Temporal patterns in the social network of core units in Rwenzori Angolan colobus monkeys: Effects of food availability and interunit dispersal

Abstract

Multi-level societies are complex, nested social systems where basic social groups (i.e., core units) associate in a hierarchical manner, allowing animals to adjust their group sizes in response to variables such as food availability, predation, or conspecific threat. These pressures fluctuate over time and examining the extent to which this variation affects the clustering of core units into different tiers may be instrumental in understanding the evolution of multi-level societies.The goal of our study was to determine the degree of temporal variability in interunit associations in a multi-level society of Rwenzori Angolan colobus monkey (Colobus angolensis ruwenzorii), and to determine the social and ecological factors that underlie association patterns. The C. a. ruwenzorii multi-level society consists of at least three tiers, with core units clustering into clans that share a home range in a band tier.We performed social network analyses on 21 months of association data from 13 core units (totaling 139 identifiable individuals) at Lake Nabugabo, Uganda. We described the patterns of variation in core-unit associations over time and investigated how changes in rainfall, food availability, and interunit dispersals were correlated with these associations over the short-term (month to month) and long-term (year to year).Although clans were relatively stable, larger-scale changes in association patterns included the formation of an all-male unit and the transfer of one core unit between clans (within the band tier). Seasonally, core units associated significantly more when fruit, their preferred food source, was abundant (i.e., social networks were denser and more clustered) and there was no direct effect of rainfall seasonality or young leaf availability. Male dispersals also occurred more during periods of high fruit availability, suggesting that greater band cohesion allowed males to prospect and transfer between core units. Once males transferred, their previous and new units associated significantly more with one another than with other core units for 1-2 months postdispersal. The dispersal of five males from one core unit to another in a different clan co-occurred with this core unit switching its clan affiliation.By examining temporal shifts in social network structure among core units, this study shows the interconnected roles that food availability and dispersal have in shaping the C. a. ruwenzorii multi-level social system. Our findings highlight how ecological conditions can drive association patterns, impact interunit relationships, and influence social organization.

Keywords: colobines; core unit associations; food availability; food competition; male dispersal; social network analysis.

FIGURE 1

An adult male Colobus angolensis ruwenzorii at Nabugabo, Uganda. (Photo credit: Samantha Stead)

FIGURE 2

Comparison of Colobus angolensis ruwenzorii core units’ clan associations across years at Nabugabo, Uganda, depicted by dendrograms created using hierarchical cluster analysis (CCC = 0.891; SOCPROG: Whitehead, 2009). * indicates the movement of a core unit between clans. The “Year 1 – 2017–2018” dendrogram has been adapted from figure 1a in Stead and Teichroeb (2019)

FIGURE 3

Stability of the social network between core units of Colobus angolensis ruwenzorii at Nabugabo, Uganda from August 2017 to May 2019 in both the (a) short‐term and (b) long‐term, as illustrated by the cosine similarity (a) to the previous month and (b) to the first month of the study period. Shaded areas indicate the 95% highest density interval from 100 bootstrap samples of the observed data

FIGURE 4

(a) Available data on ecological conditions (rainfall, young leaf and fruit availability) with (b) social network density measures, (c) node strength, and (d) the timing of male dispersal events from August 2017 to July 2019 in a band of Colobus angolensis ruwenzorii at Nabugabo, Uganda. Core unit association levels are represented by (b) and (c). For (b) the shaded area represents the density range expected due to chance encounters, and was calculated using data‐stream permutations using 95% CI

FIGURE 5

Association indices for the Colobus angolensis ruwenzorii core units at Nabugabo, Uganda that males transferred between (solid lines) for the 3 months following each male dispersal event from August 2017 to May 2019. Dashed lines represent expected (baseline) levels of association, given how much these core units (i.e., those involved in male transfers) continued to associate with other core units not involved in the male dispersal event. (*) indicates association indices that were significantly higher in the core units with male transfer than would be expected at α = 0.05