Relationship Quality is defined by the pattern and frequency of interactions. Cords and Aureli proposed that it can be described by three components: Value, Compatibility and Security. In captive animal groups, social behaviour might be influenced by limited space and human interventions and therefore might differ in some aspects from natural social behaviour. In semi-captive enclosures the space limitation is more moderate and human interventions are normally restricted to the bare minimum required. Apart from the missing gene flow, which might lead to inbreeding, the social behaviour should be almost natural. I studied the relationship quality of a semi-captive group of Japanese macaques. To gain knowledge of the paternal kinship I did paternity tests and I resolved the paternity of 43 out of 85 Landskron-born individuals. Three-fourths of the parent pairs were not related, in two cases the relatedness was based on r=0.125 and in one case on r>0.125 (10 pairs relatedness uncertain). Permutation tests showed that reproduction by closely related individuals was as frequent as expected by chance. No evidence of active inbreeding avoidance could be found. Paternity was distributed over a large number of males (19 out of 34). Reproductive success was highest in some subordinated males, but maybe due to the age of these males and the attractiveness of young, unfamiliar males. On average reproductive success was higher in top ranking males. Becoming a top ranking male may entail reproductive advantage and therefore a higher fitness. Grooming occurred significantly more often in lineal related kin than in non-kin individuals. This supports the kin-selection hypothesis which says kin should favour each other because of inclusive fitness benefits. Furthermore, on dyadic level, aggression and submission occurred significantly more often in kin than in non-kin, although on the level of the focal individual less aggressive and submissive interactions were found in lineal kin.