The objective was to determine the influence of gonadotropin-releasing hormone on pregnancy rates of dairy cattle at first services, when both the timing of hormone injection and insemination were altered relative to the onset of estrus. Cows (n = 325) were assigned randomly to six groups making up a 2 X 2 X 2 incomplete factorial experiment; dose of GnRH (100 micrograms versus saline), timing [1 h (early) or 12 to 16 h (late) after first detected estrus] of AI, and timing of hormone injection (early versus late) were the three main effects. Cows were observed for estrus 4 times daily. Treatments and resulting pregnancy rates were: 1) hormone injection early plus AI early (35%), 2) hormone injection late plus AI early (34%), 3) saline injection early plus AI early (30%), 4) hormone injection late plus AI late (30%), 5) hormone injection early plus AI late (46%), and 6) saline injection late plus AI late (43%). Pregnancy rate in the first four groups (32%) was less than that in the latter two groups (44%). Concentrations of LH in serum were greater for cows given hormone or saline injections in early estrus than for cows injected with either hormone of saline during late estrus. Concentrations of LH in serum 2 h after GnRH were elevated above those of controls, whether GnRH was injected during early or late estrus. Neither concentrations of LH during estrus nor concentrations of progesterone 8 to 14 d after estrus explained the possible antifertility effect of GnRH given during late estrus.(ABSTRACT TRUNCATED AT 250 WORDS)