| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
2 Department of Evolution, Ecology and Marine Biology, University of California, Santa Barbara, California 93106 USA; and 3 Division of Biological Sciences, 105 Tucker Hall, University of Missouri, Columbia, Missouri 65211-7400 USA
Received for publication November 20, 1998. Accepted for publication September 28, 1999.
ABSTRACT
Competition among pollen grains for the fertilization of ovules can play an important role in determining the male and female reproductive success of flowering plants. To examine the influence of pollen-donor genotype on male reproductive success, hand-pollinations were conducted on Clarkia unguiculata and the siring success of pollen-donor plants was compared between donors homozygous for different allelomorphs of the allozyme PGI (phosphoglucoisomerase). Donors homozygous for the B allele sired more seeds than C-allele donors. Single-donor crosses indicated that C-donor-sired seeds are aborted more often than are B-donor-sired seeds, suggesting that the B-allele donor's advantage in mixed pollinations was a result of differential abortion. A negative relationship between pollen load and the siring success of B-allele donors implies that pollen from B-allele donors has reduced performance relative to C-allele donors when pollen loads are high. These data demonstrate consistent differences in siring success between individuals homozygous for different alleles at a single locus and suggest that variation at the Pgi locus may be maintained by a post-pollination trade-off.
Key Words: Clarkia • gametophytic competition • Onagraceae • Pgi • pollen performance • reproductive success • siring success
Outcrossing angiosperms typically produce far more pollen grains than ovules. One consequence of the production of surplus pollen is that flowers may receive more pollen than is required for full seed set, such that pollen grains must compete for the fertilization of a limited number of ovules. An increasing number of experimental studies indicate that the proportions of seeds sired do not reflect a random sample of the pollen available on stigmas (e.g., Pfahler, 1967
; Marshall and Ellstrand, 1986
; Snow and Mazer, 1988
; Cruzan, 1990
; Snow and Spira, 1991a, b
; Quesada et al., 1991
; Nakamura and Wheeler, 1992
; Radha et al., 1993
; Rigney et al., 1993
; Carney, Cruzan, and Arnold, 1994
; Baker and Shore, 1995
; Cruzan and Barrett, 1996
). Variance among pollen donors in siring success when all else is equal (e.g., pollen deposition) is attributed to variation in pollen performance (germination and growth rate) and the frequency of abortion of seeds sired by those donors (Stephenson and Bertin, 1983
; Willson and Burley, 1983
; Willson, 1990
).
Currently, our understanding of the factors that determine siring success following gametophytic competition is limited (Lyons et al., 1989
; Marshall and Folsom, 1991
). If variance among donors in siring success is due to heritable variation in traits that influence competitive ability and siring success, then selection is expected to lead to an evolutionary response and potentially the fixation of favored alleles. Here, we present data from an experiment on Clarkia unguiculata Lind. (Onagraceae) designed to test the null hypothesis that siring success is random among competing pollen donors with different genotypes at a single locus.
Several studies have demonstrated nonrandom siring success following gametophytic competition among unrelated pollen donors. Crosses involving competition between donors from different subspecies or cultivars have resulted in nonrandom siring success (Devries, 1924
; Quesada et al., 1991
; Carney, Cruzan, and Arnold, 1994, 1996
). It is also common to find differences among outcross donors in the proportion of seeds sired when the outcross pollen is competing against self pollen (Bowman, 1987
; Montalvo, 1992
; Sork and Schemske, 1992
; Johnston, 1993
; Jones, 1994
). Differences in siring success have also been found among genetically distinct outcross donors following gametophytic competition (Sari-Gorla, Ottaviano, and Faini, 1975
; Marshall and Ellstrand, 1986
; Epperson and Clegg, 1987
; Snow and Mazer, 1988
; Marshall, 1990, 1991
; Nakamura and Wheeler, 1992
; Snow and Spira, 1991a, b
; Rigney et al., 1993
).
In contrast to evidence that siring success varies among genetically distinct donors in several plant species, other studies have failed to support the hypothesis that there is a genetic basis to either pollen performance, seed abortion, or siring success. Snow and Mazer (1988)
found that there was no detectable response to selection for increased pollen competitive ability in Raphanus raphanistrum. Mazer (1987)
found significant differences among donors in in vitro pollen performance of Raphanus raphanistrum, but found no significant variation among the same donors in siring success following pollinations in which pollen from a single donor was placed on each flower.
We conducted hand pollinations to compare the siring success of pollen donors with different genotypes at the locus coding for the production of the allozyme phosphoglucoisomerase (PGI). Despite the fact that allozymes are often assumed to be neutral genetic markers for gametophytic competition studies (Marshall and Ellstrand, 1986
; Bertin, 1990
; Cruzan, 1990
; Montalvo, 1992
; Johnston, 1993
; Rigney et al., 1993
; Carney, Cruzan, and Arnold, 1994
) there are reasons to suspect that different Pgi alleles could confer differential siring success to pollen donors in competition.
PGI plays an important role in metabolism by catalyzing the first step of glycolysis. Modifications of this enzyme in Clarkia xantiana have resulted in reduced production of starch and glucose (Kruckeberg et al., 1989
). Moreover, different genotypes of Pgi have been associated with different metabolic properties (e.g., thermal stability and Km) in a variety of organisms (Watt, 1977, 1983
; Jones, 1985
). However, Gottlieb and Greve (1981)
found no differences in the metabolic properties of plants with different Pgi alleles. The metabolic role of this enzyme could link the Pgi genotype of a pollen-donating plant with the siring success of that plant. If different alleles for Pgi result in different levels of metabolic activity among pollen donors in C. unguiculata, then pollen from donors with different genotypes differ with respect to performance. Specifically, the process of pollen tube growth is likely to be influenced by metabolic rate because tube growth involves the enzymatic breakdown of stylar tissue and the production of constituents of the pollen-tube wall (Heslop-Harrison, 1987
). Alternatively, variance in metabolic activity of embryos could lead to differential abortion of those embryos. Slow development of some embryos as a result of reduced metabolic activity relative to other embryos in the same ovary could reduce the proportion of maternal resources acquired by the slower embryos and result in their abortion (Snow, 1994
). Thus, embryos fertilized by donors with different Pgi genotypes may be aborted at different frequencies.
The experiment presented here was designed to determine whether there are differences between pollen donors with different Pgi genotypes in: (1) average siring success, (2) the proportion of sired seeds that are aborted, viable, and undeveloped, and (3) pollen performance.
MATERIALS AND METHODS
Clarkia unguiculata is a species of annual plant that is common on steep hillsides of oak woodlands in California. The large, showy flowers attract native bees, honeybees, bumblebees and occasionally hummingbirds as pollinators (MacSwain, Raven, and Thorp, 1973
). All of the parental plants in this experiment were derived from seed collected from a single population of C. unguiculata (Santa Barbara County, California, USA). Although the outcrossing rate has not been quantified in the source population of the current study, other populations of Clarkia unguiculata are predominantly outcrossing (Vasek, 1965
). Further, the large protandrous flowers in this population strongly suggest an outcrossing mating system since other large-flowered, protandrous, closely related Clarkia spp. have been found to be outcrossing (Moore and Lewis, 1965
; Holtsford and Ellstrand, 1992
).
We collected seed from 60 randomly chosen maternal plants. Ten seeds from each of the 60 maternal families were germinated in Metro mix (40% vermiculite, 40% peat moss, 20% bark), then the seedlings were transferred 2 wk later to plastic tubes (3 cm diameter, 15 cm length) filled with pure fritted clay. The seedlings were placed in a glasshouse, watered twice daily, and fertilized (25% solution of Peterson's Excel 15-5-15) once weekly for the remainder of the study.
We used starch gel electrophoresis to determine the Pgi genotype of each seedling. We extracted tissues from young leaves and followed electrophoretic procedures with System I buffers used by Gottlieb (1981)
.
In C. unguiculata, Pgi has two cytosolic loci (C1 and C2) with multiple alleles at each locus (Gottlieb, 1977
; Gottlieb and Ford, 1996
). At Pgi-C1 there are three alleles in this population that we refer to as A, B, and C, where A is the most anodally migrating and C is the least anodally migrating of the bands in gels. Two pollen donors, one homozygous for the B allele and one for the C allele, were assigned to each pollen recipient. There were five pollen recipients, which were homozygous for B, and five recipients homozygous for C. A different pair of pollen donors was used for each pollen recipient, i.e., donors were nested within recipients. Each of these parental plants was from a different field-collected maternal sibship. Since the source population was probably highly outcrossed we believe that the parental plants were not closely related to each other. The assumption of unrelated parents is not critical, however, to any of the hypotheses tested.
We conducted single-donor and two-donor (competitive) pollinations. The two-donor pollinations were conducted by mixing pollen from both donor plants assigned to each recipient in a 1:1 ratio prior to pollination. We standardized the amounts of pollen in the mixture from the two plants by either counting each of the pollen grains taken from the donors (for the low pollen load treatment) or by removing equal lengths of anther from each of the two donors and stripping all of the pollen from those anther sections (for the medium and high pollen load treatments; see below). In C. unguiculata there is a strong relationship between the number of pollen grains in an anther sac and the length of the anther sac (y = 180.5x -23.4, df = 27, R2 = 0.955). We measured anthers using a dissecting microscope, cut the anther with a scalpel, and then stripped the pollen from the anther sacs with a dissecting needle. The pollen from each of the two donors was placed on a glass microscope slide and thoroughly mixed with two dissecting needles <3 min. before pollination. A dissecting needle was also used to transfer pollen to the stigmas.
We applied variable pollen loads to recipient stigmas to distinguish between pre- and post-zygotic mechanisms of nonrandom mating. This experimental design is a modification of one developed by Cruzan and Barrett (1996
; see also Bertin, 1990
). If a donor has superior pollen performance relative to a competing donor, then when pollen is limiting there will be a positive relationship between the percentage of seeds sired by the superior donor and pollen load. At very low pollen loads, the siring success should be ~50% for both donors (assuming a 1:1 mixture of pollen) since even slow pollen tubes may fertilize ovules when pollen is limiting. However, when the total pollen load is greater than that required for full seed set, the donor with superior pollen performance should sire >50% of the seeds. In contrast, if nonrandom siring success is a result of differential abortion then there should not be a positive relationship between pollen load and siring success of the superior donor, assuming there is no frequency dependence in abortion rates.
Low, medium, and high pollen loads differed in the total number of pollen grains in the mixed pollen load, but in all three treatments the two donors contributed equal amounts of pollen. The mean number of pollen grains transferred per cross (±1 SE) in a pilot study was 76.9 ± 3.4 for low, 189.4 ± 8.1 for medium, and 505.9 ± 21.5 for high pollen loads. Based on an estimate of 95 ovules per ovary (SE = 2.6, N = 22, range = 76–124), the average pollen:ovule ratios for the pollen load treatments are 0.81:1 for low, 1.99:1 for medium, and 5.32:1 for high.
Each pollen load was replicated three times for a total of nine pollinated flowers per recipient plant. There were ten recipient plants. The same sequence of pollination treatments was used in all ten sets. The first three flowers were pollinated with low, medium, and high pollen loads, respectively. In this way, the pollination treatments were alternated through all nine pollinated flowers per recipient. All recipient flowers were emasculated within 24 h of bud break. Clarkia unguiculata are highly protandrous and the onset of stigma receptivity is ~2 d after bud break. In the case of high pollen load crosses, we saturated the stigma with pollen. As a result, it was not always possible to use all of the pollen available on the microscope slide.
To obtain an accurate count of the number of pollen grains deposited in each experimental pollination, we cut off the pollinated stigmas from each flower 24 h after pollination, preserved them in alcohol, and later counted the number of pollen grains on them. The pollen loads were applied as one of three treatment levels in the experiment, but due to variation in the actual number of grains transferred, pollen load was treated as a continuous variable in all subsequent analyses.
We used cellulose acetate gel electrophoresis to score seed genotype (Hebert and Beaton, 1993
). Ten seeds from the basal half and ten seeds from the stylar half of each fruit were chosen randomly and assayed. We assayed seeds that had been hydrated on wet filter paper for a minimum of 8 h at 4°C. The seeds were first crushed in the same extraction buffer used for the starch gels. The crude extracts were then run on cellulose acetate gels using a pH 8.3 Tris-Glycine buffer system (Hebert and Beaton, 1993
). The gels were run for 20 min at 200 V. Twelve seeds were assayed per gel with two lanes per gel devoted to seeds from single-donor crosses of each of the two competing donors from the pertinent set as controls. The gels were then stained with agar overlays following Hebert and Beaton (1993)
. Protein products from Pgi-C1 and Pgi-C2 form interlocus heterodimers (Gottlieb, 1977
; Gottlieb and Ford, 1996
). However, all 30 parent plants (20 pollen donors and ten recipients) were homozygous for a common allele at the Pgi-C2 locus so that gel scoring was simplified. The results from electrophoresis of the seeds were used to estimate the total percentage of seeds sired in each fruit by the B-allele donor (hereafter SSB) and the C-allele donor (hereafter SSC) and the percentage sired by each allele in the basal and stylar halves of each fruit.
We also conducted single-donor crosses between the recipient of each set and each of the two competing donors to determine the frequency of seed abortion following both types of cross. In these pollinations, we simply saturated a recipient stigma with pollen from one of the two donors per set. One pollination was conducted per donor. Once the fruits matured, we collected them before they dehisced and counted the number of undeveloped ovules and late-aborted and viable seeds per fruit. The fruit contents from two-donor pollinations were also categorized in this way. Late-aborted seeds are smaller and lighter in color than viable seeds. Undeveloped ovules, which appear as very small (<0.1 mm diameter) yellow disks, may result from early abortion (Nakamura and Stanton, 1987
) or lack of fertilization.
Analysis
The results were analyzed to determine whether the proportion of seeds sired by each donor differed significantly from the relative proportions of each type of pollen in the pollen load (50%). We calculated the mean SSB for each pair of competing donors by calculating the average SSB value across all nine replicate fruits from each of ten sets (recipient and two nested donors) of plants. Raw SSB values were arcsine square-root transformed prior to analysis to improve the normality of the data (Zar, 1984
). The ten mean SSB values from the ten sets of plants were than analyzed with a t test for significant deviation from a mean of 50%.
We examined the relationship between pollen load and SSB by conducting an analysis of covariance on arcsine square-root transformed data (SAS, 1994
). Plant set and recipient genotype were included as discrete variables in the analysis and pollen load was included as a covariate. The SSB values from only 81 fruits were included in the analysis because nine fruits failed to produce the 20 seeds needed for determination of donor siring success.
In order to compare the two types of single-donor crosses (B-allele donor and C-allele donor) for percentage of seeds that were viable, undeveloped, and aborted, we conducted a nonparametric Wilcoxon signed-rank test for paired data (Zar, 1984
). In single-donor crosses, because we crossed each of the competing donors in a set with the same pollen recipient from that set, we paired the percentage counts from the two single-donor fruits per set. These analyses tested the hypothesis that the percentage of viable, aborted, and undeveloped seeds per fruit was the same between paternal genotypes for a given maternal plant. All statistical analyses were conducted with the JMP statistical package (SAS, 1994
).
RESULTS
Single-donor crosses
The majority (>50%) of seeds produced from single-donor crosses were viable. However, the percentage of viable seeds per fruit varied among maternal plants and among donor types (Fig. 1). On average, roughly a third of the seeds per fruit were undeveloped (Table 1). Fewer than 15% of the seeds per fruit were late aborted in all single-donor crosses. There were significantly more seeds aborted in single-donor crosses when the donor was homozygous for the C allele vs. the B allele (T = -21.5, P < 0.03, N = 10; Fig. 1, Table 1). In addition, there was a trend for higher abortion percentages of seeds from crosses of C-allele donors with C-allele recipients relative to the other three types of crosses (Table 1). The difference between donor types in the percentage of viable seeds was marginally nonsignificant (T = 18.5, 0.05 < P < 0.06, N = 10). There was no difference between donor types in the total mass of a random sample of ten seeds from each single donor cross (t = -1.463, N = 10, P = 0.178).
|
|
|
|
|
|
|
DISCUSSION
In two-donor competitive pollinations, pollen parents homozygous for different Pgi alleles sired different numbers of seeds per fruit despite the deposition of equal amounts of pollen from the two donors on the same stigma. In a sample of 20 individuals, C. unguiculata plants homozygous for the B allele sired a greater number of seeds compared to C-allele donors. The average percentage of the mature viable seeds produced per fruit sired by the B-allele donor was 56%. An even larger percentage of seeds (mean = 62.2%) were sired by the B-allele donor in the basal half of fruits. These results imply that genotype at a single locus can influence male reproductive success in C. unguiculata. This study demonstrates that pollen donor genotype at the Pgi-C1 locus can influence fitness components through strictly post-pollination processes: competition for ovule fertilizations, and seed maturation. These results are consistent with previous studies that found genetic effects on siring success (for a review see Stephenson et al., 1992
) and single gene effects on pollen performance (Schiefelbein et al., 1993
; Azpiroz et al., 1998
).
An alternative explanation is that differences in siring success among donors were due not to genotype at the Pgi locus but rather were due to genotype at loci in linkage disequilibrium with the Pgi-C1 locus. However, the thirty plants used in this experiment were derived from 30 unrelated maternal sibships. The experimental design randomized donor and recipient genotype except for the B and C alleles at Pgi-C1. If outcrossing rates are high and population substructure is low, then linkage disequilibrium should be low. The population from which all the parental plants descended was also sufficiently large (thousands of individuals) to minimize the effects of drift.
Pre-zygotic vs. post-zygotic mechanisms
The results from the manipulation of pollen load do not fit any simple models of pre- versus post-zygotic mechanisms for the siring success advantage of B-allele donors. There was a negative relationship between the siring success of one type of donor (B-allele) and pollen load as expected if the pollen performance of the C-allele donors is superior to that of B-allele donors (Cruzan and Barrett, 1996
). However, the intercept of the best-fit line for SSB vs. pollen load is above 50%. If the seeds sired by the different donors are not aborted differentially, then half of the seeds should be sired by each donor following pollinations with small pollen loads. The relationship between siring success and pollen load may best be explained by a combination of pre- and post-zygotic mechanisms.
The differences in siring success between B- and C-allele donors must be caused in part by differential seed abortion. In the single-donor crosses, which controlled for recipient identity, the mean proportion of late-aborted seeds was 7.8% for C-allele donor crosses and 3.4% for B-allele donor crosses (Fig. 1). Moreover, crosses in which both parents were C-allele homozygotes had the highest abortion rates of all four possible parental combinations.
Differential abortion of seeds sired by different donors has been demonstrated in other plant species following several types of crosses. Previous experiments on gametophytic competition have attributed nonrandom siring success to post-zygotic mechanisms following crosses among closely related species or races (Gadish and Zamir, 1987
; Abbo and Ladizinsky, 1994
), between self and outcross pollen donors (Bowman, 1987
; Manasse and Pinney, 1991
; Montalvo, 1992
; Rigney et al., 1993
; Jones, 1994
) and, like the current study, among outcross donors within a species (Marshall and Ellstrand, 1988
; Cruzan, 1990
; Rocha and Stephenson, 1990
; Rigney et al., 1993
; Manicacci and Barrett, 1996
). However, previous work has not found a significant difference in abortion rates attributable to genotype at the loci used to determine parentage.
More frequent abortion of seeds sired by C-allele donors in two-donor crosses may result from selective abortion by the recipient. A number of investigators have argued that variation in siring success among competing donors can be caused by selective abortion of fertilized ovules by the recipient (Westoby and Rice, 1982
; Mazer, 1987
; Marshall and Ellstrand, 1988
; Marshall, 1991
). If C. unguiculata recipients selectively allocated resources to developing seeds sired by B-allele relative to C-allele donors or B-bearing embryos coerced more resources from their maternal plants, then this could explain the observed patterns of siring success in the two-donor crosses. However, selective abortion cannot explain higher abortion percentages of C-allele relative to B-allele donor seeds in single-donor crosses. Alternatively, the recipient may have no influence on which seeds are aborted, such that the differential abortion of C-allele seeds results strictly from embryo genotype. Seeds sired by B-allele donors may be more effective at competing for resources than seeds sired by C-allele donors. It is also possible that the C allele may be detrimental to developing embryos. Higher rates of abortion have been associated with pollen donor genotypes in Lillium sp. (Cave and Brown, 1954, 1957
). Either differential resource acquisition or embryo lethality could explain the relatively high abortion rates of seeds sired by C-allele pollen. The increased abortion percentage of seeds from two C-allele parents relative to one parent suggests that some fraction of the negative effects of the C allele are additive.
In contrast to the results of other studies (Bertin, 1990
; Cruzan and Barrett, 1996
), the siring success of the superior competitor (SSB) decreased as pollen loads increased rather than increasing or remaining constant. The negative relationship between SSB and pollen load suggests that despite the higher mean siring success of B-allele donors, C-allele donors produce pollen that outperforms pollen from B-allele donors. If pollen from C-allele donors germinated faster or grew faster than pollen tubes from B-allele donors, but sired seeds that were aborted more frequently than seeds sired by B-allele donors, then SSC should be positively related to pollen load with a y-intercept less than 50%, as the data show.
Position effects
We found significant fruit position effects on the siring success of competing donors, as has been observed in other studies (Marshall and Ellstrand, 1988
; Quesada, Winsor, and Stephenson, 1993
; Carney, Hodges, and Arnold, 1996
). B-allele donors sired a larger percentage of basal seeds than stylar seeds (Table 3). Marshall and Ellstrand (1988)
demonstrated that under favorable growing conditions, some pollen donors of Raphanus sativus differentially fertilize ovules in basal and middle positions of the fruit relative to stylar positions. These position differences during fertilization translate into higher fitness for the donors, because the embryos in these positions are less likely than stylar embryos to be aborted under stressful environmental conditions. Positional differences in the likelihood of abortion have also been found in other plant species (Bawa and Webb, 1984
; Rocha and Stephenson, 1990, 1991
).
Seeds sired by B-allele donors may be aborted less frequently because pollen from these donors is more likely to fertilize basal vs. stylar ovules. However, the differences in seed abortion between the stylar and basal halves were small compared to differences found in other studies. The percentage of seeds aborted in the stylar half of each fruit was higher than in the basal half when all three pollen load treatments were considered together (basal = 6.0%, stylar = 6.5%), but the difference is not statistically significant (Table 1). Considering the pollen load treatments separately, abortion percentages were significantly higher in stylar vs. basal fruit halves only in the high pollen load treatment (7.8 vs. 5.4%; Wilcoxon signed-rank test: T = -65.5, P < 0.035, N = 28; Table 1). Overall, these results suggest that B-allele donors may experience a small increase in siring success by fertilizing basal ovules, which are slightly less likely to be aborted than ovules in the stylar half. However, the preponderance of seeds sired by the B-allele donor cannot be explained by these position effects alone. Seeds sired by the C-allele donor are aborted more frequently than B-donor sired seeds regardless of position, based on the results of single-donor crosses.
Maintenance of variation in Pgi
Given the preponderance of B-allele-sired seeds following two-donor pollinations, selection should favor the spread of B alleles relative to C alleles and ultimately lead to the fixation of the B allele. In contrast with this prediction, Pgi is polymorphic in natural populations of Clarkia. In the population surveyed for this study both the B and C allele were present at the Pgi-C1 locus. Gottlieb (1973)
found that four populations of Clarkia rubicunda were all variable at the C1 locus for Pgi. If the reproductive success of B-allele homozygotes is higher than C-allele homozygotes in natural populations, then some process must be maintaining the C allele despite selection against it.
Previous discussions of maintenance of the variation in Pgi in both plants and animals have focused on selection for different alleles under a variety of environmental conditions. In many studies of both plants and animals, Pgi allele frequencies were found to vary consistently along environmental gradients (Shumaker and Babble, 1980
; Schuster, Alles, and Mitton, 1989
; Shea, 1990
). These studies support the argument that variation at the Pgi locus is maintained by the differential success of different Pgi genotypes in different environments (Riddoch, 1993
). Our results suggest an alternative explanation for the maintenance of variation at the Pgi locus in C. unguiculata through the effects of Pgi genotype on differential siring success of competing pollen donors. The negative relationship between SSB and pollen load suggests that, in the field, the siring success of pollen donors homozygous for one or the other alleles studied here will partially depend on pollen load. In all but the highest pollen loads, the B-allele donor sired a majority of the seeds. It may be that C alleles are maintained in field populations because under certain pollination conditions (i.e., stigmas saturated with pollen) the siring success advantage of donors homozygous for the B allele disappears. In fact, C-allele pollen may have an advantage under high pollen loads (Fig. 3). Variable pollination conditions could result from year-to-year or site-to-site variation in pollen availability or pollinator availability. Future studies would benefit from measuring pollen load variation across time and space in natural populations. In addition, future studies should consider the interplay between selection operating on Pgi allele frequencies through reproductive mechanisms (e.g., gametophytic competition) and nonreproductive mechanisms (e.g., differential plant growth and survivorship).
FOOTNOTES
1 The authors thank Susan Mazer, Bob Warner, Scott Hodges, and John Endler for their constructive comments on early versions of this manuscript. This work was funded by grants to SET from the California Native Plant Society and the National Science Foundation (DEB 95-20611) and a University of Missouri Research Board grant (93-060) to TPH. 
4 Author for correspondence, current address: Department of Biology, Amherst College, Amherst, MA 01002 USA (e-mail: setravers{at}amherst.edu
).
LITERATURE CITED
Abbo, S., and G. Ladizinsky. 1994 Genetical aspects of hybrid embryo abortion in the genus Lens L. Heredity 72: 193–200.[ISI]
Azpiroz, R., Y. Wu, J. C. Locascio, and K. A. Feldmann. 1998 An Arabidopsis brassinosteroid-dependent mutant is blocked in cell elongation. Plant Cell 10: 219–230.
Baker, A. M., and J. S. Shore. 1995 Pollen competition in Turnera ulmifolia (Turneraceae). American Journal of Botany 82: 717–725.[CrossRef][ISI]
Bawa, K. S., and C. J. Webb. 1984 Flower, fruit, and seed abortion in tropical forest trees: implications for the evolution of paternal and maternal reproductive patterns. American Journal of Botany 71: 736–751.[CrossRef][ISI]
Bertin, R. I. 1990 Paternal success following mixed pollinations of Campsis radicans. American Midland Naturalist 124: 153–163.
Bowman, R. N. 1987 Cryptic self-incompatibility and the breeding system of Clarkia unguiculata (Onagraceae). American Journal of Botany 74: 471–476.[CrossRef][ISI]
Carney, S. E., M. B. Cruzan, and M. L. Arnold. 1994 Reproductive interactions between hybridizing irises: Analyses of pollen-tube growth and fertilization success. American Journal of Botany 81: 1169–1175.[CrossRef][ISI]
———, S. A. Hodges, and M. L. Arnold. 1996 Effects of differential pollen-tube growth on hybridization in the Louisiana irises. Evolution 47: 1432–1445.
Cave, M. S., and S. W. Brown. 1954 The detection and nature of dominant lethals in Lilium. II. Cytological abnormalities in ovules after pollen irradiation. American Journal of Botany 41: 469–483.[CrossRef][ISI]
———, and ———. 1957 The detection and nature of dominant lethals in Lilium. III. Rates of early embryogeny in normal and lethal ovules. American Journal of Botany 44: 1–8.[CrossRef][ISI]
Cruzan, M. B. 1990 Variation in pollen size, fertilization ability and postfertilization siring ability in Erythronium grandiflorum. Evolution 44: 843–856.
———, and S. C. H. Barrett. 1996 Post-pollination mechanisms influencing mating patterns and fecundity: an example from Eichhornia paniculata. American Naturalist 147: 576–598.
Devries, H. 1924 Preferential fertilization in Oenothera lamarckiana. Botanical Gazette 77: 73–79.
Epperson, B. K., and M. T. Clegg. 1987 First-pollination primacy and pollen selection in the morning glory Ipomoea purpurea. Heredity 58: 5–14.
Gadish, I., and D. Zamir. 1987 Differential zygotic abortion in an interspecific Lycopersicon cross. Genome 29: 156–159.
Gottlieb, L. D. 1973 Enzyme differentiation and phylogeny in Clarkia franciscana, C. rubicunda and C. amoena. Evolution 27: 205–214.
———. 1977 Evidence for duplication and divergence of the structural gene for phosphoglucose isomerase in diploid species of Clarkia. Genetics 86: 289–307.
———. 1981 Gene numbers in species of Asteraceae that have different chromosome numbers. Proceedings of the National Academy of Sciences, USA 78: 3726–3729.
———, and V. S. Ford. 1996 Phylogenetic relationships among the sections of Clarkia (Onagraceae) inferred from the nucleotide sequences of PgiC. Systematic Botany 21: 45–62.
———, and L. C. Greve. 1981 Biochemical properties of duplicated isozymes of Phosphoglucose Isomerase in the plant Clarkia xantiana. Biochemical Genetics 19: 155–172.
Hebert, P. D., and M. J. Beaton. 1993 Methodologies for allozyme analysis using cellulose acetate electrophoresis. Helena Laboratories, Beaumont, Texas, USA.
Heslop-Harrison, J. 1987 Pollen germination and pollen-tube growth. In K. L. Giles and J. Prakash [eds.]. International Review of Cytology 107: 1–78.
Holtsford, T. P., and N. C. Ellstrand. 1992 Genetic and environmental variation in floral traits affecting outcrossing rate in Clarkia tembloriensis (Onagraceae). Evolution 46: 216–225.[CrossRef][ISI]
Johnston, M. O. 1993 Tests of two hypotheses concerning pollen competition in a self-compatible, long-styled species (Lobelia cardinalis: Lobeliaceae). American Journal of Botany 80: 1400–1406.[CrossRef][ISI]
Jones, K. N. 1994 Nonrandom mating in Clarkia gracilis (Onagraceae): a case of cryptic self-incompatibility. American Journal of Botany 81: 195–198.
Jones, T. W. A. 1985 The kinetics and thermal stability of phosphoglucoisomerase isozymes of ryegrass (Lolium spp). Physiologia Plantarum 63: 365–369.[CrossRef]
Kruckeberg, A. L., H. E. Neuhaus, R. Feil, L. D. Gottlieb, and M. Stitt. 1989 Decreased-activity mutants of phosphoglucose isomerase in the cytosol and chloroplast of Clarkia xantiana impact on mass-action ratios and fluxes to sucrose and starch and estimation of flux control coefficients and elasticity coefficients. Journal of Biochemistry 261: 457–468.
Lewis, H., and M. E. Lewis. 1955 The genus Clarkia. University of California Publications 20: 241–392.
Lyons, E. E., N. M. Waser, M. V. Price, J. Antonovics, and A. F. Motten. 1989 Sources of variation in plant reproductive success and implications for concepts of sexual selection. American Naturalist 134: 409–433.[CrossRef][ISI]
MacSwain, J. W., P. H. Raven, and R. W. Thorp. 1973 Comparative behavior of bees and Onagraceae. IV. Clarkia bees of the western United States. University of California Press, Berkeley, California, USA.
Manasse, R. S., and K. Pinney. 1991 Limits to reproductive success in a partially self-incompatible herb: fecundity depression at serial life-cycle stages. Evolution 45: 712–720.[CrossRef][ISI]
Manicacci, D., and S. C. H. Barrett. 1996 Fertility differences among floral morphs following selfing in tristylous Eichhornia paniculata (Pontederiaceae): inbreeding depression or partial incompatibility? American Journal of Botany 83: 594–603.[CrossRef][ISI]
Marshall, D. L. 1990 Non-random mating in a wild radish Raphanus sativus. Plant Species Biology 5: 143–156.
———. 1991 Non-random mating in wild radish: variation in pollen donor success and effects of multiple paternity among one to six-donor pollinations. American Journal of Botany 78: 1404–1418.[CrossRef][ISI]
———, and N. C. Ellstrand. 1986 Sexual selection in Raphanus sativus: experimental data on nonrandom fertilization, maternal choice, and consequences of multiple paternity. American Naturalist 127: 446–461.[CrossRef][ISI]
———, and ———. 1988 Effective mate choice in wild radish evidence for selective seed abortion and its mechanisms. American Naturalist 131: 739–756.[CrossRef][ISI]
Marshall, D. L., and M. W. Folsom. 1991 Mate choice in plants: an anatomical to population perspective. Annual Review of Ecology and Systematics 22: 37–64.
Mazer, S. J. 1987 Parental effects on seed development and seed yield in Raphanus raphanistrum implications for natural and sexual selection. Evolution 41: 355–371.[CrossRef][ISI]
Montalvo, A. M. 1992 Relative success of self and outcross pollen comparing mixed- and single-donor pollinations in Aquilegia caerulea. Evolution 46: 1181–1198.
Moore, D. M., and H. Lewis. 1965 The evolution of self-pollination in Clarkia xantiana. Evolution 9: 104–114.
Nakamura, R. R., and M. L. Stanton. 1987 Cryptic seed abortion and the estimation of ovule fertilization. Canadian Journal of Botany 65: 2463–2465.
———, and N. C. Wheeler. 1992 Pollen competition and paternal success in Douglas fir. Evolution 46: 846–851.[CrossRef][ISI]
Pfahler, P. L. 1967 Fertilization ability of maize pollen grains. II. Pollen genotype female sporophyte and pollen storage interactions. Genetics 57: 513–521.
Quesada, M., C. D. Schlichting, J. A. Winsor, and A. G. Stephenson. 1991 Effects of genotype on pollen performance in Cucurbita pepo. Sex Plant Reproduction 4: 208–214.
———, J. A. Winsor, and A. G. Stephenson. 1993 Effects of pollen competition on progeny performance in a heterozygous cucurbit. American Naturalist 142: 254–266.
Radha, M. R., R. Vasudeva, S. G. Hegde, K. N. Ganeshaiah, and R. U. Shaanker. 1993 Components of male gametophytic competition in Vigna unguiculata 1. walp. Evolutionary Trends in Plants 7: 29–36.[ISI]
Riddoch, B. J. 1993 The adaptive significance of electrophoretic mobility in phosphoglucose isomerase (PGI). Biological Journal of the Linnaen Society 50: 1–17.
Rigney, L. P., J. D. Thomson, M. B. Cruzan, and J. Brunet. 1993 Differential success of pollen donors in a self compatible lily. Evolution 47: 915–924.[CrossRef][ISI]
Rocha, O. J., and A. G. Stephenson. 1990 Effect of ovule position on seed production, seed weight, and progeny performance in Phaseolus coccineus 1. Leguminosae. American Journal of Botany 77: 1320–1329.[CrossRef][ISI]
———, and ———. 1991 Order of fertilization within the ovary in Phaseolus coccineus 1. Leguminosae. Sexual Plant Reproduction 4: 126–131.[ISI]
Sari-Gorla, M., E. Ottaviano, and D. Faini. 1975 Genetic variability of gametophyte growth rate in Maize. Theoretical and Applied Genetics 46: 289–294.[ISI]
SAS. 1994 JMP statistics and graphics manual. SAS Institute, Cary, North Carolina, USA.
Schiefelbein, J., M. Galway, J. Masucci, and S. Ford. 1993 Pollen tube and root-hair tip growth is disrupted in a mutant of Arabidopsis thaliana. Plant Physiology 103: 979–985.[Abstract]
Schuster, W. S., D. L. Alles, and J. B. Mitton. 1989 Gene flow in Limber Pine: Evidence from pollination phenology and genetic differentiation along an elevational transect. American Journal of Botany 76: 1395–1403.[CrossRef][ISI]
Shea, K. L. 1990 Genetic variation between and within populations of Engelmann spruce and subalpine fir. Genome 33: 1–8.
Shumaker, K. M., and G. R. Babble. 1980 Patterns of allozymic similarity in ecologically central and marginal populations of Hordeum jubatum in Utah. Evolution 34: 110–116.[CrossRef][ISI]
Snow, A. A. 1994 Postpollination selection and male fitness in plants. American Naturalist 144: S69–S83.[CrossRef][ISI]
———, and S. J. Mazer. 1988 Gametophytic selection in Raphanus raphanistrum a test for heritable variation in pollen competitive ability. Evolution 42: 1065–1075.[CrossRef][ISI]
Snow, A. A., and T. P. Spira. 1991a Differential pollen-tube growth rates and nonrandom fertilization in Hibiscus moscheutos Malvaceae. American Journal of Botany 78: 1419–1426.[CrossRef][ISI]
———, and ———. 1991b Pollen vigor and the potential for sexual selection in plants. Nature 352: 796–797.[CrossRef]
Sork, V. L., and D. W. Schemske. 1992 Fitness consequences of mixed-donor pollen loads in the annual legume Chamaecrista fasciculata. American Journal of Botany 79: 508–515.
Stephenson, A. G., and R. I. Bertin. 1983 Male competition, female choice and sexual selection in plants. In L. Real [ed.], Pollination biology. Academic Press, Orlando, Florida, USA.
———, T. Lau, M. Quesada, and J. A. Winsor. 1992 Factors that affect pollen performance. In R. Wyatt [ed.], Ecology and evolution of plant reproduction. Chapman and Hall, New York, New York, USA.
Vasek, F. C. 1965 Outcrossing in natural populations. II. Clarkia unguiculata. Evolution 19: 152–156.
Watt, W. B. 1977 Adaptation at specific loci. I. Natural selection on Phosphoglucose isomerase of Colias butterflies: biochemical and population aspects. Genetics 87: 177–194.
———. 1983 Adaptation at specific loci. II. Demographic and biochemical elements in the maintenance of the Colias PGI polymorphism. Genetics 103: 691–724.
Westoby, M., and B. Rice. 1982 Evolution of the seed plants and inclusive fitness of plant tissues. Evolution 36: 713–724.[CrossRef][ISI]
Willson, M. F. 1990 Sexual selection in plants and animals. Trends in Ecology and Evolution 5: 210–214.
———, and N. Burley. 1983 Mate choice in plants: tactics, mechanisms and consequences. Princeton University Press, Princeton, New Jersey, USA.
Zar, J. 1984 Biostatistical analysis, 2nd ed. Prentice-Hall, Englewood Cliffs, New Jersey, USA.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
x plant) were homozygous for the B allele (first five in the series), and even-numbered recipients were homozygous for the C allele (second five in the series)
