IN THIS ESSAY, I explore the interplay of two of the most important conceptual frameworks in community ecology—island biogeography and food web ecology (figure 6.1). My goal is to lay out steps toward their synthesis—with the ultimate objective being to stimulate the fuller development of what we might call “trophic island biogeography.” I start by sketching key insights at the heart of each paradigm, and point out ways they were already related (albeit for the most part implicitly, or sketchily) in the famed 1967 monograph by Robert MacArthur and and E.O. Wilson, The Theory of Island Biogeography. I then use simple modifications of the canonical model of colonization and extinction on an island presented in that monograph to consider questions such as top-down effects of predators on the species-area relationships of prey, and bottom-up effects of prey on food chain length and predator species-area relationships. Next, I consider a number of interesting complications which arise when bottom-up and top-down effects occur simultaneously, and in particular emphasize the potential importance of island area as a moderator of intrinsically unstable trophic interactions. To round off the paper, I briefly discuss a number of areas of active inquiry in community ecology that will be important for a fully developed trophic island biogeography, and then conclude by reflecting on how trophic interactions in fragmented landscapes in some ways resemble, and in other ways radically differ from, those in isolated oceanic islands.
A central question posed in the opening chapters of MacArthur and Wilson’s monograph was: What factors govern variation in the number of species found on islands, as a function of island area and distance from continental source pools? Their answer, the “equilibrium theory,” as portrayed in the model on the left side of figure 6.1, focused on colonization and extinction (Schoener, this volume). This theory embodies two crucial insights that go well beyond island biogeography. First, communities at all spatial scales are dynamic. Viewed over the grand span of earth history, local communities (“local” denotes the spatial scale where individuals potentially interact, for instance by competition) assemble via colonization from external sources (augmented by occasional in situ speciation) and are depleted by extinctions (Graham et al. 1996). MacArthur and Wilson (1967) argued that a similar dynamism occurs even over shorter time scales. Subsequent literature has often focused on the celebrated, and indeed controversial, hypothesis by MacArthur and Wilson (1967) that communities are at or near equilibrium, so the number of species remains roughly constant in the face of continual turnover in composition. But the deeper message that communities are dynamic does not depend on the assumption of equilibrium. Long-term censuses on both islands and continents often reveal extinctions and recolonizations over short time scales (Williamson 1981, Schoener, this volume). Extinctions can be deterministic—due to disturbance, succession, interspecific interactions, or shifts in climate—or simply the stochastic winking in and out of rare community members. Unraveling the mechanics of community assembly and disassembly mandates a close focus on colonization and extinction, which are thus essential for understanding all communities, whether or not they reach equilibrium.
Figure 6.1. Two of the most important conceptual paradigms in community ecology—island biogeography and food web ecology. The text explores the question of how these paradigms are related. The right panel is a simplified food web of the Serengeti ecosystem (Holt et al. 2008).
Second, space matters. Most ecology textbooks show how the curves in figure 6.1 (left) vary with island area and distance. Colonization should reflect an island’s distance from sources of colonists and the ability of species to traverse dispersal barriers. This insight was not new to MacArthur and Wilson (1967), but they did elegantly articulate the logic of demographic influences on colonization, as well as stepping stones and other determinants of colonization rates, using quantitative approaches that set a high standard for subsequent ecological theory. Within continents, spillover of species among habitats can boost local diversity; the absence of such spillover may lead to lower diversity on islands than on comparably sized mainland areas (MacArthur and Wilson 1967, pp. 16 and 115; Holt 1993, Rosenzweig 1995). Second, the area of an island influences extinction rates. This is partly simply because larger areas harbor more individuals—a “pure area” effect—and partly because larger areas contain more distinct habitats, which can buffer extinctions and sustain specialized niches—an “environmental diversity” effect. The pure area effect can reflect two processes. If a species’ density is constant, its absolute numbers will scale with island area; smaller populations face larger dangers of extinction from demographic risk and other factors (Schoener, this volume). Moreover, if colonization is analogous to random sampling from a continental fauna, as small islands have few total individuals they in effect are a small sample and so could contain few species by chance alone (Schoener, personal communication). The emphasis on space was a fundamental insight provided by the theory of island biogeography that still resonates throughout both basic ecology and applied arenas such as conservation biology (Laurance, this volume).
The second canonical paradigm in figure 6.1—the food web—goes back at least to Charles Elton, with an intellectual lineage running through Lindeman, Hutchinson, Cohen, Pimm, and many others up to the present. The powerful metaphor of communities as interactive webs has stimulated an enormous amount of creative work. For instance, one can view webs as abstract networks of connections and focus on efficient descriptors describing those patterns (e.g., Martinez 1992). Or one can attach dynamical equations to each node (e.g., Yodzis 1998) and explore the implications of web structure for issues such as the relationship between stability and complexity (e.g., McCann 2000, Kondoh 2003), the vulnerability of webs to disturbance, invasion, and the extinction of resident species (e.g., Dunne et al. 2002), and the relative strength of top-down and bottom-up forces.
What is the relationship between these two ecological paradigms? Until recently, very little. Classical studies of food webs paid scant attention to the influence of spatial processes on food web structure and dynamics. The excellent monograph on food webs by Stuart Pimm (1982), for instance, deals with space only with respect to how distinct habitats can lead to food web compartmentalization. Tom Schoener (1989) in an important paper did provide an insightful discussion of how food chain length might be influenced by island size, and his paper helped stimulate some growth in this area (for reviews see Holt and Hoopes 2005, Polis et al. 2004). But until quite recently (Amaresakare 2008), analyses of spatial patterns and processes have overall been a rather minor theme in the food web literature.
Conversely, I think it is fair to say that classic island biogeography theory (and its modern descendant, metacommunity theory [Hubbell 2001, Holyoak et al. 2005]) largely emphasized the “horizontal” structure of communities, such as potential competition between members of a guild or taxon, with little attention given to food webs per se. Yet although MacArthur and Wilson (1967) do not directly discuss food webs, it should be noted that they do state that the extinction curve should be concave because of “interference” among species; interference might well include predation, as well as exploitative and interference competition (the concavity in the extinction curve may also arise because of variation in species-specific rates; see Schoener, this volume). Moreover, they do touch upon trophic interactions in two short, but telling, passages. In chapter 5, “Invasibility and the variable niche,” the section titled “The closed community” comments on how predators influence coexistence. “Each of the conditions for reduction of diversity—competitors too similar, species too rare, predators too rare (or too common)—can prevent invaders from colonizing.” This statement suggests that local food web interactions can govern colonization. In chapter 6, “Evolutionary changes following colonization,” one reads “impoverishment of diversity often leads to lack of effective predators. This is because the K of predators is considerably lower than that of their prey, so they are precariously rare even on large islands.” One way to parse this passage is that trophic structure (and in particular trophic rank) influences extinction. The second sentence in this quotation implies the first, in the sense that, if predators are differentially vulnerable to extinction, then communities with low diversities on islands are particularly likely to lack predators. Sampling effects could also play a role; effective predators may be absent in species-poor assemblages by chance alone. An alternative interpretation of the first sentence is that the impoverishment of prey diversity itself leads to a lack of effective predators. One mechanism leading to this is the increase of predator abundance with prey species richness, permitting predators to more effectively limit any particular prey population. This is apparent competition (Holt 1977, Holt and Lawton 1994), an indirect interaction among alternative prey species arising from a predator’s numerical response to the entire suite of prey in its diet. The basic idea hinted at in chapter 6 of The Theory of Island Biogeography is thus that trophic structure and rank can influence extinction rates.
Hence, food web interactions may govern the two basic processes of island biogeography theory—colonization and extinction. Conversely, local food web structure itself should reflect these same processes. All local food webs are assembled by colonization, and depleted by extinction, both of which are spatially mediated processes. A recognition of the interplay of these two paradigms suggests that the time is ripe for their fusion into a “trophic island biogeography.” As a start toward such a theory, it is useful to take the simplest version of the MacArthur-Wilson equilibrial theory, and ask how a consideration of trophic position influences its predictions for broad categories such as “predators and prey,” or “specialist and generalist predators.” The next sections present several complementary approaches to this theme.
As a simple start, with a food web in hand, by using various protocols (e.g., counting links up from the base, or using stable isotopes; Post and Takimoto 2007), one can assign a trophic rank to each species and then contrast “predators” (a set of high-ranked species) to “prey” (a set of low-ranked species). There could be systematic population-level attributes correlated with trophic rank that directly influence colonizing ability or extinction risk. For instance, predators are often rarer than their prey (Spencer 2000), and thus, ceteris paribus, more likely to go extinct on small islands due to demographic and environmental stochasticity. Figure 6.2 shows how these considerations influence a noninteractive model of island communities. The model is
where Si is the number of species in a given trophic set i (for now, respectively, predator or prey), Ki is the number of species in this set i in the mainland species pool, Ii is the colonization rate per species, and Ei is an extinction rate. (For simplicity, I assume that colonization and extinction rates are linear.) We assume extinction declines with the logarithm of island area A, i.e., dEi/dlog (A) < 0. The equilibrial species richness in trophic set i is
Figure 6.2. The MacArthur-Wilson equilibrial model applied to predators and prey. A. As explained in the text, as a deliberately oversimplified starting point, we assume a non-interactive community in which we have taxonomic or functional grounds to separate “predators” from “prey.” For simplicity, we assume immigration rates are equivalent for these two classes. If predators are typically less dense than prey, this may not affect extinction rates on a large island much, but would make predators much more sensitive than their prey to reduced island size. B. Predators are present on both large and small islands. In the example shown, increased extinction due to predation reduces the effect of island area upon prey species richness (after Rydberg and Chase 2007).
[a = log(A)]. If the strength of the species-area relationship for trophic set i is
(apt for any relationship that is roughly a power law, S = cAz), after a little manipulation we have
(the vertical lines denote absolute value; log here refers to natural log). The numerator measures the sensitivity of extinction rates to island area. Start with islands large enough that all species have low extinction rates. As island size decreases, it may be reasonable to expect extinction rates for predators to increase more sharply than for prey, simply because predators tend to be relatively rare. Due to demographic stochasticity, a decline in a predator from 1000 to 100 individuals should increase extinction risk much more than a proportional decline in its prey from 10,000 to 1000, and so the numerator of (6.4) should be larger for predators. As indicated in figure 6.2A, this leads to the very simple prediction that there should be a stronger species-area relationship for predators than for prey. This prediction is not watertight, for z also depends on the rates in the denominator of (6.4). If extinction rates are high, few species will be present, and z-values will all be low, so there would only be minor, nearly undetectable differences between predators and prey. Equation (3.6) in Schoener (this volume; see also Schoener 1976a) relates the number of species present on an island to their aggregate density. If the total density is independent of species richness (a zero-sum assumption), this equation predicts that, over a given range of island areas, among taxa with comparable colonization and extinction rates and source pool diversities, those taxa with lower aggregate densities will show stronger species-area relationships than do taxa with higher aggregate densities. Consistent with this prediction, Schoener (1976a) notes that. in general, birds with relatively low z have relatively high summed population densities and vice versa; in particular, raptors have relatively high z.
The effect expected for distance is less clear. Predators are often larger than their prey and might behaviorally avoid physical transport processes that could take them across water gaps; this reduces colonization. A low immigration rate in the denominator of (6.3) inflates the impact of area sensitivity on extinction, and so increases z for predators. By a comparable argument, one expects a stronger species-distance relationship for predators than for prey. Some evidence matches this prediction. Shulman and Chase (2007) showed in experimental mesocosms that the ratio of predator to prey species declined with distance from a source pond (figure 6.3). Yet some predators are highly mobile, readily crossing barriers that impede prey. Greater mobility at higher trophic ranks should weaken species-area and species-distance relationships for predators, compared with their prey. The whole issue of how trophic rank influences colonization cries out for more empirical study and mechanistic modeling.
The model of figure 6.2 is a reasonable place to start, but it blatantly ignores the fact that the fates of predator and prey are closely intertwined. Comparable arguments pertain to any grouping of species into sets that differ in colonization and extinction rates (e.g., large-vs. small-body species in the same trophic level; species near the edges of their climatically defined geographical ranges vs. species near their range centers). The next section presents a first step toward incorporating trophic interdependencies.
Sometimes, predators may be distributed largely independently of island/patch area and distance. Humans, for instance, deliberately or inadvertently introduce predators onto islands, or into islandlike habitats (e.g., trout have been introduced into isolated glacial lakes in New Zealand). The distribution of these predators should then be largely independent of prey species. How does such extrinsically determined predation modify prey colonization and extinction dynamics? The incidence and abundance of prey on islands can be strongly influenced by predation. This is particularly dramatic for introduced alien predators (Salo et al. 2007), but also occurs for predators and prey with a shared evolutionary history. Adler and Levins (1994) note that rodent numbers often increase with decreasing island area, and suggest that this reflects predator presence and abundance. An excellent example comes from islands in the Thousand Island Region of the St. Lawrence River, where occupancy and density of the short-tailed shrew (Blarina brevicauda) decline with distance from the mainland, and conversely occupancy and density of the meadow vole (Microtus pennsylvanicus) increase (Lomolino 1984). Blarina disperses poorly across open water and ice; this explains its absence on distant, small islands. Blarina is also a voracious generalist predator, so given that it can colonize, its persistence may largely be independent of the vole. Conversely, when the shrew is present, it can limit or even eliminate Microtus. Thus, the vole exhibits ecological release on islands when freed of Blarina predation (Lomolino 1984). Likewise, Nordstrom and Korpimaki (2004) showed in Fennoscandia that introduced minks are constrained to islands close to sources, and that mink predation in turn leads to a positive relationship between island bird species richness and distance. The presence of predators may act synergistically with disturbance to elevate prey extinction risks (Schoener et al. 2001). Experiments also show that predators can substantially reduce prey colonization success (Schoener and Spiller 1995, Kotiaho and Sulkava 2007).
Figure 6.3. Predator-prey ratios vary systematically with distance from a source. Aquatic mesocosms (plastic tubs) were placed at varying distances from a pond, a source for aquatic insect colonists. In the left column, open and closed columns are respectively predators and prey; the top is raw data, the bottom, rarefied data. The right column is the ratio of predator to prey species There is a strong signal of distance from the pond on the trophic composition of the mesocosms, with predator species richness declining relatively more strongly at large distances (from Shulman and Chase 2007).
Several authors have modified the basic MacArthur-Wilson (1967) model by adding top-down impacts of consumers onto prey extinction and colonization rates. Olff and Ritchie (1998) examined how herbivory influences plant species richness, where the presence of the herbivore is governed by extrinsic factors (e.g., as in livestock husbandry). They used a graphical model comparable to figure 6.2 to illustrate how grazing alters species richness by shifting colonization and extinction curves. For instance, by disturbing soil, herbivores open sites for germination, thus potentially boosting colonization. When grazers selectively attack competitive dominants, they may relax competition and reduce local extinctions (Harper 1969). Conversely, if grazers are unselective and grazing pressure is sufficiently intense, or competitively dominant plants can tolerate grazing better than can competitively inferior species, herbivores can boost extinction rates (Lubchenco 1978). Increases in extinction due to predation are likely common. For instance, Schoener and Spiller (1996) showed experimentally that predatory lizards directly depress spider prey species richness by elevating extinctions.
Ryberg and Chase (2007) recently modified the simple noninteractive model given by equation (6.1) by assuming that predators elevate extinction rates of prey by a constant additive amount, independent of island area. Here, I generalize their approach, allowing both intrinsic extinctions and extinctions from predation to vary with island area, as follows:
The equilibrial species richness is
Ryberg and Chase (2007) predict that, if predators uniformly and additively increase per species extinction rates of prey, islands with predators will have a more shallow species-area relationship than islands without predators. Manipulation of (6.6) shows that
If predators elevate extinction uniformly across all islands, the second term in the numerator is zero, and there is an additional positive term in the denominator. This implies a lower z-value due to predation (figure 6.2B). If predation-driven extinctions increase with island size, the species-area relationship of the prey will be even weaker; decreased extinctions permitted by increasing island size will tend to be canceled out by increased extinctions from predation. Conversely, if extinction rates from predation are magnified on small islands, the effect of island size on species richness may be enhanced.
Equation (6.6) assumes that the most natural way to represent the impact of predation upon prey extinction is via an additive term. This is mathematically convenient, but does not as yet follow from any more microscopic derivation. Alternatively, one could assume that predators alter extinction rates multiplicatively by x, so that the extinction rate of the prey is x(a)Ei(a) (T. Schoener, personal communication). After substitution, and manipulation, we find that
If the impact of predation upon prey extinction is independent of island area, x > 1 implies that predation increases the strength of the species-area relationship in prey.
Further study is required to determine whether (6.7) or (6.7′) provides the most “natural” or parsimonious representation of predation impacts upon prey extinction. But empirically there is support in the literature for the effects of predators on prey z-values going in both directions. Support for the prediction that predation flattens the species-area relationship comes from Ryberg and Chase (2007), who examined distributional patterns in two island-like habitats: orthopteran richness in Ozark glades (open rocky outcrops within a forest matrix), with and without the insectivorous collared lizard Crotaphytus collaris; and man-made ponds, with and without fish as predators on zooplankton. In both cases, for larger patch sizes, islands without predators clearly contained a greater richness of prey species than did islands with predators, and the former also had higher z-values. At low ranges of areas in both study systems, however, contrary to the model predictions (and as noted by Ryberg and Chase), the species-area relationships converged, suggesting minimal or no impact of predation upon prey species richness on small islands, or even possibly a slight positive effect. An area dependence in the impact of predation could reflect several factors. One such factor is that, among islands occupied by predators, their densities may decline sharply with decreasing island size (as shown in Lomolino [1984] for Blarina). For generalist predators like collared lizards and shrews, the reduced prey species richness expected on smaller islands may translate to a lower carrying capacity. If total mortality inflicted by predators on prey scales with predator density, the contribution of predation to extinctions in a focal group of prey species may be less important on smaller islands, because predators, even if present, tend to be rare.
But in other cases the impacts of predators on prey on small islands, compared to on large islands or continents, may be severe. Schoener and Spiller (1999) used removal experiments in the Bahamas to show that lizard predators much more strongly reduce spider density and species richness on small islands than on large islands. Several distinct mechanisms could be at play (and Schoener and Spiller [1999] suggest still others). Resources available for the prey themselves may be limited on small islands. If so, prey cannot tolerate as much predation and still persist, and even if they do persist it may only be at a lower abundance. Reduction to low densities by predation aggravates the risk of stochastic extinctions, just because absolute abundances are low on small islands. Fewer refuges may be available on small islands, making prey more vulnerable to exclusion from persistent generalist predators. Finally, generalist predators may be able to persist on just a few prey species, which permits the predators to drive other prey species extinct. Thus, top-down effects could amplify the species-area relationship in a prey guild.
Now, I reverse the assumptions of the previous section. A food web at the very least describes bottom-up asymmetrical resource dependencies among species. For now we will assume the distribution of predators depends upon that of their prey, and for simplicity (relaxed below) assume also that, by contrast, prey distributions are independent of predation. I start by sketching the classic problem of the determinants of food chain length, focusing on specialist food chains, and then turn to the influence of trophic rank on the strength of the species-area relationship.
Understanding what limits food chain length is a long-standing puzzle in ecology. Ecological communities vary much more in species richness than in food chain length. But why? Traditional explanations are nicely summarized in Pimm (1982) and Post (2002), and these hypotheses have implications for how island size and distance might influence food chain length. For instance, energetic constraints suggest that longer food chains are expected in more productive habitats. Schoener (1989) generalized this observation and provided one way to link space to food web theory by pointing out that the total energy production of an island is productivity (energy/unit time/unit area) times area. He suggested that instead of productivity, per se, the total production contained within an island might govern the food chain length it can support—the “productive space” hypothesis. Schoener described this hypothesis as follows: “maximum food-chain lengths are determined by the amount of productive space required to allow critical component species populations [namely, ones at the top of the food web] to persist with some high probability.” The hypothesis rests on a population-size argument. Consider a continental community with a classic “pyramid of numbers,” so that density declines with increasing trophic rank in a food chain. Absolute population size is of course density times area. If we consider islands which have identical environmental conditions, but differ in area, a null model is that population size (total numbers, not density) for each species will be proportional to area. If there is a critical population size below which extinction is certain, the area at which this threshold will be reached will be larger for species at higher trophic ranks. This implies shorter food chains on smaller islands. Alternatively, assume that we compare these islands with another set of islands, which have a uniformly higher primary productivity. If this increase in production translates into a comparable increase in density at each trophic level, working through the same argument, one predicts that, with higher productivity, there is a lower critical island size below which the top predator dips below its critical abundance, than is observed on islands with lower productivity.
The productive space hypothesis is appealing, and is surely part of the story, but the jury is still out on the degree to which it entirely explains variation in food chain length among communities. Production does seem to be related to the decline in species diversity with increasing trophic rank (Rosenzweig 1995, Havens 1992, Duffy 2002), but the evidence to date suggests that it does not fully account for area effects on food chain length (Post 2002). One complication is that increased primary production may not translate neatly into proportional increases in abundance at each trophic level. For instance, shifts in species composition at lower trophic levels toward inedible species can lower the amount of production passing through to higher trophic levels. Satiation or interference competition may constrain predator numerical responses to increased food supplies. Increased production can destabilize predator-prey interactions; excursions to low densities may then aggravate extinction risks (the classic “paradox of enrichment”), particularly on small islands. Finally, spatial subsidies on small islands can elevate the food base for predators above that expected from in situ productivity (Anderson and Wait, 2001; Schoener, this volume).
An alternative way for island area (and distance) to influence food chain length involves the consideration of trophic dependencies among species, in their own right. Introducing trophic dependencies into colonization-extinction dynamics can lead to the expectation that food chain length will increase with island area. I here summarize models exploring this idea presented earlier (Holt 1993, 1996, 1997a,b, 2002; see also Schoener et al. 1995) and weave in new thoughts and examples.
All species need resources and to some degree have specialized diets. If a species arrives on an island lacking its required resources, it cannot persist. On a continent, recurrent immigration can sustain “sink” populations at sites without resources, but if the distance between the mainland and island is sufficiently great, such sink populations will be absent or vanishingly rare. Consider an unbranched food chain of “stacked specialists.” Species i has trophic rank i and feeds on species i − 1. A useful descriptor of island distributions is the incidence function (Diamond 1975), which gives the percentage of islands occupied by species i, p(i), as a function of island area, or distance to the mainland, or other island traits. In a food chain of stacked specialists, at equilibrium the incidence of species i is constrained by the incidence of all lower-ranked species on which it directly or indirectly depends. This leads to nested spatial distributions; islands without species i − 1 are guaranteed not to harbor species i, but the converse need not hold.
We now define a conditional incidence function p(i | i − 1) to be the conditional probability that species of rank i is present, given that its required resource, species i − 1, is present. Often, conditional incidence will increase with island area. Specialist herbivores, for instance, are often more likely present on larger populations of their host plants (Otway et al. 2005). The unconditional incidence function for species i is a product of conditional incidence functions, up the food chain:
With this expression, and some simple assumptions, we can draw conclusions about how food chain length should vary with area and distance. The expected food chain length is simply the sum of incidence functions, up the chain:
Assume that the incidence function for the basal species and the conditional incidence function for each higher-ranked species all increase with island area and decrease with increasing distance from the mainland. By application of the chain rule, we find that the expected food chain length also increases with area, and decreases with distance. As an example, Komonen et al. (2000) report that, following forest fragmentation, a specialist food chain supported by a bracken fungus was truncated on small forest fragments. So, with almost no biology at all, other than assuming trophic specialization and the garden variety expectation that island area and distance affect the likelihood that a species will be present, we can predict effects of island area and distance on food chain length.
As noted above, a principal motivation of MacArthur and Wilson’s monograph was to understand how species richness covaried with island area and distance. Instead of a single food chain, assume the mainland community has m “stacked specialist” chains. What is the effect of trophic rank on z? For simplicity, assume all species of rank i have the same conditional incidence function. The expected number of species of rank i is simply Si = mp (i). The strength of the species-area relationship on a log-log plot is
If conditional incidence increases with area, this expression implies that
The strength of the species-area relationship should thus increase with trophic rank.
“Stacked specialist” food chain models are a sensible starting point for the development of a theory of trophic island biogeography. But such trophic specialization does not typify most food webs, which contain a mix of tight specialists and highly generalized consumers. Developing models of multispecies webs which pay attention to the detailed pattern of trophic interactions, and how these change during community assembly to feed back onto colonization-extinction dynamics, is a significant challenge. One approach is to craft detailed community assembly models that specify rules for the explicit distribution of trophic specialization and generalization in source food webs, and then use these to assemble island communities. Here I focus instead on an alternative approach to trophic island biogeography. I ignore the details of the web of interactions and instead make broad qualitative assumptions about how diversity in one trophic level influences rates of colonization and extinction in another, using a somewhat simpler and extended version of a model presented in Holt and Hoopes (2005). The goal is to craft qualitative theoretical predictions describing how species richness scales with area, contrasting generalists with specialists, and predators with their prey.
We first assume donor control, so predators do not influence prey colonization-extinction dynamics. The prey follow model (6.1) above and show island area and distance effects. Colonization-extinction dynamics in the predators is controlled in a bottom-up fashion by the number of prey species present on an island, S, as well as by island area and distance. It is well known that there can be a codependency in species richness among trophic levels. For instance, the composition of local arthropod herbivore communities is strongly affected by plant community composition (Siemann et al. 1999, Schaffers et al. 2008). So a reasonable rule of thumb is that a more diverse prey base should be able to support a more diverse assemblage of consumers.
The number of predator species on an island is P, which can change by colonization or extinction. This is assumed to given by an expression like (6.1) above. I use a prime to denote predator immigration and extinction rates. The immigration rate of the predator guild I′ is assumed to increase with the number of prey species present on the island. Likewise, we assume the extinction rate E′ decreases with increasing island area, for a fixed number of prey species, and also decreases with an increasing number of prey species, for a fixed island area. Taking logarithms of (6.2), as before, after some manipulation it can be shown that the z-value of the predators is related to the z-value of their prey by the following compact expression:
where
The first term on the right side of expression (6.12) describes the indirect effect of area upon predator log(species richness), mediated through the species richness of the prey. The second term describes the direct effect of area upon predator extinction, controlling for prey species richness. With these expressions in hand, we can now address several qualitative issues in trophic island biogeography.
How should the z-values for specialists differ from those for generalist predators? Consider colonization. It is often reasonable to expect predator colonization to increase with prey species richness. For a specialist, colonization requires the prior presence of its required prey. On small, species-poor islands, there is a high probability that any particular prey species will be absent, precluding colonization by specialists that need it. Colonization by specialists should be more likely, the more prey species are present. For generalists, colonization may also depend positively upon prey species richness. For instance, an increased number of prey species may increase the total food supply and permit a higher initial rate of increase. If different prey species provide distinct limiting nutrients (called “obligate generalism” in Holt et al. 1999), colonizing predators may require multiple prey species to enjoy positive growth rates at all. But more usually, a generalist should be able to colonize communities containing many different subsets of the mainland prey community. If so, there may be a relatively weak effect of prey species richness upon colonization by generalists, compared to specialists.
It also seems reasonable that predator extinction rates should decrease with an increase in prey species richness. Ritchie (1999) provides a nice empirical example for prairie dog colony extinction rates, which decline with increasing plant species richness. But again, this effect may be stronger for specialists than for generalists. For specialist predators, their extinction rates can be no less than those of their required prey types—when a given prey species goes extinct, it drags all its specialist consumers with it. Generalists, by contrast, may subsist on other prey species, and so a reduction in prey species richness could imply a more modest increase in extinction rates. This should imply a lower Q for generalists, compared to specialists.
The final term in (6.12) is the direct effect of area upon predator extinction rates, controlling for prey species richness. Two factors are at play here. First, all else being equal, a decrease in area will proportionally shrink absolute population sizes. A systematic difference in the average densities of generalist vs. specialist predators would then imply a comparable difference in area sensitivity. I know of no data that directly address systematic differences in abundance as a function of degree of trophic specialization. Second, specialist predator-prey interactions are prone to unstable dynamics, with recurrent phases at low densities. Predators face a differential risk of extinction in these phases, a risk that is magnified on small islands. Moreover, as discussed below, small islands may lack spatial mechanisms that stabilize specialist predator-prey dynamics, further aggravating extinction risks of specialist predators versus generalists. It is thus plausible to hypothesize that extinction rates of specialists will be more sensitive to area, than will be the case for generalists.
These observations lead to the prediction that for a given trophic level zspecialist > zgeneralist. When will predators have a steeper species-area relationship than their prey, i.e., zpred > zprey? It is sufficient that Q > 1, which is more likely if both predator immigration and extinction rates vary strongly with prey species richness. Direct area effects on the predator can also make it possible for predator z-values to exceed those of their prey, even if Q < 1.
If one accepts the above arguments, it is overall more likely for the z-values of predators to exceed those of their prey, when predators are relatively specialized in their diets; when overall immigration rates of predators are low, relative to extinction; and, when there are additional effects of area upon predator extinction rates, arising for reasons other than the effect of area upon prey species richness.
Empirical studies of the relationship between trophic rank and the species-area relationship, where comparison is made among taxa within a given set of islands or habitat patches, reveal patterns broadly consistent with these theoretical expectations. In a nice study of how trophic specialization influences the species-area relationship, Steffan-Dewenter and Tscharntke (2000) showed that the predicted effect of trophic generalization on the magnitude of z is found in butterflies differing in dietary breadth and distributed across habitat fragments; z-values increase monotonically from butterflies which are extreme generalists, to oligophages, to tight specialists on a single host plant (figure 6.4). Trophic generalists had lower z-values (between 0.05 and 0.1) than their host plants (0.13), whereas oligophages and monophages had higher values (0.16 and 0.21). This pattern matches the above theoretical predictions. Kruess and Tscharntke (2000) report species-area relationships for herbivorous insects, and their relatively specialized parasitoids, in meadows of red clover and vetch in central Europe, and demonstrate that z is considerably higher for the parasitoids than for their hosts (figure 6.5). Holt et al. (1999) review other examples. In assemblages dominated by trophic specialists, stronger species-area relationships (higher z) typically are seen at higher trophic ranks. But generalists reveal a mix. Some examples fit, but others do not. Even generalists can show strong area effects. For instance, Spencer et al. (1999) studied effects on predator extinction in arthropod communities in temporary ponds in Israel, and found the proportion of the community comprised of generalist predators to increase strongly with log(area) (figure 6.6). John Glasser (1982) reanalyzed the classic Simberloff-Wilson (1969) study of arthropod communities on mangrove islets and found a suggestion of successional patterns in web structure. He classified species into three trophic groups: herbivores, predators, and parasites, and then plotted their colonization curves. One result (his figure 7) reveals a pronounced area effect on trophic organization: at the end of the study, the large islands E7 and E9 had a larger predator species-to-herbivore species ratio than did the small islands E1 and E2. But invertebrate predators on islands in the Gulf of California do not show a systematic increase in z-values with trophic rank (Holt et al. 1999; G.A. Polis, personal communication); consumers such as scorpions are highly generalized and have lower z-values than do lower-ranked trophic levels on the same islands (e.g., plants).
Figure 6.4. In butterflies of central Europe, there is a systematic relationship between the value of z, and the degree of trophic specialization (from Steffan-Dewenter and Tscharntke 2000).
Figure 6.5. An example of stronger species-area relationships for specialist natural enemies (parasitoids) than their prey (host insects), in meadows in central Europes (from Kruess and Tscharntke 2000).
Figure 6.6. An example of stronger species-area relationships for generalist predators than for their prey (aquatic organisms in temporary ponds) (from Spencer et al. 1999, and unpublished data provided by Leon Blaustein). Closed circles: macroscopic predators; open circles: all predators.
Piechnik et al. (2008) have recently analyzed the Simberloff-Wilson dataset in more detail, and conclude that there is a succession in niche breadth among consumers, with generalists colonizing before specialists. It is plausible, as Piechnik et al. suggest, that this reflects the sequential dependence of colonization expected for specialist consumers, which have to wait for establishment of their required resources before colonizing, as assumed in the theory sketched above. As Montoya et al. (2006) note, some community patterns may best be explained by an assembly “process whereby species sequentially partition resources as they invade an ecological community. Rare, trophically specialized species enter the community later than do generalists.”
An alternative, complementary explanation for the higher z-values shown by specialists may be that generalists are good colonists for reasons other than their ability to exploit a variety of prey. From (6.12) and (6.13), if immigration rates are higher for generalists, then even with comparable area dependencies in the rate constants, the z-values for generalists will be lower. Why might this be a reasonable expectation? Model (6.1) (et seq.) assumes a noninteractive community. If we consider competition among predators for a moment, the question that arises is what permits the coexistence of specialist and generalist consumers? Given trade-offs in exploitative ability, as a broad rule of thumb (albeit with exceptions) one expects guilds of specialists, each with skills honed to their own particular prey, to outcompete generalists. Generalists could nonetheless persist in a metacommunity, given a trade-off between competitive abilities and colonizing abilities, so that generalists arrive before specialists, say, following local disturbances. This might preadapt generalist consumers to be among the earlier colonizers onto isolated islands.
In reality, communities emerge from the interplay of both bottom-up and top-down forces, as well as “horizontal” forces (competition, mutualism). This leads to a wide range of complex and interesting issues in spatial community ecology (Amaresakare 2008), and below I explore some that must be considered en route to a fully fleshed-out theory of trophic island biogeography.
One way to proceed is to develop models that explicitly describe colonization and extinction by each species. For a moment, consider again a food chain of stacked specialists. Schoener (Schoener et al. 1995, Appendix) and I (Holt 1996, 1997) independently developed Markov chain patch occupancy models that, in the spirit of MacArthur and Wilson (1967), track colonization and extinction at each trophic level. With this model, we relax the assumption of donor control. I will not repeat the analyses here but instead summarize results. The “state” of each island is the length of its food chain. For simplicity, we assume the basal species in the chain to be an effective colonizer, i.e., its incidence is unity. A fraction of islands, P1, have just the basal species (e.g., a plant), a fraction P2 have that species and a prey species that utilizes it (e.g., an herbivore), and the remaining fraction P3 have the full food chain. The predator can colonize only after the prey has become established. If the prey species goes extinct, so does the predator; in addition, the predator might go extinct on its own. A model based on these assumptions is
(The subscript “ij” denotes “transition from state i to state j.”) At equilibrium, we can solve to examine how occupancies depend on area. We assume that extinction rates decline with increasing area, and consider three basic possibilities:
1. e21 = e31. Prey extinction is not affected by the predator. We might call this “biogeographic donor control.” In food web ecology, donor control denotes situations in which resource recruitment is independent of consumption by a consumer. If a predator does not alter prey extinctions, then even if predation is biologically significant (e.g., causes decreased local prey abundance), this will not be reflected in occupancy.
2. e21 > e31. The prey extinction rate is reduced by the predator. This seems counterintuitive, but the effect is well grounded in theory and empirical examples are known. May (1972), for instance, showed in a model of a three-link food chain that a top predator attacking an herbivore could stabilize plant-herbivore dynamics if the top predator experiences direct density dependence (e.g., from territoriality), and the herbivore on its own has weak direct density dependence and is easily saturated by its own resource. On small islands in the Baltic, for instance, voles in the absence of predation explode to high numbers and overgraze their food resources to the point of local extinction, whereas numbers stay steady and bounded away from zero when predators are present (Banks et al. 2004).
3. e21 < e31. The final possibility is for the predator to increase prey extinctions. This may be the most likely of the three logical possibilities.
For the two first possibilities, larger island area implies a longer equilibrial food chain length. In the third, food chain length can decrease with increasing island area, or proximity to the source. This paradoxical effect can arise if an increase in area strongly decreases extinctions by the predator alone. An intuitive explanation goes as follows. If the predator colonizes small islands, by assumption it goes extinct rapidly, leaving the prey behind. But on a large island, the predator may persist and grow, drive its prey extinct, and then itself go extinct, reinitializing the island with just the basal species. Averaging over food chains on all islands of a given size, one might find shorter chains on larger islands, because these are precisely the arenas where predators persist long enough to exterminate their prey. This effect is particularly likely if predators have alternative resources which permit them to persist, at least for a while, in the absence of the focal prey species.
Fundamental features of predator-prey ecology suggest that there should be strong dependencies on island area of extinction rates in food webs. Classical predator-prey theory predicts that, if predators effectively limit their prey, unstable dynamics arise with periods at low densities. On a small island there will be recurrent periods of low absolute abundances, hence elevated extinction risks. All else being equal, unstable predator-prey interactions should be more persistent on large islands. This is a pure area effect.
Another pure area effect arises because the larger the island, the less likely it will contain well-mixed populations. Many taxa are relatively sluggish, and with limited within-island dispersal, partially independent populations are likely to emerge within large islands (Holt 2002). One active area of research in community ecology at present is metacommunity ecology (Holyoak et al. 2005), which is an intellectual descendant of island biogeographic theory. A “metacommunity” is a set of local communities, connected by dispersal. In a metacommunity, colonization into a focal patch comes from other occupied patches, rather than a fixed external source. Even if one is primarily interested in islands, there are good reasons to consider the implications of metacommunity dynamics for understanding within-island processes. When a species first colonizes an island, it rarely immediately occupies the entirety of the island, but establishes a beachhead, from which it expands. If dispersal is limited within islands, one can view island area as being a proxy for the number of local sites potentially connected by within-island dispersal (Holt 1992). Larger islands in effect are larger meta-communities, comprised of more such local sites. Area effects on extinction rates reflect the diverse ways island area influences internal metacommunity dynamics.
Predator-prey models incorporating space, dispersal, and localized interactions in metacommunities are often more stable than nonspatial models (Holt 1984, Hosseini 2003), due to several distinct stabilizing mechanisms that emerge in spatially distributed systems. All these mechanisms should be sensitive to area, and so could contribute to systematic effects of island area on food web structure. There are several recent reviews of the influence of space on the persistence and stability of predator-prey and food web interactions (Hassell 2000, Briggs and Hoopes 2004, Holt and Hoopes 2005), and here I summarize key insights that seem particularly germane to island biogeography.
Even in homogeneous areas, localized interactions, limited dispersal, and stochastic variation generate heterogeneities in population abundance and interaction strengths that are broadly stabilizing (Hassell 2000, Briggs and Hoopes 2004). A large area can contain many local populations that become asynchronous in their dynamics, given limited within-island dispersal, permitting persistence of locally unstable predator-prey interactions. Experimental studies suggest that with localized predator-prey interactions, persistence is enhanced with increasing size of the arena containing the interaction (Huffaker 1958, Holyoak and Lawler 1996, McCauley et al. 2000, Ellner et al. 2001). Theoretical models predict that spatial patterns such as traveling waves emerge at scales larger than the local population, but smaller than the whole system, and these patterns can contribute to stability (Hassell et al. 1991). But these emergent spatially patterned interactions have characteristic spatial scales (Donalson and Nisbet 1999, Gurney and Veitch 2000), and so cannot be sustained on small islands (Hassell et al. 1991). Wilson et al. (1998) considered a food chain of a hyperparasitoid, a primary parasitoid, and a basal host, all interacting on a lattice, in effect an island with local dispersal and highly unstable local interactions. This theoretical study revealed that food chain persistence was strongly sensitive to lattice size. An order-of-magnitude larger lattice was needed to sustain the full tritrophic interaction, compared to the host-parasitoid interaction (figure 6.7). Larger islands also often contain internal hetereogeneities (e.g., distinct habitats) leading to spatial variation in parameters such as attack rates and intrinsic growth rates. In general, such environmental heterogeneities can stabilize predator-prey systems (Holt 1984, Hassell 2000, Schreiber et al. 2006).
Developing patch occupancy models for more complex multispecies assemblages is a challenging task, because of the proliferating number of possible states and transitions (see Holt 1997a, 2002 for complexities arising even for simple food chains in a metacommunity context). The above models just blithely ignored all the reticulate detail of the structure of the web of interactions among species. As one example of the importance of such details, food chains may in some cases be longer on larger islands not because of the sequential additions of species at increasingly higher trophic ranks, as assumed above. A given predator species may be found across all islands, but be at a realized higher trophic rank on larger islands because those islands are also occupied by additional species at various intermediate ranks (Post and Takimoto 2007). For instance, in the Midwest United States the lake trout is the top predator across a wide range of lake volumes, but it is at a higher realized trophic rank in larger lakes, which compared to small lakes have many additional species of zooplankton and smaller fish providing long chains linking phytoplankton to the trout (Post et al. 2000b).
Figure 6.7. Larger lattices (a surrogate for island area) are more likely to retain strongly interacting food chains of hyperparasitoids, parasitoids, and hosts (from Wilson et al. 1998).
In general, larger areas may permit the persistence of otherwise unstable multispecies trophic interactions. When two prey species share a common enemy, one can indirectly exclude the other locally via the numerical response of their shared enemy. But this strong apparent competition (Holt and Lawton 1994, Hamback et al. 2006) may not cause extinction in spatially extended systems, if the prey inferior at withstanding predation more effectively colonizes empty patches, or if the predator prefers the prey with faster growth (King and Hastings 2003). Bonsall et al. (2005) showed this experimentally for a parasitoid attacking two species of bruchid hosts; coexistence was prolonged when the interaction played out in a larger spatial arena. In a field study, Cronin (2007) showed experimentally that one plant hopper (Delphacoides schlochoa) strongly suppressed to the point of local extinction another plant hopper (Prokelisia crocea) (even though the two hosts occupied distinct habitats) due to numerical responses of shared parasitoids that straddled these habitats. He argued that coexistence occurred regionally because the species superior at withstanding the shared parasitoids was a poorer disperser. Such coexistence mechanisms are ineffective on small islands.
Moving to entire food webs, in his celebrated book on species invasions, Charles Elton (1958) argued that islands are prone to unstable dynamics and vulnerable to invasion, because of reduced species richness. McCann et al. (2005a,b) observe that this pattern (assuming it is true) could instead reflect the fact that island food webs are spatially constrained and so not buffered by the stabilizing mechanisms that emerge from interspecific interactions played out in expansive spatial arenas. Spencer and Warren (1996) carried out experiments on multispecies webs in small aquatic microcosms, where they compared the productive space hypothesis of Schoener (1989) with the effects of area, per se. They concluded that their results did not fit the productive space hypothesis very well, but “that spatial effects on the persistence of unstable food webs may be important.” Spatial heterogeneity permits many mechanisms to operate—predator switching among habitats, source-sink relationships, and transient refuges—stabilizing even complex food webs (Holt 1984, Post et al. 2000a, Kondoh 2003, Eveleigh et al. 2007, Goodwin et al. 2005, Gripenberg and Roslin 2007). Conversely, on small islands the inherent instability of strong trophic interactions can be unleashed and cause extinctions. On large islands, within-island metacommunity processes may help counter the many ways species-rich webs have of being locally unstable.
It is useful to provide pointers to some of the interesting and challenging complexities that need to be addressed in a mature trophic island biogeography. Many of these reflect important intellectual currents in contemporary community ecology.
Interaction modifications. There are behavioral effects by which predators can indirectly influence prey persistence in metacommunities. For example, the presence of predators in a patch can induce prey to emigrate, enhancing colonization rates into empty patches (Gilliam and Fraser 2001, Prakash and de Roos 2002), thus facilitating prey persistence. This behavioral effect needs considerable space to operate effectively, and so might help further explain why strong local predator-prey interactions can persist on large islands, but not on small islands. Many other kinds of interaction modification (Abrams 1983) could modulate colonization-extinction dynamics. For instance, nonprey can interfere with the ability of a predator to capture its prey (Vos et al. 2001, Kratina et al. 2007, van Veen 2005); this is called “associational resistance” in plant-herbivore interactions (Atsatt and O’Dowd 1976, Hamback et al. 2000, Aquilino et al. 2005, Callaway et al. 2005). Such facilitation among prey has several consequences for trophic island biogeography. As overall prey species richness increases with island area, the stability of a specialist predator-prey interaction could be enhanced, relative to a monoculture, because predators are less able to overexploit their prey. So extinction rates of specialist predators and their prey may decline on larger islands. Countering this effect, however, successful colonization by specialist predators may be inhibited in richer prey communities. Colonization rates by specialists might actually peak at intermediate island sizes, then decline on larger islands.
Moreover, predator diversity can have diverse effects on the overall consumption of prey (Casula et al. 2006). Such diversity can augment predation pressure on prey (van Ruijven et al. 2005, Snyder et al. 2006), for instance because prey have fewer places to hide or modes of behavior that permit predator avoidance. Or, predators may interfere with each other, relaxing predation on their shared prey. If predator diversity increases with prey species richness, which effect predominates will govern how prey colonization and extinction rates change with predator species richness, which can then feed back onto colonization by the predators themselves.
Ecosystem dimensions of trophic island biogeography. Flows of materials between marine and terrestrial ecosystems can profoundly impact island communities. On unproductive islands, a regular influx of subsidies from marine sources can sustain terrestrial consumers even on very small islands (Anderson and Wait 2001), which can then exploit resident prey more effectively. Conversely, pulsed subsidies can lead to periods of relaxed predation upon resident island prey (Schoener, in press). Predators can limit the abundance of species (e.g., seabirds) that are key conduits of nutrients between islands and marine environments (Maron et al. 2006).
Transients. Oceanic island communities are likely to assemble one species at a time. After a species colonizes a food web, there is often a phase of pronounced transient dynamics, where abundances deviate very sharply from long-term equilibrial values, possibly for long periods of time (Hastings 2004). For instance, when a resident predator and prey are present, and a second prey species which does not compete directly with the resident is introduced, large-amplitude cycles in all species result enroute to a long-term stable equilibrium (Holt and Hochberg 2001). Though in the long run all species mathematically persist in this deterministic model, in biological practice extinctions may occur when species pass through transient low-density troughs. Noonburg and Abrams (2005) show that in a standard model of keystone predation—where a top predator facilitates coexistence of competitors by feeding preferentially on the dominant prey—invasion by one prey species into a community with the other species initially present and at equilibrium leads to very low densities, which in practice would likely preclude realistic coexistence. All these newly recognized effects of transient dynamics should be particularly important in small oceanic islands, where absolute abundances are in any case low. By contrast, on a continental island, the initial community is carved out of the original mainland biota, and such transient dynamics emerging during assembly should be less important in determining current community structure.
Cyclic assembly processes. Theoretical and experimental studies of food web assembly reveal that local communities receiving immigrants from an external source can go through cycles, from state A to B, and back again, or from A to B to C to D . . . and finally back to A. Cyclic compositional changes are common in theory (e.g., Morton and Law 1997, Steiner and Leibold 2004). Warren et al. (2003) in a microcosm study with protists found that the community could exist in two states, which we dub A and B. Predatory species could invade A, and transform it into B, and then themselves go extinct. After B had settled down, the predators could reinvade, and take the community back to A, and again the predators went extinct. For this process (and indeed any cyclical dynamics in composition) to be maintained, there needs to be a supply from external sources (either a continent or metacommunity) for one or more species.
A plausible cyclic assembly scenario emerges from considering the implications of garden variety, uncontroversial community ecology played out on islands, which can be understood (I hope) even without equations (see figure 6.8). Consider a source where two predators share two biotic, noncompeting resource populations and stably coexist (upper left corner of the figure). Predator coexistence requires niche partitioning, which we assume suffices for coexistence but is incomplete (i.e., there is dietary overlap). Assume predator 1 has a rate of exploitation of prey 1 (denoted by α) higher than on prey 2 (α′). Reciprocally, assume predator 2 is better at exploiting prey 2 at rate α but also exploits prey 1 at rate α′. In simple cases, for instance if the two predators have linear functional and numerical responses to their prey, and the prey have logistic growth, it can be shown that an equilibrium with all species exists and is locally stable. We assume this food web module persists on the mainland, and that the two predators are effective at limiting prey abundance below carrying capacity. Despite the local stability of this four-species module at equilibrium, species losses can lead to a cascade of additional extinctions. If we consider the three-species subwebs within this four-species module, it is clear why instability looms, should a species be lost. Say a prey species is missing. If both predators are still present, we expect competitive exclusion; the predator better at utilizing that prey supplants the other. If instead one of the predators is absent, given our assumption about effective predation, exclusion due to apparent competitive advantage can occur; the prey species experiencing the lower predation rate indirectly supplants the more vulnerable prey species, mediated through the shared predator’s numerical response.
Figure 6.8. Community churning. Two predators persist on two prey species, because of resource partitioning on a mainland. Because of the reciprocal forces of resource competition and apparent competition in each three-species module, a series of colonizations and extinctions can be observed on islands, leading to a perpetual cycle in island community composition. See main text for details.
When an island community is assembled, an interesting phenomenon emerges. If colonizations are rare, in any given time period only one species is likely to colonize. We start with an empty island. The two prey species colonize first, then a predator. This predator overexploits the prey species to which it is best adapted (as measured by the attack rate), leaving it sustained by the prey to which it has the lower attack rate. When the other predator colonizes, the first predator is now competively excluded. But the resulting two-species configuration is now open for colonization by the alternative prey species (which experiences a lower attack rate than the resident), after which the resident prey species is supplanted. This in turn permits the original predator to colonize again, restarting the cycle. These alternative shifts in species composition, driven by reciprocal shifts in the relative importance of resource competition and apparent competition, can lead to a constant, if leisurely, churning in island species composition, with colonists drawn from a stable mainland community. Variability in species composition among comparable islands may reflect not just the chance vicissitudes of colonization, but emergent heterogeneities due to inherent community instabilities.
More complex webs, and parasites. The theories presented above have assumed simple patterns of trophic organization, such as simple food chains, or discrete predator and prey trophic levels, as well as species with fixed properties. Realistic food webs are often very complex in their organization, with reticulate feeding relationships among large numbers of species, and on top of this complexity, the properties of food webs also should reflect the long-term imprint of coevolution among species, as well as speciation. One important class of trophic interactions that is still poorly understood in the context of food web ecology is host-parasite interactions, but it is increasingly clear that such interactions are ubiquitous and dynamically important (Lafferty et al. 2008). There are many potential implications of parasitisim for trophic island biogeography. For instance, many host-specific pathogens have strong area effects in incidence, with an increasing probability of being present on larger patches containing more of their hosts (e.g., the smut fungus Ustilago scorzonerae on its asteraceous host, Scorzonera humilis; Colling 2004). Although rather poorly documented, in some systems it is clear that the parasite load is less on distant islands; for instance, Anolis lizards in the northern Lesser Antilles have depauperate parasite faunas (Dobson et al. 1992). The presence of parasites can have surprising effects on predator-prey interactions. For instance, if prey sustain a pathogen, selective predation on infected prey can at times increase prey numbers and also prevent devastating epidemics (Packer et al. 2003). The strength of the effect of the predator on prey numbers may be greater on small islands, because pathogens are missing there. There may even be profound evolutionary effects from depauperate parasite communities on islands. Ricklefs and Bermingham (2007) suggest that one reason the Lesser Antilles had a more modest avian radiation than either the Galapagos or the Hawaiian Islands is that the latter two archipelagoes have relatively few pathogens, which when present can prevent secondary sympatry of budding species due to disease-mediated competition.
Why did The Theory of Island Biogeography resonate so thoroughly? Many biologists find islands intrinsically fascinating, and the interplay of empirical patterns and mathematical theory in MacArthur and Wilson (1967) presented a new paradigm for ecological studies. But beyond this, the late 1960s and 1970s were a time of increasing concern among environmentalists and scientists about the serious environmental problems caused by humanity around the globe, most notably extinction threats caused by habitat destruction and fragmentation. Indeed, Macarthur and Wilson (1967) opens on this theme: their first figure is the celebrated diagram by Curtis of forest fragmentation from a section of land in Wisconsin. Scenes of tropical deforestation—rich forests replaced by depauperate cattle pastures or miles upon endless deadeningly dull miles of oil palm plantations—are depressingly familiar to any well-traveled biologist. Fragmentation creates land-locked “islands” of habitat. The conceptual perspective provided by the island metaphor sparked an explosion of work on habitat fragmentation (Harris 1984), including observational studies, theory development, and long-term landscape experiments. Examples include the ongoing Biological Dynamics of Forest Fragments Project near Manaus, Brazil, 1979–present (Bierregard et al. 2001), and my own project on secondary succession in a fragmented landscape, near Lawrence, Kansas, 1983–present (Robinson et al. 1992, Cook et al. 2005). Laurance (this volume) provides an overview of the value—and limitations—of island theory for understanding habitat fragmentation.
A ubiquitous implication of habitat fragmentation is the disruption, elimination, or magnification of preexisting trophic interactions (see Terborgh chapter). Theoretical studies (Holt 1993, Holt et al. 1999, Bascompte and Sole 1998, Sole and Bascompte 2006, Sole and Montoya 2006) suggest that species (especially specialists) at higher trophic ranks may be differentially vulnerable to fragmentation. I mentioned above several studies of species-area relationships from fragmented habitats, consistent with these predictions. Empirical studies show that parasitoids (which are often relatively specialized) are more extinction-prone than their hosts (e.g., Cronin 2004), leading to reduced parasitism on smaller or more isolated habitat fragments (Kruess and Tscharntke 1994, Elzinga et al. 2005, Steffan-Dewenter and Tscharntke 2002, Tscharntke et al. 2002, Tscharntke and Brandl 2004, Valladores et al. 2006, van Nouyys 2005). At the community level, this differential susceptibility to fragmentation can lead to reduced predator-to-prey ratios with decreasing patch area (Didham et al. 1998, Ryall and Fahrig 2006), to trophic cascades (more intense herbivory on smaller patches where prey are freed from predation; Terborgh et al. 2001), and to steeper species-area relationships for predators than prey among fragments (Hoyle 2004). So some fragmentation effects do seem to match the above predictions of trophic island biogeography about food chain length and trophic influences on the strength of the species-area relationship.
However, although island biogeography continues to provide a powerful metaphor for thinking about habitat fragmentation, with the maturation of conservation biology it has become widely recognized that this metaphor can be limited, and at times misleading. Habitat fragments in some ways are like islands, but in some ways are radically different (Ewers and Didham 2006, Watling and Donnelly 2006, Laurance chapter). Edge effects can penetrate deep into fragments (Ewers and Didham 2006). The area separating fragments is not an empty sea, a mere barrier to dispersal, but sustains communities which often utilize the fragmented habitats to some extent. Coupling of distinct habitats by consumer or resource movement is a ubiquitous landscape process (Polis et al. 1997). Even as specialist predators become less important on small fragments (as predicted by trophic island biogeographic theory), generalist predators may become ever more present. For example, Robinson et al. (1995) showed that in the Midwestern United States, nest predation on forest birds by generalist predators increased strongly with fragmentation. Rand and Louda (2006) likewise showed that insect herbivores in remnant prairie patches in Nebraska experienced more intense predation due to generalist coccinellids sustained across a broader agricultural landscape, and comparable effects emerge in a wide range of landscape studies (Ryall and Fahrig 2006, Tscharntke et al. 2005, Rand and Tscharntke 2007). So even if top-down effects on small or distant oceanic islands are arguably unimportant, they may be very strong in small or isolated habitat patches embedded in anthropogenically modified landscapes, leading to strongly synergistic effects of predation with fragmentation (Davies et al. 2004, Rand et al. 2006). Moreover, transient dynamics are a key aspect of habitat fragmentation when landscapes shift rapidly. Holt and Hochberg (2001) conjectured that habitat destruction could lead to transient spikes in natural enemy impacts in remnant patches, as mobile predators crowd into remaining suitable areas. Thies et al. (2008) empirically demonstrated this effect; reductions in the area of rape crop cultivation led to a large short-term increase in mortality imposed by parasitoids on hosts in the remnant crop patches and elevated extinction risks, because parasitoids produced over a larger area surged into these areas.
So habitat fragments are not just islands. But it is clear that the island biogeograpic perspective has played a crucial historical role in stimulating analyses of habitat fragmentation (Laurance, this volume). Moreover, as humans continue to degrade the matrix habitat separating fragments, the long-term outcome may be island-like reserves, separated by a wasteland not all that different from a sterile ocean.
The Theory of Island Biogeography was a harbinger of the current rising tide of interest in spatial patterns and processes throughout the basic and applied ecological sciences, including food web ecology. Rather than end this essay by trying to summarize the ideas presented above, I would like to conclude on a more personal note. The Theory of Island Biogeography appeared in 1967. In 1970, I had the exceptional good fortune as a sophomore at Princeton of taking “Biogeography,” taught by Robert MacArthur and Ed Fischer. Due to an improbable series of events, MacArthur became my advisor in a special university program, and he graciously took me along on his last lengthy field trip to Arizona in 1971, where I helped him carry out some of his foliage-profile measurements—he would stand in an opening in the chaparral, while I would disappear, thrashing along a randomly chosen direction he had picked into the thick, clothes-ripping grip of the scrub, carrying a checkerboard. The goal was for me to hold it up at different distances, until half the squares were hidden from his view. It was physically challenging, but I did manage to stumble across a Flammulated Owl, a few feet away from one of my sampling points—still the only one I have ever seen. In conversations over the campfire, and then in his office later, MacArthur gently guided my thinking toward an academic career in ecology (rather than physics, my major). On his sickbed in 1972, he handwrote letters of recommendation for me to deliver to Ed Wilson at Harvard, and elsewhere. I have no doubt this was instrumental in my getting into fine graduate schools. How lucky can a clueless young man from Tennessee be!
I have profited from many teachers, mentors, friends, and colleagues over the years, learning from them much related to the themes of my essay. As noted above, Robert MacArthur and E. O. Wilson had an impact on my own life, for which I am eternally grateful. In my undergraduate years, John Terborgh had (and continues to have) many influences on how I think about the world, and I recall with fondness my interactions with Henry Horn, John Bonner, and Tom Givnish. In graduate school, I was very fortunate to interact with Tom Schoener and Joel Cohen as teachers, and to have Bill Stubblefield and Russ Lande as friends. Since then, I have had the good fortune to have as professional collaborators thinking about these themes, and as friends, some of the finest scientists in the world: Gary Polis, John Lawton, Mike Hassell, Neo Martinez, David Post, Andrew Gonzalez, Stuart Pimm, Scott Robinson, Ilkka Hanski, David Steadman, George Robinson, Wendy Anderson, Scott Robinson, Manojit Roy, Rico Holdo, and Mike Barfield, among others. To all of you—thanks for the ride. I also thank the organizers for their invitation to contribute to this volume, and the University of Florida Foundation for its continued support.
Abrams, P. A. 1983. Arguments in favor of higher-order interactions. American Naturalist 121:887–91.
Adler, G. H., and R. Levins. 1994. The island syndrome in rodent populations. Quarterly Review of Biology 69:473–90.
Amarasekare, P. 2008. Spatial dynamics of food webs. Annual Review of Ecology, Evolution, and Systematics 39:49–500.
Anderson, W. B., and D. A. Wait. 2001. Subsidized island biogeography hypothesis: Another twist on an old theory. Ecology Letters 4:289–91.
Aquilino, K. M., B. J. Cardinale, and A. R. Ives. 2005. Reciprocal effects of host plant and natural enemy diversity on herbivore suppression: An empirical study of a model tritrophic system. Oikos 108:275–82.
Atsatt, P. R., and D. O’Dowd. 1976. Plant defense guilds. Science 193:24–29.
Banks, P. B., K. Norrdahl, M. Nordström, and E. Korpimäki. 2004. Dynamic impacts of feral mink predation on vole metapopulations in the outer archipelago of the Baltic Sea. Oikos 105:79–88.
Bascompte, J., and R. V. Sole. 1998. Effects of habitat destruction in a prey-predator metapopulation model. Journal of Theoretical Biology 195:383–93.
Berlow, E. L. 1999. Strong effects of weak interactions in ecological communities. Nature 398:330–34.
Bierregaard, R., C. Gascon, T. E. Lovejoy, and R. Mesquita, eds. 2001. Lessons from Amazonia: The Ecology and Conservation of a Fragmented Forest. New Haven, CT: Yale University Press.
Bonsall, M. B., J. C. Bull, N. J. Pickup, and M. P. Hassell. 2005. Indirect effects and spatial scaling affect the persistence of multispecies metapopulations. Proceedings of the Royal Society of London, Series B 272:1465–71.
Bonsall, M. B., D. R. French, and M. P. Hassell. 2002. Metapopulation structures affect persistence of predator-prey interactions. Journal of Animal Ecology 71:1075–84.
Briggs, C. J., and M. F. Hoopes. 2004. Stabilizing effects in spatial parasitoid-host and predator-prey models: A review. Theoretical Population Biology 65:299–315.
Callaway, R. M., D. Kikodze, M. Chiboshvili, and L. Khetsuriani. 2005. Unpalatable plants protect neighbors from grazing and increase plant community diversity. Ecology 86:1856–62.
Cantrell, R. S., C. Cosner, and W. F. Fagan. 2001. How predator incursions affect critical patch size: The role of the functional response. American Naturalist 158:368–75.
Casula, P., A. Wilby, and M. B. Thomas. 2006. Understanding biodiversity effects on prey in multi-enemy systems. Ecology Letters 9:995–1004.
Chaneton, E. J., and M. B. Bonsall. 2000. Enemy-mediated apparent competition: empirical patterns and the evidence. Oikos 88:380–94.
Colling, G., and D. Matthies. 2004. The effects of plant population size on the interactions between the endangered plant Scrozonera humilis (Asteraceae), a specialized herbivore, and a phytopathogenic fungus. Oikos.105:71–7 8.
Cook, W. M., J. Yao, B. L. Foster, R. D. Holt, and L. B. Patrick. 2005. Secondary succession in an experimentally fragmented landscape: Community patterns across space and time. Ecology 86:1267–79.
Cronin, J. T. 2004. Host-parasitoid extinction and colonization in a fragmented prairie landscape. Oecologia 139:503–14.
———. 2007. Shared parasitoids in a metacommunity: Indirect interactions inhibit herbivore membership in local communities. Ecology 88:2977–90.
Davies, K. F., C. R. Margules, and J. F. Lawrence. 2004. A synergistic effect puts rare, specialized species at greater risk of extinction. Ecology 85:265–71.
Diamond, J. M. 1975. Assembly of species communities. In Ecology and Evolution of Communities, ed. M. L. Cody and J. M. Diamond, 342–444. Cambridge, MA: Harvard University Press.
Didham, R. K., J. H. Lawton, P. M. Hammond, and P. Eggleton. 1998. Trophic structure stability and extinction dynamics of beetles (Coleoptera) in tropical forest fragments. Philosophical Transactions of the Royal Society of London, Series B 353:437–51.
Dobson, A. P., S. V. Pacala, J. D. Roughgarden, E. R. Carper, and E. A. Harris. 1992. The parasites of Anolis lizards in the northern Lesser Antilles. Oecologia 91:110–17.
Donalson, D. D., and R. M. Nisbet. 1999. Population dynamics and spatial scale: effects of system size on population persistence. Ecology 80:2492–507.
Duffy, J. E. 2002. Biodiversity and ecosystem function: The consumer connection. Oikos 99:201–19.
Dunne, J. A., R. J. Williams, and N. D. Martinez. 2002. Network structure and biodiversity loss in food webs: Robustness increases with connectance. Ecology Letters 5:558–67.
Ellner, S. P., E. McCauley, B. E. Kendall, C. J. Briggs, P. R. Hosseini, S. Wood, A. Janssen, M. W. Sabelis, P. Turchin, R. M. Nisbet, and W. W. Murdoch. 2001. Habitat structure and population persistence in an experimental community. Nature 412:538–43.
Elton, C. S. 1958. The Ecology of Invasions by Animals and Plants. London: Methuen and Co.
Elzinga, J. A., H. Turin, J.M. van Damme, and A. Biere. 2005. Plant population size and isolation affect herbivory of Silene latifolia by the specialist herbivore Hadena bicrurusi and parasitism of the herbivore by parasitoids. Oecologia 144:416–26.
Eveleigh, E. S., K. S. McCann, P. C. McCarthy, S. J. Pollock, C. J. Lucarotti, B. Morin, G. A. McDougall, D. B. Strongman, J. T. Huber, J. Umbanhowar, and L.D.B. Faria. 2007. Fluctuations in density of an outbreak species drive diversity cascades in food webs. Proceedings of the National Academy of Sciences U.S.A. 104:16976–81.
Ewers, R. M., and R. K. Didham. 2006. Confounding factors in the detection of species responses to habitat fragmentation. Biological Reviews 81:117–42.
———. 2008. Pervasive impact of large-scale edge effects on a beetle community. Proceedings of the National Academy of Sciences U.S.A. 105:5426–29.
Gilliam, J. F., and D. F. Fraser. 2001. Movement in corridors: Enhancement by predation threat, disturbance, and habitat structure. Ecology 82(1):258–73.
Glasser, J. W. 1982. On the causes of temporal change in communities: Modification of the biotic environment. American Naturalist 119:375–90.
Goodwin, B. J., C. G. Jones, E. M. Schauber, and R. S. Ostfeld. 2005. Limited dispersal and heterogeneous predation risk synergistically enhance persistence of rare prey. Ecology 86:3139–48.
Graham, R. W., E. L. Lundelius, M. A. Graham, E. K. Schroeder, R. S. Toomey, E. Anderson, A. D. Barnosky, J. A. Burns, C. S. Churcher, D. K. Grayson, R. D. Guthrie, C. R. Harington, G. T. Jefferson, L. D. Martin, H. G. McDonald, R. E. Morlan, H. A. Semken, S. D. Webb, L. Werdelin, and M. C. Wilson. 1996. Spatial response of mammals to Late Quaternary environmental fluctuations. Science 272:1601–6.
Gripenberg, S., and T. Roslin. 2007. Up or down in space? Uniting the bottom-up versus top-down paradigm and spatial ecology. Oikos 116:181–88.
Gurney, W.S.C., and A. R. Veitch. 2000. Self-organization, scale, and stability in a spatial predator-prey interaction. Bulletin of Mathematical Biology 62:61–86.
Hamback, P. A., J. Agren, and L. Ericson. 2000. Associational resistance: Insect damage to purple loosestrife reduced in thickets of sweet gale. Ecology 81:1784–94.
Hamback, P. A., J. A. Stenberg, and L. Ericson. 2006. Asymmetric indirect interactions mediated by a shared parasitoid: Connecting species traits and local distribution patterns for two chrysomelid beetles. Oecologia 148:475–81.
Harper, J. L. 1969. The role of predation in vegetational diversity. Brookhaven Symposium in Quantitative Biology 22:48–62.
Harris, L. D. 1984. The Fragmented Forest: Island Biogeography and the Preservation of Biotic Diversity. Chicago: University of Chicago Press.
Hassell, M. P. 2000. The Spatial and Temporal Dynamics of Host-Parasitoid Interactions. Oxford: Oxford University Press.
Hassell, M. P., H. N. Comins, and R. M. May. 1991. Spatial structure and chaos in insect population dynamics. Nature 353:255–58.
Hastings, A. 2004. Transients: The key to long-term ecological understanding? Trends in Ecology and Evolution 19:39–45.
Havens, K. 1992. Scale and structure in natural food webs. Science 257:1107–10.
Holt, R. D. 1977. Predation, apparent competition, and the structure of prey communities. Theoretical Population Biology 12:197–229.
———. 1984. Spatial heterogeneity, indirect interactions, and the coexistence of prey species. American Naturalist 124:377–406.
———. 1992. A neglected facet of island biogeography: The role of internal spatial dynamics in area effects. Theoretical Population Biology 41:354–71.
———. 1993. Ecology at the mesoscale: The influence of regional processes on local communities. In Species Diversity in Ecological Communities, ed. R. Ricklefs and D. Schluter, 77–88. Chicago: University of Chicago Press.
———. 1996. Food webs in space: An island biogeographic perspective. In Food Webs: Contemporary Perspectives, ed. G. Polis and K. Winemiller, 313–23. London: Chapman and Hall.
———. 1997a. From metapopulation dynamics to community structure: Some consequences of spatial heterogeneity. In Metapopulation Biology, ed. I. Hanski and M. Gilpin, 149–64. New York: Academic Press.
———. 1997b. Community modules. In Multitropihc Interactions in Terrestrial Ecosystems, ed. A. C. Gange, and V. K. Brown, 333–49. Oxford: Blackwell Science.
———. 2002. Food webs in space: On the interplay of dynamic instability and spatial processes. Ecological Research 17:261–73.
Holt, R. D., P. A. Abrams, J. M. Fryxell and T. Kimbrell. 2008. Reticulate food webs in space and time: Messages from the Serengeti. In Serengeti III: Human Impacts on Ecosystem Dynamics, ed. A.R.E. Sinclair, C. Packer, S.A.R. Mduma, and J. M. Fryxell, 241–76. Chicago: University of Chicago Press.
Holt, R. D., and M. E. Hochberg. 2001. Indirect interactions, community modules, and biological control: A theoretical perspective. In Evaluating Indirect Ecological Effects of Biological Control, ed. E. Wajnberg, J. K. Scott, and P. C. Quimby, 12–37. New York: CABI International.
Holt, R. D., and M. F. Hoopes. 2005. Food web dynamics in a metacommunity: Modules and beyond. In Metacommuniites: Spatial Dynamics and Ecological Communities, ed. M. Holyoak, M. Leibold, and R. D. Holt, 68–93. Chicago: University of Chicago Press.
Holt, R. D., and J. H. Lawton. 1994. The ecological consequences of shared natural enemies. Annual Review of Ecology and Systematics 25:495–520.
Holt, R. D., J. H. Lawton, G. A. Polis, and N. D. Martinez. 1999. Trophic rank and the species-area relationship. Ecology 80:1495–504.
Holyoak, M., and S. P. Lawler. 1996. Persistence of an extinction-prone predator-prey interaction through metapopulation dynamics. Ecology 77:1867–79.
Holyoak, M., M. A. Leibold, and R. D. Holt, eds. 2005. Metacommunities: Spatial Dynamics and Ecological Communities. Chicago: University of Chicago Press.
Hosseini, P. R. 2003. How localized consumption stabilizes predator-prey systems with finite frequency of mixing. American Naturalist 161:567–85.
Hoyle, M. 2004. Causes of the species-area relationship by trophic level in a field-based microecosystem. Proceedings of the Royal Society of London, Series B 271:1159–64.
Hubbell, S. P. 2001. The Unified Neutral Theory of Biodiversity and Biogeography. Princeton, NJ: Princeton University Press.
Huffaker, C. B. 1958. Experimental studies on predation: dispersion factors and predator-prey oscillations. Hilgardia 27:343–83.
King, A. A., and A. Hastings. 2003. Spatial mechanisms for coexistence of species sharing a common natural enemy. Theoretical Population Biology 64:431–38.
Komonen, A, R. Penttila, M. Lindgren, and I. Hanski. 2000. Forest fragmentation truncates a food chain based on an old-growth forest bracket fungus. Oikos 90:119–26.
Kondoh, M. 2003. Foraging adaptation and the relationship between food-web complexity and stability. Science 299:1388–91.
Kotiaho, J. S., and P. Sulkava. 2007. Effects of isolation, area, and predators on invasion: A field experiment with artificial islands. Applied Soil Ecology 35:256–59.
Kratina, P., M. Vos, and B. R. Anholt. 2007. Species diversity modulates predation. Ecology 88:1917–23.
Kruess, A., and T. Tscharntke. 1994. Habitat fragmentation, species loss, and biological control. Science 264:1581–84.
———. 2000. Effects of habitat fragmentation on plant-insect communities. In Interchanges of Insects between Agricultural and Surrounding Landscapes, ed. B. Ekbom, M. E. Irwin, and Y. Robert, 53–70. Dordrecht: Kluwer Academic Publishers.
Lafferty, K. D., S. Allesina, M. Arim, C. J. Briggs, G. De Leo, A. P. Dobson, J. A. Dunne, P.T.J. Johnson, A. M. Kuris, D. J. Marcogliese, N. D. Martinez, J. Memmott, P. A. Marquet, J. P. McLaughlin, E. A. Mordecai, M. Pascual, R. Poulin, and D. W. Thieltges. 2008. Parasites in food webs: The ultimate missing links. Ecology Letters 11:533–46.
Lomolino, M. V. 1984. Immigrant selection, predation, and the distributions of Microtus pennsylvanicus and Blarina brevicauda on islands. American Naturalist 123:468–83.
Lubchenco, J. 1978. Plant species-diversity in a marine intertidal community: Importance of herbivore food preference and algal competitive abilities. American Naturalist 112:23–39.
MacArthur, R. H., and E. O. Wilson. 1967. The Theory of Island Biogeography. Princeton, NJ: Princeton University Press.
Maron, J. L., J. A. Estes, D. A. Croll, E. M. Danner, S. C. Elmendorf, and S. L. Buckelew. 2006. An introduced predator alters Aleutian island plant communities by thwarting nutrient subsidies. Ecological Monographs 76:3–24.
Martinez, N. 1992. Constant connectance in community food webs. American Naturalist 139:1208–18.
May, R. M. 1972. Time-delay versus stability in population models with two and three trophic levels. Ecology 54:315–25.
McCann, K. 2000. The diversity-stability debate. Nature 405:228–33.
McCann, K., J. Rasmussen, and J. Umbanhowar. 2005. The dynamics of spatially coupled webs. Ecology Letters 8:513–23.
McCann, K., J. Rasmussen, J. Ubanhowar, and M. Humphries. 2005. The role of space, time, and variability in food web dynamics. In Dynamic Food Webs, ed. P. C. de Ruiter, V. Wolters, and J. C. Moore, 56–70. New York: Academic Press.
McCauley, E., B. E. Kendall, A. Janssen, S. Wood, W. W. Murdoch, P. Hosseini, C. J. Briggs, S. P. Ellner, R. M. Nisbet, M. W. Sabelis, and P. Turchin. 2000. Inferring colonization processes from population dynamics in spatially structured predator-prey systems. Ecology 2000:3350–61.
Montoya, J. M., S. L. Pimm, and R. V. Sole. 2006. Ecological networks and their fragility. Nature 442:259–64.
Morton, R. D., and R. Law. 1997. Regional species pools and the assembly of local ecological communities. Journal of Theoretical Biology 187:321–31.
Noonburg, E. G., and P. A. Abrams. 2005. Transient dynamics limit the effectiveness of keystone predation in bringing about coexistence. American Naturalist 165:322–35.
Nordstrom, M., and E. Korpimakki. 2004. Effects of island isolation and feral mink removal on bird communities on small islands in the Baltic Sea. Journal of Animal Ecology 73:424–33.
Olff, H., and M. E. Ritchie. 1998. Effects of herbivores on grassland plant diversity. Trends in Ecology and Evolution 13:261–65.
Otway, S. J., A. Hector, and J. H. Lawton. 2005. Resource dilution effects on specialist insect herbivores in a grassland biodiversity experiment. Journal of Animal Ecology 74:234–40.
Packer, C., R. D. Holt, P. J. Hudson, K. D. Lafferty, and A. P. Dobson. 2003. Keeping the herds healthy and alert: Implications of predator control for infectious disease. Ecology Letters 6:797–802.
Piechnik, D. A., S. P. Lawler, and N. D. Martinez. 2008. Food-web assembly during a classic biogeographic study: Species’ “trophic breadth” corresponds to colonization order. Oikos 117:665–74.
Pimm, S. L. 1982. Food Webs, reprinted with a new Foreword, 2002. Chicago: University of Chicago Press.
Polis, G. A., W. B. Anderson, and R. D. Holt. 1997. Toward an integration of landscape and food web ecology: The dynamics of spatially subsidized food webs. Annual Review of Ecology and Systematics 28:289–316.
Polis, G. A., M. E. Power, and G. R. Huxel. 2004. Food Webs at the Landscape Level. Chicago: University of Chicago Press.
Post, D. M. 2002. The long and short of food-chain length. Trends in Ecology and Evolution 17:269–77.
Post, D. M., M. L. Pace, and N. G. Hairston, Jr. 2000. Ecosystem size determines food-chain length in lakes. Nature 405:1047–49.
Post, D. M., and G. Takimoto. 2007. Proximate structural mechanisms for variation in food-chain length. Oikos 116:775–82.
Prakash, S., and A. M. De Roos. 2002. Habitat destruction in a simple predator-prey patch model: How predators enhance prey persistence and abundance. Theoretical Population Biology 62:231–49.
Rand, T. A., and S. M. Louda. 2006. Spillover of agriculturally subsidized predators as a potential threat to native insect herbivores in fragmented landscapes. Conservation Biology 20:1720–29.
Rand, T. A., and T. Tscharntke. 2007. Contrasting effects of natural habitat loss on generalist and specialist aphid natural enemies. Oikos 116:1353–62.
Rand, T. A., J. M. Tylianakis, and T. Tscharntke. 2006. Spillover edge effects: The dispersal of agriculturally subsidized insect natural enemies into adjacent natural habitats. Ecology Letters 9:603–14.
Ricklefs, R. E., and E. Bermingham. 2007. The causes of evolutionary radiations in archipelagoes: Passerine birds in the Lesser Antilles. American Naturalist 169:285–97.
Ritchie, M. E. 1999. Biodiversity and reduced extinction risks in spatially isolated rodent populations. Ecology Letters 2:11–13.
Robinson, G. R., R. D. Holt, M. S. Gaines, S. P. Hamburg, M. L. Johnson, H. S. Fitch, and E. A. Martinko. 1992. Diverse and contrasting effects of habitat fragmentation. Science 257:524–26.
Robinson, S. K., F. R. Thompson, T. M. Donovan, D. R. Whitehead, and J. Faaborg. 1995. Regional forest fragmentation and the nesting success of migratory birds. Science 267:1987–90.
Rosenzweig, M. L. 1995. Species Diversity in Space and Time. Cambridge: Cambridge University Press.
Russell, F. L., S. M. Louda, T. A. Rand, and S. D. Kachman. 2007. Variation in herbivore-mediated indirect effects of an invasive plant on a native plant. Ecology 88:413–23.
Ryall, K. L., and L. Fahrig. 2006. Response of predators to loss and fragmentation of prey habitat: A review of theory. Ecology 87: 1086–93.
Rydberg, W. A., and J. M. Chase. 2007. Predator-dependent species-area relationships. American Naturalist 170:636–42.
Salo, P., E. Korpimaki, P. B. Banks, M. Nordstrom, and C. R. Dickman. 2007. Alien predators are more dangerous than native predators to prey populations. Proceedings of the Royal Society of London, Series B 274:1237–43.
Schaffers, A. P., I. P. Raemakers, K. V. Sykora and C.J.F. ter Braak. 2008. Arthropod assemblages are best predicted by plant species composition. Ecology 89:782–94.
Schoener, T. W. 1976. The species-area relationship within archipelagos. In Proceedings of the 16th International Ornithological Congress (1974), ed. H. J. Frith and J. H. Calaby, 629–42. Canberra: Australian Academy of Science.
———. 1989. Food webs from the small to the large. Ecology 70:1559–89.
Schoener, T. W., and D. A. Spiller. 1995. Effect of predators and area on invasion: An experiment with island spiders. Science 267:1811–13.
———. 1996. Devastation of prey diversity by experimentally introduced predators in the field. Nature 381:691–94.
Schoener, T. W., and D. A. Spiller. 1999b. Variation in the magnitude of a predator’s effect from small to large islands. In Monografies de la societat d’historia natural de les balears 6: Ecologia de les illes, 35–66. Palma, Spain: Grafiques Miramar, S. A.
Schoener, T. W., D. A. Spiller, and J. B. Losos. 2001. Predators increase the risk of catastrophic extinction of prey populations. Nature 412:183–86.
Schoener, T. W., D. A. Spiller, and L. W. Morrison. 1995. Variation in the hymenopteran parasitoid fraction on Bahamian islands. Acta Oecologica 16:130–21.
Schreiber, S. J., R. N. Lipcius, R. D. Seitz, and W. C. Long. 2006. Dancing between the devil and deep blue sea: The stabilizing effect of enemy-free and victimless sinks. Oikos 113:67–81.
Shulman, R. S., and J. M. Chase. 2007. Increasing isolation reduces predator-prey species richness in aquatic food webs. Oikos 116:1581–87.
Siemann, E., J. Haarstad, and D. Tilman. 1999. Dynamics of plant and arthropod diversity during old field succession. Ecography 22:406–14.
Simberloff, D. S., and E. O. Wilson. 1969. Experimental zoogeography of islands: The colonization of empty islands. Ecology 50:278–96.
Snyder, W. E., G. B. Snyder, D. L. Finke, and C. S. Straub. 2006. Predator biodiversity strengthens herbivore suppression. Ecology Letters 9:789–96.
Sole, R. V., and J. Bascompte. 2006. Self-Organization in Complex Ecosystems. Princeton, NJ: Princeton University Press.
Sole, R. V., and J. M. Montoya. 2006. Ecological network meltdown from habitat loss and fragmentation. In Ecological Networks; Linking Structure to Dynamics in Food Webs, ed. M. Pascual and J. A. Dunne, 305–47. Oxford: Oxford University Press.
Spencer, M. 2000. Are predators rare? Oikos 89:115–22.
Spencer, M., L. Blaustein, and J. E. Cohen. 1999. Species richness and the proportion of predatory animal species in temporary pools: Relationships with habitat size and permanence. Ecology Letters 2:157–66.
Spencer, M., and P. H. Warren. 1996. The effects of habitat size and productivity on food web structure in small aquatic microcosms. Oikos 75:419–30.
Steiner, C. F., and M. A. Leibold. 2004. Cyclic assembly trajectories and scale-dependent productivity-diversity relationships. Ecology 85:107–13.
Steffan-Dewenter, I., and T. Tscharntke. 2000. Butterfly community structure in fragmented habitats. Ecology Letters 3:449–56.
———. 2002. Insect communities and biotic interactions on fragmented calcareous grasslands—a mini review. Biological Conservation 104:275–84.
Swihart, R. K., Z. Feng, N. A. Slade, D. M. Mason, and T. M. Gehring. 2001. Effects of habitat destruction and resource supplementation in a predator-prey metapopulation mode. Journal of Theoretical Biology 210:287–303.
Terborgh, J., L. Lopez, P. Nuñez, M. Rao, G. Shahabuddin, G. Orihuela, M. Riveros, R. Ascanio, G. H. Adler, T. D. Lambert, L. Balbas. 2001. Ecological meltdown in predator-free forest fragments. Science 294:1923–26.
Thies, C., I. Steffan-Dewenter, and T. Tscharntke. 2008. Interannual landscape changes influence plant-herbivore-parasitoid interactions. Agriculture, Ecosystems, and Environment 125:266–68.
Tscharntke, T., and R. Brandl. 2004. Plant-insect interactions in fragmented landscapes. Annual Review of Entomology 49:405–30.
Tscharntke, T., T. A. Rand, and F.J.J.A. Bianchi. 2005. The landscape context of trophic interactions: Insect spillover across the crop-noncrop interface. Annales Zoologici Fennici 42: 421–32.
Tscharntke, T., I. Steffan-Dewenter, A. Kruess, and C. Thies. 2002. Characteristics of insect populations on habitat fragments: A mini review. Ecological Research 17:229–39.
Valladores, G., A. Salvo, and L. Cagnolo. 2006. Habitat fragmentation effects on trophic processes of insect-plant food webs. Conservation Biology 20:212–17.
Valone, T. J., and M. R. Schutzenhofer. 2007. Reduced rodent biodiversity destabilizes plant populations. Ecology 88:26–31.
Van Nouhuys, S. 2005. Effects of habitat fragmentation at different trophic levels in insect communities. Annales Zoologici Fennici 42:433–47.
van Ruijven, J., G. B. De Deyn, C. E. Raaijmakers, F. Berendes, and W. H. van der Putten. 2005. Interactions between spatially separated herbivores indirectly alter plant diversity. Ecology Letters 8:30–37.
van Veen, F.J.F., P. D. Van Holland, and H.C.J. Godfray. 2005. Stable coexistence in insect communities due to density- and trait-mediated indirect effects. Ecology 86:1382–89.
Vos, M., S. Moreno Berrocal, F. Karamaouna, L. Hemerik, and L.E.M. Vet. 2001. Plant-mediated indirect effects and the persistence of parasitoid-herbivore communities. Ecology Letters 4:38–45.
Warren, P. H., R. Law, and A. J. Weatherby. 2003. Mapping the assembly of protist communities in microcosms. Ecology 84:1001–11.
Watling, J. I., and M. A. Donnelly. 2006. Fragments as islands: A synthesis of faunal responses to habitat patchiness. Conservation Biology 20:1016–25.
Williamson, M. 1981. Island Populations. Oxford: Oxford University Press.
Wilson, H. B., M. P. Hassell, and R. D. Holt. 1998. Persistence and area effects in a stochastic tritrophic model. American Naturalist 151:587–95.
Winemiller, K.O. 1990. Spatial and temporal variation in tropical fish trophic networks. Ecological Monographs 60:331–67.
Yodzis, P. 1998. Local trophodynamics and the interaction of marine mammals and fisheriers in the Benguela ecosystem. Journal of Animal Ecology 67:635–58.
