Cr. Fonseca et G. Ganade, ASYMMETRIES, COMPARTMENTS AND NULL INTERACTIONS IN AN AMAZONIAN ANT-PLANT COMMUNITY, Journal of Animal Ecology, 65(3), 1996, pp. 339-347
1. In the tropics, many plants offer housing and food for their specia
lized ant partners which, in return, offer benefit in the form of defe
nce and/or nutrients, thus forming mutualistic bonds. Such ant-plants,
also called myrmecophytes, occur together at a local scale, generatin
g community patterns of mutualistic ant-plant associations. Here, we p
resent the first fully quantitative description of an ant-myrmecophyte
community. 2. The study site in Central Amazonian tropical rainforest
had a high myrmecophyte density of about 380 ind. ha(-1). Sixteen myr
mecophyte and 25 ant species were recorded, the species abundance rank
curves being highly uneven. 3. The ant-myrmecophyte matrix was highly
compartmentalized, and a Monte Carlo simulation showed that the obser
ved pattern was not a product of chance and sample size (P < 0 . 0001)
. Cluster analyses indicated that compartments were partially explaine
d by occurrence of the ants in phylogenetically related host plants, b
ut not by habitat specificity. 4. The connectance of the ant-plant com
munity was 12%. This value seems quite low when compared with publishe
d results from other mutualistic systems (pollinator and seed-disperse
r), after controlling for the total number of interacting species. The
high frequency of null interactions in the ant-myrmecophyte system co
uld not be explained by the 'phenological non-coincidence hypothesis',
since both ant and plant partners occur together throughout the year.
5. Ant-plant interactions were highly asymmetrical: ant species had f
ewer partners than plant species and ants were more dependent on the p
lants than the reverse. These asymmetries are in the opposite directio
n to those recorded for plant-pollinators and plant-dispersers; howeve
r, they seem to be :he product of the same underlying process: differe
ntial fitness benefits between mutualistic partners. 6. The low number
of ant and plant partners per compartment, coupled with an apparently
high temporal and spatial stability of ant-myrmecophyte interactions,
suggests that compartments are the appropriate scale at which to inve
stigate coevolution in ant-myrmecophyte systems.