Forces that have biased research on competition

Why have the aspects of consumer and resource dynamics reviewed above been ignored in the vast majority of those systems where competition has been studied? A scientific approach to this question about the direction of research is not possible.

‘Assumption drag' is a commonly used term for the persistence of assumptions when they have been used often enough, regardless of evidence against their validity. It is particularly problematic for ecology, given the incredible range of different com­petitive relationships. The continued frequent use of the LV model and MacArthur’s consumer-resource model both reflect assumption drag. The prevalence of function­al responses that are functions of the abundance of a single resource is another case of assumption drag within consumer-resource theory. Although specialist consumers are rare, the fact that C. S. Holling originally defined functional responses in terms of a single food/resource type has led to the dominance of such relationships in both the empirical and theoretical literatures. This tendency is reflected in the comprehensive functional response review articles by Jeschke et al. (2002, 2004).

In theoretical biology, using new or difficult mathematical techniques or deriv­ing extremely general results are both important avenues for obtaining recognition. Developing a large body of specific results dependent on simulations of many alter­native models usually does not translate into career advancement for theoreticians.

Research on the functional responses of consumer species is a good example of how the structure of the research environment discourages a broader perspective. The fact that many measurements of functional responses have already been pub­lished makes any new study less likely to attract scientific attention. Because few studies since Abrams (1980a) have analysed the effect of functional response shape on competition, the functional response shapes of competing consumers are almost always unknown. An additional problem is that multi-species functional response measurements require a large number of treatments, which often makes them techni­cally impossible. The fact that competitive systems, particularly those involving biotic resources, often take a substantial amount of time to come to equilibrium has led to a marked lack of studies of organisms with relatively ‘slow’ life histories. All of this leads to a high level of ignorance regarding the range of functional responses in competitive systems involving multiple resources.

Competitive interactions involving cycling species are also rarely studied. In these cases, standard measures of interspecific effects require calculating an average over the course of a complete cycle, or over a large number of ups and downs in a sys­tem with an aperiodic attractor. Keeping a laboratory system going for such periods is seldom possible, and the many years required for field studies of cycling verte­brate populations makes studies of interactions even less likely.

The semi-random process of particular ideas becoming popular was highlighted in the quotation from David Houle in the first paragraph of this chapter. The idea of separating coexistence into the two categories of stabilizing and equalizing effects is currently attracting a huge amount of attention, but it seems unlikely to lead to any useful results for truly understanding competitive processes.

Finally, the urge to oversimplify is a potentially powerful source of bias in most research in most areas of science. Simplification in competition theory finds its ultimate embodiment in the LV model. However, even within the framework of consumer-resource models, the focus on 2-consumer-2-resource models, combined with isocline or invasibility analysis of those models, has narrowed the subject mat­ter of competition studies in a detrimental way. The overuse of simple models has likely been promoted by a seldom-stated idea that simple models are in some sense more general. This applies to the overuse of the MacArthur model within studies of consumer-resource systems. A linear functional response is intermediate between concave and convex responses, so it is thought to potentially apply to some extent to both cases. However, this justification fails if most real responses are strongly non­linear, and if nonlinear response categories share some dynamical properties that are lacking in the linear case. Both of these conditions happen to be true.

The consumer-resource models used in this book are still much too simple to be used for quantitative predictions for any natural system. However, the diversity of dynamics demonstrated by these models is very likely to be part of the repertoire of potential dynamics in real-world systems. And if they are not, then there must be additional features or restrictions in natural systems that we are currently unaware of, or have not yet studied. Knowing the properties of intermediate-complexity consumer-resource models will be part of the knowledge needed to uncover what those additional features might be.

12.7