The effects of four seasons of semen collection and of artificial insemination on conception in dairy cows were studied. The solstices and equinoxes (December, March, June and September) defined the beginning and/or end of each season. Semen was collected from 973 progeny-test bulls over 8 years at the two Norwegian AI stations at 60.8°N and 63.4°N where artificial light was used to provide a minimum photoperiod of 10 h/day. The effect of using semen of elite bulls during progeny testing and after selection as elite sires also was investigated. Norwegian Red (NRF) cows were inseminated over a 7-year period using progeny test semen and over the last 4 years of the same period using the semen of the elite sires. The probability of conception to only first inseminations for cows up to, and including, the fifth lactation was assessed by 56-day non-return rate (56d NRR) and calving rate. Two data sets were analysed which excluded cows culled within 270 days of AI or included such cows as non-calving. The reasons for culling were categorised as those for fertility problems or all other reasons. Semen was used for AI irrespective of the season in which it had been collected. Season of semen collection did not affect 56d NRR but calving rate was significantly higher (by 0.5–0.8%, approximately; P < 0.01) for semen collected in the December–March period, when photoperiod was increasing, than at other times of the year. The season in which AI was performed showed a peak of 56d NRR in spring for heifers ( P < 0.01) and in summer for parous animals ( P < 0.01). For calving rate, however, no seasonal peak was found in heifers, whereas pluriparous cows had much higher calving rates in summer and autumn/early winter than late winter and spring ( P < 0.01). Semen of elite sires resulted in higher calving rates by 0.5 (NS) to 1.9% ( P < 0.01) when used after selection than when used during progeny testing. The difference between the calving rate achieved when the semen from elite sires was used during progeny testing and after selection indicates that farmers select different classes of cows for submission to AI by progeny test bulls and sires. The 56d NRR was not as good as calving rate for assessing seasonal and other effects on conception rates.