Breakdown of an Ant-Plant Mutualism Follows the Loss of Large Herbivores from an African Savanna

Abstract

Mutualisms are key components of biodiversity and ecosystem function, yet the forces maintaining them are poorly understood. We investigated the effects of removing large mammals on an ant-Acacia mutualism in an African savanna. Ten years of large-herbivore exclusion reduced the nectar and housing provided by plants to ants, increasing antagonistic behavior by a mutualistic ant associate and shifting competitive dominance within the plant-ant community from this nectar-dependent mutualist to an antagonistic species that does not depend on plant rewards. Trees occupied by this antagonist suffered increased attack by stem-boring beetles, grew more slowly, and experienced doubled mortality relative to trees occupied by the mutualistic ant. These results show that large mammals maintain cooperation within a widespread symbiosis and suggest complex cascading effects of megafaunal extinction.

Senescent stem-galls in trees of Eremanthus erythropappus as a resource for arboreal ants

“…Ants of the genera Azteca, Crematogaster, Camponotus and Wasmannia are considered "truly arboreal" in that they nest in trees, whereas other genera display less specialized behavior and maintain nesting and foraging areas in many different substrates (Brown 2000;Silvestre et al 2003;Ribas et al 2003;Palmer et al 2008).…”

“…This type of behavior can strongly influence the occupation and hierarchization of canopies. According to Palmer et al (2008), in the African savanna, some species of Crematogaster play a key role in structuring the local arthropod community. Some may behave aggressively and significantly reduce the presence of other herbivores in the host plant (such as cerambycid borers), whereas other Crematogaster species may, on the contrary, not only tolerate the presence other insects (such as endophytics), but actively benefit from them, for example by appropriating of larvae-created cavities for nesting.…”

The Direct and Ecological Costs of an Ant-Plant Symbiosis

“…A2b, available online), and the presence of ants does not increase domatia size (M. E. Frederickson, A. Ravenscraft, L. M. Arcila Hernández, and G. A Miller, unpublished data). Unlike some other ant-plants (e.g., A. drepanolobium; Palmer et al 2008), C. nodosa exhibits little phenotypic plasticity in the production of domatia; in effect, every new branch has a hollow swelling, and the number of domatia is the same as the number of internodes on a tree. Ants change the number of domatia on a tree by affecting plant growth but do not increase the density of domatia on branches.…”

“…Numerous studies have found that ant rewards are produced sparingly, which certainly suggests that they are costly to make. For example, Heil et al (2001) and Palmer et al (2008) showed that both domatia and extrafloral nectar production can be induced by herbivory, and plants often produce extrafloral nectar and food bodies on only young, expanding leaves, which are the most vulnerable to folivory (Coley and Barone 1996). However, other previous tests of the costs of ants to plants have been inconclusive (i.e., Rudgers 2004; Rudgers and Strauss 2004;Rutter and Rausher 2004).…”

“…However, ants can also impose ecological costs on plants if they negatively affect the interactions between their host plants and other organisms in the environment. For example, ants can deter pollinators (Ness 2006) or increase herbivore loads on their host plants (Frederickson and Gordon 2007;Palmer et al 2008), such as by tending scale insects or aphids (Styrsky and Eubanks 2007). Direct and ecological costs are tightly linked (e.g., investing in ants instead of growth may reduce a plant's competitive ability; see also Heil 2002), but nonetheless we think the distinction is useful because of the different implications of direct and ecological costs for the evolution of cheating.…”