1 Pollinator limitation and seed set of Impatiens walleriana Alison Parker Department of Biological Sciences, The George Washington University ABSTRACT Plants that are pollinator limited may show important fitness relationships with factors influencing visitation. Groups of 60 bagged, natural, and hand pollinated flowers of Impatiens walleriana (Balsaminaceae) showed much higher seed set with pollination flowers produced seed pods 21%, 70%, and 100% of the time, respectively. I also found higher seed s et for high visitation sites in the Monteverde Butterfly Garden with naturally visited patches outside mean seed set inside was 146.5 per 60 stems, while outside it was only 81. Neighborhood, patch size, sunlight, and elevation are speculated to play a r ole in pollinator visitation. I compared seed set of these variables among patches throughout Monteverde, Costa Rica. With a mean of 81 seed pods per 60 stems, patches located in the vicinity of other flowering plants have a significantly higher seed set t han isolated patches, with a mean of 55.6 seed pods per 60 stems. There is a slight tendency for seed set to increase with patch size, but this relationship is not significant. Flowers located in higher sunlight produced significantly more seed pods than t hose in the shade with means of 5.6 and 2.4 pods per 5 stems, respectively. Finally, seed set increases along an altitudinal gradient. It appears that I. walleriana is pollinator limited and that nearby flowers of other species, higher sunlight and higher elevation increase the likelihood of seed set through greater visitation. RESUMEN Las plantas que estÃ¡n limitadas por sus polinizadores pueden mostrar importantes relaciones adaptativas con factores que influencian las visitas. Los grupos de 60 flores de Impatiens walleriana (Balsaminaceae) cubiertas por bolsas, silvestres y cultivadas, mostraron una producciÃ³n de semillas mucho mayor con polinizaciÃ³n las flores produjeron vainas con semillas el 21%, 70% y 100% del tiempo, respectivamente. TambiÃ©n s e encontrÃ³ una producciÃ³n mayor de semillas en los sitios con mÃ¡s visitas en el JardÃn de Mariposas de Monteverde con grupos de flores silvestres en el exterior la producciÃ³n promedio de semillas adentro fue de 146.5 en 60 tallos, mientras que afuera sÃ³l o se produjeron 81. Se especulÃ³ que los alrededores, el tamaÃ±o del grupo, la luz solar y la elevaciÃ³n podrÃan afectar las visitas de los polinizadores. Las producciones de semillas de estas variables se compararon entre grupos en Monteverde, Costa Rica. Los grupos en los alrededores de otras plantas con flores tuvieron una producciÃ³n de semillas significativamente mayor, con un promedio de 81 vainas en 60 tallos, que las flores en grupos aislados con un promedio de 55.6 vainas en 60 tallos. La producciÃ³ n de semillas presentÃ³ una tendencia leve a aumentar con el tamaÃ±o del grupo, pero esta relaciÃ³n no fue significativa. Las flores en condiciones de luz solar altas produjeron significativamente mÃ¡s vainas con semillas, con un promedio de 5.6 por cada 5 ta llos, que las flores en la sombra (con un promedio de 2.4 por cada 5 tallos). La producciÃ³n de semillas aumenta con la elevaciÃ³n. Se sugiere que I. walleriana estÃ¡ limitada por sus polinizadores y que las flores cercanas de otras especies, mÃ¡s luz solar y elevaciones mÃ¡s altas aumentan las posibilidades de producciÃ³n de semillas debido al aumento en las visitas de los polinizadores.
2 INTRODUCTION The herbaceous I. walleriana (Balsaminaceae) is an exotic weed prevalent on roadsides, forest edges, and ot her disturbed habitat throughout Monteverde, Costa Rica. The species is wind dispersed through elastic dehiscence, which results in aggregations of individuals in patches of various sizes. Visitors and likely pollinators include a variety of Lepidoptera an d Hymenoptera (Baker 1993). Sexual reproduction of flowering plants usually necessitates pollination. However, in I. walleriana , the importance of pollinators is unclear. Different species experience distinct levels of pollinator limitation, but often, t he influence of pollination is ignored in resource allocation studies (Bierzychudek 1981). Through the study of I. walleriana and its pollinator systems, I wanted to determine whether reproductive capacity, as measured in seed set, is influenced by pollina tor limitation. Pollinator behavior in natural systems is affected directly and indirectly by many abiotic factors, reflecting the abundance and preferences of these mutualistic partners. Specific characteristics of I. walleriana pollinators are unknown, but their actions can be predicted through both known and inferred characteristics. One important factor is the surrounding plant species, which can increase insect visitation to the general area through their own attractive mechanisms. Once there, the in sects are likely to exploit all available resources, therefore increasing visitation to neighboring flowers (Feinsinger 1986). The insect pollinated Alstroemeria aurea responds to changes in neighborhood due to herbivory, because the extent of neighborhood flowers have a significant impact on the amount of visitation. Like I.walleriana , A. aurea is self compatible, protandrous, and has a high dependence on pollinators (Vasquez and Simberloff 2004). Also, when isolated patches occur, individuals can experie nce increased visitation as a result of patch size. This is due to the attractive capacity of a larger quantity of flowers. Kunin (1997) found that individual bee pollinators of Brassica kaber visited more flowers in larger patches. Many studies found that reproductive success is compromised in sparse populations. Population size and germination rates are directly correlated in the prarie plant Silene regia (Menges 1991). Results of a study on the herb Lythrum salicaria indicate that there is a positive co rrelation between population size and seed production due to inadequate pollination in small aggregations (Agren 1996). Environmental variables such as sunlight and elevation can have a significant affect on insect visitation. Butterflies and small insec ts select foraging times or sites with maintain a suitable body temperature. Therefore, when variations in sunlight occur within a plant population, it can result in in dividual variation in visitation and reproductive fitness, as in the shrub Lavandula latifolia (Herrera 1995). Further, in Monteverde, Costa Rica, nectar robbery of I. walleriana decreased with increased elevation (Morris 2005). Due to this observation, pollinator visitation rates could increase on an altitudinal gradient. Can I. walleriana self pollinate? Is the amount of pollinator activity a limiting factor its reproductive ability? I predicted that insect pollination was necessary for pod production, and that in the natural environment, pollination does not occur for every flower, resulting in pollinator limitation in reproduction. Environmental conditions that influence I. walleriana pollinator visitation should then, in turn, also affect strength of
3 seed set. Because of the attraction of neighboring flowers, I predicted higher seed set in I. walleriana patches located in gardens. Similarly, I expected seed set to increase with ors as well as possible elevational protection from nectar robbery, I thought, would increase the number of pods. METHODS In order to determine the effect of pollinator visitation on seed set in I. walleriana , I selectively bagged, marked, and hand poll inated I. walleriana stems in a patch in front of the Estacion Biologica Monteverde. To show the extent of self pollination in this species, I chose sixty stems with unopened flower buds and removed all seed pods and adult flowers from the stem. I covered the stems with mesh bags made of mosquito netting. A set of control stems, also with unopened buds and removed pods and flowers, were marked to measure the effect of natural pollination. Finally, I marked and hand pollinated sixty I. walleriana adult flowe rs, using a toothpick to transfer pollen from male to female flowers. In addition, I compared I. walleriana seed set for the Monteverde Butterfly Garden where butterflies were forced to visit available flowers, to the seed set of I. walleriana outside. To do this, I counted the number of seed pods on 60 stems of I. walleriana on all the patches within the garden, as well as similar patches outside. In order to measure seed set under a variety of conditions, I compared the number of seed pods in 52 patches of I. walleriana throughout Monteverde, Cerro Plano, and San Luis. At each location I recorded the potential environmental conditions important to visitation. Patch size was measured with a simple count of I. walleriana stems. I was unable to count the st ems in larger patches, so I extrapolated the number through area measurements and the density of stems/m 2 . Elevation, relative amount of sunlight and shade, and the presence of neighboring plants were also recorded for each patch. To quantify seed set, I c ounted the number of seed pods on 60 stems for each condition. I was then able to compare a single variable while controlling for the other three. For example, patches along an elevational gradient were medium sized, covered by both sun and shade, and isol ated from other flowers. One exception from the patch method was the sunlight comparison in order to measure the effect of sunlight within a patch, I compared seed set of 60 individuals in a sunny area to 60 individuals in the shade of the same patch. RESULTS In the manipulated pollination experiments, bagged stems performed self pollination rarely just 20% produced seed pods. The naturally pollinated controls were pollinator limited, but 70% still set seed. All hand pollinated stems produced pods, suggesting that the natural controls were pollinator not resource limited.
4 FIGURE 1. Relative effects of self, natural, and hand pollination on seed set. Bagged stems self pollinated, while control stems indicated the extent of natural pollin ation. All are significantly different (Chi squared test, X 2 = 63.32, df = 2). Comparisons of seed set contained within the butterfly gardens had significantly higher seed set when compared to patches of similar neighborhood in a more natural setting (Unp aired t test, p =.0019). Inside the four butterfly gardens, there was an average of 146.5 seed pods on 60 stems ( N = 4) . In gardens with similar plant types outside, the mean was only 81 ( N = 13). Higher seed set was due to both the proportion of stems co ntaining seed pods and the number of seed pods per stem. Within the contained gardens, there was an obviously abnormal concentration of butterflies, and I was able to observe many flower visitations in a small amount of time.
5 FIGURE 2. The effect of forced visitation on seed set. Seed set is measured in the number of pods on 60 stems. Mean seed set in gardens outside of the butterfly garden ( N = 13) is significantly lower than mean seed set within the butterfly garden ( N = 4) ( Unpaired t test, p = .0019). I. walleriana patches located in the vicinity of other cultivated flowers produced significantly more seed pods per stem than comparable isolated patches (Unpaired t test, p=.0096). Neighborhood patches were located in the ga rdens of hotels and homes, near other cultivated flowering plants. All patches were medium sized, had a similar elevation, and a combination of sun and shade. Mean seed set of garden I. walleriana was 81 seed pods per 60 flowers ( N = 13), while the isolate d patch mean was only 55.6 ( N = 13). Garden stems usually had at least one seed pod, and often had more than one. Both the percentage of stems with pods and the number of pods per stem increased in garden settings. Patch size appeared to show a positive r elationship with seed set, but statistical analysis failed to show significance (Simple regression, p =.1013). However, it appeared that the size of the patch was an important variable in seed set. When I was able to sample very large patches, it seemed th at more often than not the stems contained more seed pods as a whole. The 13 smallest patches, containing from 60 to 200 I. walleriana stems, contained seed set ranging from 16 to 91 pods per 60 stems. In comparison, the 3 largest patches, containing stem numbers in the ten thousands, contain seed set ranging from 52 135 pods per 60 stems. Although these numbers seem to support a positive relationship, they are not verified statistically. Stems located in areas of higher sunlight showed higher seed pod p roduction (Unpaired t test, p =.0096). Although I only measured this trend within one patch, sets of five stems in the sun contained an average of 5.6 pods, compared to 2.4 in the shade. Throughout seed set sampling of all patches, I often noticed higher n umbers of seed pods on stems in the sun. Therefore, this trend is supported by statistics in one patch and observation in many. I also noticed more butterfly visitation in areas of high sunlight. Comparisons of seed set on an elevational gradient likewis e indicated a significant positive relationship (Simple regression, p =.0024), indicating that I. walleriana stems produce more seed pods at higher elevations. Although middle
6 elevations contains a fairly high variation in seed set values, low elevations a nd high elevations both demonstrated extremes in seed set production. FIGURE 3. Difference in seed set between patches in gardens and outside. Seed set is measured as the number of pods on 60 stems. Mean seed set values ( N =13) of patches neighboring other plants are significantly higher than isolated patches ( N =13) (Unpaired t test, p =.0096). FIGURE 4 . Effect of patch size on seed set. Seed set is measured as the number of pods on 60 stems. Pa tch size is measured as the number of stems in the patch. There is a tendency for larger patches to have a higher seed set, but the relationship is not significant (Unpaired t test, p = .1013).
7 FIGURE 5. Effect of sunlight on s eed set. Seed set is measured in seed pods on 5 individuals. Stems in sunnier areas have a significantly higher mean ( N =12) than individuals in areas of decreased sunlight ( N =12)(Unpaired t test, p =.0083). FIGURE 6 . The ef fect of elevation on seed set. Seed set measured in number of seed pods on 60 stems. There is a significant relationship between seed set and elevation (Simple regression, p =.0024).
8 DISCUSSION My first set of experiments verified that I. walleriana is a ffected by both self pollination and pollinator limitation. Since 11 of 52 bagged plants produced seed pods without pollinator visitation, it follows that in the absence of pollinators, I. walleriana plants are able to self fertilize. Therefore, I. walleri ana reproduction is not completely dependent on insect pollinators, allowing for insurance in areas of low visitation. However, much higher seed pod production occurred in the 50 I. walleriana stems available for natural pollination. Hand pollination alway s resulted in a seed pod, meaning that pollination must have failed to occur for the 15 control stems that did not set seed if pollinators had been more active within this patch, the remaining 15 could have set seed as well. These results indicate that p ollinator limitation does, in fact, directly affect seed set of I. walleriana . This conclusion is further supported by high seed set production inside of the butterfly garden. Since all other conditions are equal, this significant difference must be a resu lt of high butterfly concentration. I. walleriana is pollinator limited, so aspects of the environment that affect pollinator activity should also affect seed set. I.walleriana patches located within a garden had a significantly higher number of seed pods than those isolated from other plants. Since pollinator activity has a direct result on seed set, it is likely that this difference is the result of higher visitation. Studies of neighborhood effects show that some plant species are affected (Vasquez and Simberloff 2004), while others are not. (Feinsinger 1986). From the significant increase in seed set, I conclude that it is likely that neighborhood is important in I. walleriana reproduction. This increases the probability that cultivated I. walleriana in troduced into gardens will spread into surrounding disturbed habitat, due to higher reproductive capacity. However, it is likely that garden patches of I. walleriana are more contained and controlled than isolated patches. Seed set did increase with patch size, and although this relationship was not significant, the trend could be reflected in visitation, as well. A study on patch size of I. walleriana found that although the positive relationship between area and insect visitation was not quite significan t, there was a positive trend (Baker 1993). These results correspond with the trend that I found in seed set production. Therefore, although the relationship was not significant, it is still likely that the seed set is a result of visitation. Factors simil ar to patch size, such as density, are also likely to affect the amount of visitation (Kunin 1997). Therefore, it is possible that density comparisons would yield a significant relationship where patch size did not. Higher visitation in larger or denser pa tches would explain the enormous size of some patches that I observed in areas of disturbed habitat. Once a patch reaches a certain size, it seems, reproductive advantages could allow it to grow exponentially and reproduce until it covers the forest floor. There was a significant difference in seed set between sunny and shady areas in the same patch. It is likely that the higher seed set in sunny areas is due to preferences of butterflies and small insects because of body temperature requirements, butter flies and small insects choose to forage in high sunlight areas. Shady areas often fail to provide the energy these insects need to function (Herrera 1995). High seed set could also be a result of the tendency of sunnier areas to be located on roadsides or other easily accessible places, where pollinators are more likely to forage. Higher reproductive success in sunlight could explain I. walleriana
9 sunlight is available. Likewise, it helps to promote repr oduction and growth in these areas. Finally, seed set increased on an elevational gradient, lending support to the idea that visitation increases with altitude, as well. Because nectar robbery exists in low elevations in the Monteverde area (Morris 2005), it is possible that visitation increases at higher elevations in order to take advantage of available nectar. Species vary along elevational gradients, so the elevation seed set relationship could be a result of differences in visitation preferences of in dividual species (Bronstein et al. 2000). In addition, water availability increases altitudinally (Bronstein et al. 2000). Therefore, general health of I. walleriana individuals may increase with elevation, which in turn may increase attractive capacity an d visitation. Influences in elevation should result in more patches at higher elevations. This could explain the abundance of I. walleriana in Monteverde as a whole, as well as the prevalence of the species in the Monteverde Cloud Forest Reserve. Pollinat or limitation has implications in the genetic makeup of populations of I. walleriana . If patches located in favorable conditions for visitation have higher reproductive success, they will be represented to a greater extent in the next generation. Eventuall y, this could result in much higher population sizes in these areas. Influences of self pollination in isolated patches could result in inbreeding depression of these populations. Pollinator limitation of I. walleriana throughout varying environmental cond itions results in limited reproductive success. Since I. walleriana is known to spread to disturbed habitats quickly and easily, it is important to understand the limits in reproduction that it does have. The plant pollinator relationship between I. waller iana and its native population could differ greatly from results obtained in Costa Rica, where it is an exotic. Such is the case for Cytisus scoparius , which experiences pollinator limitation in the western United States due to an absence of natural mutual ists (Parker 1997). Information about changes in fecundity between habitats could help determine the effect of I. walleriana on native species. It could also predict the environments in which I. walleriana is likely to thrive, information that could preven t possible invasion of this exotic. ACKNOWLEDGEMENTS providing constant encouragement and ideas throughout the one that did. Thanks to Karen and Ja vier, for random discussion and ideas in the beginning, and to Ollie and Matt for running around and helping with statistics. Thanks to the EstaciÃ³n Biologica Monteverde, the Monteverde Butterfly Garden, Hotel El Bosque, the Monteverde Institute, Finca Eco lÃ³gica, Hotel de Lucia, Hotel Belmar, the Natural Wonders Walk for allowing me to use their property, and to the Monteverde residents who let me count seed set in their gardens and yards. _________________________________________________________________ _____________________ LITERATURE CITED Agren, J. 1996. Population size, pollinator limitation, and seed set in the self incompatible herb Lythrium salicaria . Ecology 77: 1779 1790.
10 Baker, E. 1993. The relationship of patch size in Impatiens walleriana to herbivory and insect visitation. EAP 1993 Course book, pp. 44 49. Bierzychudek, P. 1981. Pollinator limitation of plant reproductive effort. The American Naturalist 117: 838 840. Bronstein, J.L., Kinsman, S., Murray, K.G. 2000. Plant animal interac tions . In: Monteverde . N.M. Nadkarni and N.T. Wheelwright, eds. Oxford University Press, Oxford, p. 252. Herrera, C. M. 1995. Microclimate and individual variation in pollinators: flowering plants are more than their flowers. Ecology 76: 1516 1524. Kunin , W.E. 1997. Population size and density affects in pollination:pollinator foraging and plant reproductive success in experimental arrays of Brassica kaber . The Journal of Ecology 85: 225 234. Menges, E.S. 1991. Seed germination percentage increases with population size in a fragmented prairie species. Conservation Biology 5: 158 164. Morris, A. 2005. Functional differences between color morphs of Impatiens walleriana. CIEE, Spring 2005. Parker, I.M. 1997. Pollinator limitation of Cytisus scoparius (Sco tch Broom), an invasive exotic shrub. Sargent, S. 1990. Neighborhood effects on fruit removal by birds: a field experiment with Viburnum dentatum (Caprifoliaceae). Ecology 71: 1289 1298. Sih, H. & Baltus, M. 1987. Patch size, pollinator behavior, and p ollinator limitation in catnip. Ecology: 68: 1679 1690. Vazques, D.P. & Simberloff, D. 2004. Indirect effects of an introduced ungulate on pollination and plant reproduction. Ecological Monographs 74: 281 308. Zimmerman, J.K. and T. M. Aide. 1989. Patter ns of fruit production in a neotropical orchid: pollinator vs. resource limitation. American Botany 76: 67 73.
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xlink http:www.w3.org1999xlink xmlns:xsi http:www.w3.org2001XMLSchema-instance
leader 00000nas 2200000Ka 4500
controlfield tag 008 000000c19749999pautr p s 0 0eng d
datafield ind1 8 ind2 024
subfield code a M39-00359
Limitacin de polinizadores y la produccin de semillas de Impatiens walleriana
Pollinator limitation and seed set of Impatiens walleriana
Plants that are pollinator limited may show important fitness relationships with factors influencing visitation. Groups of 60 bagged, natural, and hand-pollinated flowers of Impatiens walleriana (Balsaminaceae) showed much higher seed set with pollination flowers produced seed pods 21%, 70%, and 100% of the time, respectively. I also found higher seed set for high visitation sites in the Monteverde Butterfly Garden with naturally visited patches outside mean seed set inside was 146.5 per 60 stems, while outside it was only 81. Neighborhood, patch size, sunlight, and elevation are speculated to play a role in pollinator visitation. I compared seed set of these variables among patches throughout Monteverde, Costa Rica. With a mean of 81 seed pods per 60 stems, patches located in the vicinity of other flowering plants have a significantly higher seed set than isolated patches, with a mean of 55.6 seed pods per 60 stems. There is a slight tendency for seed set to increase with patch size, but this relationship is not
significant. Flowers located in higher sunlight produced significantly more seed pods than those in the shade with means of 5.6 and 2.4 pods per 5 stems, respectively. Finally, seed set increases along an altitudinal gradient. It appears that I. walleriana is pollinator limited and that nearby flowers of other species, higher sunlight and higher elevation increase the likelihood of seed set through greater visitation.
Las plantas que estn limitadas por sus polinizadores pueden mostrar importantes relaciones adaptativas con factores que influencian las visitas. Los grupos de 60 flores de Impatiens walleriana (Balsaminaceae) cubiertas por bolsas, silvestres y cultivadas, mostraron una produccin de semillas mucho mayor con polinizacin las flores produjeron vainas con semillas el 21%, 70% y 100% del tiempo, respectivamente. Tambin se encontr una produccin mayor de semillas en los sitios con ms visitas en el Jardn de Mariposas de Monteverde con grupos de flores silvestres en el exterior la produccin promedio de semillas adentro fue de 146.5 en 60 tallos, mientras que afuera slo se produjeron 81. Se especul que los alrededores, el tamao del grupo, la luz solar y la elevacin podran afectar las visitas de los polinizadores. Las producciones de semillas de estas variables se compararon entre grupos en Monteverde, Costa Rica. Los grupos en los alrededores de otras plantas con flores tuvieron una produccin de semillas significativamente mayor, con un promedio de 81 vainas en 60 tallos, que las flores en grupos aislados con un promedio de 55.6 vainas en 60 tallos. La produccin de semillas present una tendencia leve a aumentar con el tamao del grupo, pero esta relacin no fue significativa. Las flores en condiciones de luz solar altas produjeron significativamente ms vainas con semillas, con un promedio de 5.6 por cada 5 tallos, que las flores en la sombra (con un promedio de 2.4 por cada 5 tallos). La produccin de semillas aumenta con la elevacin. Se sugiere que I. walleriana est limitada por sus polinizadores y que las flores cercanas de otras especies, ms luz solar y elevaciones ms altas aumentan las posibilidades de produccin de semillas debido al aumento en las visitas de los polinizadores.
Text in English.
Monteverde Biological Station (Costa Rica)
Estacin Biolgica de Monteverde (Costa Rica)
Tropical Ecology Spring 2005
Monteverde Butterfly Garden (Costa Rica)
Ecologa Tropical Primavera 2005
Jardin de Mariposas de Monteverde (Costa Rica)
t Monteverde Institute : Tropical Ecology