General themes of this book

I will argue that both the LV and many of the more recent models are indeed insuffi­cient, in that they do not reflect many of the dynamic characteristics of large classes of slightly more realistic consumer-resource systems involving shared resources.

A current preoccupation seems to be determining to what extent coexistence is due to ‘stabilizing’ and ‘equalizing’ factors, neither of which is well defined. In con­sidering the role of temporal variation in altering competition, the focus has been on cases when variation favours coexistence. Within that set of cases, the goal has been to classify the mechanism as being ‘the storage effect’ or ‘relative nonlinearity’, or some mixture of the two, following Chesson’s (1994) categorization of coexistence mechanisms. While there is no doubt that useful insights have grown out of Chesson’s approach, the goal of incorporating competition into applied disciplines like conser­vation or exploiting natural resources seems to have disappeared, and classification is of little use here.

One of the problems contributing to the development of competition theory applies much more generally to theory in ecology. It is the idea that all theory needs to be tested by determining whether some prediction of the theory applies across a wide range of real (preferably field) systems. The problem is that the theory that exists is so simplified that, without additional work on slightly more complicated models, we will not know whether a particular prediction has any chance of being ‘true’ in some sense of the word. If a prediction from theory that includes some feature is not manifested by one or more natural systems, this does not mean that the new feature should be discarded from future models; it is more likely to signal that other missing features need to be added. In other words, more theory is needed to allow empirical scientists to have some idea of what should occur (or have occurred) in the system being studied. Coming up with intelligent hypotheses to test empirically requires a large theoretical infrastructure.

The enormous range of ecological communities also means that there will never be a model that applies universally. The logical way to proceed from the LV model would be to incorporate a range of real-world features that are missing, but are likely to be present in real systems. At some point, this should provide a robust framework for choosing a range of alternative explanations for some observed population dynamical result involving competitors. MacArthur’s original consumer-resource system was an initial step in the direction of building towards such a body of theory. Modify­ing MacArthur’s system to include nonlinear components and interactions between resources were initial steps towards expanding the model, begun in the 1970s.

Suppose a new invasive species begins to increase rapidly, and an ecologically sim­ilar resident species decreases at an even more rapid rate. Can we expect the same relative changes in abundance to continue in the future? The Lotka-Volterra mod­el implies they would do so. Some of the more plausible consumer-resource models suggest that they should not, and that decreasing per individual effects on the resi­dent are more likely as the invader population grows (Abrams et al. 2008b). A useful theory of competition should at least suggest what information is required to predict the most likely consequences for the native species as the invading one increases.

Ecological systems are both extremely complex and difficult to study without alter­ing the system itself. To come up with a hypothesis about why one or more species is changing rapidly in abundance, one must have some idea of the important process­es that could be operating. This must come from theory based on simpler systems. Models that incorporate the key processes of resource use are essential in generating basic hypotheses that might underlie the observed change.

Part of the problem with ‘mechanistic models' that include more variables is that determining which one applies to a given system requires work that is not attractive to individual researchers. For example, what is the range of forms of relationships between food abundances and consumption rates (i.e., functional responses)? Which are most likely to apply to certain scenarios or species? A single study will only con­tribute a small amount towards answering these broader questions, but many such studies on a range of species would usually require much more time than that avail­able during a graduate student career or during the duration of a single research grant.

There is still hope that these historical influences on both the study of consumer­resource interactions, and the science of competition more broadly, will diminish in the future. I am not arguing that ecologists should ignore past work; in fact, I believe the opposite is true. The problem has been that, when an area drops out of favour and then returns (as competition has done), there is a tendency to ignore much of that history, and concentrate on the earliest parts. The history of empirical and theoretical approaches to competition has been characterized by abrupt changes in the degree of general interest in the topic as well as the underlying assumptions made and questions asked. These seem to indicate that a change in focus is possible. Perhaps this gives hope for the future.

This book will return to these themes, and will end with a set of recommendations for producing a more useful competition theory in the future. Such a body of theory could provide the ‘guide for action' called for by Michel Loreau in the quotation that introduced this chapter. This is Iikelyto be my last publication in this area, to which I have devoted most of my career, and I wanted to provide a unified account of articles produced in many different decades and which appeared in many different journals. As a consequence, readers will note that rather large amounts of the text deal with ‘old’ work and also with my own work.

1.7