Social grooming is often exchanged between individuals in many primate species. Rates of bidirectional (or simultaneous mutual) grooming vary across primate species, and its function is not yet fully understood. For example, mutual grooming is frequent in chimpanzees but rare in most primate species including wild bonobos. There are, however, no quantitative data available in captive bonobos.

Therefore, through the direct comparison between captive bonobos and chimpanzees, this study aimed to:

• (i) compare the frequency of mutual grooming between Pan species,
• (ii) explore and compare the function of mutual grooming, and
• (iii) discuss the rarity of this behavior in wild bonobo populations.

We tested three hypotheses following the previous literature in wild chimpanzees. The social bonding hypothesis states that mutual grooming facilitates the maintenance of strong dyadic bonds. The immediate investment hypothesis states that it serves to signal willingness to invest in the interaction. The switching hypothesis states that mutual grooming serves no function but only occurs as an overlap to change the direction of unidirectional grooming.

Our findings strongly supported the immediate investment hypothesis, but not the others. Grooming bouts that included mutual grooming were longer and more equitable than bouts without, illustrating that captive Pan species use mutual grooming to maximize their short-term benefits and increase the social value of their interaction. Captive bonobos performed mutual grooming in similar proportions and for similar functions as captive and wild chimpanzees do. This contrasts with wild bonobos who engage in this behavior only rarely.

We suggest that the differences in patterns of mutual grooming between captive and wild bonobos might be explained by different degrees of kinship or by a potential intraspecies variation.