Skip to main content
Dryad

Swimming and schooling data of golden shiners in variable water temperatures

Cite this dataset

Kuruvilla, Maria et al. (2022). Swimming and schooling data of golden shiners in variable water temperatures [Dataset]. Dryad. https://doi.org/10.5061/dryad.h44j0zpmp

Abstract

Temperature is highly influential on the physiology and behaviour of ectotherms. In fish, temperature affects social interactions such as schooling behaviour, a common defence against predation. However, the effect of temperature on the ability of schooling fish to collectively respond to a predator is unknown. Here we used a loom stimulus to simulate an approaching predator that elicited a fleeing response in schooling fish over a range of water temperatures (9-29°C) and group sizes (1-16 fish). While speed and acceleration always exhibited a positive curvilinear response to temperature, the optimal temperature at which performance peaked was different during the predation threat versus when they were unperturbed. Similarly, group-level metrics were sensitive to temperature immediately after a loom stimulus but showed no response to temperature during unperturbed swimming. The time taken for fish to respond to the loom stimulus was minimal at 20°C. The proportion of fish that startled, during a loom, peaked at 13°C – around the same temperature at which speed and acceleration were maximum. Taken together, our results suggest that ectothermic fish may be able to compensate for their slower swim speeds at lower temperatures during unperturbed swimming by increasing their sensitivity to startle in response to a predation threat. More generally, we show that in ectotherms the qualitative and quantitative effect of temperature on a behavioural trait may be dependent on the context.

Methods

Overhead video recordings of fish in shallow experimental water tanks were converted to pseudo-2D trajectories using idtracker.ai, which allows tracking of a large number of unmarked animals using deep neural networks to maintain identities and prevent error propagation after crossovers (Romero-Ferrero et al. 2019). Distance-related metrics, such as speed and acceleration, were scaled by the body length (BL) of the fish (Romero-Ferrero et al. 2019). Code from https://github.com/maria-kuruvilla/temp_collective_new was used to calculate the parameters in this dataset.

Funding