The objective was to examine (a) how pregnancy rate on one farm (500 cows) was affected by signs of oestrus and disease stressors and (b) whether pregnancy rate could be maximised by considering cow activity. The signs of oestrus and timings were recorded at artificial insemination (AI), and cow activity was monitored by neck collars. Pregnancy rate tended to be higher in animals that displayed standing oestrus (35 v 26 per cent; P=0.06) but was 10 per cent lower in those cows with an elevated somatic cell count (SCC; >200,000 cells/ml milk) within 0–4 or 4–8 weeks prior to AI (P=0.01 and 0.05, respectively), irrespective of the incidence of clinical mastitis prior to AI. Cow activity data were available for 525 inseminations (from a total of 1299). The mean interval from increased activity to AI in all cows (11 hours 32 minutes; 95 per cent CI 10 hours 40 minutes to 12 hours 24 minutes) was not different for cows that did or did not establish a pregnancy (P=0.90). The pregnancy rate improved to the average of unaffected cows if AI was delayed by about eight hours in animals with an elevated SCC 0–4 weeks prior to AI (P=0.025), indicating that, in cows with prior elevated SCC, AI could be repeated approximately eight hours later to achieve maximum pregnancy rates.