| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ecology |
Department of Biology, University of Miami, 1301 Memorial Dr., Miami, Florida 33143 USA
Received for publication October 5, 2005. Accepted for publication May 1, 2006.
ABSTRACT
The floral traits of plants with specialized pollination systems both facilitate the primary pollinator and restrict other potential pollinators. To explore interactions between pollinators and floral traits of the genus Burmeistera, I filmed floral visitors and measured pollen deposition for 10 species in six cloud forest sites throughout northern Ecuador. Nine species were primarily bat-pollinated (84–100% of pollen transfer); another (B. rubrosepala) was exclusively hummingbird-pollinated. According to a principal components analysis of 11 floral measurements, flowers of B. rubrosepala were morphologically distinct. Floral traits of all species closely matched traditional ornithophilous and chiropterophilous pollination syndromes; flowers of B. rubrosepala were bright red, lacked odor, opened in the afternoon, and had narrow corolla apertures and flexible pedicels, which positioned them below the foliage. Flowers of the bat-pollinated species were dull-colored, emitted odor, opened in the evening, and had wide apertures and rigid pedicels, which positioned them beyond the foliage. Aperture width appeared most critical to restricting pollination; hummingbirds visited wide flowers without contacting the reproductive parts, and bats did not visit the narrow flowers of B. rubrosepala. Aperture width may impose an adaptive trade-off that favors the high degree of specialization in the genus. Other floral measurements were highly variable amongst bat-pollinated species, including stigma exsertion, calyx lobe morphology, and pedicel length. Because multiple species of Burmeistera often coexist, such morphological diversity may reduce pollen competition by encouraging pollinator fidelity and/or spatially partitioning pollinator's bodies.
Key Words: chiropterophily • cloud forests • Ecuador • floral morphology • generalization vs. specialization • Lobelioideae • ornithophily • pollination syndromes
In tropical rainforests, 98% of angiosperms are estimated to be animal-pollinated (Bawa, 1990
). This poses a fundamental problem for tropical plants: How can conspecific pollen transfer be ensured in the midst of the high diversity of heterospecific flowers? The selective pressure to maximize conspecific pollen transfer is thought to have led to the evolution of various mechanisms of partitioning the pollinator resource, including divergence in flowering time (e.g., Stone et al., 1998
), differential pollen placement on the bodies of pollinators (e.g., Armbruster et al., 1994
), and specialization on different types of pollinators (e.g., Sargent and Otto, 2006
).
Since Darwin (1862)
, biologists have been documenting floral adaptations that facilitate specialization on pollinators. Extensive comparative studies across angiosperm families have revealed suites of floral traits, or pollination syndromes, that correspond to different types of pollinators (Baker, 1961
; Pijl, 1961
; Stebbins, 1970
). However, recent ecological studies suggest that flowers are often visited by diverse assemblages of animals (Herrera, 1996
; Ollerton, 1996
; Waser et al., 1996
). How can ecological generalization be so widespread and yet floral traits suggest extensive evolutionary specialization?
This apparent paradox becomes less surprising when the pool of animals observed to visit a flower is broken down based on the selective pressures they exert on floral phenotype. As a first step, species that are functionally equivalent as pollinators can be grouped together (e.g., bats vs. bees vs. birds; Armbruster et al., 2000
). Because of differences in behavior and morphology, often only a subset of the functional groups that visit a flower actually pollinate it. And of this subset, different groups can vary greatly in their effectiveness as pollinators (Schemske and Horvitz, 1984
) and in the selective pressures they exert on floral form (Wilson and Thomson, 1996
; Aigner, 2001
; Thompson, 2001
). Some may actually decrease a plant's fitness by "wasting" large amounts of pollen or by decreasing visits of more effective pollinators (Thomson, 2003
). Given such variation between pollinators, it is understandable that selective pressures would favor floral specialization on only one functional group out of all of the animals that visit a flower.
Floral traits may function not only to facilitate pollination by the primary pollinator but also to restrict other potential pollinators. Such traits may represent adaptations to prevent ineffective pollinators from "wasting" pollen that would be better transferred by the primary pollinator (Thomson, 2003
). For example, red coloration in hummingbird-pollinated flowers may evolve primarily to discourage visitation by bees and thus conserve pollen for hummingbird visits (Raven, 1972
). In support of this idea, Schemske and Bradshaw (1999)
found that a shift to red coloration in Mimulus flowers had no effect on hummingbird visitation but decreased bee visitation by 80%. Traits that restrict pollinators may also represent adaptive trade-offs; that is, in facilitating pollination by one type of pollinator, they sacrifice pollination by another. For example, the narrow tubes of hummingbird-pollinated flowers may be selected for primarily to improve the "fit" between flowers and hummingbirds and, in the process, trade off bee pollination by preventing bees from accessing the nectar (Temeles et al., 2002
; Aigner, 2004
; Castellanos et al., 2004
). Regardless of their evolutionary origins, understanding how such floral traits restrict the potential pollinator pool is critical to understanding specialization in pollination systems.
Here I explore the role of pollinators in the evolution of floral phenotype for the genus Burmeistera. While previous research indicates that bats and hummingbirds visit Burmeistera flowers, the effectiveness of hummingbirds as pollinators remains unclear (Feinsinger and Colwell, 1978
; Stein, 1992
; Muchhala and Jarrin-V., 2002
; Muchhala, 2003
). In this study I filmed flowers to identify all floral visitors and measured pollen deposition on stigmas to quantify the effectiveness of visitors as pollinators. Motivating questions include (1) For each species of Burmeistera, what percentage of pollen flow is attributable to each functional group of pollinator (bats vs. hummingbirds)? (2) How do floral traits (morphology, exposure, timing of anthesis, color, and odor) relate to these pollination systems? (3) Which of these traits appear to be most important in restricting the potential pollinators?
MATERIALS AND METHODS
Study sites and study organisms
I studied 10 species of Burmeistera in six cloud forest reserves throughout the northern Andes of Ecuador. The reserves varied in elevation from 300–2600 m a.s.l., and each contained 3–4 species of Burmeistera (Table 1). Fieldwork was carried out from March through November 2003. In total, 15 species were present in the study sites. Voucher collections for each were made and deposited in the herbarium of the Pontificia Universidad Católica del Ecuador. Pollination studies were carried out on those species abundant enough to allow adequate sample sizes, which included: B. borjensis Jeppesen, B. ceratocarpa Zahlbr., B. cyclostigmata Donn. Sm., B. cylindrocarpa Zahlbr., B. lutosa E. Wimm., B. multiflora Zahlbr., B. rubrosepala (E. Wimm.) E. Wimm, B. smaragdi Lammers, B. sodiroana Zahlbr., and B. succulenta H. Karst.
|
), with a center of diversity in the cloud forests of the Andes of Colombia and Ecuador. In Ecuador alone 35 species occur, of which 22 are endemic (Moreno and Muchhala, 2005
). Burmeistera are often misclassified as epiphytes, but in fact individuals remain rooted in the ground throughout their lives as either free-standing herbs or hemi-epiphytic herbs or subshrubs that climb nearby vegetation (Jeppesen, 1981
). Flowers of Burmeistera are zygomorphic; corollas typically have a tubular base that expands distally into a bell shape with the reproductive parts positioned above the opening (Fig. 1). Anthers are fused together to form a tube into which pollen is shed. During the male phase, pollen is released gradually through the open end of this tube via a "pump mechanism" as the style elongates within the tube (Erbar and Leins, 1995
). When fully exserted (normally by the second day), the stigma becomes receptive and the female phase begins. Flowers senesce 6–8 d after anthesis (Stratton, 1989
), but stigmas probably lose receptivity earlier; flowers typically angle downwards, and visits stop several days after the female phase begins (N. Muchhala, personal observation). Individual plants remain in flower for months, with 1–4 flowers open at a time. Full seed set probably requires thousands of pollen grains; hand-pollinated flowers of B. sodiroana produced on average 2430 seeds (±388.9, N = 8; N. Muchhala, unpublished data).
|
Pollen deposition
I measured nocturnal and diurnal pollen deposition on stigmas to quantify effectiveness of visitors in transferring pollen. I located and identified Burmeistera plants throughout each reserve and then selected a trail loop with as many individuals as possible. I marked the base of the pedicel of each flower with masking tape and a unique number. I wrapped the flower's reproductive parts (i.e., the stigma and anther tube from which it emerges) with a layer of parafilm, which I left in place throughout the male and female phases of the flower. I then affixed a small rectangle of clear, double-sided plastic tape (10 x 5 mm) to the parafilm in order to collect pollen brought by floral visitors. I removed and replaced this tape every dawn and dusk to have separate nocturnal and diurnal samples of pollen deposition. These samples were immediately placed on a slide and covered with clear, single-sided plastic tape. In the laboratory, I temporarily lifted the tape and stained the pollen with gelatin cubes containing fuchsin dye (Beattie, 1971
). I examined the slides with a light microscope to identify pollen grains present. To quantify pollen deposition, I counted a subset of the pollen grains using the following method. I cut a 5 x 10 mm hole in the middle of a 15 x 20 mm square of poster board and affixed hairs in vertical and horizontal transects through the center of the hole. For each slide, I placed this square over the tape sample and counted and identified all pollen grains along the two transects. I was able to identify Burmeistera pollen to species, with the exception of two pairs of species with indistinguishable pollen: B. succulenta and B. ceratocarpa in Yanayacu and B. succulenta and B. cylindrocarpa in Bellavista. For these species, counts of conspecific pollen deposition may be overestimated. However this does not affect the primary goal of this aspect of the study, which was to compare the relative contributions of nocturnal and diurnal pollinators.
Specialization
I combined the flower visitation and pollen deposition results to estimate the specificity of Burmeistera pollination systems. For each flower, I calculated the percentage of total pollen deposition that could be attributed to bats and to hummingbirds. I considered a species of Burmeistera that received more than 75% of pollen flow from one type of pollinator to be specialized (sensu Fenster et al., 2004
) and in this way categorized each species as (1) specialized to bats, (2) specialized to hummingbirds, or (3) generalized to multiple taxa.
Pollination syndromes
I compared floral traits of species in each of the three categories of specialization to determine whether pollination syndromes can be identified in the genus. I measured 11 different aspects of floral morphology for each species: greater corolla length (greatest length from corolla base to ends of corolla lobes); lesser corolla length (from base to the split between the two dorsal and three ventral corolla lobes); corolla tube length (from base to corolla flare); corolla tube width (greatest width of corolla tube); outer aperture width (distance between distal ends of dorsal corolla lobes); inner aperture width (width at the split between the two dorsal and three ventral corolla lobes); calyx lobe length (length of the sepal-like calyx lobes), calyx lobe width (greatest width of calyx lobe); pedicel length (from branch to hypanthium), pedicel width (where pedicel attaches to hypanthium), and stigma exsertion (from distal end of corolla tube to center of stigma). I used principal components analysis to explore how the 10 species of Burmeistera group in the morphospace defined by these 11 morphological measurements. The analysis was conducted with SPSS for Windows (SPSS Inc., Chicago, Illinois, USA) using standardized data and a matrix of correlations. I also calculated the average flower height for each species (distance from the ground to the flower), and noted flower color, odor, and timing of anthesis. For B. cyclostigmata, I was only able to measure the morphology of five flowers and the height of two flowers; for all other species sample sizes ranged from 9 to 41 for morphology and 8 to 38 for height.
RESULTS
Flower visitation
Filming demonstrated that only bats and hummingbirds visit Burmeistera flowers regularly. Bat visits occurred only at night (1830–0630 hours), while hummingbird visits occurred only during the day (0630–1830 hours). Several moth visits were recorded at night, but these animals did not contact the reproductive parts of the flowers. I also filmed a rodent that climbed a nearby branch to visit a B. cylindrocarpa flower. Such visits are a very rare event, probably because the non-woody branches of Burmeistera cannot support the weight of a rodent.
Nine of the 10 species of Burmeistera that I filmed received bat visits (Table 2). The majority of these also received hummingbird visits during the day. For these nine species, bats visited somewhat more frequently than hummingbirds on average (0.27 vs. 0.20 visits/h, respectively). The tenth species, B. rubrosepala, only received hummingbird visits (0.24 visits/h). These visits were all made by Adelomyia melanogenys. Hummingbird visits to the other nine species were mainly made by A. melanogenys (N = 17) and Coeligana torquata (N = 6), with occasional visits by Aglaiocereus coelestis (N = 3), Haplophaedia lugens (N = 1), and Phaethornis longirostris (N = 1). Bats could not be reliably identified from videos; however, bat netting demonstrated that Anoura caudifera and A. geoffroyi (Phyllostomidae: Glossophaginae) visit these flowers (N. Muchhala, unpublished data). Both of these species regularly carry Burmeistera pollen and do not discriminate between the various Burmeistera species available in a reserve; that is, there is no specialization of Burmeistera species to different bat species.
|
On several large B. sodiroana plants, C. torquata were observed to set up territories and visit all flowers in regular cycles. A flower filmed within one of these territories received 7.84 hummingbird visits/h. Because it is an extreme outlier relative to the typical hummingbird visitation rate of 0.20 visits/h, I did not include this flower in the data summarized in Table 2.
Pollen deposition
In Table 3 I report nocturnal and diurnal pollen loads on the stigmas of 10 different species of Burmeistera in six different reserves. I summarized the data as mean number of pollen grains per sample (i.e., all pollen, regardless of identity), mean number of conspecific pollen grains per sample, and mean number of heterospecific Burmeistera pollen grains per sample. Note that I counted only a portion of total pollen present (along two transects through the center of each sample) to be able to compare relative abundance; total pollen deposition on stigmas was much higher.
|
Although the majority of pollen received by these nine species was conspecific, they also received an average of 13.4 grains of heterospecific Burmeistera pollen per night (N = 450). This means that approximately four of every five grains of Burmeistera pollen deposited were conspecific grains. As might be expected, this ratio was related to the local abundance of each species; when a given species of Burmeistera was rare locally, it received less conspecific pollen and more heterospecific pollen. For example, Burmeistera lutosa received no conspecific pollen in Bellavista, where it is rare, and an average of 55 grains per sample in Golondrinas, where it is abundant (Table 3). Similarly, B. multiflora received only two grains of conspecific pollen per sample in Golondrinas and none in Otonga, while in Pahuma it received 58 conspecific pollen grains per sample (Table 3).
Nocturnal samples also occasionally contained pollen from Markea, Marcgravia, Aphelandra, bromeliads, Meriania, and Passiflora—all flowers known to be bat-pollinated (Muchhala and Jarrin-V., 2002
). Diurnal samples occasionally contained pollen from Ericaceae and Gesneriaceae, as well as several unidentified pollen morphotypes.
Specialization
The results of filming demonstrated that only bats were depositing pollen on Burmeistera flowers at night and only hummingbirds during the day. Therefore, the results of nocturnal and diurnal pollen deposition can be used to estimate the relative importance of bats and hummingbirds as pollinators. This demonstrates that none of the species of Burmeistera were generalized; all 10 received greater than 75% of pollen flow from one type of pollinator (Table 4). Nine species were primarily bat-pollinated, with bats responsible for 84–100% of observed pollen flow. The remaining species, B. rubrosepala, was exclusively pollinated by hummingbirds.
|
The flowers of B. rubrosepala also differed morphologically from the bat-pollinated species (see Fig. 1). In Table 5, I report means of flower height and the 11 measurements of flower morphology for each species. For the principal components analysis of the 11 measurements of floral morphology, B. rubrosepala sits apart in the morphospace defined by the two most important components (Fig. 2). The first component accounted for 37.4% of the variation, the second for 23.7%, and the third for 14.7%. Comparison with a broken-stick null model (Jackson, 1993
) demonstrated that only these three components have eigenvalues greater than would be expected by chance, and the third only marginally so. The first component reflects overall flower size, as many variables contributed positively to loading. Those with coefficients whose absolute value is greater than 0.700 included lesser corolla length (0.877), greater corolla length (0.839), tube length (0.823), outer corolla width (0.821), inner corolla width (0.804), and pedicel width (0.785). Because of their small corollas and thin pedicels, this component separates B. rubrosepala and the bat-pollinated B. multiflora from the other bat-pollinated species. The second component further separates the various bat-pollinated species. Tube width (0.860), sepal width (0.799), and stigma exsertion (0.749) contribute to loading the second component.
|
|
DISCUSSION
The 10 focal species of Burmeistera were all highly specialized for pollination by either bats or hummingbirds. For nine species, the nectar bats Anoura geoffroyi and A. caudifera contributed to greater than 75% of conspecific pollen flow (Table 4). The remaining species, B. rubrosepala, was exclusively pollinated by the hummingbird Adelomyia melanogenys. These results stress the importance of quantifying the effectiveness of pollinators; although hummingbirds regularly visit the flowers of all 10 species, they only effectively transfer pollen of one. In the following sections, I discuss the close correlation between floral traits and pollinators of Burmeistera, the relative importance of these traits in restricting pollinators, and possible implications of the observed variation in floral morphology amongst the bat-pollinated species.
Pollination syndromes
The floral traits of these species appear to have been selected for by the animals that pollinate them and correspond closely to traditional chiropterophilous and ornithophilous pollination syndromes (Baker, 1961
; Pijl, 1961
; Helversen, 1993
). Hummingbirds rely primarily on vision to detect flowers and have no sense of smell; correspondingly, flowers of B. rubrosepala are brightly colored and lack odor. Bats, which are active nocturnally, rely heavily on olfaction and less on vision to detect flowers; flowers of bat-pollinated Burmeistera are dull colored and emit odor nocturnally. Bats treat flowers relatively roughly, as documented by my filming of visits to Burmeistera, while hummingbirds typically extract nectar without moving the flowers; correspondingly, bat-pollinated Burmeistera have pedicels that are nearly twice as thick as those of B. rubrosepala. Only B. rubrosepala flowers consistently move during hummingbird visits; they angle slightly down, ensuring that the reproductive parts contact the hummingbird's head. By increasing pedicel flexibility, decreased pedicel width may serve to enhance the effectiveness of hummingbirds as pollinators (sensu Hurlbert et al., 1996
). Hovering mechanics also differ between these pollinators: bats flap their wings in an arc that extends well in front of their heads, while hummingbirds keep their wings behind their backs (Helversen, 1993
). Therefore, flowers of bat-pollinated Burmeistera are well-exposed and high above the ground (mean height = 140.6 ± 29.6 cm), typically projecting vertically or at a 45° angle above the plant's foliage, while flowers of B. rubrosepala are oriented horizontally under the leaves and are closer to the ground (67.3 cm). Given that nectar bats echolocate, better exposure may also be important in increasing the acoustic "visibility" of bat-flowers (Helversen and Helversen, 1999
). Finally, these pollinator types differ in the morphology of their mouthparts. To ensure proper and consistent placement of pollen, Burmeistera flowers must fit their pollinators well. The wide corolla apertures of the bat-pollinated species closely fit bat snouts (Table 5), while the narrow corolla aperture of B. rubrosepala matches the relatively thin bills of hummingbirds (Fig. 1). Hummingbirds can access the nectar of wide flowers without touching reproductive parts, as demonstrated by the filming, while narrow flowers may preclude bat visits.
The close match between floral characteristics and pollinators amongst these 10 species of Burmeistera suggests pollination syndromes reliably predict pollinators for this genus. During fieldwork in Ecuador, I observed the flowers of an additional 13 species of Burmeistera; all correspond closely to the chiropterophilous syndrome. Based on a review of herbarium collections, species descriptions, and keys to Burmeistera throughout the neotropics (Wilbur, 1975
, 1981
; Jeppesen, 1981
; Stein, 1987
), I conclude that the genus as a whole is primarily bat-adapted, with occasional secondary pollination from hummingbirds as demonstrated for B. ceratocarpa (this study) and B. tenuiflora (Muchhala, 2003
). Apart from B. rubrosepala, the Costa Rican B. parviflora appears to be the only other nonchiropterophilous species; its small, bright yellow flowers suggest adaption to bee-pollination.
Specialization
Both bats and hummingbirds occur in Ecuadorian cloud forests, and both visit flowers of the bat-pollinated species, yet Burmeistera are highly specialized to one or the other taxon. Why don't any species of Burmeistera generalize to exploit both pollinator types? In considering the evolution of specialized pollination systems, it is useful to decompose the floral phenotype into various floral traits and identify those that serve to restrict the potential pollinator pool (Wilson and Thomson, 1996
; Aigner, 2001
). For Burmeistera, color and odor probably serve primarily to facilitate rather than restrict pollination. That is, bright coloration increases hummingbird pollination without restricting bat pollination, and stronger odor increases bat pollination without restricting hummingbird pollination. Timing of anthesis similarly does not restrict visitation because the flowers remain open for several days and nights. Decreased exposure, in terms of flower accessibility and height, may restrict bat visitation. And pedicel flexibility may restrict both pollinators; bats may be unable to extract nectar from flowers with thin, flexible pedicels while thick, rigid pedicels may decrease hummingbird contact with reproductive parts. However, of all of the floral traits that differ between the ornithophilous and chiropterophilous species of Burmeistera, I believe the width of the corolla aperture is the most critical in restricting pollination by bats and hummingbirds and imposes an adaptive trade-off that favors specialization. The ideal width for bat pollination is too wide to ensure that hummingbirds contact the reproductive parts, while the ideal width for hummingbird pollination is too narrow to allow bats access to the nectar. Two lines of evidence support the importance of the inner and outer aperture widths to flower–pollinator fit. First, both measurements are highly constrained amongst bat-pollinated species, with lower coefficients of variation than all other floral measurements (Table 5). Second, they are much smaller for the hummingbird-pollinated species, falling more than six standard deviations outside of the bat-pollinated means (Table 5). I hypothesize that the fit between flower and pollinator is so important for this genus that no intermediate width exists that would adequately exploit both bats' snouts and hummingbirds' bills. Experimental manipulation of width and other floral traits would be useful to verify their influence on bat and hummingbird pollination.
Variation in bat-pollinated flowers
The degree of stigma exsertion has relatively high levels of variation amongst the bat-pollinated species of Burmeistera (11.6–29.4 mm, coefficient of variation [CV] = 0.33). This variation may be biologically important if different degrees of exsertion correspond to different sites of pollen transfer on the heads of bat visitors. For sympatric species of plants, sharing a pollinator can reduce fitness due to reproductive interference (i.e., the loss of pollen to foreign stigmas and stigma clogging by foreign pollen; Rathcke, 1983
). Variation in the site of pollen placement is one way to alleviate this cost (Nilsson et al., 1987
; Armbruster et al., 1994
). In each of the study sites, three species of Burmeistera coexist on average (range 1–4 species; Table 3). If bats were randomly distributing pollen between three co-occuring species, each would be expected to receive 33% conspecific and 66% heterospecific pollen. Results show that they actually receive approximately 80% conspecific and only 20% heterospecific pollen on average. Local divergence in the degree of stigma exsertion may contribute to such low levels of heterospecific pollen transfer.
The morphology of the calyx lobes also varies greatly between bat-pollinated species, in terms of width (1.7–13.0 mm, CV = 0.92) and length (1.9–18.2 mm, CV = 0.63) as well as the margin (entire, sinuate, or dentate) and the angle relative to the corolla (erect, patent, or reflexed). A somewhat speculative hypothesis for this remarkable diversity is that it also serves to alleviate reproductive interference, in this case by encouraging pollinator fidelity. Calyx lobes likely function as a visual or acoustic signal for bats, facilitating detection of flowers of Burmeistera amidst background foliage. For many animals, it has been demonstrated that the experience gained in detecting a cryptic food item leads to formation of a "search image" to better detect these same items later (Langley, 1996
; Zentall, 2005
). If flowers of all local species of Burmeistera appeared identical, after visiting the flower of one species and learning that it contains nectar, a bat would be just as likely to visit a heterospecific Burmeistera next. If flowers appeared different for each species, a bat would be more likely to develop a distinct search image and follow with a visit to another individual of the same species. In this way, divergence in calyx lobe morphology may serve to encourage fidelity of individual bats despite infidelity of the species as a whole. Divergence in morphology of corolla lobes would likely achieve a similar effect, but these structures are less evolutionarily labile because of their functional role in protecting the reproductive parts of developing flowers prior to anthesis and in regulating the "fit" between bat and flower after anthesis.
Finally, the relatively high variation in pedicel length (40.6–100.1 mm, CV = 0.32) reflects different methods of spatially separating flowers and foliage to increase flower accessibility. The two free-standing herbaceous species (B. multiflora and B. lutosa) have some of the shortest pedicels (40.6 and 41.8 mm, respectively); because flowers only occur on the tops of vertical stems, these lengths are adequate to position them beyond the plant's foliage. The climbing hemi-epiphytic species (B. cyclostigmata, B. cylindrocarpa, B. borjensis, B. succulenta, and B. smaragdi) have flowers that intersperse with leaves along horizontal branches; for these species, longer pedicels (53.9–74.2 mm) serve to raise flowers above the leaves at 45° angles. Burmeistera sodiroana is also hemi-epiphytic, but its flowers hang below the branches on extremely long pedicels (100.1 mm), which position them horizontally beyond the leaves (see cover photo). Flowers of the hummingbird-pollinated B. rubrosepala hang below the branches in a similar fashion, but shorter pedicels (80.5 mm) place the flowers under rather than beyond the leaves. Burmeistera ceratocarpa is sometimes herbaceous and sometimes hemi-epiphytic and has evolved a unique solution to increase accessibility for bats. Wherever flowers occur, leaves are much smaller than elsewhere on the plant; herb-like pedicel lengths (41.2 mm) suffice to position the flowers beyond these leaves.
Conclusions
This study demonstrates high specificity in the pollination systems of Burmeistera and provides another example of the varied and often highly specialized pollination modes of tropical plants (e.g., Nilsson et al., 1987
; Armbruster, 1993
; Johnson and Steiner, 1997
; Kay and Schemske, 2003
). The close match between floral traits and pollinators in this genus supports the predictive power of the traditional chiropterophilous and ornithophilous pollination syndromes. Of the traits that differ between these syndromes, the width of the corolla aperture appears most important for restricting the potential pollinator pool and may impose a trade-off that favors specialization. While floral width varies little amongst the bat-pollinated species of Burmeistera, the degree of stigma exsertion and the morphology of calyx lobes are highly variable. Divergence in these traits may reduce heterospecific pollen transfer and thus facilitate the coexistence of multiple species. This study highlights various aspects of the floral phenotype that may have been shaped by pollinator-mediated selective pressures; experimental manipulations of floral traits would be useful to verify these proposed selective pressures.
FOOTNOTES
1 The author thanks S. Armbruster, V. Briggs, T. H. Fleming, H. Greeney, L. Jost, E. Knox, C. McCain, and K. Murray for discussion and comments on manuscript drafts; J. Vizuete and A. Caiza for assistance in the field, as well as C. Arcos, C. Engelsborg, A. Hoyos, M. Kox, D. Proaño, R. Stephens, and L. Tonato; J. Clark for discovering the population of B. rubrosepala; and the following for permission to work in their reserves: H. Greeney (Yanayacu), M. E. Manteca (Golondrinas), G. Onore (Otonga), R. Parsons (Bellavista), C. Woodward and J. Meisel (Pahuma), and Fundación Jatun Sacha (Bilsa). This work was supported by a Graduate Research Fellowship from the National Science Foundation, a Student Grant from Bat Conservation International, and a Curtis Fellowship from the University of Miami. 
LITERATURE CITED
Aigner P. A.. 2001. Optimality modeling and fitness trade-offs: when should plants become pollinator specialists?. Oikos 95: 177-184.[CrossRef][ISI]
Aigner P. A.. 2004. Floral specialization without trade-offs: optimal corolla flare in contrasting pollination environments. Ecology 85: 2560-2569.[CrossRef][ISI]
Armbruster W. S.. 1993. Evolution of plant pollination systems: hypotheses and tests with the neotropical vine Dalechampia. Evolution 47: 1480-1505.[CrossRef][ISI]
Armbruster W. S. Edwards M. E. Debevec E. M.. 1994. Floral character displacement generates assemblage structure of Western Australian triggerplants (Stylidium). Ecology 75: 315-329.[CrossRef][ISI]
Armbruster W. S. Fenster C. B. Dudash M. R.. 2000. Pollination "principles" revisited: specialization, pollination syndromes, and the evolution of flowers,. Det Norske Videnskaps-Akademi I Matematisk Naturvidenskapelige Klasse, Skrifter, Ny Serie 39: 179-200.
Baker H. G.. 1961. The adaptation of flowering plants to nocturnal and crepuscular pollinators. Quarterly Review of Biology 36: 64-73.[CrossRef]
Bawa K. S.. 1990. Plant–pollinator interactions in tropical rain forests. Annual Review of Ecology and Systematics 21: 399-422.[CrossRef][ISI]
Beattie A. J.. 1971. A technique for the study of insect-borne pollen. Pan-Pacific Entomologist 7: 82.
Castellanos M. C. Wilson P. Thomson J. D.. 2004. ‘Anti-bee' and ‘pro-bird' changes during the evolution of hummingbird pollination in Penstemon flowers. Journal of Evolutionary Biology 17: 876-885.[CrossRef][ISI][Medline]
Darwin C.. 1862. On the various contrivances by which British and foreign orchids are fertilized Murray, London, England.
Erbar C. Leins P.. 1995. Portioned pollen release and the syndromes of secondary pollen presentation in the Campanulales-Asterales-complex. Flora 190: 323-338.[ISI]
Feinsinger P. Colwell R. K.. 1978. Community organization among neotropical nectar-feeding birds. American Zoologist 18: 779-795.
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.[CrossRef][ISI]
Helversen D. von von Helversen O.. 1999. Acoustic guide in bat-pollinated flower. Nature 398: 759-760.[CrossRef]
Helversen O. von.. 1993. Adaptations of flowers to the pollination by glossophagine bats. In W. Barthlott, C. M. Naumann, S.-L. K. and K. L. Schuchmann [eds.] Animal–plant interaction in tropical environments 41-59 Museum Koenig, Bonn, Germany.
Herrera C. M.. 1996. Floral traits and plant adaptation to insect pollinators: a devil's advocate approach. In D. G. Lloyd and S. C. H. Barrett [eds.] Floral biology: studies on floral evolution in animal-pollinated plants 65-87 Chapman and Hall, New York, New York, USA.
Hurlbert A. H. Hosoi S. A. Temeles E. J. Ewald P. W.. 1996. Mobility of Impatiens capensis flowers: effect on pollen deposition and hummingbird foraging. Oecologia 105: 243-246.[ISI]
Jackson D. A.. 1993. Stopping rules in principal components analysis: a comparison of heuristical and statistical approaches. Ecology 74: 2204-2214.[CrossRef][ISI]
Jeppesen S.. 1981. Lobeliaceae. In G. Harling and B. Sparre [eds.] Flora of Ecuador, No. 14 9-170 Swedish Natural Science Research Council, Stockholm, Sweden.
Johnson S. D. Steiner K. E.. 1997. Long-tongued fly pollination and evolution of floral spur length in the Disa draconis complex (Orchidaceae). Evolution 51: 45-53.[CrossRef][ISI]
Kay K. M. Schemske D. W.. 2003. Pollinator assemblages and visitation rates for 11 species of Neotropical Costus (Costaceae). Biotropica 35: 198-207.[ISI]
Lammers T. G.. 2006. Campanulaceae. In K. Kubitzki [ed.] The families and genera of vascular plants, in press Springer-Verlag, Berlin, Germany.
Langley C. M.. 1996. Search images: selective attention to specific visual features of prey. Journal of Experimental Psychology: Animal Behavior Processes 22: 152-163.
Moreno M. P. Muchhala N.. 2005. Campanulaceae. In S. León-Yánez, R. Valencia, N. Pitman, L. Endara, C. Ulloa and H. Navarette [eds.] Libro rojo de las plantas endémicas del Ecuador 2005 Herbario QCA, Pontificia Universidad Católica del Ecuador, Quito, Ecuador.
Muchhala N.. 2003. Exploring the boundary between pollination syndromes: bats and hummingbirds as pollinators of Burmeistera cyclostigmata and B. tenuiflora. Oecologia 134: 373-380.[ISI][Medline]
Muchhala N. Jarrin-V P.. 2002. Flower visitation by bats in cloud forests of western Ecuador. Biotropica 34: 387-395.[ISI]
Nilsson L. A. Jonsson L. Ralison L. Randrianjohany E.. 1987. Angraecoid orchids and hawkmoths in central Madagascar: specialized pollination systems and generalist foragers. Biotropica 19: 310-318.[CrossRef][ISI]
Ollerton J.. 1996. Reconciling ecological processes with phylogenetic patterns: the apparent paradox of plant–pollinator systems. Journal of Ecology 84: 767-769.[CrossRef][ISI]
Pijl L.. 1961. Ecological aspects of flower evolution. II. Zoophilous flower classes. Evolution 15: 44-59.[CrossRef][ISI]
Rathcke B. J.. 1983. Competition and facilitation among plants for pollination. In L. Real [ed.] Pollination biology 305-329 Academic Press, New York, New York, USA.
Raven P. H.. 1972. Why are bird-visited plants predominantly red?. Evolution 26: 674.[CrossRef][ISI]
Sargent R. D. Otto S. P.. 2006. The role of local species abundance in the evolution of pollinator attraction in flowering plants. American Naturalist 167: 67-80.[CrossRef][ISI][Medline]
Schemske D. W. Bradshaw H. D.. 1999. Pollinator preference and the evolution of floral traits in monkeyflowers (Mimulus). Proceedings of the National Academy of Sciences, USA 96: 11910-11915.
Schemske D. W. Horvitz C. C.. 1984. Variation among floral visitors in pollination ability—a precondition for mutualism specialization. Science 225: 519-521.
Stebbins G. L.. 1970. Adaptive radiation of reproductive characteristics in angiosperms. I: Pollination mechanisms. Annual Review of Ecology and Systematics 1: 307-326.[Medline]
Stein B. A.. 1987. Synopsis of the genus Burmeistera (Campanulaceae, Lobelioideae) in Peru. Annals of the Missouri Botanical Garden 74: 494-496.[CrossRef][ISI]
Stein B. A.. 1992. Sicklebill hummingbirds, ants, and flowers. Bioscience 42: 27-33.
Stone G. N. Willmer P. Rowe J. A.. 1998. Partitioning of pollinators during flowering in an African Acacia community. Ecology 79: 2808-2827.[CrossRef][ISI]
Stratton D. A.. 1989. Longevity of individual flowers in a Costa Rican cloud forest: ecological correlates and phylogenetic constraints. Biotropica 21: 308-318.[CrossRef][ISI]
Temeles E. J. Linhart Y. B. Masonjones M. Masonjones H. D.. 2002. The role of flower width in hummingbird bill length–flower length relationships. Biotropica 34: 68-80.[ISI]
Thomson J. D.. 2003. When is it mutualism?. American Naturalist 162: S1-S9.
Thompson J. D.. 2001. How do visitation patterns vary among pollinators in relation to floral display and floral design in a generalist pollination system?. Oecologia 126: 386-394.[CrossRef][ISI]
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.[CrossRef]
Wilbur R. L.. 1975. A synopsis of the Costa Rican species of Burmeistera (Campanulaceae: Lobelioideae). Bulletin of the Torrey Botanical Club 102: 225-231.[CrossRef][ISI]
Wilbur R. L.. 1981. Additional Panamanian species of Burmeistera (Campanulaceae: Lobelioideae). Annals of the Missouri Botanical Garden 68: 167-171.[CrossRef][ISI]
Wilson P. Thomson J. D.. 1996. How do flowers diverge?. In D. G. Lloyd and S. C. H. Barrett [eds.] Floral biology: studies on floral evolution in animal-pollinated plants 88-111 Chapman and Hall, New York, New York, USA.
Zentall T. R.. 2005. Selective and divided attention in animals. Behavioural Processes 69: 1-15.[CrossRef][ISI][Medline]
This article has been cited by other articles:
|
|
 |
|
  S. Marten-Rodriguez and C. B. Fenster Pollination Ecology and Breeding Systems of Five Gesneria Species from Puerto Rico Ann. Bot., July 1, 2008; 102(1): 23 - 30. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
