Variable intergroup encounters: What drives neutral and intolerant encounters in blue monkeys?

In group-living animals, between-group relations are often revealed during intergroup encounters, which may be intolerant (competitive), affiliative (cooperative), or neutral. Our study investigated factors that predict different encounter types in blue monkeys, where intolerant encounters predominate but neutral encounters also occur. Using data from 9 groups monitored over 24 months in equatorial forest, we first examined whether the availability of fruit, a preferred and important food, predicted monthly encounter rates for individual groups. We then examined how fruit availability and time of day influenced the odds that individual encounters were intolerant vs. neutral. For a subset of encounters between study groups with well-known histories, we examined the effects of additional social and location factors, which included group size, the proportion of females with infants, the time elapsed since the opposing groups had split from each other in the past years, and the encounter’s location on edge-centre axes for both groups. Finally, we examined how ecological and social factors predicted individual group behaviour (aggressive, fleeing, neutral) during encounters. The variables we examined included encounter location, group size, group size differential, time since the two groups split from each other, fruit availability, and time of day. These investigations required 3 data sets (one for each italicized item in this abstract, namely rates, encounters, and records of group behavior).

Methods
Data were collected from January 2019 through December 2020 via field observations of wild, individually recognized blue monkeys (Cercopithecus mitis stuhlmanni) living in nine groups in the Kakamega Forest, Kenya. The study groups were part of a long-term project that began in 1979. During the study period, each group was monitored on a nearly daily basis from ca. 0800-1700 hrs by members of an experienced field team that recorded all occurrences of intergroup encounters, both intolerant and neutral (see publication for definitions of these two categories). These records included the time of the encounter, whether it was intolerant or neutral, and (for intolerant encounters) a brief description of the opponent groups' behavior. All encounters included at least one study group, as we detected them by following these groups; however, opponent groups could be another study group or a non-study group.

In addition, the field team took a daily census of the group/s they monitored, as all group members could be individually identified based on natural features. If an individual adult female or juvenile was not spotted on a given day, we assumed it was present if it had been seen in the group both before and after that day. We did not make the same assumption for males, as they are more ephemeral members of social groups. To measure group size on each day, we totaled the records of individuals present (or assumed present) in the group.

A single field assistant, E. Shikanga, also collected monthly data on fruiting phenology, scoring the same ~10 individual trees from each of the 29 main feeding trees, on a semi-quantitative (0-4) scale for the presence of fruit. From these records, we computed mean monthly scores for each of the 29 species, and combined this information with data on the abundance and size of individual trees to compute a fruit availability index (FAI). The abundance and size of individual trees were assessed through vegetation surveys, which preceded the data collection period of this study. The survey enumerated and measured the diameter at breast height (DBH) of all trees with DBH >10 cm in 29 randomly placed transects (4-16 per group home range, proportional to varying home range sizes; each transect plot measured 10m x 100m). From these vegetation survey data, we calculated the mean basal area per transect for each of the 29 phenology species. To compute the monthly FAI for each study group, we multiplied the mean monthly fruit score of a given phenology species by its mean basal area per transect in that group's home range, and summed these products for the 29 monitored species. The monthly FAI values were then standardized as a Z-score, using the mean and standard deviation for each study group across the months in this study in which it was monitored. Further details are available in the associated publication.

We used the intergroup encounter records to specify the behavior of study groups involved in each encounter. Group behavior fell into one of three categories: (1) aggressive (any group member threatened, chased, or grappled with the opponent group); (2) fleeing (group fled from opponents without any members showing aggressive behaviour); or (3) neutral (neither aggressive nor fleeing). 

From the data collection and processing described above, we created three data files for analysis in the publication, all of which are included in this Dryad submission. 

(1) In "Cords-Thurau_Datafile_1_Rate_data", each line of data represents a particular group in a particular study month. We included a rate variable, based on the monthly sum of intergroup encounters recorded relative to total group observation time in that same month. Other variables in this data set are described in the ReadMe text.

(2) In "Cords-Thurau_Datafile_2_Encounter_data", each line of data represents a single intergroup encounter, classified as intolerant or neutral. Other variables in this data set are described in the ReadMe text.

(3) In "Cords-Thurau_Datafile_3_Group_behaviour_data", each line of data represents a record of the behavior of one study group in one encounter, with behavior categorized as aggressive, fleeing or neutral. Other variables in this data set are described in the ReadMe text.