| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Reproductive Biology |
Department of Biology, Valdosta State University, Valdosta, Georgia 31698 USA
Received for publication March 22, 2006. Accepted for publication January 10, 2007.
ABSTRACT
Natural hybridization plays a critical role in speciation, the maintenance of reproductive isolation, and genetic introgression. While many plant species have hybrid swarms in areas of sympatry, the lack of hybrids among closely related sympatrically distributed species suggests that strong pre- and/or postzygotic barriers exist to hybridization. Gelsemium sempervirens and G. rankinii (Gelsemiaceae) are sympatrically distributed southeastern sister taxa that have strong postzygotic barriers to hybrid formation and high levels of genetic differentiation. In this study, two sympatric populations in Lowndes County, Georgia were surveyed from 1999–2005 to assess the role of temporal and pollinator isolation as potential prezygotic barriers. The populations had mostly non-overlapping flowering periods in 2003–2005, with significant differences in time of peak flowering and length of flowering. Both species shared a similar community of flower visitors, with the apid bee Habropoda laboriosa the dominant visitor to both species. A choice experiment found that H. laboriosa visited both species but preferred G. sempervirens. The primary prezygotic barrier is temporal isolation preventing hybridization in spite of the shared pollinators. This study suggests that reliance on a shared pollinator during speciation may limit opportunity for divergent selection on flowering time.
Key Words: floral isolation • Gelsemiaceae • Gelsemium • hybridization • prezygotic reproductive isolation • sympatry • temporal isolation
The study of hybridization is important in understanding the process of natural speciation, the evolution of domesticated species of plants and animals, and the evolution of pathogens (Barton, 2001
; Hewitt, 2001
; Arnold, 2004
; Coyne and Orr, 2004
; Rundle and Nosil, 2005
). The degree of reproductive isolation among related species is an important factor influencing genetic integrity of a species and the probability of the formation of hybrids (Levin, 1978
; Grant, 1981
; Harrison, 1993
; Avise, 1994
; Arnold, 1997
; Alarcón and Campbell, 2000
; Bohs, 2000
; Chari and Wilson, 2001
; Brock, 2004
; Mráz et al., 2005
; Stökl et al., 2005
). The degree of reproductive isolation through pre- and postzygotic mechanisms is most critical for sympatric species.
In this study, I utilize a pair of taxa from a small monophyletic group [Gelsemium sempervirens (L.) Jaume Saint-Hilaire and G. rankinii Small; Gelsemiaceae (Struwe et al., 1994
)] that occur sympatrically in the southeastern United States; a single related species is found in Southeast Asia [G. elegans (Gardn. & Champ.) Bentham] (Duncan and Dejong, 1964
). Cladistic analysis of eight morphological characteristics of the three taxa in the Gelsemiaceae found that G. sempervirens and G. rankinii are sister taxa (Wyatt et al., 1993
). Based on genetic divergences, the two species split approximately 3–3.5 mya in the late Tertiary, with possible adaptation to new wetland habitats from rising sea levels associated with some of the morphological features found in the riparian G. rankinii (Wyatt et al., 1993
).
The range of G. rankinii (found along the outer coastal plain from Louisiana to North Carolina) is entirely sympatric within the much larger range of G. sempervirens (Duncan and Dejong, 1964
), which can be found further South, North, East, and West of G. rankinii. Within the area of sympatry, both species can be found growing in close physical proximity, with some populations adjoining each other along ecotones. This type of sympatry is neighboring sympatry where populations of each species are found in different niches within the same geographical area (Grant, 1981
). In general, G. sempervirens is found in dry, upland habitats, and G. rankinii occurs in swamps and river floodplains (Ornduff, 1970
). This close proximity could result in pollen movement by wide ranging insects such as larger bees that are known to visit G. sempervirens (Ornduff, 1970
; Adler and Irwin, 2005
; Greenleaf, 2005
).
Interspecific crosses in the greenhouse produced very low fruit set, and the few hybrid seeds that germinated had minimal growth; all but one died before flowering, and the hybrid had reduced fertility (Ornduff, 1970
). There are no known natural hybrids of the two species, and the two species are highly distinct in allozyme profiles, suggestive of a long period of genetic isolation (Duncan and Dejong, 1964
; Wyatt et al., 1993
). These postzygotic barriers (reduced fertility among crosses and reduced germination, survival, and fertility of hybrids) could lead to lower fitness if gene flow is substantial along the ecotones present throughout the area of sympatry. Under these conditions, natural selection should favor the evolution of prezygotic barriers that would limit gene flow (Levin, 1978
; Grant, 1981
; Nagamitsu et al., 2006
). In this study, I explored the following two hypotheses of potential prezygotic isolation:
1. Temporal isolation hypothesis. Differences in when species flower (flowering phenology) are often thought to be a common mechanism preventing gene flow between related species (Armbruster and Herzig, 1984
; McGuire and Armbruster, 1991
; Johnson et al., 1998
; Leebens-Mack and Milligan, 1998
). Although herbarium specimens from across the range of the two Gelsemium species overlap in flowering times, both Duncan and Dejong (1964)
and Wyatt et al. (1993)
commented that localized populations appeared to have less overlap in flowering times, although no quantitative data was presented. Selection against individuals that flower synchronously could result from loss of male and female gametes through heterospecific movement, leading to low fitness either through reduced fruit and seed set and the production of inviable hybrids (Ornduff, 1970
; Levin, 1971
). Because the two species are sister taxa that evolved from a common ancestor (Wyatt et al., 1993
), either disruptive selection during sympatric speciation or adaptation to local microhabitats during allopatric speciation could result in a shift in flowering phenology from the ancestral pattern (Savolainen et al., 2006
).
2. Floral isolation hypothesis. Pollinator movements between parental species are important in creating and maintaining hybrid zones (Melendez-Ackerman et al., 1997
; Bradshaw et al., 1998
; Leebens-Mack and Milligan, 1998
; Emms and Arnold, 2000
; Wesselingh and Arnold, 2000
; Wolf et al., 2001
; Campbell et al., 2002
; Johnson et al., 2005
). Although flowers are similar in shape, size, and color, the two species differ in fragrance, present in G. sempervirens but typically lacking in G. rankinii. Floral fragrance is known to be an important cue in attracting specific types of pollinators (Andersson et al., 2002
; Fenster et al., 2004
; Knudsen et al., 2004
; Stökl et al., 2005
; Johnson et al., 2005
; Miyake and Yafuso, 2005
) and could lead to different suites of floral visitors.
If the two species differ in either phenology or pollinator composition as a prezygotic reproductive isolating mechanism, this could have important consequences on reproductive success through differential effects on pollinator attraction. Because the two species flower during a climatically variable season (late winter/early spring), the species that has shifted toward the earlier flowering time may experience lower visit rates and a lower diversity of pollinators if pollinators are limited by air temperature (Orueta, 2002
; Aizen, 2004
; Alonso, 2004
; Schulz and Zasada, 2005
). Conversely, the early flowering species may benefit from few other flowering species, while the later flowering species may have a larger number of co-flowering species from which it must attract pollinators. Thus, selection to maintain reproductive isolation could lead to changes in reproductive success as a consequence.
MATERIALS AND METHODS
Study species
The Carolina jessamine, G. sempervirens, and the swamp jessamine, G. rankinii, are noted for the presence of toxic alkaloids (Duncan and Dejong, 1964
; Adler and Irwin, 2005
; Irwin and Adler, 2006
). The two species are very similar in vegetative appearance, with the strongest morphological differences found in the flowers and fruits. For example, G. rankinii has unwinged seeds, while G. sempervirens has winged seeds (Duncan and Dejong, 1964
). Both species have a genetically determined distylous breeding system, with two flower morphs: pins (long style) and thrums (short style) (Ornduff, 1970
, 1980
). Highest fruit set occurs between the different morphs of the same species, with much lower fruit set between the same morphs of the same species (Duncan and Dejong, 1964
; Ornduff, 1970
), essentially making the morph self-incompatible.
Study sites
Both species occurred sympatrically at two sites located in Lowndes County, Georgia: (1) YMCA (elevation 43 m a.s.l., location 30°51'36.10'' N, 83°19'11.77'' W). At this locality, G. rankinii occurs in the floodplain of the Withlacoochee River, and G. sempervirens occurs on adjacent uplands. (2) Grand Bay (GB). This site occurs at the intersection of highway 94 and Grand Bay Creek (41 m a.s.l., 30°46'6.56'' N, 83°8'8.56'' W). At this location, G. rankinii occurs along the flood plain of Grand Bay Creek, while G. sempervirens occurs at higher elevations outside of the flood plain. At both sites, the two species occur in close physical proximity (<50 m). Floral visitors and flowering phenology were also observed at a population of G. sempervirens located at the Lake Louise (LL) Field Station of VSU (Lowndes County, Georgia, 53 m a.s.l., 30°43'51.45'' N, 83°15'15.46'' W). Although G. rankinii occurs at this site within a bay swamp, it was difficult to access and was not studied.
At YMCA and GB, the populations of G. rankinii occurred on hydric soils of the Johnston Loam series. These soils are frequently flooded and ponded for up to 7 mo of the year, from November to May, with poor drainage and high water capacity. In contrast, the two populations of G. sempervirens occurred on nonhydric soils of the Tifton (YMCA) and Chiplay (GB) series, sandy loam soils with low to moderate water capacity, rapid to moderate permeability, lower water tables, and no flooding or ponding at any time (USDA 2006b
). Soil pH ranges from 4.5–5.5 for the Johnston soils and 3.6–6.0 and 4.5–6.0 in Tifton and Chiplay series, respectively.
Flowering phenology
Flowering phenology was examined regionally using herbarium records and locally using field sites. Valdosta State University herbarium collections of G. sempervirens and G. rankinii were examined to determine regional (southern Georgia and northern Florida) flowering periods for the two species in the area of sympatry. At the local level, the level of flowering for each study population was monitored in permanent transects of at least 600 m2 in size from 2003–2005 at YMCA and GB and from 1999–2000 and 2003–2005 at LL. Every 2 wk, I counted the number of open flowers. Because of differences in population structure, the shape and number of transects varied among populations. Because total area sampled varied among populations, the percentage of total flower production was used to estimate level of flowering. I compared the overlap in flowering intensity between and within species and years using Pianka's niche overlap index O and compared the observed pattern to a random pattern using the program EcoSim 7.0 (Krebs, 1989
; Gotelli and Entsminger, 2001
). I also compared the length of flowering, peak day of flowering, and day of 95% completed flowering between the two species using paired t tests (SigmaStat, version 2.03, SYSTAT Software, Inc., Richmond, California, USA) for each site across all years.
Observations of flower visitors
For both species, I recorded flower visitors during multiple 10-min observations. The majority of observations were made in 1999–2000 at YMCA, GB, and LL with additional observations from 2001–2005. During the observation, a randomly selected group of flowers was watched by an observer who recorded the number of visits, the number of flowers visited, and identity of the visitors. A visit consisted of entry into the focal group of flowers and visiting at least one of the focal group flowers. Observations were made during daylight hours only, between 0830–1730 hours. Air temperature was recorded using a Kestrel (Nielsen-Kellerman, Boothwyn, Pennsylvania, USA) handheld recorder.
I used Pianka's niche overlap index (Krebs, 1989
) to compare similarity of pollinators between years and species and used EcoSim 7.0 (Gotelli and Entsminger, 2001
) to evaluate the significance of this index value vs. a random distribution. Data on the sex of the bee Habropoda laboriosa visiting Gelsemium was also recorded when visible (males have a bright white clypeus on their face, while females are all black). Sex of the bee was of interest because male H. laboriosa forage earlier in the year than females (Cane, 1994
) and sex ratio of the pollinators may vary as a function of flowering time of the two Gelsemium species.
I examined the relationship between air temperature and visit rate for each species separately using linear regression. Data for this analysis was a subset of the total pollinator visit data consisting of data from 25 February and 4 and 11 March 1999 for G. sempervirens and 8 April 1999 for G. rankinii. This subset was used because it included multiple observers over the entire day, yielding a large sample, and had the greatest temperature range that may influence insect activity. Data on hourly daytime (0700–1830 hours) air temperatures during 2003–2005 were obtained from the National Climatic Data Center (2005)
. I summed the total numbers of hours above 16°C during each week of flowering and multiplied this number times the relative proportion of flowers open. I compared the total time available for pollination between the species using a paired t test within the site across all years.
Choice experiment
In 2004, potted plants of both species were purchased from local nurseries and maintained in the VSU greenhouse. In the choice experiment, I presented either one (three trials) or two plants (one trial) of G. rankinii and G. sempervirens to wild foraging bees. Plants were 1 m apart, different plants were used for each trial, and their position was rotated every 15 min during the trial. Trials were done under natural outdoor conditions on 11, 12, 14, and 18 March at VSU, LL, and the author's residence in suburban Valdosta, all sites with large populations of the bee Habropoda laboriosa (Pascarella, in press
).
In each trial, the sex and identity of the visitor was noted, the species of the plant it visited, the number of flowers available, and the total number of flowers visited was recorded. Observations were made from 0946 to 1700 hours. A visit consisted of a single bee entering the array and visiting at least one flower. The visit ended when the bee left the array and did not immediately return (while being observed). I examined visits to the two species with an assumption of equal attractiveness for number of visits and species first visited. For the one choice test that used two plants of each species, I compared movement patterns within a visit using association tests. I calculated the proportion of flowers visited to flowers available (using arcsine square-root transformed data). Summed across all trials, I calculated the preference index Manly's alpha (Krebs, 1989
). This index compares the ratio of utilized resources (flowers visited) to available resources (total flowers present). With two resource types, values >0.5 indicate preference and <0.5 indicate avoidance. I used a resampling method to calculate confidence intervals to assess the statistical significance of the values (PopTools, 2006, version 2.6, http://www.cse.csiro.au/poptools). Each index was resampled 10 000 times and the 250th and 9750th values were used to generate lower and upper 95% confidence intervals. Non-overlapping values were taken to indicate significance at P < 0.05.
RESULTS
Flowering phenology
Twenty-five nonduplicate herbarium records from the VSU herbarium spanning the years 1962–1992 of plants from southern Georgia and northern Florida indicated that G. sempervirens flowered from 16 January to 31 March (N = 16 specimens from 13 counties) and G. rankinii flowered from 28 March to 14 April (N = 19 specimens from 16 counties). There were no collections of both species during the same year within the same county. At the local level in 2003–2005, similar flowering patterns were found, with flowering of G. sempervirens typically starting in mid-February, peaking in late February to early March, and finishing by mid to late March (Fig. 1). In all 3 years, G. sempervirens initiated flowering prior to G. rankinii, and the peak bloom was not simultaneous between the species at any site or year. Of the six combinations of years and sites, 2 years had no niche overlap, while there was minimal overlap (0.3–5.3%) in the remaining four. Four of the six combinations had significantly less overlap than expected by chance (P < 0.05), while the remaining two were nearly statistically significant (P < 0.06). Flowering phenology of G. sempervirens at LL was similar to that observed for G. sempervirens at GB and YMCA (J. B. Pascarella, unpublished data).
|
Observations of flower visitors
A total of 379 h of observations were made from 1999–2005 over 55 d (Table 1). For both species, the majority of floral visits were made by blueberry bees (Habropoda laboriosa), but a variety of other visitors were observed (Table 2). Male blueberry bees made up 50% of all visits to G. sempervirens but were absent during the flowering period of G. rankinii. Pianka's index of similarity was greater than 0.96 (most were >0.99) for all within-year cross-species comparisons and for all cross-year within-species comparisons, showing a high similarity in the flower-visiting fauna. The observed similarity indices were significantly greater than expected by chance for all comparisons (P < 0.05).
|
|
|
|
Seventy-eight percent of all visits (N = 119) were to a single species, with a significant preference to G. sempervirens (65%) over G. rankinii (35%) (
2 = 10.3, df = 1, P < 0.01). Both male and female bees visited both species of Gelsemium, with females visiting G. sempervirens (66%) more often than G. rankinii (34%) (2 = 10.5, df = 1, P < 0.01). Males were more strongly biased in favor of G. sempervirens (71% vs. 29%, but not significant due to small sample size). For single species visits, females visited a greater proportion of flowers than did males for G. sempervirens (Mann–Whitney rank sum test, P < 0.001), but there was no significant difference between the sexes in the proportion of G. rankinii flowers visited. For the 22% of visits to both species, there was no significant difference in the first species visited for all bees, males, or females. For G. sempervirens, females visited a greater proportion of flowers than did males (Mann–Whitney rank sum, P < 0.001), but there was no significant difference between the sexes in the proportion of G. rankinii flowers visited.
For movement patterns within a visit (summed across all visits), males and females differed in their movement patterns (G = 12.8, df = 3, P < 0.01) with males tending to move between the two species more often than females (76 vs. 33%, respectively). However, there was no difference in movement patterns between the two Gelsemium species (35% within same species, 65% to other species). The preference index Manly's alpha indicated a significant preference for G. sempervirens (0.56, 0.49–0.63) over G. rankinii (0.42, 0.34–0.48) (mean, lower and upper 95% CI).
DISCUSSION
Temporal isolation hypothesis
This study provides quantitative support for the hypothesis that G. sempervirens and G. rankinii differ in flowering phenology. Both the herbarium data spanning 30 yr from the area of regional sympatry in southern Georgia and northern Florida and the two local sympatric populations of both Gelsemium sp. in Lowndes County, Georgia from 2003–2005 found only a minimal overlap in flowering period, with one site having no overlap during two flowering seasons. In addition, when flowering intensity is taken into account, the overlap occurs during a period of relatively little flowering for either species. This pattern differs significantly from random, suggesting that flowering phenology is an important difference between the two species. Thus, temporal isolation is a strong prezygotic isolation mechanism that limits the potential for interspecific pollen flow.
In the area of sympatry, the divergence in flowering time is likely limited by both abiotic and biotic factors. During the months of January and February, periods of freezing weather and daytime temperatures less than 16°C limit insect activity (Fig. 2) and increase the risk of flower damage, restricting the shift to earlier flowering for G. sempervirens (National Climatic Data Center, 2005
). In contrast, warmer temperatures present during the later flowering period of G. rankinii usually do not present limitations to insect flight (Fig. 2) or threaten flowers. The key biotic factor may be the reliance on the bee Habropoda laboriosa for pollination. Divergent selection on flowering time may be limited to that spanning the normal emergence and flight period of this bee species. H. laboriosa is a strictly vernal species, usually emerging in February and completing its life cycle by early April (Cane, 1994
; Cane and Payne, 1988
). Flowering too early could result in few visits if H. laboriosa has not yet emerged. On the other end, shifts to flowering later in the year may be constrained by the lack of sufficient number of H. laboriosa later than mid-April. It is possible that worker bumblebees, observed foraging on G. rankinii in 2005, may play an important role some years in pollination of G. rankinii although they did not comprise a large proportion of observed visits during this study. The lack of floral odor in G. rankinii, coupled with the denser vegetation present by late April and early May that may obscure the visual signal, may potentially limit the attractiveness of this species to other potential large bodied bees suitable as pollinators (e.g., Bombus, Xylocopa, Osmia [Adler and Irwin, 2005
]).
As shown by the meteorological data, most insect visitors are limited to air temperatures above 16°C. Adler and Irwin (2005)
found that pollinator visitation rate to G. sempervirens in the Georgia Piedmont was much lower during the colder spring of 2004 (0.0007 probes per flower per 10 min) compared to a warmer period in 2002 (0.032). They also noted that most visits were during the midday at 1000–1300 hours, likely reflecting similar temperature limitations early in the morning. Comparing visitation rates in the outer coastal plain to the Piedmont, I found pollinator visitation rates in this study to be considerably higher (0.07–0.43), reflecting either the warmer temperatures found in the outer coastal plain and/or higher insect abundance.
However, when total length of flowering and the timing of peak flowering were taken into account, there was no significant difference in time available for pollination between the two species. Because G. sempervirens has a longer flowering period than G. rankinii, this increases the chance that flowers are open during climatically favorable conditions for activity of their pollinators. In general, most winter flowering species tend to have long flowering periods relative to summer flowering species, which is related to the low probability of pollinator visits during inclement weather possible during the winter (Orueta, 2002
; Schulz and Zasada, 2004; Alonso, 2005). Based on diversity of floral visitors and high visit rates when temperatures are suitable for flight, there appears to be minimal negative consequence for the earlier flowering G. sempervirens.
In addition to minimizing heterospecific pollen contamination in areas of sympatry, the evolution of flowering phenology in Gelsemium may be related to competition with blueberries (Vaccinium sp.; Ericaceae) for the services of H. laboriosa. Blueberries are an important pollen source of H. laboriosa (Cane, 1994
; Cane and Payne, 1988
, 1993
). In central Florida, female and male H. laboriosa were noted visiting G. sempervirens, while females also visited V. corymbosum (Deyrup et al., 2002
). In southern Georgia, V. corymbosum (sensu Vander Kloet, 1980
) flowers from January to April (J. B. Pascarella, unpublished data), overlapping the flowering period of both Gelsemium species. In 10 bees caught foraging on either Vaccinium or Gelsemium and in day 1 stigmas of both Gelsemium species (10 stigmas each), I found mixed Gelsemium/Vaccinium pollen loads in nearly every bee and stigma examined (J. B. Pascarella, in press). This suggests that H. laboriosa is a frequent visitor to both Vaccinium and Gelsemium flowers where they co-occur.
The mechanism, whether environmental, genetic or a combination of the two, that regulates the onset of flowering in both species is unknown. It is clear that the habitat differences of the two species create different soil moisture regimes. Reciprocal transplant, common garden, and environmental measurements of soil moisture and temperature across the habitat gradient are needed to assess the relative roles of genetic and environmental factors on the timing of flowering. In the northern part of its range in the piedmont of Georgia, South Carolina, and North Carolina where it is not sympatric with G. rankinii, G. sempervirens flowers considerably later during late March to mid-April (Duncan and Dejong, 1964
; Ornduff, 1970
; Adler and Irwin, 2005
), the normal flowering time of G. rankinii in the coastal plain. Whether this shift is due to a simple relationship with local climate (climate zone 8b vs. zone 7b, USDA, 2006a
) or to a relaxation of selection pressures due to the absence of G. rankinii remains to be elucidated.
Pollinator isolation hypothesis
The hypothesis that lack of fragrance production in G. rankinii would result in different floral visitors was not supported as both Gelsemium species had a similar group of flower visitors. For G. sempervirens, floral visitors are highly consistent across the range of this species, with the bees Habropoda laboriosa, Bombus sp., Xylocopa sp., Apis mellifera, and Osmia sp. and with butterflies as visitors to G. sempervirens. This is the first study to report floral visitors to G. rankinii.
The observation that H. laboriosa is the most common floral visitor in North Carolina (Ornduff, 1970
), a common visitor in the Piedmont of Georgia (Adler and Irwin, 2005
, 2006
) present at the southern edge of its range in central Florida (Deyrup et al., 2002
) and the most common visitor in this study, suggests that this species plays a important role in pollination of both Gelsemium species. In a recent study, Adler and Irwin (2005)
found that H. laboriosa, present in both years of their census of floral visitors, was the most common visitor in 2004, was ranked as a good pollen carrier, and was one of the most efficient floral visitors, spending little time per flower (Adler and Irwin, 2006
). They found that G. sempervirens can be pollinated by a wide range of medium to large bees, consistent with theory on the generality of most plant–pollinator interactions (Thompson, 1994
; Waser et al., 1996
). However, given the very high relative proportion of flowers visited and the consistency of visitation across sites and years, H. laboriosa is the most important local visitor to both Gelsemium species in the outer coastal plain.
As expected, male bees, which fly earlier than females, were present on G. sempervirens but absent on G. rankinii. This differential abundance may be of importance if male bees are less effective pollinators than females. However, a preliminary study found that fruit set of G. sempervirens was not significantly related to either the gender of the visitor or the number of visits (J. B. Pascarella, unpublished data). However, if male bees have different foraging patterns than females, this could lead to changes in gene flow and seed set within fruits.
Wyatt et al. (1993)
proposed that G. sempervirens evolved floral odor during its split from its ancestor based on the lack of floral odor in G. rankinii and the Asian G. elegans, an evolutionary sequence that is uncommon. I propose that the evolution of floral odor in G. sempervirens is related to the use of the winter flowering season and the need to attract naïve pollinators that are just emerging from their nests. In areas of sympatry, by the time G. rankinii blooms, long-lived individual blueberry bees are already familiar with the floral morphology of Gelsemium due to interactions with G. sempervirens early in the year. Thus, there may be reduced selection pressures for fragrance production in G. rankinii. This type of sequential mutualism (Waser and Real, 1979
) where species share a common pollinator may also favor selection pressures that maintain similar floral morphologies between the two species.
Pollinator choice experiment
Although H. laboriosa preferred to visit G. sempervirens, both Gelsemium species were visited and some individual bees visited both species during a single visit. Thus, if both species flowered together, the potential exists for gene flow between the two species, because bees were more likely to switch species than to continue foraging on the same species. In studies that have examined the role of pollination in hybrid formation, pollinators typically move between parental species and hybrids, although often having preferences and varying in visitation rate (Leebens-Mack and Milligan, 1998
; Wesselingh and Arnold, 2000
; Wolf et al., 2001
; Campbell et al., 2002
). For species where hybrid formation is limited by postzygotic barriers, pollinator movement may result in reduced fitness. For example, Miyake and Yafuso (2005)
noted that two species of flies pollinated two sympatric species of Colocasia, leading to a reduction in reproductive success in these species. However, experimental results in a similar system in the Arctic that examined two sympatric Saxifraga species that differed in flowering periods (McGuire and Armbruster, 1991
) had no negative effect on fruit set in the earlier flowering species when it was experimentally made to flowering synchronously and in close proximity to the later flowering species. Thus, experiments that manipulate the flowering time of the two Gelsemium species are needed to determine the consequence of overlapping flowering periods in terms of potential reduced fitness due to interspecific pollen flow, stigma clogging, and hybrid inviability.
In conclusion, the maintenance of reproductive isolation between G. sempervirens and G. rankinii is through a combination of prezygotic and postzygotic barriers. This combination of barriers explains the lack of natural hybrids and the strongly differentiated allozyme profiles. For many other species pairs or complexes that currently occur in sympatry, hybrid formation is usually not as strongly limited (Chari and Wilson, 2001
; Wolf et al., 2001
; Campbell et al., 2002
). In many of these studies, species pairs involved in hybridization occur in the same habitat.
Specialization to distinct soil types may be involved in the evolution of reproductive barriers that limit hybrid formation. Savolainen et al. (2006)
proposed that adaptation to different soil types with contrasting pH levels led to diversification of an ancestral species approximately 4.5–5.5 mya in two sympatric Howea plam species on the remote Oceanic Lord Howe Island in the Pacific. The two Howea species diverge in their flowering times on the two soil types even when close geographically and sharing a similar pollination system (wind pollination), similar to the observations noted in this study. In the case of Gelsemium, Wyatt et al. (1993)
suggested that, during periods of high sea levels, greater amounts of riparian and swamp habitats appeared on the landscape and produced selection pressures that led to the diversification of the common ancestor of G. rankinii and G. sempervirens.
Once populations were isolated, the morphological features of G. rankinii such as the loss of the winged seed (found in both G. sempervirens and G. elegans), are likely related to adaptation to the seasonally flooded habitat. The loss of the winged seed is important because the single hybrid that survived to flower produced during a cross of the two species had winged seeds (Ornduff, 1970
). Winged seeds should be strongly selected against because they could lead to the movement of seeds away from the narrow, linear wetland habitats that G. rankinii prefers. In addition, plants that live in flooded environments have a wide range of molecular and cellular adaptations to flooding stress that are not found in related upland species (Subbiah and Sachs, 2003
), leading to further selection against hybrids.
FOOTNOTES
1 The author thanks Valdosta State University (VSU) undergraduates D. Alderman, D. Down, R. Kanhai, S. Lumpkin, N. Nelson, J. Thompson, A. Waldron, and S. Wright and the students in the spring 1999 Ecology course for assistance with data collection; J. Cane for suggesting studying the interaction between Habropoda laboriosa and Gelsemium; R. Carter for use of the VSU herbarium records and assistance in locating field populations; and J. Zimmerman, S. Scheiner, and several anonymous reviewers for comments on an earlier version of this manuscript. Financial support was provided by a VSU Faculty Research Grant and a grant from the Georgia Native Plant Society. 
2 Reprint requests (e-mail: jbpascar{at}valdosta.edu
)
LITERATURE CITED
Adler L. S. Irwin R. E.. 2005. Ecological costs and benefits of defenses in nectar. Ecology 86: 2968-2978.
Adler L. S. Irwin R. E.. 2006. Comparison of pollen transfer dynamics by multiple floral visitors: experiments with pollen and fluorescent dyes. Annals of Botany 97: 141-150.
Aizen M. A.. 2004. Influences of animal pollination and seed dispersal on winter flowering in a temperate mistletoe. Ecology 84: 2613-2627.
Alarcón R. Campbell D. R.. 2000. Absence of conspecific pollen advantage in the dynamics of an Ipomopsis (Polemoniaceae) hybrid zone. American Journal of Botany 87: 819-824.
Alonso C.. 2004. Early blooming's challenges: extended flowering season, diverse pollinator assemblage and the reproductive success of gynodioecious Daphne laureola. Annals of Botany 93: 61-66.
Andersson S. Nilsson L. A. Groth I. Bergström G.. 2002. Floral scents in butterfly-pollinated plants: possible convergence in chemical composition. Botanical Journal of the Linnean Society 140: 129-153.
Armbruster W. S. Herzig A. L.. 1984. Partitioning and sharing of pollinators by four sympatric species of Dalechampia (Euphorbiaceae) in Panama. Annals of the Missouri Botanical Garden 71: 1-16.
Arnold M. L.. 1997. Natural hybridization and evolution Oxford University Press, Oxford, UK.
Arnold M. L.. 2004. Natural hybridization and the evolution of domesticated, pest and disease organisms. Molecular Ecology 13: 997-1007.[CrossRef][Medline]
Avise J. C.. 1994. Molecular markers, natural history, and evolution Chapman and Hall, New York, New York, USA.
Barton N. H.. 2001. The role of hybridization in evolution. Molecular Ecology 10: 551-568.[CrossRef][Medline]
Bohs L.. 2000. Insights into the Witheringia solanacea (Solanaceae) complex in Costa Rica. II. Insect visitors and pollination biology of W. asterotricha and W. meiantha. Biotropica 32: 80-89.
Bradshaw H. D. Jr. Otto K. G. Frewen B. E. McKay J. K. Schemske D. W.. 1998. Quantitative trait loci affecting differences in floral morphology between two species of monkey-flower (Mimulus). Genetics 149: 367-382.
Brock M. T.. 2004. The potential for genetic assimilation of a native dandelion species, Taraxacum ceratophorum (Asteraceae), by the exotic congener T. officinale. American Journal of Botany 91: 656-663.
Campbell D. R. Waser N. M. Pederson G. T.. 2002. Predicting patterns of mating and potential hybridization from pollinator behavior. American Naturalist 159: 438-450.
Cane J. H.. 1994. Nesting biology and mating behavior of the southeastern blueberry bee, Habropoda laboriosa (Hymenoptera: Apoidea). Journal of the Kansas Entomological Society 67: 236-241.
Cane J. H. Payne J. A.. 1988. Foraging ecology of the bee, Habropoda laboriosa (Hymenoptera: Anthophoridae), an oligolege of blueberries (Ericaceae: Vaccinium) in the southeastern United States. Annals of the Entomological Society of America 81: 419-427.
Cane J. H. Payne J. A.. 1993. Regional, annual, and seasonal variation in pollinator guilds: intrinsic traits of bees (Hymenoptera: Apoidea) underlie their patterns of abundance at Vaccinium ashei (Ericaceae). Annals of the Entomological Society of America 86: 577-588.
Chari J. Wilson P.. 2001. Factors limiting hybridization between Penstemon spectabilis and Penstemon centranthifolius. Canadian Journal of Botany 79: 1439-1448.
Coyne J. A. Orr H. A.. 2004. Speciation Sinauer, Sunderland, Massachusetts, USA.
Deyrup M. Edirisinghe J. Norden B.. 2002. The diversity and floral hosts of bees at the Archbold Biological Station, Florida (Hymenoptera: Apoidea). Insecta Mundi 16: 87-120.
Duncan W. H. Dejong D. W.. 1964. Taxonomy and heterostyly of North American Gelsemium (Loganiaceae). Sida 1: 346-357.
Emms S. K. Arnold M. L.. 2000. Site-to-site differences in pollinator visitation patterns in a Louisiana iris hybrid zone. Oikos 91: 568-578.
Fenster C. B. Armbruster W. S. Wilson P. Dudash M. R. Thomson J. D.. 2004. Pollination syndromes and floral specialization. Annual Review of Ecology, Evolution, and Systematics 35: 375-403.
Gotelli N. J. Entsminger G. L.. 2001. EcoSim: null models software for ecology, version 7.0. Acquired Intelligence Inc & Kesey-Bear, Jericho, Vermont, USA. Website http://homepages.together.net/gentsmin/ecosim.htm [accessed February 2006].
Grant V.. 1981. Plant speciation Columbia University Press, New York, New York, USA.
Greenleaf S.. 2005. Local-scale and foraging-scale habitats affect bee community abundance, species richness, and pollination services in northern California Ph.D. dissertation, Princeton University, Princeton, New Jersey, USA.
Harrison R. G., [ed]. 1993. Hybrid zones and the evolutionary process Oxford University Press, Oxford, UK.
Hewitt G. M.. 2001. Speciation, hybrid zones, and phylogeography—or seeing genes in space and time. Molecular Ecology 10: 537-549.[CrossRef][Medline]
Irwin R. E. Adler L. S.. 2006. Correlations among traits associated with herbivore resistance and pollination: implications for pollination and nectar robbing in a distylous plant. American Journal of Botany 93: 64-72.
Johnson S. A. Bruederle L. P. Tomback D. F.. 1998. A mating system condundrum: hybridization in Apocynym (Apocynaceae). American Journal of Botany 85: 1316-1323.
Johnson S. D. Steiner K. E. Kaiser R.. 2005. Deceptive pollination in two subspecies of Disa spathulata (Orchidaceae) differing in morphology and floral fragrance. Plant Systematics and Evolution 255: 87-98.
Knudsen J. T. Tollsten L. Groth I. Bergström G. Raguso R. A.. 2004. Trends in floral scent chemistry in pollination syndromes: floral scent composition in hummingbird-pollinated taxa. Botanical Journal of the Linnean Society 146: 191-199.
Krebs C. J.. 1989. Ecological methodology Harper and Row, New York, New York, USA.
Leebens-Mack J. Milligan B. G.. 1998. Pollination biology in hybridizing Baptisia (Fabaceae) populations. American Journal of Botany 85: 500-507.[Abstract]
Levin D. A.. 1971. The origin of reproductive isolating mechanisms in flowering plants. Taxon 20: 91-113.
Levin D. A.. 1978. The origin of isolating mechanisms in flowering plants. In M. K. Hecht, W. C. Steere, B. Wallace [eds.], Evolutionary biology, vol. 11 185-317 Plenum Press, New York, New York, USA.
McGuire A. D. Armbruster W. S.. 1991. An experimental test for reproductive interactions between two sequentially blooming Saxifraga species (Saxifragaceae). American Journal of Botany 78: 214-219.
Melendez-Ackerman E. Campbell D. R. Waser N. M.. 1997. Hummingbird behavior and mechanisms of selection on flower color in Ipomopsis. Ecology 78: 2532-2541.
Miyake T. Yafuso M.. 2005. Pollination of Alocasia cucullata (Araceae) by two Colocasiomyia flies known to be specific pollinators for Alocasia odora. Plant Species Biology 20: 201-208.
Mráz P. Chrtek J. Fehrer J. Pla
ková I.. 2005. Rare recent natural hybridization in Hieracium s. str.—evidence from morphology, allozymes, and chloroplast DNA. Plant Systematics and Evolution 255: 177-192.
Nagamitsu T. Kawahara T. Kanazashi A.. 2006. Pollen-limited production of viable seeds in an endemic dwarf birch, Betula apoiensis, and incomplete reproductive barriers to a sympatric congener, B. ermanii. Biological Conservation 129: 91-99.
National Climatic Data Center.. 2005. NOAA satellite and information service Website http://www.ncdc.noaa.gov/ [accessed February 2006].
Ornduff R.. 1970. The systematics and breeding system of Gelsemium (Loganiaceae). Journal of the Arnold Arboretum 51: 1-17.
Ornduff R.. 1980. The probable genetics of distyly in Gelsemium sempervirens (Loganiaceae). Canadian Journal of Genetics and Cytology 22: 303-304.
Orueta D.. 2002. Thermal relationships between Calendula arvensis inflorescences and Usia aurata bombyliid flies. Ecology 83: 3073-3085.
Pascarella J. B. In press Foraging patterns of the southeastern blueberry bee Habropoda laboriosa (Apidae, Hymenoptera): implications for understanding oligolecty. Journal of Apicultural Research..
Rundle H. D. Nosil P.. 2005. Ecological speciation. Ecology Letters 8: 336-352.[Medline]
Savolainen V. Anstett M. C. Lexer C. Hutton I. Clarkson J. J. Norup M. V. Powell M. P. Springate D. Salamin N. Baker W. J.. 2006. Sympatric speciation in palms on an oceanic island. Nature 441: 210-213.[CrossRef][Medline]
Schulz K. Zasada J.. 2005. Annual, local, and individual variation in the inflorescence and fruit production of eastern leatherwood (Dirca palustris L. Thymelaceae). Journal of the Torrey Botanical Society 131: 292-304.
Stökl J. Paulus H. Dafni A. Schulz C. Francke W. Ayasse M.. 2005. Pollinator attracting odour signals in sexually deceptive orchids of the Ophyrs fusca group. Plant Systematics and Evolution 254: 105-120.
Struwe L. Albert V. A. Bremer B.. 1994. Cladistics and family level classification of the Gentianales. Cladistics 10: 175-205.
Subbiah C. Sachs M. M.. 2003. Molecular and cellular adaptations of maize to flooding stress. Annals of Botany 90: 119-127.
Thompson J. N.. 1994. The coevolutionary process University of Chicago Press, Chicago, Illinois, USA.
[USDA] United States Department of Agriculture.. 2006a. USDA plant hardiness zone map. Website http://www.usna.usda.gov/Hardzone/hzm-se1.html [accessed February 2006].
[USDA] United States Department of Agriculture.. 2006b. Web soil survey Website http://websoilsurvey.nrcs.usda.gov/app/WebSoilSurvey.aspx [accessed February 2006].
Vander Kloet S. P.. 1980. The taxonomy of the highbush blueberry, Vaccinium corymbosum. Canadian Journal of Botany 58: 1187-1201.
Waser N. M. Chittka L. Price M. V. Williams N. M. Ollerton J.. 1996. Generalization in pollination systems, and why it matters. Ecology 77: 1043-1060.
Waser N. M. Real L. A.. 1979. Effective mutualism between sequentially flowering plant species. Nature 281: 670-672.
Wesselingh R. A. Arnold M. L.. 2000. Pollinator behavior and the evolution of Louisiana hybrid zones. Journal of Evolutionary Biology 13: 171-180.
Wolf P. G. Campbell D. R. Waser N. M. Sipes S. D. Toler T. R. Archibald J. K.. 2001. Tests of pre- and postpollination barriers to hybridization between sympatric species of Ipomopsis (Polemoniaceae). American Journal of Botany 88: 213-219.
Wyatt R. Broyles S. B. Hamrick J. L. Stoneburner A.. 1993. Systematic relationships within Gelsemium (Loganiaceae): evidence from isozymes and cladistics. Systematic Botany 18: 345-355.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
