Intraspecific aggregation and species coexistence Comment from Murrell, Purves and Law Our recent article in TREE [1] was motivated by the lack of empirical information on the effect of spatial structure on competition. We thank Rejmanek [2] for drawing attention to several further empirical studies. Chesson and Neuhauser [3] create the impression that, in the absence of life-history tradeoffs, coexistence of competing species becomes less likely when spatial structure is considered. The basis for this is an analysis by Neuhauser and Pacala [4] that concluded that local interactions in a spatial version of the Lotka–Volterra competition model would reduce the parameter space of coexistence. We offer the following counterexample (model from [4] with small modifications) in which the spatial extension causes the coexistence of two species (Fig. 1). In this example, the first species is a stronger competitor and leads to extinction of the second in the non- spatial Lotka–Volterra system. But the distance over which interactions between species occur is shorter than that within species. This, together with the spatial segregation of the species, reduces the strength of interspecific competition sufficiently to permit either species to invade the other. We have not invoked a life- history tradeoff (e.g. the familiar competition–colonization tradeoff) to achieve coexistence here. Moreover, there is nothing intrinsic to coexistence here that says that it has to be due to niche differentiation. But, this is not to say that, if competition extends to conspecific neighbors that are more distant than heterospecifics, niche differentation is not a potential mechanism; different distances could, for instance, be caused by host-specific enemies that aggregate around parents (the Janzen–Connell hypothesis). That we get dynamics different from [4] is not surprising – the models have several different assumptions. Six extra functions are needed in the spatial version of the two-species Lotka–Volterra competition model to deal fully with local interactions and local dispersal, and each function has at least one parameter [5]. To make analysis tractable, theoreticians have had to use simplifying symmetries in the interactions; investigation of how the extended parameter space affects asymptotic and transient coexistence has barely begun. Coupling of spatial structure to population dynamics is intricate and it would be unwise to assume either that aggregation always leads to exclusion or the reverse. It is most likely that there are some conditions under which spatial structure promotes coexistence and others under which it does not: the former obviously deserve special attention. David Murrell* NERC Centre for Population Biology, Imperial College at Silwood Park, Ascot, Berkshire, UK SL5 7PY. *e-mail: d.murrell@ic.ac.uk Drew Purves Dept of Ecology and Evolutionary Biology, Princeton University, Princeton, NJ 08544, USA. Richard Law Dept of Biology, University of York, PO Box 373, York, UK YO10 5YW. References 1 Murrell, D.J. et al. (2001) Uniting pattern and process in plant ecology. Trends Ecol. Evol. 16, 529–530 2 Rejmánek, M. (2002) Intraspecific aggregation and species coexistence. Trends Ecol. Evol. 17 10.1016/S0169- 5347(02)02494-1 3 Chesson, P. and Neuhauser, C. (2002) Intraspecific aggregation and species coexistence. Trends Ecol. Evol. 17 10.1016/S0169-5347(02)02482-5 4 Neuhauser, C. and Pacala, S.W. (1999) An explicitly spatial version of the Lotka–Volterra model with interspecific competition. Ann. Appl. Prob. 9, 1226–1259 5 Law, R. and Dieckmann, U. (2000) A dynamical system for neighborhoods in plant communities. Ecology 81, 2137– 2148 Fig. 1. In the non-spatial Lotka–Volterra model (heavy solid and dashed lines), the strong competitor (red) drives the weak one (blue) to extinction. In the spatial model, by allowing interspecific interactions to occur over a shorter range than do the intraspecific interactions, the weaker competitor is able to coexist with its rival. This is shown both in a stochastic model (uneven lines) averaged over 40 realizations and also in a deterministic approximation based on moment dynamics (light solid and dashed lines). Apart from the weak and strong competition, the two species have the same parameter values. TRENDS in Ecology & Evolution 0 50 100 150 200 250 0 100 200 300 400 Density Time