The temperature in each tank varied on average by±0.26 °C (SD) of
its mean across 114 measurement events during the experiment. The
majority of fish appeared healthy and fedwell over the entire study. The
24.5 °C tank was removed from the study after a system malfunction
caused multiple mortalities. Thirteen other mortalities were recorded
among other tanks over the duration of the experiment; nine due to
system faults or animals jumping out of tanks andfourwere unexplained
and unrelated to treatments or months. Survival was therefore not
analysed as a response variable