HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Boyd, A. E. Search for Related Content PubMed Articles by Boyd, A. E. Agricola Articles by Boyd, A. E.

Breeding system of Macromeria viridiflora (Boraginaceae) and geographic variation in pollinator assemblages¹

Warren Wilson College, CPO 6074, P.O. Box 9000, Asheville, North Carolina 28815-9000 USA

Received for publication August 21, 2003. Accepted for publication August 5, 2004.

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
This study explores the association between variation in pollinator type and flower size in Macromeria viridiflora (Boraginaceae) by studying the breeding system of the plant and the pollinator effectiveness of floral visitors. Studies were conducted at two sites where plants differ in flower size and floral visitors. Breeding system studies showed that while plants are self-compatible and occasionally produce seed autogamously, pollinators are important for reproductive success in the plants. However, plants are not pollinator-limited at these sites. Combining visitation rate and pollen deposition as measures of pollinator effectiveness, I found hummingbirds to be the most effective pollinators at both sites. Although hawkmoths also pollinate the flowers, they visit the flowers less frequently and, at one of the two sites, deposit less pollen. These results are consistent with the hypothesis that geographic variation in corolla size is the result of selection by different hummingbird species.

Key Words: Boraginaceae • hawkmoth pollination • hummingbird pollination • Macromeria viridiflora • plant breeding system • pollen deposition • pollinator effectiveness • visitation rate

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
Biologists have long invoked evolution by natural selection to explain relationships between the traits of pollinating animals and those of the flowers they visit. Attracting and positioning pollinators so that effective pollen transfer occurs are necessary for the reproductive success of plants that depend on animal pollinators (Faegri and van der Pijl, 1979 ; Proctor et al., 1996). Floral tube length and spur length may directly affect the physical transfer of pollen from anther to pollinator and/or from pollinator to stigma (Faegri and van der Pijl, 1979 ; Nilsson, 1988 ; Johnson and Steiner, 1997 ; Alexandersson and Johnson, 2002 ). Therefore, there is great potential for pollinator-mediated selection on these traits (Hodges, 1997 ). Tube length and spur length have been the traits most frequently identified in intraspecific studies as exhibiting geographic variation associated with pollinator identity (Grant and Grant, 1965 ; Miller, 1981 ; Robertson and Wyatt, 1990 ; Johnson, 1994 ; Arroyo and Dafni, 1995 ; Johnson, 1997 ; Johnson and Steiner, 1997 ; Alexandersson and Johnson, 2002 ; but see Herrera [1996] for evidence to the contrary).

In a previous study, I documented geographic variation in floral and vegetative traits of Macromeria viridiflora DC. (Boraginaceae) in a study of the morphometrics of populations across the species' range, from southern Chihuahua, Mexico, north to the Sangre de Cristo Mountains of New Mexico (Boyd, 2002 ). Although variation in vegetative traits follows no perceptible latitudinal pattern, variation in flower size follows a latitudinal gradient: corollas are shortest in the northernmost populations (ranging from 4.2 to 5.2 cm) and longest in the southernmost populations (ranging from 6.7 to 8.5 cm). The populations with the shorter-flowered plants are visited by hummingbirds with short culmens, and the populations with longer-flowered plants are visited by hummingbirds with longer culmens (Boyd, 2002 ), suggesting that this geographic variation in flower size may be associated with hummingbird pollinator-mediated selection.

However, flowers of M. viridiflora are visited regularly by hawkmoths as well as hummingbirds, and where hummingbird species visiting the plants vary across sites, the hawkmoth species is consistent across sites. Either of these pollinator groups may be effective pollinators, and their relative effectiveness may vary across sites.

In the present study, I seek to clarify the association between variation in pollinator type and flower size variation in Macromeria viridiflora. The following questions are addressed: (1) How important are pollinators in the plant's breeding system? (2) What is the relative pollinator effectiveness of the floral visitors?

To determine the importance of pollinators, I conducted breeding experiments at two sites where plants differ in flower size and floral visitors, testing at each site for self-compatibility, autogamy, and the potential for pollinator visits to increase seed set.

To determine relative importance of different pollinator groups, I compared pollination effectiveness among different visitors within and between sites by analyzing frequency, duration, and pollen deposition rates of floral visits.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
Macromeria viridiflora is an herbaceous perennial found in montane pine-oak and mixed-coniferous forests. It is distributed from the mountains of the northern Sierra Madre Occidental north to the Mogollon Rim of Arizona and the southern Sangre de Cristo Mountains of New Mexico, USA, at elevations between 1500 and 3000 m a.s.l. (Turner, 1994 ). Flowers are arranged in helicoid cymes and have long (3.5–8.5 cm), trumpet-shaped, pale green corollas that fade to pale yellow with age. Flowering of Macromeria viridiflora coincides with the summer wet season (June–September), and plants die back to the ground with the onset of winter. The stigma and the five anthers are exserted from the corolla; the style extends well beyond the anthers so that the stigma is positioned farther out than the anthers. Anthers dehisce before anthesis; flowers wilt and corollas fall off the plant 2–3 d after anthesis. Each flower produces up to four pearly-white, ovoid nutlets. These nutlets have no special dispersal mechanism and fall passively from the plants when ripe.

Field work was conducted at two sites: Mt. Lemmon, in the Santa Catalina Mountains of southeastern Arizona, USA (32°25' N, 110°45' W), and the South Fork of the Little Colorado River in the White Mountains of east-central Arizona (34°05' N, 109°25' W). On Mt. Lemmon, populations were located in the picnic area in Marshall Gulch, near the head of the Marshall Gulch Trail, and along Mt. Lemmon trail #5; plants at this site had a mean floral length of 56.6 mm (Boyd, 2002 ). The South Fork study site was located on the floodplain and surrounding slopes of South Fork creek; plants at this site had a mean floral length of 46.4 mm (Boyd, 2002 ).

All statistical analyses were performed in JMP (SAS Institute, 1989–1999 ).

Breeding system
Breeding system experiments were conducted on Mt. Lemmon in July 1996 and at South Fork in July 1999. At each site, five plants with more than five flowering stems each were selected for the study; plants of this size produce sufficient flowers to permit all treatments to be conducted simultaneously. On each selected plant, one stem was chosen at random for each of five treatments. The five most mature buds on each stem were tagged for use in the study, for a total of five flowers per plant assigned to each treatment. The five treatments were (1) unbagged control; (2) bagged, not emasculated; (3) bagged, emasculated; (4) bagged, emasculated, self-pollinated by hand; (5) bagged, emasculated, outcross-pollinated by hand.

For self-pollination by hand, pollen was taken from the anthers of the flower being pollinated. For outcross-pollination by hand, pollen was taken from a flower from a plant in the population that was not otherwise being used in the breeding system study. To transfer pollen, a single anther was removed from the donor flower using a pair of forceps, and the pollen was then applied directly from this anther to the recipient flower's stigma. Hand pollinations were performed the first day that the flower was open.

Visitors were excluded from the flowers in bagged treatments 2 through 5 by covering the flowers with fine-mesh cloth bags secured with twist-ties. Bags were placed over flowers before anthesis and were removed after flowers wilted or fell from the plants. Because anthers dehisce prior to anthesis, flowers in treatments 3 through 5 were emasculated in the bud before anther dehiscence. After 21–28 d, treated flowers were assessed for fruit and seed set.

Results of treatments were compared using one-way ANOVA. Four a priori contrasts were made comparing the following treatments: (1) and (2), to determine if seed set is increased by the presence of pollinators; (3) and (4), to determine if plants produce seed through autogamy; (4) and (5), to determine if plants are self-compatible; and (1) and (5), to determine if plants are pollinator limited.

Variance associated with the results for treatment 3 differed sharply from that for the other treatments. Because ANOVA assumes homogeneity of variance, treatment 3 was omitted from the ANOVA. I used Welch's approximate t-test (Zar, 1984 ) for the first of these comparisons because it does not assume equal variances. The other three comparisons were made using ANOVA, including site as an effect.

Pollinator effectiveness
I used two measures of pollinator effectiveness: rate and duration of visits and amount of pollen deposited onto stigmas. Visitation rate gives a measure of pollinator quantity, whereas pollen deposition gives a measure of pollinator quality (Herrera, 1987 ). These measures have been combined in previous studies to quantify pollinator effectiveness (Thomson et al., 1982 ; Kearns and Inouye, 1994 ; Fishbein and Venable, 1996 ).

I observed groups of plants for flower visitation for a total of 64 h at South Fork and 117 h on Mt. Lemmon in 1999. Observations included periods throughout the day from 05:00 to 20:30 and covered a variety of weather conditions, ranging from sunny and warm to very cool and rainy. I watched flowers for nocturnal visitors during periods throughout several nights as well, but because no visitors were seen at night I limited my observations to the times above. During each observation period, I recorded number of flowers open in the observed area, duration of each visit, identity of the visitor, number of flowers visited, and any movement of the visitor between plant species. Eliminating rare pollinator groups because of small sample size, I tested for differences among visitor taxa in visitation rate (number of visits per flower per 1-h observation period) and mean visit duration. A Kruskal-Wallis test was used to compare frequency of visits to patches; Welch's ANOVA and Tukey-Kramer tests were used to compare visit duration among taxa.

To test the pollen deposition rate for each visitor species, I emasculated flowers in the bud before anther dehiscence (a total of 77 buds on Mt. Lemmon and 74 buds at South Fork, both in the summer of 1999) and then enclosed them in fine-mesh bags to exclude pollinators. After anthesis, I removed the bag and watched the flower until one visitor probed the flower. I recorded the visitor taxon, and then removed the stigma and distal 3–4 mm of the style and mounted it in melted fuchsin jelly (Kearns and Inouye, 1993 ) on a glass slide. This method preserved the stigma and pollen that had been placed there until they could be transported to the lab, where pollen grains present on each slide were counted under a microscope.

For visitation rates and amounts of pollen deposited by pollinator classes, I reported medians and interquartile ranges because of very skewed distribution of data. Because distributions were not normal, pollen deposition was compared between visitor species within each site using the Wilcoxon two-sample test. I also compared the pollen deposition by one visitor (hawkmoths) between the two sites using the same test.

To create a composite measure of pollinator effectiveness, I multiplied mean pollen deposition by mean visit duration for each floral visitor species. I used means despite the nonparametric nature of the data because the medians of zero for visitation rates by hawkmoths would result in pollinator effectiveness quotients of zero. This would not reflect the fact that while these pollinators only visit for a short while at dusk, they are effective pollinators during that time, visiting many flowers and depositing plenty of pollen.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
Breeding system
Flowers in the unbagged control treatments had the highest mean seed set (Mt. Lemmon: mean = 1.6 ± 0.24 SE nutlets; South Fork: mean = 2.6 ± 0.18 SE nutlets; Table 1). The bagged, emasculated flowers had the lowest mean seed set (Table 1) and in fact only two of 23 flowers at South Fork produced one seed each (out of four possible seeds maximum per flower; mean = 0.16 ± 0.10 SE). No seeds were produced by emasculated, bagged flowers on Mt. Lemmon, which was expected because the treatment was intended to prevent any pollen from reaching the stigma. The low level of seed set that did occur in bagged, emasculated flowers at South Fork could be due to a low frequency of agamospermy or may be due to experimental error. The latter seems more likely, especially considering that no seeds were produced by this treatment on Mt. Lemmon.

Pollinator effectiveness
At Mt. Lemmon, magnificent hummingbirds (Eugenes fulgens Swainson) and hawkmoths (white-lined sphinx moths, Hyles lineata Fabricius) were common visitors. Smaller hummingbirds (Selasphorus sp.) and small bees were rare visitors to flowers. At South Fork, rufous hummingbirds (Selasphorus rufus Gmelin) and hawkmoths (white-lined sphinx moths, Hyles lineata) were common floral visitors, and small bees were rare visitors.

Mean duration of visits by hawkmoths on Mt. Lemmon was longer than any other pollinator at either sites (mean = 6.22 s ± 0.55 SE; Welch's ANOVA: F = 11.32, P < 0.0001, Table 2). There was no difference in visit duration among other pollinators.

View larger version (35K): [in this window] [in a new window]   Fig. 1. Mean number of visits per hour by common floral visitors to Macromeria viridiflora at two study sites

Hawkmoths at Mt. Lemmon deposited fewer pollen grains per visit than did hummingbirds (Table 3; ² = 5.54, P = 0.0186). At South Fork, there was no significant difference in pollen deposition among pollinators (Table 3; ² = 2.04, P = 0.1531). Hawkmoths at South Fork deposited significantly more pollen than did their conspecifics at Mt. Lemmon (Table 3; ² = 12.43, P = 0.0004).

Overall, hummingbirds emerge as much more effective pollinators than hawkmoths, with rufous hummingbirds at South Fork being the most effective pollinators (Table 3). The marked difference between rufous hummingbirds (PE [pollinator effectiveness] = 516.8) and magnificent hummingbirds (PE = 86.5) is due to the much higher visitation rate of the former, whereas the very low value for hawkmoths at Mt. Lemmon (PE = 12.4) relative to its conspecifics at South Fork (PE = 27.1) is due to a much lower pollen deposition rate at the former site.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
Seed set in M. viridiflora was not dependent on pollen source; seed set from self-pollinated and outcross-pollinated flowers was equal, showing that these plants are highly self compatible. It is possible that some degree of inbreeding depression would emerge in seed germination rates or other measures of offspring fitness.

Autogamy can occur in these plants, as demonstrated by seed set in bagged flowers with intact anthers (treatment 2) as compared with emasculated, bagged flowers (treatment 3), although the level of seed production via autogamy is relatively low. Therefore, not only are plants self-compatible but they are also capable of producing seed in the absence of pollinators, albeit at a lower level.

The fact that these plants are self compatible and are somewhat autogamous raises the question as to whether or not attracting pollinators is important to this species. Pollinators can be important contributors to reproductive success even in self-compatible, autogamous plants if reproductive success is significantly increased by pollinator visitation. In M. viridiflora, seed set was increased at both sites by the presence of pollinators, as shown by the comparison between unbagged and bagged flowers with intact anthers. However, there was no evidence for pollinator limitation because open-pollinated control flowers had greater seed set than the hand-pollinated flowers. This may be due to seed set being limited by resources rather than pollinators, or poor pollinator success by human researchers.

The pollination effectiveness of common visitors can play a role in their potential as agents of selection. Combining measures of pollinator effectiveness, hummingbirds are apparently more effective pollinators of Macromeria viridiflora than are hawkmoths. Even at South Fork, where pollen deposition by hummingbirds was not significantly greater, the large difference in visitation rate would mean that hummingbirds were responsible for more seed production than hawkmoths.

The hawkmoths at Mt. Lemmon had a longer handling time and deposited less pollen per visit than their conspecifics at South Fork. The difference may be due to the difference in flower size (i.e., plants at South Fork have much shorter corollas than plants on Mt. Lemmon; see Boyd, 2002 ). Tongue lengths of Hyles lineata in central Colorado (north of M. viridiflora populations) averaged around 4 cm in a study by Miller (1981) , and in Hermosillo, Mexico (just northwest of the southernmost M. viridiflora populations) averaged 3.5 cm ± 0.08 SE in a study by Willmott and Burquez (1996) . These values are close to the average corolla tube length at South Fork (mean = 3.8 cm ± 0.51 SD) but shorter than the average corolla tube length at Mt. Lemmon (mean = 4.8 cm ± 0.60 SD). The moths may need more time to handle the larger flowers on Mt. Lemmon and collect the nectar. Alternatively, the flowers may have a larger standing crop of nectar on Mt. Lemmon due to lower hummingbird visitation, and thereby require longer visitation time by moths. In either case, longer visitation time could be causing the lower pollen deposition on the stigmas.

Hummingbirds at both sites were frequent and effective pollinators. The difference in bill length between the two species mirrors the difference in corolla length of the flowers at the two sites (Boyd, 2002 ); this difference probably makes it possible for the flowers to be effectively pollinated by their respective hummingbird species at both sites. In addition, the difference in hummingbird bill size may have resulted in selective pressure for geographic differentiation in corolla size in this species. Interestingly, this selection by hummingbirds apparently may have been powerful enough at Mt. Lemmon to override counter-selection by hawkmoths, given the evolution of corollas in this population that are too large for efficient pollination by the moths.

Baker (1961) has described a very similar mixed pollination system in Mirabilis froebeli, another southwestern plant. The flowers of M. froebeli are pink and have long, tubular corollas. The flowers themselves have no odor detectible by humans, but the foliage and stems exude a strong odor at dusk. Flowers open at around 18:00 and are visited by hummingbirds in the early evening and Hyles lineata at dusk, following the same pattern seen here in M. viridiflora. A more widespread study of flowering plant and pollinator assemblages in the southwest might show that this system of sharing flowers between hawkmoths and hummingbirds is widespread.

To conclude, I have shown that pollinators increase reproductive success in M. viridiflora. Flowers of M. viridiflora vary in size along a latitudinal gradient, and this geographic variation in corolla size is mirrored by geographic variation in the bill size of the major hummingbird species visiting the flowers (Boyd, 2002 ). Although hawkmoths also pollinate the flowers, they are less effective pollinators, visiting the flowers less frequently and, at one of the two sites, depositing less pollen. These results are consistent with the hypothesis that geographic variation in corolla size is the result of selection by different hummingbird species.

	FOOTNOTES

 
¹ The author thanks A. Fox, A. Martin, J. Aukema, S. Adondakis, and J. King for assistance in the field, and L. McDade, M. McIntosh, J. Aukema, and three anonymous reviewers for critical reading of the manuscript. This research was supported by funding from the American Society of Plant Taxonomists, the Department of Ecology and Evolutionary Biology, and the Research Group in the Analysis of Biological Diversification at the University of Arizona, and T & E, Inc. Grants for Conservation Research.

² Reprint requests: (e-mail: aboyd{at}warren-wilson.edu )

	LITERATURE CITED

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
Aigner P. A. 2001 Optimality modeling and fitness trade-offs: when should plants become pollinator specialists?. Oikos 95: 177-184[CrossRef][ISI]

Alexandersson R. S. D. Johnson 2002 Pollinator-mediated selection on flower-tube length in a hawkmoth-pollinated Gladiolus (Iridaceae). Proceedings of the Royal Society of London B 269: 631-636[CrossRef]

Arroyo J. A. Dafni 1995 Variations in habitat, season, flower traits and pollinators in dimorphic Narcissus tazetta L. (Amaryllidaceae) in Israel. New Phytologist 129: 135-145[CrossRef][ISI]

Baker H. G. 1961 The adaptation of flowering plants to nocturnal and crepuscular pollinators. Quarterly Review of Biology 36: 64-73[CrossRef]

Boyd A. E. 2002 Morphological analysis of Sky Island populations of Macromeria viridiflora (Boraginaceae). Systematic Botany 27: 116-126[ISI]

Faegri K. L. van der Pijl 1979 The principles of pollination ecology, 3rd rev. ed. Pergamon Press, Oxford, UK

Fishbein M. D. L. Venable 1996 Diversity and temporal change in the effective pollinators of Asclepias tuberosa. Ecology 77: 1061-1073[CrossRef][ISI]

Grant V. K. A. Grant 1965 Flower pollination in the Phlox family. Columbia University Press, New York, New York, USA

Herrera C. M. 1987 Components of pollinator "quality": comparative analysis of a diverse insect assemblage. Oikos 50: 79-90[CrossRef][ISI]

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

Hodges S. A. 1997 Floral nectar spurs and diversification. International Journal of Plant Sciences 158: S81-S88[CrossRef]

Johnson S. D. 1994 Evidence for Batesian mimicry in a butterfly-pollinated orchid. Biological Journal of the Linnean Society 53: 91-104

Johnson S. D. 1997 Pollination ecotypes of Satyrium hallackii (Orchidaceae) in South Africa. Botanical Journal of the Linnean Society 24: 225-235

Johnson S. D. K. E. Steiner 1997 Long-tongued fly pollination and evolution of floral spur length in the Disa draconis complex (Orchidaceae). Evolution 51: 45-53[CrossRef][ISI]

Kearns C. A. D. W. Inouye 1993 Techniques for pollination biologists. University Press of Colorado, Niwot, Colorado, USA

Kearns C. A. D. W. Inouye 1994 Fly pollination of Linum lewisii (Linaceae). American Journal of Botany 81: 1091-1095[CrossRef][ISI]

McDade L. A. P. Davidar 1984 Determinants of fruit and seed set in Pavonia dasypetala (Malvaceae). Oecologia 64: 61-67

Miller R. B. 1981 Hawkmoths and the geographic patterns of floral variation in Aquilegia caerulea. Evolution 35: 763-774[CrossRef][ISI]

Nilsson L. A. 1988 The evolution of flowers with deep corolla tubes. Nature 334: 147-149[CrossRef]

Robertson J. L. R. Wyatt 1990 Evidence for pollination ecotypes in the yellow-fringed orchid. Evolution 44: 121-133[CrossRef][ISI]

Rozzi R. M. K. Arroyo J. J. Armesto 1997 Ecological factors affecting gene flow between populations of Anarthrophyllum cumingii (Papilionaceae) growing on equatorial- and polar-facing slopes in the Andes of Central Chile. Plant Ecology 132: 171-179[CrossRef][ISI]

SAS Institute. 1989–1999 JMP IN version 3.2.1. SAS Institute, Cary, North Carolina, USA

Thomson J. D. W. P. Maddison R. C. Plowright 1982 Behavior of bumble bee pollinators of Aralia hispida Vent. (Araliaceae). Oecologia 54: 326-336[CrossRef][ISI]

Turner B. 1994 Synoptical study of the genus Macromeria (Boraginaceae). Phytologia 77: 393-407

Willmott A. P. A. Burquez 1996 The pollination of Merremia palmeri (Convolvulaceae): can hawk moths be trusted?. American Journal of Botany 83: 1050-1056[CrossRef][ISI]

Zar J. H. 1984 Biostasticial analysis, 2nd ed. Prentice-Hall, Englewood Cliffs, New Jersey, USA

This article has been cited by other articles:

				H. F. Sahli and J. K. Conner Visitation, effectiveness, and efficiency of 15 genera of visitors to wild radish, Raphanus raphanistrum (Brassicaceae) Am. J. Botany, February 1, 2007; 94(2): 203 - 209. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Boyd, A. E. Search for Related Content PubMed Articles by Boyd, A. E. Agricola Articles by Boyd, A. E.