Insect communities on woody fabaceous plants: A comparison of tropical and temperate dry sites

I studied insect communities on woody fabaceous plant species in the Sonoran Desert, Arizona, U.S.A., for three years. This showed that the variation in number of insect herbivore species associated with their host plants was due to the availability of food resources such as stems, leaves and flowers. Other resources such as pods (fruits) had no effect on the number of insect herbivores associated. I also tested the effect of plant abundance and area of plant distribution on the number of insect herbivores associated with their host-plants and these were not significant.;I compared the insect herbivore species richness on woody fabaceous host-plants in the Sonoran Desert (temperate) to that in host-plants in Cerrado (tropical) to evaluate the role of host-plant characteristics and predation by invertebrate carnivores on the number of insect herbivores associated with their host-plant species. For a one time censusing the insect communities on woody fabaceous shrubs in the tropical site had a depauperate insect herbivore fauna when compared to the temperate site. The resource diversity index accounted for 80% of the variation in insect herbivore species richness between plant species. Insect herbivores on tropical shrubs showed greater specialization since 68.57% of the insect species utilized only one host plant species.;I studied the distribution and abundance of insects on woody fabaceous host-plant species along an elevational gradient in Cerrado (tropics) and evaluated the role of host plant species richness on the number of insect herbivore species associated with their host-plant species. The number and abundance of insect herbivores on their host plants along an altitudinal gradient were correlated with host-plant species richness, suggesting that the number of host-plant species will determine the number of insect herbivores at any elevation.;Host-plant characteristics such as plant species richness and the availability of resources offered by host-plants were the major factors affecting the number of insect herbivore species both locally and across latitudes. We suggest that the latitudinal gradient hypothesis that predicts that the tropics is richer in insect herbivore species can only be adequately tested if insect communities on individual host plants are studied to rule out the effect of the great diversity of plants in the tropics. Our findings suggest that on a per host-plant species the tropics might be as diverse as the temperate if entire insect communities are looked at.