Farm mink (Mustela vison) have been selected on the basis of their behaviour towards man since 1988 at the Danish Institute of Agricultural Sciences and in this study were offered the possibility to mate 1 week earlier than usual. The objective was to investigate if the behaviour-related selection has affected the reproduction of the farm mink, measured as mating willingness and reproductive success. The animals belonged to three breeding groups: A: selected for curious/confident reactions (17 males, 73 females), B: selected for timid reactions (17 males, 74 females), and C: selected without any demands on reactions towards humans (18 males, 73 females). The time when an average of 50% of the population were mated was 3-6 days for group A; thus the confident were mated 1·7 to 2·1 days earlier than the timid (B) and the control (C) animals. The length of gestation was shorter in group A (46·0 (s.d. 3·0) days) than in group C (47·6 (s.d. 3·6) days) but not different from group B (47·0 (s.d. 3·2) days). Group C had a lower kit mortality from birth to day 50 (11·3%) than groups A and B (20·4 to 21·2%). No significant differences were found between the groups regarding the frequency of successful matings, the ratio of remated females, number of interrupted matings, barren females or litter size. It is concluded that 8 years of selection has led to the development of reproductive differences primarily in the time of mating readiness, whereas the differences seen in kit loss may be related more to random effects.