It might be imagined that scientific progress is made by providing answers to questions. In fact, progress often consists of replacing one question with another, more pertinent, more challenging question. In this section, we deal with an area where this is the case: the questions of how different coexisting competitors are, and how different coexisting species need to be if competition is not to eliminate one of them.
Species Species Species 1 2 3
Species Species Species 1 2 3
Figure 8.29 Resource-utilization curves for three species coexisting along a one-dimensional resource spectrum. d is the distance between adjacent curve peaks and w is the standard deviation of the curves. (a) Narrow niches with little overlap (d > w), i.e. relatively little interspecific competition. (b) Broader niches with greater overlap (d < w), i.e. relatively intense interspecific competition.
The Lotka-Volterra model predicts the stable coexistence of competitors in situations where interspecific competition is, for both species, less significant than intraspecific competition. Niche differentiation will obviously tend to concentrate competitive effects more within species than between them. The Lotka-Volterra model, and the Competitive Exclusion Principle, therefore imply that any amount of niche differentiation will allow the stable coexistence of competitors. Hence, in an attempt to discover whether this was 'true', the question 'do competing species need to be different in order to coexist stably?' greatly exercised the minds of ecologists during the 1940s (Kingsland, 1985).
It is easy to see now, however, that the question is badly put, since it leaves the precise meaning of 'different' undefined. We have seen examples in which the coexistence of competitors is apparently associated with some degree of niche differentiation, but it seems that if we look closely enough, all coexisting species will be found to be different - without this having anything to do with competition. A more pertinent question, therefore, would be 'is there a minimum amount of niche differentiation that has to be exceeded for stable coexistence?' That is, is there a limit to the similarity of coexisting species?
One influential attempt to answer this question for exploitative competition, based on variants of the Lotka-Volterra model, was initiated by MacArthur and Levins (1967) and developed by May (1973). With hindsight, their approach is certainly open to question (Abrams, 1983). Nevertheless, we can learn most about the 'limiting similarity problem' by first examining their approach and then looking at the objections to it. Here, as so often, the models can be instructive without being 'right'.
Imagine three species competing for a resource that is unidimensional and is distributed continuously; food size is a clear example. Each species has its own realized niche in this single dimension, which can be visualized as a resource-utilization curve (Figure 8.29). The consumption rate of each species is highest at the center of its niche and tails off to zero at either end, and the more the utilization curves of adjacent species overlap, the more the species compete. Indeed, by assuming that the curves are 'normal' distributions (in the statistical sense), and that the different species have similarly shaped curves, the competition coefficient (applicable to both adjacent species) can be expressed by the following formula:
where w is the standard deviation (or, roughly, 'relative width') of the curves, and d is the distance between the adjacent peaks. Thus, a is very small when there is considerable separation of adjacent curves (d/w >> 1; Figure 8.29a), and approaches unity as the curves themselves approach one another (d/w < 1; Figure 8.29b).
How much overlap of adjacent utilization curves is compatible with stable coexistence? Assume that the two peripheral species have the same carrying capacity (K1, representing the suitability of the available resources for species 1 and 3) and consider the coexistence, in between them on the resource axis, of another species (carrying capacity K2). When d/w is low (a is high and the species are similar) the conditions for coexistence are extremely restrictive in terms of the K1 : K2 ratio; but these restrictions lift rapidly as d/w approaches and exceeds unity (Figure 8.30). In other words, coexistence is possible when d/w how much niche differentiation is needed for coexistence?
a simple model provides a simple answer...
is low, but only if the suitabilities of the environment for the different species are extremely finely balanced. Furthermore, if the environment is assumed to vary, then the fluctuations will lead to variations in the K1 : K2 ratio, and coexistence will now only be possible if there is a broad range of K1 : K2 ratios leading to stability, i.e. if, roughly, d/w > 1.
This model, then, suggests that there is a limit to the similarity of coexisting competitors, and that the limit is given by the condition d/w > 1. Are these the correct answers? In fact, it seems most unlikely that there is a universal limit to similarity, or even a widely applicable one that we could express in such a simple way as d/w > 1. Abrams (1976, 1983), amongst others, has emphasized that models with competition in several dimensions, with alternative utilization curves and so on, all lead to alternative limits to similarity, and often to much lower values of d/w being compatible with robust, stable coexistence. In other words, 'd/w > 1' is a property of one type of model analysis, but not of others, and thus, almost certainly, not of nature as a whole. Furthermore, we have already seen that because of environmental heterogeneity, apparent competition and so on, exploitative competition and any niche differentiation associated with it are not necessarily the whole story when it comes to the coexistence of competitors. This too argues against the idea of a universal limit.
On the other hand, the most general messages from the early models still seem valid, namely: (i) in the real world, with all its intrinsic variability, there are likely to be limits to the similarity of coexisting exploitative competitors; and (ii) these limits will reflect not only the differences between species, but also the variability within them, the nature of the resource, the nature of the utilization curve and so on.
But is even the limiting similarity question the best question to ask? We want to understand the extent of niche differentiation amongst coexisting species. If species were always packed as tightly together as they could be, then presumably they would differ by the minimum (limiting) amount. But why should they be? We return, once again, to the distinction between the ecological and the evolutionary consequences of competition (Abrams, 1990). The ecological effects are that species with 'inappropriate' niches will be eliminated (or repelled if they try to invade), and the limiting similarity question implicitly concerned itself with this: how many species can be 'packed in'? But coexisting competitors may also evolve. Do we generally observe the ecological effects, or the combined ecological and evolutionary effects? Do they differ? We cannot attempt to answer the first question without answering the second, and the answer to that seems to be, perhaps inevitably, 'it depends'. Different models, based on different underlying mechanisms in the competitive process, predict that evolution will lead to more widely spaced niches, or to more closely packed niches or to much the same disposition of niches as those predicted by ecological processes alone (Abrams, 1990).
Two points, therefore, emerge from this discussion. The first is that it has been entirely theoretical. This is a reflection of the second point, which is that we have seen progress - but in terms of successive questions superseding their predecessors rather than actually answering them. Data provide answers - what we have seen is a refinement of questions. The latest stage in this appears to be that attempts to answer questions regarding 'niche similarity' may need to be postponed until we know more about resource distributions, utilization curves and, more generally, the mechanisms underlying exploitative competition. It is to these that we now turn.
Was this article helpful?