Role of driptips on Begonia involucrata (Begoniaceae) Ethan Householder Texas Christian University, Department of Environmental Science Abstract This study examined the dual role of multiple driptips of Begonia involucrata (Begoniaceae) in shedding excess water and minimizing erosion beneath a plant. One hundred and twenty leaves, having between one and four driptips were tested both with and without driptips to determine: the amount of water each leaf retained when a given amount was applied to its surface, the volume of soil displaced by each leaf, and the distance the soil was thrown by the drops. When driptips are removed, the amount of water retained on the leaf increases and the mean volume of soil displaced per driptip increases. When driptips are intact, two and three driptipped leaves experience a 2.5 fold decrease in amount of water retained on the leaf surface. Multiple driptips were shown in increase the amount of soil displaced, with three driptipped leaves displa cing the greatest total volume of soil. The relative abundance of each leaf morph can explained by the interplay of three selective forces: epiphyll growth, erosion underneath the plant, and leaf surface area (a larger surface area increases sunfleck capture). Two driptipped leaves seem to strike a balance between these forces, and may explain their greater abundance (68%) at an elevation of 1,545 meters. Resumen Este estudio examin el papel de mltiples puntos de goteo de Begonia involucrata (Begoni aceae) al dronar agua y al controlar erosin bajo de la planta. Ciento veinte hojas, entre uno y cuatro puntos de goteo, fueron examinados con y sin puntos de goteo para determinar el volumen de agua cada hoja retuvo despus de la aplicacin de cierto cant idad de agua y el volumen de tierra desplazada por cada hoja y punto de goteo. Cuando los puntos de goteo estn eliminados, la cantidad de agua retenida en la hoja aumenta y el volumen promedio de tierra desplazada por punto de goteo aumenta. Cuando los pu ntos de goteo estn intactos, hojas con dos y tres puntos experiencian un disminucin en agua retenida por un factor de 2.5. Multip les puntos de goteo desplazan m s tierra, llegando a un mximo con hojas de tres p untos de goteo desplazando lo m s volumen d e tierra. La abundancia relativa de cada hoja fue explicado por la interaccin de tres fuerzas selectivas: crecimiento de epiphylls, erosin bajo la planta y la superficie de la hoja (un superficie ms grande aumenta la captura de rayos de sol). Hojas con dos puntos de goteo mostraron ser lo ms efica z en combatir este s problemas y por eso, representan 68% de todas hojas que se encuentran a una elevacin de 1545msmm. Introduction Driptips are a characteristic trait of understory leaves in tropical rainforests, and yet little is known about their ada ptive significance (Richards 1996 ). Driptips have long been considered an adaptation designed to rapidly funnel water from the leaf sur face to prevent epiphyll growth, which can be a significant problem in the humid, shady environment of the forest understory (Lightbody 1985). This hypothesis is based on the assumption that epiphylls can more easily colonize a leaf that retains a thick fi lm of water
on the surface. Epiphylls are thought to be harmful to a leaf in that they shade the host leaf, and can cause wilting (Ivey and DeSilva 2000). Ivey and DeSilva (2000) demonstrated that removing the driptips increases colonization of epiphylls; however, it has also been shown that no relationship exists between epiphyll growth and driptips (Richards 1996 ). Furthermore, it is not clear whether epiphylls are harmful to lea ves, because certain nitrogen fixing epiphylls provide a limiting nutrient (nitrogen) that can be beneficial to the plant once it is taken up by the leaf (Richards 1996). Whether driptips serve to prevent epiphyll growth remains to be definitively shown. A second, less established hypothesis states that driptips function to prevent erosion. Water drops dislodge and loosen the soil beneath the plant, which is then more likely to be washed away during heavy rains. A drop collecting at the end of a driptip wil l be smaller and lighter, minimizing soil disturbance, and therefore reducing soil erosion beneath the plant (Williamson 1981). Tropical forest flora maintain luxuriant growth on thin and nutrient poor soils due to efficient recycling of nutrients, so any trait (such as driptips) that protects this valued resource is important (Terborgh 1992). Temperate rainforests with more nutrient rich soils are not characterized by driptips because nutrient los s is not a concern (Richards 1996 ). As indirect evidence of driptips functioning to reduce erosion, Williamson (1983) found that driptips on understory plants increase in length with increased distance from the ground. A drop originating from a higher source will have greater kinetic energy and disturb the soil mor e. This provides a selective force for longer, more developed driptips in higher leaves. Despite these relationships, a driptips' role in minimizing erosion remains open for research. Begonia involucrata (Begoniaceae) is a common, herbaceous, understory pl ant with between one and four driptips. It is common along trails and clearings in Costa Rica and Panama at elevations between 1,400 and 2,000m in elevation. These plants range from a few centimeters tall to two meters, and form large clusters of several i ndividual stems (Agren and Schemske, 1991). It has a shallow root system and often grows in clumps (pers. obs.). Begonia involucrata has simple alternate leaves with palmate venation that are asymmetrical with one, two, three, or four driptips. Although th e function of multiple driptips has not been studied in great detail, many ideas have been formulated. The increase in the number of driptips with increasing elevation has been considered indirect evidence for the role of B. involucrata driptips in minimiz ing epiphyll growth (Taylor 1997). This increase in driptip number per leaf is generally attributed to the increase in precipitation with elevation. A previous study on B. involucrata found that more driptips reduce the efficiency of water shedding, possibly pointing to another adaptive significance such as erosion control (Haloin, 2000). Multiple driptips may serve to minimize the disturbance of the soil beneath the plant. The relation ship of the number of driptips per leaf increases with leaf height from the ground can be explained if erosion does impose a selective pressure to have more driptips (Massa 1993). This study examined the role of multiple driptips in preventing epiphyll gro wth and controlling erosion as dual benefits and the possible effect they have in shaping the relative abundance of each leaf morph.
Materials and Methods Leaves were taken f rom the forest above the Estacin Biol gica Monteverde located on the pacific slo pe of the Tilarn mountain range 1540m, representing a Holdridge life zone of tropical montane wet forest Holdridge 196 7. Mean annual temperature is 18.5C and it receives an average of 3.191m of annual rainfall with .886m of extra mist during the wind y misty season Clark et al. 2000. Water Removal Thirty leaves of each leaf morph one, two, three, and four driptipped leaves were examined for a total of one hundred and twenty leaves. An entire B. involucrata plant was collected, stored in a vase containing water, and used within two hours to prevent wilting. Each leaf was dipped in water to remove debris and allowed to drip excess water for thirty seconds. The leaf was the n supported by a platform at an angle of 55 below horizontal to mimic the average leaf angle in the forest. A syringe ten ml maximum was filled with water to the five millimeter mark and this amount was weighed with a Fisher Scientific XT top loading balance 0.001 gr. Water was applied along the upper perimeter of the leaf at a constant rate and collected in a container. The leaf was allowed 20 seconds to drip before the collecting container was removed and the water weighed. The amount of water after application remaining on the leaf wa s calculated by subtracting the weight of water collected from the weight applied. Erosion Control Using the same leaf, the volume of soil displaced by each driptip was measured. The leaf was positioned at 55on a platform 20cm above a bowl of finely grained, leveled sand. Seventy squirts of water from a squirt bottle held 30cm away were applied to the leaf at a constant rate. Twenty seconds were allowed for the leaf to drip. The d epth d and width 2r of the depression made was measured with calipers 0.001cm, and a volume was calculated by assumi ng a cone shape calculated as 1 /3x3.14*r 2 xd. Next, the driptips were removed the edge was cut smoothly and was rounded, and the above two processes were repeated. To measure surface area, the leaves were photocopi ed, cut out, and weighed 0.001 gr. Area was calculated by making a ratio with the mass of a known area of the same paper SA = 25cm 2 / .205gxweight of copied leaf Field experiment Forty clumps of B. involucrata were located, and two pairs of two driptipped leaves on the same stem that were free of herbivory, epiphyll growth, and of approximately equal size, were tagged. The driptips of one leaf of each pair were removed, and the other leaf was unmanipulated to serve as a control. These 160 leaves were monitored and misted everyday unless it rained sufficiently to reach the forest floor for seventeen days April 13 to April 30, 2002. After the 17 days the pairs were collected and the percent epiphyll coverage was
determined. To determine the relative abundance of the four types of leaves, every leaf within a clump of leaves at 50m intervals was counted and a tally made for each leaf morph. One thousand and thirty seven leaves were counted along about one kilo meter of trail, starting Estaci n Biolgica Monteverde Sendero Principal Data Analysis An ANOVA was used to determine whether driptips, the number of driptips, or leaf area had significant effects on water removal and volume of the depression made by falling drops. Fisher's PLSD post hoc test was used to determine the difference in volume of soil displaced, water retain on the leaf, and surface area for each leaf morph with and without driptips. Difference in abundance of leaf morphs was determined with a Chi square test. Results Area and Abundance of Leaf Types Surface area of the four different leaf morphs were significantly different Fig. 1, Two Way ANOVA, F = 32.187, p < 0.0001. And Fisher's PLSD post hoc test showed leaves with one and two p < 0.0001, one and three p < 0.0001, one and four p < 0.0001, and two and three p < 0.0001 driptips had significantly different areas. The area of two and four driptipped leaves, and three and four driptipped leaves is not significantly different p = 0.0720 and p = 0.1594 respectively. Two driptipped leaves were the most abundant, making up 68% of all leaves from the field Fig. 2. Three driptipped leaves are the second most common, making up 24% of all leaves. One and four driptipped leaves are least common, each making up 4% of all leaves. Chi square test revealed a significant difference x 2 = 1124.77. Water Removal Water retained on the leaves was significantly different when the driptips were removed Fig. 3, Two Way ANOVA, F = 9.908, p < 0.0019. There was no significant difference in the amount of water retained between leaves with intact driptips One Way ANOVA, F = 2.448, p = 0.0673, however, Fisher's PLSD post hoc test revealed that leav es with one and two, one and three, two and four, and three and four driptips retained significantly different amounts of water p = 0.0001, p = 0.0065, p = 0.0014, and p = 0.0453 respectively. Erosion Control The mean volume of soil displaced per driptip is significantly different when the driptips are present Fig. 4, Two Way ANOVA, F = 239.742, p < 0.0001. The total volume of soil displaced per leaf was also significantly different Fig. 5, Two Way ANOVA, F = 403.877, p < 0.0001. Fisher's PLSD post ho c test revealed that all total volumes of soil
displaced were significantly different all having p < 0.0001 except for leaves with three and four driptips p = .6667. Field Experiment There was no measurable change in the percent coverage by epiphylls between the cut leaves and the controls after the 17 day period. Discussion Water Removal The results show that driptips do play an important role in water removal. Removal of the driptips increased the amount of water retained on the leaf in every case Fig. 3. This supports the findings of Lightbody 1985, who also hypothesized that the more water remaining of the leaf, the more ideal the leaf surface is for epiph yll growth. The increased epiphyll load could potentially reduce the health or photosynthetic capabilities of the plant Ivey and DeSilva 2000. This result illustrates that driptips do help funnel water off, but does not indicate any effects of multiple d riptips. Figure three also shows that in the cut leaves, water trapped on the leaf increases with driptip number. However, this is not the case when driptips are intact. In cut leaves, the increase in water retention is probably an effect of their greater surface area Fig. 1. Both two and three driptipped leaves experience a 2.5 fold decrease from cut leaves in the amount of water retained on the leaf when driptips are left intact, while one and four driptipped leaves only experience a 0.7 and 1.7 decre ase respectively. With driptips intact, two and three driptipped leaves are almost twice as efficient in water removal as one and four driptipped leaves. These disproportionate changes between leaves with and without driptips indicates a functional purpose of driptips in water shedding. Also, the relative abundance in the forest of each leaf morph reflects its efficiency in removing water. Two driptipped leaves are the most common 68%, and three driptipped leaves represent 24% of the leaves Fig. 2. Thes e leaf morphs are also the most efficient at water removal. One and four driptip morphs are the least efficient at removing water, and are the least common, making up eight percent of all leaves. Two and three driptipped leaves may have a selective advanta ge over their one and four driptipped counterparts and are therefore more common in the population. The results of the field experiment, however, did not support, nor refute the water shedding hypothesis proposed by Ivey and DeSilva 2000. No measurable difference in epiphyll growth could be detected between leaves with and without driptips. There are two possible reasons why no epiphylls were found. First, seventeen days may not have been sufficient to permit the colonization by ep iphylls. Second, the experiment was performed at the height of the dry season March to April. Because epiphylls are very susceptible to desiccation for lack of a root system and their exposed locations, reproduction and dispersal may have been severely l imited.
Erosion Control The presence of driptips affects the amount of soil displaced by the leaf. Removal of the driptips increases the mean volume of soil displaced 3 fold in some cases fig. 4. This indicates that driptips may affect the soil beneath t he plant dramatically. Falling drops serve to dislodge and loosen the substrate beneath the plant, causing the soil to be more susceptible to erosion as water flows down slope and transports soil particles away from the plant Williamson 1983. Furthermore erosion in areas of high rainfall, extreme slopes, and gap like conditions or clearings where the soil is more exposed is much greater Williamson 1981. Because B. involucrata has such a shallow root system, severe erosion directly underneath the plant would eventually expose the roots to desiccation and reduce stability. After uprooting over 100 plants in the field, it was apparent that roots were rarely deeper than four centimeters. Also, two of the 80 monitored plants with exposed roots had fallen ove r; suggesting that erosion of the substrate may affect plant stability. Furthermore, severe erosion underneath the plant could wash away important nutrients in the soil and reduce the fitness of the plant. Erosion is major problem in the preferred habitat of B. involucrata. This paper supports Williamson's hypothesis that driptips significantly decrease the amount of soil displaced and may be advantageous to B. involucrata. In regards to multiple driptips' role in erosion control, figure five shows that fo r one, two and three driptipped leaves, with increasing driptip number, more soil is displaced. The decrease in four driptipped leaves is due to one of the outside driptips being non functional, making it more similar to two driptipped leaves, rather than pointing to a true function of multiple drips in reducing soil displacement this suggests that multiple driptips do not minimize erosion. If the results of both sections water removal, erosion control are analyzed while keeping in mind the natural histor y, relative abundance of each morph, and size of each leaf morph, interesting conclusions can be made. There are three different selective pressures on B. involucrata which all interact to produce the observed abundance of each leaf morph. The first select ive pressure is to minimize epiphyll growth. As stated before, two and three driptipped leaves are the most efficient at shedding water off the leaf, reducing the danger of epiphyll growth and giving them a selective advantage over the other morphs. This m ay explain why 92% of all B. involucrata leaves in the forest are two and three driptipped leaves Fig. 2. Size of the leaf is also important because light is often a limiting factor in the understory. Since understory leaves receive most of their light f or photosynthesis in the form of sunflecks, it would be advantageous to have a larger surface area to capture these ephemeral beams of sunlight Rundel and Gibson, 1996. The tendency then, is to be larger, which partly explains the lack of small one driptipped leaves, and that 96% of all leaves are the larger two, three, and four driptipped leaves Fig. 3 and Fig. 2. However, this selective pressure to have a large surface area is balanced by the ne ed to rapidly funnel off water. This explains the paucity of four driptipped leaves 4% of total leaves which, although are very large for capturing sunflecks, do not shed water efficiently. The third selective pressure is erosion. As demonstrated in figu re five, multiple driptips are not important in minimizing erosion, but they may play an important role in
the shaping the relative abundance of each leaf morph. Although one driptipped leaves displace the least total volume (fig.5), they are small and ine fficient at water removal so that benefits of reduced erosion are outweighed by the costs of epiphyll growth and reduced photosynthesis. Three driptipped leaves on the other hand displace the greatest volume of soil, while two driptipped leaves displace in termediate amounts (fig.5). The fact these two morphs are equally efficient at water removal but displace unequal amounts of soil, may explain why two driptipped leaves are more common (68%), while leaves with three driptips only represent 24%. The interac tion of all these selective forces has shaped the relative abundance of B. involucrata leaf morphs. I have shown experimentally that leaves with two driptips are the most efficient at shedding water, displace relatively small amounts of soil, and are the m ost abundant leaf type in the forest. This would lead me to speculate that they strike a balance between the three selective forces better than the other three leaf types and are therefore advantageous to the fitness of B. involucrata. Acknowledgements I would like to thank Mauricio Garcia for his help, optimism, and encouragement especially during the early stages of the project and in the final revising stages. I would also like to thank both TA's, Andrew Rodstrom and Will Wieder, for their help, and t he Estacin Biol gica Monteverde. Finally, I thank all those who kept me company in that lonely, lower laboratory. Literature Cited Agren, J and D.W. Schemske. 2000. Deceit pollination in Begonia. Biotropica 23: 235-241. Clark, K.L., R.O. Lawton and P.R. B utler. 2000. The physical environment. In: Monteverde: ecology and conservation of a tropical cloud forest. Nadkarni, N.M and N.T. Wheelwright, eds. Oxford University Press, New York, New York. pp. 15 33. Haloin J. 2000. The Effect of Drip tip Number on th e Efficiency of Water Shedding in Leaves of Begonia involucrata. CIEE summer 2000: 119 123. Holdridge, L. R. 1967. Life zone ecology. Tropical Science Center, San Jos Costa Rica. Ivey C. and DeSilva N. 2000. A Test of the Function of Drip Tips. Biotropica 33: 188 191. Lightbody, J.P. 1985. Distribution of leaf shapes of Piper sp. in a tropical cloud forest: evidence for therole of drip-tips. Biotropica 13: 339-342. Massa, G.W. 1993. Multiple drip tips of Begonia involucrata: a twist on leaf adaptations to high rainfall environments. UCEAP Monteverde. Fall 1993: 3 9. Richards, P.W. 1 996 The Tropical Rainforest, pp. 96 98. Cambridge University Press, Cambrigde, England. Rundel, P. W., and Gibson, A.C. 1996. Adaptive Strategies of Growth Forms and Physiological Ecology in Neotropical Lowland Rain Forest Plants. In: Neotropical Biodiversity and Conservation. Gibson, A.C., ed. University of California, Los Angeles, Los Angeles, California, pp. 33 71.
Taylor J. 1997. Environmental influences on leaf morphology: Adaptation of Begonia involucrata to a cloud forest environment CIEE Su mmer 1997 :41 50. Terborgh J. 1992. Diversity and the Tropical Rainforest, pp. 31 52. Scientific American Library, New York. Williamson, G. B. 1981. Driptips and splash erosion. Biotropica 13: 228 231. 1983. Driptips, Drop Size and Leaf Drying. Biotropica 15: 232 234.
Number of driptips Figure 1. Average surface area of the four leaf morphs were significantly different (One way ANOVA, F = 32.187, p < 0.0001) Fishers PLSD showed that leaf one was different from the rest (p < 0.0001) and leaf two was different from leaf three (p = 0.0019). F igure 2. Abundance of each leaf morph in 1,037 leaves of B. involucrata at Estacin Biolgica Monteverde, 1540msmm.
Figure 3. Mean water shedding in the different leaf morphologies when the driptips were cut and uncut was significantly different (ANOVA F = 9.908, p < 0.0019). One way ANOVA for uncut leaves was not significant (F = 2.448, p = 0.0673) but Fishers PLSD post hoc sh owed significant differences for one and two, one and three, two and four, and three and four (p = 0.0001, p = 0.0065, p = 0.0014, p = 0.0453). Figure 4. Mean volume of soil displaced per driptip for each leaf morph when the driptips are cut and uncut were significantly different. (N = 120). (Two way ANOVA, F = 239.742, p < 0.0001) Estacin Biolgica Monteverde, Costa Rica.
Figure 5. Mean total volumen of soil displaced per leaf for each leaf morph when the driptips are cut and uncut are significantly different. (N = 120) (Two way ANOVA, F = 403.877, p < 0.0001) Fishers PLSD show that all leaf types are significantly different (p < 0.0001) except for th ree and four driptipped leaves (P = 0.6667).