The role of height and flower symmetry in bat pollination of Mucuna urens Papilionaceae Callie Vincent Department of Biology, Kenyon College _____________________________________________________________________________________ ABSTRACT The flowers of plants in the tropics have evolved mechanisms that attract bats and ensure that pollination is deposited on the bat for transfer to the next flower Altringham 1996. In particular, a recent study has shown that flowers in the Mucuna genus attract bats usi ng an acoustic nectar guide vonHelverson and vonHelverson 1999. In order to test additional factors that could possibly play a role in the pollinator plant relationship involving Mucuna inflorescence height, infructescence height, and flower symmetry we re studied for five Mucuna urens patches in the Monteverde Puntarenas, Costa Rica cloud forest. A simple regression showed no significant relationship between flowers pollinated and inflorescence height Simple Regression; p = 0.1724, R 2 = 0.021. Altern atively, a regression of fruits per infructescence versus infructescence height showed a positive relationship Simple Regression; p < .0001, R 2 = 0.254. It is likely that this significance can be attributed to more plant clutter around lower inflorescenc es, making it more difficult, but not impossible, for the bats to locate them. Finally, unpaired t tests used to analyze the relationship between flower symmetry and pollination revealed that there was not a significant difference between pollinated and no n pollinated flowers for any of the points measured unpaired t tests; a p = 0.5557, b p = 0.4036, c p = 0.4511. Possible reasons for this involve the trap lining behavior of nectar bats and nectar peaks in the flower throughout the night. High diff erences in variation between pollinated and non pollinated flowers for the point at the top of the flower can possibly be attributed to effects of pollination on floral attraction and longevity. RESUMEN Las flores de plantas en los tr picos han evolucionado mecanismos que atraen murcilagos y aseguran la transferencia de polen entre flores. Altringham 1996. Un estudio reciente mostr que flores en el gnero Mucuna atraen murcilagos usando una gua acstica de nctar vonHelverson y vonHelverson 1999. En este estudio se examinaron los factores adicionales que posiblemente expliquen la relacin de Mucuna y su polinizador, como es la altura de las flores, altura de frutas y la simetra de flores; en cinco reas diferentes de Mucuna ur ens en el bosque nuboso de Monteverde, Puntarenas, Costa Rica. Se encontr una relacin significativa entre las flores polinizadas y la altura de la inflorescencia Regresin Simple; p = 0.1724, R 2 = 0.021. Alternativamente, se encontr una relacin posit iva entre el nmero de frutos contra la altura Regresin Simple p < .0001, R 2 = 0.254. Es probable que la diferencia sea atribuida a una conglomeracin alrededor de las inflorescencias ms bajas hacindolo ms difcil, pero no imposible, para los murci lagos polinizarlos. Finalmente, no se encontr una diferencia entre la simetra del vexilium en flores polinizadas y las que no lo fueron en ninguno de los tres puntos medidos Prueba de t, no pareada; a p = 0.5557, b p = 0.4036, c p = 0.4511. Las p osibles razones para explicar esto es el comportamiento rutero de los murcilagos y los picos de produccin de nctar de las flores durante la noche. Las diferencias tan altas en la variacin entre flores fertilizadas y no en el punto ms alto del
vexilium c de la flor pueden ser atribuidas a efectos en la longevidad de las flores que ya fueron polinizadas. INTRODUCTION Flowers specialized for bat pollination are restricted primarily to the tropics LaVal and Rodr guez 2002. Unlike in temperate zones, animals in the tropics are able to participate in pollinator plant relationships, and bats are especially ideal because they are mobile and can provide long distance movement of genetic material. Flowers depending on bats for pollination frequently show distinct adaptations. They usually produce copious amounts of pollen and mucilagous nectar, and often open at dusk Endress 1994. Some flowers are white, but most have dull colors such as green or brown, and many have a characteristic, fermented smell Endress 1994. Finally, inflorescences are often produced in a relatively open space, and the corollas are generally robust in order to survive the visits of their pollinators Tschapka and Dressler 2002. Recent studie s have shown that flowers might also be adapted to acoustically attract bats. Von Helversen and von Helversen 1999 found that the bat pollinated neotropical vine Mucuna holtonii Papilionaceae directs its echolocating pollinators to its flowers by means of an acoustic nectar guide. When they moved a microphone in an arc around the concave part vexillum of the open flower, they found that most of the energy was reflected back in the direction of the sound source, indicating that the geometry of the flow er may be an adaptation to the echolocation system of its bat pollinator. A subsequent study on a similar species in the Monteverde region, Mucuna urens Papilionaceae, has supported these results and additionally found that an increase in vexillum size c orrelated with number of bat visits Macedo 2002. Echolocation is defined as the analysis by an animal of the echoes of its own emitted sound waves, by which it builds a sound picture of its immediate environment Altringham 1996. Bats emit sounds in sm all bursts and then detect objects in their surroundings by listening to the echo of each emitted pulse. Acoustic signals are produced in the larynx and emitted from the facial region, a combination of circumstances that could cause the outgoing signals to mask the much fainter echoes Fenton et al. 1995. Echolocation tends to be mentioned in terms of its role as an important perceptual modality for prey location in bat species that are predatory Simmons et al. 1979, but the Mucuna pollination system pre sents another very important role for echolocation. Von Helversen and von Helverson 1999 found Glossophaga commissarisi Phyllostomidae to be the main pollinator of M. holtonii Bats in this genus are small body length less than 60 mm, have a short f oraging distribution relative to similar genera, and use a solitary method of foraging Heithaus et al. 1975. They also appear to be most active shortly after dusk and just before dawn. Despite the fact that they have a small body size, their tongues are long and can reach deep into flowers. The tip of the tongue is equipped with brushlike papillae that are used to collect nectar by capillary action von Helversen and von Helversen 1975. A characteristic of flowers that has been shown in relation to inse ct preference, but not to bat preference, is symmetry. A study involving flower models found that insects and Hymenoptera that visited the models clearly preferred symmetrical models to
asymmetrical ones, and the ranking of visits to the models reflected a preference for large, symmetrical flowers Moller and Sorci 1997. Another investigation of flower symmetry showed that flowers visited by bumblebees were larger and more symmetrical than the nearest neighboring flower, and also that nectar production was larger in the symmetrical flowers Moller 1995. These experiments suggest that insects with a preference for symmetrical flowers will benefit in terms of resource gain from their foraging. The purpose of this study was to test two new factors that could possibly play a role in the pollinator plant relationship between M. urens and G. commissarisi bats: height and symmetry. A past study has shown that number of fruits per infructescence and height of the infructescences were each significantly different b etween three sights in the Monteverde area Mullaney 2001, but it did not test for a relationship between the two. This study investigated both flowers and fruits and their relation to height and pollination success. It also determined if symmetry plays a role in pollination, as it seems likely that a similar pattern of symmetrical preference could exist in bats as well as insects. MATERIALS AND METHODS The study was conducted between the days of April 10 and May 3 in the lower montane forest of Monteverde, Puntarenas, Costa Rica. Study organism: Mucuna urens Mucuna urens is one of the most common leguminous vines associated with both disturbed habitats and primary forest in the central highlands of Costa Rica Young 1983. Its penduliflorous in florescences hang down from the vines, and on each flower two petals constitute the keel, two lateral petals form the shorter Â€wingsÂ, and a fifth, upper petal forms the triangular Â€standardÂ, or vexillum, which is raised when the bud comes into flower vo n Helversen and von Helversen, 1999. The pollen is released explosively, meaning that it is easy to detect a pollinated flower because the keel is open and the reproductive organs exposed. Fruits are covered with a dense layer of tiny reddish brown irrita ting hairs, and both fruits and flowers can be found during the dry season Young 1983. Five patches of M. urens were used for data collection. Pollinators for the sites are likely G. commissarisi as nearby mist netting found that 28% of individuals of t hat species had M. urens pollen on their bodies D. Cohen unpublished data. a Pollination At each location, inflorescences reachable by a small ladder up to 3.5 meters were tagged using clear straws; each flower was designated a number and letter. Prior to dusk, the state of each flower open, closed, or pollinated was recorded. At 9:00 p.m., inflorescences were covered with small mesh bags in order to prevent any other types of pollinators from visiting the flowers. The following morning, the bags were removed and
the state of each flower was recorded in order to see if pollination had occurred. Inflorescences were observed from three to five nights, depending on how many open flowers were present. A regression analysis was used to analyze the data b Fruit The height and amount of fruit was measured at each of the locations used for the pollination analysis, and took place during the last two days of data collection in order to give more time to developing fruits. In this study, fruit served as a secondary m ethod of investigating pollination, under the assumption that more fruit per infructescence meant that more flowers had been pollinated. All infructescences within reaching distance were measured with a tape measure; taller ones were measured by placing a meter stick up against a tree and then stepping at least ten meters away in order to estimate the height with a ruler. Fruits on the infructescence often varied from a greenish/orange color on the newly formed fruit to a dark brown on the more developed fr uit; only the brown fruits were counted. The number of fruits on each infructescence was recorded along with the height, and a regression analysis was run on the data. c Symmetry On the final day of data collection for an inflorescence, individual flowers were collected if the inflorescence contained a sampling of both pollinated and nonpollinated flowers. The vexillums of the flower were measured with a caliper 0.01 mm at three locations on left and right sides: the bottom, a standardized midpoint, and t he top Figure 1. The bottom was measured at the lowest horizontal plane of the vexillum, the midpoint was measured at the dip in coloration found on all of the flowers, and the top point was measured halfway between the dip in coloration and the tip of t he flower. The vertical midpoint of the flower was determined by the tip. Unpaired t tests were used for each of the locations on the flower. RESULTS a Pollination Out of a total of 91 flowers analyzed, only 20 were pollinated. There was not a significant relationship between pollination and inflorescence height Simple Regression, p = 0.1724, R 2 = 0.021, N = 91 Figure 2. After determining an index which took into account the number of original open flowers on the inflorescence and the number pollinated all but six flowers fell below the index of three, meaning that pollination levels were low. Despite the fact that the levels are low, they do show that pollination was occasionally occurring at low heights.
b Fruit A total of 263 fruits were analyzed in relation to height, with an average of 2.26 fruits per infructescence SD = 1.398. Number of fruits per infructescence was positively correlated with height Simple Regression; p < .0001, R 2 = 0.254, N = 263 Figure 3. The average height of an infructescence was 402.95 cm SD = 305.774. c Symmetry A total of 58 flowers were measured for symmetry, with 31 being pollinated flowers. Symmetry was not significantly related to pollination for any of the three measured points unpaired t test; a p = 0.5557, b p = 0.4036, c p = 0.4511, N = 59 Figure 4 See Figure 1 for positions of the measurements. For points A and B, levels of variation standard deviation between pollinated and non pollinated flowers were similar, with respective differences of 0.177 and 0.026. However, the difference between v ariation for point C was 0.699, with more variation in the pollinated flowers. DISCUSSION Out of 91 analyzed inflorescences, only about 22% were pollinated, showing that flowers below the height of 3.5 meters are being pollinated at low rates and suggest ing that height does not play a role in the relationship between M. urens and Glossophaga bats Figure 2. However, the results from fruit height showed a positive relationship between the number of fruits and the height from the ground Simple Regression; p < 0.0001, R 2 = 0.021, Figure 3, indicating that higher inflorescences get pollinated more. It seems likely that the results from the inflorescences simply did not represent a large enough spatial scale in terms of height to show the complete relationsh ip between height and pollination. Following germination, M. urens grows up small understory trees, ultimately reaching lengths of more than 70 meters and heights of more than 20 meters in the canopy OÂDell 2000. Because the densest foliage in rainforest s commonly lies under 20 meters and is especially thick in the understory Terborgh 1992, it is probable that higher inflorescences are getting pollinated more because they are more likely to be in open area. To facilitate detection by bats, chiropterophi lous flowers and inflorescences often protrude from the foliage Tschapka and Dressler 2002, and in the case of M. urens inflorescences hang down from peduncles that place them in any available open areas. Although results indicate that it is advantageo us for M. urens inflorescences to hang in the open, they also show that some flowers are getting pollinated at lower heights, even if at low densities Figure 2. Because M. urens thrives in secondary growth where it may not always have open area in which to place its inflorescences, there must be mechanisms in place to aid in pollination of lower inflorescences. First, temporal differences in flowering and fruiting patterns are important in bat pollinated flowers and can be characterized by two extremes A ltringham 1996. Some plants exhibit Â€big bangÂ reproduction, producing vast numbers of flowers over just a few days, while other plants are Â€steady stateÂ and produce just a few flowers each night over a long period Altringham 1996. The flowering patter n of M. urens seems to fall in between the two
extremes in that it produces large amounts of flowers throughout the flowering season. This increases the chances of an inflorescence being discovered by bats and also increases the chance that an inflorescenc e will end up in an open area. In addition, bats will be more likely to return to a foraging area if there is a high supply of food. A second factor that could aid in the pollination of inflorescences at lower heights involves the physiological demands of nectar bats. These bats have high metabolic rates and no means of storing large energy reserves Lemke 1984, meaning that they must eat continuously during foraging time. Flowers in open areas are likely to be pollinated first, but as these food resource s are depleted, it becomes more efficient for the bats to expend the energy to find the inflorescences at lower heights in order to meet their metabolic needs. In terms of energetics, any potential food sources must be available in sufficient quality or qu antity or both to compensate for the energy expended to obtain it Lemke 1984. Lower inflorescences are likely to be surrounded by foliage and may be harder to locate, but once the bat reaches the inflorescence, the flowers are likely to contain more nect ar than the higher inflorescences that have already been foraged upon. In order to account for the energy lost in inflorescences that do not get pollinated sufficiently; M. urens may abort some inflorescences. Mangifera indica Anacardiaceae aborts many of its flowers, even if they have been fertilized, because the production of fruit is very energetically costly Hansen 1983. In a similar fashion, Cocos nucifera Aracaceae drops approximately 50 60% of its female flowers before they begin to mature into fruit Vandermeer 1983. Mucuna urens allocates much of its energy to high flower production, so it seems possible that it would only invest further energy into those inflorescences that are likely to have the highest fruit yields. Measure ments of three points on M. urens vexillums showed that there was not a difference in terms of symmetry between pollinated and non pollinated flowers Figure 4, indicating that symmetry is not a factor in the M. urens and Glossophaga pollination system. S tudies involving insect pollinators show that flowers visited by insects are more symmetrical and that nectar production is larger in symmetrical flowers Moller 1995. This suggests that insects will benefit from visiting more symmetrical flowers because they will receive a greater reward for pollination. Nectaries of bat pollinated flowers often produce large quantities of nectar Â‚ five to ten mL of nectar when full Altringham 1996 Â‚ but in the case of M. urens this nectar load is distributed among man y individual flowers. Therefore the symmetry of an individual flower may not reflect the nectar producing ability of the inflorescence as a whole, and it may not be advantageous for a bat to select a more symmetrical flower. There are several other factor s suggesting why symmetry may not be as important in M. urens flowers in terms of pollination success as it is in other species. Glossophaga species have been shown to forage along trap lines, returning to a single flower many times Altringham 1996. This study found that the average number of open flowers on an inflorescence was 5.2, suggesting that multiple flowers on a single inflorescence can be detected by the bats and foraged upon. These two factors together indicate that the bats are moving between inflorescences, feeding on numerous flowers at each. It seems probable that bats locate an inflorescence using the vexillum of one flower and then feed on others at the inflorescence because it is energetically efficient, not because particular flowers are more symmetrical. In addition, it has been shown that some bat flowers exhibit peaks in nectar production throughout the night. Bats always draw all of the
nectar from Bauhinia pauletia but a return visit after an average interval of 21 minutes allows ti me for fresh nectar to accumulate in the flowers Proctor et al. 1996. If flowers have varying amounts of nectar throughout the night, bats will benefit by moving from flower to flower in order to obtain the maximum food supply, and again symmetry would n ot be important. These hypotheses indicate that symmetry may not influence this bat pollinator relationship, but a lab study in which bats are presented with both symmetrical and nonsymmetrical flowers could provide further insight. Although symmetry mea surements were not significant between pollinated and non pollinated flowers for any of the points measured, the results from point C are interesting because they show a much greater difference in variation between flower types than do points A and B Figu re 4. Studies have shown that in several plant species the flowers are long lived when not pollinated, but show petal withering or petal abscission shortly following pollination van Doorn 1997. It seems likely that if the M. urens vexillums were similar ly affected, point C, because it is farthest from the flower, would be affected first. As the flower begins to change shape, there would be more variation among the flowers that have already been pollinated. Flowers cease to be attractive to pollinators by a range of cues such as cessation of scent production, a color change, permanent flower closure, and petal senescence van Doorn 1997. Field observations from this study indicated that after pollination, vexillums changed from green to black and began to shrivel within days, again suggesting that there would be more variation among the pollinated flowers. In addition to inflorescence height, there are several other factors that may play a role in the M. urens pollination system. In this study, the number of flowers per inflorescence ranged from one to 21. A study in the Monteverde area showed that an increase in vexillum size correlated with number of bat visits Macedo 2002. Vexillum size and total number of flowers are factors that therefore vary between inflorescences and could have implications for pollination success. Furthermore, most species in the Glossophaga genus feed on fruit and insects as a supplement to nectar Reid 1997. It seems possible that fluctuations in the populations of fruit and insects could also affect the pollination rate of M. urens Because bats are rather large pollinators they are comparably expensive partners for plants, but they also offer some unique advantages: they generall y visit a comparably small number of plant species and they may transport pollen over long distances Tchapka and Dressler 2002. The relationship between M. urens with its unique acoustic nectar guide von Helverson and von Helverson 1999 and Glossophaga species with their special techniques for locating the flowers Fenton et al. 1995 exemplifies the intricacy with which pollinator plant systems have evolved in the tropics. Past findings support the idea that although M. urens does well in fragmented ar eas, it can do significantly better in the interior of a large forest, given that sufficient area is available Mullaney 2001. Large continuous forest provides enough area to harbor populations of Glossophaga species, which roost in hollow trees, tunnels, and caves Reid 1997. The relationship between M. urens and bats stresses the importance of protecting biodiversity and large forest plots in order to conserve habitats suitable to the species and to protect the interaction between them.
ACKNOWLEDGEMENT S I would like to extend much thanks to Mauricio for his continuous help with this project, from number crunching to Spanish to brainstorming Â‚ everything. Thanks so much to my lovely TAÂs, Rick and Andrew, for having the answers to all of my endless ques tions. Rick, thanks for teaching me to always save to the computer, and Andrew, thanks for the Mucuna hike and for being so generous with the Mucuna patch at your house. I am grateful to La Estacin Biolgica Monteverde for the use of their forest and for giving us a place to work and play. I would like to thank D. Mauricio Cohen for sharing his bat information. Special thanks to my girls for all of the laughter and fun times, and thanks to the entire group for making Costa Rica such a beautiful experience. LITERATURE CITED A ltringham, J. 1996. Bats: Biology and Behavior. Oxford University Press, Oxford. Endress, P.K. 1994. Diversity and evolutionary biology of tropical flowers. Cambridge University Press, Cambridge. Fenton, M.B., D. Audet, M.K. Orbist, and J. Rydell. 1995. Signal strength, timing, and self deafening: the evolution of echolocation in bats. Paleobiology 21: 229 242. Hansen, M. 1983. Mangifera indica Mango. In: Costa Rican Natural History, D.H. Janzen, ed. The University of Chicago Press, Chicago, IL, pp. 95 97. Heithaus, E., T. Fleming, and P. Opler. 1975. Foraging patterns and resource utilization in seven species of bats in a seasonal tropical forest. Ecolog y 56: 841 854. LaVal, R. and B. Rodriguez H. 2002. Bats of Costa Rica. INBio, Costa Rica. Lemke, T. 1984. Foraging ecology of the long nosed bat, Glossophaga soricina with respect to resource availability. Ecology 65: 538 548. Macedo. M. 2002. Function and size optimization in Mucuna urens vexillum. CIEE student project, unpublished. Moller, A.P. 1995. Bumblebee preference for symmetrical flowers. Proceedings of the National Academy of Sciences of the United States of America 92: 2288 2292. Moller, A. P. and G. Sorci. 1997. Insect preference for symmetrical artificial flowers. Oecologia 114: 37 42. Mullaney, M. 2001. The fruit production of Mucuna urens in continuous forest and forest fragments. CIEE student project, unpublished. OÂDell, G. 2000. Mucuna urens : A tropical liana. In: Monteverde: Ecology and Conservation of a Tropical Rainforest. Nadkarni, N. and N. Wheelright. Ed Oxford University Press, New York, pp.72. Proctor, M., P. Yeo, and A. Lack. 1996. The Natural History of Pollination. Tim ber Press, Oregon.
Reid, F. 1997. A Field Guide to the Mammals of Central America and Southeast Mexico. Oxford University Oxford. Simmons, J.M. Fenton, and M.OÂFarrel. 1979. Echolocation and pursuit of prey by bats. Science 203: 16 21. Terborgh, J. 1992. Diversity and the Tropical Rain Forest. Scientific American Library, New York. Tschapka, M. and S. Dressler. 2002. Chiropterophily: on bat flowers and flower bats. Royal Botanic Gardens: 114 125. Vandermeer, J. 1983. Cocos nucifera Co conut. In: Costa Rican Natural History, D.H. Janzen, ed. The University of Chicago Press, Chicago, IL, pp. 85 86. Van Doorn, W. 1997. Effects of pollination on floral attraction and longevity. Journal of Experimental Botany 48: 1615 1622. Von Helversen, D. and O. von Helversen. 1975. Glossophaga soricina In: Wolf, G. ed., Encyclopedia Cinematographica Vol. E 1837: 3 17. Von Helversen, D. and O. von Helversen. 1999. Acoustic guide in bat pollinated flower. Nature 398: 759 760. Young, A.M. 1983. Seed mortality and recruitment in the forest canopy vine Mucuna urens in the central Highlands of Costa Rica. Brenesia 21: 13 25.
_____________________________________________________________________________________ Figure 1. Three points measured on the vexillum of M. urens flowers in order to determine symmetry. Point A measured the bottom, point B measured from the dip of differential coloring, and point C measured from the midpoint between the dip and the top of the vexill um. All measurements were taken on the left and right sides of the vexillum, using the tip as the vertical midline. _____________________________________________________________________________________
_____________________________________________________________________________________ Figure 2. No significant relationship was found between M. urens inflorescence height and the number of flowers pollinated per night. The following formula was used in or der to account for the fact that a flower with more open flowers should be more attractive to bats: relationship = number of flowers pollinated + 0.5 / number of open flowers *5. Simple Regression Analysis; p = 0.1724, R 2 = 0.021, N = 91 _____________ ________________________________________________________________________ ________________________________________________________________________ Figure 3. A significant relationship was found between M. urens infructescence height and the number of fru its on each infructescence. Only dark brown fruits were counted. Simple Regression Analysis; p < 0.0001, R 2 = 0.254, N = 263. ________________________________________________________________________
_____________________________________________________________________________________ Figure 4. Absolute values of the mean differences between leaf and right sides of the vexillum in M. urens flowers. Measurements were taken at three locations: a bottom b midpoint, and c top see Figure 1. All flowers started as open the night before measurements: pollinated flowers N = 31 were pollinated during the night and nonpollinated flowers N = 27 remained open in the morning. Error bars on all graphs sh ow one standard deviation. Unpaired t test; a p = 0.557; b p = 0.4036; c p = 0.4511, N = 58 _____________________________________________________________________________________