The timing of ovulation induction is usually decided according to estradiol plasma concentrations and follicle size. We administered human chorionic gonadotropin (hCG) when at least three follicles of 16 mm or more in diameter and adequate estradiol plasma concentrations were detected. We studied the percentage of mature oocyte-cumulus-corona radiata complexes, estradiol and progesterone concentrations in a heterogeneous sized follicle population (range 10-20 mm, n = 90) to perform a retrospective analysis of the adequacy of criteria adopted for the timing of ovulation induction. Plasma and follicular fluid were obtained from 20 normo-ovulating women (aged 28-37 years) treated with gonadotropin releasing hormone analogs (GnRH-a) and human menopausal gonadotropin (hMG) for in vitro fertilization (IVF). No correlation was found between the mean individual follicular fluid estradiol concentration (500-5640 nmol/l) and the respective maximum concentration in plasma (2-16 nmol/l). The estradiol concentration was similar in all follicles. Total follicular fluid estradiol concentration was found to be correlated with follicular fluid volume (r = 0.771, p < 0.01). On the day of hCG administration, the concentration of estradiol in the plasma but not the follicular fluid was correlated with the number of oocyte-cumulus-corona radiata complexes collected (p < 0.01) and the number of mature complexes (p < 0.01). At oocyte pick-up, the plasma concentration of progesterone was correlated (p < 0.01) with number of complexes collected and the number of mature complexes. The percentage of mature complexes collected (77.5%) was higher than suggested by the number of leading follicles. This indicates that our criteria for administering hCG were adequate and that heterogeneous follicle size does not exclude a high rate of mature oocyte-cumulus-corona radiata complexes.