Risk-taking behavior related to mercury contamination in a high Arctic seabird

Understanding how animal personality traits are modified by environmental stressors, including chemical contaminants, is of increasing importance given rapid anthropogenic environmental change. In this context, we explored whether mercury (Hg) contamination is associated with neophobia and risk-taking behavior in little auks (or dovekies, Alle alle), an Arctic seabird facing altered Hg exposure due to climate change. To quantify parental risk-taking and neophobia behavior, we presented novel objects of different colors at little auks nests at Hornsund, Svalbard. We quantified latency of birds to enter the nest with food under control conditions and when confronted with novel objects. We related behavior to blood Hg and baseline corticosterone (CORT), as CORT might be modulated by Hg and affect behavioral stress responsiveness. We also determined repeatability and asked whether birds investing highly in reproduction displayed reduced neophobia. Little auks displayed neophobia, with latency to enter the nest increasing from ~40 to 80 sec on average in response to novel objects. Latency to enter was individually repeatable within and across control and novel object sessions, suggesting repeatability in cautiousness. However, neophobia (increased latency relative to controls) exhibited non-significant repeatability, perhaps due to habituation. Birds with elevated Hg for this population (range: 0.3-0.8 mg g-1 dry weight) took longer to enter the nest upon first appearance in control and novel object sessions, suggesting elevated cautiousness, but did not show higher neophobia or reduced habituation. CORT negatively correlated with neophobia. Findings support prior work suggesting that Hg might alter risk-taking behavior, calling for more work on this topic in animals at high Hg exposure risk.

Methods
Novel object presentation experiments, designed to test neophobia responses, were performed at little auk nests commencing when chicks were ~8-10 days old, at the height of the chick provisioning period. Experiments consisted of two different types of sessions, control and novel object sessions. During control sessions, we set up JVC video cameras ~5 meters from nests. Cameras could have been perceived as a novel object. However, to reduce this possibility cameras and tripods were black, a color present in the environment. Furthermore, cameras were set up far enough from the nest that birds could avoid immediate proximity to them when approaching the nest entrance. The camera tripods were also set up in the field long before the onset of the experiment. We preceded to record behavior using 1-sec timelapse mode, which compressed 72 hrs of continuous recording into 3 hrs, and facilitated conservation of memory space. Control sessions allowed us to assess bird behavior at nests in the absence of novel objects. During novel object sessions, which followed control sessions, a novel object consisting of a fist sized rock wrapped in blue, red, or yellow plastic (derived from plastic garbage bags) was placed as close as possible to nest entrances, without blocking the entrance.

When viewing videos to extract data, we recorded all sightings of the focal bird (marked with color bands) near the nest. However, this dataset has been processed such that it only contains lines for every first appearance at the nest with food (appears with food) and the corresponding entrance with food, resulting in feeding (enters with food). Latency times were calculated between first appearance with food and an entrance which resulted in feeding, as detailed in the manuscript. When conducting the analysis, we subsetted the dataset to only include lines for which Activity = "appears with food", as latency times (time to enter and feed) are on these lines.

This dataset also contains all available blood mercury (Hg, µg/g dw), corticosterone (CORT, ng/mL) data and scale mass index (SMI) data.