Triplicate groups of gilthead sea bream, Sparus aurata (10.4 g), were distributed among 27 tanks (12 fish per tank) and reared in flow-through seawater. A factorial experiment (3 x 3) was designed to include a continuously fed control group and two cycled starvation groups: 1 + 3 (starved 1 d, fed 3 d), 1 + 5 (starved 1 d, fed 5 d). Each of the feeding groups was subjected to one of three feeding frequencies (2, 4, and 6 times per day) over the 60-d experiment duration. The average final weight of fish in 1 + 3 and 1 + 5 groups were significantly lower than that of the control group. Partial compensation was observed in the starved groups subjected to any of the three feeding frequencies. Regardless of the feeding frequency, control fish consumed less feed than the starved groups. The highest body protein content was found in the control group. The rate of oxygen consumption significantly increased 30 min after the feeding and the magnitude of the effect increased with the feeding frequency. These results suggest that the present cycling starvation schedules did not invoke a full compensation in gilthead sea bream.