Duplicate groups of first-feeding (4-5-d-old) Tilapia zillii (Gervais) were maintained on two feeding regimes (high and low ration size) for a total period of 550 d (~ 17 months) in environmentally-controlled aquaria. Differences in mean individual fish weight were evident between fish fed the different ration sizes after just 3 months, though these differences were not statistically significant until day 180. Thereafter, fish maintained upon the higher ration were significantly larger than fish fed a reduced ration. Mortality in high ration fish over the course of the experiment was ~60 % compared to ~88 % in low ration fish. Mortality in both treatment groups was most pronounced during the first three months (particularly during the first 30 d) but was much reduced thereafter. Spawning trials undertaken from day 450 onwards indicated that there was no significant difference between the two ration groups in terms of spawning periodicity over a discrete time period of 60 d. Mean total fecundity was significantly higher in fish on high rations though mean egg diameter and gonadosomatic index remained unchanged. Adjustment of spawning data to a common maternal fish size using one-factor ANCOVA failed to detect significant differences between the two ration levels in terms of either mean total fecundity or mean egg diameter. Stereological analysis of plastic-embedded ovarian tissue revealed that the ovaries of low ration fish possessed significantly more stage 2 and 3 (early and late perinucleolar) oocytes and significantly less stage 6/7 (late-vitellogenic/maturing) oocytes than the ovaries of fish fed the higher ration. No significant differences were detected between the two rations in terms of ovarian atresia. It is suggested that under prolonged food restriction, female T. zillii sacrifice somatic growth such that reproductive investment can be maintained.