COS 76-8
Evidence of endogenous processes in the spatial patterns of an arboreal ant in an intensifying agroecosystem

Wednesday, August 12, 2015: 4:00 PM
301, Baltimore Convention Center
Kevin Li, School of Natural Resources and Environment, University of Michigan, Ann Arbor, MI
John H. Vandermeer, Department of Ecology and Evolutionary Biology, University of Michigan, Ann Arbor, MI
Ivette Perfecto, School for Environment and Sustainability, University of Michigan, Ann Arbor, MI

Spatial patterns in ecology can be described by density-dependent (endogenous) and density-independent (exogenous) processes. Understanding the role of these processes can have important management implications in agroecological systems. Azteca sericeasur is a tropical arboreal ant species that provides biological control in Central American coffee agroecosystems through interactions with mutualist and parasitizing species. Negative parasitic interactions increase for A. sericeasur with greater nest density, opposing the tendency of ant colonies to bud to new host trees. Together, these negative interactions and budding describe an endogenous, self-regulating process in the landscape that we compared with exogenous, environmental factors, including elevation, wetness, slope, tree density, tree-clearing intensity, and density of trees with nectaries. We used point process analyses and logistic regression to describe and model nest and transition (birth and death) event patterns in a 45-ha plot from 2004 to 2012. We described the scale of significant spatial clustering using a variant of the Ripley’s K analysis. We then used point cluster process models to investigate whether endogenous processes sufficiently explained nest spatial distribution, or whether exogenous variables also played a role. To further identify significant relationships between nest transition events and endogenous and exogenous variables, we used logistic regression models.


We found that nest clusters were predicted by nest density from the previous year and not consistently by exogenous variables, except by tree density in later years. Clustering occurred among nests out to a radius of about 40m, and new nests occurred within clusters of existing nests out to a radius of approximately 30m. This clustering most likely reflects budding colonies, and is suggestive of the ant’s preferred dispersal range. However, nest deaths had no significant clustering pattern with other nests. Logistic regressions suggested that nest births were positively related with measures of nest density and proximity, but nest deaths could not be predicted by any combination of variables. These results support our hypothesis that the Azteca sericeasur system is endogenously driven, although mainly through effects on new nests. Tree density became a consistently significant predictor of nest clustering, which coincided with a shift towards more intensive agricultural management that reduced tree density by one-third. We hypothesize that this effect is due to increased spacing between host trees that caused new nests to be further apart. Increased nest spacing could counteract negative density-dependent interactions with parasitizing species if those species have a shorter dispersal range than A. sericeasur.