Argentine Ant Invasion in Fynbos
The litany of threats that exotic plant species pose to Fynbos is well known. However, in addition to being choked by plant invaders, Fynbos also faces an indirect threat from an exotic ant, the Argentine Ant Linepithema humile (formerly Iridomyrmex humilis). The Argentine Ant is a tramp species that thrives in Mediterranean-type regions around the world, such as the Western Cape and coastal California. Though mostly found in human-disturbed areas, the Argentine Ant is slowly and steadily advancing into natural veld in these regions. Argentine Ants are especially worrisome in the Western Cape because they eliminate many native ant species that may disperse the seeds of over 20% of the Fynbos flora, including over 50% of proteas. The seeds of these plants attract ants with small, white, oily appendages called eliaosomes. In exchange for these nutritious rewards, ants bury seeds in their nest middens, where they consume the elaiosome and discard the bare seeds in locations that are safe from rodent seed predators and fire. Although invasive species may replace the role of the species they eliminate, Argentine Ants do not disperse seeds. We know from previous research that few unburied seeds survive the gauntlet of seed predators without assistance from native ants. Because many ant-dispersed species depend on soil-stored seeds to regenerate their populations after fire, there has been much concern that the loss of native ants from invaded areas could greatly compromise Fynbos recovery after fire.
Will the effects of the invasion ripple through Fynbos communities, culminating in the loss or decline of ant-dispersed plants as many of us fear? Predicting the magnitude of such an ecological chain reaction is notoriously difficult even in the simplest ecosystems, let alone in Fynbos where the diversity of ant and plant species is astoundingly high. The research I report on here expands on previous work by William Bond and colleagues by examining how specific the dispersal relationships are among the various ant and protea species found in Fynbos. In particular, my research addresses the following questions: Are there only a few ant species, such as the pugnacious ants (Anoplolepis spp.), responsible for most of the seed dispersal services in Fynbos? Or are many ant species involved in seed dispersal? The more seed-dispersing partners a plant species has, the less worrisome the loss of any one seed disperser would be. However, even if many ant species are involved in seed dispersal, ant species may differ in their efficiency in dispersing seeds of different plant species. Understanding the roles of the different ant species in seed dispersal is a necessary step for predicting which plant species are most vulnerable in invaded areas.
To investigate these questions, I have directed my efforts at studying the relationships among several species of ant-dispersed proteas found in the Kogelberg Biosphere Reserve and four very common Fynbos ant species. Two species are eliminated by the Argentine Ant (Pugnacious Ant Anoplolepis custodiens and Cape Nutcracker Ant Pheidole capensis) and two appear to coexist (Meranoplus peringuey and Black Garden Ant Tetramorium quadrispinosum). I am focusing on proteas because of their importance in Fynbos communities both in terms of their abundance and diversity (though clearly many other plant families would meet these criteria). A decline in these species would signal an overall change in the composition of many Fynbos communities. Despite the overwhelming diversity of ant-dispersed species within proteas, there are two general types of seeds - the large-seeded type (e.g. Leucospermum, Diastella, Mimetes, Paranomus, some Sorocephalus) and small-seeded type (e.g. Serruria, Spatalla, some Sorocephalus). These two groups of proteas are characterized by different elaiosome types and their seeds differ in weight by an order of magnitude. Are species within each of these two general seed groups responding to the Argentine Ant invasion in a similar way?
To examine seed dispersal patterns of large- and small-seeded proteas, I conducted seed-choice experiments in which I presented single seeds of three large-seeded and three small-seeded species outside of several nest entrances of the four ant species and the Argentine Ant. Overall, I found that the two ant species persisting in invaded areas (Tetramorium and Meranoplus) dispersed only 15% of the large seeds placed outside of their nests compared to 70% dispersed by those ant species displaced by Argentine Ants (Anoplolepis and Pheidole). The Argentine-Ant resistant and vulnerable ant species dispersed small seeds equally well, while Argentine Ants did not disperse either type of seed. Although there was some overlap in the sizes of seeds that the different ant species dispersed, I found that each ant species was most efficient at dispersing seeds of a certain size. For example, Meranoplus dispersed only small-seeded species, preferring the second smallest protea species I presented. In contrast, Anoplolepis was the only ant species capable of moving the largest of the large-seeded species (e.g. Ls conocarpodendron), and consistently dispersed twice as many large seeds than small seeds. This means that small-seeded species have a wider range of seed dispersers available to them than do the large-seeded species. Therefore, large-seeded species may be far more vulnerable than small-seeded species in areas in which Argentine Ants are eliminating native ants.
Probability of seed dispersal of large and small-seeded proteas by four native Fynbos ant species and the Argentine Ant
The protea species used in the study included (in descending order of
Are large-seeded proteas declining more in areas invaded by the Argentine Ant relative to the small-seeded species as suggested above? To measure the impact of the invasion, I have been monitoring post-fire seedling recruitment patterns after a prescribed burn in uninvaded and invaded Fynbos adjacent to farmlands in the Elgin basin. Argentine Ants have moved into natural Fynbos from these disturbed areas. Atlassers are well aware that this region is extraordinarily rich in ant-dispersed proteas (11 species co-mingling in my study plots!). At this site the pre-burn densities of adult plants did not differ between the invaded and uninvaded areas, suggesting that any differences in post-fire seedling recruitment levels between these areas would reflect the cumulative impacts of the invasion since the last fire cycle at this site (~30 years ago). As predicted from the seed dispersal data above, I found that seedling densities of three common large-seeded species combined were half as great as the seedling densities of these same species in the uninvaded areas. However, the abundance of three small-seeded species did not differ between invaded and uninvaded sites. These data illustrate the composition of the plant community is changing in invaded areas in the study area.
In summary, there is both good and bad news. The good news is that small-seeded proteas are dispersed in ant communities altered by the Argentine Ant. The likely reason is that at least two of their seed dispersers are capable of coexisting with the Argentine Ant. However, large-seeded species appear to be declining in areas where their primary seed dispersers have been exterminated by the Argentine Ant. Although the threat of the Argentine Ant pales in comparison to exotic plants, the alteration of even small areas could spell the decline of several narrowly restricted large-seeded species of protea.
It is with deepest gratitude that I thank the following people for support of my research: Cape Nature Conservation, Mark and Amida Johns, Hamish Robertson, and William Bond.
Back PAN 47