Work Description

Date: 1 August, 2019 Dataset Title: A kinematic analysis of Micrurus coral snakes reveals unexpected variation in stereotyped anti-predator displays within a mimicry system Dataset Creators: Talia Y. Moore, Shannon M. Danforth, Joanna G. Larson, Alison R. Davis Rabosky Dataset Contact: Talia Y. Moore taliaym@umich.edu Funding: Key Points: - We compare intraspecific and interspecific variation in thrashing kinematics during Micrurus sp. snake anti-predator displays - We find more variation than expected, given the strong selective pressures Research Overview: 1. Warning signals in chemically defended organisms are critical components of predator-prey interactions, often requiring multiple coordinated display components for a signal to be effective. When threatened by a predator, venomous coral snakes (genus Micrurus) display a vigorous, non-locomotory thrashing behaviour that has been only qualitatively described. Given the high-contrast and often colourful banding patterns of these snakes, this thrashing display is hypothesized to be a key component of a complex aposematic signal under strong stabilizing selection across species in a mimicry system. 2. By experimentally testing snake response across simulated predator cues, we analysed variation in the presence and expression of a thrashing display across five species of South American coral snakes. 3. Although the major features of the thrash display were conserved across species, we found significant variation in the propensity to perform a display at all, the duration of thrashing, and the curvature of snake bodies that was mediated by predator cue type, snake body size, and species identity. We also found an interaction between curve magnitude and body location that clearly shows which parts of the display vary most across individuals and species. 4. Our results suggest that contrary to the assumption in the literature that all species and individuals perform the same display, a high degree of variation persists in thrashing behaviour exhibited by Micrurus coral snakes despite presumably strong selection to converge on a common signal. This quantitative behavioural characterization presents a new framework for analysing the non-locomotory motions displayed by snakes in a broader ecological context, especially for signalling systems with complex interaction across multiple modalities. Methodology: (For detailed methodology, please see the BioRxiv link below) All animal-related procedures have been approved by the University of Michigan Institutional Animal Care and Use Committee (Protocols #PRO00006234 and #PRO00008306) and the Peruvian government SERFOR (Servicio Nacional Forestal y de Fauna Silvestre; permit numbers: 029-2016-SERFOR-DGGSPFFS, 405-2016-SERFOR-DGGSPFFS, 116-2017-SERFOR-DGGSPFFS). We collected data during five field expeditions in the Amazonian lowlands of Peru from March 2016 to December 2018, at Villa Carmen, Los Amigos, Madre Selva, and Santa Cruz Biological Stations (Fig. 1A). We recorded anti-predator behavior in a pop-up behavioral arena constructed of corrugated plastic (Fig. 1B [included in paper, not in dataset]) illuminated by a string of led lights attached to the inner surface at the top edge of the arena walls. We placed snakes into the arena, one at a time, and immediately began recording their behaviour. Between each snake we cleaned the arena with unscented soap and water. We covered the inner surface of the arena with an adhesive transparent plastic film (Con-Tact, Rubbermaid) to facilitate thorough and rapid cleansing and preserve the visual fiducial markings. We recorded snake behavior using either two or three GoPro (San Mateo, California) Hero 4+ Black or three Hero 5+ Black cameras filming from overhead and lateral views (see Figure 1B for camera positions [included in paper, not in dataset]) at 30, 60, or 120 frames per second, depending on the lighting conditions. We used three different stimuli to elicit anti-predator behaviors: overhead looming, pulsed vibration, and a tactile stimulus through physical contact. To simulate avian predation threat, we quickly moved a piece of cloth across the top of the arena to create the visual looming and pressure wave stimuli produced by a swooping bird. To simulate a large mammal predator, we used the vibration produced by a cellular phone and placed it in contact with the arena. To test for response after contact with a predator, we used a 1m snake hook to tap the snake�s tail. We randomized all treatments across individuals, and we recorded snake behaviors for up to two minutes, allowing one minute of time to rest between the 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 Matlab to analyze the snake curvature (see GitHub link below) -We wrote code in R to analyze the durations and test for the effects of different factors on expression of anti-predator behaviors Files contained here: 26 items: -all.trials.csv contains a record of whether or not the snake exhibited anti-predator behavior during the trial, and the conditions of the trial --The first column, DateTime, indicates the date and time (YYYYMMDD-HHMM) on a 24 hour clock for the start of the trial. --Site indicates where the snake was captured (VC = Villa Carmen, LA = Los Amigos, IQ = Iquitos) --RAB indicates the field record number for each snake --Sex indicates the sex of the snake, if known. A more up-to-date record can be found in Table 1 of the paper (not included in this supplementary dataset) --Genussp indicates the Genus and the species of the snake, separated by an underscore --SVL indicates the snout-vent length in mm --response is a binary variable that indicates (1) an anti-predator thrashing response or (0) no anti-predator response, including escape --Setting indicates the location of the behavioral test. (F) indicates in a bucket in the field at the time and location of capture, (A) indicates in the pop-up arena in the laboratory near the capture site, (L) indicates in a bucket in the laboratory near the capture site. -23 video files are named for the species tested and an identifier corresponding to figures in the paper. The naming scheme is M_species_RAB####_ and an optional HHMM code to discern between multiple trials from the same individual. These values correspond to the information in all.trials.csv. -Supporting Information.docx contains plots and captions associated with the paper. --Supporting Figure 1 shows the results from several tests to determine how different factors affect coral snake likelihood of expression of anti-predator behavior. --Supporting Figures 2-5 show the duration of response and mean magnitude of curvature for all trials examined in this study, sorted by species. --The caption for Supporting Video 1 corresponds to the M_lemniscatus_RAB3587_1925.mp4 video. Related publication(s): -Talia Y. Moore, Shannon M. Danforth, Joanna G. Larson, Alison R. Davis Rabosky (BioRxiv) A kinematic analysis of Micrurus coral snakes reveals unexpected variation in stereotyped anti-predator displays within a mimicry system http://biorxiv.org/cgi/content/short/754994v1 -GitHub Repository at https://github.com/sdanfort/snake_curvature Use and Access: To Cite Data: Talia Y. Moore, Shannon M. Danforth, Joanna G. Larson, Alison R. Davis Rabosky (2019) A kinematic analysis of Micrurus coral snakes reveals unexpected variation in stereotyped anti-predator displays within a mimicry system [Data set] University of Michigan, Deep Blue