An African elephant (Loxodonta africana).

Blogging on Peer-Reviewed ResearchPart of what makes ecology such a fascinating subject is the complexity of interactions between species, especially when things don’t happen in exactly the manner we might initially hypothesize. Such is the case with a paper just published in the journal Science by Palmer et al. called “Breakdown of an Ant-Plant Mutualism Follows the Loss of Large Herbivores from an African Savanna,” which shows that the loss of large herbivores initiates a cascade of ecological changes, changes which have some important effects for acacia trees and the ants that live in them. Indeed, some of the largest animals in Africa are vitally important to some of the smallest.

As is well understood, trees provide homes and resources for a large diversity of animal species (often invertebrates), and in a number of instances trees have formed mutualistic relationships with some species (i.e. both the tree and the species inhabiting it benefit from the interaction). The tree that plays a major role in the study is the whistling thorn tree (Acacia drepanolobium), a tree common to clay-rich soils in East Africa that secretes nectar from areas near the base of leaves and produces hollow, swollen thorns at branch nodes. The hollow thorns house any of four species of ants, Crematogaster mimosae, Crematogaster sjostedti, Crematogaster nigriceps, or Tetraponera penzigi, the different species competing with each other for the use of the trees for homes and food. Although each species does defend the trees, they are not equally aggressive and do not have the same habits.

Crematogaster mimosae aggressively defends the tree against herbivores and depend upon the swollen thorns to raise young, house workers, and sometimes care for scale insects which produce honeydew. Crematogaster sjostedti does not use the hollow thorns, however, and instead lives in stem cavities formed by the larvae of long-horned beetles and are less aggressive than C. mimosae. C. mimosae is the dominant species occupying about 52% of all trees at the studied sites, with C. sjostedti occupying about 16% of the trees. C. sjostedti is still more dominant than Crematogaster nigriceps or Tetraponera penzigi, though, both employing behaviors to lessen the chance that their trees will be taken over by C. mimosaue or C. sjostedti. C. nigriceps controls plant growth at bud locations and kills apical meristems (thus reducing the lateral spread of the branches of the tree) and T. penzigi destroys nectar-producing parts of the tree to reduce the attractiveness of the tree to the other Crematogaster ants (T. penzigi also is the only ant of the four species that does not leave the tree in search of other food, relying instead on pollen, fungal spores, and other resources on the tree itself).

While C. mimosae might be the dominant ant under natural conditions, the researchers found that excluding the trees from browsing by large herbivores had drastic effects on the C. mimosae populations. In the plot of land where large herbivores (herbivores over 15kg) were excluded, most of the trees stopped producing as much nectar and the large hollow thorns used by C. mimosae for homes. Some of the trees kept producing nectar and hollow thorns, however, and these were the trees inhabited by C. nigriceps, the ant that nipped the apical meristems of the tree to keep it from growing too close to other trees. This pruning behavior may have been simulating browsing by herbivores, the production of nectar and hollow thorns by a plastic response by the trees to herbivores feeding on the trees. The loss of hollow thorns and nectar from many of the trees caused the abundance of C. mimosae to drop by about 30% (as well as a reduction in individual colony size), however, and C. sjostedti became the most abundant ant as a result (the populations of the other two species showed little to no change). The behavior of C. mimosae also changed, the ants becoming more likely to tend scale insects (insects that suck sap from the trees) and less likely to defend the trees against disturbance, and the trees also pay a price for these changes.

As noted before, C. sjostedti experienced increased abundance with the weakening of C. mimosae due to changes in the trees, but the trees protected from herbivore browsing and inhabited by C. sjostedti were weaker than trees that were browsed by large herbivores and inhabited by C. mimosae. Why should this be? If you will recall C. sjostedti ants rely upon excavations made into the tree by beetles for their homes, and it turns out that C. sjostedti ants actually help to facilitate the beetles in burrowing into the trees. C. mimosae and C. nigiceps actively protect their trees against the beetles. Indeed, the activities of the destructive beetles being important to the life cycle of C. sjostedti and so these ants are likely to let the beetles do as they please even though it is to the detriment to the tree (trees inhabited by C. sjostedti suffered twice the mortality of other trees in the study).

If large herbivores continue to be extirpated throughout Africa, then, an ecological shift is likely to occur in areas where ants rely on the defensive products (hollow thorns and sap) of trees that are browsed by the large mammals. Indeed, it could be predicted that ants that actively protect trees like C. mimosae would decline in abundance while ants like C. sjostedti that rely on another parasitic insect would thrive, resulting in trees occupied by C. sjostedti experiencing higher rates of mortality as a result. The interactions are certainly complex, but the study does reflect that the activities of large African herbivores do have an important effect on the landscape, their removal beginning a series of changes that may likely be to the detriment of not only the trees they fed on but also the creatures that inhabited those trees.

[Also check out this report on the new research on NPR.]


Palmer, T.M.; Stanton, M.L.; Young, T.P.; Goheen, J.R.; Pringle, R.M.; Karban, R. (2008) “Breakdown of an Ant-Plant Mutualism Follows the Loss of Large Herbivores from an African Savanna.” Science Vol. 319, pp. 192-195


  1. #1 AnimalEnthusiast
    January 11, 2008

    This article is very interesting. Thanks for the information!!

  2. #2 Caledonian
    January 11, 2008

    I noticed that articles on this topic in general media tended to include statements like “no one would have imagined that stopping browsing would harm the trees”.

    Given that disturbing any aspect of an ecology is likely to disrupt its evolved beneficial relationships, how could any researcher have said such a blatantly dumb thing?

  3. #3 Laelaps
    January 13, 2008

    Caledonian; The impact of megaherbivores on the makeup of their habitats, specifically the intersections of grassland/woodland/forest in Africa, has long been supposed. The problem has been that evidence has been difficult to pin down, but as you say I’m surprised that so many reports have feigned ignorance. While the level of complexity the researchers found might have been unexpected, I don’t see why anyone would think that the extirpation of large herbivores would leave any habitat unchanged. I guess it’s all about making a story seem more impressive, yet another trend in the mainstream media that leaves a bad taste in my mouth.