to document the country's castles on an unprecedented scale, providing an additional resource for historical analysis. After the Second World War the historical analysis of British castles was dominated by Arnold Taylor, R. Allen Brown and D. J. Cathcart King. These academics made use of a growing amount of archaeological evidence, as the 1940s saw an increasing number of excavations of motte and bailey castles, and the number of castle excavations as a whole went on to double during the 1960s. With an increasing number of castle sites under threat in urban areas, a public scandal in 1972 surrounding the development of the Baynard's Castle site in London contributed to reforms and a re-prioritisation of funding for rescue archaeology. Despite this the number of castle excavations fell between 1974 and 1984, with the archaeological work focusing on conducting excavations on a greater number of small-scale, but fewer large-scale sites. The study of British castles remained primarily focused on analysing their military role, however, drawing on the evolutionary model of improvements suggested by Thompson earlier in the century. In the 1990s a wide-reaching reassessment of the interpretation of British castles took place. A vigorous academic discussion over the history and meanings behind Bodiam Castle began a debate, which concluded that many features of castles previously seen as primarily military in nature were in fact constructed for reasons of status and political power. As historian Robert Liddiard has described it, the older paradigm of "Norman militarism" as the driving force behind the formation of Britain's castles was replaced by a model of "peaceable power". The next twenty years was characterised by an increasing number of major publications on castle studies, examining the social and political aspects of the fortifications, as well as their role in the historical landscape. Although not unchallenged, this "revisionist" perspective remains the dominant theme in the academic literature today. Notes References Bibliography Abels, Richard Philip and Bernard S. Bachrach. (eds) (2001) The Normans and their Adversaries at War. Woodbridge, UK: Boydell. Amt, Emilie. (1993) The Accession of Henry II in England: royal government restored, 1149–1159. Woodbridge, UK: Boydell Press. Andrews, Malcolm. (1989) The Search for the Picturesque. Stanford, US: Stanford University Press. Armitage, Ella S. (1912) The Early Norman Castles of the British isles. London: J. Murray. OCLC 458514584 Bailey, Elaine. (2003) "Building for growth", in Fairweather and McConville (eds) (2003) Barry, Terry. (1995) "The Last Frontier: Defence and Settlement in Late Medieval Ireland", in Barry, Frame and Simms (eds) (1995) Barry, Terry, Robin Frame and Katharine Simms. (eds) (1995) Colony and Frontier in Medieval Ireland: Essays Presented to J.F. Lydon. London: Hambledon Press. Bartlett, Thomas. (2010) Ireland: A History. Cambridge: Cambridge University Press. Bradbury, Jim. (2009) Stephen and Matilda: the Civil War of 1139–53. Stroud, UK: The History Press. Brindle, Steven and Kerr, Brian. (1997) Windsor Revealed: New light on the history of the castle. London: English Heritage. Brown, James Baldwin. (1823) Memoirs of the Public and Private Life of John Howard, the Philanthropist. London: T. and G. Underwood. OCLC 60719334 Brown, R. Allen. (1962) English Castles. London: Batsford. OCLC 1392314 Bruce, David M. (2010) "Baedeker – the Perceived 'Inventor' of the Formal Guidebook: a Bible for Travellers in the 19th Century", in Butler and Russell (eds) (2010) Bull, Stephen. (2008) "The Furie of the Ordnance": Artillery in the English Civil Wars. Woodbridge UK: Boydell Press. Butler, Richard and Rosylyn A. Russell (eds) (2010) Giants of Tourism. Wallingford, UK: CAB. Butler, Lawrence. (1997) Clifford's Tower and the Castles of York. London: English Heritage. Carpenter, David. (2004) The Struggle for Mastery: The Penguin History of Britain 1066–1284. London: Penguin. Clark, Geo. T. (1874) "The Defences of York", in The Archaeological Journal, XXXI (1874) Clark, Geo. T. (1884) Medieval Military Architecture. London: Wyman & Sons. OCLC 2600376/ Cooper, Thomas Parsons. (1911) The History of the Castle of York, from its Foundation to the Current Day with an Account of the Building of Clifford's Tower. London: Elliot Stock. OCLC 59345650 Cottrell, Peter. (2006) The Anglo-Irish War: The Troubles of 1913–1922. Botley, UK: Osprey Publishing. Coulson, Charles. (1994) "The Castles of the Anarchy", in King (ed) (1994) Coulson, Charles. (2003) Castles in Medieval Society: Fortresses in England, France, and Ireland in the Central Middle Ages. Oxford: Oxford University Press, Creighton, Oliver Hamilton and Robert Higham. (2003) Medieval Castles. Princes Risborough, UK: Shire Publications. Creighton, Oliver Hamilton. (2005) Castles and Landscapes: Power, Community and Fortification in Medieval England. London: Equinox. Cruickshanks, Eveline. (ed) (2009) The Stuart Courts. Stroud, UK: The History Press. Curnow, P.E. and E.A. Johnson. (1985) "St Briavels Castle", in Chateau Gaillard: études de castellologie médiévale. Caen: Centre de Recherches Archéologiques Médiévales. Danziger, Danny and John Gillingham. (2003) 1215: The Year of the Magna Carta. London: Coronet Books. Delafons, John. (1997) Politics and Preservation: a policy history of the built heritage, 1882–1996. London: Chapman and Hall. Dobres, Marcia-Anne and John E. Robb. (eds) (2000) Agency in Archaeology. London: Routledge. Duffy, Christopher. (1997) Siege Warfare: The Fortress in the Early Modern World, 1494–1660. London: Routledge. Dunbar, John G. (1999) Scottish Royal Palaces: the Architecture of the Royal Residences during the Late Medieval and Early Renaissance Periods. East Lothian, UK: Tuckwell Press. Eales, Richard. (2003) "Royal power and castles in Norman England", in Liddiard (ed) (2003a) Eales, Richard. (2006) Peveril Castle. London: English Heritage. Eltis, David. The Military Revolution in Sixteenth-Century Europe. London: Tauris. Emery, Anthony. (1996) Greater Medieval Houses of England and Wales, 1300–1500: Northern England. Cambridge: Cambridge University Press. . Emery, Anthony. (2006) Greater Medieval Houses of England and Wales, 1300–1500: Southern England. Cambridge: Cambridge University Press. Fairweather, Leslie and Seán McConville. (eds) (2003) Prison Architecture: Policy, Design, and Experience. Oxford: Architectural Press. Fielding, Theodore Henry. (1825) British castles: or, a compendious history of the ancient military structures of Great Britain. London: Rowlett and Brimmer. OCLC 6049730 Fox, Lionel. (2001) The English Prison and Borstal Systems: an account of the prison and Borstal systems. London: Routledge. Gerrard, Christopher M. (2003) Medieval Archaeology: Understanding Traditions and Contemporary Approaches. London: Routledge. Glendinning, Miles, Ranald MacInnes and Aonghus MacKechnie. (2002) A History of Scottish Architecture: from the Renaissance to the Present Day. Edinburgh: Edinburgh University Press. . Gilmour, Tony. (2007) Sustaining Heritage: giving the past a future. Sydney: Sydney University Press. . Gomme, Andor and Alison Maguire. (2008) Design and Plan in the Country House: from castle donjons to Palladian boxes. Yale: Yale University Press. Goodrich, Samuel Griswold. (2005) Recollections of a Lifetime Or Men and Things I Have Seen in a Series of Familiar Letters to a Friend. Kessinger. Gravett, Christopher and Adam Hook. (2003) Norman Stone Castles: The British Isles, 1066–1216. Botley, UK: Osprey. Greene, Kevin and Tom Moore. (2010) Archaeology: An Introduction. Abingdon, UK: Routledge. Grenier, Katherine Haldane. (2005) Tourism and Identity in Scotland, 1770–1914: Creating Caledonia. Aldershot, UK: Ashgate. Harding, Christopher, Bill Hines, Richard Ireland and Philip Rawlings. (1985) Imprisonment in England and Wales: a Concise History. Beckenham, UK: Croon Helm. Harrington, Peter. (2003) English Civil War Fortifications 1642–51. Oxford, UK: Osprey Publishing. Harrington, Peter. (2007) The Castles of Henry VIII. Oxford, UK: Osprey Publishing. Harris, John. (2007) Moving Rooms: the Trade in Architectural Salvages. Yale: Yale University Press. Hassard, John Rose Greene. (1881) A Pickwickian Pilgrimage. Boston: Osgood. OCLC 3217047. Hayton, David. (2004) Ruling Ireland, 1685–1742: politics, politicians and parties. Woodbridge, UK: Boydell Press. . Hearne, Thomas. (1711) The itinerary of John Leland the Antiquary. Oxford: The Theatre. OCLC 655596199 House of Commons Public Accounts Committee. (2009) Maintaining the Occupied Royal Palaces: twenty-fourth report of session 2008–09, report, together with formal minutes, oral and written evidence. London: The Stationery Office. Hull, Lise E. (2006) Britain's Medieval Castles. Westport: Praeger. Hull, Lise E. and Whitehorne, Stephen. (2008) Great Castles of Britain & Ireland. London: New Holland Publishers. Hull, Lise E. (2009) Understanding the Castle Ruins of England and Wales: How to Interpret the History and Meaning of Masonry and Earthworks. Jefferson, US: MacFarland. Hulme, Richard. (2008) "Twelfth Century Great Towers – The Case for the Defence," The Castle Studies Group Journal, No. 21, 2007–8. Huscroft, Richard. (2005) Ruling England, 1042–1217. Harlow: Pearson. . Impey, Edward and Geoffrey Parnell. (2000) The Tower of London: The Official Illustrated History. London: Merrell Publishers. Insall, Donald. (2008) Living Buildings: architectural conservation : philosophy, principles and practice. Victoria, Australia: Images Publishing. Janssens, G. A. M. and Flor Aarts. (eds) (1984) Studies in Seventeenth-Century English Literature, History, and Bibliography. Amsterdam: Rodopi. Johnson, Matthew. (2000) "Self-made men and the staging of agency", in Dobres and Robb (eds) 2000 Johnson, Matthew. (2002) Behind the castle gate: from Medieval to Renaissance. Abingdon, UK: Routledge. Jones, Nigel R. (2005) Architecture of England, Scotland, and Wales. Westport, US: Greenwood Publishing. King, D. J. Cathcart. (1991) The Castle in England and Wales: An Interpretative History. London: Routledge. King, Edmund. (ed) (1994) The Anarchy of King Stephen's Reign. Oxford: Oxford University Press. Lenihan, Pádraig. (2001) "Conclusion: Ireland's Military Revolution(s)", in Lenihan (ed) (2001) Lenihan, Pádraig. (ed) (2001) Conquest and Resistance: War in Seventeenth-Century Ireland. Leiden, Netherlands: BRILL. Liddiard, Robert. (ed) (2003a) Anglo-Norman Castles. Woodbridge, UK: Boydell Press. . Liddiard, Robert (2003b) "Introduction", in Liddiard (ed) (2003a) Liddiard, Robert. (2005) Castles in Context: Power, Symbolism and Landscape, 1066 to 1500. Macclesfield, UK: Windgather Press. Lowenthal, David. (1985) The Past is a Foreign Country. Cambridge: Cambridge University Press. Lowenthal, David. (1996) Possessed by the Past: the heritage industry and the spoils of history. New York: Free Press. Lowry, Bernard. Discovering Fortifications: From the Tudors to the Cold War. Risborough, UK: Shire Publications. Mackenzie, James D. (1896) The Castles of England: Their Story and Structure, Vol II. New York: Macmillan. OCLC 504892038 Mackworth-Young, Robin. (1992) The History and Treasures of Windsor Castle. Andover, UK: Pitkin. Mallgrave, Harry Francis. (2005) Modern Architectural Theory: a Historical Survey, 1673–1968. Cambridge: Cambridge University Press. McConville, Seán. (1995) English Local Prisons, 1860–1900: Next Only to Death. London: Routledge. McNeill, Tom. (2000) Castles in Ireland: Feudal Power in a Gaelic World. London: Routledge. Morris, Richard K. (1998) "The Architecture of Arthurian Enthusiasm: Castle Symbolism in the Reigns of Edward I and his Successors", in Strickland (ed) (1998) Morris, Richard K. (2010) Kenilworth Castle. London: English Heritage. Musty, A. E. S. (2007) Roaring Meg: Test Firing a Copy of Colonel Birch's Civil War Mortar. Hereford, UK: Archaeological and Archival, with Mainmast Conservation. Mynors, Charles. (2006) Listed Buildings, Conservation Areas and Monuments. London: Sweet and Maxwell. Nicolson, Adam. (1997) Restoration: The Rebuilding of Windsor Castle. London: Michael Joseph. O'Dwyer, Frederick. (1997) The Architecture of Deane and Woodward. Cork, Ireland: Cork University Press. . Omar, Paul J. (ed) (2008) International insolvency law: themes and perspectives. London: Ashgate. Parker, John Henry. (1882) Some account of domestic architecture in England from Edward I to Richard II. Oxford: J. H. Parker. OCLC 11791317 Pettifer, Adrian. (2000) Welsh Castles: a Guide by Counties. Woodbridge, UK: Boydell Press. Pettifer, Adrian. (2002) English Castles: a Guide by Counties. Woodbridge, UK: Boydell Press. Prestwich, Michael. (2001) "The Garrisoning of English Medieval Castles", in Abels and Bachrach (eds) (2001) Pounds, Norman John Greville. (1994) The Medieval Castle in England and Wales: a social and political history. Cambridge: Cambridge University Press. Purton, Peter. (2009) A History of the Early Medieval Siege, c.450–1200. Woodbridge, UK: Boydell Press. Prestwich, Michael. (1988) Edward I. Berkeley and Los Angeles: University of California Press. . Rakoczy, Lila. (2007) Archaeology of destruction: a reinterpretation of castle slightings in the English Civil War. York: University of York (PhD thesis) Rajak, Harry. (2008) "The culture of bankruptcy", in Omar (ed) (2008) Reid, Stuart. (2006) Castles and Tower Houses of the Scottish Clans, 1450–1650. Botley, UK: Osprey Publishing. Robinson, John Martin. (2010) Windsor Castle: the Official Illustrated History. London: Royal Collection Publications. Rowse, Alfred Leslie. (1974) Windsor Castle in the History of the Nation. London: Book Club Associates. Rudge, Thomas. (1803) The history of the county of Gloucester brought down to the year 1803. Gloucester, UK: Harris. OCLC 260199931 Ryan, Chris. (2003) Recreational Tourism: Demand and Impacts. Clevedon, UK: Channel View Publications. Scott-Garrett, C. "Littledean Camp," in Transactions of the Bristol and Gloucestershire Archaeological Society, 1958, Vol. 77 Shawcross, William. (2009) Queen Elizabeth: the Queen Mother: the official biography. London: Macmillan. Simpson, Grant G. and Bruce Webster. (2003) "Charter Evidence and the Distribution of Mottes in Scotland", in Liddiard (ed) (2003a) Stell, Geoffrey. (2000) "War-damaged Castles: the evidence from Medieval Scotland", in Chateau Gaillard: Actes du colloque international de Graz (Autriche), 22–29 août 1998. Caen, France: Publications du CRAHM. Strickland, Matthew. (ed) (1998) Armies, Chivalry and Warfare in Medieval England and France. Stamford, UK: Paul Watkins. . Stokstad, Marilyn. (2005) Medieval Castles. Westport, US: Greenwood Press. Tabraham, Chris J. (2004) Edinburgh Castle: Prisons of War. Historic Scotland. Tabraham, Chris J. (2005) Scotland's Castles. London: Batsford. Tabraham, Chris J. and Doreen Grove. (2001) Fortress Scotland and the Jacobites. London: Batsford. Taylor, Arnold. (1989) Caernarfon Castle. Cardiff, UK: Cadw. Thompson, Alexander. (1912) Military Architecture in England during the Middle Ages. London: H. Frowde. OCLC 458292198. Thompson, M. W. (1991) The Rise of the Castle. Cambridge: Cambridge University Press. Thurley, Simon. (2009) "A Country Seat Fit For a King: Charles II, Greenwich and Winchester", in Cruickshanks (ed) 2009. Timbs, John and Alexander Gunn. (2008) Abbeys, Castles and Ancient Halls of England and Wales: Their Legendary Lore and Popular History, Volume 3. Alcester, UK: Read Books. Tite, Catherine. (2010) Portraiture, Dynasty and Power: Art Patronage in Hanoverian Britain, 1714–1759. Amherst, US: Cambria Press. Toy, Sidney. (1933) "The round castles of Cornwall", in Archaeologia 83 (1933) Toy, Sidney. (1985) Castles: Their Construction and History. New York: Dover Publications. Turner, Ralph V. (2009) King John: England's Evil King? Stroud, UK: History Press. Turner, Rick. (2006) Chepstow Castle. Cardiff, UK: Cadw. Walker, David. (1991) "Gloucestershire Castles," in Transactions of the Bristol and Gloucestershire Archaeological Society, 1991, Vol. 109 West, T. W. (1985) Discovering Scottish Architecture. Aylesbury, UK: Shire Publications. Whyte, Ian D. and Kathleen A. Whyte. (1991) The Changing Scottish Landscape, 1500–1800. London: Routledge. Wiener, Martin J. (1994) Reconstructing the Criminal: Culture, Law, and Policy in England, 1830–1914. Cambridge: Cambridge University Press. Williams, J. Anthony. (1984) "No-Popery Violence in 1688- Revolt in the provinces", in Janssens and Aarts (eds) (1984) Williams, William H. A. (2008) Tourism, Landscape, and the Irish character: British Travel Writers in Pre-Famine Ireland. Wisconsin: Wisconsin University Press. Zuelow, Eric. (2009) Making Ireland Irish: Tourism and National Identity since the Irish Civil War. New York: Syracuse University Press. . Further reading Dempsey, Karen; Gilchrist, Roberta; Ashbee, Jeremy; Sagrott, Stefan; Stones, Samantha (2019), "Beyond the martial façade: gender, heritage and medieval castles", International Journal of Heritage Studies, . Goodall, John (2011), The English Castle, New Haven: Yale University Press. Higham, Robert; Barker Philip A. (1992), Timber Castles, London: Batsford. Marshall, Pamela (2002), "The ceremonial Function of the Donjon in the Twelfth Century", Château Gaillard. Etudes de castellologie médiévale, 20: 141–151. McNeill, Tom (1997), Castles in Ireland: Feudal Power in a Gaelic World, London: Routledge. Speight, Sarah (2000), "Castle warfare in the Gesta Stephani", Château Gaillard. Etudes de castellologie médiévale, 19: 269–274. Wheatley, Abigail (2004), The Idea of the Castle in Medieval England, York: York Medieval Press. External links Cadw English Heritage The National Trust The National Trust for Scotland An Taisce, the National Trust for Ireland The Castle Studies Group A photo record of Castles in England, Scotland, Wales and Ireland Category:Castles in Ireland Category:Articles containing video clips Castles

Q: find, save and remove duplicate values from more then one array first array [0]=> Brian [1]=> A [2]=> Leo [3]=> A [4]=> Mike second array [0]=> 1 [1]=> 2 [2]=> 3 [3]=> 4 [4]=> 5 I want to check if in first array there are duplicates, if yes, save only the first occurrence of that value, the other removes, remember those keys, and delete them from the second array too. In the end i want to have first array [0]=> Brian [1]=> A [2]=> Leo [3]=> Mike second array [0]=> 1 [1]=> 2 [2]=> 3 [3]=> 5 I tried with this but second array does not have duplicates so it won't work for both array: array_values(array_unique($array)); A: You did array_values(array_unique($array)); This will give unique values of $array, but you won't find which index of second array needs to be unset. Approach #1: Your best shot is a simple for loop with an isset check. If we find the value already present in our $set(a new temp array), we unset that index from both original arrays, else we preserve it. Snippet: <?php $arr1 = [ 'Brian', 'A', 'Leo', 'A', 'Mike' ]; $arr2 = [ 1,2,3,4,5 ]; $set = []; foreach($arr1 as $key => $value){ if(!isset($set[$value])) $set[$value] = true; else{ unset($arr1[$key]); // foreach creates a copy of $arr1, so safe to unset unset($arr2[$key]); } } print_r($arr1); print_r($arr2); Demo: https://3v4l.org/BQ9mA Approach #2: If you don't like for loops, you can use array wrappers to do this. You can use array_combine to make first array values as keys and second array values as arrays. Note that this would only preserve the latest key value pairs, so we do a array_reverse to only maintain first occurrence pairs. Snippet: <?php $arr1 = [ 'Brian', 'A', 'Leo', 'A', 'Mike' ]; $arr2 = [ 1,2,3,4,5 ]; $filtered_data = array_combine(array_reverse($arr1),array_reverse($arr2)); print_r(array_keys($filtered_data)); print_r(array_values($filtered_data)); Demo: https://3v4l.org/mlstg

Your Cheats All Cheats - Latest First Drumstick HoldingAdded 20 Dec 2008, ID #3665 If you're a real drummer and it's your first time playing or you're not used to holding the drumsticks the real way for real drums , hold them differently.Playing the "real" way dosen't feel right on rockband.You'll find a way to hold it. In the menu screen where it says Rockband in big letters , push red, then yellow, then blue. Then push red, red, followed by blue lue, blue and then red, yellow,blue again. Do this quickly and the screen will say that all songs are unlocked. Warning, this fades away when the xbox is turned off and you cannot save the game when it is in use! Have fun This is a good little trick if you don't want to lose fans in Band World Tour when your band fails a song. Unplug one of your wired instruments and using a different instrument controller select 'Return to Main Menu' so you exit out of your career and return to the Main menu. When you now resume your Band career you won't have lost any fans.

// Copyright (c) Microsoft. All rights reserved. // Licensed under the MIT license. See LICENSE file in the project root for full license information. using Microsoft.Protocols.TestTools.StackSdk.Asn1; namespace Microsoft.Protocols.TestTools.StackSdk.ActiveDirectory.Adts.Asn1CodecV3 { /* MessageID ::= INTEGER (0 .. maxInt) */ [Asn1IntegerBound(Max = 2147483647L, Min = 0)] public class MessageID : Asn1Integer { public MessageID() : base() { } public MessageID(long? val) : base(val) { } } }

Pivotal Pathogenic and Biomarker Role of Chlamydia Pneumoniae in Neurovascular Diseases. Chlamydia Pneumoniae (C. Pn) is an obligatory intracellular bacterium that is associated with respiratory tract infections like pneumonia, pharyngitis and bronchitis. It has also been implicated in cerebrovascular (stroke) as well as cardiovascular diseases. The most possible pathway via which C. Pn elicits its pathogenesis could be via activation of Human Vascular Smooth Muscle Cells (VSMCs) proliferation resulting in the stimulation of Toll-Like Receptor-4 (TLR-4) and/or phospho-44/42(p44/p42) Mitogen-Activated Protein Kinases (MAPK). It is also established that tyrosine phosphorylation of IQ domain GTPase-activating protein 1 (IQGAP1) also contributed to C. Pn infection-triggered Vascular Endothelial Cell (VEC) movements via the SRC tyrosine kinase inhibitor PP2 (4-amino-5-(4-chlorophenyl)-7-(t-butyl)pyrazolo[3,4-d]pyrimidine) resulting in angiogenesis. It is also proven that restricted inflammatory cell infiltrates as well as apoptosis have been linked to C. Pn or C. Pn-specific proteins in atherosclerotic plaques of patients with stroke. It is further an evidence that C. Pn enters the cerebral vasculature during the initial infection and worsen atherosclerosis either directly or indirectly. Chronic, persistent C. Pn infection is also capable of triggering the secretion of Chlamydial Heat Shock Protein 60 (cHSP60) in the vessel wall resulting in augmentation of inflammation. C. Pn also aids in the activation of explicit cell-intermediated immunity within plaques. Macrophages in the carotid plaques co-exist with CD4+ lymphocytes which are capable of triggering the release of pro-inflammatory cytokines resulting in the augmentation of atherogenic development during C. Pn infection. C. Pn actively participated in the modification of both histones H3 and H4 during chromatin analysis via the interleukin 8(IL-8) gene facilitator as well as conscription of nuclear factor kappa-B(NF-κB) or NF- κB/p65 complex and polymerase II (Pol II). This review, therefore, focuses on the crucial involvement of C. Pn in the pathogenesis of cerebrovascular events.

Predictive value of interferon-γ release assays and tuberculin skin testing for progression from latent TB infection to disease state: a meta-analysis. Given the current lack of effective vaccines against TB, the accuracy of screening tests for determining or excluding latent TB infection (LTBI) is decisive in effective TB control. This meta-analysis critically appraises studies investigating the positive and the negative predictive value (PPV and NPV, respectively) from a test-determined LTBI state for progression to active TB of interferon-γ release assays (IGRAs) and the tuberculin skin test (TST). We searched MEDLINE, EMBASE, and Cochrane bibliographies for relevant articles. After qualitative evaluation, the PPV and NPV for progression of commercial and “in-house” IGRAs and the TST for persons not receiving preventive treatment in the context of the respective IGRA studies were pooled using both a fixed and a random-effect model. Weighted rates were calculated for all study populations and for groups solely at high risk of TB development. The pooled PPV for progression for all studies using commercial IGRAs was 2.7% (95% CI, 2.3%-3.2%) compared with 1.5% (95% CI, 1.2%-1.7%) for the TST (P < .0001). PPV increased to 6.8% (95% CI, 5.6%-8.3%) and 2.4% (95% CI, 1.9%-2.9%) for the IGRAs and the TST, respectively, when only high-risk groups were considered (P < .0001). Pooled values of NPV for progression for both IGRAs and the TST were very high, at 99.7% (95% CI, 99.5%-99.8%) and 99.4% (95% CI, 99.2%-99.5%), respectively, although they were significantly higher for IGRAs (P < .01). Commercial IGRAs have a higher PPV and NPV for progression to active TB compared with those of the TST, especially when performed in high-risk persons.