| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Reproductive Biology |
2Department of Biology, University of New Mexico, Albuquerque, New Mexico 87131 USA; and 3Baker Hall Residential Academic Program, Campus Box 176, University of Colorado, Boulder, Colorado 80309-0176 USA
Received for publication March 2, 2001. Accepted for publication June 26, 2001.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: Brassicaceae • mate choice • pollen competition • Raphanus sativus • seed paternity • sexual selection • wild radish
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
; Charnov, 1979
; Willson, 1979, 1982, 1990
; Queller, 1983, 1987
; Stephenson and Bertin, 1983
; Willson and Burley, 1983
; Snow, 1994
; Delph and Havens, 1998
). However, the suggestion that sexual selection might occur under these circumstances is controversial for a number of reasons (e.g., Charlesworth, Schemske, and Sork, 1987
; Lyons et al., 1989
; Grant, 1995
). These include the possibility that mating may be nonrandom due to the degree of relatedness of the pollen donors with maternal plants, the difficulty in distinguishing pollen competition from mate choice, and the problems in applying greenhouse-based results to the more complicated situations that are likely to occur in the field (Winsor, Peretz, and Stephenson, 2000
).
Despite these difficulties, the possibility of sexual selection in plants remains worth studying. Sexual selection has been powerful in animals, producing elaborate behaviors and morphologies (e.g., Andersson, 1994
). While more controversial in plants, it is logical to suggest that pollen donors that produce more successful pollen and maternal plants that can mate with superior pollen donors will both be favored by selection. Thus, rather than dismiss this interesting possibility, our goal here is to address some of the difficulties in applying sexual selection to plants.
A fundamental prerequisite for sexual selection is that pollen donors sire unequal numbers of seeds. Several studies have shown that when pollen from two or more donors is applied to stigmas, seed paternity is, in fact, nonrandom (Bemis, 1959
; Barnes and Cleveland, 1963
; Pfahler, 1967, 1974a, b, 1975
; Pfahler and Linskens, 1972
; Levin, 1975
; Ottaviano, Sari-Gorla, and Mulcahy, 1975
; Sari-Gorla, Ottaviano, and Faini, 1975
; Currah, 1981
; Ottaviano, Sari-Gorla, and Arenair, 1983
; Sarr, Fraleigh, and Sandmeir, 1983
; Marshall and Ellstrand, 1986, 1988
; Bertin, 1990
; Cruzan, 1990
; Young and Stanton, 1990
; Marshall, 1991, 1998
; Snow and Spira, 1991, 1996
; Rigney et al., 1993
; Björkman, Samimy, and Pearson, 1995
; Mitchell and Marshall, 1998
; Pasonen et al., 1999
). However, it is often difficult to assess whether pollen competition or mate choice determines the outcome of mating in these cases and in some cases, the applicability of the results to field situations is not clear.
One reason that it is difficult to apply greenhouse results to the field is that over the course of a field season the opportunity for and context of nonrandom seed paternity may change. Resource levels, the number of previously filled fruits, and the identity of the fathers of previously filled fruits will likely all change over the season. These factors may affect the competence of the maternal tissue to distinguish among pollen donors and the amount of nonrandom mating that is possible. An example of what might occur is the following scenario: early in the season, plants have produced little seed and pollen and still have resources to do both. However, the plant has little "experience" of the level of pollen availability or the quality of available pollen donors; pollinator availability may be low. The pollen donors may also be in peak condition. Later in the season, more fruits are being filled, so resources may be less available, but plants still have sufficient resources to produce pollen and set seeds. Maternal plants have some mating "experience." By "experience" we mean that plants have received adequate pollen or not, filling fruits are vigorously drawing maternal resources or not, and fruits are variable in quality or not. At the end of the season, resources may be quite limited, the maternal plants and pollen donors are senescing and the ability of pollen donors to demonstrate mate quality and maternal plants to assess it may be limited. Time to fill additional fruits would also be limited.
There is some evidence that this model of changes over the season can apply in the field. First, it is not uncommon to find that, especially for annuals, fruit abortion rates, seeds per fruit, and/or seed size change over the season (e.g., Stephenson, 1981
; Marshall, Levin, and Fowler, 1985
; Mazer, Snow, and Stanton, 1986
; Stanton, 1987a
; Richardson and Stephenson, 1991
; Susko and Lovett-Doust, 2000
). These changes imply that resources are less available. Second, fruit filling can be influenced by the identity or number of mates that sire the seeds (Marshall and Ellstrand, 1986, 1988
; Marshall, 1991
). More relevantly, fruit filling can be influenced by the identity of neighboring fruits (Marshall and Oliveras, 1990
), suggesting that the mating history of a plant matters. Finally, the condition of maternal plants is known to affect the paternity of seeds (Marshall and Ellstrand, 1988
; Marshall and Fuller, 1994
).
Thus, it is worth exploring changes in seed paternity over time on the same plants. Variation in seed paternity as maternal condition changes would give evidence that the maternal tissue plays a role in the outcome of mating. An extreme change in seed paternity over time (e.g., seed paternity starting as nonrandom and becoming random or the rank order of pollen donor success changing) would suggest that the range of conditions under which sexual selection might occur in the field is limited. In contrast, temporal changes in seed paternity that do not alter the rank order of pollen donor performance would suggest that differences in pollen donor success persist over a range of conditions.
In the experiment reported here, we address those issues, particularly in relation to our studies of wild radish, Raphanus sativus. In previous work with this species, application of mixtures of pollen has resulted in differential seed siring success among pollen donors (Marshall and Ellstrand, 1986, 1988
; Marshall, 1991, 1998
; Marshall and Fuller, 1994
), and this difference is not due to direct or subtle effects of the incompatibility system (Marshall, 1998
). Thus, pollen competition and mate choice may occur. Changing the condition of maternal plants by reducing water availability or removing some leaves alters the proportions of seeds sired by pollen donors, but not the overall rank order of pollen donor performance (Marshall and Ellstrand, 1988
; Marshall and Fuller, 1994
; Marshall and Diggle, 2001
). Thus, it is likely that the maternal tissue plays a role in the outcome of mating, but there is still room for absolute differences in pollen donor performance.
We extend those results here by examining the relative performance of pollen donors over time on greenhouse plants that matured hundreds of fruits. In addition, we consider whether fruit and seed filling are affected by the pattern of pollinations on a branch. Thus, we address whether changes in maternal condition and changes in the recent pollination history of a maternal plant affect the outcome of subsequent pollinations.
Specifically, we first ask whether resources for reproduction appear to become more limited over time and then address the following questions: (1) Does the success of pollen donors change over time on maternal plants as measured by number of seeds sired after mixed pollinations? (2) Does the success of pollen donors change based on the recent pollination history of maternal plants (as measured by number of seeds sired after mixed pollinations)? (3) Are there overall differences in pollen donor performance that persist across changes in maternal condition?
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
), weedy annual that is common along roadsides, in ditches, and in agricultural fields in California, USA. Considerable study of the mating system (Ellstrand, 1984
; Ellstrand and Marshall, 1986
; Ellstrand, Devlin, and Marshall, 1989
) and pollination biology (e.g., Kay, 1976
; Stanton, 1987a, b
; Stanton and Preston, 1988
; Ashman, Galloway, and Stanton, 1993
) provide an extensive background for the present work. In addition, certain practical details make wild radish an attractive study organism: due to physiological self-incompatibility (Sampson, 1957; Hinata and Nishio, 1980
; Pundir, Abbas, and Al-Attar, 1983
), flowers need not be emasculated for experimental crosses; it is possible to do many crosses on each plant; seeds are large enough to weigh individually and to obtain electrophoretic genotypes; and electrophoretic systems for paternity analysis of seeds are well developed (Ellstrand, 1984
).
Experiments
We grew 125 wild radish plants from field-collected seeds in a greenhouse at the University of New Mexico, Albuquerque, New Mexico, USA. After screening for genotypes at a phosphoglucose isomerase (PGI) locus and tests for compatibility, five maternal lineages (45h, 53c, 58h, 8f, and 7m) and three pollen donors (B, E, and G) were selected for the experiment. All plants were homozygous for the PGI locus, and the pollen donors had different PGI phenotypes. The maternal plants were homozygous for the PGI alleles labeled 22, 33, or 44 and pollen donors B, E, and G were homozygous for PGI alleles 33, 22, and 44, respectively. The overlap in allele identity between maternal plants and pollen donors would be a problem if these plants were self-compatible. However, they are not. In addition to scoring plants for the PGI locus, we screened for compatibility of the pollen donors with the maternal plants, but compatibility among the maternal plants was not tested.
We planned to use a total of ten maternal plants, two siblings from each of the five lineages. This design gave us an array of maternal families and some ability to test for effects of maternal lineage. However, one of the plants in lineage, 58h, was slow to flower. To start the experiment quickly so that many crosses could be completed before the plants senesced, we assumed that this slower plant had the same compatibility relationships as its sibling. We were incorrect; the slower flowering plant was incompatible with pollen donor E and we had to exclude it from analysis. Thus, our results are based on nine maternal plants, two siblings from each of four lineages and one sibling from the fifth lineage. Given the high frequency of multiple mating in the field populations from which these seeds were collected (Ellstrand and Marshall, 1985
), these are likely half-siblings.
The three pollen donors were used to make three kinds of single pollinations (B-only, E-only, and G-only) and one mixed pollination (B + E + G) on each maternal plant. Both single and mixed pollinations were made by collecting pollen in a small petri dish, mixing with a tissue-wrapped forceps, and applying the pollen. For mixed pollinations, pollen was collected from equal numbers of flowers for each of the three pollen donors. This method results in the application of hundreds of pollen grains per stigma (M. Folsom, University of New Mexico, personal communication). At least 500 crosses were made on each of the ten maternal plants. The four kinds of crosses were performed in specific orders on each maternal plant so that both the location of crosses on the branch (near to the main stem or far from the main stem) and the paternity of the neighboring fruits varied. On average, 20 crosses were performed on each of 25 branches on each of the maternal plants and equal numbers of B, E, G, and mixed (M) crosses were made. The location of crosses along the branch presumably affected the resources available to fill the fruit, as, for example, the twentieth cross on a branch would have far more competitors than the first cross on a branch. The location of crosses on a branch also represented a temporal sequence because one to two flowers were available for crossing on a branch each day. Thus, the twentieth cross on a branch was performed 2–3 wk after the first cross. Because new flowers are produced distally on inflorescence branches, flowers more proximal to the main stem were pollinated earlier than more distal flowers.
The location and timing of crosses were used as indices of time in the fruiting season and resource availability in two ways. First, we considered the effect of the order in which crosses were performed on a branch. An examination of fruit survivorship over time (Fig. 1) showed that there was an early period of high fruit survivorship followed by a longer period of decreasing fruit survivorship. The final period showed low and fairly stable fruit survivorship. Based on examination of this figure, we divided crosses into four equal groups. We created these groups separately for each maternal plant, since each plant had a different mean number of crosses per branch. Some plants produced many flowers on fewer branches and some plants produced fewer flowers per branch on more branches. For each plant, we defined four groups of crosses along a branch. Each group of crosses (called cross group) was set as one quarter of the mean number of crosses per branch for that plant.
|
Pollinations varied both by when they were performed (indicated by the order of crosses on a branch and the order in which branches were used) and the identity of the fathers of nearby fruits. For example, fruits pollinated by donor B were surrounded by fruits sired by donor B (sequence BBB), donor E (EBE), donor G (GBG), or by mixed pollinations (MBM). Since there were four kinds of crosses that could each be surrounded by four kinds of neighboring crosses, the pollinations could be divided into 16 categories. However, in creating long sequences of crosses on branches, some crosses did not fit into these patterns, i.e., the neighboring fruits were pollinated by two different donors (e.g., GBE). These fruits were not used in analyses of neighbor effects, but were used in analyses of effects of location of crosses along branches.
The identity of the two neighboring fruits of any target fruit was categorized in two ways. First, we created a variable that indicated whether the neighboring fruits were pollinated by the same or different pollen donors. We called this variable neighbor type (e.g., for cross order BBB, neighbor type = "same" and for cross order GBG, neighbor type = "different.") Finally, we created a variable that indicated the identity of the pollen donor that was used to pollinate the two surrounding fruits. We called this variable neighbor identity.
We measured resources allocated to reproduction in several ways. Patterns of fruit set were monitored during fruit development. At maturity, all fruits were collected and seeds were counted, categorized as good or aborted, and weighed to the nearest 0.1 mg. Good seeds were round and of normal color. Aborted seeds were small, shriveled, and black.
We measured differential success of pollen donors in mixed pollinations by examining the paternity of seeds. Seed paternity for those crosses was assessed through starch gel electrophoresis of the seed tissue followed by scoring for a locus of PGI (Ellstrand, 1984
).
Analysis
Our first concern in examining the data was to decide whether to treat maternal plant and pollen donor as random or fixed effects. Because the plants were not selected at random (they were selected based on compatibility relationships and PGI phenotype), we decided to treat these variables as fixed, even though this limited our conclusions to the plants used in this study.
Our second concern was whether to represent maternal effects as the five maternal lineages or as the nine maternal plants. Because we had unequal numbers of plants for each lineage and because the behavior of plants within lineages was very different, we used individual maternal plant rather than maternal lineage.
Not all of our questions could be answered by a single large analysis of variance because some fruits did not fit into the neighbor identity categories. Overall effects of the maternal plant and time or resource level (cross group and branch category) were examined using the entire data set. We used mean individual seed mass, good seed number per fruit, and total seed mass per fruit (calculated as individual seed mass x number of good seeds per fruit) as dependent variables in an analysis of variance where maternal plant, cross group, branch category, and their interactions were the independent variables. Due to strong interactions between maternal plant and independent variables, the analyses were repeated for individual maternal plants where appropriate. Generally, these latter analyses are mentioned in the text but not shown.
To examine seed paternity across plants and time, we used all fruits that matured from mixed pollinations. For each fruit that was pollinated with a mixed pollen load, we calculated the number of seeds within that fruit sired by each of pollen donors B, E, and G. We then calculated the proportion of the seeds within the fruit that was sired by each pollen donor by dividing the number of seeds sired by a particular donor by the number of good seeds per fruit. This created three values for each fruit: the proportions of seeds sired by donors B, E, and G. However, since the proportions of seeds sired by the three donors will always sum to one, we could not use all three values in an analysis. Therefore, we reduced our data set by eliminating proportion of seeds sired by donor E (Donor E was selected randomly) and performed the analysis on the remaining two donors. The arcsine square-root transformation of proportion of seeds sired was used as the dependent variable in analyses of variance where pollen donor (B or G) and its interactions with maternal plant, cross group, and branch category were the independent variables. Where appropriate analyses were repeated for each maternal plant.
Subsequent to this analysis, we performed separate ANOVAs for each pollen donor. Proportion of seeds sired was the dependent variable, and maternal plant, cross group, branch category, and their interactions were the independent variables. Since these three analyses test the same hypotheses, a Bonferroni correction was used to adjust the P values that were considered significant (Rice, 1989
).
Comparison of reproductive variables and seed paternity among neighbor types and neighbor identities could be made using only that subset of the data for which these variables could be defined. Fruits for which the surrounding fruits were sired by two different pollen donors (e.g., sequence EBG) were thus excluded from these analyses. Allocation to reproduction was examined in analyses of variance for which individual seed mass, number of good seeds per fruit, and total seed mass per fruit were the dependent variables and maternal plant, neighbor type or neighbor identity, and their interactions were the independent variables. Seed paternity was examined in analyses of variance in which the arcsine, square-root transformation of proportion of seeds sired per fruit (by donor B or G) was the dependent variable, and pollen donor and its interactions with maternal plant, neighbor type, or neighbor identity were the independent variables. For seed paternity, there were interaction effects between maternal plant and neighbor identity, so those analyses were repeated for individual maternal plants.
We also used analyses of variance to test whether pollen donors differed in characters other than seed paternity. Individual seed mass after mixed pollination was the dependent variable in an analysis of variance where maternal plant, pollen donor, and their interaction were the independent variables. Effects of pollen donors in single pollinations were compared in other analyses. Mean individual seed mass, number of good seeds per fruit, and total seed mass per fruit after single pollination were used as dependent variables in analyses of variance where maternal plant, pollen donor, and their interaction effects were used as independent variables.
In the analyses described above, the models used to evaluate patterns of individual seed mass were slightly different than those used to evaluate number of seeds per fruit or total seed mass per fruit. The number of good seeds per fruit was used as a covariate in analyses of individual seed mass in order to account for trade-offs between individual seed mass and seed number per fruit.
Patterns of fruit abortion across cross groups and branch categories were compared using chi-square analyses. The hypotheses that frequencies of aborted and nonaborted fruits were equal among these groups were tested.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
The percentage of fruits that survived to maturity decreased along branches such that later pollinations were less likely to produce fruit. (Fig. 1). This result supports the hypothesis that resources for fruit production are less available to the later than to the earlier crosses on branches. Initially high rates of fruit maturation (>80%) declined to <50% by the 20th cross. Similarly, the percent of fruits aborted was low in the first two cross groups (each group represents 1/4 of the crosses performed on a branch), a bit higher in the third group, and quite high in the fourth (Table 1). These differences were significant.
|
|
While percentage fruits aborted varied among branch categories, it decreased rather than increased with increasing category number (Table 3). Both number of good seeds per fruit and total seed mass per fruit declined from the first to the subsequent branch categories, while individual seed mass declined from the first to the second and third branch categories. However, individual seed mass was greater in the fourth branch category, perhaps due to compensation. The overall effects of branch category on all three variables were small, but statistically significant (Table 2).
|
Seed paternity and changes over time
Overall, the three pollen donors differed strongly in the number of seeds they sired after mixed pollinations (Tables 4 and 5). This pattern is indicated both by comparisons of the total number of seeds sired by each pollen donor (Table 4) and by examination of the mean proportion of seeds per fruit sired by two of the three pollen donors after mixed pollinations (Table 5).
|
|
|
|
Separate examination of the performance of pollen donors also indicates that seed paternity varied over time (Table 8). For both donors B and G, proportion of seeds sired per fruit varied significantly among cross groups. The interaction effects with maternal plant were considerably weaker than in the combined analysis.
|
|
Effects of the recent pollination history
The recent pollination history of plants did not affect seed and fruit filling. Neither individual seed mass, number of seeds per fruit, nor total seed mass differed among crosses where the neighboring fruits were pollinated by the same donor or by different donors (ANOVAs, P >> 0.05). There was also no effect of the identity of the pollen donors that were used to pollinate neighboring fruits on these variables (ANOVA, P >> 0.05).
Whether neighboring fruits were pollinated by the same or by different pollen donors had no effect on the outcome of mixed pollinations. There was no significant pollen donor by neighbor type interaction effect and no significant pollen donor by maternal plant by neighbor type interaction effect on number of seeds sired per fruit.
In contrast, there was a modest potential for neighbor identity to affect the outcome of mating after mixed pollination. While there was no overall effect of neighbor identity on seed paternity (Table 9), examination of individual maternal plants revealed significant pollen donor by neighbor identity effects for two maternal plants from two different lineages (for the pollen donor x neighbor identity interaction effect, P < 0.0435 and P < 0.0212 for plants 8f10 and 7m2, respectively).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
The success of pollen donors varied over time as resources declined; the number of seeds sired per fruit by three pollen donors used in mixed pollination changed significantly over time. The pattern of this change is intriguing (Fig. 2). Number of seeds sired per fruit was most even among the pollen donors in the earliest pollinations (cross group 1), suggesting that "inexperienced" or resource-rich plants showed less discrimination among pollen donors. Differential pollen donor success peaked in the second cross group and tended to decline toward the end of the fruiting period. Thus, plants near then end of the flowering season might be less capable of distinguishing among pollen donors. This interpretation is consistent with our earlier studies, which showed that when maternal plants were under stress, seed paternity was more even among pollen donors than when plants were in better condition (Marshall, 1988
; Marshall and Fuller, 1994
).
Changes in mating among branch categories were more difficult to interpret, possibly because branch category was a less clear measure of plant status than cross group. Fruits in later cross groups are produced under different conditions than fruits in early cross groups. Later fruits must compete for resources with more mature fruits closer to the main stem of the plant and closer to resources stored in the roots. However, because there may be some independence among branches, and some early branches that completed fruit filling before later branches began flowering, later branches may not always be at a resource disadvantage.
Consideration of the changes in mating pattern is interesting in light of several studies with animals. Work on several species has indicated that inexperienced females distinguish less among potential mates than do more experienced females (Bakker and Milinski, 1991
; Downhower and Lank, 1994
; Collins, 1995
; Gabor and Halliday, 1997
; Rosenquist and Houde, 1997
). This is analogous to the change we saw from the first to the second and third cross groups. While mating experience does not have the same meaning for plants and animals, plant physiology is likely affected by the presence of growing fruits, an indication of whether mating has occurred. And, over a longer time scale, pollinator service may typically have been weaker early in the season when plants have just begun to flower. This might reduce selection for mate choice early in the season.
The trend toward more equal pollen donor performance from the second to fourth cross groups fits some models developed for mate choice in animals (Real, 1990
). In sequential mate choice models with a finite time horizon, which would apply to a plant whose resources are being depleted by fruit filling during a finite reproductive season, discrimination among mates is predicted to decline toward the end of the mating period. However, the trend toward more equal performance was not strong.
Although we have interpreted mating patterns in light of changes in maternal condition, it is also possible that changes in pollen donor condition occurred over the length of the experiment. It is known that such changes can affect pollen performance (Sharma et al., 1990
; Young and Stanton, 1990
; Lau and Stephenson, 1993, 1994
; Havens, 1994
; Quesada, Bollman, and Stephenson, 1995
; Mazer and Gorchov, 1996
; Delph, Johannsson, and Stephenson, 1997
). Certainly pollen donors aged and may have senesced, even though we tried to keep the plants in good condition. However, we do not think that pollen donor condition can completely explain the results. First, the largest change in pollen donor success was between the first and second cross group. This change occurred relatively early in the experiment when pollen donors were still in peak condition. Second, when we measured pollen production over time in other experiments, pollen size and number did not change over time periods as short as the few weeks that it took to do pollinations in cross groups 1 and 2 (D. L. Marshall and P. J. Gegick, personal observation).
The recent pollination history along branches had very little effect on mating. Neither individual seed mass, number of seeds per fruit, nor total seed mass per fruit were affected by the identity of the fathers of neighboring fruits. These results are in conflict with an earlier study that demonstrated a rare pollen donor effect (Marshall and Oliveras, 1990
). In that study, fruits sired by pollen donors that were rare along branches were filled to greater masses than fruits sired by pollen donors that were common along branches. The results likely differ because the scale at which we examined pollination history differed. In the earlier study, we considered the identity of the fathers of six fruits on either side of a target fruit. However, in this study we considered only the identity of immediately neighboring fruits.
In contrast to the results on fruit filling, there was some evidence that seed paternity varied with the recent pollination history of a branch. For the two plants that showed significant effects of neighboring fruit identity, there was evidence that any given pollen donor sired a lower proportion of seeds in a fruit when the neighboring fruits had been pollinated with pollen from that donor. The significant effects are in the same direction as those in our earlier study in which fruit filling patterns favored rare pollen donors (Marshall and Oliveras, 1990
).
It is important to note that, despite the complications of time during the experiment and pollination history of branches, the overall pattern of mating was clear. Pollen donor B sired many more seeds than pollen donors E or G. The shifts in mating performance over time and the shifts in mating performance associated with recent pollination history did not obscure this result.
While we addressed the potential effects of variation in mating patterns during a season, our experiment did not address the problem of variable pollen abundance in the field. We applied quite large pollen loads to ensure that pollen competition and mate choice were possible. However, smaller pollen loads might be typical of the field. When pollen loads are quite small, opportunities for differential pollen donor success are more limited and mating may be closer to random (Mitchell, 1997
). Fortunately, another recent experiment with wild radish suggests that large pollen load size does not confound the interpretation of our experiment. When mixed pollen loads of three sizes (40, 115, and 258 grains per stigma) were applied to wild radish flowers, seed paternity was unaffected by the amount of pollen applied (Marshall et al., 2000
). These pollen loads are well within the range that has been found in field populations (Ashman, Galloway, and Stanton, 1993
).
There has also been concern that differential seed paternity will be controlled by the relatedness of maternal plants and pollen donors, resulting in strongly variable pollen donor performance across maternal plants (e.g., Waser et al., 1987
). If this is true, overall differences in pollen donor success could be rare in the field. Our experiment did not specifically address this possibility. Nonetheless, pollinations were performed on nine maternal plants from five different lineages, so we can comment on success of pollen donors across plants. While there were maternal plant by pollen donor interaction effects on seed paternity, these effects were weaker than the main effect of pollen donor identity. Although there were some switches in rank order of pollen donor performance among maternal plants, the most successful donor did best on seven of nine plants and the least successful donor did worst on seven of nine plants. This result is consistent with all of our previous work. Although small effects of maternal plant by pollen donor interactions on seed paternity have been common, the differences among pollen donors in number of seeds sired have always been stronger (Marshall and Ellstrand, 1986, 1988
; Marshall, 1991, 1998
; Marshall and Fuller, 1994
).
In conclusion, our results show that maternal condition can alter mating patterns and hint that recent pollination history can affect mating. The changes in mating with variation in maternal condition suggest that the degree of discrimination among mates varies over time. Plants just beginning to produce fruits showed more equal seed paternity among pollen donors than plants that were a bit older. Plants nearing the end of the reproductive season tended to show a reduction in discrimination among pollen donors. There was also a suggestion that the outcome of mating could be influenced by the identity of neighboring fruits. These patterns point to interesting subtleties in the ability of maternal plants to influence the outcome of mating. However, they do not overshadow the strong differences in pollen donor success across the mating season. Despite all of the intriguing complexities, there was still a strong winner among the pollen donors.
|
FOOTNOTES |
|---|
4 Author for reprint requests (marshall{at}unm.edu
).
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Ashman T. L. F. Galloway M. L. Stanton 1993 Apparent vs. effective mating in an experimental population of Raphanus sativus. Oecologia 96: 102-107[CrossRef][ISI]
Bakker T. C. M. M. Milinski 1991 Sequential female choice and the previous male effect in sticklebacks. Behavioral Ecology and Sociobiology 29: 205-210[CrossRef][ISI]
Barnes D. K. R. W. Cleveland 1963 Genetic evidence for nonrandom fertilization in alfalfa as influenced by differential pollen tube growth. Crop Science 3: 295-297
Bemis W. P. 1959 Selective fertilization in lima beans. Genetics 44: 555-562
Bertin R. I. 1990 Paternal success following mixed pollinations of Campsis radicans. American Midland Naturalist 124: 153-163[CrossRef][ISI]
Björkman T. C. Samimy K. J. Pearson 1995 Variation in pollen performance among plants of Fagopyrum esculentum. Euphytica 82: 235-240[CrossRef][ISI]
Charlesworth D. D. W. Schemske 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
Collins S. A. 1995 The effect of recent experience on female choice in zebra finches. Animal Behavior 49: 479-486[CrossRef]
Cruzan M. B. 1990 Variation in pollen size, fertilization ability, and postfertilization siring ability in Erythronium grandiflorum. Evolution 44: 843-856[CrossRef][ISI]
Currah L. 1981 Pollen competition in onion (Allium cepa L.). Euphytica 30: 687-696[CrossRef][ISI]
Delph L. F. K. Havens 1998 Pollen competition in flowering plants. In T. R. Birkhead and A. P. Moller [eds.], Sperm competition and sexual selection, 149–173. Academic Press, New York, New York, USA
Delph L. F. M. H. Johannsson A. G. Stephenson 1997 How environmental factors affect pollen performance: ecological and evolutionary perspectives. Ecology 78: 1632-1639[CrossRef][ISI]
Downhower J. F. D. B. Lank 1994 Effect of previous experience on mate choice by female mottled sculpins. Animal Behavior 47: 369-372[CrossRef]
Ellstrand N. C. 1984 Multiple paternity within the fruits of the wild radish, Raphanus sativus. American Naturalist 123: 819-828[CrossRef][ISI]
Ellstrand N. C. B. Devlin 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, USA 86: 9044-9047
Ellstrand N. C. D. L. Marshall 1985 Interpopulation gene flow by pollen in wild radish, Raphanus sativus. American Naturalist 126: 606-616[CrossRef][ISI]
Ellstrand N. C. D. L. Marshall 1986 Patterns of multiple paternity in populations of Raphanus sativus. Evolution 40: 837-842[CrossRef][ISI]
Gabor C. R. T. R. Halliday 1997 Sequential mate choice by multiply mating smooth newts: females become more choosy. Behavioral Ecology 8: 162-166
Grant V. 1995 Sexual selection in plants: pros and cons. Proceedings of the National Academy of Sciences, USA 92: 1247-1250
Havens K. 1994 Clonal repeatability of in vitro pollen tube growth rates in Oenothera organensis (Onagraceae). American Journal of Botany 81: 161-165[CrossRef][ISI]
Hinata K. 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
Kay Q. O. N. 1976 Preferential pollination of yellow-flowered morphs of Raphanus raphanistrum by Pieris and Eristalis spp. Nature 261: 230-232[CrossRef]
Lau T. C. A. G. Stephenson 1993 Effects of soil nitrogen on pollen production, pollen grain size, and pollen performance in Curcurbita pepo (Curcurbitaceae). American Journal of Botany 80: 763-768[CrossRef][ISI]
Lau T. C. A. G. Stephenson 1994 Effects of soil phosphorus on pollen production, pollen size, pollen phosphorus content and the ability to sire seeds in Curcurbita pepo (Cucurbitaceae). Sexual Plant Reproduction 7: 215-220[ISI]
Levin D. A. 1975 Gametophytic selection in Phlox. In D. L. Mulcahy [ed.], Gamete competition in plants and animals, 207–217. North Holland, Amsterdam, The Netherlands
Lyons E. E. N. M. Waser M. V. Price J. Antonovics 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. 1988 Post pollination effects on seed paternity: mechanisms other than microgametophyte competition operate in wild radish. Evolution 42: 1256-1266[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]
Marshall D. L. 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
Marshall D. L. J. J. Avritt M. Shaner R. L. Saunders 2000 Effects of pollen load size and composition on pollen donor performance in wild radish, Raphanus sativus (Brassicaceae). American Journal of Botany 87: 1619-1627
Marshall D. L. P. K. Diggle 2001 Mechanisms of differential pollen donor performance in wild radish, Raphanus sativus (Brassicaceae). American Journal of Botany 88: 242-257
Marshall D. L. 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]
Marshall D. L. N. C. Ellstrand 1988 Effective mate choice in wild radish: evidence for selective seed abortion and its mechanism. American Naturalist 131: 736-759
Marshall D. L. 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]
Marshall D. L. D. A. Levin N. L. Fowler 1985 Plasticity in yield components in response to fruit predation and date of fruit initiation in three species of Sesbania. Journal of Ecology 73: 71-81[CrossRef]
Marshall D. L. D. M. Oliveras 1990 Is regulation of mating within branches possible in wild radish, Raphanus sativus?. Functional Ecology 4: 619-627
Mazer S. J. D. L. Gorchov 1996 Parental effects on progeny phenotype in plants: distinguishing genetic and environmental causes. Evolution 50: 44-53[CrossRef][ISI]
Mazer S. J. A. A. Snow M. L. Stanton 1986 Fertilization dynamics and parental effects upon fruit development in Raphanus raphanistrum: consequences for seed size variation. American Journal of Botany 73: 500-511[CrossRef][ISI]
Mitchell R. J. 1997 Effects of pollen quality on progeny vigor: evidence from the desert mustard Lesquerella fendleri. Evolution 51: 1679-1684[CrossRef][ISI]
Mitchell R. J. 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]
Mulcahy D. L. 1979 The rise of the angiosperms: a genecological factor. Science 206: 20-23
Ottaviano E. M. M. Sari-Gorla 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
Ottaviano E. M. M. Sari-Gorla D. L. Mulcahy 1975 Genetic and intergametophytic influences on pollen tube growth. In D. L. Mulcahy [ed.], Gamete competition in plants and animals, 125–134. North Holland, Amsterdam, The Netherlands
Pasonen H. L. P. Pulkkinen M. Käpylä A. Blom 1999 Pollen-tube growth rate and seed-siring success among Betula pendula clones. New Phytologist 143: 243-251[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
Pfahler P. L. 1974a Fertilization ability of maize (Zea mays L.) pollen grains. IV. Influence of storage and the alleles at the shrunken, sugary, and waxy loci. In H. F. Linskens [ed.], Fertilization in higher plants, 15–25. North-Holland, Amsterdam, The Netherlands
Pfahler P. L. 1974b Pollen genotype studies in maize (Zea mays L.). In H. F. Linskens [ed.], Fertilization in higher plants, 3–14. North-Holland, Amsterdam, The Netherlands
Pfahler P. L. 1975 Factors affecting male transmission in maize (Zea mays L.). In D. L. Mulcahy [ed.], Gamete competition in plants and animals, 115–124. North Holland, Amsterdam, The Netherlands
Pfahler P. L. H. F. Linskens 1972 In vitro germination and pollen tube growth of maize (Zea mays L.) pollen. VI. Combined effects of storage and alleles at the waxy (wx), sugary (su1), and shrunken (sh2) loci. Theoretical and Applied Genetics 42: 136-140[CrossRef][ISI]
Pundir N. S. R. F. Abbas 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]
Queller D. C. 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. K. Bollman A. G. Stephenson 1995 Leaf damage decreases pollen production and hinders pollen performance in Cucurbita texana. Ecology 76: 437-443[CrossRef][ISI]
Real L. 1990 Search theory and mate choice. I. Models of single-sex discrimination. American Naturalist 136: 376-404[CrossRef][ISI]
Rice W. R. 1989 Analyzing tables of statistical tests. Evolution 43: 223-225[CrossRef][ISI]
Richardson T. E. A. G. Stephenson 1991 Effects of parentage, prior fruit set and pollen load on fruit and seed production in Campanula americana L. Oecologia 87: 80-85[CrossRef][ISI]
Rigney L. P. J. D. Thomson M. B. Cruzan J. Brunet 1993 Differential success of pollen donors in a self-compatible lily. Evolution 47: 915-924[CrossRef][ISI]
Rosenquist G. A. Houde 1997 Prior exposure to male phenotypes influences mate choice in the guppy, Poecilia reticulata. Behavioral Ecology 8: 194-198
