The effect of ketone body status on occurrence of first ovulation during early lactation was assessed in 84 multiparous dairy cows under field conditions. Animals were equally distributed across 8 farms and were controlled by the same herd fertility monitoring program. Cows were visited twice antepartum and 6 times postpartum at weekly intervals between 5:30 and 8:30 AM. On these occasions, body condition scores and milk yields were measured, blood and milk samples were taken, cows were gynecologically examined, and parameters of reproduction were determined. The onset of first ovulation was specified by milk progesterone determination and rectal palpation. Cows starting postpartum ovarian cyclicity within or after 30 d were classified as early and late responders (ER and LR, respectively). Resumption of the estrous cycle within 30 d postpartum is considered optimal under practical conditions, and classification based on this threshold value resulted in groups of equal size and equal distribution of ER + LR cows within farms. Ketone bodies measured were beta-hydroxybutyrate in serum and acetoacetate and acetone in serum and milk. Blood serum and milk ketone body concentrations during the first 6 wk of lactation were higher in LR than in ER, whereas plasma glucose and nonesterified fatty acid and milk fat, protein and urea concentrations did not differ between groups. Maximal concentrations of ketone bodies from parturition to first ovulation were better predictors of the onset of the estrous cycle than mean or minimal concentrations over the same period. Milk acetone and serum beta-hydroxybutyrate concentrations provided the most reliable information with regard to resumption of ovarian activity of all ketone bodies.