Studies of the thermal dependence of locomotor performance in ectotherms have provided extensive data on species differences, but often have neglected the time of day at which the test organism is usually active. To compare performance abilities among species that are active at different times of day, do we need to measure performance at the time of day that each species is normally active, or can we adopt the (logistically more convenient) alternative of testing all taxa at the same time (i.e. during daylight hours)? We scored swimming speeds of six species of Australian elapid snakes, incorporating both diurnal and nocturnal species, at a range of temperatures encompassing the usual conditions experienced during activity, and at night as well as by day. Nocturnal species swam faster by night than by day, whereas the reverse was true for diurnal taxa. The magnitude of species differences in speed depended on test temperatures as well as time of day. Thus, interspecific comparisons of locomotor abilities need to consider not only the differing activity temperatures normally experienced by species active at different times of day, but also circadian rhythms in performance.