Function and size optimization in Mucuna urens vexillum Mark Macedo Department of Biology, Occidental College, Los Angeles, CA ________________________________________________________________ Abstract This study examines the role of the vexillum as an aural nectar guide in the species Mucuna urens. Field manipulations show that the vexillum of M. urens enhances visitation and pollen removal X 2 = 52.8, p< .001. Subsequent observational studies conducted on bat visited hummingbird feeders support this finding as well X 2 = 10.67, p< .005. Data also show an increase in vexillum size correlated with number of bats visits X 2 = 32.06, p < .001. Two explanati ons are offered for why M. urens and other bat pollinated flowers in the genus have not developed as large a vexillum as possible. Investment return optimization is one possible explanation. Niche partitioning of potential pollinators is another. Resumen E ste estudio examina el papel del "vexillum" como una gua auricular de nctar en la especie Mucuna urens. Las manipulaciones del campo muestran que el vexillum de M. urens aumenta las visitas y la remocin de polen X2 = 52.8, P<.001. Observaciones en los comederos de colibres que fueron visitados por los murcilagos apoyaron estas observaciones X2 = 10.67, P <. 005. Los datos mostraron que un aumento en el tamao de vexillum tuvo correlacin con el nmero de visitas de los murcilago X2 = 32.06, P <. 0 01. Dos alternativas se ofrecen para explicar porque M. urens y otras flores polinizadas por murcilagos en mismo gnero no han desarrollado un vexillum tan grande como es posible. La optimizacin del regreso de la inversin y la divisin de nicho de los polinizadores potenciales son las posibles explicaciones. Introduction One of the most pressing challenges plants face is finding an efficient method of reproducing sexually. In areas like the Tropics, where a large number of plant species are rare and live in a highly heterogeneous environment, a plant's ability to encounter pollen from conspecifics may be especially difficult. In response to this challenge, plants have developed specialized relationships with insects, mammals, and birds to move pollen Endress 1994. Bawa 1990 estimates that nearly all flowering plants in tropical lowland rain forest are pollinated by animals. With the large number of plants and potential pollinators in the Tropics, evolving very specific plant pollinator interactions greatly increases the likelihood that pollen from conspecifics will rea ch a given individual Feinsinger 1983. This has led to specialized features to attract specific pollinators, commonly referred to as pollination syndromes van der Pijl 1960. Pollination syndromes serve to attract those pollinators that will provide the most effective pollination service Feinsinger 1983. Pollination syndromes can be generalized by
pollinator. Bat pollinated plants, for example, are typically large and often a dull or white in color. They characteristically have a strong odor that somew hat resembles fermented fruit and produce large quantities of nectar Janzen 1975. Bat pollinated flowers also usually flower at night and contain some sort of receptacle to hold pollen as a reward for the bat. Finally, many bat pollinated plants devise s ome way of reducing clutter around the inflorescences, such as hanging the flowers free of the leaves Altringham 1996. In addition to the generalized pollination syndromes, some flowers also have nectar guides to direct their pollinators to the flower. M any types of insect pollinated flowers are known to use UV light patterns as visual nectar guides Brehm and Krell 1975; Penny JH 1983; van der Pijl 1960. The insects that pollinate these flowers are able to see UV light, and are essentially guided toward s the flowers by the visual cues provided by the flower. One of the more interesting and rare examples of a nectar guide, however, is that of an aural guide. The bat pollinated liana Mucuna holtonii Papillionoideae was shown by von Helverson O and von He lverson D 1999 to contain an aural nectar guide that helps bats find the small green flowers using echolocation. The study was done at the La Selva field station in Costa Rica. Flowers of M. holtonii contain a structure called a vexillum also known as a standard that has a concave structure and sits on top of the flower fig. 1. This structure was hypothesized to reflect the echolocation calls of bats, which helps the bat locate the flower. Removing the vexillum resulted in much lower rates of pollinat ion, and tests using microphones supported the hypothesis that M. holtonii does have a structure that acts as a nectar guide by reflecting the calls of bats von Helverson O and von Helverson D 1999. More than 1500 meters above the M. holtonii study site at La Selva there grows a different species in the same genus, Mucuna urens McDade and Hartshorn 1994, Diller GW, 2000. M. urens, like M. holtonii, is known to be bat pollinated and also has a vexillum structure like that of M. holtonii. The p urpose of this study is to examine the role of the vexillum in M. urens, and to test different sizes of vexillum for size optimization. Materials and Methods Site Description: All studies were conducted in the lower montane cloud forest in Monteverde, Costa Rica. The entire study took place from mid April to mid May 2002. The first part of the study was conducted at the Estacin Biolgica Monteverde. All feeder experiments were conducted at the hummingbird gallery, 50 meters from the entrance to the Monteverde reserve. Both sites are at nearly the same elevation, approximately 1550 meters above sea level Masters K, pers. comm.. Mist nets set up between hummingbird feeders for u se in two concurrent studies frequently caught three species of nectivorous bats, Glosso phaga commisserissi Anoura geoffroyi, and Hylonycteris underwoodi, and one species of frugivorous bat, Artibeus toltecus Majewski JJ 2002 ; Miller RM 2002. All of th ese bats are in the family Phyllostomatidae. Of these, the likely pollinators of M. urens are G. commisserissi and H. underwoodi. A. geoffroyi is most likely too large for the relatively small M. urens flower and A. toltecus is a fruit eating bat LaVal, Pers. Comm.; Diller GW, 2000.
Study organism: Mucuna urens: M. urens is a relatively common liana around the Monteverde region and is most often seen high in the forest canopy or along roadsides where bats have easy access to the flowers von Helverson and von Helverson, 1999. Individuals tend to be fairly far away from each other, so bats are the ideal vectors for M. urens, since bats are known to travel up to 16 km in one night to visit a plant Janzen 1975. Inflorescences of M. urens take the form of a long pendant raceme Endress 1994. Up to 20 flowers may be found on a single inflorescence. When a flower on an inflorescence matures, the vexillum open s, also exposing the keel and wings. The pollination strategy of M. urens is explosive. Visitation by a bat triggers the keel to open, which projects the pollen onto the rear of the bat and exposes the sexual organs of the flower Endress 1994. Thus, visi tation by bats is easy to note because of the change in appearance of the flowers. a Testing functionality of M. urens vexillum: The first part of the experiment is necessary to verify that M. urens vexillum serves the same function as the vexillum in M. holtonii as a nectar guide for bats. The methodology of the original La Selva study was copied as closely as possible. Since M. urens flowers open sequentially through the course of a night, the number of open flowers available for visitation was optimize d by covering all mature inflorescences with a large mesh bag between noon and 6:00 pm von Helverson O and von Helverson D 1999. The bags were then removed between 8:30 and 9:00 pm. One half of the flowers had the vexillum cut off and the other half were left as control. Sampling was continued for five nights and a total of 215 flowers were tallied. b Testing visitation and vexillum at feeders: In order to determine whether using detached M. urens vexillum on feeders would solicit more visits by bats; one hummingbird feeder with four holes was assembled. Attached directly behind two of the holes on opposite sides of the feeder were actual M. urens vexilla. The other two holes were left unaltered as a control. The feeder was then obs erved for up to three hours a day between 6:00 and 9:00 pm for four days. A red filter was put on a small 40 watt light, so that just enough red lights were produced to see the feeder. For every bat that came to the feeder, the specific hole visited was wr itten down. Every fifteen minutes to half hour, the feeder was rotated a quarter turn to minimize the possibility of placement affecting the data. c Testing optimal vexillum size at feeders with actual vexilla: This test was done to test the effect of us ing paper vexilla of varied size along with an actual vexillum. Ten actual vexilla were measured and both the height and width were found to be 2.5 . 1 cm. The average of 2.5 was used for subsequent vexilla created using bond paper as a basis for medium p aper vexilla. Two paper vexilla were created, one 5 cm in diameter and height, which is twice the size of an actual M. urens, and one a little more than half the size of the actual, about 1.5 cm in height and diameter. Fake vexilla made of notebook paper w ere shaped into the general form of an actual vexilla and attached to the same four hole feeder as the first feeder experiment. A real vexilla from a
plant was also attached and one hole was left open as a control. Data were collected using the same observ ational methods used in the first part of the feeder experiment. d Testing optimal vexillum size with paper vexilla: This observational study was virtually identical to c with two important exceptions. The first was that the M. urens vexillum was replaced by a paper one of the same dimensions, so that the feeder had three paper vexilla, one small, one medium M. urens sized, and one large. Dimensions for the small and large remained the same. As in all other studies, one hole was left as a control. The second difference was the use of more resilient note card paper instead of the more flimsy notebook paper vexillum. The time of observation for this and the next study was moved to 8:00 to 11:00 pm e Testing optimal vexillum size with waterproof vexillum: This study was identical to d except for the use of contact paper around the vexillum to minimize damage by rain. All results were analyzed with a chi square test X 2 to determine significance. Results a Functionality of vexillum Results of the five days of data were shown to be significant X 2 = 52.8, p<.001. A total of 46% of the control flowers were visited while only 3% of the modified flowers were visited. When compared to the original study by von Helverson 1999, the samp le size is considerably smaller, 524 compared to 215, but the overall pattern of pollination is consistent Figure 2. b Viability at feeders Observational results with M. urens vexillum and control yielded significant results X 2 = 10.67, p< .005. A total number of 225 bats were observed. Out of total visits, 61% were to the holes with vexillum attached Figure 3. A total of two hours were spent at the feeders for three nights. Distinguishing type of bat visiting the feeder was impossible. Many bats seemed a little hesitant to feed from the feeder. Often a bat would fly to about half a meter from the feeder, and turn away. ItÂ€s hard to say, however, if this is a function of the light, the vexillum, or if this is part of their normal behavior c Artificial vs. real vexillum Results from this study once again are significant X 2 = 12.93, p<.005. Sample size was 98 bats. The numbers for the large and small artificial vexillum are nearly identical, though; in general, all holes with vexillum show a greater number of visits than the control Figure 4; small 26% of visits, Mucuna 39%, Large 21%, Control 14%. Light rains severely affected the paper vexillum, which were destroyed and remade several
times through the course of the night. d Optimal size with vexilla of paper These results turned out not to be significant X 2 = 4.59, critical value 7.85. A total of 162 bats were observed. These data, however, are fairly suspect due to problems with rain and the large vexillum. In this study in particular, the rain affected the paper nectar guide by saturating it with water and causing it to fall forward, blocking the hole. This was not always visible immediately, and since it rained constantly throug h entire sampling time, data from the large portion of the study was probably severely affected. When the large vexillum is not included into the final calculations, a significant difference is found between the medium vexillum vs. the small plus the contr ol X 2 = 8.01, p < .025, cv = 5.99. The small and control had nearly identical numbers of bat visits. The graph Figure 5 also looks very similar to the graph in figure 6, with the exception of a sizable dip in the number visiting the hole with the large vexillum. One fin al part was added to the study to combat the effects of the rain on previous feeder studies. e Optimal size with waterproof vexillum These results ended up being significant X 2 = 32.06, p < .001, cv = 7.81. The sample size was 325 bats. This graph shows a clear numerical trend towards larger vexillum, with the biggest jump being between the medium and small vexillum. There is also a clear, though smaller, jump between the large and medium vexillum Figure 6. Note that in c, d, and e, the control i s always one of the least popular visited. Discussion From the collective data presented in my report, there is little doubt in my mind that the vexillum in M. urens serves as an aural nectar guide, as was shown for M. holtonii von Helverson and von Helverson, 1999. Though the visitation figures are different between the two studies, the same pattern is clear. There are simply too many variables to accurately predict why only 48% of my control flowers were visited while those at La Selva had a nearly 90% visitation rate. Accounting for variations in plant species, time of year, number of flowers, varied topography, bat population structure, rain, a full moon, and the fact that the plant in Monteverde is more widespread than in La Selva makes concludin g anything about the fact that only half of the unmodified flowers were pollinated nearly impossible LaVal, pers. comm.. In general, however, the results of both studies were very clear. In addition, these results were further supported by the data from the later studies at the hummingbird feeders. There is an interesting extra factor in all of the hummingbird feeder studies that may have had an effect on the total number of visitors to each hole, but probably not on the final numbers. The fact that A. ge offroyi is too large to use M. urens would presumably mean that A. geoffroyi could ignore any reflectance from M. urens vexilla LaVal pers. comm.. This could have several effects. In my opinion, the most likely is that the bats would simply ignore the ve xillum and go to the holes indiscriminately. There is also the possibility, however, that they would actually avoid the vexillum and go towards the control hole. When all the data are looked at and weighed, however, it seems more likely that A. geoffroyi is probably feeding indiscriminately. If this is the case, then A. geoffroyi only impacts the data
when total number of bats are considered, not individual holes. This should not impact my data. There is a possibility, discussed in detail later in this pap er, that the larger vexillum did attract the larger A. geoffroyi to the feeders more effectively than other studies did. At the hummingbird feeders, the initial data comparing modified holes to unmodified holes yielded the expected trend. The fact that jus t the vexillum and no other part of the flower was used to attract a significant amount of bats further supports the findings of the first part of the study, as well as the La Selva study. The first test with the artificial paper vexilla showed nothing mor e than paper could be used to simulate the function of the nectar guide. Nothing more can be concluded from this study due to the fact that rain badly warped the large vexillum, rendering it useless for regular intervals. Still, the data clearly showed tha t the holes with vexilla were preferred over the control. This is consistent with all other studies. The fact that paper was shown to effectively serve as a vexillum indicates that the vexillum on M. urens serves mainly as a reflective structure and probab ly does not contain any undetected chemical or visual cues to the bat. The final two studies, probably the two most dependable data sets, show a clear trend toward favoring larger vexillum size. The first of the two suffered once again from the effects of rain on the large paper vexillum. The medium, M. urens sized vexillum performed well, however. Especially when the data from the large vexilla is thrown out, it is clear that paper is an acceptable substitute for actual M. urens vexillum, given the same re sult as the previous experiment. The final study clearly shows the trend favoring largeness in vexilla Graph E. This is a fascinating trend which begs the question why M. urens has not evolved to have larger flowers and take advantage of the increased nu mbers of visits. One possible explanation for this is that the plant is maximizing the output from the vexillum while still minimizing the investment needed in the flower. This is supported by the fact that there was virtually no difference between the con trol and small vexillum, while a significant difference was found between the medium and small. This indicates that, for a relatively small amount of biomass, visitation increases by 56%. When more investment is put into the vexillum, however, the advantag e given to the plant, while numerically greater only grants a 22% increase. This suggests that the M. urens sized flower is optimized to get the most out of the least possible material. It is interesting to note that other plants in the genus Mucuna also have vexillum approximately the same size Table 1; Stevens et al. 2001; Woodson et al. 1980. This suggests that selection everywhere has converged upon 2 3 cm as an optimum size of vexillum. Evolutionarily, the sub family Papilionoidea is widespre ad throughout the world, and is pollinated by just about every type of pollinator known. The vast majority of Papilionoideae are insect pollinated and the vexillum works as a visual flag for the insects Judd et al. 1999. It has been hypothesized that bat pollinated flowers in general originated from insect pollination and evolved from there to bat specialization Stebbins 1970. It is not too much of a stretch to imagine the visual flag being modified through evolutionary time to reflect bat echoes. The f act that a large amount of the bat pollinated species have vexillum of nearly the same size reveals a remarkable amount of convergence. Selection must strongly favor the current flower size of Mucuna in bat pollinated species. The genus Mucuna has undergon e remarkable adaptive radiation. Insects, bats, birds, and possums all visit Mucuna, but data about differences in vexilla size is not readily available Endress 1994. However, in the s pecies Mucuna rostrata, which probably is visited by hummingbirds, the vexillum has grown to nearly 5 cm long. This further suggests that a 2 3 cm vexillum is a specific adaptation to bat pollinated flowers. Other species in the genus Mucuna that are not visited
by bats probably have highly varied Mucuna vexillum sizes. A se cond explanation takes into account the larger nectar bat, Anoura geoffroyi that was found at the site where sampling was done. It was assumed that since A. geoffroyi would not normally visit M. urens, it was unlikely that the bats would respond to vexilla. It is possible, however, that the larger bat might respond to the larger vexillum. There may very well be nectar guides in larger flowers that A. geoffroyi pollinate. If so, they would act similarly to the vexillum in M. urens, but on a larger scale. Putting on an artificially large vexillum onto a feeder may reflect the echolocation of the bat in a way that closely resembles a species with a larger flower, one that A. geoffroyi could visit. Much more information is needed, though, test this hyp othesis. Research on other nectar guides in different bat pollinated plants is scarce. In order to confidently support this theory, aural nectar guides would have to be found in other, larger plants. Looking at most bat pollinated plants, a common characte ristic is that they are cone or trumpet shaped Janzen 1975. I propose that these flowers are the actual nectar guide, formed by the petals of the flower. The concave shape of the flower serves to reflect back echolocations, just as the vexillum did in Mu cuna. It is interesting to note that in areas of the old world tropics where bats do not echolocate; Mucuna species that are visited by bats do not possess raised and concave vexillum von Helverson and von Helverson 1999. It would be worthwhile to compar e other bat pollinated flowers from these areas with those in the new world tropics. In the broader context of plant pollinator evolution, my results suggest an optimization of vexillum size. With a larger nectar guide, the plant simply can not justify the extra investment required for a small increase in pollinator visits. In addition, M. urens has selected for pollination by smaller bats, which gives it a few very efficient pollinators. Making the nectar guide larger may attract other bats, who have th eir own suite of specific plants that they pollinate. This makes them much less efficient pollinators that do not necessarily visit conspecifics. Acknowledgements I would like to thank Alan, Andrew, and Richard for providing so much assistance with this project. I would also like to thank Jay and Rhett for keeping me company up at the gallery. I would also like to thank the owners of the hummingbird gallery for allowing this project. No bats were harmed in the making of this project. _______________________________________________________________________________________________ Literature Cited Altringham J.D. 1996. Bats: biology and behavior, pp. 222 229. Oxford University Press. Oxford, England. Bawa K.S. 1990. Plant pollinator in teractions in tropical rain forests. Ann. Rev. Ecol. Syst. 21:399 422 Brehm B. G. and Krell D. 1975. Flavonoid Localization in Epidermal Papillae of Flower Petals: A Specialized Adaptation for Ultraviolet Absorption. Science 190 4220: 1221 1223. Diller, G.W. 2000. Mucuna urens, a tropical liana in Monteverde. Nadkarni N.M. and
N.T. Wheelwright Eds., pp. 72 Oxford University Press, England. Endress P.K. 1994. Diversity and evolutionary biology of tropical flowers, pp. 141 142,297 298. Cambridge Universi ty Press. Cambridge, England. Feinsinger P. 1983. Coevolution and pollination in Coevolution. Ed. Futuyma, DJ and Slatkin M. pp. 284 287. Sinauer Associates Inc. Mass. Janzen D.H. 1975. Ecology of Plants in the Tropics. Institute of Biology's Studies in Biology, no. 58. London, England, pp 18 19 JuddW.S., C.S.Campbell, E.A. Kellogg, P.F Stevens. 1999. Plant systematics: a phylogenetic approach, pp. 282 288. Sinauer Associates Inc. Mass. Majewski J.J. 2002. Factors Affecting Arthropod Loads on Nectivorous Phyllostomatidae. CIEE: Tropical Ecology and Conservation. This Edition. Miller R.M. 2002. Pollen Loads of Cloud Forest Nectivorous Bats. CIEE: Tropical Ecology and Conservation. This Edition. McDade L.A. and G.S. Hartshorn. 1994. La Selva Biological Stat ion in La Selva. Ed. McDade L.A., Bawa K.S., Hespenheide H.A., Hartshorn G.S. pp. 6. University of Chicago Press, Chicago. Penny J.H.J. 1983. Nectar guide colou r contrast: A possible relationship with pollination strategy. New Phytologist 954: 707 721. Stebbins G.L., 1970. Adaptive radiation of reproductive characteristics in angiosperms, I: pollination mechanisms. Annual Review of Ecology and Systematics. 1: 307 326. Stevens, WD., C. Ulloa, A. Pool, and O.M. Montiel Eds.. 2001. Flora de Nicaragua. Tom o I. Missouri Botanical Garden Press, St. Lous, MO. Van der Pijl L. Ecological Aspects of Flower Evolution. I. phyletic evolution. Evolution 14 4: 403 416. Von Helverson D. and O. von Helverson 1999. Acoustic guide in bat pollinated flower. Nature 398: p p 759 760. Woodson Jr., R.E. R.W. Scery and collaborators. 1980. Flora of Panama. Annals of the Missouri Botanical Garden, St. Louis. Vol.67, #3