Four weeks before their predicted mean second lactation calving date, 60 spring-calving Holstein-Friesian cows were blocked into groups of six on the basis of their predicted calving date and body condition score, and allocated at random to one of six dietary treatments in a factorial design: ad libitum grass silage, ad libitum grass silage plus barley straw or ad libitum grass silage plus 3 kg of concentrates, was offered for four weeks before the expected calving date, and after calving they were offered either 4 kg or 8 kg of concentrates plus ad libitum grass silage for eight weeks. On average, the first luteal activity occurred in all the groups at 29 days after calving. Seventeen of the cows had an atypical first plasma progesterone profile; 12 were anovulatory, three had prolonged luteal phases and two became anovulatory after having been cyclic. The cows offered grass silage only before calving had a significantly shorter mean (se) interval between calving and second luteal activity (44·9 [2·1] days), than the cows offered grass silage and straw (53·5 [1·9] days) or grass silage and concentrates (51·5 [3·2] days) (P<0·05). After calving none of the 28 cows offered grass silage and 4 kg of concentrates started cycling before day 21, whereas five of the 30 cows offered grass silage and 8 kg of concentrates cycled before day 21 (P<0·05).