We carried out behavioural observations to investigate the function of long-term investment in vocal display by fallow bucks during the breeding season. The measures of long-term investment used were the date of initiation of vocal activity, the number of days vocal during the breeding season, and the proportion of time spent vocalizing. We analysed data from 3 years (1993-1995) to assess the relationship between the date of initiation of vocal activity, and the number of days vocal, and age, dominance rank and mating success. Observations from a sample of focal males in 1996 were used to determine the effect of the proportion of time vocal during the breeding season on dominance rank and mating success. The majority of socially immature males (</=3 years old) did not vocalize; among socially mature males (>/= 4 years old), dominance rank was more important than age in explaining variation in vocal activity. The onset of vocal activity by fallow bucks was not a direct consequence of the presence of mating opportunities since the first males became vocal more than 3 weeks before any matings occurred. Long-term investment in vocal activity did not alter the dominance relationships that had been established between males before they became vocal. When we considered all mature males from 3 years of observations, the majority of matings were achieved by those that had high rank, initiated vocal activity early during the breeding season and remained vocal on most days. For the 1-year sample of mature males, the factor most highly correlated with mating success was the proportion of time that males spent vocalizing during the rut. Thus we have shown a strong relationship between the time invested in vocal display by fallow bucks and their mating success. Copyright 1999 The Association for the Study of Animal Behaviour.