Work Description

Date: 15 November, 2019 Dataset Title: Convergence and divergence in anti-predator displays: A novel approach to quantitative behavioural comparison in snakes Dataset Creators: Alison R. Davis Rabosky, Talia Y. Moore, Ciara M. Sanchez-Paredes, Erin P. Westeen, Joanna G. Larson, Briana A. Sealey, Bailey A. Balinski Dataset Contact: Talia Y. Moore taliaym@umich.edu Funding: This research was supported with funding from the University of Michigan to ARDR and TYM and the Packard Foundation to Dan Rabosky. Key Points: - We present a novel modular ethogram system for characterizing the anti-predator behaviors of snakes - Our ethogram system has low observer bias and infinite extension across any species - This ethogram system enables comparative behavioral analyses Research Overview: Animals in nature use diverse strategies to evade or deter their predators, including many vivid behavioural displays only qualitatively described from field encounters with natural predators or humans. Within venomous snake mimicry, stereotyped anti-predator displays are suggested to be a critical component of the warning signal given by toxic models and thus under strong selection for independent convergence in mimetic species. However, no studies have systematically quantified variation in snake anti-predator displays across taxonomically broad clades to test how these behaviours evolve across species within a phylogenetic comparative methods framework. Here we describe a new, high-throughput approach for collecting and scoring snake anti-predator displays in the field that demonstrates both low observer bias and infinite extension across any species. Then, we show our method�s utility in quantitatively comparing the behaviour of 20 highly-divergent snake species from the Amazonian lowlands of Peru. We found that a simple experimental setup varying simulated predator cues was very successful in eliciting anti-predator displays across species and that high-speed videography captured a greater diversity of behavioural responses than described in the literature. We also found that although different display components evolve at different rates with complicated patterns of covariance, there is clear evidence of evolutionary convergence in anti-predator displays among distantly related elapid coral snakes and their colubrid mimics. We conclude that our approach significantly advances opportunity for future analyses of snake behaviour, kinematics, and the evolution of anti-predator signals more generally, especially macroevolutionary analyses across clades with similarly intractable behavioural diversity. Methodology: (For detailed methodology, please see the published paper) All field methods were approved by the University of Michigan Institutional Animal Care and Use Committee (Protocols #PRO00006234 and #PRO00008306) and the Servicio Nacional Forestal y de Fauna Silvestre (SERFOR) in Peru (permit numbers: 029-2016-SERFOR-DGGSPFFS, 405-2016-SERFOR-DGGSPFFS, 116-2017-SERFOR-DGGSPFFS). We measured the anti-predator response of field-captured snakes across four field stations in the Amazonian lowlands of eastern Peru between October 2016 and December 2018: Villa Carmen Biological Station (850m elevation, latitude: -12.89, longitude: -71.40), Los Amigos Biological Station (270m elevation, 12.56, -70.10), Santa Cruz Biological Station (100m elevation, -3.52, -73.18), and Madre Selva Biological Station (100m elevation, -3.69, -72.46). We collected snakes through a combination of standardised drift-fence lines with both pitfall and funnel traps (Rabosky et al., 2011) and hand-foraging during both the day and night. To ensure that initial anti-predator response was recorded at moment of capture, field survey personnel carried 20L buckets (38.50 cm tall, 34 cm diameter) during all transect or trail surveys and trapline checks. As soon as each snake was safely captured or removed from a trap, we placed it into the bucket and recorded behaviour for up to one minute at 120 frames per second (fps) and with a resolution of at least 1080 pixels. After capture, we kept snakes in cloth bags in a quiet, low disturbance part of the laboratory for between 12-24 hours and handled them as a little as possible between capture and completion of experiments in the field laboratory. To record behaviour in a more controlled setting, we placed snakes into a runway arena [described in the paper Figure 1C]. we used a set of 2-3 GoPro Hero 4 or 5 cameras to film 2 overhead views and added a lateral view by mounting a camera flush against a transparent acrylic window in later models. We scheduled trials to reflect natural activity patterns by species, with diurnal snakes tested between 11am and 5pm and nocturnal snakes recorded from 5pm to 11pm. Each snake was then exposed to a standardised set of predator cues designed to simulate different information types snakes might use to assess predation risk. To simulate an avian attack from overhead, we created �looming� cues by swooping a black or white t-shirt above the arena to create both a shadow and pressure wave over the snake. A white shirt represented the countershading of a bird in the far distance, whereas trials with black shirts represented a bird passing over the animal at a close distance. To simulate the footfalls of an approaching large-bodied terrestrial predator like a mammal, we created �vibration� cues by playing a pulsed vibration (1 sec pulse, 1 sec no pulse) from a smartphone through a wooden plank underneath the bottom of the arena. To simulate a true attack by either predator class, we created a �tactile� cue in which we tapped the tail and body with a snake hook. We randomised the order of cue presentation across individuals and recorded the snake�s response to each of these cues for one minute trials with one minute of rest in between trials. Instrument and/or Software specifications: -We used GoPro Hero4+ Black or Hero5+ Black cameras to record video -We used Quicktime7 to watch videos -We wrote code in R to analyze and compare the ethograms Files contained here: Note: All video and ethogram data are named with prefixes that denote the time at which the video was captured, the species of the animal, and either the camera angle or the identity of the scorer. YYYYMMDD-HHMM-Genusspecies-[gCamera]/[Initials of Scorer].[mp4]/[csv] # items: -Zipped file "EthogramData" includes the following subfolders: --Subfolder "EthogramData>Data Video ethograms" contains 20 .csv files that include the ethograms generated for the analyses included in our study. --Subfolder "EthogramData>Training video ethograms" contains 27 .csv files used to measure inter-individual differences in scoring the same 9 videos. -Zipped file "Heatmaps" contains 36 .pdf files that represent the correlation between behavioral states in different behavioral categories. Each file represents one part (head, body, or tail) of one individual examined for this study. --Zipped file "Data Videos" contains 48 .mp4 files corresponding to the 20 trials analyzed for this study. The suffix "-g" indicates different camera views. Videos without suffixes had only one camera view per trial. --Zipped file "Training videos" contains 9 .mp4 files corresponding to the trials used to measure inter-individual differences in scoring. -"SnakeEthogramTemplate.xlsx" is an excel template used to collect ethogram data. -"SnakeBehaviorDemoVideo.mov" is a video that shows some stereotypical anti-predator behaviors and how they were characterized. -"EthogramOptions.csv" includes a list of all behavioral states currently identified for each body part. This is not meant to be a definitive list, but includes many behavioral states associated with common anti-predator displays. This list may be added to or reduced depending on the specific needs of a study. -"Head Behaviors.pdf", "Body Behaviors.pdf", "Tail Behaviors.pdf" include drawings that correspond to each of the behavioral states listed in EthogramOptions.csv and SnakeEthogramTemplate.xlsx -"Video for Figure 3.zip" includes a video corresponding to the data depicted in Figure 3 of the paper [not in dataset]. This video shows a Xenodon severus anti-predator display. Related publication(s): - Alison R. Davis Rabosky, Talia Y. Moore, Ciara M. Sanchez-Paredes, Erin P. Westeen, Joanna G. Larson, Briana A. Sealey, Bailey A. Balinski (in review) Convergence and divergence in anti-predator displays: A novel approach to quantitative behavioural comparison in snakes. BioRxiv https://doi.org/10.1101/849703 Use and Access: To Cite Data: Alison R. Davis Rabosky, Talia Y. Moore, Ciara M. Sanchez-Paredes, Erin P. Westeen, Joanna G. Larson, Briana A. Sealey, Bailey A. Balinski (2019) Convergence and divergence in anti-predator displays: A novel approach to quantitative behavioural comparison in snakes. [Data set] University of Michigan, Deep Blue