Fifty-four lean genotype crossbred gilts were allocated at 118 days of age among three diets with different protein concentrations to give lysine: energy (g/MJ digestible energy) ratios: high (0·9), medium (0·6) and low (0·3) given twice daily at 2·9 × maintenance energy. At 160 days of age, gilts were treated with exogenous gonadotropin (PG600™) and animals were examined daily for signs of oestrus. Animals were slaughtered after the second oestrus, if they had shown behavioural oestrus, or at the age of 212 days. Reproductive tracts were recovered for counting of corpora lutea and albicantia. L gilts were lighter than M or H gilts at puberty induction (80, 95, 97 (s.e. 0·73) kg for L, M, H respectively, P < 0·001), with greater backfat thickness (10·8,10·0, 9·2 (s.e. 0·21) mm P2,P < 0·001), lesser longissimus muscle depth (57·4, 65·9, 64·3 (s.e. 0·77) mm, P < 0·001) and poorer food conversion ratio during rearing (3·87, 2·48, 2·42 (s.e. 0·098) kg food per kg gain, P < 0·001). There was no statistically significant difference in the total number of animals that responded to the puberty induction, although L had a greater latency to oestrus than H. Ovulation rate at this induced oestrus was significantly lower in L gilts than in M or H gilts (12·5, 17·3, 21·5 (s.e. 1·32), P < 0·02). A lower proportion ofL, compared with M or H gilts showed spontaneous ovulation in a subsequent cycle (0·15, 0·75, 0·77, %2 = 12·72, P < 0·005). It proved possible, by means of low protein (lysine) diets, to increase body fat reserves in breeding gilts but protein restriction in the rearing phase negatively affected aspects of reproductive performance.