Published Food Webs Research Articles

VARIATION OF TROPHIC FRACTIONS AND CONNECTANCE IN FOOD WEBS

We present and illustrate methods of analyzing food web structure based on generalized linear models. Applied to the published food webs of Schoenly et al., these methods reveal strong associations of trophic fractions and connectance with the number of species in the web. Furthermore, the nature of these relationships depends on the habitat type of the web. Measures of goodness-of-fit of the statistical models reveal pronounced overdispersion, i.e., web-to-web variation in trophic fractions and connectance that greatly exceeds the variation attributable to the number of species and habitat type of the web. These results suggest it may be difficult to identify laws governing food web properties over a wide range of organisms and habitats.

Marine and continental food webs: three paradoxes?

Carbon stocks and flows give a picture of marine and continental biotas different from that based on food webs. Measured per unit of volume or per unit of surface area, biomass is thousands to hundreds of thousands of times more dilute in the oceans than on the continents. The number of described species is lower for the oceans than for the continents. One might expect that each species of organism would therefore feed on or be consumed by fewer other species in the oceans than on the continents. Yet in reported food webs, the average oceanic species interacts trophically with more other species than the average terrestrial or aquatic species. Carbon turnover times imply that the mean adult body length of oceanic organisms is 240 to 730 times shorter than that of continental organisms. By contrast, in reported food webs, marine animal predators are larger than continental animal predators, and marine animal prey are larger than continental animal prey, by as much as one to two orders of magnitude. Estimates of net primary productivity (NPP) per unit of surface area or per unit of occupied volume indicate that the oceans are several to hundreds of times less productive than the continents, on average. If NPP limited mean chain length in food webs, oceanic food chains should be shorter than continental chains. Yet average chain lengths reported in published food webs are longer in oceans than on land or in fresh water. In reconciling these unexpected contrasts, the challenge is to determine which (if any) of the many plausible explanations is or are correct.

Temporal Variation in Food Web Structure: 16 Empirical Cases

Analysis of food web structure and temporal dynamics is essential to understanding energy flow and population dynamics of species, and may contribute to conservation, wildlife management, and disease and pest control. This report synthesizes all the observational studies of food web dynamics to which we have access. Most published food webs are static and cumulative: they depict information gathered over many occasions. A web observed over a single, relatively short time period is time specific. Here we analyze the relation between cumulative and time—specific versions of webs in 16 published cases. Fourteen of the 16 webs are from detritus—based habitats that harbor large fractions of arthropod species: carcasses, tree fluxes, felled logs, treeholes, dung pads, and an acidic pond. The other two webs describe soybean fields and the arctic tundra. These webs are presented here in a consistent format and are analyzed in four ways. First, we quantified temporal trends and levels of variation in nine web properties: the percentages of species in the web that are top predators (%T), intermediate species (%I), and basal species (%B); the ratio of number of prey species to number of predator species (P); the mean chain length (µ); the product of species richness and connectance (S × C); and the numbers of total species, newly arriving species, and local extinctions. In most webs %I and %T fluctuated widely; the latter generally increased in time or remained constant, while the former correspondingly decreased or remained constant. Since the number of basal species usually varied little, changes in %B were inversely associated with changes in species richness over successional and seasonal time scales. Predictable changes in P, µ and S x C accompanied the changes in %B, %I, and %T. The numbers of total species, new arrivals, and local extinctions displayed no consistent increasing or decreasing trends. Second, we compared cumulative and time—specific webs from the same habitat to determine which properties, if any, of time—specific webs might be predicted from cumulative webs. In cumulative webs, P, µ, and %T came closest to the median of the values from time—specific webs, followed by %I, S x C, and %B Cumulative webs, which usually appear in general ecology textbooks, overestimate S x C and underestimate %B relative to time—specific versions. In five studies cumulative webs were completed when the last or next—to—last samples were taken; additional sampling in these cases would probably have uncovered more species. Third, opportunistic species were removed from four time—specific webs to determine how these species influenced web structure. Removing one top—feeding opportunistic species from each web caused a dramatic rise in %T, small reductions in %I, S × C, µ, and P, and a negligible rise in %B. A single opportunistic species, even though it makes only rare and brief visits to a habitat, can dramatically reshape web structure. Fourth, properties of cumulative and time—specific versions of the 16 food webs were compared to properties of cumulative webs in two published web catalogs. The cumulative versions of the 16 webs grossly resemble the cumulative webs in both prior catalogs, but the median S × C is greater and the median %B is lower in the 16 cumulative webs than in either prior catalog. Even for these two statistics, the median value for the 16 cumulative webs falls well within the range of variation of both prior catalogs. The time—specific webs in the 16 cases differ from those of the two prior catalogs somewhat more than do the cumulative webs. Comparisons between time—specific and cumulative versions of a web, one system at a time, are more sensitive than rough comparisons between collections of webs because the methods used to define species and links are (usually) consistent within a study.

Food Web Patterns: Trivial Flaws or the Basis of an Active Research Program?

Paine (1988) has written a persuasive evaluation of food web research. While his criticisms may be muted and focus on his work, they cast doubt on all papers analyzing previously published food web data, including many of our papers. The criticisms are telling because they are thoughtful and because Paine has written for so long and so effectively about food webs. Do Paine's comments indicate serious flaws in our studies of food webs? We feel the answer is yes: that analyses of previously published food webs have many limitations. However, neither of us has published a primary paper analyzing existing food web data since Pimm (1980a). What we have attempted to do since then is to address many of the issues Paine has raised. What we shall attempt here is, first, the placement of studies of existing food web data into a historical perspective. Then we shall review subsequent work, using both data collected expressly to test food web theories and experiments. Where we disagree with Paine is in our assessment of the health of food web research. Like the argument over whether a glass is half empty or half full, we agree with Paine's observations, but draw from them much more optimistic conclusions.

Static and dynamic explanations for patterns in food webs

Ten interesting patterns can be discerned in compendia of published food webs, and are not all easily dismissed as artefacts of poor data. A variety of theoretical explanations have been put forward to explain one, or more, of these patterns; for example, five reasons have been advanced for why food chains are short. Two bodies of theory each satisfactorily account for a majority (but not all) of the patterns, but draw on very different assumptions. One group of theoretical explanations is based on dynamic interactions between species, modelled by Lotka-Volterra equations. The other takes a static view of food web assembly, and assumes a trophic cascade, or hierarchy of feeding links, in which species body sizes appear to play a crucial role. On present evidence, it is not possible to distinguish between the relative contributions of either of these possibilities in the creation of structure in real food webs.

Invertebrate predator-prey body size relationships: an explanation for upper triangular food webs and patterns in food web structure?

It has been suggested by Cohen and Newman (1985) that many of the patterns in published food webs can be derived from a stochastic model in which the species are arranged in a trophic hierarchy (the 'cascade model'). We suggest that, if predators are larger than their prey, a trophic hierarchy can be generated on the basis of body size Empirical evidence from the literature shows that there is a positive relationship between predator and prey size for a range of invertebrates and that predators are usually larger than their prey. Using experimental data on an aquatic food web we show that body size can lead to the type of trophic hierarchy used in the cascade model, suggesting that many food web patterns may be a product of body size. This conclusion is discussed with respect to the limitations of the food web data and the relationship between 'static' and 'dynamic' models of web structure.

Are Food Webs Divided into Compartments?

SUMMARY compartmented. (3) Alternative food web models are briefly described. These incorporate biologically more realistic assumptions, and do not neccessarily predict that food webs are more likely to be stable if they are divided into blocks. (4) Compartments exist in food webs if the interactions within the web are grouped into subsystems: that is, if species interact strongly only with species in their own sub- systems, and interact little, if at all, with species outside it. (5) Drawing on a number of alternative approaches, we test the null hypothesis that real food webs are not significantly more compartmented than chance alone dictates. (6) Analyses of published food webs show that subsystems can only be detected where the webs span major habitat divisions, for example a forest and a prairie, or adjacent freshwater and terrestrial habitats. These compartments are imposed by the natural histories of the component species. There are no grounds for believing that dynamical constraints, i.e. a requirement for persistent natural food webs to be stable, play any part in imposing compartments.