Species Interactions

Communities can be characterized in terms of the relationships among species, most commonly trophic (feeding) interactions (i.e., food webs). Clearly, the most complete description of the community would include all possible interactions (including indirect interactions) among the total number of species (e.g., Polis 1991a). In practice, this is difficult to accomplish, even in relatively species-poor communities (Camilo and Willig 1995, Polis 1991a, 1991b, Reagan et al. 1996) because of the largely unmanageable number of arthropod species (Table 9.1) and lack of complete information on their interactions. More commonly, research focuses on subsets or simplified representations of the community.

The simplest approach to community description emphasizes interactions between only a few species (e.g., plant-herbivore or predator-prey interactions). In particular, many studies have addressed the relatively distinct assemblages of arthropods based on individual plant species (e.g., Richerson and Boldt 1995, Schowalter and Ganio 1998) or soil/litter resources (e.g., J. Moore and Hunt 1988, Seastedt et al. 1989). This approach maximizes description of interactions among a manageable number of relatively resource-specific herbivores or detritivores and their associated predators and parasites. Detailed descriptions at this level have been useful for identifying and comparing factors affecting these trophic interactions (e.g., chemical defenses; see Chapters 3 and 8), for evaluating the co-evolutionary patterns of speciation between insects and their hosts (e.g., Becerra 1997), and for comparing trophic interactions among community types (e.g., comparing phenological responses of insect herbivores to leaf emergence in tropical and temperate forests; Coley and Aide 1991). However, this approach emphasizes relatively linear trophic relationships (i.e., food chains) and does not address linkages among members of different component communities.

Broader subcommunities can be identified. For example, Hunt et al. (1987) described the trophic interactions among arthropod and microbial species composing the litter subcommunity of a grassland ecosystem. J. Moore and Hunt (1988) subsequently noted that relatively discrete component communities supported by particular resource bases (bacteria, fungi, or plant roots) could be distinguished within this broader subcommunity (Table 9.2). Similarly, individual plant species represent resource bases for relatively discrete component communities of associated arthropods and other organisms in the above-ground subcommunity (Curry 1994). Resource-based component communities are


The proportion of energy and nitrogen derived from the bacteria, fungal, and root (including mycorrhizal fungi) resource channels by different faunal groups in the North American shortgrass steppe.

Resource Channel

Faunal Group Bacteria Fungi Roots

Protozoa Flagellates Amoebae Ciliates

Nematodes Bacteriovores Fungivores Root-feeders Omnivores Predators


Mycophagous Collembola Mycophagous oribatid mites Mycophagous prostigmatid mites Nematophagous mites Predaceous mites

100 0 0

100 0 0

100 0 0

100 0 0

0 90 10

0 0 100

100 0 0

69 3 28

0 90 10

0 90 10

0 90 10

67 4 30

40 39 2

From J. Moore and Hunt (1988) by permission from Nature, © 1988 Macmillan Magazines, Ltd.

linked to each other by generalist herbivores and predators. Similarly, the canopy and soil/litter subcommunities are linked by species that feed above ground but pupate in the soil or feed on litter resources but disperse and bask on foliage and by predators and detritivores that move among substrates in search of resources.

The most inclusive approach to community description is represented by interaction webs, in which all species are connected by arrows indicating interactions. Relatively few communities are composed of sufficiently few species to depict all interactions conveniently. Hot springs and other communities subject to extreme abiotic conditions usually are composed of a few tolerant algal and invertebrate species (N.C. Collins et al. 1976). Communities composed of relatively few invertebrate and vertebrate species characterize some aquatic ecosystems (e.g., vernal pools, riffles, etc.). However, even the desert communities described by Polis (1991a) were composed of >103 arthropod species, most of which had not been studied sufficiently to provide complete information on interactions. A number of studies have addressed trophic interactions (i.e., food webs), although even trophic interactions are poorly known for many species, especially insects.

A number of techniques have been used to identify trophic relationships. Early studies of food web structure tracked radioisotopes through trophic exchanges (e.g., Crossley and Howden 1961). Stable isotopes or other tracers also can be tracked through feeding exchange (e.g., Christenson et al. 2002). Furthermore, animal tissues reflect the stable isotope ratios of their diet, with slight enrichment of 15N with increasing trophic level (Bluthgen et al. 2003, Ponsard and Arditi 2000, Scheu and Falca 2000, Tayasu et al. 1997). However, interpretation of trophic interactions depends on the isotopic homogeneity of the diet (Gannes et al. 1997). Selective feeding on particular substrates can affect 13C enrichment in animals (Santruckova et al. 2000). Adams and Sterner (2000) reported that 15N enrichment was linearly related to dietary C:N ratio, which could vary sufficiently to indicate as much as a 2-trophic level separation, potentially leading to misiden-tification of trophic level for particular species.

Advances in molecular techniques have provided new tools for identifying interactions among species in communities. Enzyme-linked immunosorbent assay (ELISA) techniques involve development of antibodies against enzymes from potential food sources. These antibodies can be used to precipitate enzymes in gut samples containing the target food source. Irby and Apperson (1988) and Savage et al. (1993) used ELISA to identify associations between various mosquito species and their particular amphibian, reptile, bird, and mammal hosts. Agusti et al. (1999a) demonstrated the utility of this technique for detecting prey, Helicoverpa armigera, in heteropteran, Dicyphus tamaninii, Macrolophus caligi-nosus, and Orius majusculus, predator gut contents. More recently, polymerase chain reaction (PCR) and DNA amplification techniques have been used to illuminate feeding relationships (Suh et al. 2003). Broderick et al. (2004) used this methodology to describe the microbial community in gypsy moth, Lymantria dispar, midgut and to demonstrate that bacterial composition was influenced by the plant species composition of the diet. Agusti et al. (1999b), Y. Chen et al. (2000), Hoogendoorn and Heimpel (2001), and Zaidi et al. (1999) demonstrated that PCR and DNA amplification can be used to identify prey species in gut contents for 12-28 hours after predator feeding. Although these techniques can help identify feeding relationships, developing the sequence library to distinguish all potential prey in the field presents a challenge.

Several properties have appeared to characterize food webs (see Briand and Cohen 1984, Cohen and Palka 1990, Cohen et al. 1990, Martinez 1992, May 1983, Pimm 1980,1982, Pimm and Kitching 1987, Pimm and Lawton 1977,1980, Pimm and Rice 1987, Pimm et al. 1991, Polis 1991b, Reagan et al. 1996). However, food web analysis usually has been based on combination of all insects (often all arthropods) into a single category, in contrast to resolution at the individual species level for plants and vertebrates. Polis (1991b) and Reagan et al. (1996) increased the resolution of arthropod diversity to individual "kinds," based on taxonomy and similar phylogeny or trophic relationships, for evaluation of food web structure in desert and tropical rainforest communities, respectively. They found that the structure of their food webs differed from that of food webs in which arthropods were combined. Goldwasser and Roughgarden (1997) analyzed the effect of taxonomic resolution on food web structure and found that food web properties reflected the degree of taxonomic resolution. The following properties of food webs, based on analyses with insects or arthropods as a single cat egory, are evaluated with respect to challenges based on greater resolution of arthropod diversity.

Was this article helpful?

0 0

Post a comment