Title: Causes and consequences of coexistence in the Vachellia drepanolobium ant-plant mutualism
Abstract: This thesis focuses on a mutualism between the East African acacia tree Vachellia drepanolobium and the species of canopy-dwelling ants that inhabit it. The tree provides the ants with nesting space in the canopy and extra-floral nectar, and in return the ants defend the tree from herbivores. Several different ant species compete vigorously with each other for this canopy nesting space. Despite this, coexistence is maintained among those species at fine spatial scales, apparently by a colonization-competition tradeoff: ant species that are the best at colonizing unoccupied trees are the worst at defending those trees from neighboring competitors, and vice versa.
In Chapter 1, to determine the causes of these different competitive and colonizing abilities among the ant species, I used RADseq, a next-generation sequencing technique, to determine the colony structure of each ant species. This enabled me to reject a primary hypothesis for why the ant species differ in competitive ability, which is that the more competitive ants have more than one egg-laying queen per colony, and therefore these queens can produce much larger colonies than other ant species that have only one queen. The RADseq technique is sufficiently powerful that by sampling about 6-8 worker ants per tree, I could estimate the number of queens in the colony as well as the number of times that each queen had mated. I found that more competitive species do not have more egg-laying queens in their colonies.
In Chapter 2, I looked for alternative explanations for differences in competitive and colonizing ability. I measured how individual colonies divide their biomass among different castes, and how these castes are distributed throughout the tree. I found that the three main ant species inhabiting V. drepanolobium differ significantly in colony composition as well as the distribution of castes throughout the tree, and some of these differences are likely to shape competitive and colonizing abilities.
I also performed a broader scale survey of V. drepanolobium and its mutualist ant species across their range in Kenya. In Chapter 3, I show that this mutualism varies widely across Kenya, for both the tree and the ant. I also use the measurements from Chapter 2 to estimate how the biomass of each partner varies in the landscape, and compares to other parties in the system, such as vertebrate herbivores.
In Chapter 4, I present a RADseq-based landscape genomics study of V. drepanolobium and its ant inhabitants. This analysis uncovered how differences in ant behavior at small spatial scales were recapitulated at landscape scales. I also explored whether abiotic factors influence the different ranges of the tree and ant species. I found that geography contributed to population structure in the tree and in each of the ants, but that environmental distance had a weak or negligible effect on the population structures of the tree and of two of the three associated ant species. Moreover, the tree had a population structure distinct from its ant inhabitants. These results suggest that population structure in this system may be influenced more by dispersal limitation than by local adaption of specific partners to each other or to the abiotic environment.