Finally, ratios of species and links from basal to intermediate to top trophic levels (where basal species are prey only, intermediate species are prey and predators, and top predators
273 I. APProAChES To DESCrIBIng CoMMUnITIES
have no predators) are expected to be constant (Briand and Cohen 1984). This implies a large proportion of top predators, which are expected to comprise 29% of all species in a given community, and prey to predator ratios should be < 1.0 (Briand and Cohen 1984).
As shown for the properties discussed above, this property reflects poor resolution of arthropod diversity. Top predators appear to be common because they are easily distin- guished vertebrate species, whereas poor taxonomic resolution at basal and intermediate levels underrepresents their diversity. Reagan et al. (1996) reported that in a rain forest food web that distinguished “kinds” of arthropods, representation of basal and intermedi- ate species was 30% and 70% of all species, respectively, and the proportion of top preda- tors was < 1%. Polis (1991b) also reported that top predators were rare or absent in desert communities. Both Polis (1991b) and Reagan et al. (1996) reported that ratios of prey spe- cies to predator species are much greater than 1.0 when the true diversity of lower trophic levels is represented.
Although the properties of food webs identified by early theorists may be flawed, to the extent that arthropod diversity was not resolved adequately, they represent hypotheses that stimulated considerable research into community organization. Future advances in food web theory will reflect efforts to address arthropods at the same level of taxonomic resolution as for other taxa.
c. functional organization
A third approach to community description is based on the guild, or functional group, concept (Cummins 1973, C. Hawkins and MacMahon 1989, Kửrner 1993, Root 1967, Sim- berloff and Dayan 1991). The guild concept was originally proposed by Root (1967), who defined a guild as a group of species, regardless of taxonomic affiliation, that exploit the same class of environmental resources in a similar way. This term has been useful for studying potentially co-evolved species that compete for, and partition use of, a com- mon resource. The largely equivalent term, functional group, was proposed by Cummins (1973) to refer to a group of species having a similar ecological function. Insects, as well as other organisms, have been combined into guilds or functional groups based on the similarity of their response to environmental conditions (e.g., Coulson et al. 1986, Fielding and Brusven 1993, Grime 1977, Root 1973) or of effects on resources or ecosystem pro- cesses (e.g., Romoser and Stoffolano 1998, Schowalter et al. 1981c, Siepel and de Ruiter- Dijkman 1993). This method of grouping is one basis for pooling “kinds” of organisms, as discussed above.
Pooling species in this way has been attractive for a number of reasons (Root 1967, Simberloff and Dayan 1991). First, it reflects the compartmentalization of natural com- munities (see above) and focuses attention on sympatric species that share an ecologi- cal relationship, e.g., those competing for a resource or affecting a particular ecological process, regardless of taxonomic relationship. Second, it helps resolve the multiple usage of the term “niche” to refer both to the functional role of a species and the set of condi- tions that determines its presence in the community. Use of guild or functional group to refer to species’ ecological role(s) permits limitation of the term niche to refer to the conditions that determine species presence. Third, this concept facilitates comparative studies of communities which may share no taxa but do share functional groupings, e.g., herbivores, pollinators, detritivores, etc. Guild or functional groupings permit focus on a particular group, with specific functional relationships, among community types. Hence, researchers avoid the necessity of cataloging and studying all species represented in the
community, a nearly impossible task, before comparison is possible. Functional groupings are particularly useful for simplifying ecosystem models to emphasize the effects of func- tional groups with particular patterns of carbon and nutrient use on fluxes of energy and matter. Nevertheless, this method for describing communities has been used more widely among aquatic ecologists than among terrestrial ecologists.
The designation of functional groupings is largely a matter of convenience and depends on research objectives (e.g., C. Hawkins and MacMahon 1989, Kửrner 1993, Simberloff and Dayan 1991). For example, defining “the same class of resources” or “in a similar manner” is ambiguous. Each species represents a unique combination of abilities to re- spond to environmental conditions and to affect ecosystem processes, i.e., species within functional groups are similar only on the basis of the particular criteria used to distinguish the groups. Characterization of functional groups based on effect on primary production, effect on carbon flux, or effect on biogeochemical cycling would involve different combi- nations of species.
Insects are particularly difficult to categorize because functional roles can change seasonally (e.g., wasps switching between predation and pollination) or during matu- ration (e.g., sedentary herbivorous larvae becoming mobile pollinating adults, aquatic larvae becoming terrestrial adults, etc.), and many species are too poorly known to have functional roles assigned to them. Nearly all Lepidoptera can be assigned to a plant- feeding functional group, but various species would be assigned to different functional groups on the basis of the plant part(s) affected (e.g., foliage, shoots, or roots). Clearly, functional groups can be subdivided to represent a diversity of resource exploitation strategies or subtle differences in ecological effects. For example, the plant-feeding
“functional group” could be divided into subgroups that selectively feed on ruderal, competitive, or stress-adapted plant hosts (Fielding and Brusuen 1995). The foliage- feeder guild can be divided into subgroups that fragment foliage, mine foliage, or suck cellular fluids, feed on different plant species, etc., with each subgroup affecting energy and matter fluxes in a different manner. Luh and Croft (1999) developed a computer algorithm to classify predaceous phytoseiid mite species into functional groups (spe- cialist vs. generalist predators). The computer-generated classification confirmed the importance of the combination of life history traits that had been used previously to distinguish functional groups.
The species included in a particular functional group should not be considered redun- dant (Beare et al. 1995, Lawton and Brown 1993), but rather complementary, in terms of ensuring ecological functions. Schowalter et al. (1999) reported that each functional group that was defined on the basis of feeding type included species that responded positively, negatively or non-linearly to moisture availability. Species replacement within functional groups maintained functional organization over an experimental moisture gradient, but changes in species would result in differences in pathways and rates of energy and matter fluxes (see Chapter 4).
Changes in the relative abundance or biomass of functional groups can signal changes in the rate and direction of ecological processes. For example, changes in the relative proportions of filter-feeder vs. shredder functional groups in aquatic ecosystems affect the ways in which detrital resources are processed within the stream community and their contribution to downstream communities. Similarly, changes in the relative proportions of folivores vs. sap-suckers affect the flux of nutrients as solid materials vs. liquid (e.g., honeydew) and their effect on the detrital community (e.g., Schowalter and Lowman 1999, Stadler and Müller 1996, Stadler et al. 1998).
275 II. PATTErnS of CoMMUnITy STrUCTUrE
The functional group concept permits a convenient compromise in dealing with diver- sity, i.e., sufficient grouping to simplify taxonomic diversity while retaining an ecologically relevant level of functional diversity. Therefore, the functional group approach has be- come widely used in ecosystem ecology.
II. pAtteRnS of communIty StRuctuRe
A central theme of community ecology has been the identification of patterns in commu- nity structure across environmental gradients in space and time (see also Chapter 10). The diversity of community types at landscape and regional scales has been a largely neglected aspect of biodiversity, but is important to the maintenance of regional species pools and metapopulation dynamics for many species. In addition, the mosaic of community types on a landscape may confer conditional stability to the broader ecosystem, in terms of rela- tively consistent proportions of community types over time (see Chapters 10 and 15).
Identification of patterns in community organization has become increasingly impor- tant to population and ecosystem management goals. Introduction of exotic insects to combat noxious pests (weeds or other insects) requires that attention be paid to the ability of the biocontrol agent to establish itself within the community and to its potential effects on non-target components of that community. Efforts to conserve or restore threatened species require consideration and maintenance of the underlying community organiza- tion.
Depending on the descriptive approach taken (see above), patterns have been sought in terms of species diversity, food web structure, or guild or functional group composi- tion. Unfortunately, comparison of data among communities has been hampered by the different approaches used to describe them, compounded by the variety of sampling tech- niques, with their distinct biases, that have been used to collect community data. For ex- ample, sweep netting, light trapping, interception trapping, pitfall trapping, soil coring, canopy fumigation and branch bagging are among the techniques commonly used to sam- ple terrestrial arthropods (Leather 2005). These techniques differ in their representation of nocturnal vs. diurnal flying insects, arboreal vs. soil/litter species, and sessile vs. mobile species, etc. (e.g., Blanton 1990, Leather 2005, Majer and Recher 1988, Southwood 1978).
Variation in the mesh size of sampling nets affects the representation of aquatic species (Storey and Pinder 1985). Relatively few studies have used the same or similar techniques, to provide comparative data among community types or locations. Some proposed pat- terns have been challenged as subsequent studies provided more directly comparable data or increased the resolution of arthropod taxonomy (e.g., C. Hawkins and MacMahon 1989, Polis 1991b, Reagan et al. 1996). Disturbance history, or stage of post-disturbance recovery, also affects community structure (e.g., Harding et al. 1998, Schowalter et al.
2003, E. Wilson 1969, see Chapter 10). However, the history of disturbance at sampled sites often is unknown, potentially confounding any interpretation of differences in com- munity structure. Nevertheless, apparent patterns that have been identified at a variety of spatial scales may serve as useful hypotheses to guide future studies.
A. global patterns
Communities can be distinguished on a taxonomic basis at a global scale because of the distinct faunas among biogeographic realms (A. Wallace 1876). However, similar commu- nity types on different continents often are dominated by unrelated species with similar
attributes, termed ecological equivalence. For example, grassland communities on every continent should show similar food web structure and functional group organization, reflecting similar environmental conditions, regardless of taxonomic representation. A number of studies have indicated global patterns in community structure that are related to latitudinal gradients in temperature and moisture, and to the ecological history of adap- tive radiation of particular taxa.
Latitudinal gradients in temperature and precipitation establish a global template of habitat suitability, as discussed in Chapters 2 and 7. Equatorial areas, characterized by high sun angle and generally high precipitation, provide favorable conditions of light, temperature and moisture, although seasonal patterns of precipitation in some tropical areas create periods of adverse conditions for many organisms. The strongly seasonal cli- mate of the temperate zones requires specific adaptations for survival during unfavorable cold periods, thereby limiting species diversity. The harsh conditions of temperate deserts and high latitude zones generally restrict the number of species that can be supported or that can adapt to these conditions.
Species richness generally decreases with latitude for a wide variety of taxa (R. Dunn et al. 2009, Gaston 2000, Price 1997, J. Stout and Vandermeer 1975, Wiens et al. 2006, Willig and Lyons 1998). Latitudinal gradients are especially pronounced for insects, with some studies suggesting that the tropics support several million undescribed arthropod species (Erwin 1995, May 1988, E. Wilson 1992), depending on scale-dependent estimates of specialization of herbivorous groups among plant species (Gering et al. 2007). Latitudi- nal trends may not be reflected by all taxa (e.g., aphids, Dixon 1985) or component com- munities (Vinson and Hawkins 1998). Although L. Dyer et al. (2007) reported that the larval diets of tropical Lepidoptera were more specialized than those of temperate forest caterpillars, contributing to higher diversity of this group in tropical forests, Novotný et al.
(2006) found similar levels of specialization between tropical and temperate Lepidoptera and concluded that the greater diversity of this group in the tropics reflected the greater di- versity of plants. Lewsinsohn and Roslin (2008) conducted a meta-analysis of studies that compared temperate and tropical herbivore diversity and found that correlation between plant and herbivore diversity explained 60% of the variation in insect species richness.
They concluded that higher insect diversity in the tropics reflects the greater diversity of host plants. Vinson and Hawkins (1998) reviewed the literature for stream communities and concluded that species richness is highly variable, and no strong latitudinal trends are apparent. Furthermore, Willig and Lyons (1998) showed that latitudinal gradients can result from chance. Nevertheless, a number of hypotheses have been proposed to explain latitudinal gradients in species richness.
Terborgh (1973) showed that the apparent trend in species richness with latitude can reflect increasing land area toward the equator. He noted that climate is relatively con- stant across a wide belt between latitudes 20° N and S but shows a distinct gradient above those latitudes. Combining climate and surface area gradients yielded a latitudi- nal gradient in habitat area available within each climate class, with a preponderance of global surface area in tropical habitats. These data suggest that the gradients in species richness reflect the habitat area that is available for within-habitat speciation (see discus- sion below).
Latitudinal gradients in species richness also may reflect greater primary produc- tivity in the tropics (Rosenzweig and Abramsky 1993, Tilman and Pacala 1993, Waide et al. 1999; see below). Hutchinson (1959) proposed that animal diversity is related to the energy available in ecosystem primary production. D. Currie (1991) subsequently
277 II. PATTErnS of CoMMUnITy STrUCTUrE
demonstrated that North American patterns of plant and vertebrate diversity were related to environmentally-available energy. More recently, A. Allen et al. (2002) and J. Brown et al. (2004) proposed a metabolic theory of ecology that explained latitudinal gradients in diversity as a result of relationships between temperature, body size and metabolic rate that determine the maximum population sizes attainable, given the energy and nutrients available in the ecosystem. All other factors being equal, available energy will support more small organisms than larger ones, and warm environments will support more organ- isms than cold environments. Algar et al. (2007) and B. Hawkins et al. (2007) tested the prediction of metabolic theory that the natural logarithm of species richness (of a variety of data sets for trees, blister beetles, tiger beetles, butterflies, amphibians and reptiles) is a linear function of temperature. All taxa tested showed a curvilinear relationship to tem- perature, rather than the predicted linear relationship, indicating that energy availability alone is not a sufficient explanation for latitudinal gradients in species richness.
Finally, evolutionary time may explain latitudinal gradients for some taxa. Wiens et al.
(2006) found that tree frog diversity was strongly correlated with time since colonization of a region, but not with latitude per se. However, since these frogs originated in tropical South America and spread to temperate regions relatively recently, there has been more time for speciation in the tropics than in temperate regions. R. Dunn et al. (2009) evalu- ated the asymmetry in ant species richness between northern and southern hemispheres (Fig. 9.8) and concluded that the greater climate change since the Eocene in the northern hemisphere had resulted in more extinctions and reduced species richness relative to the southern hemisphere.
fIG. 9.8 Latitudinal trends in a) mean annual precipitation and temperature, b) temperature range, c) local species richness of ants, and d) regional richness of ant genera. Negative latitudes are for the southern hemisphere. Generic richness is derived from lists of species and genera from countries and smaller political regions and presented for comparison. From R. Dunn et al. (2009) with permission from John Wiley & Sons.
Superimposed on the latitudinal gradients are the relatively distinct biogeographic realms identified by Wallace (1876). These realms reflect the history of continental break- up, with southern floras and faunas being largely distinct from northern floras and faunas (see Chapter 7). However, the southern continents show a varied history of reconnection with the northern continents that has resulted in invasion, primarily by northern species.
The proximity of North America and Eurasia has facilitated movement of species between these land masses, leading to development of a Holarctic species component, especially within the arctic and boreal biomes. Whereas many genera, and even some species, occur throughout the Holarctic realm, the flora and fauna of Australia have remained relatively distinct as a result of continued isolation.
Species richness also may be related to geological time. E. Wilson (1969) suggested that coevolution should improve the efficiency of total resource exploitation and lead to further increase in coexisting species over time. In other words, a habitat or resource that has persisted for a longer period of time would acquire more species than a more recently derived habitat or resource. Birks (1980) found that the residence time of tree species in Britain was strongly correlated with the diversity of associated insect species. Tree species that had a longer history of occurrence in Britain hosted a larger number of species than did tree species with shorter residence times. Again, because residence time is correlated with area of occurrence (habitat area), the effects of these two factors cannot be distin- guished easily (Price 1997, see below).