Polygyny does not explain the superior competitive ability of dominant ant associates in the African ant-plant, Acacia (Vachellia) drepanolobium

Abstract

The Acacia drepanolobium (also known as Vachellia drepanolobium) ant-plant symbiosis is considered a classic case of species coexistence, in which four species of tree-defending ants compete for nesting space in a single host tree species. Coexistence in this system has been explained by trade-offs in the ability of the ant associates to compete with each other for occupied trees versus the ability to colonize unoccupied trees. We seek to understand the proximal reasons for how and why the ant species vary in competitive or colonizing abilities, which are largely unknown. In this study, we use RADseq-derived SNPs to identify relatedness of workers in colonies to test the hypothesis that competitively dominant ants reach large colony sizes due to polygyny, that is, the presence of multiple egg-laying queens in a single colony. We find that variation in polygyny is not associated with competitive ability; in fact, the most dominant species, unexpectedly, showed little evidence of polygyny. We also use these markers to investigate variation in mating behavior among the ant species and find that different species vary in the number of males fathering the offspring of each colony. Finally, we show that the nature of polygyny varies between the two commonly polygynous species, Crematogaster mimosae and Tetraponera penzigi: in C. mimosae, queens in the same colony are often related, while this is not the case for T. penzigi. These results shed light on factors influencing the evolution of species coexistence in an ant-plant mutualism, as well as demonstrating the effectiveness of RADseq-derived SNPs for parentage analysis.

Keywords: Acacia drepanolobium; Crematogaster; Tetraponera; Vachellia drepanolobium; ant‐plant; coexistence; colonization; competition; mutualism; polygyny.

Figure 1

Ants inhabit Acacia drepanolobium trees on black cotton soils. On these soils, A. drepanolobium may account for 95% or more of trees, as shown on the left‐hand side (Young et al., 1997). Ants live in hollow, swollen thorns and patrol the tree against herbivores (right). Photographs: NEP (left) and JHB (right)

Figure 2

Average relatedness of workers between trees is close to zero, but relatedness within trees is high, and differs among ant species. (a) shows relatedness between ants on different trees (between‐tree comparisons); (b) shows relatedness between ants on the same tree (within‐tree comparisons). The species are arranged left to right in order from most to least competitively dominant. Boxplots show the median and inter‐quartile range for each species. Dots underlying each boxplot show the average relatedness between each pair of trees (a) or within each tree (b); they are jittered horizontally better to show their distribution. Lines above the boxplots denote significant differences between species as follows: *p < .05; **p < .01; ***p < .001. In (a), although the distributions are significantly different among the species, none are significantly greater than zero

Figure 3

Ant species vary in Polygyny Index, Polyandry Index, and number of male mates per queen. (a) The number of queen genotypes recovered from each tree; (b) the number of male genotypes recovered from each tree; (c) the number of male genotypes recovered from the offspring of each recovered queen (including only those recovered queens with a minimum of four offspring sampled in our data set). The species are arranged left to right in order from most to least competitively dominant. Boxes show median and inter‐quartile ranges. Dots behind each plot show the number of genotypes recovered from each tree (from each queen for c); the values are jittered slightly to help display the data. Lines above the boxplots denote significant differences between species as follows: *p < .05; **p < .01; ***p < .001