Deag, J. M. 1980. Interactions between males and unweaned barbary macaques: testing the agonistic buffering hypothesis. Behaviour 75, 54-81.

Observations of wild Barbary macaques in the Moroccan Middle Atlas revealed more male interaction with unweaned monkeys than has been described for other Cercopithecoidea. 2. Adult, subadult and juveniles males were all involved. On the birth of a new season's babies the males interacted with them and virtually ignored the infants (still unweaned) from the previous birth season. They continued to interact with the new babies when these were infants. 3. Adult and subadult males approached mothers with babies, sat with them, removed their babies and cared for them. Babies were rarely threatened, or attacked, but in agonistic interactions infants were treated more like juveniles. 4. Close non-agonistic interaction between males occurred in the presence of babies (type b interactions). This was initiated by males taking babies to other males, or by males without babies approaching those with them. A framework (based on the most consistently repeated behavioural elements) is presented for describing the seven phases of these complex interactions. 5. Elements of male-care and type b interactions often punctuated play bouts between infant and juvenile males and babies. The interactions involving males, babies and non-maternal females are also briefly described. They were far less frequent than type b interactions and female-care interactions. 6. Individual males and babies did not interact with each other equally often. Particular male-baby partnerships occurred regularly in the context of both male care and type b interactions. The males' rank did not determine these relationships. 7. Examination of nearest neighbour records revealed that the presence of a baby near a male was associated with a clumping of other individuals (particularly other males) near the male. Infants did not have this effect. 8. Most friendly approaches given to and received from males (particularly those to and from other males) were in the presence of babies. 9. There was a relationship between agonistic rank and approaches between males in the presence of babies. Between unlike age classes and between adults, an approacher carrying a baby was usually subordinate. Most approaches to males in possession of babies were by subordinates but there were exceptions. 10. Males carried babies away from and left babies with males both dominant and subordinate to themselves. This showed how limited competition for babies was between males. 11. There was a higher probability of a male being threatened if he made a friendly approach in the presence of a baby. However, it seemed that males found the risk of receiving an aggressive response acceptable because it was coupled with a higher probability of non-agonistic interaction with males who could rarely be approached without a baby present. 12. Low ranking males approaching high ranking males positioned themselves so as to judge the response of the approached male and they also drew the baby to his attention. Several behaviour patterns were used to maintain proximity with other males. Most male-male grooming occurred in the presence of babies. 13. The results confirmed the importance of a baby's presence in the dynamics of male-male interaction and justify the use of the term 'agonistic buffering'. The behaviour between males of considerably different rank (e.g. between subadults and adults; juveniles and adults) most clearly fitted the agonistic buffering hypothesis. The behaviour between males of more similar rank (e.g. adults and adults) was less consistent and did not fit the hypothesis so closely.