In all areas where they have been studied, male roe deer are believed to have a territorial mating system, although few quantitative studies have been conducted and there remains considerable debate about the function of male roe deer territories. We observed 139 aggressive interactions between male roe deer in Storfosna Island (Norway) during one territorial season (March-August). We recognised seven rank levels of escalation according to the potential danger of the behaviour. On the basis of the number of escalation levels included in the interactions, the complexity of the fights was also scored. We recorded the presence of other individuals during the interaction, the age, the antler size, the territorial status and the residency status of the two contestants and tested how these variables affected escalation, complexity and outcome of the fights. Most of the interactions ended with low levels of escalation, and physical contact occurred only in fights between two territorial bucks. The escalation was also affected by the difference in antler size index (the bucks escalated more when the difference in antler size was smaller) and increased with an increasing number of female deer present during the interaction. The resident buck won in 81% of the fights. When it drew or lost, it was generally both inferior in age and antler size, and the duration and escalation of the interactions were higher. However, even when a fight was lost, no territory loss occurred. These results are consistent with the evolutionary game theory and the proposed low risk-low gain strategy of roe deer bucks.

R. 2005. Age-specific reproductive behaviours in male roe deer Capreolus capreolus. Acta Theriologica 50: 445-452.The behaviour of active male roe deer Capreolus capreolus Linnaeus, 1758 was studied in a wild population on Storfosna island (Norway), during the pre-territorial and territorial seasons (February-August) of 1994. Observed behaviours were grouped in seven categories (mating, aggression, display, marking, vigilance, moving and maintenance), and the proportion of time spent in each behaviour by three age classes (yearlings, subadults and adults) was compared. Time spent in each behaviour varied significantly between months for all male age classes. From February to August, time spent in maintenance decreased, while time spent in vigilance increased. The time spent in mating increased both in adult and subadults but did not vary significantly for yearlings throughout the months. When comparing age classes within a month, there were no significant differences in time spent in each behaviour. Surprisingly, adult territorial males did not invest significantly more time in mating, with respect to yearlings and subadults. This may be consistent with the oligogynous mating system of roe deer, in which the reproductive success is divided amongst several years and their territoriality, which could be a tactic to reduce the competition for mating.