| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Biological Sciences, Wellesley College, Wellesley, Massachusetts 02481 USA; and Rocky Mountain Biological Laboratory, Box 519, Crested Butte, Colorado 81224 USA
Received for publication March 7, 2000. Accepted for publication June 6, 2000.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: Antirrhinum • assortative mating • Bombus • flower color • natural selection • pollination
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
; Fry and Rausher, 1997
; Morgan and Schoen, 1997
). Previous investigations of pollinator-mediated selection on single-locus traits have predicted directional selection against one allele; nonetheless the apparently deleterious alleles are maintained (Stanton, Snow, and Handel, 1986
; Jones, 1996
), illustrating some of the difficulties of studying selection in natural populations. In Ipomoea purpurea a flower-color polymorphism is maintained in an unexpected way: when the white morph is relatively rare, pollinators discriminate against it, resulting in an increased rate of selfing. Because inbreeding depression is weak, more copies of the white allele than the dark allele are transmitted to the next generation via selfed progeny (the "Fisher effect"; Fisher, 1941
; Fry and Rausher, 1997
). If showy flowers typically are adaptations to attract pollinators, we should be able to find examples of alleles that decrease rates of pollinator visitation to plants and are therefore driven to lower frequencies by pollinator-mediated selection.
Faced with variation in floral morphology within a plant population, pollinators may ignore the variation, treating plants randomly with respect to the variable trait, or they may adjust their behavior in any of several possible ways. One of the alternative morphs may receive (1) more plant visits, by which we mean approaches to a plant resulting in at least one entry into a flower, (2) more flower visits per plant visit, and/or (3) longer flower visits. Any of these pollinator responses may influence the relative fitnesses of the alternative floral morphs. An increase in plant visits is likely to increase both pollen donation and receipt of outcross pollen (e.g., Galen, 1989
; Young and Stanton, 1990
). Also, plants that attract more visits may have a greater diversity of mates (Galen, 1992
; Harder and Barrett, 1996
), if pollinators carry pollen from different donors and distribute a plant's pollen to different recipient plants on the various foraging trips. Increasing the number of flowers probed per plant visit will generally result in more pollen donated and received per flower, since the average flower will receive more visits (Galen and Stanton, 1989
; Harder and Thomson, 1989
; Harder, 1990
; Young and Stanton, 1990
; Mitchell and Waser, 1992
). An increase in the quantity of pollen transferred may be compromised by a decrease in effective quality, however, as more pollen will be transferred within a plant (Harder and Barrett, 1995
; Hoc and Garcia, 1997
). The duration of pollinator visits to individual flowers is an important determinant of the amount of pollen picked up or deposited in some species (e.g., Conner, Davis, and Rush, 1995
; Thostesen and Olesen, 1996
) but not others (Mitchell and Waser, 1992
). The relationship will depend on how pollen is packaged and dispensed (Harder and Thomson, 1989
), and on the behavior of the pollinator while inside the flower (Neff and Simpson, 1990
; Mitchell and Waser, 1992
). It may be more beneficial for a plant to export smaller amounts of pollen on a larger number of visitors (Harder and Thomson, 1989
).
In addition to influencing the quantity and quality of pollen export and receipt, pollinator behavior directly affects the pattern of mating among floral morphs in a polymorphic population. If pollinators truly ignore the variation, so that the morphs are visited in proportion to their representation in the population and the sequences of plants visited are random with respect to the polymorphism, then the pattern of mating among morphs will be random in the population (barring self-fertilization). Random mating is a useful null hypothesis that can be addressed directly by testing for deviations from Hardy- Weinberg genotype proportions in the offspring generation. Assortative mating among floral morphs results from pollinator constancy, in which a pollinator visiting a plant of one morph is more likely to move to another of the same morph than would be expected based on morph frequencies in the population (Waser, 1986
). The term "constancy" is often used broadly to designate the tendency of most pollinators to restrict their visits to a subset of the available flower types, thereby encompassing both assortative transitions among plants and nonrandom preferences (e.g., Chittka, Thomson, and Waser, 1999
). Therefore I will use the term "within-bout constancy" to indicate the more narrow meaning, restricted to assortative transitions among plants. Positive assortative mating also can result from heterogeneous preferences among pollinators (Jones, 1997
). Individuals overvisiting one morph will pick up mostly that morph's pollen and carry it to stigmas primarily of the same morph, and other individuals do the same for the other morph(s).
We are using a simplified model system to examine pollinator response to a single-gene floral polymorphism, and the consequences of pollinator behavior for plant reproduction. Viability selection has become much better understood through testing of models with experimental populations, for example Sewall Wright's classic studies of single loci in Drosophila (e.g., Wright and Kerr, 1954
). The primary goal of our research project is to achieve a similar understanding of fertility selection, so that we can predict accurately the generation-by-generation microevolution of a floral trait in experimental populations, using observations of pollinator behavior and an appropriate fertility selection model. Once the effects of assortative mating and selection are understood for evenly spaced, randomized populations, they can be applied to more realistic scenarios.
The specific questions addressed here are the following: (1) Do bees forage randomly among the floral color morphs? If not, how do they deviate from random behavior? (2) Does pollen export or receipt by individual flowers depend on any or all of the observed components of pollinator behavior (number of visits, average visit duration, total duration of visits)? (3) How do the offspring genotype and allele frequencies compare to the parental population's equal allele frequencies and Hardy-Weinberg genotype frequencies? (4) How does experimentally augmenting nectar amounts in one color morph affect bee behavior, pollen transfer, and offspring genotype and allele frequencies?
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). The genetics of snapdragon floral characteristics have been studied for decades (e.g., Wheldale, 1907
; Dayton, 1956
; Stubbe, 1966
), and Antirrhinum is an important model plant for the study of floral development (e.g., Schwarz-Sommer et al., 1990
; Coen and Meyerowitz, 1991
; Bradley et al., 1996
; Luo et al., 1996
). Snapdragons are easily grown, have no special germination requirements, and bloom 10 wk from sowing. Antirrhinum majus flowers are hermaphroditic and protandrous. Anthers dehisce on the first day that the corolla becomes accessible to pollinators and the stigma becomes receptive on the 3rd or 4th d. Individual flowers remain open for 5–15 d, with a lack of pollination extending the bloom time (Jones and Reithel, unpublished data). Plants in our field experiments averaged 3.6 flowers open at a time on a single inflorescence. The complex flowers exclude small insects, as the weight of relatively large bees such as bumble bees is required to depress the lower lip and gain access to nectar and pollen. The energy cost of flight is very high in bumble bees, constraining their behavior to maximize foraging efficiency (Heinrich, 1975
). They can use floral pigments as cues that help reduce searching and handling time (Waser and Price, 1981
).
The genetic basis of floral pigment formation has been most thoroughly studied in Antirrhinum and Petunia (Dooner, Robbins, and Jorgensen, 1991
; Martin and Gerats, 1993
). Coen and coworkers are studying the genetics of pigment biosynthesis in Antirrhinum (e.g., Almeida et al., 1989
); two of their genetic lines are being used in our study (courtesy of R. Carpenter, John Innes Institute, Norwich, UK). Incolorata is a recessive mutation in the gene encoding flavanone 3-hydroxylase, a required enzyme in the biosynthetic pathway of anthocyanin pigments in pink ("wild-type") snapdragons (Martin et al., 1991
). We chose incolorata instead of other white mutants because it blocks anthocyanin biosynthesis downstream of chalcone synthase (Jackson, Roberts, and Martin, 1992
), and so avoids the pleiotropic effects caused by the loss of chalcone synthase (e.g., Mo, Nagel, and Taylor, 1992
). Homozygous incolorata plants have white flowers with two small yellow marks on the lower lip. Pink flowers also have the yellow marks, which are thought to serve as nectar guides for bees (Müller, 1876)
. Sulfurea is a recessive mutation of a second gene that causes the yellow pigment to be expressed over the entire face of the flower rather than being restricted to the nectar guides. This yellow pigment is primarily aureusidin (Seikel and Geissman, 1950
), an anthochlor of the aurone class of flavonoid pigments (Scogin, 1983
). On a homozygous incolorata background, segregating sulfurea produces white and yellow phenotypes in Mendelian proportions (K. Jones, unpublished data). Sulfurea flowers may be more attractive to bees because their yellow color contrasts better with background vegetation to the insect eye (Kevan, 1983
). On the other hand, the loss of contrast with the nectar guide marks may reduce their appeal to bumble bees, as demonstrated for the analogous yup gene in Mimulus (Schemske and Bradshaw, 1999
). Thus in a population polymorphic for sulfurea, either phenotype could be favored by pollinators, allowing greater flexibility in the questions the system is used to address.
Methods
We observed pollinator visits to experimental arrays of potted snapdragon plants placed in subalpine meadows at the Rocky Mountain Biological Laboratory (RMBL), elevation 2900 m, in Gunnison County, Colorado, USA (107° W, 39° N). The plants were visited by freely foraging bumble bees, primarily Bombus appositus workers, with a small percentage of visits from B. flavifrons, B. occidentalis, and B. californicus. The array rows were staggered so that every interior plant had six equidistant neighbors, each 0.5 m away. To the best of our knowledge, there were no Antirrhinum plants within at least a 5 km radius of the sites prior to our introduction of the experimental arrays (they do not grow in the wild in Colorado), so the bees most likely were inexperienced with snapdragons. Plants were watered daily.
From 20 June to 13 July 1994 and 11 to 28 July 1995, we observed all pollinator visits to arrays of potted plants. The 1994 array had 48 plants: 24 Liberty Yellow (commercial cultivar) and 24 Liberty Crimson (deep red). The 1995 array had 36 plants: 18 white and 18 yellow offspring from a backcross of white heterozygotes with yellow homozygous sulfurea plants. Two observers moved around the outside of the array to keep each bee in clear view, from 0900 until 1700 daily. When the plants were not being observed, during rainstorms and overnight, they were covered to prevent visitation; bees generally were not foraging in the surrounding meadows during these times. We defined flower visit duration as the time between the entrance of a bee's head into the corolla tube, after the two corolla lips were pried apart, and the reappearance of her head as she backed out of the flower. Visit duration was recorded for as many of the flower visits as possible (
70% of the visits). Thus, for each flower in the array we obtained a record of the number and spacing of visits and the duration of most visits.
To estimate pollen export and receipt and the pattern of pollen transfer among color morphs, array plants were assigned randomly to one of the following treatments, so that each color had three treatments in equal numbers: (1) flowers emasculated upon opening, stigmas collected at the end of day 5 after opening; (2) lower right anther collected upon opening, lower left anther collected at the end of day 3 after opening, stigma collected at the end of day 5; and (3) dye applied to dehiscing anthers, with blue dye in yellow flowers and pink dye in white flowers. All open flowers on a plant received the same treatment. Stigmas were examined for dye particles under 45x magnification immediately after collection, then squashed on a slide with glycerine jelly and stained with basic fucshin for pollen counts. We recorded the order of magnitude of dye particles of each color, and the number of pollen grains on the stigma, with counts over 1000 recorded as 1000+. Anthers were placed individually in microcentrifuge tubes in the field, with 1 mL of 95% ethanol added later that day. Tubes were stored for several months, then sonicated with 10 µL of detergent solution and vortexed to remove all pollen from the anthers. Six 10-µL samples from each tube were counted on a grid under 45x magnification. The amount of pollen exported from a flower was recorded as the difference between the amounts from the anther collected on day 1 and the corresponding anther on day 3.
In order to explore more fully the effects of strong pollinator preferences on pollen movement, we performed a nectar manipulation in 1997. At two of the sites (Barr and Avery, chosen randomly) we added 10 µL of 30% sucrose solution to all open yellow flowers at 0900 and 1300 daily; the Rosy site had no nectar manipulation. We made the same kinds of behavior observations as previously, except that arrays were not observed continuously as before, but for 0900 to 1200 and 1300 to 1700 time blocks on alternate days, with some weather-induced variation (observations covered one-third to one-half of the total time of pollinator activity on each array). The dates of the 1997 observations were 26 June to 10 July (Rosy site), 1–16 July (Barr site), and 15– 31 July (Avery site). Each array contained 48 plants, the F2 generation of a cross between homozygous sulfurea (yellow) and homozygous wild-type (white) on an incolorata background (no anthocyanins). The frequency of the sulfurea allele was 0.5, and the phenotype ratio was 3:1 white : yellow. Genotype and allele frequencies in the offspring generation were compared to the parental Hardy-Weinberg genotype ratios and 1:1 allele frequencies to check for nonrandom mating and fitness differences between morphs (see below).
Visitation data were analyzed by testing for heterogeneity of preference among foraging bouts for each array, followed by tests for within-bout constancy within homogeneous groups of foraging bouts (see Jones, 1997
, for the complete methodology). Differences between flower colors in number of plant visits, visits per flower, number of flowers per plant visit, average duration per flower visit, and average number of flowers open per day were tested for each array with two-tailed t tests. Dependence of the number of pollen grains exported or received on the number of visits, average duration of visits, and total visit duration to the flower were tested with simple regressions. A two-factor ANOVA tested whether recipient flower color and/or dye color (indicating donor flower color) explained a significant fraction of the variance in dye receipt, with a significant interaction term indicating assortative dye transfer among color morphs.
The offspring generation was analyzed for the 1997 Rosy and Barr arrays. All seeds produced by flowers that entered female phase during the 2-wk observation period were counted and averaged over all fruits per plant for comparisons between colors. This procedure eliminated any differences between plants in numbers of female flowers produced during the relevant time period, but the more robust individuals could still have an advantage through female function by producing more mature seeds per fruit or heavier seeds. Typical full seed set per fruit was 500 seeds, and most of the variance in total seed mass per fruit was due to seed number rather than individual seed mass (Jones and Reithel, unpublished data). To estimate the fraction of offspring sired by pollen carrying the sulfurea allele, three fruits were randomly chosen from each yellow-flowered plant and each white-flowered plant known to be heterozygous (via test crosses). Among ten seeds per fruit grown to bloom, the fraction of yellow-flowered offspring for each plant was recorded. Offspring flower color directly reflected pollen genotype for yellow (homozygous recessive) maternal plants. For white heterozygote mothers we assumed that pollen carrying the sulfurea allele sired twice the fraction of offspring that were yellow, since half of their offspring received the dominant wild-type allele maternally. We have seen no segregation distortion in hundreds of hand crosses with these lineages. To calculate genotype and allele frequencies for the offspring, we assumed that homozygous wild-type plants received the same fraction of pollen carrying the sulfurea allele as the heterozygotes did, because their white-flowered phenotypes were indistinguishable.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
|
In 1995, foraging trips among yellow- and white-flowered plants were heterogeneous [S = 238 (S is similar to chi-square; see Jones, 1997
), N = 145 trips, P < 0.001]. The foraging bouts could be divided into two groups, depending on whether the bee visited yellow more than half the time in a bout. There was no significant within-bout constancy within the homogeneous group favoring yellow [f(Y) = 0.75] (
2 = 1.5, P > 0.5). Within the homogeneous group favoring white [f(Y) = 0.30] there was significant negative within-bout constancy (more transitions between unlike colored plants than expected by chance: 2 = 13.5, P = 0.005).
In the two 1997 arrays in which nectar was added to yellow flowers, bee preferences again were heterogeneous (Avery: S = 225, N = 95 bouts, P < 0.001; Barr: S = 176, N = 91, P < 0.001). In both arrays the foraging bouts could be split into two homogeneous groups: those in which yellow was visited more than expected (25% of plants were yellow-flowered) and those with less than or equal to random visitation. Within none of those four groups was there significant within-bout constancy (2 = 0.3, 0.3, 1.8 and 0.5, P > 0.5 for each). At Rosy, where no nectar was added, visit rates were low and the overall frequency of visits to yellow was only 13% (again the expectation was 25%), preferences were not significantly heterogeneous (S = 33.9, N = 41 bouts, P = 0.75). Nor was there significant within-bout constancy (2 = 1.2, P > 0.5) at Rosy.
Pollen receipt and export
Overall, the best predictor of the number of pollen grains received on a stigma was the number of bee visits to the flower (Table 2). This was true for emasculated flowers in both 1994 and 1995, as well as for intact flowers in 1995. There was a great deal of (presumably self) pollen deposited on stigmas of intact flowers (all flowers on a plant received the same treatment), as shown by the substantially greater intercepts in the regression equations for nonemasculated compared to emasculated flowers, in both years. In 1995, the slopes of the regression equations were similar between emasculated and nonemasculated flowers, indicating similar amounts of pollen deposited per visit or per second of visit duration. In 1994, none of the measured components of visitation explained much of the variance in pollen receipt in intact flowers, perhaps because our upper limit of 1000 pollen grains counted constrained the variance.
|
The total duration of visits to a flower was the best predictor of the number of pollen grains exported from the flower in both 1994 and 1995 (Table 3). In 1994 the average visit duration was the more important component of total duration in determining pollen export, whereas the number of flower visits was more important in 1995.
|
, for details).
Offspring genotype and allele frequencies
Genotype and allele frequencies of the next generation in the 1997 Rosy and Barr arrays were concordant with pollinator behavior on those arrays. At Rosy, where no nectar was added to either color but white-flowered plants on average were larger, had more flowers open, and received more bee visits, the frequency of the sulfurea allele dropped from 0.50 to 0.44. Superimposed upon a drop in allele frequency was assortative mating (due to heterogeneity of preference and/or selfing), so that the frequency of heterozygotes dropped from 0.50 to 0.41, homozygous wild-types increased from 0.25 to 0.37 and homozygous sulfurea remained near 0.25 at 0.23 (Table 4A). White-flowered plants produced more than their expected proportion of seeds (79% instead of 75%), probably because they were more robust than yellow plants in that array, on average. Since mating was assortative and there were more white than yellow-flowered plants (3:1 Hardy-Weinberg ratio), white-flowered plants also had an advantage through male function as their mates produced more offspring. Overall, pollen carrying the wild- type allele sired 70% of the seeds from white-flowered females and 30% of the seeds from yellow-flowered females (that the percentages are reciprocals is mere coincidence). The total contribution of the wild-type allele through male function was (0.70 x 0.79) + (0.30 x 0.21) = 62%, in contrast to the frequency in the parent population of 50%.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Bees exhibited remarkably strong preferences for either yellow or red flowers in mixed populations, very rarely moving between morphs. Some individuals crawled or hovered around red flowers, seemingly unable to find an entrance, and thereafter restricted their visits to yellow flowers. The few that managed to find a way into red flowers found a relative gold mine of nectar, as red flowers went unvisited until late afternoon of the third day of regular visits to the array. Also, bees who entered red flowers spent three times longer inside them than bees in yellow flowers, on average. Thus we believe that the experiences of individuals shaped their preferences. Such extreme preference differences among foraging bouts resulted in nearly complete assortative transfer of dye particles, suggesting that almost all pollen transfer occurred within color morphs. This is the strongest case that we have seen of assortative pollinator movements with respect to a flower color polymorphism.
Among yellow- and white-flowered plants, bumble bee preferences were weaker but again were heterogeneous in all but one 1997 array. Dye transfer, measured in 1995, was assortative in that flowers of each color received more dye from the same floral morph than from the other morph. Both of these observations suggest that mating among yellow and white morphs was assortative, but there was not as strong a separation of mating pools as in the red and yellow population. Offspring genotypes in both 1997 populations showed a loss of heterozygotes relative to Hardy-Weinberg expectations, supporting the assortative mating scenario. Self-pollination also would reduce heterozygote frequencies and was probably a factor in our populations, likely explaining the loss of heterozygotes in the Rosy population where there was no heterogeneity of preference or within-bout constancy and a low overall rate of pollinator visitation. We will need to identify selfed progeny in order to better understand the role of assortative mating in determining genotype frequencies in our populations.
When one floral morph received more visits and/or longer visits than the other, our data suggest that it had a fitness advantage through superior pollination success. All of the significant relationships between the number or duration of visits to flowers and pollen export and receipt had positive slopes. We showed that plants receiving more outcross pollen per stigma produced more seeds per fruit under greenhouse conditions, where the ranges of numbers of pollen grains on stigmas and seeds per fruit were similar to those observed in our experimental populations in the field. Pollen export can be a reasonable predictor of relative siring success (Broyles and Wyatt, 1990
; Galen, 1992
; Ashman, 1998
; but see Stanton, Ashman, and Galloway, 1992
; Kobayashi, Inoue, and Kato, 1999
). Note that the increase in number of visits per flower of the favored color morph was always due to more visits to the snapdragon plants. Only in one array did the number of flowers probed per plant visit differ between morphs, and that array (Barr in 1997) had nectar added to yellow flowers. Thus it seems that the increase in flower visits was not accompanied by an increase in the rate of self-pollination, as probably would have been the case if the favored morph had more flowers visited per plant visit (Harder and Barrett, 1995
).
The number of visits to a flower and the average duration of those visits are not independent of each other (Jones, Reithel, and Irwin, 1998
). When visit duration is a function of the standing crop of nectar in a flower, then the more recently the flower was visited the shorter the next visit will be. We found that the duration of bumble bee visits to snapdragon flowers was positively correlated with the time since the flowers were last visited, providing the basis for a trade-off between the number and duration of flower visits (Jones, Reithel, and Irwin, 1998
). In the yellow- and red-flowered population (Barr in 1994), yellow-flowered plants received many more visits, but the visits to red flowers were of much longer duration, suggesting a trade-off between the quantity and quality of pollinator visits. The trade-off was much weaker among yellow- and white-flowered plants in 1995, when there was substantial variation in nectar production rate among plants and thus no negative correlation between numbers of visits a flower received and their average duration, unless plant-level effects were removed (Jones, Reithel, and Irwin, 1998
). In the two 1997 arrays where nectar was added to yellow flowers, both the number and duration of visits to yellow flowers increased markedly.
When both visit number and duration influence pollination success, modeling will be especially useful to predict the outcome of differential pollinator visitation for relative pollination and reproductive success of the alternative floral morphs. Establishing relationships between components of visitation and pollen export and receipt, such as the regression equations for the snapdragons, is the first step. Connecting pollen receipt with seed production and pollen export with siring success are next, using controlled hand pollinations and pollen competition experiments and/or measuring all of these parameters in the field (including paternity analysis). These two steps enable prediction of the male and female components of reproductive success from observations of pollinator behavior.
Ultimately, in order to understand changes in genotype and allele frequencies for a floral trait, an estimate of the pattern of mating among phenotypes is needed. When pollinators show neither heterogeneity of preference nor within-bout constancy, outcross pollen transfer should be random with respect to the floral polymorphism. In our experimental populations heterogeneity of preference resulted in assortative transfer of fluorescent dye particles among yellow and red flowers (1994), and among yellow and white flowers (1995, despite significant negative within-bout constancy). Further work is needed to establish a relationship between the degree of heterogeneity of preference (and/or within-bout constancy) and the strength of assortment, in order to predict the magnitude of deviations from Hardy-Weinberg genotype proportions.
Our models of how bumble bee behavior influences mating and reproduction in snapdragons are based on experimental populations with regular spacing (0.5 m apart) and random mixing of floral phenotypes. We found that many of the bees actively skipped over flowers of one color to preferentially visit another. Perhaps at lower plant densities fewer bees would show a preference, or those that do might sample the alternative phenotype more frequently (weaker preference by a similar fraction of the pollinators). Lower plant densities may also encourage bees to visit more flowers per plant, increasing rates of self-pollination (Van Treuren et al., 1993, 1994
; Karron et al., 1995
; Bosch and Waser, 1999
). Clumped distributions of floral morphs are likely to increase rates of assortative mating through any of several mechanisms: (1) pollinators that prefer one morph or the other based on past experience may sample the alternative morph less frequently due to lower encounter rates, (2) pollinators with no active preference for one floral morph may disproportionately visit one or the other morph within each foraging bout, depending on where they forage, resulting in passive heterogeneity of preference over time, and (3) if clumps enhance preferential visitation to either morph, nectar standing crops may become different between morphs in each patch, so that new visitors (e.g., newly emergent bumble bee workers) are more likely to form a preference. Thus more pollinators are likely to have a preference, and preferences are likely to be stronger, in populations with clumped distributions of floral morphs. Testing these hypotheses in experimental populations should be straightforward. It will be especially interesting to compare the effects of density and phenotype distribution over many species. The various kinds of pollinators are likely to respond differently, and factors such as floral complexity that affect the ease of learning and handling time on the flowers may alter the effects of these population variables in predictable ways.
A conspicuous and important variable in natural populations that our study only begins to address is the relative frequency of floral morphs. The magnitude and even the direction of pollinator preferences might depend on the relative frequency of floral morphs. For example, there is evidence that bumble bees prefer more common color morphs when the flowers are rewarding (see Smithson, in press, for review). Such behavior would accelerate the loss of the less frequent form, unless it can maintain or even increase its presence via self-fertilization, as in Ipomoea (Brown and Clegg, 1984
). Assortative mating can magnify a reproductive advantage for the more common morph, as we saw in our 3:1 white : yellow snapdragon population at Rosy. If the mating pool is restricted to within phenotypes to some extent, this will limit mating opportunities for the rare morph more so than for the common morph. Because assortative mating is likely, especially in clumped natural populations, this may impede a rare morph from successfully invading a population.
It is not easy to intuit how assortative mating and frequency dependent preferences interact and influence plant reproduction. This is where fertility selection models tailored to plant- pollinator systems will be especially valuable. Most models of selection with assortative mating are based primarily on polymorphic animal species such as ladybugs (e.g., Scudo and Karlin, 1968
; O'Donald et al., 1984
). It will be fascinating to learn how fertility selection in animal-pollinated plants differs from selection in animals that more directly choose their mates.
|
FOOTNOTES |
|---|
2 Author for reprint requests (kjones{at}wellesley.edu
)
3 Current address: Department of Ecology and Evolutionary Biology, University of California, Irvine, California 92697 USA
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Ashman, T.-L. 1998 Is relative pollen production or removal a good predictor of relative male fitness? An experimental exploration with a wild strawberry (Fragaria virginiana, Rosaceae). American Journal of Botany 85: 1166–1171[Abstract]
Bosch, M., and N. M. Waser. 1999 Effects of local density on pollination and reproduction in Delphinium nuttallianum and Aconitum columbianum (Ranunculaceae). American Journal of Botany 86: 871–879
Bradley, D., R. Carpenter, L. Copsey, C. Vincent, S. Rothstein, and E. Coen. 1996 Control of inflorescence architecture in Antirrhinum. Nature 379: 791–797[CrossRef][Medline]
Brown, B. A., and M. T. Clegg. 1984 The influence of flower color polymorphisms on genetic transmission in a natural population of the common morning glory, Ipomoea purpurea. Evolution 38: 796–803[CrossRef][ISI]
Broyles, S. B., and R. Wyatt. 1990 Paternity analysis in a natural population of Asclepias exaltata: multiple paternity, functional gender, and the "pollen-donation hypothesis." Evolution 44: 1454–1468[CrossRef][ISI]
Campbell, D. R., N. M. Waser, and E. J. Meléndez-Ackerman. 1997 Analyzing pollinator-mediated selection in a plant hybrid zone: hummingbird visitation patterns in three spatial scales. American Naturalist 149: 245–315
Chittka, L., J. D. Thomson, and N. M. Waser. 1999 Flower constancy, insect psychology, and plant evolution. Naturwissenschaften 86: 361– 377[CrossRef][ISI]
Coen, E. S., and E. M. Meyerowitz. 1991 The war of the whorls: genetic interactions controlling flower development. Nature 353: 31–36[CrossRef][Medline]
Conner, J. K., R. Davis, and S. Rush. 1995 The effect of wild radish floral morphology on pollination efficiency by four taxa of pollinators. Oecologia 104: 234–245[CrossRef][ISI]
Dayton, T. O. 1956 The inheritance of flower colour pigments. I. The genus Antirrhinum. Journal of Genetics 54: 249–260
Dooner, H. K., T. P. Robbins, and R. Jorgensen. 1991 Genetic and developmental control of anthocyanin biosynthesis. Annual Review of Genetics 25: 173–199[CrossRef][ISI][Medline]
Fisher, R. A. 1941 Average excess and average effect of gene substitution. Annals of Eugenics 11: 53–63[ISI]
Fry, J. D., and M. D. Rausher. 1997 Selection on a floral color polymorphism in the tall morning glory (Ipomoea purpurea): transmission success of the alleles through pollen. Evolution 51: 66–78[CrossRef][ISI]
Galen, C. 1989 Measuring pollinator-mediated selection on morphometric floral traits: bumble bees and the alpine sky pilot, Polemonium viscosum. Evolution 43: 882–890[CrossRef][ISI]
———. 1992 Pollen dispersal dynamics in an alpine wildflower Polemonium viscosum. Evolution 46: 1043–1051[CrossRef][ISI]
———, and M. L. Stanton. 1989 Bumble bee pollination and floral morphology: factors influencing pollen dispersal in the alpine sky pilot, Polemonium viscosum (Polemoniaceae). American Journal of Botany 76: 419–426[CrossRef][ISI]
Harder, L. D. 1990 Pollen removal by bumble bees and its implications for pollen dispersal. Ecology 71: 1110–1125[CrossRef][ISI]
———, and S. C. H. Barrett. 1995 Mating cost of large floral displays in hermaphrodite plants. Nature 373: 512–515[CrossRef]
———, and ———. 1996 Pollen dispersal and mating patterns in animal- pollinated plants. In D. G. Lloyd and S. C. H. Barrett [eds.], Floral biology: studies on floral evolution in animal-pollinated plants, 140–190. Chapman and Hall, New York, New York, USA
———, and J. D. Thomson. 1989 Evolutionary options for maximizing pollen dispersal of animal-pollinated plants. American Naturalist 133: 323–344[CrossRef][ISI]
Heinrich, B. 1975 The energetics of pollination. Annual Review of Ecology and Systematics 6: 139–170
Hoc, P. S., and M. T. A. Garcia. 1997 Floral biology and reproductive system of Phaseolus augusti (Fabaceae). Beiträge zur Biologie der Pflanzen 70: 121–140
Jackson, D., K. Roberts, and C. Martin. 1992 Temporal and spatial control of expression of anthocyanin biosynthetic genes in developing flowers of Antirrhinum majus. Plant Journal 2: 425–434[CrossRef][ISI]
Jones, K. N. 1996 Fertility selection on a discrete floral polymorphism in Clarkia (Onagraceae). Evolution 50: 71–79
———. 1997 Analysis of pollinator foraging: tests for nonrandom behaviour. Functional Ecology 11: 255–259[CrossRef][ISI]
———, J. S. Reithel, and R. E. Irwin. 1998 A trade-off between the frequency and duration of bumblebee visits to flowers. Oecologia 117: 161–168[CrossRef][ISI]
Karron, J. D., N. N. Thumser, R. Tucker, and A. J. Hessenauer. 1995 The influence of population density on outcrossing rates in Mimulus ringens. Heredity 75: 175–180[ISI]
Kevan, P. G. 1983 Floral colors through the insect eye: what they are and what they mean. In C. E. Jones and R. J. Little [eds.], Handbook of experimental pollination biology, 3–30. Van Nostrand Reinhold, New York, New York, USA
Kobayashi, S., K. Inoue, and M. Kato. 1999 Mechanism of selection favoring a wide tubular corolla in Campanula punctata. Evolution 53: 752–757[CrossRef][ISI]
Luo, D., R. Carpenter, C. Vincent, L. Copsey, and E. S. Coen. 1996 Origin of floral asymmetry in Antirrhinum. Nature 383: 794–799[CrossRef][Medline]
Martin, C., and T. Gerats. 1993 Control of pigment biosynthesis genes during petal development. Plant Cell 5: 1253–1264
———, A. Prescott, S. Mackay, J. Bartlett, and E. Vrijlandt. 1991 Control of anthocyanin biosynthesis in flowers of Antirrhinum majus. Plant Journal 1: 37–49[ISI][Medline]
Mather, K. 1947 Species crosses in Antirrhinum I. Genetic isolation of the species majus, glutinosum and orontium. Heredity 1: 175–186
Mitchell, R. J., and N. M. Waser. 1992 Adaptive significance of Ipomopsis aggregata nectar production: pollination success of single flowers. Ecology 73: 633–638[CrossRef][ISI]
Mo, Y., C. Nagel, and L. P. Taylor. 1992 Biochemical complementation of chalcone synthase mutants defines a role of flavonols in functional pollen. Proceedings of the National Academy of Sciences, USA 89: 7213– 7217
Morgan, M., and D. Schoen. 1997 The role of theory in an emerging new plant reproductive biology. Trends in Ecology and Evolution 12: 231–234[CrossRef]
Müller, H. 1876 On the relations between flowers and insects. Nature 15: 178–180
Neff, J. L., and B. B. Simpson. 1990 The roles of phenology and reward structure in the pollination biology of wild sunflower (Helianthus annuus L., Asteraceae). Israel Journal of Botany 39: 197–216[ISI]
O'Donald, P., M. Derrick, M. Majerus, and J. Weir. 1984 Population genetic theory of the assortative mating, sexual selection and natural selection of the two-spot ladybird, Adalia bipunctata. Heredity 52: 43– 61[ISI]
Schemske, D. W., and H. D. Bradshaw. 1999 Pollinator preference and the evolution of floral traits in monkeyflowers (Mimulus). Proceedings of the National Academy of Sciences, USA 96: 11910–11915
Schwarz-Sommer, Z., P. Huijser, W. Nacken, H. Saedler, and H. Sommer. 1990 Genetic control of flower development by homeotic genes in Antirrhinum majus. Science 250: 931–936
Scogin, R. 1983 Visible floral pigments and pollinators. In C. E. Jones and R. J. Little [eds.], Handbook of experimental pollination biology, 160– 172. Van Nostrand Reinhold, New York, New York, USA
Scudo, F. M., and S. Karlin. 1968 Assortative mating based on phenotype: I. Two alleles with dominance. Genetics 63: 479–498[ISI]
Seikel, M. K., and T. A. Geissman. 1950 Anthochlor pigments. VII. The pigments of yellow Antirrhinum majus. Journal of the American Chemical Society 72: 5725–5730[CrossRef]
Smithson, A. In press Pollinator preference, frequency-dependence, and floral evolution. In L. Chittka and J. Thomson [eds.], Cognitive ecology of pollination. Cambridge University Press, Cambridge, UK
Stanton, M. L., T.-L. Ashman, and L. F. Galloway. 1992 Estimating male fitness of plants in natural populations. In R. Wyatt [ed.], Ecology and evolution of plant reproduction, 62–90. Chapman and Hall, New York, New York, USA
——–, A. A. Snow, and S. N. Handel. 1986 Floral evolution: attractiveness to pollinators influences male fitness. Science 232: 1625–1627
Stubbe, H. 1966 Genetik und Zytologie von Antirrhinum L. sect. Antirrhinum: Veb. Gustav Fischer, Verlag, Jena, Germany
Thostesen, A. M., and J. M. Olesen. 1996 Pollen removal and deposition by specialist and generalist bumblebees in Aconitum septentrionale. Oikos 77: 77–84[CrossRef][ISI]
Van Treuren, R., R. Bijlsma, H. J. Ouborg, and M.M. Kwak. 1994 Relationships between plant density, outcrossing rates and seed set in natural and experimental populations of Scabiosa columbaria. Journal of Evolutionary Biology 7: 287–302
———, ———, ———, and W. Van Delden. 1993 The effects of population size and plant density on outcrossing rates in locally endangered Salvia pratesis. Evolution 47: 1094–1104[CrossRef][ISI]
Waser, N. M. 1986 Flower constancy: definition, cause, and measurement. American Naturalist 127: 593–603[CrossRef][ISI]
———, and M. V. Price. 1981 Pollinator choice and stabilizing selection for flower color in Delphinium nelsonii. Evolution 35: 376–390[CrossRef][ISI]
Wheldale, M. 1907 The inheritance of flower colour in Antirrhinum majus. Proceedings of the Royal Society of London B 79: 288–305
Wright, S., and W. E. Kerr. 1954 Experimental studies of the distribution of gene frequencies in very small populations of Drosophila melanogaster. II. Bar. Evolution 8: 225–240[CrossRef][ISI]
Young, H. J., and M. L. Stanton. 1990 Influences 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:
|
|
 |
|
  J. E. Carlson Male-biased nectar production in a protandrous herb matches predictions of sexual selection theory in plants Am. J. Botany, April 1, 2007; 94(4): 674 - 682. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. E. Hoballah, T. Gubitz, J. Stuurman, L. Broger, M. Barone, T. Mandel, A. Dell'Olivo, M. Arnold, and C. Kuhlemeier Single Gene-Mediated Shift in Pollinator Attraction in Petunia PLANT CELL, March 1, 2007; 19(3): 779 - 790. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. D. Karron, R. J. Mitchell, and J. M. Bell Multiple pollinator visits to Mimulus ringens (Phrymaceae) flowers increase mate number and seed set within fruits Am. J. Botany, September 1, 2006; 93(9): 1306 - 1312. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. L. Kaczorowski, M. C. Gardener, and T. P. Holtsford Nectar traits in Nicotiana section Alatae (Solanaceae) in relation to floral traits, pollinators, and mating system Am. J. Botany, August 1, 2005; 92(8): 1270 - 1283. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. E. Weis and T. M. Kossler Genetic variation in flowering time induces phenological assortative mating: quantitative genetic methods applied to Brassica rapa Am. J. Botany, June 1, 2004; 91(6): 825 - 836. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Cayenne Engel and R. E. Irwin Linking pollinator visitation rate and pollen receipt Am. J. Botany, November 1, 2003; 90(11): 1612 - 1618. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||
0.10