Explain how resource partitioning can promote long-term coexistence of competing species

Transitivity/intransitivity in competitive assemblages

While species coexistence is generally viewed as one potential outcome between two competing species, numerous community members may be involved in this process. For example, higher-order interactions among a dominant competitor and multiple weaker competitors (i.e., transitive interactions) may facilitate increased population growth of rare competitors. Experiments with three Drosophila species (one superior and two inferior competitors) and a shared predator showed that the dominant competitor declined due to frequency-dependent predation, thereby allowing the weaker competitors to increase in abundance (Worthen, 1989). While these experiments did not explicitly test the invasibility criterion, their results suggest that the role of predator-mediated interactions in insect coexistence warrants further investigation. This may merit special consideration given widespread shifts in the occurrence and distributions of predators in response to global change. Lastly, it may be that competing species have discrete competitive advantages and disadvantages (i.e., intransitive interactions), such that each species is a superior and inferior competitor relative to different competing species. However, these complex species interactions are challenging to document in nature and have only been theoretically demonstrated for nectivorous ants (reviewed in Siepielski et al., 2018).

Mechanisms of Coexistence

The previous discussion emphasized constraints on species coexistence arising from interspecific interactions. Rules of dominance are important conceptual tools that quantify these constraints and help identify biological traits leading to dominance. However, even in simple microcosms, coexistence can occur, and most natural communities are rich in species. Species in principle may coexist when any of the assumptions leading to the competitive exclusion principle are violated. This suggests three classes of mechanisms promoting species coexistence of potentially competing species in a local community:

Species may coexist in a closed, temporally constant world if they experience different limiting factors; this includes classical niche partitioning of resources, as well as mechanisms involving predation and parasitism, and direct interference.

Species may coexist, even though they experience the same limiting factor, if the environment is temporally variable and species respond differently to this temporal variation (temporal niche partitioning).

Species may coexist if the environment is spatially open or interactions are localized; the implications of space for coexistence can include spatial niche partitioning at scales broader than the local community, mechanisms such as colonization–competition tradeoffs in metapopulations, and microscale habitat partitioning.

From the 1950s to the mid-1970s, stimulated largely by G.E. Hutchinson and his brilliant student Robert MacArthur, most community ecologists emphasized classical niche partitioning in studies of species coexistence. In recent years, the balance of attention has shifted markedly to a broader range of coexistence mechanisms. Ecologists now believe that maintenance of diversity – coexistence writ large – often depends on spatial dynamics in open communities, food web interactions (including predation and parasitism), and nonequilibrial dynamics reflecting either extrinsic temporal variation or the endogenous instability of complex ecological systems. Moreover, as noted above, some ecologists have concluded that in natural communities, many species are effectively equivalent and can co-occur, without necessarily permanently coexisting, in the sense of tending to increase when rare.

IV.B. Temporal Heterogeneity

As resources vary in both space and time, species could partition their resource environment temporally as well as spatially. Temporal heterogeneity in resource availability is often not considered in discussions of maintenance of diversity, but could prove central to our understanding of species coexistence. Differences in the timing of plant developmental events, such as germination, flowering, and fruit drop, could play a critical role in mechanisms of species coexistence. In some communities that have existed for a long time, there is evidence for clear separation in flowering between groups, e.g., Peter Ashton's work on staggered mast-flowering in the Dipterocarpaceae in the tropical rain forests of Southeast Asia. In the tallgrass prairie of the American Midwest (communities that have a long evolutionary history and great species diversity), we have demonstrated clearly how these species separate their life history events along the entire growing season, i.e., phenological separation. One can recognize three fairly distinct groups of flowering species (Fig. 5). Also, in these grasslands one can observe the separation of the phenology of the introduced species, Poa pratensis. Although it is now considered by many investigators to be part of the tallgrass prairie, it is an introduced species (from European grasslands) that grows well in the cool portion of the season. Thus, it is able to grow at two distinct times of the year: before and after the main growing season of its competitors, the native, warm-season grasses of the community (Fig. 6). In this way, this species separates its growth activity from that of the native species and reduces competition for critical resources.

This kind of phenological separation is also evident in some early successional communities where species generally have broad niches (see earlier discussion in Section IV.A). In old-field communities in the American Midwest, Chenopodium album delays its growth and reproduction much later than all other species in the community (Bazzaz, 1996), thereby reducing competition during the main portion of the growing season. In this case, it is not clear that competition has shaped these phenological responses, as species in this early successional community have not necessarily been present together over evolutionary time. Many arguments about coexistence assume that competition has shaped the niches of co-occurring species in the past (see earlier discussion). However, in communities of unknown history, the question remains whether a species has modified its phenology by competition with the native species or was preadapted to function in these communities. An example of such preadaptation is found in the shifted daily flowering behavior of the native Erigeron annuus and the introduced Lactuca scariola. In winter annual communities of the midwestern United States, Lactuca opens its flowers in the morning hours 7:00 to 10:00 a.m. and requires a large number of pollinators. After 10:00 a.m., however, Erigeron starts to open and Lactuca closes its flowers (Fig. 7). As a result, the pollinators shift their foraging to Erigeron. Thus, while both species use the same pollinators, the daily separation in the opening of the flowers ensures that both are pollinated. At first glance, this may appear to be a convincing example of niche differentiation. However, when examined carefully, one finds that the introduced Lactuca developed this flowering behavior in its native habitat in Europe and not by competition for pollinators in the successional fields of the Midwest.

Temporal differences in species' traits may extend to very early stages of the life cycle. The formation of seed banks that can persist in the soil for a long time may also lead to maintenance of species diversity. A persistent seed bank may buffer species' responses to short-term variation in environmental conditions and prevent irreversible population crashes. Weedy species are known to form persistent seed banks. In good years for a particular species, many seeds are produced and enter the soil seed bank. In bad years, in contrast, few seeds are produced. The species can, therefore, partition the time axis to contribute the most to future generations in years when conditions are favorable for growth and reproduction. The development of this sort of persistent seed bank has been mathematically modeled and is often called the storage effect (see work by Peter Chesson). Coexistence is promoted if species have differing sensitivities to temporal variation in environmental conditions.

Further niche axes may potentially be created through the combination of temporal heterogeneity in multiple resources (see discussion in Section III). Because plant responses to a particular resource are contingent on prevailing local environmental conditions, e.g., temperature and relative humidity, the time course of resource availability patterns could critically determine a plant's performance. If resource availability is high when other conditions are favorable (i.e., resources are congruent), then plants may be able to make good use of that resource. If, however, conditions are unfavorable and plant activity is inhibited to some extent, then plants may not be able to take full advantage of the resource in ample supply. Species might have different levels of tolerance for the extent of resource congruency in a given plant community (Bazzaz, 1996). Additional research by our laboratory at Harvard Forest has demonstrated that, for seedlings in a canopy gap where environmental conditions show distinct temporal patterns, resource congruency could play an important role in permitting species coexistence (see papers by Peter Wayne and Gary Carlton).

VII Conclusions

Plant currencies that impact on fitness (e.g., sequestered nutrients, biomass, meristems) may be allocated to growth, survival, or reproductive functions. The three-way trade-off implied by this allocation defines three corresponding fundamental components of competitive ability in plants (Fig. 8.2) and hence, the potential for intransitive relative competitive abilities of genotypes at the local neighborhood level, thus promoting multispecies coexistence at the community scale, despite intense competition and even within a relatively homogeneous physical environment. Identifying the relevant spatial scales for these patterns and processes, therefore, is of particular importance, but so too is the identification of relevant temporal scales. Measurement and comparisons of competitive abilities of plants are incomplete and likely to be misleading unless they encompass both survival and fecundity allocation components spanning at least one complete generation (Aarssen and Keogh, 2002). In addition, species coexistence involving the aforesaid mechanism is not expected to be as theoretically stable as when the minority species always has an advantage (e.g., as with traditional niche differentiation) (Tokeshi, 1999). Theoretically, a “random-walk” to exclusion could be expected eventually for some species that coexist only because of intransitive relative competitive abilities. Nevertheless, the time required for this within real vegetation may be so long that species are more likely to be excluded from a local community for other reasons first (e.g., habitat destruction or climate change). During this time, therefore, species unequivocally coexist (i.e., avoid exclusion despite intense competition), regardless of the fact that there is no theoretical potential for indefinite stable equilibrium (Aarssen, 1983, 1989, 1992; Hubbell and Foster, 1986; Hubbell, 2001). In terms of practical applications (e.g., conservation of biodiversity), the time scales over which it is most relevant to interpret and understand the mechanisms that promote species coexistence and hence, maintain biodiversity within vegetation, are the finite time intervals over which real habitats and real resident species within them, actually exist.

The challenge for future studies is to obtain quantitative measures of competitive ability for plants in terms of all three fundamental fitness components (Fig. 8.2) and to use these data to test predictions about not only which genotypes and species have the traits necessary to cause the competitive exclusion of others, but also which genotypes and species have the traits necessary to prevail under intense competition and hence, coexist with other genotypes and species that are similarly equipped.

Summary

Species interactions are classified by the direction of their effects and are divided into the direct pairwise categories of competition, predator–prey, mutualism, commensalism, and amensalism (Table 1). The strength of an interaction or the amount an interaction affects the population size of its participants, can be determined by experimentally removing one species and observing the population response of the second species and vice versa.

A simple mathematical theory for competition predicts that the amount of resource overlap between two coexisting, competing species must be small (Figure 2). A simple mathematical theory for predation predicts that for coexisting predators and prey, population sizes will oscillate through time (Figure 3). In both cases, theory may oversimplify the opportunities for species coexistence. Factors such as variation in resources, space, and time, as well as the complexity of a habitat, can minimize resource overlap among competing species or dampen predator–prey oscillations. Mutualistic or positive interactions do not have a historic theory for predicting when and where they will occur, but mutualisms are common in nature.

Pairwise interactions between species can select for changes in morphological, physiological, and behavioral traits through evolutionary time. Coevolution occurs when there is reciprocal genetic change between two interacting species. Coevolution in competitive systems can lead to character displacement. Coevolution in predator–prey systems can lead to an arms race of capture and escape. Also, coevolution in mutualistic systems can lead to highly specialized associations between mutually benefiting species.

Communities are composed of many species, and indirect as well as direct interactions control community composition (Figure 5). Indirect interactions include trophic cascades, apparent competition, and the modification of a direct interaction. A keystone is a species that strongly influences the biodiversity of its community via indirect and direct effects (Table 2). Keystones are often, but not always, found at the top of a trophic chain or food web. Both indirect and direct interactions can influence the evolution of community members.

Humans are causing widespread extinction of species, and the loss of species is leading to the decline of many goods and services provided by species interactions in natural systems. Human stresses on the environment, such as climate change, alteration of geochemical cycles, and destruction of habitat, can break down species associations because individual species, not communities, respond to changes in the environment.

Why Neutrality?

The ultimate goals of neutral theory, and of community ecology more generally, are to increase scientific understanding of species assemblages and to predict or manage their future behavior. How can hundreds of different tree species apparently coexist in a single 50-ha forest plot? What processes structure species abundance distributions (SADs) and species–area relationships (SARs) at different spatial scales? What governs the temporal dynamics of ensembles of species? There are both pure scientific benefits to developing this understanding, and practical conservation and management benefits as well. At what rate will species be lost with habitat destruction? How should conservation reserves be designed to maximize biodiversity? The development of quantitative models like neutral theory can address these questions too.

The classical ecological perspective is that species coexist through their differences: each species has its own niche, and any two species too similar in form and function compete, with only the fitter surviving. The formalization of this competitive exclusion principle and the subsequent development of niche theory have given ecologists an intuitive view of how communities are structured, often borne out by empirical evidence.

There are, however, two clear challenges to this niche-structured picture. The first is empirical: in highly diverse communities it is difficult to reconcile the number of species with the number of niches. The second difficulty is theoretical: deriving predictions for classical ecological patterns like the SAR and SAD from niche theories is not only difficult, but also seems contingent on the complexities of a given community. The remarkable similarity of these and other patterns across multiple communities has led to many statistical models, but deriving these patterns from mechanistic processes in niche-based models has proved extremely difficult. Two questions therefore motivate neutral theory:

Is there an alternative general explanation for species coexistence that does not rely on species occupying distinct niches?

Is it possible to generate quantitative predictions of biodiversity patterns (such as SADs and SARs) from a mechanistic theory?

The neutral theory of biodiversity provides possible answers to these questions. To address the first question, neutral theory steps away from the idea of potentially indefinite coexistence and toward transient coexistence, also known as co-occurrence. If observed species richness arises from a simple balance between speciation and extinction, then coexistence no longer depends on each species having a distinct role in the community. Neutral theory answers the second question directly by accurately predicting certain biodiversity patterns from simple mechanistic models. This provides evidence that similar mechanisms may be responsible for generating biodiversity patterns across different communities, despite large intercommunity variation in species identities, interactions, and environmental conditions. It is important to note that neutral theory does not deny the existence of nonneutral processes; it postulates that they are not important for predicting community structure at certain spatial and temporal scales.

Key Characteristics of Neutral Models

Although the definition of neutrality provides an overarching guiding principle, specific neutral models are needed to generate predictions and test the theory, and the definition allows scope for substantial variation among these models. The three fundamental ecological processes that most neutral models include are stochastic drift, speciation, and dispersal. The authors give a schematic picture of a neutral model in Figure 1.

Figure 1. Schematic diagram of the processes of birth, death, speciation and dispersal, shown on a spatially explicit lattice. The death of an individual opens a space on the lattice, which is followed by a birth event: the space is colonized by the offspring of an existing organism. In spatially explicit models, dispersal limitation constrains the likelihood that a given organism's offspring will successfully colonize the space, as a function of separation. The lattice structure here imposes a zero-sum constraint, so that the total number of individuals is fixed. Subsequent models have relaxed this zero-sum constraint, so that the total number of individuals fluctuates around a fixed average.

Stochastic Drift

In typical neutral models, stochastic birth and death processes approximate the complex, multifaceted and typically unpredictable fine-scale processes that occur in real ecosystems. The demographic stochasticity that governs the dynamics of such models is also referred to as drift. Drift is expected to drive the dynamics of species abundances when stabilizing forces, characteristic of deterministic niche theories, are weak. Some neutral models constrain the drift process by enforcing a zero-sum rule at the community level, whereas others do not.

Speciation

The definition of neutrality does not make any reference to the fundamental ecological process of speciation, but speciation may be introduced to balance extinction arising from stochastic drift, facilitating the maintenance of diversity. Point speciation, the mode of speciation in most published neutral models, allows a small but finite probability that an individual birth event will give rise to a new species. This is a schematic representation of true speciation processes, and more recent developments have added a variety of ways to model speciation in neutral theory: ‘random fission’ is an analog of allopatric speciation, whereas ‘protracted’ speciation distinguishes incipient from well-established species.

Dispersal

Neutral models treat dispersal in different ways. Hubbell's original approach has a two-scale spatial structure (see Neutral Models), in which dispersal is treated as immigration from a metacommunity into a local community. In other neutral models, dispersal is modeled as a spatially explicit process, in which case it is possible to give a more realistic treatment of neutral dynamics on a contiguous landscape.

V. Conclusions

Theory, experiments, and natural history all suggest that communities are not tightly organized, so that species packing may not be strongly relevant in nature. In short, the popular image of an ecological community as an airplane in which each part has a vital role for the integrity of the whole is dubious. There may be groups of species, in which each one is closely connected to a few others but only loosely connected to other groups.

The term “community” was once extremely useful and is still of pedagogic value if carefully used. As reified objects for research, the concept of communities is now threatening to become what Simberloff and Dayan have referred to as a “panchreston,” an idea as likely to generate confusion as enlightenment.

The prospect of a general species packing theory has melted away. Despite the unreality of the models, they did direct attention to what seemed to be a real phenomenon and encouraged experimentalists and field biologists to ask important questions.

Differences in species richness can be partially explained by a multiplicity of factors that do not necessarily relate to each other so as to permit formation of a single coherent theory. Each of the multiplicity of theories of diversity focuses on one or a few of the empirical factors that are known to enhance or diminish the possibility of species coexistence.

Many of these can be seen in laboratory experiments, which have the advantage of clarity but may be of questionable applicability. For example, local geometry complexity influences species diversity in nature as well as in the laboratory.

Certainly in some cases the term “species packing” is used in the sense of species being squeezed into a space. In several cases individuals of particular species in species-rich regions are believed to have a narrower range of activities than individuals of the same species in species-poor situations. Diet or nest sites may be more restricted. In these cases the individual organisms can be imagined to have been constricted by some packing process, like individual pillows in a crate but even this has two possible meanings depending on whether we are concerned with the population level or with individuals.

In comparing different locales, the range of variation among organisms within a unispecific natural population may be reduced when more other species are present. In the case of comparisons of different islands or lakes, this might be tentatively attributed to species packing.

Returning to the initial analogy of packing actual objects, an island is clearly a container. But if packing means filling the ecological space, either by pillows or tumblers, we would not expect islands or speciose lakes to easily admit invasive species. In lakes there may be enormous species richness of fishes, as in the ancient African lake cichlids, but I don't know of any comparisons showing that species-rich lakes show less within species, among individual variation, than species-poor lakes. Is the attempt to crowd multispecific collections of fish together in the same lake equivalent to packing glass tumblers rather than pillows? There is a general impression that species rich systems seem at least as likely to be invaded by exotic species as species-poor systems, violating our sense of what packing might mean.

If the container walls are not apparent but among individuals variation is reduced in one or more populations of a species, is this a sign of species packing? Does the mere fact that among individuals variation is reduced when more species are present imply that there must exist a container wall which may not be obvious?

The overall conclusion is that the theory of species packing does not conveniently predict very much about natural systems but that the images of packing that it generates do informally suggest interesting phenomena to look for.