Overdispersion of ant colonies: a test of hypotheses.

The nest locations of two ant species in the Colorado Desert are intraspecifically overdispersed. Intraspecific overdispersion has been thought to represent strong intraspecific competition. Here we consider this hypothesis along with three competing hypotheses: microhabitat selection by foundress queens, predation on foundress queens, and predation on established colonies. To test these hypotheses five types of data were collected: (1) the forager population sizes of Veromessor pergandei and Pogonomyrmex californicus, (2) the response of the territory use of V. pergandei to varying levels of food, (3) the encounter rates of conspecifics and other ant species to foundress queens artificially placed near and far from conspecific colonies, (4) predation on colonies as a function of colony spacing, and (5) the relationship between the plant microhabitat at the nest and colony spacing. The results show that established colonies have no apparent selectivity for a particular type of plant microhabitat nor do foundress queens show avoidance or attraction toward conspecifics. V. pergandei workers show only a slight ability to find V. pergandei queens that are artificially placed near their entrances. Certain spiders are the most common ant predators on our study area. Direct observations on spiders indicate that colonies with closer neighbors are not prone to a higher risk of predation. In addition, the estimates of the death rate of workers from a mark-recapture technique indicate that colonies with closer neighbors lose similar numbers of workers as compared to colonies with further neighbors. In favor of the competition hypothesis, the summed size of intraspecific nearest neighbor pairs are larger for colonies that are spaced further apart than those colonies that are spaced closer together. We also develop an index of foraging directionality for the column foraging species V. pergandei. Using this measure, we find that nearest neighbors tend to avoid foraging toward each other. The response of territory use to food levels was tested with experiments involving patches of cracked wheat. These experiments showed that patches away from nearest neighbors were found significantly sooner than patches toward nearest neighbors. The above five sets of data together suggest that resource competition and perhaps queen predation by established colonies account for the intraspecific spatial patterns of these species.