Plant Biology ISSN 1435-8603

RESEARCH PAPER

Anther and stigma morphology in mirror-image flowers of Chamaecrista chamaecristoides (Fabaceae): implications for buzz pollination
´ ´ ´ G. Arceo-Gomez1,2, M. L. Martınez1, V. Parra-Tabla3 & J. G. Garcıa-Franco1
´ ´ ´ 1 Departamento de Ecologıa Funcional, Instituto de Ecologıa, Xalapa, Veracruz, Mexico 2 Department of Biological Sciences, University of Pittsburgh, Pittsburgh, PA, USA ´ ´ ´ ´ ´ ´ ´ ´ 3 Departamento de Ecologıa Tropical, Campus de Ciencias Biologicas y Agropecuarias, Universidad Autonoma de Yucatan, Itzimna, Merida, Yucatan, Mexico

Keywords Poricidal anther dehiscence; sonication frequency; stigma trichomes; stigmatic cavity. Correspondence ´ G. Arceo-Gomez, University of Pittsburgh, Department of Biological Sciences 225 Clapp Hall, 4249 Fifth Avenue Pittsburgh, PA 15260, USA. E-mail: gea13@pitt.edu Editor M. Ayasse Received: 9 September 2009; Accepted: 24 December 2009. doi:10.1111/j.1438-8677.2010.00324.x

ABSTRACT Enantiostyly (mirror-image flowers) is usually associated with buzz pollination. In buzz-pollinated flowers, pollen is released through terminal pores after bees vibrate the stamens. Several studies have evaluated the function of ‘buzzing’ in pollen release, but less attention has been paid to the effect of buzzing on pollen capture and deposition on stigmas. Evaluating the mechanism of pollen dispersion in buzzpollinated flowers is important because it may affect mating patterns and reproductive success. In this study, we analysed the morphology of sexual organs (anther and stigma) using electron microscopy, and determined the relationship between sexual organ structure and pollen capture function through experimental manipulations of buzz-pollinated flowers of Chamaecrista chamaecristoides, as well as vibration frequencies on floral visitors. Pollen release occurs through two terminal pores at the tip of the stamens. However, unlike most angiosperms that have their stigmatic surface exposed, C. chamaecristoides presents a stigmatic surface inside a cavity covered by trichomes. Experimental manipulations showed that effective fertilisation is only achieved when the style is vibrated, suggesting that buzzing is not only important for pollen release but also for pollen capture and deposition on the stigma. This result, in addition to vibration frequency analysis, suggests that although all floral visitors buzz flowers only those that buzz at higher frequencies achieve effective fertilisation. The anatomical features of sexual organs in flowers of C. chamaecristoides demonstrate that this species possesses a highly specialised, elaborate morphology, with both genders selected for traits that promote buzz pollination.

INTRODUCTION Ever since Darwin’s seminal work, the evolution of floral traits has been widely interpreted as adaptations that facilitate animal pollination, especially morphological traits that influence pollen removal and deposition (Herrera 1996; Fenster et al. 2004). Seed set depends on the successful interaction between pollen and stigma, so this is one of the most important processes in angiosperms, and future generations are dependent on its successful operation (Heslop-Harrison 2000). Most angiosperms have an exposed stigmatic surface that may facilitate the process of pollen capture and deposition (Dulberger et al. 1994). However, in some species in the genera Cassia, Senna and Chamaecrista in the Fabaceae, access to the stigmatic surface is protected by a fringe of trichomes (Dulberger et al. 1994). The size and number of trichomes and the degree to which the stigmatic surface is exposed vary depending on species (Dulberger 1981; Dulberger et al. 1994; Tucker 1996). This type of stigma is commonly associated with buzz pollination (Buchmann

1983), a pollination system in which bees vibrate their indirect flight muscles to produce vibrations to release pollen from the anthers of a flower, which occurs in about 6% of angiosperms (Buchmann 1983; Corbet et al. 1988; Harder & Barclay 1994; Laporta 2005). The presence of trichomes on the style has been described as a mechanism to prevent desiccation of stigmatic exudate, the secretion that facilitates pollen germination inside the stigmatic cavity (Dulberger et al. 1994). However, other studies (Wang et al. 2005) have also reported a stigmatic surface with surrounding fringe hairs, and suggested that these help to guide pollen inside the stigma, thus facilitating automatic self-pollination. In spite of the different interpretations suggested to date, the functional and adaptive significance of this type of stigma remain unclear. The variation associated with different types of stigma clearly shows the variety of mechanisms that flowers have evolved to enhance pollen capture and deposition (Dulberger et al. 1994). In this sense, several studies have analysed how buzzing affects pollen release (Buchmann 1974, 1983; Thorp
19

Plant Biology 13 (Suppl. 1) (2011) 19–24 ª 2010 German Botanical Society and The Royal Botanical Society of the Netherlands

chamaecristoides. we examine the stigma and anther morphology of buzz-pollinated Chamaecrista chamaecristoides using electron microscopy. subhumid climate has a mean annual precipitation between 1200 and 1500 mm..). and (iv) open un-manipulated flowers (control). Parra-Tabla & Garcıa-Franco & Estes 1975.. we again randomly selected and collected one flower from each morph (right. Chamaecrista chamaecristoides flowers.Buzzing stigma: implications for effective fertilisation ´ ´ ´ Arceo-Gomez.versus right-styled flowers) of different individuals (out-crossing) located 5–10 m apart. Not all treatments were conducted in the same plant due to the reduced number of flowering plants available and the different numbers of open flowers per plant at the time manipulations were performed. Flowers remain open for only 1 day and ´ are visited by pollen-collecting bees (Garcıa-Franco J. The study population of C. 96º22¢W. large yellow flowers (Fig. enantiostylous species (with both left. and followed this procedure twice in a day. In each of these. Flowers were collected and observed at two distinct times in order to determine if the degree to which the stigmatic surface is exposed varies during the time that flowers remain open (1 day). However. 1) provide no nectar reward. unpublished data). 20 days on an individual plant (Arceo-Gomez G.and left-styled flowers). assessing successful pollination by means of fruit set. giving a total of 40 flowers. and the pollen is released through apical pores in the anthers (poricidal anther dehiscence) following vibration by large bees. examined and photographed for each flower. Evaluation of buzzing on pollen deposition The study site is located in the Coastal Research Center La Mancha (CICOLMA. unpublished data). This plant is a monomorphic. Evaluation of stigma and anther morphology Chamaecrista chamaecristoides (Fabaceae) is a perennial shrub 50–70-cm tall.. Forty flowers were manipulated in each pollination treatment. chamaecristoides grows on the mobile coastal dunes located at La Mancha centre (Martinez 2003). Each manipulation was considered a unit of replication. In this study. The arrows show in (A) the left-styled morph. METHODS Study species mean yearly temperature of 26 °C (Castillo-Campos & Medina 2002). A B Fig.l. and observed in the laboratory with an electron microscope (JSM-5600LV-JEOL scanning electron microscope. Stigma and anther morphology were observed. Martınez. C. 20 Plant Biology 13 (Suppl. Tokyo. we had 20 flowers from each morph. the stamen. Study site In order to examine the morphology of the sexual organs of C. (ii) vibrating the style at a high frequency (>2000 Hz). King & Buchmann 1996) and how vibration frequencies used by pollen-collecting bees may affect pollendispensing rate (Harder & Barclay 1994. much less attention has been paid to the effect of buzzing and vibration frequencies emitted by floral visitors on pollen capture by the stigma. chamaecristoides usually flowers between July and October (rainy season) and flowering lasts ´ for ca. Japan).G. that is endemic to sand dunes along the ´ Pacific and Gulf of Mexico (Moreno-Casasola 1986. and determine the implications of this particular morphology for pollen release. All hand-pollinations were performed between floral morphs (left. Fenster 1995). 1) (2011) 19–24 ª 2010 German Botanical Society and The Royal Botanical Society of the Netherlands .s. Buchmann 1996). Mexico (19º36¢N. and a The role of buzzing in pollen deposition and fertilisation was studied through four types of experimental hand-pollination: (i) vibrating the style at a low frequency (£400 Hz). <50 m a. Wolfe & Estes 1992. The self´ compatible (Arceo-Gomez G. The warm. we randomly selected 10 plants to account for floral morphology variability within the population.and right-styled flowers on the same individual) having the buzz-pollination syndrome. Martınez et al. the style and in (B) the right-styled morph. personal observation). 1994). Thus. 1. capture and deposition on stigmas. at 07:00 and 17:00. JEOL. taken through a dehydration series. (iii) pollination without vibration. Centro de Investigaciones Costeras La Mancha) in the state of Veracruz. Flowers were collected and preserved in standard formaldehyde-acetic acid-ethanol solution (1:1:8) in the field. as reported for other buzz-pollinated species (Thorp & Estes 1975.

We used these as intermediate frequencies of those reported for bees and frequencies previously used to quantify pollen removal (up to 4000 Hz) (Harder & Barclay 1994.e. which account for the 30%. and Protoxaea sp. Plant Biology 13 (Suppl. 2. Euglossa sp. The development of fruits on all manipulated and un-manipulated flowers (fruit set) was recorded for each treatment as a measure of reproductive success. In the case of artificial vibration. Apical zones of the (A) right and (B) left style (in the morning). Two vibrating instruments were used: a tuning fork for low vibrations (£400 Hz) and a portable commercial electric shaver for high frequency vibrations (>2000 Hz)... 23% and 19% of total visits. Un-manipulated flowers were marked. 1) (2011) 19–24 ª 2010 German Botanical Society and The Royal Botanical Society of the Netherlands 21 . Vibration was only used to assess pollen capture and deposition on the stigma through trichome relaxation. the fringe of trichomes is tightly closed in both morphs. i. All pollinations (with and without vibration) were performed after cutting the anther tips and placing the anther tube directly on top of the style. both fringed by trichomes covering the stigmatic cavity. We used Chi-square analyses to compare fruit production in the different treatments and control.´ ´ ´ Arceo-Gomez. Martınez. Early in the morning. thus covering it. King & Buchmann 1996). we controlled the confounding effect that the amount of pollen released due to the different vibration frequencies could have on the success of fruit maturation. and analysed using Sound Ruler (2007). and left without any manipulation (controls). because we wanted to test whether buzz pollination occurred independent of the frequency of vibration. chamaecristoides flowers.. although trichomes are more relaxed the stigmatic cavity still closed and the stigmatic surface is not exposed.. This structural design occurs in both floral morphs. and in order to compare artificial with natural vibrations emitted by floral visitors and assess possible implications for reproductive success. Vibrations were recorded and fundamental frequencies were analysed (Harder & Barclay 1994) as described above for artificial vibrations. Hence.. This type of manipulation placed large amounts of pollen in close contact with the tip of the style without the need of vibration for pollen release (this was confirmed from previous observations under a microscope). A B C D Fig. Exomalopsis sp. and vibration frequencies were recorded almost immediately in a closed laboratory (100 m from the field) to avoid interference with any other noise. which are less frequent visitors. thus simulating the natural buzzing of bees. The ‘defensive buzzes’ elicited are biophysically identical to floral sonications (King & Buchmann 2003). Bees were collected in the field while they were visiting C. In addition to the above. unpublished data). In all cases. all styles were vibrated for 3 s. we also examined the vibration frequencies of the six most frequent visitors to C. Recordings were made within 5 cm of the microphone in a closed laboratory to avoid interference with any other noise. respectively (Arceo-Gomez et al. and Florilegus sp. Differences in vibration frequencies between the two vibration treatments were evaluated by recording both vibrations in a ‘wav’ file. Ptiloglossa sp. Vibrations were recorded by inducing bees to buzz inside an entomological net.. manipulated flowers were covered with fine mesh bags before and after manipulation to ensure that ours was the only treatment to which the plants had been exposed. chamaecristoides flowers: Eulaema sp. Frequencies were recorded in the Windows Sound Recorder program. (C) Left and (D) right style tip (in the afternoon). Parra-Tabla & Garcıa-Franco Buzzing stigma: implications for effective fertilisation Artificial vibration was performed at two frequencies. RESULTS Stigma and anther morphology The stigma of Chamaecrista chamaecristoides flowers is located inside a stigmatic cavity that is fringed with short trichomes in the apical zone of the style.

Lewis & Gibbs (1999) described a specialised pollen delivery system in Caesalpinia calycina and Caesalpinia pluviosa in which pollen is scraped from the bee abdomen by the stigmatic fringe hairs and forced into the stigmatic chamber. where pollen is released through two terminal pores located on the anther tips. presumably because of wilting as they only last for 1 day. 3A and B). and their function and impact on pollination and reproduction has been even less studied. 2C and D) appear looser. However. In C. In fact. 2007). This could also be true for C. Frequencies up to 4000 Hz have previously been used to evaluate the role of vibration frequency on pollen release (King & Buchmann 1995). Proenca 1992. bee species Eulaema sp. In contrast. while the tuning fork had a vibration frequency of 400 Hz (±32. 2007). The electric shaver had a vibration frequency of 2990 Hz (±43. Only 5% of the flowers that were vibrated with the tuning fork set fruit. Dulberger et al.001). Exomalopsis sp. Overall. 2B) flowers. 2). these floral traits have been studied in only a few species (Owens & Lewis 1989. and that the frequency of vibration plays an important role in reproduction: high vibration 22 Plant Biology 13 (Suppl. Euglossa sp. personal observation). 30% of the hand-pollinated flowers that were vibrated with the electric shaver set fruit. P < 0. it allows us to test whether buzz pollination occurred independent of the frequency of vibration. The role of buzzing in pollen deposition The two artificial vibration methods employed resulted in markedly different vibration frequencies. Dramatic differences were observed in fruit production among the experimental pollinations (X2 = 21. Although trichomes were loose in the afternoon. chamaecristoides.38. mean frequency (Hz) 731 624 538 191 164 128 range (Hz) 667–796 538–796 452–581 150–230 150–193 107–193 SD ±59 ±98 ±35 ±74 ±24 ±40 N 6 6 6 3 3 3 differences of each structure. 1) (2011) 19–24 ª 2010 German Botanical Society and The Royal Botanical Society of the Netherlands . which also reflects the functional Table 1. 1994. 1994. None of the hand-pollinations performed without vibration produced fruit. Protoxaea sp. DISCUSSION The large diversity of flower forms in angiosperms corresponds with the variation in pollen morphology and stigma structure (Endress 1994). Mean (±SD) vibration frequencies emitted by floral visitors of Chamaecrista chamaecristoides. Similarly. 3. Similar types of stigma have been reported for some other species of Fabaceae in the genera Cassia. allowing pollen grains to enter the stigmatic cavity. We observed that pollen is ejected through the two terminal pores located at the tips of the anthers in both floral morphs (Fig. Such action by floral visitors enhances the contact of pollen with the stigmatic surface and facilitates subsequent fertilisation. 2007) and for some species of Malpighiaceae (Sigrist & Sazima 2004). Marazzi et al. N = 3). the stigmatic cavity remained closed and the stigmatic surface inside was not exposed. so pollen deposition on the stigmatic surface can be only achieved through vibration.(Fig. This was similar to the percentage of naturally pollinated flowers (27%) that set fruit (Fig.. as reported for other buzz-pollinated flowers (Buchmann 1983. or to keep pollen grains inside the stigmatic cavity. right. The stigmatic cavity does not open by itself during the time flowers remain open (see results Fig. N = 3). chamaecristoides only occurs when the style is vibrated (none of the hand-pollinations performed without vibration produced fruits). chamaecristoides mechanic vibration emitted by pollinators is required to loosen the trichomes. the Fabaceae present specialised floral structures for buzz pollination that mainly include anther form and structure. Although the frequency of the electric shaver is much higher than natural frequencies emitted by bees. Marazzi et al. Vibration frequencies emitted by distinct floral visitors ranged from 128 to 731 Hz (Table 1). 4). Parra-Tabla & Garcıa-Franco A B Fig. chamaecristoides presents a stigmatic cavity at the tip of the style that is covered by short trichomes. Senna and Chamaecrista (Dulberger et al.Buzzing stigma: implications for effective fertilisation ´ ´ ´ Arceo-Gomez. The anthers of Chamaecrista chamaecristoides have poricidial dehiscence. df = 3. but in this species pollen would be removed by the stigmatic fringe ´ hairs from the backside of the visiting bees (Arceo-Gomez G. 2A) and left-styled (Fig. 1994). Martınez. The differences found among pollination treatments in terms of reproductive success suggest that effective pollination in C. C. Terminal anther pores of (A) right and (B) left anthers from which pollen is released. contrary to most angiosperms that have an exposed stigmatic surface (Dulberger et al. stigma architecture (Marazzi et al. In the afternoon. but in some cases. Sigrist & Sazima (2004) found that the stigmatic cover was removed mechanically by the floral visitors during pollination in Malpighia species. the trichomes of both types of flower (Fig. Endress 1997). Ptiloglossa sp. Wang et al. (2005) suggested for Caulokaempferia coenobialis that the hairs located along the style surface originally functioned to dislodge pollen grains from visitors. Florilegus sp. In this sense.

because ´ they lack the ability to vibrate (e. American Journal of Botany.. unpublished data). 2.. Chapman H. Instituto ´ ´ de Ecologıa A. Vergara for bee determination. minimising the likelihood of receiving mixed pollen loads (pollen from different sources) inside the stigma. only achieved a 1% fruit set ´ (Arceo-Gomez et al.S. 1350–1360. Corbet S. Scientific and Academic Editions. Guevara. Little R. Dulberger R.. the adaptive significance of floral form continues to be revealed. Martınez.A. Mexico. New York. Endress P. Other studies have suggested that the presence of trichomes surrounding the stigma only protects the secretion required for pollen germination from desiccation (Dulberger et al. In this study. Senna and Chamaecrista: morphological variation. our results show that C. Our work with C. (1988) Vibratory pollen collection and flower form: bumble-bees on Actinidia. Borago and Polygonatum. so these bee species may be the main pollinators of C.. Marazzi et al.. 4. Williams I. Buchmann S.L. and thus allowing out-crossing through vibration by the appropriate pollinators.. A. Only pollinators large enough to vibrate with enough force will achieve effective pollination. Special thanks to T. (Eds). Natural vibration frequencies emitted by floral visitors of C. Arboles y arbustos de la Reserva Natural de la Mancha. REFERENCES Buchmann S. A high out-crossing rate was observed through the use of genetic markers in a related species of Chamaecrista (Fenster 1991. (1981) The floral biology of Cassia didymobotrya and C. USA: pp 125–142. The conservation of bees. 147–155. M. In conclusion. Fruit-set is the percent of flowers pollinated that produce a fruit. frequency produced substantially more fruits than low vibration frequency. O’Toole C.g. Handbook of experimental pollination biology. Buchmann S. USA.L. 68. Bulletin of the Southern California Academy of Science. (1996) Competition between honey bees and native bees in the Sonoran Desert and global bee conservation issues. Even after many years of investigation. where flowers were only vibrated but pollen was not in contact with the style tip. Symphytum. In: Matheson A. Fenster et al.. these frequencies were estimated for floral visitors to flowers where vibration is only necessary for pollen release. 2007). (1974) Buzz pollination of Cassia quiedondilla (Leguminosae) by bees of the genera Centris and Melipona. Cambridge University Press. personal observation). respectively). Apis) (Arceo-Gomez G. Westrich P.. (1994) Diversity and evolutionary biology of tropical flowers.B. Medina M. Parra-Tabla & Garcıa-Franco Buzzing stigma: implications for effective fertilisation Fig. chamaecristoides. showed that higher vibration frequencies remove more pollen from anthers than lower vibration frequencies. manipulations at 400 Hz only set 5% of fruits) may play a more important role in achieving fertilisation. chamaecristoides range from 128 to 731 Hz. In addition Harder & Barclay (1994).. auriculata (Caesalpinaceae). Angeles for helping with electron microscopy procedures. Floral visitors that buzz at higher frequencies (over 400 Hz. (19%). Dulberger R. 81. (1994) The stigmatic orifice on Cassia. a fitting tribute to Charles Darwin on this bicentennial of his birth. Koptur. Xalapa. Our results suggest that. buzz at frequencies above 400 Hz (731 and 624 Hz. N = 40 flowers for each pollination treatment (all cross pollinations)...L. Hernandez for helping during fieldwork and C. Eulaema sp (30%). USA: pp 73–113. Bawa K. 171–173. However.H. function during pollination and possible adaptive significance. Smith M. 1390–1396. trichomes may help to prevent pollen interference. The flowers will not release pollen to just any visitor. and higher frequencies may be needed to loosen the trichomes and allow pollen deposition into the stigmatic cavity. a graduate scholarship provided by CONACYT (No: 206021 to GAG) and INECOL financial resources (Ref: 10144 to JGGF). Liu & Koptur 2003). two of the three most frequent floral visitors. (2002). NY.. 23 Plant Biology 13 (Suppl.´ ´ ´ Arceo-Gomez. 2003). The flowers are highly dependent of buzz pollinators. at least in C.K.E. Functional Ecology. a six-fold increase (30% versus 5%). without taking into account stigma morphology (Fenster & Sork 1988. Academic Press. Manfred Ayasse and an anonymous reviewer for their useful corrections and suggestions on the manuscript. ACKNOWLEDGEMENTS ´ We thank S. R. Reproductive success of the different hand-pollinations performed (see methods for explanation).C. Ver.L. chamaecristoides flowers have a highly elaborate and adapted morphology for buzz pollination. (Eds). unpublished data). Laez and G. Being unaware of this morphological trait could lead to misleading results and to a wrong interpretation of the mating system in these species. In: Jones C.J. Although other studies have reported vibration frequencies £400 Hz for other bee species (Harder & Barclay 1994. Saville N. New York. Vallejo-Marın. 1) (2011) 19–24 ª 2010 German Botanical Society and The Royal Botanical Society of the Netherlands . Buchmann 1996). (1983) Buzz pollination in angiosperms. it is important to mention that the presence and correct position of pollinators inside the flower is also essential. chamaecristoides.. suggesting differential pollination efficiency between bee species. This work was supported by SEMARNAT-CONACYT (Ref: 23669 to MLM). New York. since previous manipulations. nor can they receive pollen unless the visitor vibrates the flower at the correct frequency. ´ Castillo-Campos G. and vibration is necessary not only for pollen release but also for pollen capture and deposition inside the stigma. 1994. Buchmann S. Veracruz. which account together for the 49% of total ´ visits (Arceo-Gomez et al. and Florilegus sp. Other studies have evaluated the mating system of Chamaecrista species by means of standard hand-pollination procedures. 73. This mechanism could also prevent self-pollination via floral visitors that only move stray pollen within the flower. chamaecristoides illustrates the importance of understanding the morphology of a flower’s sexual organs for complete understanding of reproductive biology in a plant species. American Journal of Botany.

79. 1898–1903. 43–53. (2004) Pollination and reproductive biology of twelve species of Neotropical Malpighiaceae: stigma morphology and its implications for the breeding system.J. (1994) Sobrevivencia ´ ´ y crecimiento de plantulas de un arbusto endemico de dunas ´ ´ costeras ante condiciones de sequıa. 24 Plant Biology 13 (Suppl.J.R. (1996) Floral traits and plant adaptation to insect pollinators: a devil’s advocate approach. Harder L... Martınez. Liu H. 99–118.D. Wang Y..B.P. 26. (2003) Mesosomal vibration by Bombus and Xylocopa.R. 67–76. 93–105. (2005) Self-pollination by sliding pollen in Caulokaempferia coenobialis (Zingiberaceae). 295–305.Buzzing stigma: implications for effective fertilisation ´ ´ ´ Arceo-Gomez. Acta Botanica Mexicana. Annual Review of Ecology.E. American Journal of Botany. Functional Ecology. 53–62. (1999) Reproductive biology of Caesalpinia calycina and C.S. 371–391. International Journal of Plant Sciences. Fenster C. 5–13. 410–422. Plant Systematics and Evolution.sourceforge. Thomson J. 166.L.. (2004) Pollination syndromes and floral specialization. (1988) Effect of crossing distance and male parent on in vivo pollen tube growth in Chamaecrista fasciculata. 49–61.. (1992) Buzz pollination – older and more widespread than we think? Journal of Tropical Ecology. (Eds). Sork V. (2005) Floral biology and reproductive system of enan´ tiostylous Senna corymbosa (Caesalpiniaceae).. 333– 345. 85. pluviosa (Leguminosae) of the caatinga of northeastern Brazil.J. (1997) Relationships between floral organization. (1991) Gene flow in Chamaecrista fasciculata (Leguminosae) II. Chamaecrista and Senna (Leguminosae: Caesalpinioideae). (1986) Sand movement as a factor in the distribution of plant communities on a coastal dune system. 206. 48. 46–50. 753–759. (2000) Control gates and micro-ecology: the pollen–stigma interaction in perspective. 82.. Herrera C. Functional Ecology. 375–403. (1995) Mirror image flowers and their effect on outcrossing rate in Chamaecrista fasciculata (Leguminosae)..M. Wolfe A. Journal of the Kansas Entomological Society.G. 57. Armbruster W. 76. Vekemans X. Buchmann S. (2007) Diversity in anthers and stigmas in the buzz-pollinated genus Senna (Leguminosae. Fenster C. (1989) Taxonomic and functional implications of stigma morphology in species of Cassia. Senna. but not Apis (Hymenoptera: Apidae).net/main/ (accessed August 2009). Renner S. Sazima M. Chapman & Hall..B. 168.R.D. 8.. International Journal of Plant Sciences. (2003) Quantifying gene flow from spatial genetic structure data in a metapopulation of Chamaecrista fasciculata (Leguminosae). (1995) Bumble bee-initiated vibration release mechanism of Rhododendron pollen. 1) (2011) 19–24 ª 2010 German Botanical Society and The Royal Botanical Society of the Netherlands .. 65. American Journal of Botany. (1992) Pollination and the function of floral parts in Chamaecrista fasciculata (Fabaceae). Evolution. Ver. Fenster C.L. Available from http:// soundruler.B. Lewis G. 217.R. Gibbs P.. Zhang D. Endress P. Moreno-Casasola P.6 software. Sound Ruler (2007). 687–711. ejects pollen from poricidal anthers. 75. Plant Systematics and Evolution. New York.K. 314–317. 0.. (1994) The functional significance of poricidal anthers and buzz pollination: controlled pollen removal from Dodecatheon. Plant Ecology. 163. Conti E. In: Lloyd D.. Dudash M. Owens S. (1975) Intrafloral behavior of bees on flowers of Cassia fasciculata. 509–517. (1996) Trends in evolution of floral ontogeny in Cassia sensu stricto. Annals of Botany. 1180– 1187. Floral biology: studies on floral evolution in animalpollinated plants. Chen Z. Thorp R.L. Martinez M.. 90. King M. Koptur S.K. (2003) Facilitation of seedling establishment by an endemic shrub in coastal sand dunes.. Marazzi B. 83. 8. Evolution and Systematics.L. 45.W.. Gene establishment. Fenster C. Proenca C. Barrett S.C. Cassinae). 168.B. Moreno-Casasola P.J.. 33–41. King M. and Chamaecrista (Leguminosae: Caesalpinioideae: Cassieae: Cassiinae): a study in convergence. 82. USA: pp 65–87.9.B... Plant Systematics and Evolution. 1407–1411. American Journal of Botany. Heslop-Harrison Y.C.. Parra-Tabla & Garcıa-Franco Endress P. Tucker S. Laporta C.. Annals of Botany. 995–1007.S. Hardy O.. Conflicts of interest added after online publication: The authors have declared no potential conflicts. Buchmann S. 10. 94. American Journal of Botany. Estes J. Fenster C.J.. Wilson P. 53. Evolution. (2003) Breeding system and pollination of a narrowly endemic herb of the lower Florida keys: impacts of the urban–wildland interface. 175–184. American Journal of Botany.H.M. Lewis G. ´ ´ Martınez M. 115–120. architecture and pollination mode in Dillenia (Dilleniaceae). 449–456.D. Rincon E. 35. Estes J. Journal of the Kansas Entomological Society. Revista de Biologıa Tropical. Sigrist M. Buchmann S. American Journal of Botany. King M. (1996) Sonication dispensing of pollen from Solanum laciniatum flowers.. Barclay R. Vegetatio.