Burning bridges: priority effects and the persistence of a competitively subordinate acacia-ant in Laikipia, Kenya.

Theoretical models suggest that subordinate competitors may rely on strong colonization ability and/or high persistence (e.g., the ability to resist invasion) as a strategy to coexist with competitively dominant species. While strong colonization ability among subordinate competitors has been widely documented, we know less about the role of persistence in facilitating species coexistence. In upland East Africa, four species of acacia-ants (Crematogaster sjostedti, C. mimosae, C. nigriceps, Tetraponera penzigi) compete for possession of Acacia drepanolobium host trees. Despite a strong dominance hierarchy, the four acacia-ant species coexist at fine spatial scales. Here we present evidence that T. penzigi, the least aggressive competitor, modifies host trees in two ways that reduce the probability of aggressive takeover by neighboring colonies. First, T. penzigi workers destroy virtually all leaf nectaries on their host trees. Second, T. penzigi workers create and maintain entryways into their swollen thorn domiciles that are too small to allow entry by their Crematogaster competitors. In a 2×2 factorial experiment, we manipulated nectar availability and swollen-thorn entryway size to determine the influence of these factors on the probability of aggressive displacement by a dominant competitor (C. mimosae) in staged conflicts. Addition of artificial nectaries and enlargement of swollen-thorn entryways on T. penzigi-occupied trees increased the probability of aggressive displacement of T. penzigi by C. mimosae from swollen thorns 14-fold and 8-fold, respectively. Further, empty saplings with nectaries destroyed by T. penzigi workers were colonized by half as many C. mimosae workers as saplings where nectaries were left intact. Our results demonstrate that T. penzigi's unusual strategy of nectary destruction and the maintenance of small entryways in swollen thorns produce priority effects, effectively reducing the probability that T. penzigi colonies will be displaced from host trees by more dominant competitors.