The role of ovarian secretions in modulating positive oestrogen feedback on LH release in late prepubertal and sexually mature gilts was studied. Gilts were either ovariectomized (OVX) at 230 days of age (sexually mature, OVX 230, control), ovariectomized at 160 days of age (pre-pubertal, OVX 160) or ovariectomized at 160 days of age and given either low, high or very high oestradiol substitution therapy (two or three Silastic implants, 5 or 8 cm in length) from 160 days to 230 days (OVX 160 + 2E5, OVX 160 + 2E8 and OVX 160 + 3E8, respectively). The LH surge responses to three i.m. injections of oestradiol benzoate given at intervals of 12 h (in total 10 micrograms oestradiol benzoate kg-1 bodyweight) were compared at 260 days. An additional group of sexually mature gilts was ovariectomized at 260 days (OVX 260) and challenged with oestradiol benzoate at 360 days. LH surges with peak LH concentrations below pretreatment values (resembling LH responses to oestradiol benzoate in immature gilts) were classified as immature, and those with peak LH concentrations above pretreatment values as mature. LH concentrations before treatment with oestradiol benzoate were reduced (P < 0.05) in OVX 160 + 2E8 and OVX 160 + 3E8 gilts compared with OVX 230 (control). When compared with the control group, the time to LH surge peak was longer in OVX 260 gilts, LH peak amplitude was reduced in the OVX 160 group and the area under the curve of the LH surge (ng LH ml-1 plasma (48 h)-1) was less in all other groups (P < 0.05). Classification of LH surges by mature:immature criteria indicated a high mature:immature ratio in OVX 230 gilts (6/1) and a low ratio in OVX 160 (1/7) and OVX 260 (1/6) gilts. The long-term effect of ovariectomy was partially overcome in the OVX 160 + 2E5 group (mature:immature = 3/3). However, previous exposure to supraphysiological concentrations of oestradiol (as in groups OVX 160 + 2E8 and OVX 160 + 3E8) resulted in a high percentage of animals (66% and 100%, respectively) that did not respond to the oestradiol benzoate challenge with an LH surge. Immediately before oestradiol benzoate administration (30 days after removal of implants) plasma oestradiol concentrations in these two groups were still high (P < 0.05) compared with all other groups. GnRH-induced (0.2 microgram kg-1 bodyweight) LH secretion, evaluated 10 days after treatment with oestradiol benzoate, was depressed (P < 0.05) in OVX 160 + 3E8 gilts, but not affected by age at ovariectomy. We suggest that continuous ovarian secretion is necessary for the final maturation of the LH surge mechanism in late prepubertal gilts and also for maintaining the full functionality of this mechanism in sexually mature gilts. The range over which ovarian oestrogens tune the final maturation of this process appears to be rather narrow; however, other ovarian factors may contribute to the final maturation. Furthermore, continued exposure to supraphysiological concentrations of oestradiol was found to be detrimental to the development of the LH surge mechanism.