Plant evolutionary ecology


The advent of the use of molecular and genomic approaches in plant evolutionary biology has led to a quantum leap forward in our understanding of the genetic underpinnings of plant adaptation and speciation. For example, access to individual genes that influence characters such as disease and herbivore resistance (Bishop et al., 2000; Kroyman et al., 2003; Mauricio et al., 2003) and flowering phenology (Stinchcombe et al., 2004) has allowed determination of the relative roles of natural selection and genetic drift in shaping ecologically important traits. Genomic analysis has revealed that repeated polyploidization has been a major factor in angiosperm evolution (Vision et al. 2000), the implications of which are just beginning to be investigated (Adams et al., 2003). Population genetic analyses of nucleotide variation are beginning to demonstrate the importance of regulatory sequences in morphological evolution (Doebley, 1993; Purugganan, 2000). Finally, extensive genetic mapping and QTL analyses are beginning to reveal the genetic architecture of species differences (Fishman et al., 2002; Hodges et al., 2002) and of reproductive isolation (Bradshaw et al., 1998; Schemske & Bradshaw, 1999).

‘It is important to recognize that genetics is only half the story of plant evolution.’

Impressive as these achievements are, however, it is important to recognize that genetics is only half the story of plant evolution. Although sequence analysis can reveal much about the kind of genetic variation that natural selection uses to mold evolutionary change, it can often tell us very little about why particular variants are chosen by selection and others excluded. Addressing this issue usually requires understanding the ecological basis of adaptation, the nature of the diverse interactions between a plant and its physical and biotic environment, and how these interactions generate the particular patterns of natural selection that favor one variant over another. In addition, it requires understanding how specific adaptations, such as a plant's mating system, can themselves influence genetic variation and generate selection on other characters. Ultimately, a complete understanding of evolutionary change will require the unification of both the molecular/genomic and the ecological approaches. In this issue, we feature a set of reviews that focuses primarily on ecological aspects of evolution, examining both the causes and consequences of mating system evolution, and the evolution of interactions between plants and their natural enemies.

Mating system evolution

One remarkable evolutionary feature of angiosperms is their great diversity of mating systems. Explaining the evolution of this diversity has been a long-standing issue in evolutionary biology. One aspect of this diversity is the presence, in many species, of polymorphisms for floral morphology. Most work aimed at attempting to understand the evolution and maintenance of such polymorphisms has concentrated on species exhibiting heterostyly, in which intramorph mating is restricted by both morphological adaptations and physiological incompatibility mechanisms (Charlesworth, 1979). Equally interesting, however, are floral polymorphisms that involve morphology but not intramorph incompatibility. Barrett & Harder (pp. 45–53 in this issue) describe aspects of the evolution of such a system in the genus Narcissus (daffodils). One issue they address is the controversy over the evolutionary route to distyly and tristyly (Charlesworth & Charlesworth, 1979; Lloyd & Webb, 1992a,b): did physiological incompatibility arise first and establish two (or three) mating types, followed by morphological differentiation of the mating types due to selection for reduction in pollen wastage? Or was morphological differentiation initially selected for as a means to increase the efficiency of pollen deposition, and only afterwards reinforced by physiological incompatibility with morphs? Phylogenetic analysis of floral morphology in Narcissus is shown to favor one of these explanations over the other. In addition, Barrett & Harder demonstrate that the absence of within-morph physiological incompatibility allows for deviations from equal morph ratios at equilibrium. They describe how details of floral morphology and environmental variation help explain both interpopulation and interspecies variation in morph frequencies.

A long-standing mystery of mating-system evolution in plants is the question of what determines whether a population evolves to be largely outcrossing, largely selfing, or exhibit a mixture of outcrossing and selfing. Theoretical analyses have long suggested that the magnitude of inbreeding depression (ID) has a significant influence on the evolutionary outcome (Lloyd, 1979; Holsinger, 1988). More recently, however, it has become realized that a crucial factor in this process is whether natural selection and inbreeding combine to generate associations among loci affecting fitness and loci affecting the mating system. If such associations are strong, alleles that increase selfing may be favored despite a high average level of ID (Uyenoyama et al., 1993). Unfortunately, very little is understood about the prevalence and magnitude of these associations in natural plant populations, primarily because straightforward diagnostic assays have been lacking. Kelly (pp. 55–62) addresses this problem by describing how minor modifications of the standard quantitative genetics design for detecting variation in family level inbreeding depression can be used to examine the magnitude of these associations. In addition, he provides a simple test for determining whether ID is caused primarily by overdominance or deleterious alleles.

One of the features of plant mating systems for which we are closest to understanding both the genetic underpinnings and the ecological factors molding it is self-incompatibility. Incompatibility systems that have been extensively characterized typically exhibit tens to hundreds of segregating alleles that are responsible for self-recognition (or, perhaps more properly, non-self-recognition). Almost certainly, it is a form of frequency dependence in which rare alleles have a mating advantage with respect to pollen donation that maintains this extensive allelic diversity (Wright, 1939). However, there is a curious genetic aspect to this type of self-incompatibility system. Each incompatibility ‘allele’ actually consists of at least two genes: one determining pollen specificity, and one determining pistil specificity. Alleles at these two loci are held in almost complete linkage disequilibrium. Although these facts have been understood for some time, there has been little consideration of how these ‘alleles’ were built up and how linkage disequilibrium is maintained. Uyenoyama (pp. 63–70) provides a unified conceptual framework for understanding these and other features of the genetic organization of self-incompatibility ‘alleles’, and at the same time describes how this framework can be extended to understanding the evolution of genetic regions involved in mating-type recognition in organisms as diverse as plants, fungi and mammals.

Evolution of plant–enemy interactions

Next to competition, perhaps the most prevalent type of biotic interaction experienced by plants is attack by natural enemies, including bacterial, viral and fungal pathogens and herbivores. In the case of pathogens, recent theoretical work makes it clear that the expected coevolutionary responses (e.g. level of virulence attained by pathogens, level of resistance attained by plants) to selection imposed on each other by plants and their natural enemies is greatly influenced by both the mechanism of disease transmission (e.g. whether it is frequency- or density-dependent) and the fitness effects of the disease (primarily on viability vs primarily on fecundity) (Anderson & May, 1982; Frank, 1996; Lipsitch et al., 1996; O’Keefe & Antonovics, 2002). One issue that remains unclear, however, is whether knowledge of these two aspects of a particular plant disease are sufficient to predict the trajectory of coevolution, or whether there are other factors that must also be considered – how many ecological factors must be included in order to delimit categories of evolutionary interactions between plants and their enemies? For example, if two diseases are both ‘venereal diseases’, diseases that exhibit frequency-dependent transmission and that reduce host fecundity but not viability, are they likely to evolve in similar ways? By considering the implications of classifying a disease as being sexually transmitted, Antonovics (pp. 71–80) provides a tentative answer to this question: ecological and evolutionary dynamics are likely to depend very much on a myriad of details of the disease-transmission process. Moreover, he shows very persuasively how plants can be used as model systems for developing a conceptual framework for understanding disease evolution.

A general principle guiding much research examining the evolution of plant–enemy interactions is the idea that the trajectory of coevolution between a plant and an enemy can be greatly influenced by the biotic community within which that interaction is embedded. In particular, whether coevolution is primarily pairwise or diffuse (Janzen, 1980; Fox, 1981) is expected to influence both the rates at which plants and their enemies coevolve, as well as the extent and nature of resistance at evolutionary equilibrium (Gould, 1988; Rausher, 1996). Although a number of approaches for distinguishing between pairwise and diffuse coevolution have been suggested (Faeth, 1986; Strauss, 1991; Hougen-Eitzman & Rausher, 1994; Iwao & Rausher, 1997), there remain questions about the appropriate criteria that should be used to distinguish between pairwise and diffuse, as well as the types of traits to which these criteria should be applied. Strauss et al. (pp. 81–89) attempt to reconcile these disagreements by refining the criteria for diffuse selection and outlining a quantitative-genetic experimental approach for deciding whether these criteria are met.


Evolutionary change is simultaneously a genetic phenomenon and an ecological phenomenon. Molecular dissection of adaptations provides information on the kinds of genetic changes that are used by natural selection to produce adaptation. At the same time, however, ecological investigations of interactions between a plant and its environment are required for understanding the cause and magnitude of natural selection that has led to the fixation of some genotypes over others. The research on plant mating systems and plant–enemy interactions highlighted in this issue illustrates the continuing importance of the latter approach and how it can be combined with molecular and genomic approaches to yield maximum insight.