| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
0 Department of Biology, University of New Mexico, Albuquerque, New Mexico 87131 USA
Received for publication October 22, 1999. Accepted for publication February 1, 2000.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: Brassicaceae • pollen competition • Raphanus sativus • seed paternity • sexual selection • wild radish
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). If this force has been equally powerful in plants, our explanations of the evolution of plant reproductive characters must consider sexual selection. There is also interest in sexual selection in plants because the opportunity for differential pollen donor success could be present whenever there are more pollen grains available than are necessary to fertilize all of the ovules. However, despite this interest (e.g., Mulcahy, 1974, 1979
; Charnov, 1979
; Willson, 1979, 1982, 1990
; Stephenson and Bertin, 1983
; Willson and Burley, 1983
; Queller, 1983, 1987
; Snow, 1994
), the possibility of sexual selection in plants remains controversial (e.g., Charlesworth, Schemske, and Sork, 1987
; Lyons et al., 1989
; Grant, 1995
). One important source of controversy is concern over whether tests of differential pollen donor performance in the greenhouse can translate to the field.
Comparisons of field to experimental conditions are particularly essential to understanding the most frequently studied mechanism of sexual selection in plants, pollen competition. We might expect that the amount and importance of pollen competition will vary with pollen load size (Mitchell, 1997
). At very low levels of pollen availability there may be no pollen competition because virtually all pollen grains will fertilize ovules. Once a threshold of pollen grain number is reached and not all pollen grains can fertilize ovules, pollen competition (and mate choice) can become important. However, at very high pollen loads, the amount of pollen competition might not increase further if the stigma is already saturated with pollen (Mitchell, 1997
). Since the size of pollen loads on stigmas might vary considerably in the field, the opportunity for pollen competition might vary as well (e.g., Snow, 1986
).
In fact, attempting to affect the amount of pollen competition by varying pollen load size has been a critical feature of many studies that sought to test the connection between pollen competition and plant fitness. These studies assume that increasing pollen load size will increase pollen competition and that more intense pollen competition will produce more vigorous progeny. The latter assumption is reasonable because of the high level of similarity of gene expression in pollen tubes and growing seedlings (e.g., Willig, Basche, and Mascarenhas, 1988
; Ottaviano and Mulcahy, 1989
; Mascarenhas, 1990
; Hormaza and Herrero, 1992
). However, when pollen competition has been altered by variation in pollen density, effects on offspring vigor have been mixed (e.g., Ter-Avanesian, 1978a, b
; Mulcahy and Mulcahy, 1975
; Ottaviano, Sari-Gorla, and Arenair, 1983
; Lee and Hartgerink, 1986
; Stephenson, Winsor, and Schlichting, 1986
; Winsor, Davis, and Stephenson, 1987
; Davis, Stephenson, and Winsor, 1987
; Ottaviano, Sari-Gorla, and Villa, 1988; Ramstetter and Mulcahy, 1988
; Bertin, 1990a; Snow, 1990, 1991
; Richardson and Stephenson, 1992
; Quesada, Winsor, and Stephenson, 1993, 1996a, b
; Palmer and Zimmerman, 1994
; Bjorkman, 1995
; Johannsson and Stephenson, 1997
; Mitchell, 1997
; Niesenbaum, 1999
).
Results may vary for several reasons. First, selection during mating need not directly affect offspring vigor. Finding that varying the pollen competitive regime does not alter resulting progeny vigor might simply mean that the genes affecting mating behaviors, such as pollen tube growth and ability to fertilize seeds, are not strongly correlated with the genes affecting seedling growth. Second, interpretation of progeny growth may be complicated by maternal effects since ovules that are fertilized early, by fast-growing pollen tubes, may be differentially provisioned by the maternal plant (Charlesworth, 1988
; Delph, Weining, and Sullivan, 1998
). Third, in some cases the pollination treatments used might not have provided sufficient variation in the pollen competitive regime (e.g., Snow, 1990
; Mitchell, 1997
).
An alternative to varying pollen load size and examining progeny growth is to use genetic markers to detect the paternity of seeds. In these studies, mixtures of pollen are applied to stigmas and the numbers of seeds sired by various pollen donors are counted. This allows direct detection of differences in seed siring ability and studies of this sort have frequently revealed variation among pollen donors in their performance across an array of females (Marshall and Ellstrand, 1986, 1988
; Bertin, 1990b; Cruzan, 1990
; Young and Stanton, 1990a
; Marshall, 1991, 1998
; Rigney et al., 1993
; Snow and Spira, 1991, 1996
; Bjorkman, Samimy, and Pearson, 1995
; Mitchell and Marshall, 1998
). However, these studies also have limitations. Depending on the number of genetic markers available and the cost of scoring them, the number of pollen donors per cross can be quite limited. Often a non-random set of pollen donors must be chosen to facilitate use of genetic markers. And, in many of these studies, including our own (e.g., Marshall and Ellstrand, 1986, 1988
; Marshall, 1991, 1998
), the pollen loads used were fairly large. This ensures that pollen competition is possible, but does not allow us to make broad generalizations about the significance of pollen competition in the field.
One solution to the problems presented by the two sorts of methods is to combine them into single experiments. For example, it would be very useful to know the range of pollen load sizes over which differential pollen donor performance occurs. However, empirical studies of the effects of both size and composition of pollen loads on seed paternity are surprisingly few.
There is some evidence that variation in pollen donor success can occur at a range of pollen load sizes. Altering pollen availability by diluting pollen of Douglas-fir had no effect on the proportion of seeds sired by different donors, suggesting that nonrandom mating was consistent across the range of pollen dilutions used (Nakamura and Wheeler, 1992
). And, attrition of pollen tubes occurred at a similar rate across several pollen load sizes in Prunus avium (Hormaza and Herrero, 1996
). This suggests that any differential pollen tube growth was similar across the pollen load sizes used. In contrast, comparison of success of pollen from different morphs of Eichhornia paniculata showed that success of pollen from the appropriate morph varied with pollen load size for two of three morphs (Cruzan and Barrett, 1996
). While this is a mechanism that favors outcrossing rather than a form of sexual selection, it is a case where pollen donor performance varies with pollen load size. Clearly, further studies that vary both pollen load size and composition are needed to document the range of conditions under which sexual selection might occur in plants.
Our previous studies of Raphanus sativus have shown that pollen donors sire unequal numbers of seeds when large, mixed pollen loads are applied to stigmas (Marshall and Ellstrand, 1986, 1988
; Marshall, 1991, 1998
; Marshall and Fuller, 1994
) and that differential mating success is consistent across maternal plants (Marshall, 1998
). Most of our experiments were done in the greenhouse with large pollen loads and our single experimental garden study also used large pollen loads (Marshall and Fuller, 1994
). Thus, it has become increasingly critical to ask whether the nonrandom mating we have observed can occur over a wide range of pollen densities.
Here, we report on an experiment with the California wild radish in which we varied both the size and composition of pollen loads and addressed four questions. The first two were methodological prerequisites. (1) Were we able to produce pollen loads of significantly different sizes? (2) Were the smallest pollen loads small enough to limit competition? The final two questions concern the impact of variation in pollen load size. (3) Did nonrandom mating success among pollen donors occur over a range of pollen load sizes and compositions? (4) Did the range of pollen load sizes and compositions over which differential pollen donor success occurred include the kinds of pollen loads that are likely to occur in the field?
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
; Stanton, 1987a, b
; Stanton and Preston, 1988
; D. Marshall, personal observation). This variety in pollinator service likely leads to a range of pollen load sizes in the field.
Study of the wild radish mating system is facilitated by several characteristics. Plants have sporophytic self-incompatibility (Sampson, 1957
; Hinata and Nishio, 1980
; Pundir, Abbas, and Al-Attar, 1983
). Thus, self-fertilizations do not complicate studies of seed paternity. Fruits are multiseeded, making multiple paternity of fruits possible. Flowers are numerous so that crosses can be repeated on each plant sufficient times to ensure accuracy. Seed mass, 
10 mg, is large enough that individual seeds can be weighed precisely and genotyped without germination. Wild radish also has high levels of isozyme polymorphism, making paternity analysis of seeds possible (Ellstrand, 1984
; Ellstrand and Marshall, 1985a, b, 1986
; Devlin and Ellstrand, 1990
).
Results from our greenhouse study can be compared to field populations because field plants typically have the opportunity for differential pollen donor success. Most field-collected fruits have seeds sired by several donors (Ellstrand, 1984
; Ellstrand and Marshall, 1986
; Ellstrand, Devlin, and Marshall, 1989
), and this multiple paternity of fruits probably results from deposition of mixed pollen loads through pollen carryover rather than from sequential deposition of pollen from several donors (Marshall and Ellstrand, 1985
). Field studies also indicate that, while prepollination characters such as flower size affect mating, final seed paternity cannot be explained solely by prepollination events (Stanton et al., 1992
).
Experimental design
In order to select a set of plants for which the experiment was possible, we started by planting seeds from 100 maternal families. Of these 96 families were from fruits collected in the field near Riverside, California, USA. The other four families were from greenhouse crosses that were know to carry a rare allele of PGI (phosphoglucose isomerase). We started with three radish seeds per 21-cm pot. Seeds were pretreated for 1 h with a 40 mg/mL solution of giberellic acid and planted into a soil mixture consisting of 4:1:1 sand:peatmoss:perlite. Throughout the rest of the experiment plants were watered 1–3 times daily and fertilized twice weekly with Peter's 20:20:20 fertilizer and once weekly with micronutrients.
Plants were next screened in several ways to allow selection of a set of maternal plants and pollen donors for continued use. As soon as plants had more than one true leaf, tissue was collected and the PGI phenotype of each plant was scored by starch gel electrophoresis (Ellstrand, 1984
). All plants that were homozygous for any allele of PGI were retained while all heterozygotes were discarded. Where more than one homozygote was available per pot, plants were transplanted so that each pot contained only one plant. Next, three flowers per plant were scored for pollen viability using Alexander's stain (Alexander, 1980
), and a group of potential pollen donors with at least 95% pollen viability was selected for compatibility testing. We next chose several potential maternal plants for their size and general health and tested potential pollen donors for compatibility. Crosses were considered compatible if pollen from the donor could sire multiseeded fruit on the target maternal plant. Based on consideration of PGI phenotype, pollen viability, compatibility, and general health, five maternal plants and four pairs of pollen donors (a total of eight pollen donors) were selected. These plants all came from different maternal lineages, all of the pollen donors were compatible with all of the maternal plants, and within each pair of pollen donors, the two donors had different, homozygous PGI phenotypes. The remaining plants were discarded.
Each of the pollen donor pairs was used in the following crosses on each maternal plant: three pollen load sizes (low, medium, and high) and three ratios of pollen types (4:1, 1:1, and 1:4). Varying pollen load size should alter the opportunity for pollen competition and varying ratio of pollen types provides a wider range of competitive situations. Each cross was replicated five times for a total of 900 crosses (5 maternal plants x 4 pollen donor pairs x 3 load sizes x 3 ratios x 5 replicates). Note that all pollinations used mixed pollen loads. The ratios of pollen types were made by altering the number of flowers from which pollen was collected. For 1:1 mixes, we used two flowers from each donor and for 4:1 and 1:4 mixes we used four flowers from one donor and one flower from the other donor in a pair. Because these mixtures are based on flower number, which it was possible to count for each mixture, and not on actual pollen grain number, which it was not possible to count for each mixture, differences in pollen grain production per flower altered the actual ratios of pollen grains in each pollen load.
Pollen loads were applied using metal applicators of three different sizes: a number 0 insect pin for small loads, a piece of 22 gauge copper wire for medium loads, and a piece of baling wire (14.5 gauge) for large loads. Pollen was collected into small petri dishes and mixed. The pollen applicator was brushed through the pollen and tapped on the side of the petri dish. The tip of the applicator was then touched once to the stigma of a recipient flower. Recipient flowers were emasculated prior to pollination. Considerable practice resulted in very accurate pollen load sizes and, at several points in the experiment, additional test flowers were pollinated and scored for pollen load size to maintain accuracy. Pollen load sizes on stigmas were counted by collecting stigmas, staining with Alexander's stain (Alexander, 1980
), and scoring under a compound microscope. Counts of pollen grains on unpollinated stigmas of emasculated flowers (mean = 22 grains/stigma) were used to correct the pollen load sizes.
To be able to consider differential pollen production in some of our analyses, we counted and measured the number of pollen grains per flower. Five nearly open buds were collected from each pollen donor on one date near the end of the experiment. These buds were dried in microcentrifuge tubes, pollen was loosened by sonicating in a small amount of ethanol and Tween 80TM (a surfactant), the solution was diluted to 20 mL with saline solution, and four subsamples from each flower were counted using our Elzone model 280pc (Particle Data, Inc., Elmhurst, Illinois, USA) particle counter. The four subsamples were averaged for analysis. In retrospect, it may have been preferable to collect the buds for pollen counts at intervals during the experiment. However, the crosses were completed in 37 d, during which time the plants were in excellent condition. The pollen donors were well watered and fertilized and did not senesce during this period. And, in a different experiment, pollen was counted for 300 plants on two dates, 33 d apart. Pollen grain production per flower on those two dates was not significantly different (P < 0.24).
Once fruits matured, they were collected and opened to score the following data: number of seeds per fruit, number of aborted seeds per fruit (seeds that were very small, dark brown and wrinkled), and mass of each seed (to the nearest 0.01 mg). Then the PGI phenotype of each seed was scored by starch gel electrophoresis in order to ascertain the paternity of each seed from each hand pollination; 4938 seeds from 848 fruits were successfully weighed and scored for paternity.
Data analysis
To address whether we had produced significant variation in pollen load size, we performed an analysis of variance on the number of pollen grains per stigma in test pollinations, performed throughout the experiment. Pollen grains/stigma was the dependent variable and pollen load size the independent variable.
To discover whether we had produced a useful range of pollen load sizes, we asked whether individual seed size, number of good seeds per fruit (defined as total seed number per fruit minus the number of aborted seeds), and total seed mass per fruit varied with pollen load size. Total seed mass per fruit was calculated as mean seed mass x seeds/fruit. We used number of good seeds per fruit rather than total number of seeds per fruit to control for the possibility that some of the aborted seeds might not have been fertilized. However, whether we used good seeds per fruit or total seeds per fruit did not affect the results of the analyses described below.
The three measures of female reproductive success were the dependent variables in analyses of variance where maternal plant, pollen donor pair, pollen load size, ratio of the two pollen types in each pollination, and their interactions were the independent variables. Since maternal plants and pollen donors were selected for particular characteristics, all variables were treated as fixed Since we tested three, related variables, significance levels were adjusted with a sequential Bonferroni correction (Rice, 1989
). If we were able to produce pollen loads small enough that not all ovules were fertilized, we expected to find that the number of good seeds per fruit decreased significantly with pollen load size. Compensation in seed mass might result in an opposite effect on individual seed mass. We also did a separate one-way ANOVA for the effect of pollen load size on the number of good seeds per fruit for each of the five maternal plants. In this case, significance levels for testing our hypothesis were adjusted for the five plants used with a sequential Bonferroni correction (Rice, 1989
).
We used ANOVAs to assess whether pollen load size and ratio of pollen types affected seed paternity. We compared seed paternity among categories by examining the number of seeds per fruit sired by each pollen donor in a pair. Because seed paternity was scored for each seed, our initial data set had one line of data for each seed. However, crosses were done on flowers rather than seeds, so we wanted to analyze the results of crosses at the same level (the mature fruit). Therefore, we calculated the number of seeds sired in each fruit pollinated for the each of two pollen donors in a pair. This produced a data set that contained two lines of data for each fruit, one line for the number of seeds sired by each donor. Number of seeds sired per fruit by pollen donors was then used as the dependent variable in analyses of variance where maternal plant, pollen donor, pollen load size, ratio of pollen types and their interactions were used as independent variables. As described above, all variables were treated as fixed. These analyses were performed separately for each pair of pollen donors. For this reason, significance levels were corrected for the four tests of each hypothesis by a sequential Bonferroni correction (Rice, 1989
). We used the fruit as the level of analysis since each cross produced one fruit.
Finally, we addressed whether pollen production per flower had an impact on seed paternity. First, we used ANOVAs where pollen donor was the independent variable and pollen grains per flower was the dependent variable to ask whether the mean number of pollen grains per flower differed among the pollen donors in each pair. Next, we compared observed seed paternity to expected seed paternity as calculated from pollen production. Thus, for the two donors in a pair, we calculated the proportion of pollen that each would contribute to a pollen mix based on their average pollen production per flower. These proportions were used to calculate expected seed paternity. Expected seed paternity was compared to observed seed paternity by means of chi-square tests.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
These pollen load sizes fell well within the range observed in a field populations near Davis, California. In those populations, mean pollen load size was 112 ± 102 grains per flower (Young and Stanton, 1990b
) and stigmas were quickly saturated with the 90 grains that were considered necessary for pollen competition to occur (Ashman, Galloway, and Stanton, 1993
).
Pollen load size vs. seed set per fruit
The pollen load sizes we achieved appear to have been in a meaningful range since mean number of good seeds per fruit differed significantly among pollen load sizes (Tables 1 and 2). The pattern of means is as expected if the lowest pollen loads were too small for complete seed set since the number of good seeds per fruit was less for the low pollen loads than for the medium or high pollen loads.
|
|
The differences in the number of good seeds per fruit among maternal plants have consequences for the interpretation of the experiment since plants that set fewer seeds per fruit, perhaps due to reduced ovule number, may need fewer pollen grains for full seed set. Because of this possibility, we also examined the effect of pollen load size on seed number per fruit for each of the five maternal plants. In dividing the total sample by five, we reduced our power to detect significant effects of pollen load size. Nonetheless, low pollen loads resulted in significantly reduced seed set per fruit on two of the five plants considered individually (Table 3). These two plants, A and E, were the maternal plants with the greatest number of seeds per fruit. Two of the other plants, B and C, showed trends in the expected direction, seed set tended to be lower for the low pollen loads. However, for plant D, the plant with the lowest mean seed set per fruit, there was no tendency for low pollen loads to produce fewer seeds per fruit. Thus, our lowest pollen load was adequate to reduce seed set on most of the plants, but not on the plant with the fewest seeds per fruit.
|
|
|
|
|
Pollen production
Although average pollen grain production per flower varied from a high of 110 000 grains per flower for donor H to a low of 58 390 grains per flower for donor E, there was a significant difference in pollen grains per flower among the donors of a pair only for pair 3 (Table 5). However, the sample size for testing differences in pollen production was modest (five flowers per plant) and we did not want to discount potential mechanisms for differences in pollen donor performance. Therefore, we considered whether adjusting expected seed paternity for pollen production per flower explained differences in seed siring ability among pollen donors for all four pairs of donors (Table 6). For pairs 1 and 2, seed paternity differed among pollen donors even when pollen production was considered. However, for pairs 3 and 4, differences in pollen grain production per flower appear to be the source of the differences in number of seeds sired since observed and expected seed paternity are not significantly different when pollen production is taken into account.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
To test whether small pollen load sizes might reduce differences in pollen donor performance, it was first necessary to produce pollen loads of significantly different sizes. We were able to produce pollen loads ranging from 40 to 258 grains per stigma using a variety of pollen applicators. Second, it was essential to produce small pollen loads that were sufficiently small to reduce the potential for pollen competition (Mitchell, 1997
). Our data on seed set per fruit indicate that it is likely we accomplished this goal. Looking across all five maternal plants, the smallest pollen loads produced fruits with significantly fewer seeds. However, the smallest pollen loads may not have been quite small enough to prevent microgametophye competition in plants with low mean seed set per fruit. The maternal plant with the lowest mean seed number per fruit did not have a reduced seed set at the small pollen load size. Thus, we likely produced diminished pollen competition on most, but not all, of our maternal plants. Since the inclusion of plant D, the plant that did not have any reduction in seed set at the low pollen load, might have biased our results, we removed the data for plant D and repeated the analyses of seed paternity. The results of the new analysis were the same as in the analysis of the full data set.
Our examination of seed paternity indicated that differential seed siring success among pollen donors occurred over the entire range of pollen load sizes and compositions that we studied. We expected to find that ratio of pollen types within a pollen load would alter seed paternity as we had found the same result in our earliest work with wild radish (Marshall and Ellstrand, 1986
). Yet, even though ratio of pollen types had a strong effect on seed paternity, it did not mask the effect of pollen donor identity.
The observation that the difference in ability to sire seeds between competing pollen donors was unaffected by pollen load size is new and somewhat surprising given the range of pollen load sizes that we used. This suggests that nonrandom seed paternity can occur over a wide range of conditions.
The mechanism by which differential seed paternity was produced appears to be clear in two cases. For two of the pairs of pollen donors, differences in pollen grain production per flower appear to be the strongest effect on seed paternity. However, for the other two pollen donors, aspects of pollen quality or the interaction between growing pollen tubes and stylar tissue must have effects on seed paternity. This confirms our previous observations that pollen production per flower can be an important pollen donor characteristic, but that it is not the only difference among pollen donors that affects their ability to sire seeds (Marshall, 1991, 1998
). In fact, it seems remarkable that pollen donor E of pair 3 sired seeds at all under some of the conditions. Given its low pollen production relative to donor F, in the crosses where the ratio of flowers used was 1:4 for E:F and the pollen load was small (about 40 grains per stigma), donor E would have had, on average, only five pollen grains in a pollen load, yet it still sired some seeds. The effect of differential pollen production in the field would depend on the correlation between number of pollen grains produced and number of pollen grains delivered to compatible stigmas.
Although the difference in pollen production per flower among donors was an unplanned aspect of the experiment, it does not alter our interpretation of the consequences of variation in pollen load size. The effect of the variation in pollen production was to alter the actual ratios of pollen applied from that estimated by the number of flowers used per pollen donor per cross. However, the pollen load sizes were not affected and for every pair there were pollen loads in which each donor had the most or least abundant pollen. Our primary conclusion about the effect of pollen load size on seed paternity is unaffected. Neither pollen load size by ratio nor pollen donor by pollen load size by ratio effects on seed paternity approached significance. The F values were uniformly low across the pollen donor pairs (Table 4). And, the effect of pollen load size on seed paternity was small, nonsignificant, and showed no intriguing trend in all four pollen pollen donor pairs whether or not differential pollen production was an important source of variation in seed paternity.
Since seed paternity was unequal at pollen load sizes as low as 40 pollen grains per stigma, it seems likely that pollen loads in the field would often be sufficient to produce differential pollen donor success. In fact mean pollen load size for populations near Davis, California (112 ± 102 grains per flower) was considerably larger than our smallest pollen load size. Here, we show that nonrandom seed paternity occurs across a range of pollen load sizes and that, in fact, differential pollen donor success occurs well below the reported mean pollen load size in the field. Of course, our seeds came from a different location in California, but the types of pollinators that visit plants are quite similar (D. Marshall, unpublished data).
Our observation of nonrandom seed paternity across a range of pollen load sizes may help to explain the disparate results seen in previous experiments. Earlier work with the closely related R. raphanistrum showed that progeny of pollen loads that were small (60 grains per stigma) and large (300 grains per stigma) did not differ in early growth characters (Snow, 1990
). These results were used to suggest that pollen competition did not play an important role. However, our study suggests that seed paternity may have been nonrandom at both of those pollen load sizes.
In reviewing studies of the relationship between pollen load size and progeny growth, Mitchell (1997)
pointed out that studies which failed to find such a relationship may not have had sufficiently small pollen load sizes to eliminate pollen competition. Our results support that interpretation and suggest caution in interpreting the outcome of pollination with small pollen loads when seed paternity has not been measured.
While our observation that seed paternity is nonrandom after pollination with quite small pollen loads is novel, the results leave several important questions unanswered. First, our smallest pollen loads were not quite small enough to reduce seed set on all of the maternal plants. Thus, we are developing techniques to apply even smaller amounts of pollen. Second, the number of maternal plants and pollen donors studied was modest. We are planning larger studies. Third, measuring only seed paternity and not offspring growth does not allow us to directly compare our work to the previous studies that evaluated pollen competition by growing progeny. To bridge this gap we need to measure seed paternity and examine offspring growth. Fourth, using only two pollen donors per cross is convenient for experimental design and electrophoresis. However, we know that wild radish plants in the field can have at least four mates per fruit (Ellstrand, 1984
; Ellstrand and Marshall, 1986
). With more pollen donors, the number of pollen grains of the most competitive donor may be quite few in a small load and that donor may sire few seeds (Cruzan and Barrett, 1996
). To test for this possibility, we need to use some small pollen loads with a larger number of donors. Finally, while we have discussed our experiment in terms of altering the opportunity for microgametophyte competition, changing pollen load size alters the opportunity for both mate choice and microgametophyte competition. Future experiments will address these gaps.
Despite these limitations, our confidence that differential pollen donor success occurs across a wide range of conditions in wild radish is increasing. We have illustrated nonrandom seed paternity with many sets of plants (Marshall and Ellstrand, 1986, 1988
; Marshall, 1991, 1998
; Marshall and Fuller, 1994
), we have shown that nonrandom seed paternity can occur in field as well as greenhouse plants (Marshall and Fuller, 1994
), and we have demonstrated that seed paternity can be unequal among pollen donors in both healthy and drought stressed plants (Marshall and Ellstrand, 1988
). More recently, we found that differential seed siring can be consistent across maternal plants and is not simply an artifact of the self incompatibility system (Marshall, 1998
). Our current results suggest that differential seed paternity can occur at pollen load sizes that are frequently achieved in field populations. Thus, selection during mating can likely occur in the field.
|
FOOTNOTES |
|---|
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Andersson, M. 1994 Sexual selection. Princeton University Press, Princeton, New Jersey, USA.
Ashman, T., L. F. Galloway, and M. L. Stanton. 1993 Apparent vs. effective mating in an experimental population of Raphanus sativus. Oecologia 96: 102–107.
Bertin, R. I. 1990a Effects of pollination intensity on Campsis radicans. American Journal of Botany 77: 178–187.[CrossRef][ISI]
———. 1990b Paternal success following mixed pollinations of Campsis radicans. American Midland Naturalist 124: 153–163.
Bjorkman, T. 1995 The effect of pollen load and pollen grain competition on fertilization success and progeny performance in Fagopyrum esculentum. Euphytica 83: 47–52.
———, C. Samimy, and K. J. Pearson. 1995 Variation in pollen performance among plants of Fagopyrum esculentum. Euphytica 82: 235–240.
Charlesworth, D. 1988 A comment on evidence for pollen competition in plants and its relationship to progeny fitness. American Naturalist 132: 298–302.[CrossRef][ISI]
———, D. W. Schemske, and V. L. Sork. 1987 The evolution of plant reproductive characters: sexual vs. natural selection. In S. C. Stearns [ed.], Evolution of sex, 317–335. Birkhauser, Basel, Switzerland.
Charnov, E. L. 1979 Simultaneous hermaphroditism and sexual selection. Proceedings of the National Academy of Sciences, USA 76: 2480–2484.
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 Postpollination mechanisms influencing mating patterns and fecundity: an example from Eichhornia paniculata. American Naturalist 147: 576–598.
Davis, L. E., A. G. Stephenson, and J. A. Winsor. 1987 Pollen competition improves performance and reproductive output of the common zucchini squash under field conditions. Journal of the American Society of Horticultural Science 112: 711–716.
Delph, L. F., C. Weining, and K. Sullivan. 1998 Why fast growing pollen tubes give rise to vigorous progeny: the test of a new mechanism. Proceedings of the Royal Society of London, Series B 265: 935–939.[CrossRef]
Devlin, B., and N. C. Ellstrand. 1990 Male and female fertility variation in wild radish, a hermaphrodite. American Naturalist 136: 87–107.[CrossRef][ISI]
Ellstrand, N. C. 1984 Multiple paternity within the fruits of the wild radish, Raphanus sativus. American Naturalist 123: 819–828.
———, B. Devlin, and D. L. Marshall. 1989 Gene flow by pollen into small populations: data from experimental and natural stands of wild radish. Proceedings of the National Academy of Sciences 86: 9044–9047.
———, and D. L. Marshall. 1985a Interpopulation gene flow by pollen in wild radish, Raphanus sativus. American Naturalist 126: 606–616.
———, and ———. 1985b The impact of domestication on distribution of allozyme variation within and among cultivars of radish Raphanus sativus L. Theoretical and Applied Genetics 69: 393–398.[CrossRef][ISI]
———, and ———. 1986 Patterns of multiple paternity in populations of Raphanus sativus. Evolution 40: 837–842.
Grant, V. 1995 Sexual selection in plants: pros and cons. Proceedings of the National Academy of Sciences, USA 92: 1247–1250.
Hinata, K., and T. Nishio. 1980 Self-incompatibility in crucifers. In S. Tsunoda, K. Hinata, and C. Gomez-Campo [eds.], Brassica crops and wild allies: biology and breeding, 223–234. Japanese Scientific Societies, Tokyo, Japan.
Hormaza, J. I., and M. Herrero. 1992 Pollen selection. Theoretical and Applied Genetics 83: 663–672.[ISI]
———, and ———. 1996 Dynamics of pollen tube growth under different competition regimes. Sexual Plant Reproduction 9: 153–160.[ISI]
Johannsson, M. H., and A. G. Stephenson. 1997 Effects of pollination intensity on the vigor of the sporophytic and gametophytic generation of Cucurbita texana. Sexual Plant Reproduction 10: 236–240.
Kay, Q. O. N. 1976 Preferential pollination of yellow-flowered morphs of Raphanus raphanistrum by Pieris and Eristalis spp. Nature 261: 230–232.[CrossRef]
Lee, T. D., and A. P. Hartgerink. 1986 Pollination intensity, fruit maturation pattern, and offspring quality in Cassia fasciculata (Leguminosae). In D. L. Mulcahy, G. B. Mulcahy, and E. Ottaviano [eds.], Biotechnology and ecology of pollen, 417–422. Springer-Verlag, New York, New York, USA.
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]
Marshall, D. L. 1991 Nonrandom 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]
———. 1998 Pollen donor performance can be consistent across maternal plants in wild radish (Raphanus sativus, Brassicaceae): a necessary condition for the action of sexual selection. American Journal of Botany 85: 1389–1397.
———, and N. C. Ellstrand. 1985 Proximal causes of multiple paternity in wild radish, Raphanus sativus. American Naturalist 126: 596–605.
———, and ———. 1986 Sexual selection in Raphanus sativus: experimental data on non-random 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 mechanism. American Naturalist 131: 736–759.
———, and O. S. Fuller. 1994 Does nonrandom mating among wild radish plants occur in the field as well as in the greenhouse. American Journal of Botany 81: 439–445.[CrossRef][ISI]
Mascarenhas, J. P. 1990 Gene activity during pollen development. Annual Review of Plant Physiology and Plant Molecular Biology 41: 317–338.[CrossRef][ISI]
Mitchell, R. J. 1997 Effects of pollen quality on progeny vigor: evidence from the desert mustard Lesquerella fendleri. Evolution 51: 1679–1684.
———, and D. L. Marshall. 1998 Nonrandom mating and sexual selection in a desert mustard: an experimental approach. American Journal of Botany 85: 48–55.[Abstract]
Mulcahy, D. L. 1974 Correlation between speed of pollen tube growth and seedling height in Zea mays L. Nature 249: 491–493.[CrossRef][ISI]
———. 1979 The rise of the angiosperms: a genecological factor. Science 206: 20–23.
———, and G. B. Mulcahy. 1975 The influence of gametophytic competition on sporophytic quality of Dianthus chinensis. Theoretical and Applied Genetics 46: 277–280.
Nakamura, R. R., and N. C. Wheeler. 1992 Pollen competition and paternal success in douglas-fir. Evolution 46: 846–851.[CrossRef][ISI]
Niesenbaum, R. A. 1999 The effects of pollen load size and donor diversity on pollen performance, selective abortion, and progeny vigor in Mirabilis jalapa (Nyctaginaceae) American Journal of Botany 86: 261–268.
Ottaviano, E. M., and D. L. Mulcahy. 1989 Genetics of angiosperm pollen. Advances in Genetics 26: 1–64.
———, M. Sari-gorla. and I. Arenair. 1983 Male gametophytic competition in maize: selection and implications with regard to breeding systems. In D. L. Mulcahy and E. M. Ottaviano [eds.], Pollen: biology and implications for plant breeding, 367–374. Elsevier, New York, New York, USA.
———, ———, and M. Villa. 1988 Pollen competitive ability in maize: within population variability and response to selection. Theoretical and Applied Genetics 76: 601–608.[ISI]
Palmer, T. M., and M. Zimmerman. 1994 Pollen competition and sporophyte fitness in Brassica campestris: does intense pollen competition result in individuals with better pollen? Oikos 69: 80–86.[CrossRef][ISI]
Pundir, N. S., R. F. Abbas, and A. A. A. Al-Attar. 1983 Pollen germination and pollen tube growth in Raphanus sativus L. following self- and cross-pollinations. Phyton 43: 127–130.[ISI]
Queller, D. C. 1983 Sexual selection in a hermaphroditic plant. Nature 305: 706–707.[CrossRef]
———. 1987 Sexual selection in flowering plants. In J. W. Bradbury and M. B. Andersson [eds.], Sexual selection: testing the alternatives, 165–181. Wiley, New York, New York, USA.
Quesada, M., J. A. Winsor, and A. G. Stephenson. 1993 Effects of pollen competition on progeny performance in a heterozygous curcurbit. American Naturalist 142: 694–706.[CrossRef][ISI]
———, ———, and ———. 1996a Effects of pollen competition on the reproductive performance in cucurbit hybrids (Cucurbitaceae): F1 and backcross generations. Canadian Journal of Botany 74: 1113–1118.
———, ———, and ———. 1996b Effects of pollen selection on progeny vigor in a Cucurbita pepo x C. texana hybrid. Theoretical and Applied Genetics 92: 885–890.[CrossRef][ISI]
Ramstetter, J., and D. L. Mulcahy. 1988 Consequences of pollen competition for Aurelolaria flava seedlings. Bulletin of the Ecological Society of America 69: 269–270.
Rice, W. R. 1989 Analyzing tables of statistical tests. Evolution 43: 223–225.[CrossRef][ISI]
Richardson, T. E., and A. G. Stephenson. 1992 Effects of parentage, prior fruit set and pollen load on fruit and seed production in Campanula americana L. Evolution 46: 1731–1739.[CrossRef][ISI]
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]
Sampson, D. R. 1957 The genetics of self-incompatibility in the radish. Journal of Heredity 48: 26–29.
Snow, A. A. 1986 Pollination dynamics in Epilobium canum (Onagraceae): consequences for gametophytic selection. American Journal of Botany 73: 139–151.[CrossRef][ISI]
———. 1990 Effects of pollen-load size and number of donors on sporophyte fitness in wild radish (Raphanus raphanistrum). American Naturalist 136: 742–758.[CrossRef][ISI]
———. 1991 Effects of pollen-load size on sporophyte competitive ability in two Epilobium species. American Midland Naturalist 125: 348–355.[CrossRef][ISI]
———. 1994 Postpollination selection and male fitness in plants. American Naturalist 144: S69-S83.[CrossRef][ISI]
———, and T. P. Spira. 1991 Differential pollen-tube growth rates and nonrandom fertilization in Hibiscus moscheutos (Malvaceae). American Journal of Botany 78: 1419–1426.[CrossRef][ISI]
———, and ———. 1996 Pollen-tube competition and male fitness in Hibiscus moscheutos. Evolution 50: 1866–1870.
Stanton, M. L. 1987a The reproductive biology of petal color variants in wild populations of Raphanus sativus L. I. Pollinator response to color morphs. American Journal of Botany 74: 178–187.[CrossRef][ISI]
———. 1987b The reproductive biology of petal color variants in wild populations of Raphanus sativus L. II. Factors limiting seed production. American Journal of Botany 74: 188–196.[CrossRef][ISI]
———, T. Ashman, L. F. Galloway, and H. J. Young. 1992 Estimating male fitness of plants in wild populations. In R. Wyatt [ed.], Ecology and evolution of plant reproduction: new approaches, 62–90. Chapman and Hall, New York, New York, USA.
———, and R. E. Preston. 1988 Ecological correlates of petal size variation in wild radish, Raphanus sativus (Brassicaceae). American Journal of Botany 75: 528–539.[CrossRef][ISI]
Stephenson, A. G., and R. I. Bertin. 1983 Male competition, female choice, and sexual selection in plants. In L. Real [ed.], Pollination biology, 109–149. Academic Press, Orlando, Florida, USA.
———, J. A. Winsor, and C. S. Schlichting. 1986 Evidence for non-random fertilization in the common zucchini, Cucurbita pepo. In M. Cresti, P. Cori, and E. Pacini [eds.], Sexual reproduction in higher plants, 333–338. Springer, Berlin, Germany.
Ter-Avanesian, D. V. 1978a The effect of the number of pollen grains used in fertilization. Theoretical and Applied Genetics 52: 77–79.[ISI]
———. 1978b Significance of pollen amount for fertilization. Bulletin of the Torrey Botanical Club 105: 2–8.[CrossRef][ISI]
Willig, P. R., D. Bashe, and J. P. Mascarenhas. 1988 Analysis of the quantity and diversity of messenger RNA's from pollen and shoots of Zea mays. Theoretical and Applied Genetics 75: 75–83.
Willson, M. F. 1979 Sexual selection in plants. American Naturalist 113: 777–790.[CrossRef][ISI]
———. 1982 Sexual selection and dicliny in angiosperms. American Naturalist 119: 579–583.[CrossRef][ISI]
———. 1990 Sexual selection in plants and animals. Trends in Ecology and Evolution 5: 210–214.
———, and N. Burley. 1983 Mate choice in plants. Princeton University Press, Princeton, New Jersey, USA.
Winsor, J. A., L. E. Davis, and A. G. Stephenson. 1987 The relationship between pollen load and fruit maturation and the effect of pollen load on offspring vigor in Curcurbita pepo. American Naturalist 129: 643–656.
Young, H. J., and M. L. Stanton. 1990a Influence of environmental quality on pollen competitive ability in wild radish. Science 248: 1631–1633.
———, and ———. 1990b Influence of floral variation on pollen removal and seed production in wild radish. Ecology 71: 536–547.[CrossRef][ISI]
This article has been cited by other articles:
|
|
 |
|
  V. A. Koelling and K. Karoly Self-pollen interference is absent in wild radish (Raphanus raphanistrum, Brassicaceae), a species with sporophytic self-incompatibility Am. J. Botany, May 1, 2007; 94(5): 896 - 900. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. G. M. Shaner and D. L. Marshall How robust is nonrandom |
