Jump to ContentJump to Main Navigation
What Is Biodiversity?$

James Maclaurin and Kim Sterelny

Print publication date: 2008

Print ISBN-13: 9780226500805

Published to Chicago Scholarship Online: February 2013

DOI: 10.7208/chicago/9780226500829.001.0001

Show Summary Details
Page of

PRINTED FROM CHICAGO SCHOLARSHIP ONLINE (www.chicago.universitypressscholarship.com). (c) Copyright University of Chicago Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in CHSO for personal use (for details see http://www.chicago.universitypressscholarship.com/page/privacy-policy). Subscriber: null; date: 17 July 2018

Explorations in Ecospace

Explorations in Ecospace

(p.106) 6 Explorations in Ecospace
What Is Biodiversity?

Maclaurin James

Sterelny Kim

University of Chicago Press

Abstract and Keywords

This chapter focuses on local ecological communities, and on whether local communities are structured, organized systems; that is, systems whose organization has important effects on the identity and abundance of the local biota. In analyzing the idea that communities are indeed structured systems, it considers the claim that communities control their own membership and the claim that they have biologically important collective properties. If these ideas are vindicated, we do need more than species information. We need information about organization and variation in that organization from community to community. In the chapter's “units-and-differences” framework, it asks whether local ecological communities are themselves units, and, if so, what are the relevant similarities and differences among them.

Keywords:   ecological communities, ecological systems, ecosystem, structured systems

6.1 Ecological Systems

So far in this book our main focus has been on evolutionary dimensions of diversity, and on the causes and consequences of that diversity. A central theme has been the relationship between species richness and other dimensions of biodiversity, and the extent to which the biodiversity of a system is captured by information about the identity, demography, and evolutionary relationships of the species in the system. While species richness does not determine these other dimensions, and may not always be a good surrogate for them, there are important causal and theoretical links between species richness, morphological disparity, and plasticity. In the previous chapter we looked at recent work on evolution and development. We discussed the relationships between species, evolutionary morphospace, and the supply of variation to evolution. In this chapter our focus changes to ecology. We explore ecological notions of diversity and the relationship between ecological and evolutionary systems.

We begin with a familiar question: What more might we want to know about the biodiversity of an ecological system over and above its species composition and facts about variation and plasticity within those species? Ecology is an enormous and complex field in its own right, so we will pursue this general theme through a specific example. This chapter focuses on local ecological communities, and on whether local communities are structured, organized systems; that is, systems whose organization has important effects on the identity and abundance of the local biota. In analyzing the idea that communities are indeed structured systems, we will consider the claim that communities control their own membership and the claim that they have biologically important collective properties. If these ideas are vindicated, we do need more than species information. We need information about (p.107) organization and variation in that organization from community to community. In this chapter's “units-and-differences” framework, we ask whether local ecological communities are themselves units, and, if so, what are the relevant similarities and differences among them.

Ecologists study (among much else) the interaction among populations, and among those populations and their environments. Their aim is to understand the distribution and abundance of organisms. They study the processes that determine distribution and abundance at many spatial and temporal scales. But perhaps most attention has been focused on local communities, on the avifauna of a particular wood, or on the invertebrates on a particular beach. As Robert Ricklefs has put it, much of ecology has been organized around a model of “local determinism” (2004). On these models, the abundance and composition of local communities is essentially controlled by the causal characteristics of that community itself. Thus, much ecology has been local community ecology, and we shall follow that lead. We do so with some reluctance, for one of the most interesting recent developments in ecology is a shift away from local determinism to macroecological models (for a recent review of the many different proposals about the natural units of ecology, see Jax 2006). On these macroecological models, the profile of a local community (the species present and their abundance) is driven mostly by the characteristics of regional biotas. A forest patch in Ecuador is more species rich than a similar-sized patch in England because the Ecuadorian regional biota is immensely more rich; it is not the characteristics of the patch itself that primarily explain this difference.1 We return to the relationship between local and regional structure briefly in the final section, but our focus on local determinism implies that this chapter should very much be thought of as a preliminary study of ecological diversity.

Local determinism could be true in two ways. The identity and abundance of the organisms in a local patch might be controlled by the abiotic environmental features of the patch: rainfall, temperature, soil profiles, wind exposure, and the like. Alternatively, it might be controlled by interactions between the organisms present, interactions that favor some potential residents and exclude others. (Obviously, mixed models are possible.) Thus one important issue in ecology is whether the distribution and abundance of organisms is in part explained by characteristics of ecological systems themselves. Are local communities organized systems that make available space for some populations and exclude others, thus regulating their own membership? Ecologists began with the view that local systems were organized systems in this sense. Charles Elton's theory of the niche, the first theory of ecological niches, took (p.108) niches to be the ecological equivalents of economic roles in human social systems. The organization of a particular community made certain ways of life available within it, but not others (Griesemer 1992; Worster 1994). But there were early dissenters who argued that distribution and abundance is largely explained individualistically (Gleason 1926). Different species have different tolerances to physical conditions, different resource requirements, and different levels of vulnerability to biological threats or physical disturbances. The distribution of organisms is largely explained by these species' independent responses to variation in the environment, especially the physical environment. If individualist models of ecology are vindicated, information about the presence and abundance of species captures the ecologically relevant biodiversity of biological systems.

There is clearly some truth in the individualist idea; the distributions, abundances, and evolutionary trajectories of particular species are profoundly influenced by the physical features of their environments. Arid, nutrient-poor Australian environments have a very different biota from that of the (barely) temperate rainforests of the west coast of New Zealand's South Island. To some extent, biological variation across space and time is a response to physical variation across space and time. In this sense, the investigation of ecological diversity—biotic variation across habitats—is a calibration of the physical parameters that affect the distribution of species. In a well-known example of work of this kind, Robert Whittaker has argued that much of the variation in species composition across different habitats can be explained by just mean annual temperature and rainfall (1975, 167).

To the extent that the distribution and success of organisms can be explained by their autonomous response to such physical variables, the individualist program in ecology will be vindicated. We will not need to appeal to ecological systems, to the structure and organization of communities, to explain those facts. Rather, distribution, abundance, and fate are explained by the interactions between species' evolved phenotypes and their environment. Suppose, for example, that the replacement of rimu-kahikatea forest by southern beech as one goes south on New Zealand's west coast can be explained by these species' differential responses to temperature, rainfall, wind, and soil. If so, we would not need to appeal to features of rainforest community organization—to features of the ecological system—to explain these species' distributions and abundances. Individualists recognize that many organisms need biologically made resources, but their bet is that the biological tolerances of local populations are, for the most part, quite coarse. Of course there are specialists. Glossy black cockatoos (Calyptorhynchus (p.109) lathami) require mature casuarina trees, and some caterpillars will lay their eggs on only one species of plant. But while organisms depend for resources on their biological as well as their physical environment, in the individualist view they do not typically depend on a specific array of interacting populations. Species do not really care who their neighbors are.

In assessing the plausibility of this individualist line of thought in ecology, it is important to distinguish between a phenomenological and a causal view of local communities and kinds of communities. There is no doubt that local communities—these assemblages of plants and animals found in association on distinctive habitat patches—are part of the descriptive phenomenology of ecology. Moreover, there is a reasonably natural and predictive taxonomy of habitat patches. For example, when we find out that there are tidal mudflats near Moruya (on the New South Wales coast of Australia) we have a fair idea of the plants and animals we can expect to see: mangroves, samphire, a suite of distinctive birds (herons, waders, and the like) and so on. Coastal wetlands on the east coast of New South Wales vary one from another, but nonetheless they support a broadly similar range of species. These similarities allow field guides and similar tools to distinguish among woodland and closed forest, wetlands, grasslands, coastal heathlands and sand dunes, and arid and semiarid areas. Flora and fauna differ in characteristic and fairly repeatable ways that are captured by these descriptions. We can identify certain types of community by statistical patterns of association among species; woodlands and wetlands have different inhabitants. In identifying community types this way, we say nothing about the processes that produce these identifiable and repeated associations. So communities are important to ecology in this minimal sense; local patches have stable natural histories that typically do not vary dramatically from year to year. And a given local patch will resemble some other local patches well enough for there to be useful taxonomies of habitat types. The modest view of these local systems, then, is that they have reasonably stable natural history profiles, and that fact enables us to make some reasonably reliable qualitative predictions about their overall biological composition. Different phenomenological communities may just reflect differences in interactions between species and their physical environments. But they would still be useful surrogates: allowing inference from community type to species composition.

Thus the phenomenology of local communities is important because it reveals what we want to explain and protect. It is important for a second reason. Differences in the natural histories of these phenomenological communities are symptoms of important ecological processes. For (p.110) example, in Wellington there is a so-called mainland island, the Karori Reserve. This is a chunk of remnant bushland that survived around Wellington's former water reservoir. It has now been enclosed with predator-proof fences, and it is the site of a major effort to extirpate exotic mammals and weeds. The local Web site has reported the striking results of these changes: endangered animals, like the little spotted kiwi (Apteryx owenii), have been successfully reintroduced to this area. And birds like the tui (Prosthemadera novaeseelandiae)—once only just hanging on in the city—have rebounded. The differences between this reserve and other areas of local bushland (and from its former self) are only too obvious. It is full of native plants and animals. They are infested with possums, hedgehogs, rats, feral cats, and an assortment of weeds. At the Wilton's Bush Reserve, only a kilometer or two from Karori, the difference is obvious even in the course of a short walk. Wilton's Bush is quite rich in native vegetation, but the forest is almost silent.

Even if local communities have no causally salient properties that drive ecological processes, it is no surprise that ecological and conservation biology journals are full of descriptions of local communities. The differences between them reveal ecological processes in action: competition, predation, response to physical disturbance, and invasion. The local biology students learn their survey techniques at Karori, looking for mouse droppings and counting birdcalls on transects. Phenomenological descriptions of local communities help set and test an explanatory agenda for ecology and, often, for conservation biology as well. The difference between the Karori Reserve and Wilton's Bush is not the result of deliberate environmental manipulation to test for the ecological consequences of introduced predators; Wilton's Bush is not a deliberate control plot. But when conservation biology meets ecology, a standard experimental probe is to match “no-intervention” communities with “intervention” communities. Thus the devastating effects of foxes on small to medium size marsupials has been established by contrasting communities in which fox numbers are suppressed by baiting, with communities without fox control, and surveying the marsupial fauna of interest (Kinnear et al. 2002).

The effects of crucial biological processes are often revealed in this way, by comparing communities that differ (as far as we know) only in one important respect. For example, one important debate in contemporary ecology is about contingency. Contingent systems are sensitive to unpredictable events, and hence their future trajectories are unpredictable. One form of contingency is “path dependence.” A community's future is path dependent if (for example) the order in which migrants arrive makes a major difference to the community that is ultimately (p.111) assembled. If order effects were important, we would not be able to predict the future trajectories of island communities because their local ecology would depend on the accidents of arrival order. If species A and B were to arrive together, B would exclude A. But if A arrives first, it has a good chance of preventing B from establishing. It clearly matters whether path dependence is ecologically important, and the best way of testing for path dependence is by comparing communities with similar early histories to determine whether their futures are similar when they differ only or mostly in the order in which colonists arrive. The Krakatau islands have provided the opportunity to make just these comparisons, as the different remnants of Krakatau allow comparisons of different islands at the same time, and of the same island at different times (for further eruptions have turned the assembly clock back to zero).2

So one way of thinking about ecological diversity is in terms of a phenomenological ecospace. The dimensions of that space include salient measures of the physical environment. For terrestrial communities, these are rainfall, temperature, soil structure, and the like. Such an ecospace will have biological dimensions, too, specifying the presence and abundance of the dominant vegetation (by species or by functional group), and likewise for other trophic layers. Two wetlands in southeast Australia will end up near neighbors in such a space, in virtue of their physically similar substrates and the presence of similar organisms in similar numbers. The dimensions, then, are the dimensions of descriptive ecology: physical environmental variables, vegetation cover, and the animals living in and on the vegetation. As with morphospace, though, a total ecospace is of high and somewhat arbitrary dimensionality. Would we have a dimension for every duck in the regional biota? A dimension for every soil element, or just an aggregate measure of fertility? Ecologists will typically be interested in comparing communities with respect to just a few dimensions, and those few will depend on the purposes of the comparison. If we are interested in the impact of foxes on small marsupials, the most crucial dimensions will be fox abundance and small marsupial abundance, though if we think other factors might exacerbate or mitigate the effects of foxes, we will have to include these too (for example, density of cover, other predators). For fox-baiting studies, the boundaries of the community are defined by the boundaries of fox-baiting, for we are interested in the effects of baiting in that region, and that is true whether or not the limits of baiting coincide with an overt phenomenological change in the local ecology. For other purposes, we would represent the same habitats quite differently. A botanist interested in the causes of eucalypt dieback would choose different dimensions of comparison.

(p.112) Even if the individualists are right about local communities, a good descriptive taxonomy of local communities would be a good tool for both conservation biology and ecology. It would deliver an easy to use surrogate for alpha and beta species richness,3 and a valuable probe for assessing the impact of ecological processes. But there is a more ambitious project, one that takes local communities themselves—as distinct from the populations that comprise them—to have causally salient properties. The crucial question here concerns the extent to which local communities are functionally organized systems. Consider, for example, Black Mountain, a eucalypt woodland community in Canberra, and one of Sterelny's local patches. Is this an organized biological system? Not if individualism is right. If the Black Mountain community is an assembly of populations whose sizes and prospects of persistence are largely independent of one another, if its components have impacts on their environment that are mostly independent of their neighbors, and if it is an assembly of populations with varying ranges that somewhat overlap, then the Black Mountain community would merely be part of the descriptive phenomenology of biology. It would be a “unit” in something like the sense that a genus of duck species is a unit, rather than the sense in which a species is a unit. Identifying Black Mountain as a eucalypt woodland on the southwestern slopes of New South Wales would give conservation biologists a good idea of its alpha diversity. The differences between it and otherwise similar phenomenological communities calibrate the power of ecological mechanisms. But there would be questions it makes no sense to ask. The community would have no objective bound in space or time, and nor would the community as a whole have explanatorily salient features. So from the fact that such communities and community types can be characterized phenomenologically, it by no means follows that communities have autonomous, biologically important properties; it by no means follows that they have organizational or structural properties that help explain the distribution and abundance of organisms.

In brief, the individualist idea has led to a very serious debate about the extent to which communities are organized systems, whether community structure filters the species present in a local patch, excluding some and admitting others, and whether that same structure determines (or strongly constrains) the abundance of those populations that are present. In the language of ecological theory, there has been a debate about the extent to which communities are structured by “assembly rules,” rules that specify those species that can, and those that cannot, co-occur with one another in local communities.4 It is possible (p.113) that, say, a local population of banksias and another of eucalypts on Black Mountain are associated spatially only because both populations happen to tolerate the temperature, soils, and rainfall characteristic of this location. If each of the Black Mountain populations is more or less indifferent to the presence of others, then this assemblage is merely a “community of indifference.”5

Communities of indifference are merely phenomenoiogical communities; populations within them are spatially associated only because they happen to tolerate similar physical conditions. Such “communities” are not organized, structured systems. On this view, there will be a more or less deterministic explanation of why particular species are represented on Black Mountain—soils, rainfall, and temperature make it hospitable to some members of the regional species pool but not others. But these explanations will be relatively independent of one another. An explanation of the composition of the community is no more than the sum of the explanations of the presence of each member of the community. Communities of indifference have no causally salient organization. Yet, if community regulation is important, so that membership and abundance is filtered by the structure of the community itself, then communities are not just part of the phenomenology of biology. The populations present and interacting in a particular local habitat constrain the range of potential members of that community. The current status of this idea is the focus of 6.3.

There is a second challenge to the idea that ensembles in a local patch are just communities of indifference. Communities are real, causally important ecological systems if they have emergent or ensemble properties; if, for example, a forest dominated by pines has properties that are not just an extrapolation of the properties of individual pine trees. This idea is controversial, and we will return to it in 6.4. Within ecological theory, the idea that communities have ensemble properties has been explored in many ways. We shall do it by considering the diversity-stability hypothesis, the idea that more diverse communities are more stable. According to this hypothesis, diverse communities are less perturbed by external disturbance, and they return to a predisturbed condition more readily than less diverse ones. Diversity in this context is a property of the community itself, and so, in some version of the stability-diversity hypothesis, is stability. In diverse communities the overall productivity suffers less in (for example) drought, but individual populations may fluctuate as profoundly as those in less diverse communities. The idea here is that communities themselves (as distinct from the organisms and groups that compose them) have causally salient properties.

(p.114) 6.2 Communities, Ecosystems, and Ecosystem Functions

We think that communities are causally important, but that particular communities vary in the causally salient properties they have and the degree to which those properties are causally salient. We begin, though, with a conceptual prologue: function and functional organization in ecology. Function in ecology is not like function in evolutionary biology or functional morphology. In those fields, functions derive from selective history (Wright 1973; Millikan 1989; Godfrey-Smith 1994). The ponyfish has a light-emitting organ, and the function of the light the organ generates is to prevent the ponyfish from being visible from below, silhouetted darkly against a lighter background. In matching the illumination radiating down from above, the ponyfish is concealed from predators. The ponyfish shines to be invisible. In making this claim about the function of the light-emitting organ, we make a claim about selective history. Ancestral ponyfish with such organs survived better than those without them (or with less well-tuned organs) because they were less often seen from below (Williams 1997).

It is not likely that we can explain functional roles in local communities in a parallel way. In the early history of ecology, the idea that communities were like organisms was taken quite seriously. Frederic Clements 1936) thought of communities as systems in a very rich sense, as akin to superorganisms. He based this on his view of ecological succession. Succession organized plant communities in a robust way, so that even after very severe disturbance, a homeostatically preserved equilibrium, the climax community, would be rebuilt (Cooper 2003). But no one would now defend a view of functional organization of communities modeled on the functional organization of organisms. Not only are organisms much more tightly integrated and bounded than the typical community, but also, as a rule, local assemblages do not have selective histories. They are not part of lineages. Communities are not elements of a population of competing communities, and they do not have daughter communities that resemble their parents. If a selective history is necessary for communities to have organization or structure, then most assemblages of populations are not ecological systems.

However, as Robert Cummins has shown, there is an alternative view of function and organization. A part of a system has a Cummins-function when its activity makes a distinctive, stable contribution to the operation of the system as a whole (Cummins 1973; Godfrey-Smith 1993, 1994). Thus in many Australian woodlands, eucalypt litter has the Cummins-function of making fire more likely. This is a stable, regular contribution of this component of a woodland system to the overall behavior of that (p.115) System. So local communities may be functionally organized, structured systems because their components have Cummins-functions. For example, there has been important work in ecology, beginning with Robert Paine 1966), on the role of keystone predators in maintaining diversity. They do so by limiting populations that would otherwise out-compete others (for a review, see Power et al. 1996). Keystone species have Cummins-functions; starfish are not selected to maintain diversity by eating mussels, nor has there been between-community selection for mechanisms that maintain diversity. But within that community, this is a stable effect of this particular population (see Box 6.1). Community ecologists often analyze communities in terms of guilds or functional groups, which are components of a community intermediate between a community as a whole and a local population (Naeem 1998). They are sets of populations playing specific roles within a community: browsing, pollination, or seed dispersal. Such guilds and functional groups are identified by their Cummins-function (Blondel 2003).

Ecosystem ecology, in particular, has been centrally concerned with identifying and explaining Cummins-functions. There has been a (p.116) historic divide between community ecology, aiming to explain the identity and abundance of local species populations, and ecosystem ecology, aiming to explain the flow of material and energy through the local system (Golley 1993). For example, ecosystem ecologists study the flow of crucial nutrients like phosphorous and nitrogen from the soil into organisms and back into the soil. The organisms responsible for these flows—the detrivores that consume litter and make soils—are performing Cummins-functions; they make a stable, repeatable contribution to the behavior of the system as a whole. Despite this historic divide between ecosystem and community, it will not be pivotal to our discussion. John Odling-Smee and his co-workers have argued that the distinction between community ecology and ecosystem ecology is eroded once community ecologists recognize the niche-constructing role of organisms and populations (Odling-Smee et al. 2003). Organisms do not just eat, breed, and die. They reorganize their environment. Hence, an explanation of the presence, abundance, and activities of local populations will also explain the biotically caused flow of materials and energies through that local system. Once the role of organisms in niche construction is recognized, the distinction between community ecology (focusing on the distribution and size of populations) and ecosystem ecology (focusing on the flow of matter and energy through a habitat) becomes much less sharp.

In terms of this framework, then, phenomenological communities are organized systems only if they are stable, bounded, and with enduring global features of biological importance to which particular components make a regular contribution. Arguably, they are organized systems in this sense if they are regulated, that is, constrained in membership and numbers by their Cummins-functional organization, or if they have causally important emergent properties.

6.3 Individualism and Community Regulation

Local populations do not live independently of one another. Species depend on the local biology; there can be no echidnas without ants. But individualists think that species have broad-banded biological conditions of existence. Most particularly, competitive interactions do not determine community make-up; species are not typically excluded by other species with similar resource-use profiles. So they are skeptical about the predictive importance of an important organizing idea in ecology, the principle of competitive exclusion. The principle itself states that species with the same resource requirements cannot indefinitely co-occur; one will be competitively superior and drive the other (p.117) to extinction. Generalizing this, species with similar requirements will have strong competitive interactions, and will tend to exclude one another. These results are based more on models than on observations of natural systems, and many ecologists doubt that real habitats are sufficiently uniform and stable to reach the equilibrium at which exclusion takes place (for reviews, see Kingsland 1985 and Cooper 2003). Suites of parrots, of honeyeaters, and of insectivores coexist on Black Mountain and they do so (according to this line of thought) because real habitats are heterogeneous; many populations extend over patches that contain relevant environmental variation. So one species of thornbill does not exclude the others. Moreover, they are fluctuating. They are not filtered by competitive exclusion, because the world intervenes before local assemblages reach their theoretical equilibriums. On this individualist view, phenomenological communities are typically associations of overlapping populations. Such phenomenological communities do not have determinate boundaries. Moreover, though these populations are not fully independent of one another, they interact weakly. Populations do not regulate one another, nor do they impose hard-to-penetrate filters on community membership.

How plausible is this conception of communities? In particular, is it consistent with the readily observed, qualitative stability of local ensembles of populations? As Greg Cooper discusses at some length, there is a line of thought in ecological theory that infers regulation from stability.6 Stable ensembles, the thought goes, are internally organized through competitive interactions. The stability that makes field guides possible cannot be explained by abiotic factors. Their impact is too variable. Rainfall, for example, varies dramatically from season to season, and so too does the incidence of fire. The Black Mountain biota does not inhabit a physically invariant landscape. Yet the Black Mountain community is roughly stable in both composition and abundance. That fact is best explained by the hypothesis that communities are regulated by “density-dependent” biotic interactions. The size of some given population—say, superb fairy wrens on Black Mountain—will fluctuate within bounds only if the factors that limit the fairy wren population become more intense as the population rises, and less intense as it falls. An obvious candidate for such a factor is competition between the wrens for limited resources. The more wrens, the harder such limits bite. Competition is bound to get more intense as population size increases, and less intense as it dips.

In briet, stability is the result of a “balance ot nature”, a balance deriving from the internal regulation of communities. The qualitative stability of natural and artificial ecosystems shows the importance of density-dependent factors. If the forces that affect a local population act (p.118) independently of its size, it would be an amazing coincidence if abundance did not change over time. Very slight tendencies to increase or decrease result in crashes or plagues. If populations persist, something must damp down such fluctuations. Yet abiotic factors are not sensitive to population size. The impact of flood, fire, or drought—and external disturbances more generally—is not sensitive to the size of the populations on which they impact. An oil spill will destroy a seabird rookery without regard to the number of birds present. We can infer from the qualitative stability of communities that they are networks of biological interaction that filter membership and that constrain the demography of their members.7

Cooper is rightly skeptical of this whole class of arguments; they depend on a crucial ambiguity (Cooper 2003). There is an undemanding sense of “stable,” where it means something like “the persistence of community membership.” On this reading, communities are indeed typically stable, as is Black Mountain, whose species composition is similar year by year. But while most communities most of the time show a fair degree of persistence of community membership, that does not establish the existence of internal regulating mechanisms. Over shorter periods, other mechanisms can explain persistence. The crucial point is that communities are often demographically open. Thus a local population may persist by recruiting from neighboring communities. The stability of demographically open communities can be the result of such metapopulation dynamics. If, for example, echidna populations vary independently of one another in a cluster of adjacent communities, a population fluctuating toward extinction can be rescued by migration from a neighboring community whose numbers happen to be surging. Migration between communities can protect unregulated communities from random walking to extinction. The effects of density-dependent internal regulation can be coarsely mimicked by a metapopulation of unregulated communities, provided that metapopulation is spread over a heterogeneous landscape and provided that migration from one population to another is possible.

Populations without density dependence can persist for many generations. Even if competition, prédation, and other density-dependent ecological mechanisms are not important, so long as the trajectory of populations within a cluster are independent of one another, the stability of the metapopulation ensemble will be greater than the stability of a typical population within the ensemble (Baguette 2004; Hanski 2004; Murdoch 1994). We do not know the extent to which metapopulation dynamics explain the evident stability on which field guides depend. But the existence of this mechanism means that we cannot assume that persisting communities are internally regulated. Moreover, we know (p.119) there are qualitatively stable associations that can hardly be the result of strong interactions between the local residents of a community. There are field guides to estuaries and other habitats where many of the birds are migrants; they are winter residents. Many of the waders found on Australasian tidal mudflats breed in the far north. And while banding studies suggest that these birds are faithful to their breeding zones, there is no reason to believe that the same birds—the godwits, the knots, the turnstones, and curlews—return year after year to Foxton estuary on the east coast, north of Wellington, or to Miranda, south and east of Auckland.8 The stability of these associations is presumably explained by a stability in the flow of resources through these systems.

A more demanding notion defines stability not just in terms of community membership but also of population size. If communities are at true equilibrium, with population sizes varying only in minor ways around a mean to which they typically return, then they must indeed be regulated. But so understood, there is no reason to believe communities are typically stable. In short, if there is evidence of limited variation around a mean in the population sizes of components of the community then we do indeed have evidence of equilibrating mechanisms. But it remains to be shown that local assemblages are typically characterized by limited movement around a mean.

Time to take stock. It is an obvious truth of ecology that local assemblages are fairly stable over short periods of time.9 Farming would be an impossible activity if that were false. But while this fact is suggestive, in itself it is not sufficient to show that local assemblages are typically ensembles of strongly interacting and thereby stabilized populations. If stability just consists in the persistence of community membership, such persistence may be explained by metapopulation effects. The observed phenomenology of stability does not show individualism to be mistaken. What, though, of emergent properties? Perhaps communities have causally important properties that equilibrate features of the community as a whole: diversity, productivity, or ecosystem services (for example, the flow of crucial nutrients like nitrogen and phosphorous from organisms to soils and back). This is an important idea to which we turn in the next section.

6.4 The Emergent Property Hypothesis

In this section we discuss the idea that local communities have causally salient properties by exploring a family of famous hypotheses that link the diversity of a community to its stability. Communities (the thought goes) have emergent properties. An ensemble has emergent (p.120) properties if it has features that are not simple reflections of the properties of its parts. The notion of an emergent property is not inherently mysterious or spooky. No one doubts that organisms have emergent properties. Organisms are built from cells (plus some of their products), and it is quite clear that the fitness of an organism (for example) is an emergent property of the cell ensemble in its environment. There is no way that the fitness of a particular eucalypt is a simple reflection of the cells, cell population, and their products out of which the eucalypt is built. The fact that organisms are composed of cells was an epochal biological discovery; there is no understanding how organisms work without understanding how cells work. But it does not follow that we can understand how organisms work just by understanding how cells work. The compositional structure of ensembles constrains their behavior, and hence is crucial for understanding system-level behavior. However, it may not be true that understanding the parts from which an ensemble is built suffices for understanding the ensemble.

Thus, in discussing the relationship between an ensemble and its constituent elements, philosophers distinguish between a supervenience claim and an explanatory claim. The supervenience claim is that there can be no change in system-level properties without change in the properties of the parts; the Karori Reserve community cannot become better buffered against invasion by exotic weeds unless there is some change among the particular populations that make up the community. The explanatory claim is that all important system-level behavior can be explained, and can only be explained, by explaining the behavior of the parts. For such cases as the relation between organismal properties and those of cells, or the relationship between communities and the populations within them, the supervenience claim is uncontroversial. But the explanatory claim is controversial. (See Jackson and Pettit 1992; Sterelny 1996; for the claim that emergent properties are inescapably spooky, and science should never posit their existence, see Rosenberg 2006.)

The crucial idea of the emergent properly hypotnesis is that niese emergent properties are causally important; they drive ecological processes. The diversity-stability hypothesis is one attempt to show this. The thought that diversity adds stability to a community has enormous intuitive plausibility. Diversity adds redundancy, and hence allows that community to survive fluctuations in the fortunes of its members. If only one population on Black Mountain pollinates the gum tree, Eucalyptus rossii, and if it were to suffer a serious decline, rossii would be unable to recruit new plants into its population. However, if there were a suite of eucalyptus pollinators, a fluctuation in one population would not (p.121) ramify through the community as a whole. Redundancy buffers disturbance, and diversity adds redundancy.

This appealing picture seemed to be undermined by the theoretical work of Robert May 1973). His models showed that more diverse communities were less stable, not more stable. The last decade or so has seen a revival of the diversity-stability hypothesis and its close relative, the idea that more diverse communities are more productive. Interestingly, both May and his critics identify diversity with species richness. From the perspective of this book, it is clear that his account of diversity involves an important simplifying assumption. Rather than taking up this simplifying assumption, the critics have sidestepped May's result, taking issue with May's account of stability. May took the diversity-stability hypothesis to be a hypothesis about population size; in more diverse communities, the populations of the component species are more stable. However, David Tilman and others have argued that community-level properties are more stable in more diverse communities. Tilman proposed changing the focus from community composition to ecosystem processes, and to the connections between those processes and community composition. In particular, Tilman argued that the biomass of more diverse communities is more stable10 than that of less diverse ones. For somewhat similar reasons, a diversity-productivity relationship looks plausible. Habitats are heterogeneous in space as well as time. A habitat patch will exhibit small-scale variation in its physical and biological characteristics, and so (the thought goes) in different micropatches different species will be more efficient. This helps explain how communities can retain diversity (as competitive superiority will not produce a monoculture) and explains why more diverse communities are more productive. They are more likely to include the species that are best suited to the various local micropatches spread through the habitat.

Tilman's crucial theoretical idea is that of compensation. If one population declines in numbers, another population, using somewhat similar resources, expands, and hence stabilizes the overall productivity of the community. Importantly, the idea that populations compensate for one another's fluctuations does not depend on controversial ecological assumptions. In particular, it does not depend on the idea that population decline is caused by the competitive superiority of the expanding species. We get community-level stability despite population-level volatility because individual populations have somewhat overlapping resource requirements but quite different environmental tolerances. Tolerance differences explain why populations fluctuate out of synchrony. When frost (for example) causes one population to contract, the resources that the larger population used are now available (even if (p.122) only as space), and so another population can expand. The overall effect is to partially stabilize the overall productivity of the community. There is empirical evidence that supports this cluster of ideas. Tilman's own empirical work concentrates on Minnesota grassland plots, though he also reports African data supporting similar conclusions. Species-rich plots resisted drought better, overall biomass varied less in species-rich plots, and species-rich plots returned to the predrought biomass more rapidly than species-poor plots (Tilman 1996, 358; Tilman et al. 2006). In summary, Tilman argues that both theoretical models and experimental findings support diversity-stability hypotheses when these are taken to be hypotheses about communities rather than populations (Lehman and Tilman 2000; Tilman 1996, 1999; Tilman et al. 2005).

Thus the diversity-stability hypothesis has empirical support, and, most importantly, it is based on undemanding theoretical assumptions. There is a near-consensus in ecology that, in some measure, there is a positive relationship between diversity and stability (see the consensus report, Hooper et al. 2005). There is even some suggestion from the study of fossil reef systems that this diversity-stability relationship can be documented over very long time periods (Kiessling 2005; Naeem and Baker 2005).11 However, there are problems. The experimental evidence in favor of the diversity-stability relationship depends on measuring plant biomass, but there are serious doubts about whether these ensemble relationships hold when we consider the interactions between plants and animals, and among animals. When our attention shifts to herbivores and those that eat them, resource exploitation efficiency may not be a stabilizing mechanism. To the contrary, enhanced resource use can cause Overexploitation and hence productivity collapses (Loreau et al. 2001, 807).12 David Wilson rightly points out that we should be cautious about inferences from individual efficiency to efficiency of the system as a whole. Evolutionary mechanisms reward actions that shrink the pie, so long as those that shrink it get a larger chunk of the smaller pie (Wilson 1997).

In short, though the diversity-stability hypothesis (and the related diversity-productivity hypothesis) is plausible, it is not yet demonstrable that more diverse communities are more stable (or productive). That is one reason to be wary of the conclusion that communities have causally important ensemble properties. There is a second reason for caution about this inference. Even if more diverse communities are more stable, it is not clear that they are more stable because they are more diverse. Diversity may be a symptom of causally relevant properties of individual populations rather than a causally important property of ensembles. We need to distinguish redundancy effects from sampling effects. First, redundancy. (p.123) Suppose the overall productivity of the community depends on a set of key processes. These include the acquisition of energy by primary producers; the flow of minerals to and from the abiotic substrate; decomposition by the detrivores; and the flow of organic material from organism to organism via predation, herbivory, and similar activities. Ecosystem function depends on these key processes, and ecosystems that are more diverse, and hence have a variety of species with rather different tolerances that can compensate for one another driving these processes, are thereby more stable. If this is the right story, there is redundancy that matters in the system, and diversity itself is genuinely causally important (Naeem 1998).

However, there is an alternative possibility: the sampling effect. It may be that stability depends on the presence in the ensemble of some specific species. Suppose that what we know is that more diverse communities are more likely to resist exotic invasion. Resistance might depend on the presence of a species with a specific biological profile. All else equal, rich communities are more likely to contain such a species than a species-poor community. They have more tickets in the relevant biological lotteries (Wardle 1999). Variation in traits in the community is an ensemble property. If a community has an extensive range of phenotypes, many of which contribute causally to using the available resources efficiently (hence, for example, making it more difficult for potential invaders to establish), then the diversity of the community is itself causally crucial. Not so if diversity just increases the chance that a key taxon is present, and resistance to invasion depends on that taxon.13

So to establish an emergent property hypothesis, the covariation between the emergent property and its apparent effect must be robust, not limited to a few kinds of systems. And the relationship must be genuinely causal. Tilman's hypothesis and its relatives remain plausible, but no definitive case for the diversity-stability hypothesis has yet been made. The same is true of its close relative, the diversity-productivity hypothesis. Nonetheless, for reasons that will emerge in 6.6, we think the case for causally salient, emergent properties is very strong once our attention shifts to the effects organisms, populations, and functional groups have on the habitat in which they live. Some of these effects, it seems to us, are both critical to the biological profile of habitat patches, and are ensemble effects. Perhaps the clearest example is nutrient cycling: the processes through which crucial minerals flow from and to the soil; processes mediated by countless fungi, microbes, invertebrates, and plants. Moreover, these processes are synergistic rather than additive; in the recycling process the actions of one functional group create the input for another. To the dung, its beetle.

(p.124) 6.5 Boundaries

If communities are ecological systems with causally salient properties, then, presumably, they have objective boundaries too. Thus if diversity really buffers a community against disturbance, there must be boundaries: a zone after which we stop counting, as addition of diversity there makes no difference to the extent of buffering here. Likewise, if communities are networks of interacting and self-regulating populations, there has to be a fact of the matter about which populations are part of a given network. But perhaps there are no such facts. Should we think of Karori Reserve as a single community? It is 252 hectares, and it is quite topologically varied. It is not large as the bird flies, and many birds that have been introduced to the reserve forage outside it. But for skinks, geckos, and most invertebrates, this is a sizable and diverse patch. So should we think of this as a single community, or as a multicommunity ensemble? The reserve's Web site implicitly treats it as two conjoined communities, one centered on a wetland, and the other on the regenerating forest. But skeptics doubt that these questions have objective answers (see, for example, Parker 2004).

At this point, it is important to set aside a potential contusion. Communities need not have sharp boundaries for them to have real boundaries. As with evolutionary biology, the existence of intermediate cases is no challenge by itself to the idea that—for example—communities are networks of populations whose demographic trajectories are under mutual influence. An extended family of white-winged choughs that live on the grounds of Australia's Commonwealth Scientific and Industrial Research Organisation (CSIRO), adjacent to, but foraging occasionally, on Black Mountain is somewhat influenced by events on Black Mountain. But by the interaction test, it may be neither a member of that community, nor not a member.

A more serious problem is the thought that populations will typically overlap rather than coincide, because the boundaries of particular populations will depend on their powers of dispersal, and these will vary from species to species. Roughly speaking, a population is a group of organisms of the same species that are potential mates (or rivals for mating opportunity). Mating capacity and mating rivalry depends on the mobility of organisms and their gametes. So consider Karori Reserve again. A local kaka population may overlap with a tui population, a local boobook owl population, and a number of skink and gecko populations. Moreover, if there are doubts about how to count communities in the Karori Reserve, those doubts will be much greater when we consider communities not bounded by sharp physical discontinuties. (p.125) Black Mountain is much larger: thousands rather than hundreds of hectares. It is quite diverse. The topology is varied; fire has created habitat patchiness; there are important differences in microclimate; and some of it has been farmed until quite recently. But for the most part, there are no sharp changes as one moves across this patch, no abrupt differences that will matter to most of the species present, keeping local populations congruent with one another. There is not much reason to expect the dynamics of echidna populations to match those of larger and more mobile organisms, or those of smaller and less mobile ones. Black Mountain kangaroos may well compete directly for resources and breeding opportunities with kangaroos on the O'Connor Ridge (about a kilometer to the north). Echidnas are less mobile, so O'Connor Ridge echidnas are probably a source population for Black Mountain echidnas, buffering that group against population collapses rather than competing with them for scarce resources. Here is the challenge. Communities are systems with causally consequential problems only if they have objective boundaries. But they do not seem to have such boundaries.

Richard Levins and Richard Lewontin argue that community boundaries are defined by interaction patterns rather than sharp changes in physical conditions (Levins and Lewontin 1985, 138). On the Levins-Lewontin conception, communities are systems of strongly interacting populations, where “strong” and “weak” interaction are understood comparatively: the members of a community interact strongly with one another by comparison to influences on and from populations outside the community. In their view, communities are more or less closed networks of interacting populations. Their boundaries are zones in which interactions become fewer and weaker. They are zones in which biological events—local increases and dips in population—have less impact on the populations of community members. Even so, systems of strongly interacting populations will tend to occupy an identifiable physical space. Suppose populations of goannas, currawongs, and fairy wrens interact strongly, with the effects of goannas on fairy wrens mediated by their effects on currawongs (goannas are large monitor lizards that prey on currawongs. Currawongs are large corvids that prey on fairy wrens). In most cases, the strong interaction condition will imply that the territories of the three populations largely coincide. If they do not— if, say, the goannas and currawongs only intersect moderately—many currawongs will never encounter a goanna and vice versa. As a rule of thumb, interaction requires proximity. Think of such characteristic ecological interactions as predation, herbivory, mutualistic exchange of nutrients, and pollination. None of these are interactions at a distance. (p.126) Communities of strongly interacting populations will be roughly spatially identifiable.

Thus the view that communities are organized systems does not presuppose that they are bounded by zones at which biologically important abiotic conditions change markedly. Nor does it presuppose that we can always determine whether a given population is part of a community. But it does presuppose that patterns of interaction are clumped, that most populations are parts of networks whose members interact with one another more strongly than they interact with populations outside the network. It is far from obvious that this condition is typically met. It is quite likely that ecological interactions are not clumped in ways that enable us to identify bounded communities, even taking into account the fact that community boundaries are vague.

It is hard to tell just how serious this problem is, for there are ecological processes that can generate patchiness across a habitat. Organisms do not just passively experience their environment; they actively change it. Organisms in part construct their own niches (Odling-Smee et al. 2003). This is one mechanism (as Paul Griffiths has pointed out to us) through which an initially fairly homogenous territory can turn into a mosaic of quite different patches. Niche construction is one mechanism that can magnify an initial difference between patches, beginning a cascade that takes us from initially similar systems to a mosaic of quite different patches. Suppose, for example, that by chance eucalypts rather than acacias happen to initially predominate in one zone. Eucalypts have different environmental effects than acacias. They grow more slowly, but they live longer and eventually afford many animals homes in the hollows that form in them. They support very different pollinators; honeyeaters visit their flowers but not those of acacias. They produce very different litter. So an initial difference can generate quite marked differences between adjoining patches, thus generating two somewhat closed networks of interacting populations. We certainly cannot be confident that niche construction effects will increase landscape-scale heterogeneity, creating mosaic effects that bring populations of many species into rough spatial alignment with one another. But it is one possibility.

6.6 The Space of Population Assemblages

There is no definitive case for causally salient community properties. Regardless, we think that many ensembles have such properties. Organisms are profound agents of transformation both of their own and others' environments,14 and recognizing that fact greatly strengthens (p.127) the case for thinking that communities are structured and have ensemble properties. Populations within a community can be linked via niche construction networks. One population can influence another by changing important features of the physical environment. Trees buffer the wind and modulate the impact of storms while providing shelter to many organisms (Jones et al. 1997). These indirect ecological links expand the range of potential interactions in communities. Populations act on one another via the physical changes they induce. Litter recycling is the cleanest example. Plants produce litter as a by-product of their life: fallen leaves, twigs, bark. A host of organisms live by consuming the litter, and as a consequence of these actions, they return crucial materials to the soil. This is absorbed by the vegetation, which in turn produces more litter (Odling-Smee et al. 2003, 318–22). Moreover, niche construction often involves ensemble effects. Soils are made organically, but not by any single population. A vast suite of very different animals, plants, and fungi make soils. Ants and other burrowing animals turn over and redistribute soils; trees and other plants stabilize soils; fungi, microbes, and a vast army of small invertebrates make soil by consuming litter. Thus the discussion in 6.3 of community regulation and of assembly rules understates the case for causally salient, system-level properties by focusing so exclusively on density dependent forces, of which competition and predation are the prime examples (Callaway 1997). The individualist view of ecology looks much more plausible when niche construction is neglected.

We have contrasted the idea that communities are causally salient, internally regulated, bounded systems with the individualist idea that they are mere aggregates of overlapping populations that happen to have fairly similar physical and biological tolerances (and hence are merely phenomenological systems). But as the last few sections have noted, we have here a spectrum of possibilities. The crucial factors that distinguish an assemblage of indifference from a functionally organized community come in degrees. Thus an assemblage may be internally regulated to some extent. The power of internal regulation will depend on the proportion of the component populations that interact strongly enough to influence abundance in the community, the strength of those interactions, and their stability in the face of outside disturbances. Likewise, an assemblage may have causally salient emergent properties to some extent. Suppose, for example, that more diverse communities really are more stable. But stability comes in degrees and in different forms. The importance of the stabilizing effect of diversity will depend on the degree to which diversity buffers the community against disturbance, the range of properties that are buffered against disturbance, and (p.128) the kind of disturbances whose effects are muted. Boundedness, too, comes in degrees; a network of interacting populations can be more or less closed; more or less spatially coincident rather than merely intersecting. Moreover, and most importantly, these factors may be partially independent of one another. If, for example, the stabilizing effects of diversity depend on compensation, an assemblage can have causally important emergent properties without being internally regulated.

We began this chapter by identifying phenomenological communities and noting that they play an important role as biodiversity surrogates. Identifying a community as a river wetland, an alpine grassland, or as a southeastern slopes eucalypt woodland gives us a reasonable guide to its alpha diversity. Identifying physically adjacent communities as phenomenologically distinct—a riverine community next to grassland habitat—likewise gives us a reasonable guide to their beta diversity. So in an undemanding sense, local ecological communities are units that we should recognize and count, and there are important differences between landscape-level ecological systems that contain many different phenomenological communities and those that are more homogenous. Beta diversity, for example, will be much higher, and invasion and perhaps other disturbances are less likely to spread uniformly through the landscape. Keeping track of phenomenological community diversity is likely to be predictively important for ecology and conservation biology, whether or not local communities are organized systems. We do not expect this claim to be controversial, so in this chapter we have concentrated on the much more controversial issue: whether local communities are units that must be recognized in causal explanations of ecological processes, and, if so, how they differ one from another. Hence our focus on the causal questions: are communities organized systems, with their own effects on their own membership and abundance? If so, do they differ systematically from one another in their organizational properties?

We are not in a position to answer this question, but we think we have developed a useful framework for its investigation. The features that make communities explanatorily salient come in degrees and are potentially independent of one another. These dimensions define an ecospace; different local communities will differ from one another in this space, not because of their phenomenological differences—differences in membership and physical environment—but because of their differences in causal organization. Thus we think it is productive to think of specific local communities as occupying differing positions in a 3-D space: the three dimensions being boundedness, internal regulation, and emergent property effects, rather than physical gradients (p.129) like temperature or rainfall. These dimensions define a space of possibilities: specific ensembles at specific times and places will be more or less bounded; more or less internally regulated; have or lack important system level properties—properties like buffering against disturbance, nutrient cycling, fire-resistance—to some degree (as in fig. 6.2). A maximally indifferent assemblage in one corner of the space would consist of a set of populations that merely overlap, which do not significantly influence one another's demographic prospects, and which have no important collective impact on their environment. In the opposite corner of the space, there would be assemblages consisting of spatially coincident populations strongly influencing one another's demographic fates, and with important ensemble effects.

We like this way of representing the nature of communities, for it suggests an important research agenda. One set of questions will be about the circumstances under which communities of indifference become organized (and vice versa). Are sharp abiotic gradients important in aligning populations of different species or could niche construction effects also generate the right kinds of environmental patchiness? Disturbance, too, might play a role in generating patchiness; fire and flood are sources of sharp abiotic gradients. However, there is also a line of thought suggesting that disturbance can turn integrated communities into communities of indifference. There is suggestive (though far from decisive) evidence that Pleistocene communities are closer to being communities of indifference than earlier paleoecologies. Ecological associations

Explorations in Ecospace

FIGURE 6.2. The space of population assemblages.

(p.130) between species were unstable over the Pleistocene, but they may have been less so in earlier environments (Valentine and Jablonski 1993; Coope 1994). The idea here is that the intensity of Pleistocene climatic fluctuations exceeded a threshold, causing community organization to break down, replacing regulated communities with unregulated ones. If true, this is clearly very important.

A second set of questions concerns the dimensions of ecospace. It's quite possible that our three dimensions already lump together organizational features that should be considered separately. For example, internal regulation might be achieved via niche construction, but it may also be the result of strong competitive interactions or obligatory mutualisms. It is not obvious that these characteristics should be aggregated into a single scale. Likewise, there may be a number of important and conceptually independent ensemble properties. The niche construction literature indicates that ensemble effects on physical features of the habitat may be important. For example, different suites of vegetation have markedly different effects on the water table and salinity of local systems. Western Australia, in particular, is suffering severe salinity problems because woodland was replaced by pasture, causing the water table to rise. Surface water then evaporates in hot dry seasons, leaving a salt residue. It may be that there are several dimensions here, not just one. Moreover, there are candidate dimensions we have not considered at all. One is openness to migration from the regional species pool. Island biogeography (and its more nuanced descendants) suggests that openness contributes importantly to richness and to stability (Ricklefs and Schluter 1993; Ricklefs 2004).

A third set of questions concerns the distribution of actual assemblages in ecospace, and in particular, whether there are correlations between phenomenological community types and location in our causal ecospace. Desert communities are phenomenologically similar, but are they in roughly the same space? Are all open woodlands? These questions offer an alternative approach to the vexed problem of the contingency of ecology we noted in 6.1. Even if the specific composition of communities is sensitive to the accidents of arrival and establishment, their structural properties may be predictable. Suppose, for example, that grasslands with different histories are nonetheless clumped in a particular volume of ecospace; imagine that they have ensemble properties but are not tightly regulated. This would indicate that in one important respect the processes that assemble grassland communities are not contingent. They build communities with similar structural properties, even if those communities have different members. Alternatively, the case for contingency would be strengthened if grassland (p.131) communities were scattered through ecospace. One response to the difficulties involved in generating precise predictions about the changes in communities over time has been to argue that ecological theory was focused on the wrong spatial scale: ecologists should develop predictions about regional rather than local processes (Gaston and Blackburn 1999; 2000). A different response is to develop predictions about more abstract or global features of local communities rather than predicting the fates of particular actual and possible populations within them. In effect, we have seen this in Tilman's and Naeem's responses to May; they make predictions about overall productivity or stability rather than about the fate of particular populations. The suggestion here is a generalization of that approach to the contingency problem.

If individualism is essentially correct, information about species composition and numbers (together with information about the physical environment) captures ecological dynamics; there is no extra biological structure we need information about to explain ecological patterns. There would be no independent ecological ingredient to biodiversity. That fits the fact that ecological diversity has typically been characterized phenomenologically: by appeal to the physical parameters of the habitat (shallow, wave-influenced vs. deep water benthic communities) or by the predominant biota (pine forests, grasslands). As we have emphasized, such phenomenological characterizations are predictively useful. We can make a fair guess at what we can expect to see in a grassland. Grass, for example, will not be a surprise. But we think that characterizing local assemblages in terms of their positions in this abstract space opens up a research agenda in ecology in ways that phenomenological characterizations do not. It is a way of investing the open empirical possibility that there is extra biological structure that plays a central role in explaining ecological patterns. We think it is likely that there is such structure, and hence we need to go beyond species abundance and distribution information in explaining and predicting ecological patterns. But since we have focused entirely on one spatial scale in ecology, that of local ecological communities, we certainly have not proved that in this chapter. Indeed, we have not done much more than begin a preliminary sketch of this extra biological structure.

We now return to conservation biology, armed with an appreciation of the power of species-information-based measures of biodiversity, of the limits of that conception of biodiversity, and of the difficulty of supplemented species-based models in a disciplined and tractable way.


(1.) See, for example, Gaston (2000).

(2.) For a good discussion of these issues, see Ward and Thornton (2000). They argue that early in their reassembly, Krakatau island communities do not show much path dependency. Predictable, good colonizers arrive early and everywhere. Order effects exemplify contingency only if the order of arrival is itself unpredictable. But communities then diverge as a result of chance differences in colonization. These differences have caused at least a significant delay in reaching of the equilibrium condition that seems to be dominated by Dysoxylum forest.

(3.) That is, within-patch richness versus between-patch richness; more on this in chapter 7.

(4.) This debate using this terminology began with an important paper by Jared Diamond (1975) on New Guinea bird faunas, in which he argued that competition for resources structured communities. Birds whose needs were too similar could not both be present, but nor would communities have, at equilibrium, underutilized resources either: species are densely packed into communities.

(5.) Jay Odenbaugh (2006) has introduced some useful terminology to describe different ensembles. “Gleasonian” communities contrast with “Hutchinsonian” and “Clementsian” communities. These community types are named for the ecologists who famously defended different views of communities. Gleason was an early American ecologist who argued that populations respond to their environments largely independently of one another. Hutchinsonian and Clementsian communities are distinguished by the strength of the interactions among their component populations. Clementsian communities, as Odenbaugh defines them, have strongly interacting components, whereas populations in a Hutchinsonian community may interact weakly.

(6.) For an insightful discussion of this complex of ideas, see Cooper (1993; 2001; 2003) and Pimm (1991).

(7.) We have presented this argument from persistence to regulation as if it were a single view. In reality, there is a family of views that accord the “balance of nature” a central role in regulating the internal organization of communities. In part, this (p.184) is the obvious point that stability comes in degrees. But it is also due to the fact that as Stuart Pimm and others have shown, the notion of stability is itself ambiguous; see Pimm (1991) and Lehman and Tilman (2000). There is a family of notions, and hence a family of ideas, about the extent to which communities are stable. But according to all these views, communities have a genuine organization. The organisms already present and interacting will make certain roles available and will foreclose others. For a particularly good discussion of the range of stability-conceptions that are lumped together, see pp. 118–21 of Sarkar (2005).

(8.) These are two of New Zealand's richer and more species-rich tidal mudflat communities—though nowhere near as rich as equivalent Australian systems.

(9.) Although there are not very many studies demonstrating this; for one, see Woodward (2002).

(10.) As we have noted, stability comes in many forms (Sarkar 2005, 118–21). The literature that has grown up around Tilman focuses on variance around a mean over time.

(11.) This data would have to be interpreted very cautiously. Paleoecological data does not record events on the same spatial and temporal scales as ecological studies on local communities.

(12.) This issue is fraught and controversial. There is a rich literature on food webs and the effect of the complexity and depth of food webs on stability. See, for example, Pomeroy (2001), Montoya et al. (2003), and Thebault and Loreau (2003).

(13.) The lottery effect raises subtle issues about causation as well as ecological mechanism. Drift has often been interpreted as a random change in populations over time, but recently, Patrick Forber and Ken Reisman have pointed out that on a manipulationist conception of causation, drift is a cause of population change. We can intervene on population size to change the importance of drift; we can manipulate the impact of drift on evolution. In small populations, it is more important. So drift is a lever we can use to influence evolutionary trajectories, and that makes it a cause. On a similar criterion, even if the lottery effect is the mechanism of the diversity-stability connection, species richness is the cause of stability. We can manipulate the impact of the lottery effect, and hence stability, by manipulating species richness. See Reisman and Forber (2005).

(14.) Levins and Lewontin made this point forcefully, and their ideas have been further developed by those working on niche construction and ecological engineering (Levins and Lewontin 1985). See also Jones et al. (1997) and Odling-Smee et al. (2003).