Atlantic salmon (Salmo salar L.) females (2 SW), maturing for the first time, were reared under one of three temperature regimes (high: 14.3 +/- 0.5 degrees C; natural: 10.6 +/- 1.0 degrees C; and cold: 6.9 +/- 1.0 degrees C) in combination with one of two experimental treatments; an injection of GnRH analogue (GnRHa) contained in biodegradable microspheres, or a sham injection (microspheres only). The six experimental groups were then reared under simulated natural photoperiod for 4 weeks. Blood samples were drawn for analysis of plasma steroid levels and the fish were inspected for ovulation weekly. Batches of stripped eggs were incubated in triplicate incubators in raceways until the eyed stage. Treatment with GnRHa resulted in a substantial advancement and synchronization of ovulation at all temperatures, while exposure to cold water also appeared to advance ovulation slightly. While 75% (warm and cold) to 90% (natural) of GnRHa fish ovulated during the 4-week trial, only 30% of sham-treated females exposed to cold water, and none of the sham-treated fish held at higher temperatures, ovulated during this period. Survival rates of embryos to the eyed-stage were significantly higher for broodstock exposed to cold water. Plasma levels of testosterone (T), 17beta-oestradiol (E2), and 17alpha,20beta-dihydroxy-4-pregnen-3-one (17,20betaP) were all significantly affected by treatment with GnRHa and, to a lesser extent, temperature. The efficiency of GnRHa in counteracting the negative effects of high temperature on ovulation and the associated changes in circulating sex steroids suggest that temperature inhibition operates at least in part at the brain or pituitary.