SUMMARY The administration of a high dose of androgen on a single day to a neonatal female rat has been shown repeatedly to induce persistent vaginal cornification (PVC). However, this type of treatment does not parallel the continuous androgen secretion present in the male, and the high doses that have been used could represent a pharmacological and not a physiological effect. Experiments were carried out to determine the minimal effective dose of testosterone propionate (TP) needed to cause PVC when the androgen is administered to the neonatal rat for the first 10 days of life or as a long-acting ester. Injection of 1, 3 or 9 μg TP on days 1–10 of life in female rats induced PVC in adulthood. All three doses were found to be more effective than a testicular transplant on day 1. In female rats injected with low doses of TP twice daily for the first 10 days of life, PVC was shown between 90 and 100 days of life in 21 out of 22 animals given 0·5 μg TP/injection, and in eight out of 22 animals given 0·05 μg TP/injection. In an experiment where female rats were given a single injection of 0·1, 1·0 or 10·0 μg TP, or 0·1 or 1·0 μg testosterone cyclopentylpropionate (TC, a long-acting androgen) on the first day after birth, PVC occurred at 90–100 days of age in 15 of the 18 animals which were injected with 10 μg TP, in none of the 17 animals which were injected with 1 μg TP, and in 10 of 11 animals which were injected with 1 μg TC. The effects of all treatments on vaginal opening, first oestrus, ovarian weight, body weight and sexual behaviour are reported. The use and implications of low dose regimens are discussed in relation to the construction of an experimental model for the study of sexual differentiation.