Repeatability and reliability of the provoked escape response test for estimating fish maximum swimming speeds

In fish, burst swimming performance is essential for escaping predators, capturing prey, and overcoming velocity barriers. Forced swimming trials often underestimate performance, whereas volitional swimming tests provide more ecologically relevant data, but require costly infrastructure. A cost-effective alternative is the provoked escape response test, where an object is dropped near the fish to trigger an escape response, and maximum swimming speed is estimated over a pre-determined time-period. The robustness of this test has, however, been little investigated. Here we test its repeatability at both the individual and group levels by running two repeated trials on 118 individuals of the Padanian goby (Padogobius bonelli) on two consecutive days. The estimated maximum swimming speed was not repeatable at the individual fish level, suggesting that the test may be unreliable when individual fish performance measures are important. The testing protocol did consistently show that large fish outperform small fish and remains a useful, simple, and cheap method for estimating average swimming performance and for comparing the performance between groups. Lower performance in the second compared to the first test indicates that fish experience, such as habituation or stress, can influence estimated maximum swimming speeds, suggesting that absolute estimates should be used with caution.