Male Dispersal and Association Patterns
There was a significant correlation between the availability of fruits and the number of males transferring between core units in a given month (Spearman: Rho = 0.50, Z = 2.33, Monte-Carlo approximatedp -value = 0.048). Conversely, there was no relationship between the number of male transfers and rainfall (Rho = 0.14, Z = 0.71, P = 1.00), or the availability of young leaves (Rho= 0.22, Z = 1.05, P = 0.87). Thus, male transfers were most likely to occur when fruits were abundant and core units aggregated. Furthermore, the core units in which males transferred between were more likely to keep associating (i.e., maintaining a high AI) than would be expected given their baseline level of association with other units. Analyses revealed significantly higher associations (P < 0.05) of the dispersal dyad for 1-2 months post-dispersal than would be expected given their association with control core units (Fig. 4). However, we found that by the third post-dispersal month, all dispersal dyad AIs were no longer significantly different from the null. It is noteworthy that the core unit that switched its clan association from clan 1 to clan 2 between the two sampling periods (Newtonia) may have had this increase in association with clan 2 because five males dispersed from this core unit to a unit in clan 2 (Fagara).