Natural selection against white petals in Phlox.
One common type of polymorphism involves rare white-flowered plants in populations of plants with pigmented flowers (e.g., Delphinium nelsonii, Waser and Price 1981; Phlox pilosa, Levin and Kerster 1970; Digitalis purpurea, Ernst 1987; Echium plantagineum, Burdon et al. 1983; Ipomoea purpurea, Brown and Clegg 1984; Crocus scepusiensis, Rafinski 1979). White corollas may be attributed to one recessive gene (Eschscholzia californica, Frias et al. 1975; Lupinus pilosus, Pazy 1987; Lupinus nanus, Horovitz 1969; Ipomoea purpurea, Epperson and Clegg 1987a; Justicia simplex, Jain and Joshi 1962; Clarkia xantiana, Moore and Lewis 1965; Clarkia unguiculata, Vasek 1968), to one dominant gene (Digitalis purpurea, Ernst 1987; Raphanus raphanistrum, Stanton et al. 1989), or to multiple genes (Lawrence and Price 1940; Grant 1975). The genetic basis for white petals may vary among plants within the same population (Endymion non-scriptus, Stickland and Harrison 1977).
Corolla color variation is of particular interest, because pollinators have keen color vision (Kevan 1983) and can differentiate between corolla color variants (Kay 1978). Their ability to differentiate may result in assortative pollination (Kay 1976, 1982; Levin and Watkins 1984) or discrimination against certain variants (Levin 1972a, Waser and Price 1983). Variants discriminated against will be at a selective disadvantage as a result of lower seed-set (Harding 1970; Waser and Price 1981) or lower paternity (Stanton et al. 1989). These variants also may have higher selfing rates and thus be more inbred than favored variants (Brown and Clegg 1984).
It is tempting to assume that the scarcity of white-flowered plants in local populations is due to pollinator-mediated selection against them. One striking example of such selection is in Delphinium nelsonii, where white-flowered plants produce substantially fewer seeds per flower than plants with pigmented flowers (Waser and Price 1981). Pollinators undervisit albino flowers, because their nectar rewards are more time-consuming to obtain than those of pigmented flowers (Waser and Price 1983).
The present study explores the basis for the paucity of white-flowered plants in Phlox drummondii Hook. Most populations contain only plants with pigmented petals. A small percentage of populations contain white-flowered plants, which rarely constitute over I % of the population. This Phlox is pollinated by butterflies, moths, and hawkmoths, which can differentiate between white and other colors (Levin and Schaal 1970; Levin 1972a), and in garden trials may discriminate against white-flowered plants (Levin 1972a).
A demographic approach was employed to determine whether red-flowered and white-flowered variants of P. drummondii differed in components of fitness or overall fitness. The performance of red- and white-flowered plants from five source populations was studied in various native Phlox sites. Evidence is presented that white-flowered plants from all sources are at a selective disadvantage owing primarily to reduced survivorship to flowering and reduced flower production. There is no indication that pollinators are selective agents.
MATERIALS AND METHODS
The Plant.--Phlox drummondii is a winter annual endemic to central and southeastern Texas. It is almost exclusively self-incompatible (Levin 1985). Seeds germinate from October through December after heavy rains. Plants reach reproductive maturity in March, and flower through early June. They produce from a few to over 100 flowers, each having three ovules. On average, about 70% of the ovules develop into seeds (Levin 1972a). The species typically displays pink flowers based on 3-5 diglucosides of delphinidin and cyanidin or red flowers containing these anthocyanins plus pelargonidin 3-5 diglucosides (Harborne and Smith 1978). White flowers are dictated by a single, recessive gene (Kelly 1920, present study).
Field Studies.--We analyzed the performance of white- and red-flowered plants from four Texas and one Florida population. The locations of these populations, their abbreviations in tables 1 and 2, and the frequencies of the white variant in them are as follows: 2 mi west of Gonzales, Texas, GON, (0.006); 4 mi south of Luling, Texas, LUL, (0.006); 2 mi south of Bastrop, Texas, 2SB, (0.008); 3 mi west of Bastrop, Texas, 3WB, (0.006); and Zephyr Hills, Florida, FLA, (0.012). The Texas populations are native, whereas the Florida population is part of a large system of P. drummondii introduced into the southeastern United States early in this century.
True-breeding progeny were derived from red- and white-flowered parents from each population by crosses between eight to ten plants of each color morph in numerous combinations in the greenhouse. All crosses were made between plants from the same original population. Crosses among white-flowered plants yielded only white-flowered progeny and crosses among red-flowered plants yielded only red-flowered progeny. Seeds of [F.sub.1]s were generated 6-9 mo before planting. Seeds from the various maternal parents were combined into a composite seed source. By using [F.sub.1] seed, we reduce the possibility that one variant was more inbred than the other.
Seeds of red- and white-flowered [F.sub.1]s from each population were planted back into the original Texas Phlox sites as follows. One hundred fifty seeds of each color morph were sown in a quadrat at the center of the site that contained six rows spaced 1 m apart. Each row is a block within the quadrat. Twenty-five seeds of each morph were randomly assigned to positions 10 cm apart in each row. This procedure allowed a comparison of red- and white-flowered plants embedded in their native sites, where they were surrounded by various ecological associates.
In August 1981, seeds from red- and white-flowered Gonzales plants were sown at the Gonzales Phlox site, and seeds from red- and white-flowered Luling plants were sown at the Luling Phlox site. In August 1988, seeds from red- and white-flowered plants from each of the LUL, the 2SB, and 3WB populations were planted in their home sites and in the sites of the other two populations. Zephyr Hills seeds were planted in each of the three aforementioned Texas sites. We included progeny from the Zephyr Hills population to assess the relative performance of the variants at sites very distant and presumably different from the source population. The "away" sites provided additional environments in which to measure the relative performance of the variants. We chose to assess performance in additional environments, because "home" environments vary among years in ways represented in part by the additional environments. We wished to determine whether white-flowered plants are at a liability over a range of environments suitable for the species.
As a result of quadrat destruction by gophers and vehicles, complete data sets in 1989 were obtained only from the LUL seeds planted at home; the 2SB seeds at home, at 3WB, and at LUL; the 3WB seeds planted at home and at 2SB, and the Zephyr Hill seed planted at LUL and at 3WB. All told, we obtained complete data sets $rom ten quadrats, two from 1981-1982 and eight from 1988-1989.
In 1981-1982 and 1988-1989, the sites were visited every 4 wk from early October through the end of the growing season in early June to score emergence, survival flower number, and fruit-set for each plant. Seed-set per plant was estimated by multiplying the number of fruits by the average number of seeds in samples of 50 fruits per row for each type of plant. Typically there are 2.6-2.7 seeds per fruit. The net reproductive rate was determined for each morph in each quadrat. The net reproductive rate ([R.sub.0]) is an estimate of the average number of seeds produced per seed planted. It is the product of the emergence proportion, survivorship to flowering and the mean number of seeds per flowering plant (fecundity). The relative fitness of the white-flowered variant in each quadrat is the [R.sub.0] of the white variant divided by the [R.sub.0] of the red variant.
The analysis of the data for discrete life-history stages was based on values for each row (i.e., each replicate) within a quadrat. The germination and survivorship data for each row were investigated by an analysis of variance on arcsine transformed values, and the mean flower number and mean fecundity (fruit number X mean seed-set per fruit) of plants in each row were analyzed by an analysis of variance on logtransformed values. Values were log-transformed in order to normalize their distribution.
Performance of Gametophytes.--The performance of gametophytes with the alternate color alleles was determined by backcrossing red-flowered heterozygotes to white-flowered plants (homozygotes). If neither gametophyte were favored, we would expect to obtain a Mendelian ratio (1:1) in the progeny. Backcross families were generated for the Gonzales and Zephyr Hills plants.
Test for Allelism.--Although white petals are controlled by recessive alleles (Kelly 1920), it remained to be determined whether white petals in all study populations are dictated by the same allele. To address this issue, crosses were made in all combinations between five white-flowered plants from each of the five populations. If this phenotype were dictated by alleles at the same locus, all progeny would have white petals. If different alleles at the same locus or different loci were involved, progeny would have red petals.
Demography Experiments.--The performance of the red and white variants are summarized in table 1. In 1982, the Gonzales white variant had significantly lower survivorship than its red counterpart, and at Luling white had significantly lower flower production, fecundity and [R.sub.0] than the red (table 2). The performance of red- and white-flowered plants across sites in 1982 were compared with a two-factor mixed-model ANOVA, morph being a fixed effect and site being a random effect. Row is treated as a blocking factor. Morph is tested over the row x morph interaction. The two morphs did not differ in performance at any single life-history stage.
[TABULAR DATA 2 OMITTED]
In 1988-1989 white-flowered plants were substantially less fit than their red-flowered counterparts wherever they were planted (table 1). Though often inferior in many respects, the white morph usually was not significantly different from the red in single performance measures at a given site or with seed from a given source population (table 2). The performance of fitness components across sites were compared with a three-factor mixed model ANOVA, with morph and planting site being fixed effects, and seed source a random effect. Row was a blocking factor. This analysis revealed significant differences between morphs in flower number ([F.sub.3,80] = 19.02; P < 0.001), fruit-set ([F.sub.3,80] = 7.16; P < 0.001), fecundity ([F.sub.3,80] = 5.93; P = 0.001); and [R.sub.0] ([F.sub.3,80] = 8.14; P < 0.001).
Two approaches were taken to determine whether there were overall differences in fitness between the morphs in the ten experimental plantings. One involved combining probabilities of significance from each red-white pair of morphs planted at each site as determined by the aforementioned ANOVAs (Sokal and Rohlf 1981). For all life-history features considered, [chi square] 20 was less than 31.41 critical value. The other approach was a Wilcoxon signed-ranks test based on the mean value for stage-specific performance for each morph at each site. This test did show significant differences. The white-flowered morph had significantly lower survivorship ([T.sub.s] = 2, P < 0.01), flower number ([T.sub.s] = 6, P < 0.05), fecundity ([T.sub.s] = 2, P < 0.01), [R.sub.0] ([T.sub.s] = 0, P < 0.01), and relative fitness ([T.sub.s] = 0, P < 0.01) than its red-flowered counterpart.
Performance of Gametophytes.--Progeny ratios were obtained for 20 heterozygote x white crosses among Gonzales plants and for 11 heterozygote x white crosses among Zephyr Hills plants. There were no significant deviations from Mendelian expectations in any family. Of the 366 progeny scored from Gonzales, 199 were red-flowered and 167 were white-flowered. The deviation from a 1:1 ratio was not statistically significant ([chi square] = 2.89). Of the 158 progeny scored from Zephyr Hills, 89 were red-flowered and 69 were white-flowered. The deviation from a 1:1 ratio was not statistically significant ([chi square] = 2.53). Accordingly, we may conclude that there was no selective differential between gametophytes bearing different corolla color alleles.
Test for Allelism.--Crosses between five white-flowered plants from each site were made in all possible combinations. Each crossing combination yielded at least 15 progeny. All of the progeny from all of the combinations were white-flowered. Accordingly, the genetic basis for white flowers is the same in all of the populations studied.
The white-flowered variant was at a selective disadvantage to the red-flowered variant in every planting, its source and planting site notwithstanding. The mean relative fitness of the white-flowered variant was 0.62; the range was 0.41 to 0.88. For the most part, the relative performance of the white variant varied little among seed sources or planting sites.
The selective disadvantage of the white variant was independent of site quality as measured by its net reproductive rate. This is best seen at LUL, where its [R.sub.0] was about six times greater in 1982 than in 1989, yet its relative fitness was 0.58 in 1982 and 0.53 in 1989.
The selective disadvantage of the white variant cannot be attributed to discrimination by pollinators. The mean seed-set per flower and the percentage of flowers setting fruit were similar for each variant within each site and across sites. If pollinators were discriminating against white petals, there was ample opportunity for it to be expressed, because fruit-set was pollinator limited. This limitation was demonstrated with pollen augmentation experiments in the same year at two study sites (2SB and 3WB). Augmented flowers had over 85% fruit-set compared to the approximately 55% in unmanipulated flowers (Plitmann and Levin MS.). This limitation also was evident in the number of pollen tubes perpistil at the 2SB site in spring 1989. In a sample of 150 pistils from red-flowered native plants, the mean number of tubes per pistil was 2.6 (Levin MS).
The lack of pollinator discrimination against certain petal colors is by no means unique to Phlox. It also has been observed in Cirsium palustre (Mogford 1978), Aquilegia caerulea (Miller 1981), Platystemon californicus (Hannan 1981), Lotus corniculatus (Jones et al. 1986), and Ipomopsis aggregata (Elam and Linhart 1988).
The lower fitness of the white-flowered Phlox is due to its reduced survivorship and fecundity, the latter being a consequence of reduced flower production. This in turn is a reflection of smaller average size, because flower number is strongly correlated with biomass (r = 0.93, Leverich and Levin 1979). The lower survivorship of the white morph may reflect a competitive inferiority. In dense greenhouse mixtures of white and pigmented plants, the frequency of the white variant declined by approximately 50% during the flowering period (Bazzaz et al. 1982).
The most likely explanation for the reduced vigor of white-flowered Phlox drummondii is pleiotropy. Deleterious pleiotropic effects often are associated with major mutations and prevent their increase in natural populations (Wright 1980; Charlesworth et al. 1982; Lande 1983). The pleiotropic effects of a pigment mutation has been demonstrated in tomato where a recessive mutation hp (which enhances fruit lycopene and beta-carotene production) also causes increased seedling mortality, brittle stems, and premature defoliation (Jarret et al. 1984). The pleiotropic effect of mutation at an anthocyanin locus is indicated in chimeric plants where white-flowered sectors show less growth than pigmented ones (Van Fleet 1969). In Ipomaea purpurea, white flower homozygotes at the W locus are vegetatively larger and produce more flowers and seeds than darkly pigmented homozygotes (Rauscher and Fry 1993). Flower color genes in this species also affect stem pigmentation, which is manifested early in the life cycle (Schoen et al. 1984). In Clarkia unguiculata, flower-color genes also control seedling pigmentation (Bowman 1987). The expression of flower-color loci in juvenile plants opens the possibility for selection on these loci prior to flowering. Many mutations affecting the anthocyanin biosynthetic pathway in corn have major effects on growth and viability (Coe et al. 1988).
Reduced fitness unrelated to pollinator discrimination has been observed in rare white-flowered variants in other genera. Burdon et al. (1983) reported that the white variant of Echium plantagineum was an inferior competitor to the common purple-blue variant in greenhouse and field experiments. The frequency of white-flowered plants in natural populations was an inverse function of plant density, being 17 times less frequent in the most dense populations than in the most sparse ones. The average biomass of pigmented plants was nearly twice that white-flowered plants in a natural population sampled in two growing seasons. In Digitalis purpurea, the red-flowered variant produced about twice the number of flowers as the white (Ernst 1987). The white variant also required more water, phosphorus, potassium and magnesium than the red. Moreover, leaves of the former were unable to correctly orient themselves when light was scarce, whereas red-flowered plants did so.
Given the strength of selection against white-flowered Phlox, the question arises as to the mechanism responsible for its retention in natural populations. Mutation is one possibility. Mutation rates in excess of 0.001 have been demonstrated for genes affecting flower pigmentation in many species. In Ipomae purpurea such mutations are associated with a mutator (unstable) gene that is close to the gene conferring white petals (Epperson and Clegg 1987a). High germinal mutation rates to white-flowered phenotype have been described in Medicago sativa (Bingham and Clement 1989; Talbert and gingham 1989) and Glycine max (Palmer et al. 1989), and in both species may involve transposable elements. Unstable alleles producing germinal mutations to white flowers also are known in several other species (cf. gingham and Clement 1989). In addition to germinal mutations, the unstable genes in the aforementioned species also produce variegated flowers.
We have not compared the performance of pigmented homozygotes and heterozygotes in Phlox, and thus cannot attest to the presence of heterozygote advantage or the lack thereof. Heterozygote advantage might be responsible for maintaining recessive, pink-flower genes in the face of selection in Lupinus nanus (Harding 1970) and for maintaining an orange-yellow corolla color polymorphism in Eschscholzia californica (Fries et al. 1975).
We wish to thank M. Scioli for assistance with the data analysis. We are grateful to N. Waser, C. Galen, and anonymous reviewers for suggesting ways to improve this paper.
Bazzaz, E. A., D. A. Levin, and M. R. Schmierbach. 1982. Differential survival of genetic variants in crowded populations of Phlox. Journal of Applied Ecology 19:891-900.
gingham, E. T., and W. M. Clement. 1989. Alfalfa transposable elements and variegation. Developments in Genetics 10:552-560.
Bowman, R. N. 1987. Cryptic self-incompatibility and the breeding system of Clarkia unguiculata (Onagraceae). American Journal of Botany 74:471-476.
Brown, B. A., and M. T. Clegg. 1984. Influence of flower color on genetic transmission in a natural population of the common morning glory, Ipomaea purpurea. Evolution 38:796-803.
Burdon, J. J., D. R. Marshall, and A. H. D. Brown. 1983. Demographic and genetic changes in populations of Echium plantagineum. Journal of Ecology 71:667-679.
Charlesworth, B., R. Lande, and M. Slatkin. 1982. A neo-Darwinian commentary on macroevolution. Evolution 36:474-498.
Coe, E. M., M. G. Neuffer, and D. A. Hoisington. 1988. The genetics of corn. Pp. 81-258 in G. F. Sprague and J. W. Dudley, eds. Corn and corn improvement. American Society of Agronomy, Madison, Wisc.
Compton, S. G., S. G. Beesley, and D. A. Jones. 1988. Variation in the colour of the keel petals in Lotus corniculatus L. 5. Successional differences in the distribution of dark-keeled plants. Heredity 61:235-245.
Elam, D. R., and Y. B. Linhart. 1988. Pollination and seed production in Ipomopsis aggregata: Differences among and within flower morphs. American Journal of Botany 75:1262-1274.
Epperson, B. K., and M. T. Clegg. 1987. Instability at a flower color locus in the morning glory. Journal of Heredity 78:346-352.
Ernst, W. H. O. 1987. Scarcity of flower colour polymorphism in field populations of Digitalis purpurea L. Flora 179:231-239.
Frias, L. D., R. Godoy, R Iturra, S. Koref-Santibanez, J. Navarro, N. Pacheco, and G. L. Stebbins. 1975. Polymorphism and geo-graphic variation of flower color in Chilean populations of Eschscholzia californica. Plant Systematics and Evolution 123:185-198.
Grant, V. 1975. Genetics of Flowering Plants. Columbia University Press, New York.
Hannan, G. L. 1981. Flower color polymorphism and pollination biology of Platystemon californicus Benth. (Papaveraceae). American Journal of Botany 68:233-243.
Harborne, J. B., and D. M. Smith. 1978. Correlation between anthocyanin chemistry and pollination ecology in the Polemoniaceae. Biochemical Systematics and Ecology 6:127-130.
Harding, J. 1970. Genetics of Lupinus. II. The selective disadvantage of the pink flower color mutant in Lupinus nanus. Evolution 24:120-127.
Horovitz, A. 1969. Effect of flower color variations on the mating system in some forms of Lupinus nanus Dougl. (ex Benth.). Ph.D. diss. University of California, Davis.
Horovitz, A., and D. Zohary. 1966. Spontaneous variegation for perianth colour in wild Anemone coronaria. Heredity 21:513-515.
Jain, S. K., and B. C. Joshi. 1962. Local differentiation in some natural populations of Justicia simplex. Genetics 47:789-781.
Jarret, R. L., H. Sayama, and E. C. Tigchelaar. 1984. Pleiotropic effects associated with the chlorophyll intensifier mutations high pigment and dark green in tomato. Journal of the American Society for Horticultural Science 109:873-878.
Jones, D. A., S. G. Compton, T. J. Crawford, W. M. Ellis, and I. M. Taylor. 1986. Variation in the colour of the keel petals in Lotus corniculatus L. 3. Pollination, herbivory and seed production. Heredity 57:101-112.
Kay, Q. O. N. 1976. Preferential pollination of yellow-flowered morphs of Raphanus raphanistrum by Pieris and Erastilis species. Nature 261: 230-232.
--. 1978. The role of preferential and assortative pollination in the maintenance of flower colour polymorphisms. Pp. 175-190 in A. J. Richards, ed. The pollination of flowers by insects. Academic Press, London.
--. 1982. Intraspecific discrimination by pollinators and its role in evolution. Pp. 9-28 in J. A. Armstrong, J. M. Powell, and A. J. Richards, eds. Pollination and evolution. Royal Botanical Gardens, Sidney.
Kelly, J. P. 1920. A genetical study of flower form and flower color in Phlox drummondii. Genetics 5:189-248.
Kevan, R G. 1983. Floral colors through the insect eye: what they are and what they mean. Pp.3-30 in C. E. Jones and R. J. Little, eds. Handbook of experimental pollination ecology. Van Nostrand Reinhold, New York.
Kyhos, D. W. 1971. Evidence of different adaptations of flower color variants of Encelia farinosa (Compositae). Madrono 21:49-61.
Lack, A. J., and Q. O. N. Kay. 1987. Genetic structure, gene flow and reproductive ecology in sand-dune populations of Polygala vulgaris. Journal of Ecology 75:259-276.
Lande, R. 1983. The response to selection on major and minor mutations affecting a metrical trait. Heredity 50:47-65.
Leverich. W J.. and D. A. Levin. 1979. Age-specific survivorship and reproduction in Phlox drummondii. American Naturalist 113:881-903.
Levin, D. A. 1972a. The adaptedness of corolla color variants in experimental and natural populations of Phlox drummondii. American Naturalist 106:57-70.
--. 1972b. Low frequency disadvantage in the exploitation of pollinators by color variants in Phlox. American Naturalist 106:453-460.
--.1985. Reproductive character displacement in Phox. Evolution 39:1275-1281.
Levin, D. A., and Z. Bulinska-Radomska. 1988. Effects of hybridization and inbreeding on fitness in Phlox. American Journal of Botany 75:1632-1639.
Levin, D. A., and H. W. Kerster. 1970. Phenotypic dimorphism and population fitness in Phlax. Evolution 24:128-134.
Levin, D. A., and B. A. Schaal. 1970. Corolla color as an inhibitor of interspecific hybridization in Phlox. American Naturalist 104:343-353.
Levin, D. A., and K. Schmidt. 1985. Dynamics of a hybrid zone in Phlox: an experimental demographic investigation. American Journal of Botany 72:1404-1409.
Levin, D. A., and L. Watkins. 1984. Assortative mating in Phlox. Heredity 53:595-602.
Lawrence, W. J. C., and J. R. Price. 1940. The genetics and chemistry of flower color variation. Biological Review 15:35-58.
Miller, R. B. 1981. Hawkmoths and the geographic pattern of floral variation in Aquilegia caerulea. Evolution 35:763-774.
Mogford, D. J. 1974. Flower color polymorphism in Cirsium palustre. 1. Heredity 33:241-256.
--. 1978. Pollination and flower colour polymorphism, with special reference to Cirsium palustre. Pp. 191-199 in A. J. Richards, ed. The pollination of flowers by insects. Academic Press, London.
Moore, D. M., and H. Lewis. 1965. The evolution of self-pollination in Clarkia xantiana. Evolution 19:104-114.
Palmer, R. G., B. R. Hedges, R. S. Benavente, and R. W. Groose. 1989. w4-mutable line in soybean. Developments in Genetics 10:542-551.
Pazy, B. 1987. Flower-color polymorphism in Lupinus pilosus in Israel. Plant Breeding 99:327-329.
Rafinski, J. N. 1979. Geographic variability of flower colour in Crocus scepusiensis. Plant Systematics and Evolution 131:107-125.
Rauscher, M. D., and J. D. Fry. 1993. Effects of a locus affecting floral pigmentation in Ipomaea purpurea on female fitness components. Genetics 134:1237-1247.
Schoen, D. J., D. E. Giannasi, R. A. Ennos, and M. T. Clegg. 1984. Stem color and pleiotropy of genes determining flower color in the common morning glory. Journal of Heredity 75:113-116.
Sokal, R. R., and S. J. Rohlf. 1981. Biometry, 2d ed. W. H. Freeman, San Francisco.
Stace, H. M., and Y. J. Fripp. 1977. Raciation in Epacris impressa. III. Polymorphic populations. Australian Journal of Botany 25:325-336.
Stanton, M. L., A. A. Snow, S. N. Handel, and J. Bereczky. 1989. The impact of flower-color polymorphism on mating patterns in experimental populations of the wild radish (Raphanus raphanistrum L.). Evolution 43:335-346.
Stickland, R. G., and B. J. Harrison. 1977. Precursors and genetic control of pigmentation. 3. Detection and distribution of different white genotypes of bluebells (Endymion species). Heredity 39:327-333.
Talbert, L. E., and E. T. gingham. 1989. Genetic characterization of a mutable line in alfalfa (Medicago sativa L.). Journal of Heredity 80:407 410.
Van Fleet, D. S. 1969. An analysis of the histochemistry and the function of anthocyanin. Advancing Frontiers of Plant Science 23:65-89.
Vasek, E C. 1968. Outcrossing in natural populations. IV. A comparision of outcrossing estimation methods in E. T. Drake, ed. Evolution and environment. Yale Univeristy Press, New Haven, Conn.
Waser, N. M., and M. V. Price. 1981. Pollinator choice and stabilizing selection for flower color in Delphinium nelsonii. Evolution 35: 376-390.
--. 1983. Pollinator behavior and natural selection for flower color in Delphinium nelsonii. Nature 302:422-424.
Wilken, D. H., and S. T. Allard. 1986. Intergradation among populations of the Ipomopsis aggregata complex in the Colorado frontrange. Systematic Botany 11:1-13.
Wright, S. 1980. Genic and organismic selection. Evolution 34:825-843.
|Printer friendly Cite/link Email Feedback|
|Author:||Levin, Donald A.; Brack, Ellen T.|
|Date:||Oct 1, 1995|
|Previous Article:||A history of host associations and evolutionary diversification for Ophraella (Coleoptera: Chrysomelidae): new evidence from mitochondrial DNA.|
|Next Article:||Fine-scale spatial structure: correlations for individual genotypes differ from those for local gene frequencies.|