Abstract: We investigated dyadic grooming relationships in a captive group of bonobos (Pan paniscus) and questioned what social function grooming fulfils in the 'market of services and favors'. Hereto we examined which of two theoretical models – grooming for support (Seyfarth, 1977, 1980) or grooming according to the similarity principle (de Waal & Luttrell, 1986) – best accounted for the observed grooming distribution. Similarity in traits did not correlate with increased grooming or close proximity among the individuals. Therefore, the similarity hypothesis was rejected. Seyfarth's model of rank-related grooming was largely confirmed. The animals distributed their grooming according to the rank of the receivers. We found an exchange between grooming and receipt of support. There was more grooming up than down the hierarchy. However, not all predictions about rank-related competition over grooming were confirmed. We found that dyadic grooming reciprocity indeed increased with decreasing rank distance. Yet, there was no increase of grooming within the dyad with decreasing rank distance and high ranking individuals were not competed over at the highest rates. The observed correlation between grooming and support received represents an important fit with Seyfarth's prediction, but does not allow for conclusions about underlying causal processes. Other causal explanations, besides the 'groom to receive support' hypothesis, that could explain a similar correlation are discussed.