The objective of this study was to determine the relative importance of seasonal changes in ovulation rate, fertilization rate and embryo survival as the cause of reduced lambing rates in ewes mated in February compared with those mated in November. The study was conducted at 57°N using mature Mule ewes and Suffolk rams. Sixty ewes were allocated equally to five groups: unbred (UB) or mated at a natural oestrus during November (N) or February (F) by natural (N) or cervical artificial (A) insemination. Groups were maintained separately at pasture supplemented with hay. A raddled vasectomized or non-vasectomized ram was present with UB, NN and NA groups from 26 October 1995 to 1 January 1996 and with UB, FN and FA groups from 25 January 1996 to 31 March 1996. Ewes marked by the ram were recorded twice a day, and those in groups NN, NA, FN and FA were inseminated at their second behavioural oestrus. For all ewes, blood samples were obtained once a day from introduction of the vasectomized rams until 30 days after mating (groups NN, NA, FN and FA) or 20 days after the first oestrus (group UB), and ovulation rate was measured by laparoscopy 7 days after the first oestrus. For ewes in groups NN, NA, FN and FA, ovulation rate was measured again after the second oestrus and ova were recovered from six ewes per group for assessment of fertilization before autotransfer. Pregnancy and lambing rates were recorded at term. Mean (± SE) dates of the first recorded oestrus for ewes in groups NN, NA and UB, and FN, FA and UB were 4 ± 1.1 November and 4 ± 0.9 February, respectively, and intervals between the first and second oestrus were 16 ± 0.2 and 17 ± 0.3 days (P < 0.01), respectively. Ovulation rates were 2.6 ± 0.08 and 2.0 ± 0.05 (P < 0.001), and peripheral progesterone concentrations during the luteal phase were 8.5 ± 0.25 and 7.6 ± 0.31 ng ml-1 (P < 0.05), for November and February, respectively. The difference in peripheral progesterone concentration was not solely attributable to the difference in ovulation rate. There was no significant effect of month or method of insemination, or of embryo recovery and autotransfer procedures on pregnancy rates and the proportion of ewes that became pregnant were NN 0.92, NA 0.83, FN 0.67 and FA 0.75. For ewes undergoing embryo recovery and autotransfer, ova recovered per corpus luteum were 1.00, 0.93, 1.00 and 0.92, fertilized ova per ovum recovered were 0.69, 0.92, 1.00 and 0.83, and lambs born per corpus luteum were 0.62, 0.79, 0.78 and 0.58 for NN, NA, FN and FA groups, respectively. There were no significant seasonal effects on fertilization rate or embryo survival. It is concluded that a seasonal decline in ovulation rate is the primary cause of reduced lambing rates in ewes mated in February compared with those mated in November. Pregnancy rates were high after mating in both periods and were not enhanced by the use of cervical insemination.