Main content area

Effects of Seed Predation by a Tropical Ant on Competition Among Weeds

Risch, Stephen J., Carroll, C. Ronald
Ecology 1986 v.67 no.5 pp. 1319-1327
Capsella, Cenchrus ciliaris, Cynodon dactylon, Daucus carota, Festuca arundinacea, Leptochloa dubia, Panicum, Paspalum distichum, Pennisetum clandestinum, Solenopsis geminata, asexual reproduction, biomass, feeding preferences, fire ants, grasses, habitat destruction, harvesting, nests, plant competition, planting, seed predation, seedlings, seeds, weeds, Central America, Mexico
The native fire ant, Solenopsis geminata, harvests small—sized seeds, especially grasses, in disturbed habitats in wet tropical areas of Mexico and Central America. When common, this ant can lower overall abundance of many weedy species, grasses in particular. We studied experimentally the effect of the ant on pairwise competition between weed plants, using a replacement series design; the pairs were (1 Paspalum distichum and Daucus carota, (2) Leptochloa dubia and Capsella bursapastoris, (3) Panicum clandestinum and Pennisetum ciliare, and (4) Cynodon dactylon and Festuca arundinacea. When the ant strongly preferred seeds from one of the plant species, total plant biomass was significantly lower (by up to 50%) in the plots with ants for about the first 50 after plants. Subsequently, the nonpreferred species increased, and 83 d after planting, total plant biomass was about the same in the presence and absence of ants. A crop growing with ants could thus benefit from reduced competition early in the season. For one of the plant pairs, ants reversed the course of plant competition and changed which of the species was eventually excluded from the plots. For two of the weed pairs, when the ant preferentially ate the seeds of the inferior competitor, the inferior competitor disappeared more rapidly. For the fourth pair, ants created a stable equilibrium between two species, whereas otherwise one of the plant species would have disappeared. Feeding preference studied conducted in the field demonstrated that the amount of each seed type removed by the ants was strongly influenced by the amount and kinds of other seeds in the immediate area. A strongly preferred seed was removed less frequently if it occurred in the midst of nonpreferred seeds. Nonpreferred seeds were taken much more readily if they occurred with preferred seeds. The importance of this "seed environment" effect was less significant in determining final composition of a weed mixture than: (1) inherent preference of the ants for certain weed species, or (2) relative competitive ability of the weeds in the mixture. Although S. geminata can increase the proportion of dicots in a weed assemblage, several factors limit this process: (1) as grasses become rare, a large fraction of their seeds will fall into the protective cover of dicots, (2) as grasses become rare, S. geminata may move their nests to areas of higher food abundance, and (3) the vegetative reproduction of some grasses occurs independently of seed removal by ants, or may even be accelerated by the reduced competition with grass seedlings.