Groups of goldfish were held on a 12L:12D photoperiod and fed at one of four different circadian times (light onset at 6, 12, or 18 hr after light onset) for 46 days. The animals were then allowed to free-run on constant light or dark without food or disturbance for five days, and activity rhythms were remote monitored . Most groups maintained a circadian locomotor rhythm entrained to the approximate time of feeding. No consistent differences among fish on the different feeding regimes or between those on constant light or dark were noted.