Sex determination and the evolution of dioecy from monoecy in Sagittaria latifolia (Alismataceae)

The role of mutations of small versus large effect in adaptive evolution is of considerable interest to evolutionary biologists. The major evolutionary pathways for the origin of dioecy in plants (the gynodioecy and monoecy-paradioecy pathways) are often distinguished by the number of mutations involved and the magnitude of their effects. Here, we investigate the genetic and environmental determinants of sex in Sagittaria latifolia, a species with both monoecious and dioecious populations, and evaluate evidence for the evolution of dioecy via gynodioecy or monoecy-paradioecy. We crossed plants of the two sexual systems to generate F1, F2 and backcross progeny, and grew clones from dioecious populations in low-and high-fertilizer conditions to examine sex inconstancy in females and males. Several lines of evidence implicate two-locus control of the sex phenotypes. In dioecious populations sex is determined by Mendelian segregation of alleles, with males heterozygous at both the male- and female-sterility loci. In monoecious populations, plants are homozygous for alleles dominant to male sterility in females and recessive to female sterility in males. Experimental manipulation of resources revealed sex inconstancy in males but not females. These results are consistent with predictions for the evolution of dioecy via gynodioecy, rather than the expected monoecy-paradioecy pathway, given the ancestral monoecious condition.     

Androdioecy and the evolution of dioecy

The likelihood that dioecy could evolve via androdioecy is examined. It is concluded that female-sterility mutations are unlikely to be able to invade populations of self-compatible hermaphrodite species, even if the resources that an hermaphrodite devotes to seed production can be diverted to yield increased survival and also to increase male fertility. These findings are in agreement with the great rarity of androdioecy. Claimed cases of androdioecy are reviewed. All of the species in question appear to be functionally dioecious, with females retaining substantial anther vestiges. It is argued that this morphological androdioecy is in no way indicative of a previous functionally androdioecious state. The details of the reproductive biology of many of these species seem rather to be consistent with their having evolved dioecy via gynodioecy.
The rarity of androdioecy, as a route to the evolution of dioecy, suggests that re-allocation of reproductive resources is unlikely to be the sole factor of importance, and supports an important role for inbreeding avoidance. The fact that females in some dioecious species retain anthers of substantial size, containing considerable quantities of pollen, gives further support to the view that male-sterility mutations can sometimes be favoured even when little or no resources are re-allocated to male functions. This is impossible without substantial selfing and inbreeding. It is therefore concluded that inbreeding avoidance is generally important in the evolution of dioecy, though reallocation of reproductive resources is also necessary.     

Pollen limitation and the evolution of androdioecy from dioecy

Androdioecy is an unusual breeding system in which populations consist of separate male and hermaphrodite individuals. The evolution of androdioecy is still poorly understood; however, there is evidence from several androdioecious species that the breeding system may have evolved from dioecy (males and females). This article presents a simple deterministic model showing that androdioecy can evolve from dioecy under a broad range of realistic conditions. For the evolution of androdioecy from dioecy, hermaphrodites must be able to invade the dioecious population. Then, males must be maintained, while females are eliminated. Hermaphrodite invasion is favored when females are pollen limited and hermaphrodites have high overall fertility and are self-fertile. Male maintenance is favored when hermaphrodites resemble females, having high seed production and low pollen fitness, and when the selfing rate is not too high. These conditions were satisfied over a broad and realistic range of parameter values, suggesting that the evolution of androdioecy from dioecy is highly plausible.     

Repeated evolution of dioecy from androdioecy in Acer

The evolution of breeding systems was studied in the genus Acer, with special attention to the origin of androdioecy and dioecy, using a phylogenetic approach. Parsimony and maximum-likelihood techniques were used to infer the ancestral character state and trends in the evolution of breeding systems. Information on breeding systems was obtained from the literature, and phylogenetic relationships were taken from three published phylogenies. Although a general trend from duodichogamy to dioecy through heterodichogamy has been proposed for the genus Acer, our results show that a general trend is not detected when phylogenetic relationships are taken into account. Dioecy appeared as a derived state that evolved at least three times and never reversed towards other states. Three different paths to dioecy have been followed in the genus Acer: from heterodichogamous androdioecy; from heterodichogamous trioecy; and from dichogamous subdioecy. Therefore, although the best documented cases of evolution of androdioecy indicate that this breeding system evolves from dioecy, in the genus Acer the opposite situation occurs (androdioecy leading to dioecy). Here we discuss the role of inbreeding avoidance and sexual specialization as selective forces driving the evolution of dioecy in the genus Acer.     

Can classifications of functional gender be extended to all land plants?

Land plants (bryophytes, seedless vascular plants and seed plants) may have combined or separate sexes, and this variation also may occur at two life-cycle stages. Thus plants show variation in individuals’ attainment of fitness via sperms versus eggs (functional gender) and the diversity of gender morphs found in populations. We extend D.G. Lloyd's classification of flowering plant gender to all land plants, with three main functional classes according to whether populations are dimorphic or monomorphic for gender (i.e., populations consist of either one or two distinct sex classes), and at which life cycle stages this occurs: (1) sporophyte-dimorphic, (2) sporophyte-cosexual and gametophyte-dimorphic, and (3) gametophyte-cosexual. In dimorphic sporophytes and gametophytes, morphs that reproduce mostly as females and males may be constant (dioecy) or inconstant (gynodioecy, androdioecy, trioecy). We suggest that examining the sex conditions of seedless plants using a functional perspective will reveal a diversity of sexual systems largely analogous to those found in seed plants. An extended suite of model plants with different biological attributes will allow new tests of existing models of mechanisms that select for different sexual systems, and may lead to important new questions in the field, some of which we suggest here.     

Disassortative mating, sexual specialization, and the evolution of gender dimorphism in heterodichogamous Acer opalus

In sexually polymorphic species, the morphs are maintained by frequency-dependent selection through disassortative mating. In heterodichogamous populations in which disassortative mating occurs between the protandrous and protogynous morphs, a decrease in female fitness in one morph is hypothesized to drive sexual specialization in the other morph, resulting in dimorphic populations. We test these ideas in a population of the heterodichogamous species, Acer opalus . We assessed both prospective gender of individuals in terms of their allocations and actual parentage using microsatellites; we found that most matings in A. opalus occur disassortatively. We demonstrate that the protogynous morph is maintained by frequency-dependent selection, but that maintenance of males versus protandrous individuals depends on their relative siring success, which changes yearly. Seeds produced later in the reproductive season were smaller than those produced earlier; this should compromise reproduction through ovules in protandrous individuals, rendering them male biased in gender. Time-dependent gender and paternity analyses indicate that the sexual morphs are specialized in their earlier sexual functions, mediated by the seasonal decrease in seed size. Our results confirm that mating patterns are context-dependent and change seasonally, suggesting that sexual specialization can be driven by seasonal effects on fitness gained through one of the two sexual functions.     

Gender variation and the evolution of dioecy in Wurmbea dioica (Liliaceae)

A major obstacle for empirical tests of hypotheses concerning the evolution of dioecy in flowering plants is the limited number of species that possess both cosexual and dioecious populations. Wurmbea dioica (Liliaceae) is a diminutive, fly-pollinated geophyte native to temperate Australia. Marked geographical variation of floral traits is evident, particularly with respect to sex expression. A survey of phenotypic gender in 45 populations from Western Australia (WA), South Australia (SA), Victoria (Vic) and the Australian Capital Territory (ACT) revealed two contrasting patterns. Populations in SA, Vic, and ACT were uniformly dimorphic for gender, containing female and male plants, whereas populations in WA were either monomorphic or dimorphic. In most dimorphic populations varying numbers of male plants produced hermaphrodite flowers (male inconstancy). There was a significant negative relationship between female frequency and the proportion of inconstant male plants. Depending on region and population, male plants produced more flowers of larger size than females. In WA monomorphic populations often occurred on rich, moist soils at high density, whereas dimorphic populations were more commonly found at lower density on shallow soils in drier areas. In an area of sympatry, plants with contrasting sexual systems flowered at different times and were ecologically differentiated. The patterns of gender variation in W. dioica indicate that dioecy has evolved via the gynodioecious pathway. The spread of females in monomorphic populations may be favoured where ecological conditions result in increased selfing and inbreeding depression in hermaphrodites.     

Evolution of outcrossing in experimental populations of Caenorhabditis elegans

Caenorhabditis elegans can reproduce exclusively by self-fertilization. Yet, males can be maintained in laboratory populations, a phenomenon that continues to puzzle biologists. In this study we evaluated the role of males in facilitating adaptation to novel environments. For this, we contrasted the evolution of a fitness component exclusive to outcrossing in experimental populations of different mating systems. We introgressed a modifier of outcrossing into a hybrid population derived from several wild isolates to transform the wild-type androdioecious mating system into a dioecious mating system. By genotyping 375 single-nucleotide polymorphisms we show that the two populations had similar standing genetic diversity available for adaptation, despite the occurrence of selection during their derivation. We then performed replicated experimental evolution under the two mating systems from starting conditions of either high or low levels of diversity, under defined environmental conditions of discrete non-overlapping generations, constant density at high population sizes (N = 10⁴), no obvious spatial structure and abundant food resources. During 100 generations measurements of sex ratios and male competitive performance showed: 1) adaptation to the novel environment; 2) directional selection on male frequency under androdioecy; 3) optimal outcrossing rates of 0.5 under androdioecy; 4) the existence of initial inbreeding depression; and finally 5) that the strength of directional selection on male competitive performance does not depend on male frequencies. Taken together, these results suggest that androdioecious males are maintained at intermediate frequencies because outcrossing is adaptive.     

How does adaptation sweep through the genome? Insights from long-term selection experiments

A major goal in evolutionary biology is to understand the origins and fates of adaptive mutations. Natural selection may act to increase the frequency of de novo beneficial mutations, or those already present in the population as standing genetic variation. These beneficial mutations may ultimately reach fixation in a population, or they may stop increasing in frequency once a particular phenotypic state has been achieved. It is not yet well understood how different features of population biology, and/or different environmental circumstances affect these adaptive processes. Experimental evolution is a promising technique for studying the dynamics of beneficial alleles, as populations evolving in the laboratory experience natural selection in a replicated, controlled manner. Whole-genome sequencing, regularly obtained over the course of sustained laboratory selection, could potentially reveal insights into the mutational dynamics that most likely occur in natural populations under similar circumstances. To date, only a few evolution experiments for which whole-genome data are available exist. This review describes results from these resequenced laboratory-selected populations, in systems with and without sexual recombination. In asexual systems, adaptation from new mutations can be studied, and results to date suggest that the complete, unimpeded fixation of these mutations is not always observed. In sexual systems, adaptation from standing genetic variation can be studied, and in the admittedly few examples we have, the complete fixation of standing variants is not always observed. To date, the relative frequency of adaptation from new mutations versus standing variation has not been tested using a single experimental system, but recent studies using Caenorhabditis elegans and Saccharomyces cerevisiae suggest that this a realistic future goal.     

Environmental stress and the evolution of dioecy: Wurmbea dioica (Colchicaceae) in Western Australia

Stressful ecological conditions have been implicated in the evolution of separate sexes in plants. Gender dimorphic species are often found in drier habitats than their sexually monomorphic relatives, and gynodioecious populations appear closer to a dioecious state as resources, particularly water, become limiting. This pattern could result if dry conditions decrease the relative seed fitness of cosexual plants, allowing female plants to become established in monomorphic populations. We studied geographical variation in gender expression and biomass allocation among 12 monomorphic and dimorphic populations of Wurmbea dioica along a latitudinal precipitation gradient in southwestern Australia to provide insight into mechanisms by which aridity might favor transitions between sexual systems. Plants in monomorphic and dimorphic populations exhibited contrasting gender expression and patterns of biomass allocation in areas with different levels of precipitation. Among dimorphic populations, lower precipitation was associated with a higher frequency of female plants, and reduced allocation to female function by hermaphrodites during flowering. In contrast, stress conditions had no effect on female allocation at flowering in monomorphic populations. Across latitudes, unisexuals and cosexuals exhibited consistent differences in above ground traits, with cosexuals having larger leaves, taller stems and larger flowers. Although all plants were smaller under drier conditions, cosexuals decreased above ground allocation to vegetative and reproductive structures with decreasing latitude. In contrast, unisexuals increased allocation to reproduction in drier areas at the expense of below ground size. Aridity was associated with reduced flower size among all gender classes, but not with changes in flower number. These data do not support the hypothesis that resource limitation of female allocation in cosexual populations favors the establishment of gender dimorphism in W. dioica. Alternative hypotheses, involving higher selfing rates and enhanced survival of unisexuals relative to cosexuals under resource-limited conditions, are discussed as possible explanations for the origin of dioecy in W. dioica. This revised version was published online in August 2006 with corrections to the Cover Date.     

The balance between mutators and nonmutators in asexual populations

Mutator alleles, which elevate an individual's mutation rate from 10 to 10,000-fold, have been found at high frequencies in many natural and experimental populations. Mutators are continually produced from nonmutators, often due to mutations in mismatch-repair genes. These mutators gradually accumulate deleterious mutations, limiting their spread. However, they can occasionally hitchhike to high frequencies with beneficial mutations. We study the interplay between these effects. We first analyze the dynamics of the balance between the production of mutator alleles and their elimination due to deleterious mutations. We find that when deleterious mutation rates are high in mutators, there will often be many "young," recently produced mutators in the population, and the fact that deleterious mutations only gradually eliminate individuals from a population is important. We then consider how this mutator-nonmutator balance can be disrupted by beneficial mutations and analyze the circumstances in which fixation of mutator alleles is likely. We find that dynamics is crucial: even in situations where selection on average acts against mutators, so they cannot stably invade, the mutators can still occasionally generate beneficial mutations and hence be important to the evolution of the population.     

The effect of population bottlenecks on mutation rate evolution in asexual populations

In the absence of recombination, a mutator allele can spread through a population by hitchhiking with beneficial mutations that appear in its genetic background. Theoretical studies over the past decade have shown that the survival and fixation probability of beneficial mutations can be severely reduced by population size bottlenecks. Here, we use computational modelling and evolution experiments with the yeast S. cerevisiae to examine whether population bottlenecks can affect mutator dynamics in adapting asexual populations. In simulation, we show that population bottlenecks can inhibit mutator hitchhiking with beneficial mutations and are most effective at lower beneficial mutation supply rates. We then subjected experimental populations of yeast propagated at the same effective population size to three different bottleneck regimes and observed that the speed of mutator hitchhiking was significantly slower at smaller bottlenecks, consistent with our theoretical expectations. Our results, thus, suggest that bottlenecks can be an important factor in mutation rate evolution and can in certain circumstances act to stabilize or, at least, delay the progressive elevation of mutation rates in asexual populations. Additionally, our findings provide the first experimental support for the theoretically postulated effect of population bottlenecks on beneficial mutations and demonstrate the usefulness of studying mutator frequency dynamics for understanding the underlying dynamics of fitness‐affecting mutations.     

Transmission rates and adaptive evolution of pathogens in sympatric heterogeneous plant populations

Diversification in agricultural cropping patterns is widely practised to delay the build-up of virulent races that can overcome host resistance in pathogen populations. This can lead to balanced polymorphism, but the long-term consequences of this strategy for the evolution of crop pathogen populations are still unclear. The widespread occurrence of sibling species and reproductively isolated sub-species among fungal and oomycete plant pathogens suggests that evolutionary divergence is common. This paper develops a mathematical model of host-pathogen interactions using a simple framework of two hosts to analyse the influences of sympatric host heterogeneity on the long-term evolutionary behaviour of plant pathogens. Using adaptive dynamics, which assumes that sequential mutations induce small changes in pathogen fitness, we show that evolutionary outcomes strongly depend on the shape of the trade-off curve between pathogen transmission on sympatric hosts. In particular, we determine the conditions under which the evolutionary branching of a monomorphic into a dimorphic population occurs, as well as the conditions that lead to the evolution of specialist (single host range) or generalist (multiple host range) pathogen populations.     

Evolutionary consequences of gender plasticity in genetically dimorphic breeding systems

A functional view of gender helps evolutionary biologists evaluate the mechanisms underlying breeding-system evolution. Evolutionary pathways from hermaphroditism to dioecy include the intermediate breeding systems of gynodioecy and androdioecy. These pathways start with the invasion of unisexual mutants, females or males, respectively, followed by alteration of the hermaphrodites to allocate more to the sexual function that the unisexuals lack. Eventually, hermaphrodites become unisexual and dioecy has evolved. Some species evolving along these pathways stop short of completing this second step, or even revert back from dioecy. We evaluate the hypothesis that gender plasticity is involved in these transitions to and from dioecy. Evidence from studies of subdioecious species that have evolved along the gynodioecy pathway suggests that gender plasticity occurs and stabilizes subdioecy by lowering the cost of producing seed. Factors influencing species evolving toward androdioecy, or reverting to androdioecy from dioecy, appear to be more varied and include reproductive assurance, herbivory and gender plasticity. In general, gender specialization appears to be favored in resource-poor environments regardless of which pathway is taken to dioecy.     

Reproductive biology of Phillyrea angustifolia L.(Oleaceae) and effect of galling-insects on its reproductive output

The breeding system of Phillyrea angustifolia is examined in several Spanish populations, both in the southeastern Iberian Peninsula and in the Balearic IslandS. Hand-pollination experiments performed on one of the mainland and on one of the island populations demonstrated that the plant is functionally androdioeciouS. Pollen from males was several times more fertile than pollen from hermaphrodite flowers or self-pollen in the island population, but not so in the Iberian Peninsula. The interpretation of such a result is that androdioecy must have an adaptive advantage in an island system, where inbreeding depression is expected to be greater. Male individuals are much less common than hermaphrodites in all populations studied, thus supporting the prediction made by the models for the maintenance of androdioecy. Sex expression in an individual plant differs between populations. The frequency of sexes was purely bimodal in one population, whereas it was not so in another. Pollen from both male and hermaphrodite flowers appears to be morphologically different, although it remains unknown at what stage of the fertilization process the difference becomes functional. Fruitset in the hermaphrodite individuals was always less than 10%.Most fruits abort at an early stage and there is also a great flower abscission. Fruitset does not appear to be influenced by any of the plant traits describing size or fecundity. A high unpredictability of flowering in an individual was also observed in all populations.P.angustifolia is attacked by a cecidomyiid fly that induces the formation of galls in the ovaries (producing ‘deforme’ fruits which have been confused until recently with parthenocarpic fruits), thus reducing the plant reproductive output. The production of such galls (most of them in the hermaphrodite individuals but a small proportion also in males) is highly variable among plants and among populations, representing from 0 to 97% of the initiated fruitS. Such variability in gall incidence does not respond to variability in plant size, fecundity, distance to flowering conspecifics or time that flowers are available for fly oviposition. A high consistency in gall production was found in both the mainland and the island populations. The greater abundance of galls in the islands compared with the mainland populations is interpreted as a result of the higher temperatures in the former which might influence the activity period of the flies.     

Evolutionary branching under asymmetric competition

I investigate the evolution of a continuous trait, such as body size or arms level, which affects the outcome of competitive contests such that the contestant with the larger trait value has a higher probability of winning. I show that a polymorphism of distinctly different strategies can evolve in an initially monomorphic population even if mutations have only small phenotypic effect. In a simple Lotka-Volterra-type model of asymmetric competition, I derive the conditions under which two strategies can gradually evolve from a single ancestral strategy; the evolution of higher level polymorphisms is studied by numerical analysis and computer simulations of specific examples. High levels of polymorphism may build up during evolution. The coevolution of strategies in polymorphic populations, however, may also lead to extinction, which decreases the level of polymorphism. I discuss whether the evolution of several haploid strategies from a single initial strategy may correspond to the evolution of several sympatric species in a diploid outbreeding population.     

Continuous In Vitro Evolution of Ribozymes That Operate Under Conditions of Extreme pH

Continuous in vitro evolution methods were used to study the behavior of an evolving population of RNA ligase ribozymes in response to selection pressures involving conditions of extreme pH. The starting population consisted of randomized variants of a ribozyme that had been optimized for activity at pH 8.5. The ribozymes were subjected to repeated rounds of selective amplification under progressively more acidic or more alkaline conditions. The two final evolved populations of ribozymes were able to operate at either pH 5.8 or pH 9.8, respectively. Representative individuals from the two final populations were isolated and characterized. The low-pH ribozyme exhibited a 10-fold increase in catalytic rate at pH 5.8 compared to the starting molecule. The high-pH ribozyme retained its structural integrity and activity at pH 9.8, whereas the starting molecule was denatured under this condition. These findings demonstrate that a population of functional macromolecules can adapt to stringent environmental conditions through the acquisition of relatively few mutations. The results establish continuous in vitro evolution as a useful model system for exploring the evolution of enzymatic function in extreme environments. Present address (Henriette Kühne): Cardinal Health, 2950 Trade Place, San Diego, CA 92126, USA     

LOSS OF SEX IN CLONAL POPULATIONS OF A FLOWERING PLANT,DECODON VERTICILLATUS (LYTHRACEAE)

The loss of traits that no longer increase fitness is a pervasive feature of evolution, although detailed studies of the genetic, developmental, and evolutionary factors involved are few. Most perennial plants practice both sexual and clonal reproduction, and it has been hypothesized that populations with little sexual recruitment may lose the capacity for sexual reproduction by fixing mutations that disable one or more of the many processes involved in sex. The clonal, tristylous aquatic plant, Decodon verticillatus, exhibits marked geographical variation in sexual recruitment. Populations at the northern limit of the range are usually monomorphic for style length consist of single genotypes, and produce almost no seed, due, in part, to environmental conditions that inhibit pollination, fertilization, and seed maturation. Controlled crosses in a greenhouse provided evidence for greatly reduced sexual capacity in an exclusively clonal, monomorphic population. Plants from this infertile population produced only 3–18% as many seeds per pollination as fertile populations. Observations of pollen tube growth indicated that infertility is due to severe reductions in pollen tube numbers both early after pollination and later when pollen tubes were traversing the ovary, due primarily to the inability of pistils to support normal tube growth. A three-year greenhouse experiment comparing fertility, survival, and growth of F1 progenies produced from reciprocal crosses between plants from the infertile population and those from nearby fertile populations suggested that the genetic basis for infertility is simple and may involve a single recessive mutation. In addition, the results did not reveal any association between infertility and other aspects of survival and vegetative vigor. The infertile genotype was likely fixed in the population through founder effect rather than indirect selection resulting from antagonistic pleiotropy or direct selection of advantages associated with reduced investment in sexual reproduction. A broader comparison of sexual fertility in 15 clonal, monomorphic populations and five genotypically diverse, trimorphic populations under greenhouse conditions revealed substantial infertility in all but one monomorphic population. Populations varied somewhat in the stage at which infertility was expressed, however, pollen tube growth was impaired in all populations. These results provide strong support for the hypothesis that complex traits like sex are degraded by mutation when they no longer increase fitness.     

Widespread functional android in Mercurialis annua L. (Euphorbiaceae)

The widespread coexistence of male and monoecious (cosexual) plants in Spanish, Portuguese and Moroccan populations of Mercurialis annua , an annual wind-pollinated ruderal, represents an important case of functional androdioecy, a rare breeding system in plants and animals. In M. annua , both males and cosexes disperse fully competent pollen. Quantitative gender varies discontinuously between males and cosexes, with males producing a mean of 6.09 times as much pollen as cosexes. It appears that gender is determined by a simple developmental switch, with male and cosexual inflorescences differing markedly in morphology: staminate flowers are borne on erect peduncles in males and in tight spiral clusters around a subsessile pistillate flower in cosexes. Males do not differ from cosexes in their biomass, but they are significantly taller, principally as a result of their greater internode lengths. The cosexual inflorescence is strongly protogynous so that outcrossing is favoured in dense stands, but seed-set is assured in cosexes isolated from prospective mates because of their ability to self-fertilize. Males typically occur at frequencies of less than about 30% in androdioecious populations, in accordance with theoretical predictions for functional androdioecy. In the genus Mercurialis , dioecy is the ancestral condition and monoecy and androdioecy, which occur in polyploid populations of M. annua , are derived. I argue here that androdioecy is most likely to evolve in plants (1) from dioecy, (2) in wind-pollinated species, and (3) in species with a colonizing habit. These predictions are also consistent with the limited published data available for other species.     

THE DISSOLUTION OF A COMPLEX GENETIC POLYMORPHISM: THE EVOLUTION OF SELF-FERTILIZATION IN TRISTYLOUS EICHHORNIA PANICULATA (PONTEDERIACEAE)

Eichhornia paniculata (Pontederiaceae) displays a wide range of outcrossing levels as a result of the dissolution of the tristylous genetic polymorphism and the evolution of semihomostyly. Population surveys, comparison of fitness components of the style morphs, and computer simulations were used to investigate the breakdown of tristyly and the selective mechanisms responsible for the evolution of self-fertilization. Of 110 populations surveyed in northeast Brazil and Jamaica, 53% were trimorphic, 25% were dimorphic, and 22% were monomorphic for style morph. The short (S) morph was underrepresented in trimorphic populations and absent from nontrimorphic populations. The mid (M) morph predominated in dimorphic populations and was the only morph in monomorphic populations. Stamen modifications promoting selfing, associated with semihomostyle evolution, were largely confined to the M morph. They were rare in trimorphic populations, common in dimorphic populations, and often fixed in monomorphic populations. Stochastic simulations and comparisons of fruit set in natural populations indicate that founder events, population bottlenecks, and lowered fertility of the S morph due to an absence of long-tongued pollinators can each account for loss of the S morph from trimorphic populations. A reduced level of disassortative mating can accentuate the rate at which the S morph is lost by both random and deterministic processes. Nontrimorphic populations occur at the geographical margins of the region surveyed and tend to be smaller and less dense than trimorphic populations. These observations and the higher fruit set of the M morph relative to the L morph in dimorphic populations suggest that reproductive assurance, favoring selfing variants of the M morph under conditions of low pollinator service, has been of primary importance in the origin of most monomorphic populations. Where pollinator service is reliable, however, automatic selection of selfing genes, aided by mating asymmetries between the morphs, can cause the M morph to spread to fixation in dimorphic populations.