Group living is a strategy utilized by organisms as diverse as bacteria, insects, fish, and mammals. It confers the advantage of better access to mates, food, and protection from predation, but also the disadvantage of increased competition for resources and higher visibility to predators. As an evolutionary response to one or more of these pressures, socially-living individuals may organize themselves into particular configurations to decrease the probability of events that could be detrimental to their own fitness. Determining the key factor or factors that drive the selection for a specific aggregation pattern poses a challenge because there are often not clear boundaries defining the group. However, in arboral bird colonies there lies a unique opportunity to investigate the architecture and disposition of their living arrangement, as the limits of the colony are constrained to one tree. Our study aimed to determine the factors that weigh most heavily in the determination of the nest arrangement (or spatial distribution) of three species of weaver birds, Pseudonigrita arnaudi, Pseudonigrita cabanisi, and Plocepasser mahali, of the African woodlands at the MPALA research station in Laikipia, Kenya. We specifically asked if the spatial distribution of their nests within a given tree is determined by the tree architecture or the weaverbird population in the tree.
Using data that we collected from 515 nests in 16 acacia trees, we computationally modeled the factors that determine the spatial relationships between the nests in a given tree. An approach using Principle Component Analysis and Random Forests yielded canopy size, number of nests, and distance between nests as the determining factors. These results all relate to the available space within a given tree, thus pointing to tree architecture as the key factor determining nest arrangement. Future work on the order of nest construction will address whether nest location is solely based on available space or also on interaction between neighboring birds. In addition, environmental conditions should be analyzed to look at the potential influence of solar radiation and wind speed, as suggested by field observations.