SEARCH

SEARCH BY CITATION

Keywords:

  • diurnal pollinators;
  • gene flow;
  • nocturnal pollinators;
  • pollination;
  • Silene alba

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • • 
    Silene alba exists in natural metapopulations throughout its range and is visited by a suite of both diurnal and nocturnal pollinators. Pollen-mediated gene flow may help reduce genetic isolation of subpopulations. Here, we compared the relative effects of nocturnal vs diurnal pollinators on pollen-mediated gene flow in subpopulations separated by two distance treatments.
  • • 
    We established populations consisting of genetically marked individuals in an old field in Tennessee (USA). Electrophoretic examination of seedlings produced by plants exposed to nocturnal, diurnal and control pollinator treatments and separated by either 20 or 80 m allowed us to directly measure pollen-mediated gene flow.
  • • 
    Gene flow was more common between populations separated by only 20 m. Nocturnal pollinators were responsible for most gene flow between populations, regardless of distance. Diurnal pollinators played only a small role in pollen-mediated gene flow.
  • • 
    The results suggest that nocturnal pollinators are better than diurnal pollinators at moving pollen between small S. alba subpopulations. However, their effectiveness declines as the distance between subpopulations increases, making them relatively ineffective at moving genes between isolated subpopulations.

Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Small populations suffer increased risk of extinction as a result of both demographic and genetic factors (Ellstrand, 1992; Clarke & Young, 2000; Holsinger, 2000; Richards, 2000a). Most species are not distributed continually over space and time but instead occur as sometimes very small subpopulations with unique genetic and demographic characteristics. These subpopulations are connected via gene flow. In angiosperms, gene flow is accomplished during two different life stages: the pollen and seed stages. Gene flow via these two mechanisms may operate at different spatial scales such that one mode may be more important at small and the other at larger scales (McCauley, 1997). Pollen-mediated gene flow is important for outcrossing plants, so much so that many plant species have become specialized for particular animal pollinators (Wyatt, 1983; Ellstrand, 1992). Characteristic suites of floral traits, known as ‘pollination syndromes’ (Baker & Hurd, 1968; Wyatt, 1983), are associated with specific groups of pollinators.

When plants have evolved to depend on animals to move their pollen, the degree of gene flow will be limited in part by the behavior of the animal disperser. Some pollinators behave differently in small plant populations than in large ones. For example, in three species of pollen-limited orchid (Orchis spitzelii Sauter, O. palustris Jacq., Anacamptis pyramidatis (L.) Rich.) studied in Sweden, pollinator visits increased with increasing plant population size (Fritz & Nilsson, 1994). A similar result was found for bumblebees visiting Lychnis viscaria; bees preferred large over small populations but probed more flowers within plants when in small populations (Mustajärvi et al., 2001). In particularly small populations of animal-pollinated plants, mate numbers decrease with decreasing population size and may drop to a point at which pollinator service deteriorates, resulting in population extinction (Bronstein et al., 1990; Fritz & Nilsson, 1994). Further, some pollinators (e.g. bees) forage primarily within rather than between patches (Waddington, 1983) and may change behavior in response to habitat fragmentation. For example, bumblebees pollinating Betonica officinalis visited plants in fragments less often and for less time than they visited plants in unfragmented patches (Goverde et al., 2002). The degree of population isolation can also influence pollinator behavior. For example, pollinators were less abundant but visited more plants and flowers within isolated populations of Delphinium nuttallianum compared with populations that were not isolated (Schulke & Waser, 2001).

Many of these studies designed to examine the response of pollinators to small plant populations have focused on a single type of pollinator (e.g. bumblebees; Mustajärvi et al., 2001; Goverde et al., 2002). In spite of the conceptual cleanliness of the pollination syndrome concept, most insect-pollinated plants are generalists rather than specialists in terms of their attractiveness to pollinators and are visited by many species (Pellmyr & Thompson, 1996; Waser et al., 1996; Johnson & Steiner, 2000; Fenster et al., 2004). Different types of pollinators exhibit different foraging behaviors. In contrast to bumblebees, some pollinators such as hummingbirds, butterflies and large moths (e.g. noctuids and sphingids) spend more time moving between patches and travel longer distances in search of pollen or nectar (Schmitt, 1980; Courtney et al., 1981; Miyake & Yahara, 1998). For plants that are generalists, pollinated by a diversity of animals, population genetic structure may be influenced by the proportion of pollen transferred by different pollinators (Schmitt, 1980). To fully understand the role of pollen movement in gene flow in small plant populations therefore requires decomposition of the effects on gene flow of different types of pollinators with different behaviors. This kind of study is particularly important in light of the current ‘pollination crisis’ (Allen-Wardell et al., 1998; Kearns et al., 1998; but also see Ghazoul, 2005) so that we can understand the likely impact of declines of certain pollinators on plants.

In this study, we examined the relative contributions of nocturnal vs diurnal pollinators in facilitating gene flow between small Silene alba Poiret populations experiencing different degrees of spatial isolation. This species has been the subject of extensive long-term research in which natural populations in a 25 × 25 km study area in Giles County, VA, USA have been followed through space and time (Antonovics et al., 1994). Subpopulations of S. alba are typically spaced within 160 m of one another (Richards, 2000a). These subpopulations vary widely in size, from as few as a single plant to > 255 plants, and are ephemeral, showing substantial turnover, although the population as a whole persists. Small subpopulations experience much higher rates of extinction than do large subpopulations (Antonovics et al., 1994). These dynamics are characteristic of a metapopulation (Levins, 1970) in which the long-term persistence of the species is dependent upon the pattern of colonization and extinction of subpopulations. Newly colonized subpopulations of S. alba can be very small, consisting of as few as five individuals (Antonovics et al., 1994), and experience severe inbreeding depression (Richards, 2000a).

In S. alba, fruit set and gene flow are both influenced by the distance between subpopulations. In a manipulative experiment, Richards et al. (1999) found that mean fruit set, a measure of pollination success, dropped from 0.25 fruits per female flower in populations separated by 20 m to 0.05 fruits per flower in populations separated by 80 m. Similarly, the immigration rate of pollen into populations was 47% for populations separated by 20 m but dropped to 6% in populations separated by 80 m. Comparison of population genetic structure at this spatial scale using both nuclear and cytoplasmically inherited markers revealed that in S. alba pollen movement has a much larger role than seed movement in bringing about gene flow (McCauley, 1994, 1997).

Silene alba flowers have a long corolla tube, white petals, and a sweet odor, traits typical of plants pollinated by nocturnal moths (Baker & Hurd, 1968; Young, 2002). However, flowers are visited by a wide variety of both nocturnal and diurnal pollinators including noctuid, sphingid and geometrid moths as well as syrphid flies, a variety of bees, and butterflies (Shykoff & Bucheli, 1995; Altizer et al., 1998; Young, 2002). Some studies report higher flower visitation rates for nocturnal pollinators (Shykoff & Bucheli, 1995) and other studies for diurnal pollinators (Altizer et al., 1998). Nocturnal pollinators such as large moths are long-distance fliers (Linhart & Mendenhall, 1977) that have been observed feeding in widely spaced populations of S. alba (Altizer et al., 1998), whereas many diurnal pollinators such as bees tend to forage primarily within rather than between flower patches (Altizer et al., 1998). Further, many nocturnal pollinators rely on olfaction to discover even small patches of flowers (reviewed in Kelber et al., 2003). In contrast, many diurnal pollinators rely on vision to detect flowers and are generally more attracted to large floral displays (Handel, 1983). For these reasons, we predicted that, for small populations of S. alba, nocturnal pollinators would play a larger role in pollen-mediated gene flow than would diurnal pollinators, especially for populations isolated by greater distances. Specifically, we set out to determine (1) the relative role of nocturnal vs diurnal pollinators in pollen-mediated gene flow between subpopulations of S. alba, and (2) whether larger, longer flying nocturnal pollinators are better at moving pollen longer distances than their smaller, shorter flying diurnal counterparts.

Materials and Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Study species

The white campion, Silene alba Poiret [= S. latifolia (Miller) Krause =S. pratensis (Spreng) Gren & Godr. =Lychnis alba Miller =Melandrium album (Miller) Garcke] (Caryophyllaceae), is a dioecious short-lived perennial herb native to Eurasia and introduced to North America early in the 19th century (McNeill, 1977; Young, 2002). The species has been used as a model system for studying plant metapopulations from both demographic and genetic perspectives (Antonovics et al., 1994; McCauley, 1994, 1997; Richards, 2000b). In the south-eastern United States, peak flowering times span from May to mid-July, although flowering can take place across the entire summer period (Altizer et al., 1998). Male plants produce smaller flowers but usually have more flowers open at a time than do females (Alexander & Maltby, 1990; Shykoff & Bucheli, 1995; Altizer et al., 1998). The total number of flowers open on a plant depends on plant size (E. Barthelmess, personal observation). Flowers open in the late afternoon and remain open into the middle of the next day, providing ample opportunity for visits by both nocturnal and diurnal pollinators. Both male and female flowers produce nectar; nectar production is generally highest between 18:00 and 22:00 h but nectar is produced during both diurnal and nocturnal periods of pollinator activity (Witt et al., 1999). Once ovules are fertilized in female flowers, the flowers wilt. Female plants halt flower production once they have set fruit (Altizer et al., 1998). Fruits (capsules) mature in late summer and contain on the order of 300 small seeds (similar to poppy seeds; Richards, 2000a). Silene alba flowers can become infected with the anther-smut disease produced by Ustilago violacea, a fungal pathogen that causes diseased flowers to produce spores rather than pollen and which is transmitted between plants by insect pollinators (Alexander & Antonovics, 1988). No diseased plants were observed in our study.

Experimental design

We conducted this experiment between May and August 1998 in an old field at the Shelby Bottoms Park in Nashville, Tennessee, USA (36°11′−41′ N, 86°41′−43′ W). The site was within the floodplain of the Cumberland River in an area off limits to the public. Natural vegetation in the old field was sparse because the area had been used previously for both removal and storage of fill dirt. Visual inspection before and throughout the duration of the study of an approximately 1-km2 area of the park including the study site revealed no naturally occurring S. alba plants in the area, although the plant occurs regularly throughout Tennessee.

To compare the relative effects of diurnal vs nocturnal pollinators on short vs long-distance gene flow, we established four pairs of S. alba populations (Fig. 1). Each of the eight resulting populations consisted of 10 plants (two males and eight females). Plants within each population were randomly assigned to one of three treatments: nocturnal, diurnal or control. One male plant was assigned to the nocturnal treatment and the other to the diurnal treatment. Of the eight female plants, two were assigned to the control treatment, three to the nocturnal treatment and the remaining three to the diurnal treatment. Plants within each population were contained in 15-cm-diameter pots spaced 1 m apart. We removed all pre-existing flowers from each plant upon placing the plant in the field at the outset of the experiment. To prevent browsing by deer, each population was protected within a 3.0 m × 4.0 m × 1.85 m tall fenced enclosure created by hanging monofilament netting (7.5 cm mesh) from PVC poles. Within pairs, populations were spatially arranged in one of two distance treatments. In two of the four pairs, we separated populations by a distance of 20 m and in the other two pairs populations were separated by 80 m. We separated pairs of populations from each other by a minimum distance of 140 m (Fig. 1).

image

Figure 1. Experimental design. We established four pairs of Silene alba populations (eight populations in total) in Shelby Bottoms Park, Nashville, Tennessee (USA). Within pairs, populations were spaced either 20 or 80 m apart. Each pair of populations was spaced at least 140 m from other pairs. Each population (see inset) consisted of eight female and two male plants. In each population, two female plants served as controls (no circles), while three were assigned to a nocturnal pollinator (gray circles) and three to a diurnal pollinator (open circles) treatment. One male plant was also assigned to the nocturnal treatment and one to the diurnal treatment. All nocturnal or diurnal plants within a population were homozygous at either the first (11) or second (22) allele at the malate dehydrogenase locus. Plants were spaced 1 m apart within populations. The relative positioning of plants by sex and by treatment within each population was by random assignment.

Download figure to PowerPoint

Plants used in the experiment came from a stock collection derived from wild plants from across the eastern United States and grown from seed at Vanderbilt University. Silene alba is polymorphic at the malate dehydrogenase (MDH) locus, with as many as four different MDH alleles present in the stock collection. We selected plants for the experiment that were homozygous for either the first or second allele type, determined before the start of the experiment via starch gel electrophoresis (following methods in Werth, 1985 and McCauley, 1994). So that we could compare the relative rates of gene flow from within and between population pairs, all plants subjected to the nocturnal and diurnal pollinator treatments within one population shared the same homozygous MDH genotype. Plants in the paired population were also all homozygous, but at a different MDH genotype. Thus, we deemed heterozygous progeny to be the product of gene flow from a plant in the neighboring population and homozygous progeny to be the product of gene flow from a plant within the home population. We did not have enough homozygous plants to use for the control pollinator treatments, and hence we used heterozygotes. However, because control plants were female, there was no risk of confusing outcrossed progeny with progeny produced by genes from within the home population.

Upon establishment of all populations in June, plants were given 2 d in the field to adjust to relocation before beginning the experiment. Thereafter, we visited plants twice daily, once in the morning and once in the evening. Between 06:00 and 07:00 h each day, we covered plants in the nocturnal treatment with bridal veil nets suspended over tomato cages to prevent pollination but left plants in the diurnal treatment uncovered. Between 19:00 and 20:00 h each day, we covered the plants in the diurnal treatment with similar nets and removed nets from the plants in the nocturnal treatment. Tomato cages were used to suspend nets above all flowers to prevent accidental pollination resulting from pollen falling onto nets. Control plants were never covered by nets. On a few occasions we observed bumblebees beginning to visit flowers as we finished uncovering plants in the diurnal treatment but in the evenings as we unbagged the nocturnal plants we generally did not observe any pollinators. Approximately every second day for the duration of the experiment we counted the number of open flowers on each plant and counted and removed all mature seed capsules. We stored seed capsules in individual envelopes labeled with the identity of the parental plant and the population from which the capsule was taken. During the fall of 1998, we planted seeds from each capsule and grew them to seedling stage in the Vanderbilt University glasshouse. When plants reached seedling stage, we collected leaf samples from each plant and stored them in a −80°C freezer before electrophoresis. During the winter and spring of 1999 we screened leaf samples at the MDH locus using starch gel electrophoresis.

Data analysis

We checked response variables for normality and transformed those that deviated with either the square root (number of capsules produced and number of capsules screened) or arcsin (fruit set and proportion of capsules outcrossed) transformation. Means from arcsin-transformed variables were backtransformed before plotting. We then used a combination of two-factor analysis of variance and repeated measures analysis of variance models run with jmp statistical software (version 5.1; SAS Institute, Inc., Cary, NC, USA) to test our null hypotheses. The first factor, distance, included the 20- and 80-m treatments. The second factor, pollinator type, included nocturnal, diurnal and control treatments. For the sake of simplicity, we assumed µ = 0.05 for tests of statistical significance. However, for a number of our tests, statistical power, measured as the probability of detecting a false null hypothesis, was somewhat low (ranging from 0.17 to 0.8) so we consider tests with P-values between 0.05 and 0.10 as potentially biologically, if not statistically, significant.

We assessed the degree of outcrossing as a function of distance and pollinator type using two different two-factor analysis of variance tests, each with a different response variable. First we compared the average per plant proportion of outcrossed capsules (number of outcrossed capsules divided by total number of capsules for a plant) across plants. An initial test of the full two-factor analysis of variance (ANOVA) model revealed no significant interaction term, so we removed the interaction term from the model. For the second test we used the total number of outcrossed capsules as the response variable.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Flower and capsule production and fruit set

Repeated measures ANOVA revealed that the number of flowers produced during the experiment fluctuated significantly over time and declined for all treatments toward the end of the experiment (Table 1a; Fig. 2a,b). The number of flowers differed significantly between plants subjected to the three pollinator exposure treatments. At any given time during the course of the experiment, plants in the diurnal treatment generally produced more flowers than plants in the nocturnal or control treatments. This trend held for both the 20- and 80-m distance treatments. There was no significant change in flower number among the different distance treatments, nor was there a significant distance by pollinator by time interaction term (Table 1a). All plants had few flowers at the first point in the census. Nocturnal and control plants continued to produce few flowers while diurnal plants produced generally more flowers at any given point during the flower census (Fig. 2a,b). When we examined the total number of flowers produced per plant over the course of the entire experiment as a function of pollinator type and distance, we found a significant effect of pollinator treatment but not of distance on the number of flowers produced (Fig. 3a). Plants subjected to the diurnal pollinator treatment produced more than twice as many flowers as did plants in the control or nocturnal treatments. The pollinator type by distance interaction was not significant (Table 2a).

Table 1.  Results from repeated-measures analysis of variance comparing the effects of pollinator treatment (nocturnal, diurnal or control) and distance treatment (20 or 80 m) on (a) the average number of Silene alba flowers available at each site over time and (b) the average number of capsules collected from each site over time
(a)
EffectWilks λFdfP
Time15.2012, 70.0007*
Pollinator × time0.08 1.5124, 140.2117
Distance × time1.19 0.6912, 70.7236
Pollinator × distance × time0.12 1.0824, 140.4472
(b)
EffectWilks λFdfP
  • *

    Significant at µ = 0.05.

  • df, degrees of freedom.

Time7.32 5, 140.0015*
Pollinator × time0.421.5110, 280.1866
Distance × time0.270.75 5, 140.5997
Pollinator × distance × time0.770.4010, 280.9365
image

Figure 2. Flower and capsule production in Silene alba populations in Tennessee (USA). (a, b) Flower production over time for the different distance and pollinator treatment groups in the experiment. The y-axis shows the average number of flowers counted from the nocturnal (solid circles connected by solid line), diurnal (open circles connected by heavy dashed line) and control (open squares connected by light dashed line) pollinator treatments in the (a) 20-m and (b) 80-m distance treatments. (c, d) Capsule production over time for the different distance and pollinator treatment groups. Capsule production in (c) the 20-m and (d) the 80-m distance treatments is shown. The x-axis for all parts of the figure shows the dates of the flower or capsule censuses. Error bars indicate the standard error of the mean.

Download figure to PowerPoint

image

Figure 3. Average number of (a) flowers and (b) capsules produced by individual Silene alba plants over the course of the experiment as a function of distance (20 vs 80 m) and pollinator treatment. Cross-hatched bars, plants exposed to nocturnal pollinators; open bars, plants exposed to diurnal pollinators; diagonally hatched bars, plants exposed to all pollinators (controls). Note the difference in scale in the y-axes. Error bars indicate the standard error of the mean. Bars sharing the same letter are not statistically different.

Download figure to PowerPoint

Table 2.  Results of two-factor analysis of variance comparing the effects of pollinator treatment (nocturnal, diurnal or control) and distance (20 and 80 m) on (a) the total number of Silene alba flowers and (b) the total number of capsules produced over the course of the experiment
(a)
SourcedfSum of squaresFP
Pollinator type 28168.449.290.0003*
Distance 1 177.190.40310.5280
Pollinator × distance 2 442.770.50370.6069
Error58   
(b)
SourcedfSum of squaresFP
  • *

    Significant at µ = 0.05.

  • df, degrees of freedom.

Pollinator type 21061.777.130.0017*
Distance 1  64.310.860.3565
Pollinator × distance 2 147.770.990.3796
Error58   

The number of capsules harvested at a census period generally increased over time (Fig. 2c,d), a statistically significant effect (Table 1b). In both the 20- and 80-m distance treatments, nocturnal and control plants generally produced more capsules than did diurnal plants. When the total number of capsules produced per plant over the course of the experiment was compared as a function of pollinator treatment and distance, there was a significant effect of pollinator treatment but not of distance on the number of capsules produced by each plant (Table 2b). Multiple comparison of means using Tukey's highest significant difference test showed no difference in the number of capsules produced by plants in the nocturnal or control treatments but showed significantly fewer capsules produced by plants in the diurnal pollinator treatment (Fig. 3b).

We used fruit set, defined as the per plant proportion of capsules produced per flower, as a measure of pollination success. Across all experimental treatments, average per plant fruit set was 0.440 ± 0.041 (standard error of the mean). Plants in the 20- and 80-m distance treatments averaged between 0.44 and 0.48 for proportion fruit set and did not differ statistically from one another (F1,58 = 0.23, P = 0.63). However, plants in the diurnal pollinator treatment had significantly lower fruit sets than did plants in either the nocturnal or control treatments, averaging only 12% as many capsules per flower as plants from the nocturnal treatment (F2,58 = 33.61, P < 0.0001; Fig. 4). There was no significant pollinator by distance interaction (F2,58 = 0.90, P = 0.41).

image

Figure 4. Average per plant fruit set in Silene alba plants as a function of distance (20 vs 80 m) and pollinator treatment. Cross-hatched bars, nocturnal treatment; open bars, diurnal treatment; diagonally hatched bars, control treatment. Error bars indicate the standard error of the mean. Bars sharing the same letter are not significantly different.

Download figure to PowerPoint

Gene flow

We definitively genotyped a total of 487 seedlings from 168 different capsules in order to measure rates of gene flow among our different populations. There were no significant differences in the number of capsules screened as a function of pollinator type, distance, or pollinator by distance interaction (pollinator: F1,36 = 1.91, P = 0.18; distance: F1,36 = 01.11, P = 0.30; pollinator × distance: F1,36 = 0.80, P = 0.38). We screened an average of 4.2 ± 0.6 capsules per plant and an average of 2.7 ± 0.1 seedlings per capsule. There was no difference in the average number of seedlings screened per capsule by factor, nor was there a significant interaction term (pollinator: F1,36 = 1.35, P = 0.25; distance: F1,36 = 0.63, P = 0.43; pollinator × distance: F1,36 = 0.18, P = 0.68). Further, there were no differences in the number of seedlings or capsules screened per population as a function of distance (seedlings: F1,12 = 0.50, P = 0.49; capsules: F1,12 = 0.56, P = 0.47), pollinator type (seedlings: F1,12 = 03.17, P = 0.10; capsules: F1,12 = 3.13, P = 0.10) or the distance × pollinator type interaction (seedlings: F1,12 = 0.17, P = 0.69; capsules: F1,12 = 0.37, P = 0.55).

A total of 20 of the 168 capsules (11.9%) showed evidence of outcrossing and pollination from a source other than the maternal population. Of these, 10 capsules produced seedlings with identical, heterozygous genotypes that had to have resulted from outcrossing events, produced by fertilization from pollen from outside of the maternal population. In the remaining 10 capsules, all but one of the seedlings grown from the capsule were homozygous (not outcrossed), with one (but only one) seedling grown from the capsule showing a heterozygous and thus outcrossed genotype, indicating mixed paternity within the capsule resulting from fertilization by pollen originating from the maternal and the neighboring experimental populations.

There was a borderline effect of both pollinator treatment (F1,37 = 3.31, P = 0.077) and distance (F1,37 = 3.84, P = 0.058) on the proportion of outcrossed capsules. The mean percentage of outcrossed capsules for plants subjected to the nocturnal pollinator treatment was 8.35 ± 0.62% vs 0.54 ± 0.82% for plants in the diurnal treatment, a 15.5% difference (Fig. 5a). The proportion of outcrossed capsules was also much higher in the 20-m than in the 80-m distance treatments. Plants in the 20-m treatment groups produced an average of 8.74 ± 0.73% outcrossed capsules vs only 0.44 ± 0.70% outcrossed capsules for plants in the 80-m treatment group, a 19.9% difference (Fig. 5b). Nevertheless, across the board, outcrossing rates were low.

image

Figure 5. Average proportion of outcrossed capsules produced per Silene alba plant as a function of (a) pollinator treatment (nocturnal vs diurnal) and (b) distance (20 vs 80 m). Error bars indicate the standard error of the mean. Both the pollinator and distance effects showed borderline statistical significance. We were not able to measure outcrossing rates in the control plants, so control plants are not included here.

Download figure to PowerPoint

We compared the total number of outcrossed capsules per population. Distance, pollinator type and the distance × pollinator type interaction terms were all significant (Table 3). Populations in the 20-m distance treatment produced an average of 2.0 ± 0.44 outcrossed capsules, whereas those in the 80-m distance treatment produced only 0.5 ± 0.44 capsules. Plants exposed to diurnal pollinators produced substantially fewer outcrossed capsules (0.25 ± 0.44) compared with plants exposed to nocturnal pollinators (2.25 ± 0.44). The mean number of outcrossed capsules was substantially higher, at 3.75 ± 0.63, for plants in the 20-m distance treatment exposed to nocturnal pollinators than for any other treatment group. Averages for the other treatment groups ranged between 0.25 and 0.75 outcrossed capsules (Fig. 6). When considering only populations in the 80-m distance treatment, the mean number of outcrossed capsules produced by plants exposed only to nocturnal pollinators ( = 0.75 ± 0.63 capsules) was three times as high as the mean number produced by plants exposed only to diurnal pollinators ( = 0.25 ± 0.63 capsules), although this difference was not statistically significant (Fig. 6).

Table 3.  Two-factor analysis of variance comparing the main effects of pollinator type (nocturnal vs diurnal) and distance (20 vs 80 m) and the pollinator type × distance interaction on the number of outcrossed capsules produced in Silene alba populations in Tennessee (USA)
Source of variationdfSum of squaresFP
  • *

    Significant at µ = 0.05.

  • df, degrees of freedom.

Distance 10.888 5.360.039*
Pollinator type 11.90311.490.005*
Distance × pollinator type 10.888 5.3610.039*
Error121.99  
image

Figure 6. Mean number of outcrossed capsules per Silene alba population as a function of distance (20 vs 80 m) and pollinator exposure (nocturnal (cross-hatched bars) vs diurnal (open bars)). Error bars indicate the standard error of the mean. Both of the main effects and their interaction term were statistically significant. Bars sharing the same letter are not significantly different from each other.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Our major finding is that nocturnal and diurnal pollinator guilds did not have equal effects on pollen-mediated gene flow between small S. alba populations. To be effective in moving pollen between populations, a pollinator must not only visit flowers and collect pollen in one population but must also be effective in depositing that pollen onto the receptive stigmas of flowers in another population. Although plants subjected to the diurnal pollinator treatment produced significantly more flowers than did plants in either the nocturnal or the control treatments, the same plants produced significantly fewer capsules, the first indication that diurnal pollinators were less effective in pollinating flowers. Further, plants exposed only to diurnal pollinators had reduced per plant fruit sets compared with other plants. Finally, nocturnal pollinators were responsible for substantially higher rates of gene flow than were diurnal pollinators. Our findings have important implications in (1) clarifying the role of nocturnal vs diurnal pollinators on gene flow in a well-studied model experimental system, (2) understanding the role of pollen-mediated gene flow in small populations, and (3) broadening our understanding of the concept of pollination syndromes.

The results of our study, the first to use genetic markers to explicitly measure pollen-mediated gene flow by nocturnal and diurnal pollinators, are well corroborated by other research. We observed more flowers but fewer capsules produced by plants exposed only to diurnal pollinators. This result is probably related to differential resource allocation by plants to flowers rather than capsules when pollination is limited. Wright & Meagher (2003) found that flower production in S. alba is related to pollination; in their study, average daily flower production was almost three times higher for nonpollinated than for 100% pollinated plants. In the related species Silene vulgaris, the number of flowers declined significantly as capsule production increased (Colosi & Cavers, 1984).

We measured pollination success as fruit set, the per plant number of capsules produced per female flower, and found that fruit set was considerably lower in plants exposed to diurnal rather than nocturnal pollinators. In a study of S. alba populations in Colorado that used seed set as a measure of pollination success, seed set in flowers visited by only nocturnal pollinators was more than twice that of flowers visited by only diurnal pollinators (Young, 2002). Capsules resulting from visitation by diurnal bees produced fewer seeds than did capsules resulting from visitation by nocturnal moth pollinators. Further, moths moved fluorescent dye powder (an experimental analog for pollen) more than five times the distance that bees did (Young, 2002). Two studies examined patterns of pollinator visitation in S. alba in the context of transmission of anther-smut fungus, a venereal disease in plants (Shykoff & Bucheli, 1995; Altizer et al., 1998). Again using fluorescent dye powder as a pollen analog, Shykoff & Bucheli (1995) found that only 8% of plants with open flowers received dye during the day, whereas 39% received dye during the night. Our observed pattern, in which plants in the nocturnal pollinator treatment group showed 15.5% more outcrossing than plants in the diurnal treatment group, suggests that use of dye as a pollen analog accurately portrays the trend in the direction and magnitude of the difference between pollination by nocturnal and diurnal pollinators but that it may overestimate actual rates of gene flow.

Several factors could explain the increased effectiveness of nocturnal pollinators in our system. First, S. alba plants show a typical moth pollination syndrome (Baker & Hurd, 1968) and are perhaps simply more attractive to nocturnal than to diurnal pollinators. This supposition would be supported by higher flower visitation rates by nocturnal than by diurnal pollinators. Although we did not measure pollinator visitation rates in our study, Shykoff & Bucheli (1995) found that nocturnal pollinators had significantly higher flower visitation rates than did diurnal pollinators. However, Altizer et al. (1998) found the opposite. The degree to which a plant attracts pollinators is likely to change depending on the total pollination milieu, the combination of flowering plant species and pollinator species present in the same space and time, which may explain different visitation rates between nocturnal and diurnal pollinators in different studies.

Even if nocturnal pollinators visit S. alba flowers less frequently than do their diurnal counterparts, they could be more effective at transmitting genes if they have a greater tendency to transfer pollen between plants. There are several examples of plant species for which nocturnal pollinators have a greater per-visit effectiveness than do diurnal pollinators (reviewed in Young, 2002). For example, in the Japanese honeysuckle Lonicera japonica, diurnal bees removed but also consumed more pollen than did nocturnal moth pollinators, resulting in an overall lower pollen transfer efficiency for bees (Miyake & Yahara, 1998). Because we were only measuring instances of gene flow between rather than within populations, the increased gene flow by nocturnal pollinators likely results from a greater tendency of nocturnal pollinators in our system to forage between rather than within patches. Although nocturnal pollinator behavior has been the subject of fewer studies than the behavior of diurnal pollinators and is thus much less well understood, Young (2002) showed that, in a S. alba system, nocturnal pollinators carry fluorescent dye further than do diurnal pollinators, though she was only able to measure dye movement on a scale of centimeters rather than the tens of meters scale used in this study. A large body of evidence exists to indicate that, in other plant systems, diurnal pollinators such as bumblebees and honey bees have a tendency to spend more time foraging within than between patches (Waddington, 1983). Further work is called for on the role of different types of pollinators in moving pollen between rather than within populations.

We measured gene flow receipt as the proportion of outcrossed capsules produced in a population. Using this measure, we found relatively low rates of outcrossing compared with the work of Richards et al. (1999). The highest rate of gene flow receipt we measured was 8.74%, whereas Richards et al. found that gene flow receipt reached as high as 100% in populations spaced 20 m apart near large source populations. The reduced amounts in our study may be related to fewer pollinators in the area or to the small size of adjacent populations serving as pollen sources. Nocturnal moth pollinators are known to fluctuate widely in abundance within and between seasons (Pettersson, 1991).

The distance by which populations are separated from one another also has an important role in determining rates of pollen-mediated gene flow. Our finding that plant populations separated by only 20 m have much higher rates of gene flow than do populations separated by 80 m is consistent with the findings of Richards et al. (1999) who looked at the effects of population size and degree of population isolation on pollen-mediated gene flow in S. alba. Here we hypothesized that nocturnal pollinators, by virtue of their long-distance flight behavior, would be better at facilitating gene flow in populations separated by longer distances than would diurnal pollinators. The trend we observed, in which nocturnal pollinators were responsible for three times as much gene flow as diurnal pollinators for plants in the 80-m distance treatments, is consistent with our hypothesis but not statistically significant. Basically, gene flow rates were so low for the 80-m populations that it would be very difficult to statistically confirm a difference in effect between nocturnal and diurnal pollinators. Nevertheless, we did find a significant pollinator type by distance effect on the rate of gene flow. This finding is particularly important when placed in the context of the interplay among dispersal, extinction and colonization in plant metapopulations. Changes in the frequency of different types of pollinators could have enormous implications for metapopulation structure if key pollen vectors decline in number. The numerical decline in pollinators could result from direct anthropogenic disturbance of the pollinators (Kearns et al., 1998) or from habitat fragmentation if plant populations occurring in fragments are less able to attract (because of isolation) or support (because of population size) an adequate number of pollinators (Moody-Weis & Heywood, 2001; Donaldson et al., 2002).

Finally, although we clearly determined that for our S. alba system nocturnal pollinators played a much greater role in between-population gene flow than did diurnal pollinators, diurnal pollinators nevertheless contributed to pollen-mediated gene flow. This finding supports the notion advanced by Waser et al. (1996) that plant pollination systems are actually more generalized than the traditional ‘pollination syndrome’ view might suggest. Thus, although the apparent pollination syndrome of a plant may well be suited to a suite of particular pollinators, the concept of pollination syndrome should be used not as a strict rule but as a guide for the design of experiments that empirically test the role of different pollinators in mediating gene flow. Remarkably little is known about the foraging behavior, both within and between patches, of the large moths that function as nocturnal pollinators in this system.

Implications

‘Genetic rescue’ (Richards, 2000b) via pollen-mediated gene flow may reduce the negative consequences of inbreeding depression and thus reduce the likelihood of population extinction in small, isolated plant populations. Although we showed that nocturnal pollinators are considerably more effective than diurnal pollinators at moving genes in our system, their effectiveness declined in populations separated by 80 m. At fine spatial scales, nocturnal pollinators are likely the key agents of pollen movement in S. alba metapopulations. As the frequency of pollen-mediated gene flow decreases with increasing subpopulation isolation, plants must rely on other mechanisms (e.g. demographic rescue via immigration of new individual plants) in order to persist through time.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We thank Vanderbilt University undergraduates G. Gallaspy and S. White for assistance in genotyping progeny. We thank the Nashville MetroParks Office for permission to work at Shelby Bottoms Park and N. Panshin for assistance in the field and general logistical support. We thank A. Pai and two anonymous reviewers for helpful comments on the manuscript. Financial support was provided by the National Science Foundation through a research grant (DEB-9610496) to DEM.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Alexander HM, Antonovics J. 1988. Disease spread and population dynamics of anther-smut infection of Silene alba caused by the fungus Ustilago violacea. Journal of Ecology 76: 91104.
  • Alexander HM, Maltby A. 1990. Anther-smut infection of Silene alba caused by Ustilago violacea: factors determining fungal reproduction. Oecologia 84: 249253.
  • Allen-Wardell G, Bernhardt P, Bitner R, Burquez A, Buchmann S, Cane J, Cox PA, Dalton V, Feinsinger P, Ingram M, Inouye D, Jones CE, Kennedy K, Kevan P, Koopowitz H, Medellin R, Medellin-Morales S, Nabhan GP, Pavlik B, Tepedino V, Torchio P, Walker S. 1998. The potential consequences of pollinator declines on the conservation of biodiversity and stability of food crop yields. Conservation Biology 12: 817.
  • Altizer SM, Thrall PH, Antonovics J. 1998. Vector behavior and the transmission of anther-smut infection in Silene alba. American Midland Naturalist 139: 147163.
  • Antonovics J, Thrall PH, Jarosz A, Stratton D. 1994. Ecological genetics of metapopulations: the Silene-Ustilago plant-pathogen system. In: RealL, ed. Ecological genetics. Princeton, NJ, USA: Princeton University Press, 146170.
  • Baker HG, Hurd PD. 1968. Intrafloral ecology. Annual Review of Entomology 13: 385414.
  • Bronstein JL, Gouyon P-H, Gliddon C, Kjellberg F, Michaloud G. 1990. The ecological consequences of flowering asynchrony in monoecious figs: a simulation study. Ecology 71: 21452156.
  • Clarke GM, Young AG. 2000. Introduction: genetics, demography and the conservation of fragmented populations. In: YoungAG, ClarkeGM, eds. Genetics, demography and viability of fragmented populations. Cambridge, UK: Cambridge University Press, 16.
  • Colosi JC, Cavers PB. 1984. Pollination affects percent biomass allocated to reproduction in Silene vulgaris (Bladder campion). American Naturalist 124: 299306.
  • Courtney SP, Hill CJ, Westerman A. 1981. Pollen carried for long periods by butterflies. Oikos 38: 260263.
  • Donaldson J, Nänni I, Zachariades C, Kemper J. 2002. Effects of habitat fragmentation on pollinator diversity and plant reproductive success in renosterveld shrublands of South Africa. Conservation Biology 16: 12671276.
  • Ellstrand NC. 1992. Gene flow by pollen: implications for plant conservation genetics. Oikos 63: 7786.
  • Fenster CB, Armbruster S, Wilson P, Dudash MR, Thomson JD. 2004. Pollination syndromes and floral specialization. Annual Review of Ecology, Evolution, and Systematics 35: 375403.
  • Fritz AL, Nilsson LA. 1994. How pollinator-mediated mating varies with population size in plants. Oecologia 100: 451462.
  • Ghazoul J. 2005. Buzziness as usual? Questioning the global pollination crisis. Trends in Ecology and Evolution 20: 367373.
  • Goverde M, Schweizer K, Baur B, Erhardt A. 2002. Small-scale habitat fragmentation effects on pollinator behaviour: experimental evidence from the bumblebee Bombus veteranus on calcareous grasslands. Biological Conservation 104: 293299.
  • Handel SN. 1983. Pollination ecology, plant population structure, and gene flow. In: RealL, ed. Pollination biology. Orlando, FL, USA: Academic Press, 163212.
  • Holsinger KE. 2000. Demography and extinction in small populations. In: YoungAG, ClarkeGM, eds. Genetics, demography and viability of fragmented populations. Cambridge, UK: Cambridge University Press, 5474.
  • Johnson SD, Steiner KE. 2000. Generalization versus specialization in plant pollination systems. Trends in Ecology and Evolution 15: 140143.
  • Kearns CA, Inouye DW, Waser NM. 1998. Endangered mutualisms: the conservation of plant–pollinator interactions. Annual Review of Ecology and Systematics 29: 83112.
  • Kelber A, Balkenius A, Warrant EJ. 2003. Colour vision in diurnal and nocturnal hawkmoths. Integrative and Comparative Biology 43: 571579.
  • Levins R. 1970. Extinction. In: Gerstenhaber, M, ed. Some mathematical problems in biology. Providence, RI, USA: American Mathematical Society, 77107.
  • Linhart YB, Mendenhall JA. 1977. Pollen dispersal by hawkmoths in a Lindenia rivalis Benth. population in Belize. Biotropica 9: 143.
  • McCauley DE. 1994. Contrasting the distribution of chloroplast DNA and allozyme polymorphism among local populations of Silene alba: implications for studies of gene flow in plants. Proceedings of the National Academy of Sciences, USA 91: 81278131.
  • McCauley DE. 1997. The relative contributions of seed and pollen movement to the local genetic structure of Silene alba. Journal of Heredity 88: 257263.
  • McNeill J. 1977. The biology of Canadian weeds. 25. Silene alba (Miller) E. H. L. Krause. Canadian Journal of Plant Science 57: 11031114.
  • Miyake T, Yahara T. 1998. Why does the flower of Lonicera japonica open at dusk? Canadian Journal of Botany 76: 18061811.
  • Moody-Weis JM, Heywood JS. 2001. Pollination limitation to reproductive success in the Missouri evening primrose, Oenothera macrocarpa (Onagraceae). American Journal of Botany 88: 16151622.
  • Mustajärvi K, Siikamäki P, Rytkönen S, Lammi A. 2001. Consequences of plant population size and density for plant–pollinator interactions and plant performance. Journal of Ecology 89: 8087.
  • Pellmyr O, Thompson JN. 1996. Sources of variation in pollinator contribution within a guild: the effects of plant and pollinator factors. Oecologia 107: 595604.
  • Pettersson MW. 1991. Pollination by a guild of fluctuating moth populations: option for unspecialization in Silene vulgaris. Journal of Ecology 79: 591604.
  • Richards CM. 2000a. Inbreeding depression and genetic rescue in a plant metapopulation. American Naturalist 155: 383394.
  • Richards CM. 2000b. Genetic and demographic influences on population persistence: gene flow and genetic rescue in Silene alba. In: YoungAG, ClarkeGM, eds. Genetics, demography and viability of fragmented populations. Cambridge, UK: Cambridge University Press, 271292.
  • Richards CM, Church S, McCauley DE. 1999. The influence of population size and isolation on gene flow by pollen in Silene alba. Evolution 53: 6373.
  • Schmitt J. 1980. Pollinator foraging behavior and gene dispersal in Senecio (Compositae). Evolution 34: 934943.
  • Schulke B, Waser NM. 2001. Long-distance pollinator flights and pollen dispersal between populations of Delphinium nuttallianum. Oecologia 127: 239245.
  • Shykoff JA, Bucheli E. 1995. Pollinator visitation patterns, floral rewards and the probability of transmission of Microbotryum violaceum, a venereal disease of plants. Journal of Ecology 83: 189198.
  • Waddington KE. 1983. Foraging behavior of pollinators. In: RealL, ed. Pollination biology. Orlando, FL, USA: Academic Press, 213241.
  • Waser NM, Chittka L, Price MV, Williams NM, Ollerton J. 1996. Generalization in pollination systems, and why it matters. Ecology 77: 10431060.
  • Werth CR. 1985. Implementing an isozyme laboratory at a field station. Virginia Journal of Science 36: 5376.
  • Witt T, Jürgens A, Geyer R, Gottsberger G. 1999. Nectar dynamics and sugar composition in flowers of Silene and Saponaria species (Caryophyllaceae). Plant Biology 1: 334345.
  • Wright JW, Meagher TR. 2003. Pollination and seed predation drive flowering phenology in Silene latifolia (Caryophyllaceae). Ecology 84: 20622073.
  • Wyatt R. 1983. Pollinator–plant interactions and the evolution of breeding systems. In: RealL, ed. Pollination biology. Orlando, FL, USA: Academic Press, 5195.
  • Young HJ. 2002. Diurnal and nocturnal pollination of Silene alba (Caryophyllaceae). American Journal of Botany 89: 433440.