The potential for selection on pollen colour dimorphisms in Nigella degenii: morph-specific differences in pollinator visitation, fertilisation success and siring ability

Two subspecies of Nigella degenii (Ranunculaceae) possess a dimorphism in pollen colour and vary extensively in frequency of the two morphs in natural populations. Here we investigate the role of selection on pollen colour during the pollination phase in the two subspecies and its potential contribution to the maintenance of this colour variation. In a combination of common garden experiments and field observations, we obtained data on pollinator visitation rates and explored the effect of pollen colour on fertilisation success and siring ability under conditions of low vs. high pollen competition. In experimental gardens, naïve pollinators responded differently to plants with different pollen colour, but the favoured morph varied between dates and locations, and colour morphs were not visited in a frequency-dependent manner. Donor plants with dark pollen had a reproductive advantage (higher seed set) in single-donor pollinations, but the realised siring ability (measured by progeny morph ratio) was highly variable between different two-donor crosses with no general bias towards the light or dark morph. Therefore, although the dark pollen type appears to have a general selective advantage in terms of fertilisation success, our data are also consistent with a scenario involving the maintenance of both colour morphs, particularly under conditions of high pollen competition, a variable genetic background and/or spatial or temporal variation in the pollinator fauna.     

Evolution and maintenance of pollen-colour dimorphisms in Nigella degenii: habitat-correlated variation and morph-by-environment interactions

Dimorphism in pollen colour is rare among flowering plants, but occurs in two geographically and morphologically distinct subspecies of Nigella degenii (Ranunculaceae). We evaluated the role of genotype-by-environment interactions in the maintenance of two pollen morphs within each of these subspecies. Morph frequencies in a number of populations were related to current habitat conditions, and an extensive common-garden experiment involving both optimal and stressful conditions (drought and nutrient deficiency) was carried out. The putatively derived (dark) pollen morph of N. degenii ssp. barbro has a higher frequency on slopes facing north or east than on slopes facing south or west. Plants of the dark morph also have a higher mortality under drought stress or nutrient deficiency. Data available for N. degenii ssp. jenny provide little evidence for habitat-correlated variation in morph frequency or morph-specific differences in fitness under optimal and stressful growth conditions. Our results suggest that morph-by-environment interactions in mortality could contribute to the maintenance of pollen-colour dimorphisms in N. degenii ssp. barbro.     

Pollen colour morphs take different paths to fitness

Colour phenotypes are often involved in communication and are thus under selection by species interactions. However, selection may also act on colour through correlated traits or alternative functions of biochemical pigments. Such forms of selection are instrumental in maintaining petal colour diversity in plants. Pollen colour also varies markedly, but the maintenance of this variation is little understood. In Campanula americana, pollen ranges from white to dark purple, with darker morphs garnering more pollinator visits and exhibiting elevated pollen performance under heat stress. Here, we generate an F2 population segregating for pollen colour and measure correlations with floral traits, pollen attributes and plant‐level traits related to fitness. We determine the pigment biochemistry of colour variants and evaluate maternal and paternal fitness of light and dark morphs by crossing within and between morphs. Pollen colour was largely uncorrelated with floral traits (petal colour, size, nectar traits) suggesting it can evolve independently. Darker pollen grains were larger and had higher anthocyanin content (cyanidin and peonidin) which may explain why they outperform light pollen under heat stress. Overall, pollen‐related fitness metrics were greater for dark pollen, and dark pollen sires generated seeds with higher germination potential. Conversely, light pollen plants produce 61% more flowers than dark, and 18% more seeds per fruit, suggesting a seed production advantage. Results indicate that light and dark morphs may achieve fitness through different means—dark morphs appear to have a pollen advantage whereas light morphs have an ovule advantage—helping to explain the maintenance of pollen colour variation.     

Nonindependent domestication of the two rice subspecies, Oryza sativa ssp. indica and ssp. japonica, demonstrated by multilocus microsatellites

The origins of the Asian cultivated rice Oryza sativa from its wild ancestor O. rufipogon have been debated for decades. The question mainly concerns whether it originated monophyletically or polyphyletically. To shed light on the origins and demographic history of rice domestication, we genotyped a total of 92 individual plants from the two O. sativa subspecies and O. rufipogon for 60 microsatellites. An approximate Bayesian method was applied to estimate demographic parameters for O. rufipogon vs. O. sativa ssp. indica and O. rufipogon vs. O. sativa ssp. japonica. We showed that the japonica subspecies suffered a more severe bottleneck than the indica subspecies and thus a greater loss of genetic variation during its domestication. Across microsatellite loci there is a significant positive correlation in the reduction of genetic diversity between the two subspecies. The results suggest that completely independent domestication of indica and japonica subspecies may not explain our data and that there is at least partial sharing of their ancestral populations and/or recent gene flow between them.     

Pollen competition between morphs in a pollen‐color dimorphic herb and the loss of phenotypic polymorphism within populations

Flower color polymorphism is relatively uncommon in natural flowering plants, suggesting that maintenance of different color morphs within populations is difficult. To address the selective mechanisms shaping pollen‐color dimorphism, pollinator preferences and reproductive performance were studied over three years in Epimedium pubescens in which some populations had plants with either green or yellow pollen (and anthers). Visitation rate and pollen removal and receipt by the bee pollinator (Andrena emeishanica) did not differ between the two color morphs. Compared to the green morph, siring success of the yellow morph's pollen was lower, but that of mixtures of pollen from green and yellow morphs was lowest. This difference, corresponding to in vivo and ex vivo experiments on pollen performance, indicated that pollen germination, rather than tube growth, of the green morph was higher than that of the yellow morph and was seriously constrained in both morphs if a pollen competitor was present. A rare green morph may invade a yellow‐morph population, but the coexistence of pollen color variants is complicated by the reduced siring success of mixed pollinations. Potential pollen competition between morphs may have discouraged the maintenance of multiple phenotypes within populations, a cryptic mechanism of competitive exclusion.     

Genetic variability in Plantago species in relation to their ecology

Summary Genetic variation in leaf and inflorescence morphology and in generative development within the species Plantago major has been analysed by means of crosses between members of two different subspecies. The variable characters chosen are supposed to be important for determining the ecological differences between the subspecies and other ecotypes. The analyses of F₂'s indicated that a substantial number of loci controlling the above mentioned characters are situated near the Pgm-1 locus, forming a gene complex. This gene complex can exist in at least three different forms in ssp. pleiosperma, ssp. major lawn type and ssp. major roadside type, respectively. In addition, some important factors for ecotypic differentiation are situated in the neighbourhood of the Got-1 locus and in a linkage group containing three other allozyme loci. These linkages between allozyme loci and fitness-affecting loci can explain the restriction of some enzyme alleles to a particular subspecies.Grassland Species Research Group Publication No. 50     

Inheritance of seed coat color genes in <Emphasis Type="Italic">Brassica napus</Emphasis> (L.) and tagging the genes using SRAP,SCAR and SNP molecular markers

Seed coat color inheritance in Brassica napus was studied in F₁, F₂, F₃ and backcross progenies from crosses of five black seeded varieties/lines to three pure breeding yellow seeded lines. Maternal inheritance was observed for seed coat color in B. napus, but a pollen effect was also found when yellow seeded lines were used as the female parent. Seed coat color segregated from black to dark brown, light brown, dark yellow, light yellow, and yellow. Seed coat color was found to be controlled by three genes, the first two genes were responsible for black/brown seed coat color and the third gene was responsible for dark/light yellow seed coat color in B. napus. All three seed coat color alleles were dominant over yellow color alleles at all three loci. Sequence related amplified polymorphism (SRAP) was used for the development of molecular markers co-segregating with the seed coat color genes. A SRAP marker (SA12BG18388) tightly linked to one of the black/brown seed coat color genes was identified in the F₂ and backcross populations. This marker was found to be anchored on linkage group A9/N9 of the A-genome of B. napus. This SRAP marker was converted into sequence-characterized amplification region (SCAR) markers using chromosome-walking technology. A second SRAP marker (SA7BG29245), very close to another black/brown seed coat color gene, was identified from a high density genetic map developed in our laboratory using primer walking from an anchoring marker. The marker was located on linkage group C3/N13 of the C-genome of B. napus. This marker also co-segregated with the black/brown seed coat color gene in B. rapa. Based on the sequence information of the flanking sequences, 24 single nucleotide polymorphisms (SNPs) were identified between the yellow seeded and black/brown seeded lines. SNP detection and genotyping clearly differentiated the black/brown seeded plants from dark/light/yellow-seeded plants and also differentiated between homozygous (Y2Y2) and heterozygous (Y2y2) black/brown seeded plants. A total of 768 SRAP primer pair combinations were screened in dark/light yellow seed coat color plants and a close marker (DC1GA27197) linked to the dark/light yellow seed coat color gene was developed. These three markers linked to the three different yellow seed coat color genes in B. napus can be used to screen for yellow seeded lines in canola/rapeseed breeding programs.     

Adaptive associations between total body color dimorphism and climatic stress‐related traits in a stenothermal circumtropical Drosophila species

Abstract Low desiccation resistance of Drosophila ananassae reflects its rarity outside the humid tropics. However, the ability of this sensitive species to evolve under seasonally varying subtropical areas is largely unknown. D. ananassae flies are mostly lighter during the rainy season but darker and lighter flies occur in the autumn season in northern India. We tested the hypothesis whether seasonally varying alternative body color phenotypes of D. ananassae vary in their levels of environmental stress tolerances and mating behavior. Thus, we investigated D. ananassae flies collected during rainy and autumn seasons for changes in body melanization and their genetic basis, desiccation‐related traits, cold tolerance and mating propensity. On the basis of genetic crosses, we found total body color dimorphism consistent with a single gene model in both sexes of D. ananassae. A significant increase in the frequency of the dark morph was observed during the drier autumn season, and body color phenotypes showed significant deviations from Hardy‐Weinberg equilibrium, which suggests climatic selection plays a role. Resistance to desiccation as well as cold stress were two‐ to three‐fold higher in the dark body color strain as compared with the light strain. On the basis of no‐choice mating experiments, we observed significantly higher assortative matings between dark morphs under desiccation or cold stress, and between light morphs under hot or higher humidity conditions. To the best of our knowledge, this is the first report on the ecological significance of seasonally varying total body color dimorphism in a tropical species, D. ananassae.     

FLOWER COLOR POLYMORPHISM AND POLLINATION BIOLOGY OF PLATYSTEMON CALIFORNICUS BENTH. (PAPAVERACEAE)

Five genetically controlled flower color morphs in Platystemon californicus tend to occur in distinct geographic regions, suggesting regional selection of morphs. I examined the reproductive biology of P. californicus to determine whether color morphs are subject to selection due to differences in pollinator attractiveness. Plants of P. californicus have a high pollen: ovule ratio and are highly self-incompatible. Solitary bees (Andrenidae and Halictidae) are the most effective pollen vectors, but show inconsistent discrimination among color morphs. Platystemon californicus is unusual among annual self-incompatible plants in that wind is an important pollen vector. As a result of wind pollination, most polymorphic populations show no intermorph seed-set differences. Therefore, the geographic distribution of flower color morphs cannot be explained by differential attractiveness of color morphs to pollinators in different regions. Any selective value of color morphs must be due to linkage of flower color with as yet undetected morphologically or physiologically adaptive characters.     

Sequential variation in the components of reproductive success in the distylousJasminum fruticans (Oleaceae)

A 2-year study of three natural populations of the distylousJasminum fruticans showed that mean fruit and seed production were significantly greater in shortstyled plants (thrums) than in long-styled plants (pins). In this study, we investigated the role of four sequential factors which may differentially influence fruit and seed set in the two floral morphs: (1) differences in flowering phenology, (2) a limitation of pollen transfer towards pins, (3) a differential capacity of the two morphs to act through famale and male function and (4) differential fruit abortion in the two morphs. Fruit set was significantly influenced by differences in flowering phenology although there were no differences in flowering time between the two morphs. supplementary pollinations in a natural population significantly increased fruit set and reduced the difference in fruit set between the two morphs in relation to controls, indicating a limitation on pollen transfer which was most severe towards pin stigmas. In reciprocal crosses, seed set was significantly dependent on the paternal and maternal identity of the pin parent. There was no significant variation among thrums for their performance as male or female parent. Furthermore, individual pin plants with relatively high percentage seed set as female parents gave poor seed set as male parents and vice-versa. Whereas fruit removal had no effect on seed number in thrum plants, a greater proportion of viable seeds were produced on pin plants which were left to naturally mature their fruits than on pins which had fruits artificially removed, suggesting the occurrence of selective fruit abortion in pins but not in thrums. The initially greater maternal fitness of thrums due to their greater success as pollen recipients may thus be opposed by increased viable seed set in the pins due to factors acting after the pollination stage. The relative reproductive success of floral morphs in the distylousJ. fruticans is thus differentially influenced by ecological factors occurring at different stages of the reproductive process.     

Discovery of gynoecium color polymorphism in an aquatic plant

Flower color polymorphlsm exhibited by natural populations provides an opportunity for understanding the evolutionary mechanisms contributing to the diversity of floral morphology.However,little is known about the color polymorphism of female organs in flowering plants.Here we report gynoecium color polymorphism in Butomus umbellatus (Butomaceae),an emergent,aquatic monocot.Populations from Mishan,northeastern China comprised two morphs; gynoecia are either pink,as observed in other areas,or white.We measured floral traits and female fecundity in the two gynoecium color morphs in the field.There was no significant difference in plant height,pedicel length,and flower size including petal,sepal and gynoecium between the two morphs,but plants with pink gynoecia had wider inflorescence stalks,larger inner whorl anthers and produced more pollen and ovules than those with white gynoecia.Correspondingly,we found that seed production was significantly higher in the pink than in the white morph.This new finding suggested selection against white gynoecia in part because of low fecundity,consistent with the rarity of the white gynoecium morph in this species.     

Discovery of androecium color polymorphism in Epimedium pubescens with habitat preference of anther/pollen color in the genus

Aims Interspecific and intraspecific variation in flower color in natural populations provides an opportunity for us to understand the evolution and maintenance of diversity of floral traits. Compared to corolla color, little is known about the color polymorphism of sexual organs in flowering plants. To explore evolutionary transitions of androecium color and polymorphism within species, interspecific and intraspecific variation in androecium (anther and pollen) color in the genus Epimedium (Berberidaceae) was investigated.Methods To explore the geographical patterns of anther/pollen color variation in Epimedium species, data of 45 species were collected and their phylogeny was constructed based on available DNA sequences. To investigate whether intraspecific variation in androecium color relates to habitat preference, three environmental factors were measured in the field population of Epimedium pubescens in northeastern Sichuan, China, which plants had green or yellow androecia. Vegetative and reproductive traits of this species were compared between the two color morphs.Important findings Androecium (anther and pollen) color polymorphism in field populations of Epimedium pubescens is reported here where nine populations are monomorphic with a green androecium but three populations are dimorphic with individuals having either a green or a yellow androecium. Inflorescence stalk height, stalk diameter, leaf number, flower number and spur length (as well as spur and nectar volume) were not significantly different between two morphs. Compared to the yellow morph, the green morph had relatively larger leaves and anthers, but smaller sepals. The green morph produced more pollen and larger seeds, but the same number of ovules. Seed set was not significantly different between green and yellow morph. Investigations of environmental factors in the color dimorphic populations of E. pubescens indicated that the green morph was more likely to occur in habitats with relatively lower light intensity. The distribution survey of 45 Epimedium species showed that species with a green androecium tended to appear at lower elevations. Comparative phylogenetic analysis showed that transitions from yellow to a green androecium or to androecial color dimorphism occurred at least seven times. This genus, characterized by anther color diversity and containing some species with anther color polymorphism, provides a model system in which to study the evolution and maintenance of colorful sexual organs in flowering plants.     

POLLINATION AND SEED PRODUCTION IN IPOMOPSIS AGGREGATA: DIFFERENCES AMONG AND WITHIN FLOWER COLOR MORPHS

The pollination of red, pink, and white color morphs of Ipomopsis aggregata was evaluated to assess whether ethological isolation based on pollinator color discrimination may occur. We observed animal visitors, assessed pollen delivery, seed set per fruit, percentage of flowers setting fruit, nectar production, and timing of flower opening for different color morphs in the Front Range of Colorado. Based on traditional zoophilous flower classifications, we expected hummingbirds to pollinate red-flowered I. aggregata subsp. collina and hawkmoths to pollinate white-flowered I. aggregata subsp. Candida. However, ethological isolation does not appear to occur among color morphs of I. aggregata in the Front Range. Hummingbirds visited red-flowered plants in excess overall, and, to a lesser extent, so did hawkmoths. Both hummingbirds and hawkmoths visited all color morphs and probably transferred pollen among them. Pollen delivery data and a day-night bagging experiment also suggest that pollinators do not necessarily behave as predicted by flower classifications. In addition, there is little evidence for major differences between red, white, and pink flowers in any aspects of reproductive biology. Indeed, most variation occurs within a given color morph.     

Reproductive allocation and resource compensation in male-sterile and hermaphroditic plants of Plantago lanceolata (Plantaginaceae)

Gynodioecy is a breeding system in which hermaphrodites coexist with male steriles. Theoretical models predict that without any compensation in female fitness male steriles will disappear from a population due to their reproductive disadvantage. In the present study I investigated whether male-sterile (MS), partially male-sterile (IN), and hermaphroditic (H) plants of Plantago lanceolata differed in reproductive growth and allocation. Offspring of three interpopulation crosses segregating all three sex morphs were grown under nitrogen-limited conditions in a growth chamber. Independent of the genetic background MS plants attained a higher vegetative and reproductive dry mass and a higher reproductive output than H plants, whereas IN plants had intermediate values. When corrected for the mass of the pollen, the dry mass differences between the sex morphs were much reduced but still present. However, when whole-plant allocation was expressed on the basis of nitrogen, the differences between the sex morphs disappeared. Thus the sex morphs took up similar amounts of nitrogen but distributed them differently. The MS and IN plants used the nitrogen saved by not producing pollen for additional vegetative as well as reproductive growth. The data presented in this study suggest that resource compensation is one of the main mechanisms responsible for the maintenance of MS and IN plants in gynodioecious P. lanceolata.     

Reproductive ecology of distylous Palicourea Padifolia (Rubiaceae) in a tropical montane cloud forest. I. Hummingbirds' effectiveness as pollen vectors

The adaptiveness of distyly has been typically investigated in terms of its female function, specifically pollen receipt. However, pollen loads on stigmas can only provide moderate support for Darwin's hypothesis of the promotion of legitimate crosses. To determine the effectiveness of hummingbirds as pollen vectors between floral morphs and the consequences in terms of male (pollen transfer) and female function (pollen receipt) in Palicourea padifolia (Rubiaceae), floral visitors, their foraging modes, and temporal patterns of floral visitation were observed and documented. Differences in pollen and stigma morphology, pollen flow, rates of pollen deposition, and/or stigmatic pollen loads were then evaluated for their contribution toward differences in reproductive output between floral morphs. A pollination experiment with stuffed hummingbirds that varied in bill size was done to evaluate the contribution of bill variation toward differences between floral morphs in pollen receipt and pollen transfer and female reproductive output. Anthers of long-styled flowers contained significantly more and smaller pollen grains than those of short-styled flowers, independently of corolla and anther lengths. The shape and orientation of the stigma lobes differed between morphs and were significantly longer among short-styled flowers. Hummingbird visitation rates did not differ significantly between floral morphs, and foraging movements from focal plants towards neighboring plants were independent of floral morph. Stigmatic pollen loads under field conditions and those after controlled hummingbird visitation, along with rates of pollen accumulation through the day indicated that stigmas of short-styled flowers receive proportionately more legitimate (intermorph) pollen grains than did those of long-styled flowers. However, the species of hummingbird was marginally significant in explaining variation in pollen deposition on stigmas. Lastly, intermorph pollinations of P. padifolia resulted in significant differences in fruit production between floral morphs, independent of pollination treatment and pollinator species; short-styled flowers proportionately developed almost twice the number of fruits developed by long-styled flowers.     

Predation and the maintenance of color polymorphism in a habitat specialist squamate

Multiple studies have addressed the mechanisms maintaining polymorphism within a population. However, several examples exist where species inhabiting diverse habitats exhibit local population-specific polymorphism. Numerous explanations have been proposed for the maintenance of geographic variation in color patterns. For example, spatial variation in patterns of selection or limited gene flow can cause entire populations to become fixed for a single morph, resulting in separate populations of the same species exhibiting separate and distinct color morphs. The mottled rock rattlesnake (Crotalus lepidus lepidus) is a montane species that exhibits among-population color polymorphism that correlates with substrate color. Habitat substrate in the eastern part of its range is composed primarily of light colored limestone and snakes have light dorsal coloration, whereas in the western region the substrate is primarily dark and snakes exhibit dark dorsal coloration. We hypothesized that predation on high contrast color and blotched patterns maintain these distinct color morphs. To test this we performed a predation experiment in the wild by deploying model snakes at 12 sites evenly distributed within each of the two regions where the different morphs are found. We employed a 2×2 factorial design that included two color and two blotched treatments. Our results showed that models contrasting with substrate coloration suffered significantly more avian attacks relative to models mimicking substrates. Predation attempts on blotched models were similar in each substrate type. These results support the hypothesis that color pattern is maintained by selective predation.     

POLLEN APERTURE POLYMORPHISM AND GAMETOPHYTE PERFORMANCE IN VIOLA DIVERSIFOLIA

Pollen aperture polymorphism is studied in Viola diversifolia, where all plants produce three- and four-apertured pollen grains. We tested whether there are genetic differences among plants for the proportions of the different pollen morphs, and whether the morphs differ in gametophytic performance. Results show that the more apertures a pollen grain has, the more quickly it germinates but that few-apertured pollen grains have faster growing pollen tubes and longer life expectancies. The proportions of the different pollen morphs, together with pollen tube growth rates, may be inherited traits based on differences among maternal families. These results suggest that the different pollen morphs are favored in different pollination ecology situations. The production of several pollen morphs by the same individual could therefore be evolutionarily advantageous.     

The effects of loss of sex in clonal populations of an endangered perennial Coptis teeta (Ranunculaceae)

The reproductive biology of two subspecies of Coptis teeta was studied. The species is an endangered, herbaceous perennial, endemic to the Mishmi Hills in Eastern Himalaya . We investigated reproductive strategies of the subspecies, their breeding systems and reproductive output. They had contrasting reproductive strategies; ssp. teeta reproduces sexually and produces fewer ramets, while ssp. lohitensis propagates mostly vegetatively through a higher number of ramets . There is a resource adjustment between vegetative and sexual functions with a greater allocation to the former, particularly in ssp. lohitensis. Manual crosses showed that the species is self- and cross-compatible, but is essentially an autogamous selfer. Seed set in the species is generally low, with higher output in ssp. teeta and lower in ssp. lohitensis. The low fecundity is ascribed to a number of pre- and postfertilization events, including unsuitability of floral features for attracting pollinators, high pollen sterility and high rates of fruit and seed abortion. Sexual reproduction does not contribute significantly towards population increase in this species. This could be an important factor contributing to its rarity.  © 2003 The Linnean Society of London, Botanical Journal of the Linnean Society , 2003, 143 , 47−54.     

Genetic, Comparative Genomic, and Expression Analyses of the Mc1r Locus in the Polychromatic Midas Cichlid Fish (Teleostei, Cichlidae Amphilophus sp.) Species Group

Natural populations of the Midas cichlid species in several different crater lakes in Nicaragua exhibit a conspicuous color polymorphism. Most individuals are dark and the remaining have a gold coloration. The color morphs mate assortatively and sympatric population differentiation has been shown based on neutral molecular data. We investigated the color polymorphism using segregation analysis and a candidate gene approach. The segregation patterns observed in a mapping cross between a gold and a dark individual were consistent with a single dominant gene as a cause of the gold phenotype. This suggests that a simple genetic architecture underlies some of the speciation events in the Midas cichlids. We compared the expression levels of several candidate color genes Mc1r, Ednrb1, Slc45a2, and Tfap1a between the color morphs. Mc1r was found to be up regulated in the gold morph. Given its widespread association in color evolution and role on melanin synthesis, the Mc1r locus was further investigated using sequences derived from a genomic library. Comparative analysis revealed conserved synteny in relation to the majority of teleosts and highlighted several previously unidentified conserved non-coding elements (CNEs) in the upstream and downstream regions in the vicinity of Mc1r. The identification of the CNEs regions allowed the comparison of sequences from gold and dark specimens of natural populations. No polymorphisms were found between in the population sample and Mc1r showed no linkage to the gold phenotype in the mapping cross, demonstrating that it is not causally related to the color polymorphism in the Midas cichlid.     

Maladaptive mate choice maintained by heterozygote advantage

Common buzzards (Buteo buteo) show a plumage polymorphism that appears to be maintained by heterozygote advantage and allows a maladaptive form of mate choice to persist. The light and dark morphs have a much lower fitness than the presumed heterozygous intermediate morph, but are replenished through Mendelian segregation in intermediate-intermediate pairs. Light and dark morphs could maximize their fitness by mating light with dark to produce all intermediate offspring, but instead choose partners of their own color, thereby producing broods of minimally fit homozygotes. Such maladaptive behavior argues forcefully against mate choice based on "good genes," and its persistence is best explained by heterozygote advantage maintaining the polymorphism coupled with nongenetic mate choice based on sexual imprinting. Modeling different patterns of mate choice shows that random mating and preference for own morph fit our data poorly, whereas preference for mother's morph yields a good fit.