The niche concept revisited: mechanistic models and community context.
Although most textbooks emphasize the use of the niche concept in community ecology (e.g., Ricklefs 1979, Begon et al. 1990, Pianka 1994), the concept (or a near synonym) is used by ecologists working at most levels of ecological organization. Thus, physiological ecologists often work to identify environmental conditions (including factors other than resources) that affect an organism's performance (and implicitly, some component of its fitness). Similarly, much of population biology is concerned with identifying limiting factors of the environment that can alter the dynamics of populations (whether density dependent or not). Biogeographers consider how environmental conditions can constrain the distributions of taxa, and ecosystems ecologists seek ways of describing how the functional traits of taxa (whether species or functional groups) alter ecosystem processes or structure. The niche concept is not always used explicitly in theories at all these levels of organization, but it provides an important, if sometimes implicit, connection between these disparate fields of ecology that justifies its status as a fundamental ecological concept (Cherrett 1989, Real and Brown 1991). Despite its strong synthetic role and its crucial importance in community theory, the niche concept remains unclear: "most [ecologists] would agree that niche is a central concept of ecology, even though we do not know exactly what it means" (Real and Levin 1991).
In this paper, I re-examine the niche concept using "mechanistic" models of species interactions. I suggest that there are two closely related but potentially distinct types of trade-offs that affect species interactions related to the niche and I argue that each type of trade-off is implicitly related to the two different niche concepts. The first corresponds to Grinnell's (1917) and Hutchinson's (1957) concepts concerning environmental requirements of species, whereas the second is more closely related to Elton's (1927) and MacArthur and Levins' (1967) concepts concerning the short-term impacts of species on resource use. I conclude with the suggestion that the niche concept be modified to combine these two narrower concepts and I employ this broader concept to illustrate the link between current "mechanistic" community theory and conventional "niche theory."
A Historical Evaluation
Schoener (1989), Griesemer (1992), and Colwell (1992) have provided recent historical overviews of the niche concept that contrast with earlier views of Allee et al. (1949), Whittaker et al. (1973), and Hutchinson (1978). These discussions focus largely on distinguishing between "habitat" and "functional" aspects of the niche and on the importance given to the relation of different species' niches to each other (see also James et al. 1984). In the discussion below I will focus instead on the distinction between environmental requirements and environmental impacts of species to highlight the dichotomy between the responses of organisms to the environment and their effects upon it (see Goldberg 1990 for a related analysis of aspects of competition in plants). My conclusions are more similar to those of Whittaker et al. (1973), who seem to have anticipated many of the discomforts with the niche concept that motivated my analysis, but who did not focus on an explicit distinction between a species' requirements and its impacts.
When Grinnell (1917) described the niches of organisms, he emphasized the multitude of environmental requirements of species and ignored, or only mentioned in passing, the potential impacts on other organisms. Grinnell included all of the commonly listed environmental "limiting factors" in his formal descriptions of species' niches. These included microhabitats, abiotic factors, resources, and predators, and he emphasized the combination of physiological and behavioral adaptations that allow species to respond to such factors. In an era when field experiments were uncommon, he explicitly tried to identify conditions that allowed a given species to exist by examining the union set of conditions in all the environments where it was found. This approach seems entirely consistent with other population-based approaches, such as Andrewartha and Birch's (1954, 1984) "ecological web" and Shelford's "environmental complex" (1913), and it illustrates the great importance that ecologists have placed on describing such factors at least since the time of Darwin (1859). These concepts appear synonymous with Hutchinson's (1957, 1978) "fundamental niche," which he defined as a multi-dimensional "hyper-volume" describing the conditions where an organism's expected absolute fitness is at least zero, in a conceptual space whose axes include all of the environmental variables affecting that species. Hutchinson argued that this definition would "completely define its ecological properties" (Hutchinson 1957: 416), though it is easy to see ways in which this might not be so (e.g., the incidental impacts of large grazers on plant soil structure by compaction).
Elton used the niche concept in a distinctly different way to describe the effect that a given species has on the environment, sometimes including abiotic factors. Although Elton (1927) discussed "limiting factors" such as those described by Shelford (1913) and Grinnell (1917) at length, devoting an entire chapter (Chapter 2) of his book to them, that chapter does not discuss the niche concept. Instead, Elton devotes a section of a separate chapter on community relations (Chapter 3) to the niche. There, he describes the niche as an increasingly fine description of the "role" of a species in the community, beginning with its trophic level. It is hard to know if the primary importance he placed on trophic level reflected Elton's view that this was one of the more important ways to describe the resource requirements of a species, or if he viewed it as one of the more important ways to describe which aspects of the ecosystem it was most likely to affect. Elton's discussion of limiting factors and species requirements in his Chapter 2 shows that he appreciated the tremendous importance of abiotic factors on species, but his explicit negation of most of them in defining the niche of a species (Elton 1927: 64) suggests a focus on what he calls "roles" or "what they are doing." Though Elton may have meant to include species requirements in his definition of the niche (thus anticipating much of the following argument), I believe Elton's focus was on the impact of a species on the ecosystem for the following reasons:
1) The niche was "defined to a large extent by [a species'] size and food habits." Elton gave tremendous significance to body size as a component of the niche, not because it influenced requirements (as might have been suggested by reference to the effects of body size on thermoregulation, for example) but because body size influenced what resources could be consumed and what predators could be dangerous for a given organism.
2) Elton's niche concept was intimately related to the trophic pyramid of numbers, food chains, and food cycles, and patterns of relative body size between predators and prey describing interrelations among species. Elton specifically notes that niches are important because they "enable us to see how very different animal communities may resemble each other in the essentials of organization." This does not seem related to an interest in the general set of all factors that can allow a species to exist in different environments. It rather seems to reflect an interest in the similarity of impacts of species across community types rather than similarities in their requirements. Elton described species with very distinct requirements (e.g., Arctic foxes and African hyenas) as having the "same" niche because they ate similar things. It is hard to imagine Grinnell saying the same thing, though Grinnell did claim identity of niches between species like jerboas and kangaroo rats that presumably had very similar requirements as well as similar resources in different parts of the world.
3) Elton (1927: 64) specifically states that most niche relations do not involve abiotic factors. However, he later mentions examples wherein species have impacts on abiotic factors: the roles of earthworms and other burrowing organisms (including land-crabs) in affecting soil structure, and the roles of Sand Martins and Bee-eaters in digging out sand banks. He also discusses the role of holothurians in converting corals to "mud," which illustrates that "the coral-eating niche has a geological significance" (Elton 1927: 68).
Thus Elton's view of the niche was much more oriented toward describing how an organism affected the environment (primarily by consuming resources and serving as resources for higher trophic levels) than in describing what environmental factors affected the organism. I argue below that this view corresponds more strongly with MacArthur and Levins' use of the concept describing resource consumption than with either Hutchinson's definition or Grinnell's concept. Determining the difference between the two versions of the concept is difficult because resources must be consumed if they are needed to satisfy any "requirements." Furthermore, on a long-term quasi-equilibrium level, the full impact of a species on the environment depends on a complex interaction of both of these niche concepts with features of the environment as discussed below.
Hutchinson contributed to the confusion by using the concept ambiguously even in his seminal "Concluding Remarks" paper (Hutchinson 1957; see also Hutchinson 1978). Although Hutchinson used a definition of the niche explicitly focused on requirements, his use of the niche in discussing competition was much more compatible with a concept related to impacts and roles of species. In his definition of the fundamental niche, Hutchinson specifically included quantitative axes for every factor that could influence fitness. Thus there could be an axis quantifying temperature, the concentration of hydrogen ions (pH), etc. However when discussing resources, Hutchinson illustrated his point in a qualitatively different way. Instead of having an axis that described the levels or concentration of each potential resource (a natural extension of his use of the concept for abiotic factors), he used axes that served as quantitative descriptions of resources (e.g., food particle size, or distribution along a habitat gradient). This difference is important because information about resource levels was sacrificed in order to reduce the dimensionality of the problem by using a single axis to describe (actually to serve as an "index" for) a potentially large number of different resources. As discussed by Schoener (1989) and others (see Giller 1984), Hutchinson's approach created problems in understanding the association between resources and fitness and could not deal with resources that could not be described by simple quantitative axes. Considering that, without knowing their density, it is impossible to determine if resources are adequate to allow survival and reproduction, there is a severe deviation in his use of the concept from his original definition, an observation that is commonly made even by undergraduate students exposed to Hutchinson's paper (M. Leibold and E. Simms, personal observation). The analysis described below, however, shows that niche relations among species can be analyzed with an approach entirely consistent with Hutchinson's definition without resort to ignoring resource-level effects on fitness.
In taking an approach different than Hutchinson's, Maguire (1973) tried to re-describe the niche in terms of resource levels in a fashion identical to that described below. Hutchinson (1978) seems to have agreed with the compatibility of Maguire's approach with his niche concept (citing it as "doubtless the most important contribution to the concept of the niche since the death of Robert MacArthur"), but did not seem to recognize the significance of the distinct use of resource axes in the two approaches. The theoretical basis for this distinction initiated by Maguire (1973) combined with more recent theoretical work provides a framework for making this distinction explicit.
A Mechanistic Analysis of Resource Niche Relations
Following Hutchinson's work it was widely recognized (see Begon et al. 1990, Abrams 1992) that there is a distinction between resources (defined as depletable, potentially limiting, factors in the environment) and other environmental factors that limit populations (e.g., temperature, or disturbances). Approaches that explicitly consider the process of resource depletion have allowed ecological theory to move away from phenomenological models based on the highly "abstracted" (sensu Schaffer 1981) Lotka-Volterra models to more mechanistic models (e.g., MacArthur 1972, Schoener 1974, Tilman 1982). These explicit resource competition models provide an important insight into the distinctions that I believe distinguish the Grinnell/Hutchinson niche and the Elton/MacArthur-Levins niche. The essential elements of the competition model are described in detail by MacArthur (1970, 1972), Maguire (1973), and Tilman (1982) and summarized and compared by Naeem and Colwell (1991). Here I give an abbreviated description to illustrate the distinction between aspects of the biology of organisms that relate to their requirements and those that relate to their impacts on resources.
The interaction between resources and consumers depends in part on how resources affect fitness components of the consumers. This can be illustrated as shown in Fig. 1 by separating a resource-dependent component (hereafter referred to as the population growth rate, b) and a resource-independent component (hereafter referred to as the loss rate, d), recognizing that each of these components involves both reproduction and survival probabilities related to production and respiration processes. If resource levels are high enough for the growth rate to exceed the loss rate, the population will increase; if the resource levels lead to a growth rate lower than the loss rate, the population will decline. There is a set (only one point in Fig. 1, labeled [R.sup.*]) of conditions where the growth rate equals the loss rate.
Following Maguire (1973, see also Hutchinson 1978, Tilman 1982), assuming that other factors are held constant, and focusing on a pair of resources (whose densities are labeled [R.sub.1] and [R.sub.2] in Fig. 2), one can describe three sets of conditions for any consumer that depends on these two resources. If resources are too low (the clear area in Fig. 2) the consumer's density-independent fitness loss component (d) will exceed the resource-density-dependent growth rate (b). By Hutchinson's definition such conditions are "outside" the consumer's niche. In contrast, resource levels may be high enough (the hatched area in Fig. 2) that the growth rate exceeds the loss rate and the population can increase. By Hutchinson's definition, such conditions are "inside" the consumer's niche. A third set of conditions obtains when both fitness components are equal, since the organism will have an expected fitness of zero and will neither increase nor decrease. This set of conditions determines the "zero net growth isocline" or "ZNGI" (Tilman 1982) of the organism and serves to describe the "boundary conditions" that delineate the organism's Hutchinsonian niche. Though the derivation of the ZNGI does not depend on population equilibrium assumptions, MacArthur (1972), Maguire (1973), and Tilman (1982) all point out that long-term numerical responses will tend to lead to conditions on the "ZNGI" as density-dependent processes occur if the consumers deplete the resources.
However, resource-consumer relations described by the ZNGI line are not fundamentally different from relations involving any other (non-resource) factor, and the relationship between resource levels and fitness does not explicitly describe resource consumption. It is possible to have two species with identical ZNGIs but whose resource consumption rates are very different (see below) and vice versa.
Resource consumption can be described by a different set of relations illustrated by vectors relative to [R.sub.1] and [R.sub.2] as shown in Fig. 3 (see MacArthur 1972, Tilman 1982). The short-term impacts of a consumer on the two resources depend on the rates at which they consume each of the resources. The magnitude of separate vectors parallel to each resource axis can describe the expected rates at which each individual consumer consumes each resource. This can, and usually will, depend in a complex way on the abundances of the resources and will be affected by the functional responses of organisms. However, at any given set of resource densities, the net expected effect of each consumer can be described by adding the two orthogonal vectors to define an "impact vector" (Tilman refers to this as the "consumption vector" but I wish to generalize the concept below and to emphasize its role in discerning the effect of the consumer on the resources). An important point to make is that such vectors do not necessarily depend on consumer density, and that their derivation does not depend on any assumptions about population equilibria.
The relationship between impact vectors and ZNGIs depends strongly on the efficiency of the consumer in converting resources to per capita growth. Thus, two consumers might have similar ZNGIs despite having very different impact vectors if they also differ in the relative efficiency with which they convert each resource to growth. In the simplest linear model with substitutive resources (sensu Tilman 1982) comparing two species with identical loss rates and identical magnitude (but not slope) of impact vectors, a pair of species would have the same ZNGI if the ratio of their consumption rates is equal to the inverse of the ratio of their efficiencies on the two resources. The long-term population-level impact of the organism depends on an interaction between the numerical response of the organism (which depends on its ZNGI) and its per capita impact vector at equilibrium.
Similarly, the basic principles described for competition for two resources can also be used to examine interaction involving a single resource and a predator on an intermediate consumer species (Holt et al. 1994, Leibold, in press). By separating fitness components into those that depend on resources and predators from those that are density independent, we can plot the combination of resource and predator densities that result in an expected fitness [less than]0 (population declines; the open area of Fig. 4), [greater than]0 (population growth; the hatched area in Fig. 4) and fitness =0 (the ZNGI line shown in Fig. 4). These conditions describe when resource levels are high enough to mitigate the negative effects (either involving direct mortality or reduced performance of the species due to its facultative responses to predators) of predators on the expected fitness of the organism. Again, the hatched area satisfies Hutchinson's definition of the niche but does not directly describe how organisms affect their prey (via consumption) nor their predators (via serving as food). The ZNGIs describe the environmental requirements (and constraints) but do not describe the ecological impacts on the predator and resource except those that are eventually expected to occur via long-term numerical responses if the species can modify predator densities, resource levels, or both.
As before, the short-term ecological impact of the species can be described using vector notation (Leibold, in press) to partition the expected per capita effects of the organism into the effect on its resource and the effect on its predator. The vector component describing the organism's impact on its resource is identical to the one used above to describe competition for two resources. The impact on predators is described by a vector parallel to the predator axis whose magnitude is equal to the product of the death rate imposed by the predator on the organism and the predator's conversion efficiency (i.e., the per capita contribution of the organism to the predator's population growth rate). The net per capita effect of the organism on the environment (the resource and the predator) is described by the sum of the component vectors, which defines an "impact vector" (labeled C in Fig. 5). In contrast to the model for two resources where impact vectors point toward the lower left (negative impacts on both resources), here impact vectors will generally point to the upper left (positive impact on predators and negative ones on resources). Again, the derivations of the ZNGI and the impact vector do not depend on population equilibrium assumptions.
As before, species can have similar ZNGIs yet have distinct impact vectors. For example, consider a case where two species have identical minimum resource requirements in the absence of predators and are equally assimilated by the predator and have linear functional responses. Such species will have impact vectors whose slope is their mortality risk to predators divided by the feeding rates on their own resource. Species could have identical ZNGIs despite having different such vectors if the ratio of their relative assimilation efficiencies on their own food is equal to the ratio of the slopes of their impact vectors.
Again, though ZNGIs and impact vectors are related to each other, they describe two potentially distinct aspects of the relationship of an organism with external environmental factors. ZNGIs correspond closely to Hutchinson's niche concept and to Grinnell's description of biotic factors that affect the fitness of organisms (including predators), whereas impact vectors reflect the roles that species play in the community by culling resources and contributing to the support of higher trophic levels as discussed by Elton. In this case the distinction is even more striking by its relevance to food chains and the "pyramid of numbers."
Coexistence and the Niche
Regardless of which concept is involved, the niche has always been closely associated with "Gause's axiom," that two species cannot coexist if they share a single niche (Gause 1934). The theoretical models described above explicitly examine conditions for coexistence, and they identify two different criteria related to the niche. Furthermore, both aspects of the niche that can be identified by using these mechanistic models are determined by the expected (i.e., average) parameters of individuals without reference to population equilibria (though of course the prediction itself is only expected to hold if populations have come to near-equilibrium conditions): First the expected environmental requirements for an individual to complete its life cycle and replace itself, and second, the expected per capita impact of individuals on other organisms.
I argue that these two different criteria are identified with two different concepts of the niche; one associated with Grinnell (1917) and Hutchinson (1957, 1978), and the other associated with Elton (1927) and MacArthur and Levins (1967). In models of competition for two resources (e.g., MacArthur 1972, Tilman 1982) and in the model involving interactions among two species that share a common resource and predator (Holt et al. 1994), coexistence of two species requires that their ZNGIs cross. These aspects of the models describe conditions for the EXISTENCE of an equilibrium point allowing coexistence of species. Ignoring the "infinitely small" possibility that different species will have identical ZNGIs, this suggests that species must have different ZNGIs to coexist and that these differences must further be of the type where each species has higher fitness under some environmental conditions. For example, in the two-resource model each resource must benefit one species more than the other. Likewise, in the keystone-predator model, one species must be less affected by the predator and the other must be a better resource exploiter (sensu Tilman 1982) in the absence of predators. The existence of this equilibrium point does not depend on the impact vectors (except indirectly due to shared parameters common to both ZNGIs and impact vectors).
Given the existence of an equilibrium point as described above, however, the stability of the point depends critically on the impact vectors (Tilman 1982, Leibold, in press). In the case of competition for two resources, stability requires that each species consumes proportionately more of the resource that most strongly limits its growth (Tilman 1982). In the case of competition mediated by a keystone predator, stability requires that the species most strongly affected by the predator (i.e., the "resource-exploitation specialist") contribute proportionately more to the predator's growth relative to its feeding rate (M. A. Lei-bold, unpublished manuscript).
Thus the joint criteria for a stable equilibrium requires a linked set of trade-offs between species; one set must reflect a trade-off in the requirements of species, and the other must reflect a corresponding trade-off in the impacts of species. Of course, because both aspects are affected by some common parameters (direct mortality and feeding rates), both conditions for a stable equilibrium can be satisfied by trade-offs involving a single variable (relative feeding rates in the case of competition for two resources and the "impact ratio" described above for the keystone predation model). However, an important point is that trade-offs need not necessarily involve a single variable because such trade-offs could also involve assimilation and loss rate parameters.
The two examples discussed above (competition for two resources and competition mediated by a keystone predator) are only two of a wide array of reciprocally negative species interactions that can result from many different mechanisms. These examples only illustrate some of the points that characterized Grinnell's and Elton's views about niches and serve as a starting point for a similar analysis of other mechanisms. To date models of species interactions such as those described above are best understood in closed (i.e., self-sustaining local populations), homogenous (no patchiness or disturbance) situations. They can, however, serve as the basis for more complex situations, as illustrated by the work on environmental heterogeneity by Levins (1979), Chesson (1986), and Tilman (1982; see also Naeem and Colwell 1991), which illustrate the increased dimensionality of the problem when compared to the simple cases described above.
As has been often noted (e.g., Levins 1968, Tilman 1986), virtually all of the mechanisms proposed to explain long-term coexistence of species can be related to trade-offs in some components of their biology or of their "niches" (Chesson 1991). It is not yet clear how many different mechanistic models of species interactions can be analyzed in a way that separates the conditions for existence from those that determine stability of putative equilibria into the respective components of the niche concept involving requirements and impacts.
Short-term vs. Long-term Impacts of Species
One of the attractive features of the "mechanistic" approach is that biological parameters are described at the individual level (or at the level of the entire life history of individuals). Separating "requirement" and "per capita impact" components of the niche are entirely consistent with this "individual-based" approach, and the concepts do not depend on making equilibrium assumptions about the component populations of interacting species (including resources, competitors or predators). However, on a long-term basis the total impacts of a species depend on its population size, and the most predictable aspect of such long-term impacts are those that occur at the population equilibrium. In both of the models described above (involving competition for shared resources and predator-mediated competition) the long-term, equilibrium, impact of species depends more closely on their requirements than on their per-capita impacts (see Tilman 1982, Holt et al. 1994). Thus Tilman's (1982) conclusion that species with the highest competitive ability should be those with the lowest resource requirements (the "[R.sup.*] rule"), and Holt et al.'s (1994) analysis of apparent-competition ability (with an analogous "[P.sup.*]" rule) illustrates that on a population level that assumes an equilibrium, impact vectors do not play as important a role as do ZNGIs. However, a focus on describing relative competitive ability that ignores impact vectors cannot elucidate the important roles that species play in the dynamic aspects of community stability via their relative abilities to affect the environment nor can they describe the roles that species play in regulating turnover-rates of their resources or in supporting populations of their predators (see Tilman 1982, Holt et al. 1994). Furthermore, an approach focused on long-term impacts (e.g., using ZNGIs only to predict competitive ability) that ignores impact vectors cannot help make predictions about which species can coexist since it cannot predict which combinations allow for stable coexistence.
Relation of the Mechanistic Approach to Conventional "Niche Theory"
Which of the two aspects of the regulation of coexistence is included in conventional niche theoretical models? Modern "Niche Theory" focuses on examining niche relations with respect to resource consumption (see Giller 1984, for a recent review). Despite the strong personal connection between Hutchinson and MacArthur as they developed current niche theory, and acknowledging Hutchinson's ambiguous use of the concept with respect to his formal definition, I argue that modern niche theory most closely relates to impact vectors and has little to do with resource requirements (despite the observation that some definitions of competitive ability depend more on requirements than on per capita impacts as described above!).
Modern niche theory owes much of its development to the idea, proposed by MacArthur and Levins (1967), showing that Lotka-Volterra competition coefficients could be equated with diet overlap of competitors. Although subsequent work (see Giller 1984), especially by MacArthur (1970, 1972), Schoener (1974), and Abrams (1975), has shown that this is not strictly true under more complex situations, the general result has continued to be used as a practical "rule of thumb" (Chesson 1990). There has subsequently developed a large literature deriving quantitative metrics of "niche overlap" and "niche breadth" (see Petraitis 1979 for a review) and studying resource use overlap in natural communities (e.g., Cody 1968, Pianka 1973). There are two ways of looking at this body of work to see that they describe aspects of niche relations related to environmental impacts rather than to environmental requirements.
First, as shown by Vandemeer (1975), Lotka-Volterra competition coefficients are not sufficient to determine whether two species will coexist. Instead they determine whether the equilibrium point will be stable given that it exists (with both species having positive densities). Thus meeting the conditions of "permissible" niche overlap is equivalent to determining the stability of possible equilibrium points if MacArthur and Levins' (1967) isomorphism between overlap and competition strength applies. Vandermeer's result shows that permissible niche overlap is necessary for a stable equilibrium but does not guarantee the existence of such an equilibrium point (see also Persson 1985).
More to the point, Petraitis (1989) has shown that the cosine of the difference between the angles of impact vectors of two species in Tilman's (1982) competition model is equivalent to one of the specific measures of resource overlap based on the "conventional" niche model (Pianka 1973). As discussed above, impact vectors need not have a direct relation with resource requirements, because they are based on use without respect to their effects on fitness. Rather, conclusions about coexistence based on differences between diets (conventional niche theory) are equivalent to differences in per capita impacts (graphical mechanistic models) on resources. This result shows that the overlap indices of conventional niche theory do not describe the expected long-term effect of organisms on the densities of their resources (i.e., their effects on equilibrium resource densities as modeled by the intersection of ZNGIs using Tilman's terminology), and thus do not fully describe the process of interspecific competition but only describe the short-term relations in relative impacts (impact vectors only). Such confusion about competitive abilities and the importance of short-term vs. long-term effects underlies a number of current debates in ecology (e.g., disagreements between Tilman's and Grime's views of "competitive ability," see Grace 1990, and Goldberg 1990 for recent reviews).
What is Hutchinson's "realized niche"? Hutchinson (1957) introduced the "realized niche" as a concept to identify those types of niche axes that were influenced by competitors (i.e., primarily resources) and to describe how interspecific competition acted to limit populations. The notion was to determine how an environment satisfied the requirements of an organism before (the "fundamental niche" concept) and after (the "realized niche" concept) accounting for the effects of competitors. Further, Hutchinson suggested that environmental conditions describing overlap in the fundamental niches of different species could be partitioned between them to quantify the realized niche. As was pointed out by Schoener (1989), there were a number of conceptual problems associated with Hutchinson's idea of such partitioning. However, mechanistic models can be used to illustrate the basic idea.
Using Tilman's (1982) formulation of resource competition for substitutable resources we can imagine defining species niches on the basis of the Supply Point of the resources (i.e., resource densities in the absence of consumers, before accounting for the effects of competitors) as shown in Fig. 6. The only difference in the notion of the "niche" between this figure and Fig. 2 discussed above is the definition of the axes; here the axes are supply points (what resource levels would be if consumers were absent), whereas the axes in Fig. 2 are resource densities (actual densities experienced by the consumers at a given point in time), as discussed extensively by Tilman (1982). Different possible environments are classified as having supply points falling in various parts of the graph labeled with Roman numerals. As discussed by Tilman (1982), species are excluded from some such environments by interspecific competition (In Fig. 6, species a is excluded from environments whose supply point lies in region III, and species b is excluded from those found in region V). Further, species can coexist only under conditions, determined by their consumption (i.e., impact) vectors, whose supply points are found in region IV. This separation of environmental conditions that determine where species exclude each other is a way of partitioning of the "environmental space"; the relative densities of the species in the area allowing coexistence in region IV could serve as weighting factors to further describe the partitioning of the entire graph. Again, partitioning the environmental space into the different realized niches depends on the consumption vectors much more than on the ZNGIs.
A confusing issue is that the overall competitive effect of the competing species on resources is described by the single point that describes the equilibrium outcome of the set of interactions (the intersection of the ZNGIs if the species coexist). Thus this concept of the "realized niche" (defined as a subset of the original supply points) cannot be quantified from simple observations of natural communities because the position of the original supply points cannot be reconstructed without conducting an experiment (eliminating the consumers). However this concept serves to show that the minimum resource requirements play a crucial role in determining which species will win (equivalent to Tilman's "[R.sup.*] rule"; coexistence depends on which species have the lowest minimum resource requirements; Tilman 1982). A similar argument can be made to evaluate interactions among species that share predators, and species that share both resources and predators (see Holt et al. 1994).
In an apparent attempt to solve some of these problems with Hutchinson's niche concept and to make the concept operational, MacArthur and Levins (1967) redefined both the fundamental niche and the realized niche in terms of resource utilization (see also Schoener 1989). A multitude of other papers have since followed that collectively form "conventional niche theory" (reviewed by Giller 1984). Though much of the distinction between "requirements" and "impacts" outlined here was discussed by MacArthur (1967, 1970, 1972), most "niche theory" has focused on relative resource utilization without explicit reference to resource requirements (reviewed by Colwell and Fuentes 1975 and by Giller 1984).
In "conventional niche theory," resources are made available and are consumed by organisms in proportions that depend on the occurrence (and density) of other species. Thus competition describes the partitioning of resources as they become available (a rate described by resource renewal rather than a density measure of availability), into different consumer populations. The significance of Petraitis' (1989) result described above is that it shows that this approach is most relevant to the consumption ratios of species, and that despite a superficial similarity to Hutchinson's description of realized niches described above, it is not necessarily closely related to minimum food requirements.
The significance of the recent focus on "mechanistic models" (Price 1986, Schoener 1986, Tilman 1988) is that they allow results of community-level models to be defined on the basis of parameters that distinguish between the expected (i.e., average) properties of individuals (e.g., feeding rates, assimilation efficiency, and predator vulnerability) from those that are properties of the environment (e.g., maximum resource levels, disturbance level, and abiotic stress). In contrast, much conventional niche theory used a more phenomenological approach (based on Lotka-Volterra competition models) to evaluate to dynamic properties (i.e., stability properties) of competition. The two approaches have sometimes appeared to be distinct and somewhat disparate. As implied in the work of MacArthur (1970, 1972, see also Chesson 1990), I believe a synthesis is possible based on the joint consideration of Hutchinson's definition of the niche and Elton's view of the concept as follows:
1) Environmental requirements do not necessarily correspond to the per capita impacts of species on the environment (i.e., there is not a necessary one-to-one correspondence between ZNGIs and impact vectors).
2) The requirements of different species determine whether coexistence is possible, and coexistence can only occur if there is a trade-off in species' requirements. Such relationships also determine which species combinations are most likely to coexist.
3) Whether coexistence will be achieved (because it is stable) in a given environment depends on the relationship among the impacts on the environment of the species that have different requirements. Each species must have a differential effect on the factor that most limits its growth for coexistence to be stable.
4) Aspects of the niche related to environmental requirements of species most closely correspond to Grinnell's use (1917) and Hutchinson's (1957) definition of the concept as shown by Hutchinson's acceptance and discussion of Maguire's (1973) interpretation.
5) Aspects of the niche related to environmental impacts relate more closely to Elton's (1927) concept. They also correspond strongly to MacArthur and Levins (1967) development of "conventional niche theory" as shown by Petraitis' (1989) demonstration of an isomorphism between niche overlap (and breadth) measures and relations among Tilman's (1982) consumption vectors.
6) A re-interpretation of the niche concept as the union of these two niche concepts can allow greater mechanistic insights into factors that determine the outcome of species interactions as shown by the correspondence in the analysis of niche relations in models of two-resource competition and of keystone-predator mediated competition for a single resource.
The niche concept has always been meant to be general (and perhaps intractably so). Grinnell, Elton, Hutchinson, and MacArthur and Levins all conceived that it was meant to summarize either "many" or "uncountable" aspects of the biology of organisms. Distinguishing between the two different views of the niche concept clarifies it by identifying a different theoretical consequence for each component (existence of an equilibrium vs. the stability of the equilibrium). Further, it is clear that the two aspects are closely (though not necessarily) linked in many situations by sharing common parameters; in the case of trophic relations these are feeding rates and mortality risks to predators. I suggest that the term "requirement niche" be used to describe requirements (most closely corresponding to Hutchinson's definition) and that the term "impact niche" be used to describe the per capita effects of species on their environments (corresponding to Elton's "roles"). The two aspects combine to form the "total niche" in its broadest sense. In much "conventional niche theory," aspects related to impacts of species have been emphasized as the ones that regulate coexistence by affecting the stability properties of the interactions, but such considerations are irrelevant if the conditions that permit coexistence (regardless of the stability features of the equilibrium) do not apply: Tilman's (1982, 1988) "[R.sup.*]" rule-of-thumb must apply before the question of stability related to the "overlap" rule-of-thumb can even be considered.
My primary goal has been to highlight these results with the hope of clarifying the niche concept and elucidating its relationship to mechanistic models of community interactions. Clarifying this relationship also allows a much closer linking of individual ecology (involving physiology and behavior) to community models without generating "abstracted parameters" such as "competition coefficients" that are mixed properties of populations, individuals, and environments (discussed by Shaffer 1981). Though similar arguments related to resource competition were made by MacArthur (1967, 1972), these have subsequently received little attention in the context of "niche theory" and have not been extended for use with other types of species interactions (e.g., predator/prey interactions). Such a mechanistic approach is also more compatible with the current emphasis on experimental methods that combine ecological insights at the individual, population, and community levels (see Diamond 1986).
Use of the niche concept has experienced a notable decline in recent years (Schoener 1989, Colwell 1992). I believe that a major reason for this involves confusion about its role in generating testable hypotheses and its ability to synthesize basic ecological principles from different levels of ecological organization (individuals, populations, communities, and ecosystems). Chesson's (1991) call for a "Need for niches" on the basis of trade-offs that can determine coexistence and its stability, as described above, is an important way to avoid sacrificing such an important concept to the vicissitudes of increasing reductionism without compromising a strong mechanistic framework.
E. Simms, M. Wade, and E. Werner provided invaluable comments on early versions of this manuscript. I also thank J. Brown, D. Goldberg, R. Holt, M. McPeek, and T. Schoener for stimulating ideas and valuable criticisms.
Abrams, P. A. 1975. Limiting similarity and the form of the competition coefficient. Theoretical Population Biology 8: 356-375.
-----. 1992. Resource. Pages 282-285 in E. Fox-Keller and E. A. Lloyd, editors. Keywords in evolutionary biology. Harvard University Press, Cambridge, Massachusetts, USA.
Allee, W. C., A. E. Emerson, O. Park, T. Park, and K. P. Schmidt. 1949. Principles of animal ecology. Saunders, Philadelphia, Pennsylvania, USA.
Andrewartha, H. G., and L. C. Birch. 1954. The distribution and abundance of animals. University of Chicago Press, Chicago, Illinois, USA.
Andrewartha, H. G., and L. C. Birch. 1984. The ecological web: more on the distribution and abundance of animals. University of Chicago Press, Chicago, Illinois, USA.
Begon, M., J. L. Harper, and C. R. Townsend. 1990. Ecology: individuals, populations and communities. Blackwell Scientific, Boston, Massachusetts, USA.
Cherrett, J. M. 1989. Key concepts: the results of a survey of our members' opinions. Pages 1-16 in J. M. Cherrett, editor. Ecological concepts: the contribution of ecology to an understanding of the natural world. Blackwell Scientific, Oxford, England.
Chesson, P. 1986. Environmental variation and the coexistence of species. Pages 240-256 in J. Diamond and T. J. Case, editors. Community ecology. Harper and Row, New York, New York, USA
-----. 1990. MacArthur's consumer-resource model. Theoretical Population Biololgy 37:26-38.
-----. 1991. A need for niches? Trends in Ecology and Evolution 6:26-28.
Cody, M. L. 1968. On the methods of resource division in grassland bird communities. American Naturalist 102: 107-148.
Colwell, R. K. 1992. Niche: a bifurcation in the conceptual lineage of the term. Pages 241-248 in E. Fox-Keller and E. A. Lloyd, editors. Keywords in evolutionary biology. Harvard University Press, Cambridge, Massachusetts, USA.
Colwell, R. K., and E. R. Fuentes. 1975. Experimental studies of the niche. Annual Review of Ecology and Systematics 6:281-310.
Darwin, C. 1859. On the origin of species by means of natural selection, or the preservation of favoured races in the struggle for life. J. Murray, London, England.
Diamond, J. 1986. Overview: laboratory experiments, field experiment and natural experiments. Pages 3-22 in J. Diamond and T. J. Case, editors. Community ecology. Harper and Row, Cambridge, Massachusetts, USA.
Elton, C. 1927. Animal ecology. Sidgwick and Jackson, London, England.
Gause, G. F. 1934. The struggle for existence. Reprinted 1969. Hafner, New York, New York, USA.
Giller, P. J. 1984. Community structure and the niche. Chapman and Hall, London, England.
Goldberg, D. E. 1990. Components of resource competition in plant communities. Pages 27-50 in J. B. Grace and D. Tilman, editors. Perspectives on plant competition. Academic Press, San Diego, California, USA.
Grace, J. B. 1990. On the relation between plant traits and competitive ability. Pages 51-66 in J. B. Grace and D. Tilman, editors. Perspectives on plant competition. Academic Press, San Diego, California, USA.
Griesemer, J. R. 1992. Niche: historical perspectives. Pages 231-240 in E. Fox-Keller and E. A. Lloyd, editors. Keywords in evolutionary biology. Harvard University Press, Cambridge, Massachusetts, USA.
Grinnell, J. 1917. The niche-relations of the California Thrasher. Auk 34:427-433.
Holt, R. D., J. P. Grover, and D. Tilman. 1994. Simple rules for interspecific dominance in systems with exploitative and apparent competition. American Naturalist 144:741-771.
Hutchinson, G. E. 1957. Concluding remarks. Cold Spring Harbor Symposia on Quantitative Biology 22:415-427.
-----. 1978. An introduction to population ecology. Yale University Press, New Haven, Connecticut, USA.
James, F. C., R. F. Johnston, N. O. Wamer, G. J. Niemi, and W. J. Boecklen. 1984. The Grinellian niche of the wood thrush. American Naturalist 124: 17-30.
Leibold, M. A. In press. A graphical model of keystone predators in food webs: trophic regulation of abundance, incidence and diversity patterns in communities. American Naturalist.
Levins, R. 1968. Evolution in changing environments. Princeton University Press, Princeton, New Jersey, USA. 1979. Coexistence in a variable environment. American Naturalist 114:765-783.
MacArthur, R. H. 1958. Population ecology of some warblers of northeastern coniferous forests. Ecology 39:599-619.
-----. 1967. The theory of the niche. Pages 159-176 in R. C. Lewontin, editor. Population biology and evolution. Syracuse University Press, Syracuse, New York, USA.
-----. 1970. Species packing and competitive equilibria for many species. Theoretical Population Biology 1:1-11.
-----. 1972. Geographical ecology: patterns in the distribution of species. Harper and Row, New York, New York, USA.
MacArthur, R. H., and R. Levins. 1967. The limiting similarity, convergence and divergence of coexisting species. American Naturalist 101:377-385.
Maguire, B. 1973. Niche response structure and the analytical potentials of its relationship to the habitat. American Naturalist 107:213-246.
Naeem, S., and R. K. Colwell. 1991. Ecological consequences of heterogeneity of consumable resources. Pages 224-255 in J. Koasa and S. T. A. Pickett, editors. Ecological heterogeneity. Springer-Verlag, New York, New York, USA.
Persson, L. 1985. Assymetrical competition: are larger animals competitively superior? American Naturalist 126: 261-266.
Petraitis, P. S. 1979. Likelihood measures of niche breadth and overlap. Ecology 60:703-710.
-----. 1989. The representation of niche breadth and overlap on Tilman's consumer-resource graphs. Oikos 56:289-292.
Pianka, E. R. 1973. The structure of lizard communities. Annual Review of Ecology and Systematics 4:53-74.
-----. 1994. Evolutionary ecology. Fifth edition. Harper-Collins College Publishers, New York, New York, USA. Price, M. V. 1986. Introduction to the symposium: mechanistic approaches to the study of natural communities. American Zoologist 26:3-4.
Real, L. A., and J. H. Brown. 1991. Preface. Pages xii-xiv in L. A. Real and J. H. Brown, editors. Foundations of ecology: classic papers with commentaries. University of Chicago Press, Chicago, Illinois, USA.
Real, L. A., and S. A. Levin. 1991. Theoretical advances: the role of theory in the rise of modern ecology. Pages 177-191 in L. A. Real and J. H. Brown, editors. Foundations of ecology: classic papers with commentaries. University of Chicago Press, Chicago, Illinois, USA.
Ricklefs, R. E. 1979. Ecology. Second edition. Chiron, New York, New York, USA.
Schaffer, W. M. 1981. Ecological abstraction: the consequences of reduced dimensionality in ecological models. Ecological Monographs 51:383-401.
Schoener, T. W. 1974. Some methods for calculating competition coefficients from resource utilization spectra. American Naturalist 108:332-340.
-----. 1986. Mechanistic approaches to community ecology: a new reductionism. American Zoologist 26:81-106.
-----. 1989. The ecological niche. Pages 79-114 in J. M. Cherrett, editor. Ecological concepts: the contribution of ecology to an understanding of the natural world. Blackwell Scientific, Oxford, England.
Shelford, V. E. 1913. Animal communities in temperate America. University of Chicago Press, Chicago, Illinois, USA.
Tilman, D. 1982. Resource competition and community structure. Princeton University Press, Princeton, New Jersey, USA.
-----. 1988. Plant strategies and the dynamics and structure of plant communities. Princeton University Press, Princeton, New Jersey, USA.
Vandemeer, J. H. 1975. Interspecific competition: a new approach to the classical theory. Science 188:253-255.
Whittaker, R. H., S. A. Levin, and R. B. Root. 1973. Niche, habitat and ecotope. American Naturalist 107:321-338.
|Printer friendly Cite/link Email Feedback|
|Author:||Leibold, Mathew A.|
|Date:||Jul 1, 1995|
|Previous Article:||Mammals as Predators: The Proceedings of a Symposium held by the Zoological Society of London and the Mammal Society, London, 22nd and 23rd November...|
|Next Article:||Carbon isotope dynamics during grass decomposition and soil organic matter formation.|