Antipredator behavioural syndromes and nest site choice in a freshwater turtle

Behavioural syndromes, correlated suites of behaviour across time or contexts, lend insight into ecological and evolutionary causes and consequences of animal behaviour. While descriptive studies documenting the presence of behavioural syndromes across taxa have become more common, fewer studies seek to determine how behavioural syndrome expression may drive broader ecological and evolutionary processes, especially in wild populations. In this study, we leverage the complex nesting ecology of a free-ranging population of painted turtles, Chrysemys picta, to determine whether and how behavioural syndromes alter nest site choice in the field. We quantify an antipredator behavioural syndrome in this species and demonstrate, for the first time, that one measure of behavioural boldness predicts nesting behaviour. We document the presence of behavioural syndromes by demonstrating a correlation between behaviours displayed in a simulated predation trial and a novel environment assay in the field. Turtles that displayed more active defences under simulated predation were faster to explore a novel environment, signalling a consistent behaviour type across contexts. Emergence latency, one measure of boldness and exploratory behaviour, predicted the extent of canopy cover at nest sites; specifically, bolder turtles chose nest sites with less canopy cover, which has well-documented consequences for nest thermal environments and, subsequently, offspring phenotypes. Together, these results provide further evidence for the presence of behavioural syndromes in this taxon and demonstrate that behaviour type should be considered when elucidating the factors that influence nest site choice and its associated ecological and evolutionary consequences.

Methods
Field Methods: For this study, we used a population of painted turtles at the Thomson Causeway Recreational Area (TCRA) in Thomson, Illinois, U.S.A. This site is composed of grass and campsites created with concrete pads and gravel situated between a slough on the Mississippi River and a deciduous forest. The TCRA has been used as a research site for several decades to study painted turtle nesting patterns and behaviour (Delaney et al., 2020; Reinke et al., 2019). Recreational visitors were absent on the island during this study due to a campsite closure.

We conducted hourly surveys for nesting painted turtles from 29 May to 2 June 2023. In total, we surveyed 30 individuals. When we observed a turtle beginning to nest, we recorded its approximate location and continued observations from ~10 m away using binoculars such that turtles would not perceive human observers. Immediately after nest completion, one team member started the simulated predation assay while another member recorded nest characteristics. We measured the distance (m) between the nest and the nearest anthropogenic structure using a measuring tape placed at the centre of the nest. Objects deemed anthropogenic structures included concrete pads, gravel paths, electrical boxes, railroad ties, and construction materials. We measured distance from the shoreline (m) using a laser rangefinder or, in cases where the nest was very close to the shoreline, a measuring tape. We took hemispherical photographs directly above the nest and quantified percentage total canopy openness and percentage of southwest canopy openness using Gaplight Analyzer Software (Cary Institute of Ecosystem Studies, Millbrook, NY, U.S.A.) (Mitchell et al., 2015). At this study site, C. picta nest temperature and, consequently, offspring sex ratio are driven by southwest canopy cover in particular (Janzen, 1994a; Schwanz et al., 2010).

Raccoons are the most prevalent turtle predator at this field site, so we created a taxidermized raccoon outfitted with original claws and acrylic teeth to simulate realistic tactile and visual predation cues (Fig. 1). Painted turtles have reduced olfactory perception capabilities on land as they spend most of their time in water (Vieyra, 2011). Thus, for the purposes of this study, we did not scent the model predator for the assays, and instead focused on creating visual and tactile cues that painted turtles are more likely to perceive on land (Pich et al., 2019).

After the turtle had completed nesting and had begun to return to the water, we approached it from behind with the simulated predator. The same team member (M.A.C.) conducted each simulated predation assay. Time began for the 180 s assay period when the raccoon contacted the turtle’s carapace. Simulated predation began as gentle probing with the teeth and claws on and around the carapace until 90 s had passed, at which time the turtle was gently overturned (see methods: Pich et al., 2019). Probing with teeth and claws continued while the turtle’s plastron was exposed for 90 s. The raccoon was manipulated to probe the turtle constantly and continuously throughout the duration of the assay, with consistent effort between individuals. While we manipulated the turtle with the raccoon ‘predator,’ we took care to keep the turtle’s head facing away from team members to reduce visual perception of researchers. During the assay, a team member not manipulating the raccoon recorded counts of antipredator behaviour. Antipredator behaviours were grouped into either active (hissing, biting, urinating/defecating, opening mouth, head pushing) or passive (withdrawing and fleeing) categories (as per Roth et al., 2020). We considered active defences as an index of boldness (Zhao et al., 2016), with bolder turtles displaying a higher count of active defences than shyer turtles.

After the simulated predation assay, we placed the turtle inside an opaque 5-gallon (18.9-litre) plastic bucket and set the bucket away from human disturbance. The turtle remained in the bucket for 15 min to rest between behavioural assays and to minimize or eliminate any influence of the first behaviour assay on the second (see Roth et al., 2020). Limited evidence quantifying the acute stress response of painted turtles at this field site suggests that circulating levels of corticosterone begin to decrease 10 min after handling (Polich, 2018).* *After the 15 min habituation period was complete, we transferred the turtle to a plastic novel arena with a 91 cm diameter. The arena had a ~15 cm high wall to ensure the turtle could not immediately perceive external lateral visual cues. At the beginning of the assay, we placed the turtle in the middle of the arena with the head facing away from the observer, and then started a stopwatch. The observer stood 3 m away and recorded the latency for any part of the turtle’s head to emerge from the shell and be visible (emergence latency) and the latency for the turtle to move its body so that the original position of the carapace was displaced (movement latency) (Roth et al., 2020). The assay stopped when the turtle moved its body or the timer reached 180 s, whichever occurred first. We considered emergence latency and movement latency as indices for boldness, with bolder turtles having lower emergence and movement latency scores than shyer turtles (Zhao et al., 2016).