The responses of gonadotrophin gene expression, pituitary content and serum levels to castration alone and castration plus testosterone replacement (silicone elastomer implants) were compared in male rats at 10, 30, 60 and 90 days of age. Sham-operated animals served as controls. In addition, 30-day-old castrated rats were treated with dihydrotestosterone (DHT) and diethylstilboestrol (DES). When killed 7 days after castration, the increases in serum LH (six- to eightfold; P < 0.01) and FSH (two- to fourfold; P < 0.01) were similar at all ages studied. Likewise, testosterone reversed the effects of castration in a largely similar fashion at all ages. In contrast, great age-related differences were observed in the responses of gonadotrophin subunit mRNAs to the treatments. Castration increased the common alpha subunit mRNA two- to fourfold on days 10 and 30 (P < 0.01), sixfold on day 60 (P < 0.01), but not at all on day 90. Testosterone reversed the increases at all ages, but the levels were below those of controls only at 90 days (P < 0.01). The highest increases (sixfold; P < 0.01) of LH-beta mRNA were seen on days 10 and 60, the others being two- to threefold higher (P < 0.05-0.01). Testosterone reversed this effect at 60 days and suppressed LH-beta mRNA to below the control levels at other ages (P < 0.01). Castration had no effect on FSH-beta subunit mRNA at 30 and 90 days but a four- to fivefold increase was seen on days 10 and 60 (P < 0.01).(ABSTRACT TRUNCATED AT 250 WORDS)