On the Trophic Relations of Insects: A Food-Web Approach

Trophic relations of 95 insect-dominated food webs from seven habitat categories are examined and discussed in relation to current ecological theory. We identify 13 salient patterns, of which fewer than half have been reported previously from other food-web studies. (1) Analysis of trophic links in the 61 community-type webs corroborate the well-known trend that trophic diversity of insects has increased over evolutionary time. (2) Numbers of saprophages and animal parasites follow a continuum ranging from grazer-based systems (plant-herbivore, gall) and parasitoid-poor systems (aquatic, carrion, phytotelmata, wood, dung) to habitats that are detritus-based (phytotelmata, rotting wood, carrion, dung) and parasitoid-rich (gall), respectively (3) Among trophic subsets of species that determine food-web shape, only proportions of bottom species in food chains are similar to values published previously. (4) Aquatic-terrestrial comparisons reveal that aquatic webs (including phytotelmata habitats) support a richer number of trophic modes, higher proportions of bottom species, more interactions per species, and a smaller proportion of specialist consumers than the latter. (5) Cannibalistic links are generally absent in most insect webs and uncommonly recorded in the remainder of these webs, a result consistent with other food-web studies. (6) Global patterns of connectance and link numbers per species in insect-rich webs are similar to other reports even in webs with fewer investigator-defined species. This finding suggests that web participants interact with each other in broadly similar ways and in small numbers (2 2 here vs. 1 99 elsewhere). (7) Global ratios of resources (prey) to consumers (predators) are smaller (1:2) than in other webs (7:8), which suggests that for every consumer in insect-rich systems roughly one less resource is included, on the average, in its diet (8) An earlier prediction of higher prey-to-predator ratios in sink webs than in community-type webs is generally supported here. (9) Monophagy is wide-spread across both insect consumers and habitats; conversely, most insect species appear to have no more than three local consumers. Aquatic webs are exceptions in the former case. (10) Most insect families (76%) are associated with only one habitat category; conversely, five families (e.g., ants, flesh flies) are habitat generalists, occurring in four habitat types (11) Tropical webs do not, on the average, have longer food chains than those in temperate or desert climates. (12) Upper limits in chain length vary from three in carrion systems to as many as seven in gall webs: this latter value exceeds suggested upper limits for insect feeding reported elsewhere. (13) Nearly half of all the insect taxa (families) have members that are top predators in certain habitats even after vertebrates are factored in. These latter two features underscore the important roles insects may play in natural communities, roles that reach beyond occupying merely the critical middle links in ecosystems or the occasional position of top predator in unusual places such as caves.