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 M39-00070
Solberg, Erik Z.
El parasitismo dependiente de la densidad en Centropogon solanifolius (Campanulaceae) y los cambios en la longevidad floral y la fase sexual siguientes a la infestacin
Density-dependent parasitism in Centropogon solanifolius (Campanulaceae) and changes in floral and sex-phase longevity following infestation
Centropogon solanifolius flowers (Campanulaceae) are parasitized by the fly larvae of Zygothrica neolinea (Drosophilidae). The larvae burrow into the anther of the developing flower bud and eat the pollen, which changes the longevity dynamics among the infested, protandrous
flowers. This study investigated whether the flies were parasitizing in a density dependent manner. The study also explored strategies that flowers used to cope with the high infestation rates (!50% in a previous study and documented here as well). To find answers to these questions, 83 flowers were located, their infestation status noted, and nearest neighbor distance
measured. Additionally, 78 other flowers were bagged and tagged to look for changes in overall flower lifetime, as well as the longevities of the male and female phases. The results show that group dependence parasitism is operating, with 33% more grouped flowers parasitized (n = 83)
("# = 5.55, df = 1, P < 0.05). Parasitism based on distance to nearest neighbor is also observed (t = 2.261, P = 0.0266). To cope with parasitism and the loss of available pollen in the system, infested flowers shorten their male phase by approximately two days, from 4.2 2.4 to 2.5 2.1 days (t = 2.26, P < 0.21), while experiencing no significant difference in overall flower longevity (t = 0.55, P < 0.59). This results in an implied lengthening of the female phase in infected flowers, increasing the chances of pollination in the pollen-depleted system.
Las flores de Centropogon solanifolius (Campanulaceae) son parasitadas por las larvas de la mosca de Zygothrica neolinea (Drosophilidae). Las larvas hacen una madriguera en la antera del botn de la flor y comen el polen, esto cambia la dinmica de la longevidad entre las flores protandras infestadas. Este estudio investigo si las moscas estaban parasitando de una manera dependiente de la densidad de las flores. El estudio exploro tambin las estrategias que utilizan las flores para hacer frente a los altos valores de infestacin (50% en un estudio previo y documentado aqu tambin).
Text in English.
Tropical Ecology 2007
Ecologa Tropical 2007
Longevidad de la fase del sexo
Longevidad de la fase floral
t Monteverde Institute : Tropical Ecology
1 Density dependent parasitism in Centropogon solanifolius (Campanulaceae) and changes in floral and sex phas e longevity following infestation Erik Z. Solberg Department of Biology, University of Wisconsin Madison ABSTRACT Centropogon solanifolius flowers ( Campanulaceae) are parasitized by the fly larvae of Zygothrica neolinea (Drosophilidae). The larvae burrow into the anther of the developing flower bud and eat the pollen, which change s the longevity dynamics among the infested , protandrous flowers. This s tudy investigated whether the flies were parasitizing in a density dependent manner. The study also explored strategies that flowers used to cope with the high infestation rates ( ! 50% in a p revious study and documented here as well ) . T o find answers to the se questions, 83 flowers were located , their infestation status noted, and nearest neighbor distance measured. Additionally , 78 other flowers were bagged and tagged to look for changes in overall flower lifetime, as well as the longevities of the male and f emale phases. The results show that group dependence parasitism is operating, with 33% more grouped flowers parasitized (n = 83) ( "# = 5.55, df = 1, P < 0.05). Parasitism based on distance to nearest neighbor is also observed (t = 2.261, P = 0.0266). T o cope with parasitism and the loss of available pollen in the system, infested flower s shorten their male phase by approximately two days , from 4.2 Â± 2.4 to 2.5 Â± 2.1 days (t = 2.26, P < 0.21 ), while experiencing no significant difference in overall flower longevity (t = 0.55, P < 0.59) . This results in an im plied lengthening of the female phase in infected flowers , in creasing the chan ces of pollination in the pollen depleted system. RESUMEN Las f lores de Centropogon solanifolius ( Campanulaceae) son parasit adas por las larvas de mosca de Zygothrica neolinea (Drosophilidae). Las larvas hacen una madriguera en la antera del botÂ—n de la fl or y comen el polen, esto cambia la dinÂ‡mica de la longevidad entre las flores protandricas infestad a s . Este estudio investigÂ— si las moscas parasit adas dependiendo de la densidad de las flores . El estudio explorÂ— tambiÂŽn las estrategias que las flores utilizaron para enfrentarse con l os altos valores de infestaciÂ—n (=50 % en un estudio previo y documentado aquÂ’ tambiÂŽn) Para encontrar las respuestas a estas preguntas, 83 flores se localizaron, y se midio la distancia a su vecino mas cerca . Adicionalment e, otras 78 flores se empaquetaron y marcaron para buscar los cambios en la vida general de flor, asÂ’ como las longevidades de las fases masculinas y femeninas . Los resultados muestran ese parasitismo depende del grupo , con 33% mÂ‡s de parasitismo en flores agropadas (n = 83) ( " # = 5.55, df = 1, P < 0.05). El nivel de p arasitismo se basÂ— tambiÂŽn en la distancia al vecino mas cercano (t = 2.261, P = 0.027). Para enfrentarse con el parasitismo y la pÂŽrdida de polen disponible en el sistema, las flores infestadas
2 acortan su fase ma sculina por aproximadamente dos dÂ’as, de 4,2 Â± 2,4 a 2,5 Â± 2,1 dÂ’as (t = 2.26, P < 0.213), al experimentar ninguna diferencia significativa en la longevidad general de flor (t = 0.55, P < 0.59). Esto tiene como resultado un alargamiento implicado a la fase femenina , aumentando las oportunidades de polinizaciÂ—n en el sistema que tiene el polen agotado o escaso . INTRODUCTION The protandrous f lowers of Centropogon s olanifolius (Campanulaceae) , a hummingbird pollinated , understory plant in the Monteverde Cl oud Forest Reserve (Zuchowski 2005) , are parasitized by the fly Zygothrica neolinea (Drosophilidae) ( Weiss 2000 ). Zygothrica neolinea deposits its eggs into the corolla of a developing bud and larvae subsequently burrow into the anther and eat the pollen , effectively lowering male reproductive success (Weiss 2000). Infestation rates by the fly are very high in Monteverde, Costa Rica; in surveys in 1987 and 1988, between 50 71% of t he flowers were infested (Weiss 1996). C entropogon solanifolius can be found in groups or solitarily , which raises the question of whether the parasitism by Z. neolinea is density dependent . This question has not been explored previously for C. solanifolius and was raised as an important area of future research in past studies (We iss 1996). Although prevalent, density dependent predation is not ubiquitous in all biological systems . In one study, 16 of the 32 separate studies that were compiled to evaluate mathematical models of density dependence parasitism were found to have a po sitive relationship between parasitism a nd density (Lessells 1985). At the same time, another study only found a positive density dependence in 25% of the 171 insect host parasitoid interactions that were analyzed , with the remaining 75% having independenc e from density or an inverse relationship with it (Stiling 1987) . Given the variation in the prevalence of density dependent parasitism, it was important to establish whether Z. n eolinea was par asitiz ing the flowers in a density dependent manner and with w hat frequency the parasitism was occurring . Regardless of how the parasite cho o se s which flowers to lay its eggs in, C. solanifolius flowers are negatively affected by the fly parasite . Due to its pollen being eaten, its genes are not available to be pas sed on and the plant suffer s decreased reproductive success (Weiss 1996) . Because the plant s are unable to control their proximity to each other , and therefo re potential density dependent e ffects, it can cope with being parasit ized by using one of three lo gical response strategies. First, it can drop the flower before it open s , or decrease the floral longevity, effectively lowering its energy investment in the flower . Second , it can increase the longevity of the flower, allowing for a longer female phase wi thout affecting male phase longevity. Third, the overall flower longevity can stay the same, but it can shorten the male phase and lengthen the female phase, which would increase its chances of receiving pollen in the pollen depleted system. The purpose of this study is to investigate which of the coping strategies is used by C. solanifolius in response to parasitism. While previous studies have shown that the population wide average male longevity decreases in heavily parasitized populations, the longevi ty of parasitized vs. unparasitized flowers has not been documented (Weiss 1996) . By investigating the longevity of the male and female phase in nonparasitized and parasitized flowers, this study
3 will provide novel information to allow biologists to discr iminate between possible response strategies of C. solanifolius . Further , this study will evaluate whether density dependent parasitism occurs. If it does , and is coupled with a shortening of the male phase in these dense areas, as expected, uninfested flo wers in the area can increase fitness by having a higher ratio of their pollen in the system. Similarly, density dependent predation could decrease the amount of pollen in close proximity for the females , thus favoring a lengthening of the female sex phase , which would increase the chances of being pollinated . METHODS The s tudy was conducted in the forest adjacent to Estaci Â— n Biol Â— gi c a in the lower montane cloud forests of Monteverde, Costa Rica , between April 11 and May 7, 2007 . The flowers of C. solan ifolius were located by walking on Sendero DivisiÂ—n , as well as on Sendero Mirador. All of the plants were located a few meters from the main trail and were between 1700 1800 meters in elevation . For the density dependence data, 83 flowers were located. These flowers had to be open flowers because this was the most reliable state to asses s infestation . When the larvae leave the flower, ( usually right before anthesis ) , they chew a small exit hole near the base of the corolla (Weiss 2000) . FIGURE 1: Ill ustrations of a Centropogon solanifolius (Campanulaceae) m ale phase flower (Top) and flower bud (Bottom) parasitized by the fly Zygothrica neolinea (Drosophilidae). Top: An exit hole has been made at the base of the corolla by the larvae as it leaves the f lower, which happens just before anthesis ; Bottom: A bud with a latex droplet that was left by the mother during egg deposition. Modified from Weiss 2000 . Latex droplets are sometime s left by the mother when she lays eggs in the flower (Weiss 1996) , but b ecause these were never observed during this study, the only way to assess infestation status was by looking for exit holes in open flowers . Once a flower was found, the
4 distance to its nearest open flower neighbor was measured (in cm) and the infestation status of both flowers was noted . For analyzing the density dependence results, I discriminated between group and solitary flower s . To do so, the neighbor distances were added up for all flowers and the average was taken. Any flowers that were closer to their neighbor than the average value w ere considered group flowers ; those further away than the average distance were considered solitary flowers. For the flower phase longevity data, 33 plants and 78 small flower buds were used in the study. Forty fou r of the buds were tagged with an individual number and enclosed in fine mesh bags, with the bottom of the bag being fas tened shut with flagging tape. These were chosen as potentially unparasitized buds, but their status was definitively confirmed once ant hesis occurred. For this subset, small buds were chosen because they are less likely to have parasites when very young (Weiss 2000). Buds were covered with a mesh bag to prevent parasitism, and when there were multiple flowers in close proximity, a single bag was used to cover all of them . The other 34 flower buds in the study were simply tagged with labeled flagging tape to indicate that this subset would be left exposed to parasites. Because I wanted these flowers to be parasitized, finding them early in their development was not as critical. These flowers were visited five times a week , and the dates when they opened , as well what sex phase they were expressing , was noted . When th ere was a day lag between data collection , notes were made as to the likel ihood of a flower opening or changing sex during the day without observation. I had high confidence in my ability to predict correctly if a flower would open during a day that I was not going to visit, based on visual cues ( such as small cracks starting at the apex of the flower, as well as its size ) . Additionally, if a flower changed from a male phase to a female phase during the non observation day, it was easy to determine which day the change occurred. As C. solanifolius flower s change from the male to female phase, t he stigma barely protrude s from the anther on the firs t female phase day and continue s to grow out of the corolla afterwards . The day of sex change was easily inferred based on how far the female was protruding from the anther. The infes tation status of the flower was also checked each day, and only if a hole was present, the flower was marked as infested . The female stage was considered complete once either the whole flower or the s tigma fell off. RESULTS Density Dependence Among the flowers used in the density dependence data, the frequency of floral infestation was 60% (n = 83 ). There was no correlation between group size and infe s t at ion rate (R = 0 .090, P > 0.05, n = 41). Additionally, there was no relationship between one flower ' s infestation status and the infestation status of its neighbor ( ! 2 = 3.08, df = 1, P > 0.05). The percentage o f flowers in patches being infested, 66% (n = 6 8), was higher than the percentage of solitary flowers that were infested, 33% (n = 15). Thus , soli tary flowers are much less likely than group ed flowers to be infes ted by their parasites ( ! 2 = 5.55, df = 1, P < 0.05). Additionally, the flower's distance to
5 its neighbor w as significantly lower for infested flowers than for uninfes ted flowers ( t = 2.26 , P = 0 .027 ) . Sex Phase and Overall Flower Longevity The frequency of infestation for the flowers used in the longevity portion of the study was 51% (n = 78) . Seventy percent of the plants had at least one infested flower (n = 33). Among the infes ted and u ninfe s ted flowers, there was no significant difference between abortion rates ( ! 2 = 0.06, df = 1 , P > 0 .05) . There was also no significant differe nce in the overall flower longevity , or the length of the female phase, for the infested and uninfested flowers ( t = 0.55, P < 0.59 and t = 1.14, P < 0.26). The overall flower lifetimes were 13.1 Â± 3.1 d ays for infested and 12.5 Â± 3.7 days for uninfested (see figure 2 ) . The average infested female lived for 10 Â± 2.8 days, while the uninfested females li ved 8.9 Â± 3. 6 days (see figure 2 ). The m ale phase was significantly shorter in infested flowers , with the average being 2.5 Â± 2.1 days for infest ed and 4.2 Â± 2.4 days for uninfes ted (t = 2.26, P < 0. 0 22 ) (see figure 2 ) . FIGURE 2 : This bar graph shows the average lif etimes of infested and uninfested flowers, separated into male, female and total. Error bars are included. The lifetimes, in days, are as follows from left to Right: Un infested male, 4.2; infested male, 2.5 ; un infested females, 8.9; infested female 10 .0 ; u n infested total , 12.5 ; infested total , 13.1 .
6 D ISCUSSION This study and a previous investigation into parasitism of C. solanifolius by Z. neolinea, show that infestation rates are high at Monteverde. In both studies, no less than 50% of the flowers in the population are infested by the parasite (Weiss 1996). The questions that I addressed were whether high infestation rates were density dependent, and whether flowers adaptively altered their phenology to try to maximize the likelihood of pollination und er parasite mediated population sex ratio biases. The most important finding related to the spatial pattern of parasitism of C. solanifolius by Z. neolinea is that flowers in groups were more likely to be parasitized than those that were solitary. That is to say, so long as the flowers were in a group of two or more, they were more likely to be parasitized; but a group of five was no more likely to be parasitized than a group of two. Although my study was unable to detect that larger groups were parasitize d more heavily, I expect that it would be true with a larger sample size or with larger patches . My study had too fine of a scale and too small of a range (I only had groups of one to five individuals) to detect a trend. I was, however, able to detect that individuals that were closer to their neighbor were more likely to be parasitized, which strengthens the argument for distance dependent parasitism. Jan zen (1970) show ed that parasites can act in both density dependent and distance dependent manners, so my finding of distance dependent parasitism is not unique . The result of the distance dependen t parasitism i n Janzen's study was that trees in tropical forests were not likely to be found in great densities . These results were not possible for p lants of C. s olanifolius , as in my observations they seemed to congregate in well lit areas , such as tree fall gaps. These growing preferences push them towards congregations w i th high densities . Therefore, Janzen's result of having trees (or plants) found in lower d ensities is not expected for the plants of C. solanifolius . The plant's reactions to the dist ance and group density dependence will be explained momentarily , when response strategies are discussed . I am able to conclu de that distance to neighbor , and grou ping status are both important factors in determining if a flower will be parasitized. With that in mind, coupled with the fact that the parasitism was 33% greater for grouped flowers , and 82% of the flowers in my study were grouped , it is obvious that the flowers need a strategy to cope with the parasite. Pests parasitizing plants by using them as a food source has been found in many studies . For instance, Thompson (1983) found that moths of the genus Depress aria feed mostly on the flowers and immature s eeds of their host plants . Similarly, Needham (1948) found that Bide n s p ilosa was both the host and food source to a variety of insect parasites, such as Cyclorrhapous flies and several moths , which fed on the heads of the flowers. Furthermore, and similar to the flowers of C. solanifolius , B. pilos a had three species of fly larvae, Agromyza vivens, Paroxyna picciola and Xanthaciura insecta , that eat the pollen, seed sap and flower heads, respectively, after boring into the base of the corolla . Nothing is k nown about a change in flowering phenology or flower structure of B. pilosa in response to this parasitism. Another example of pests eating plants was seen in this study of C. solanifolius when a beetle pest from the family Carabidae was found eating the c orolla, anther and stigma of the study flowers ( Appendix 1).
7 Parasitic relationships can push plants and trees to adapt, through natural selection, the timing of flowering and leaf flushing . One example is Enterlobium cyclocarpum , which ha s new leaf produ ction four to eight w eeks before the beginning of the rainy season. They do this in order to have tougher leaves at the onset of the rainy season, when most of the herbivore populations increase exponentially (Janzen 1983) . The flowers of C. solanifolius a re experiencing similar pressures by floral parasites and the results here show that they also adaptively respond with changes in phenology. As previously discussed , C . solanifolius has three possible response strategies to the parasitism of its anthers. The first option is to decrease the longevity , or abort parasitized flowers before they open. My results indicated that neither of these options is employed. First, t here was no difference in abortion rates between infested and uninfested flowers. Secondl y, the floral longevity for infested and uninfested flowers was not statistically different. The next option would be to increase the overall lifetime of the flower , passing the increased longevity on to the female phase. This would allow the female to be open longer and increase its chances of being pollinated , and thus increase the reproductive success of the plant. However, I found no evidence to support this , based on the fact that the female phase and the overall longevity were not significantly differ ent for infested and uninfested flowers. The last response strategy the plants could have is to keep the overall flower longevity the same, but decrease the male phase and therefore increase the female phase . My data lend support for this s trategy . Evide nce showed that infested males lived approximately two days less than the uninfested males and that the longevity of the infested and uninfested flowers was not significantly different . Although I was unable to obtain significant information on a change in the length of the female phase , simple logic would say that if the total flo ral longevity is similar i n infested and uninfested flowers ( approximately 13 days ) , and the infested male phase is about two days less than in the uninfested flowers, then the fe male phase would have to become longer in the infested flowers to compensate for the decrease in male phase longevity (figure 2) . I expect that with a bigger sample size, I would have obtained this result of lengthening of the female phase and shortening o f the male phase, with the total lifetime of the infested and uninfested flowers being the same. Biologically speaking , the flower 's response is the most energetically favorable and will afford the highest reproductive success. When pollen is consumed by larvae, it is not available to pollinate the females. Logically, the male phase would shorten, be cause the pollen depleted male has significantly decreased reproductive potential . At the same time, the parasitized flower can increase reproductive success by extending the expression of the female phase , thereby increasing its chances of receiving pollen in the pollen depleted system. This way, the flower does not need to in crease its overall energy input; i nstead it simply shift s the energy input for each s ex phase. In reality, the most reproductively successful option for an infes ted plant would be to completely skip the male phase and only express the female phase , effectively becoming an imperfect flower. In a dense patch that is being hit heavily by th e larvae, the pollen availability will be even lower and thus using energy from the male phase to express the female phase for a
8 extended period of time greatly increases likelihood of pollination. This phenomenon of completely skipping the male phase was actually seen in three of the infested flowers of my study , as well as some of the flowers in Weiss' study (Weiss 2000) . Likewise , if a flower is uninfested, it can increase its reproductive fitness by never expressing the female stage , especially if it is in a highly dense patch where other males have been parasitized . I f there is less pollen in the system, it is reproductively favorable for a flower with a lot of pollen to put all of its energy into getting its pollen into the system. This phenomenon wa s seen in one uninfested flower in my study . In conclusion, because the flowers are more commonly found in groups and therefore more likely to be parasitized, they must converge on a strategy that will allow similar or increased reproductive fitness compa red to when they are un parasitized. The flowers of C. solanifolius seem to have evolved a s trategy to cope with the parasitism by Z. neolinea . By shifting their energy input rat ios for the parasitized flowers, t hey seem to be increasing their investment i n the female phase at the expense of the male phase, while k eep ing the overall investment in a flower the same. This behavior increases the flower's chances of be coming pollinated in the pollen depleted environment that the Z. neolinea has forced upon the plant populations. Future studies should explore a possible increase in female phase longevity in parasitized flowers. I feel this could be accomplished with a larger sample size. Additionally, it would be important to investigate further possible densit y dependence of the parasitism. D istance and group dependent parasitism were shown , and with a larger scale and larger range of group sizes, I believe density dependence would also be seen. Additionally, it might be interesting to look for the complete abs ence of male or female phases , which would suggest that the flowers may be moving towards the use of imperfect flowers . Finally, a search of the flowers for the Carabid beetle that was observed eating the flower parts, and further parasitizing the flower , may bring about more interesting flo ral longevity reactions. ACKNOWLEDGEMENTS I would like to thank Karen Masters for help in developing the project and for support in analyzing my data and statistical tests. I would also like to thank EstaciÂ—n BiolÂ—gi ca de Monteverde for use of the forests to conduct my studies. Additionally, I would like to thank Camryn Pennington and Tom McFarland for support while writing my final report. Finally, I would like to thank both Alan and Karen Masters for equipping me wi th the knowledge of Tropical Biology that allowed me to interpret my study. LITERATURE CITED Janzen, D.H. 1983. Enterlobium cyclocarpum (Guanacaste, Ear Fruit). In: Costa Rican
9 Natural History. D.H. Janzen, ed. The University of Chicago Press, Chicag o, IL, pp. 241 243 . Janzen, D.H. 1970. Herbivores and the Number of Tree Species in Tropical Forests. The American Naturalist . Vol. 104, No. 940, pp 501 507. L essells, C.M. 1985. Parasitoid F oraging: Should Parasitism be Density Dependent. The Journal of Animal Ecology, Vol. 54, No. 1. pp 27 41 . Needham , J. G. 1948. Ecological Notes on the Insect Population of the Flower H eads of Bide n s p ilosa . Ecological Monographs , Vol. 18, No. 3. Solis, A . 1999. Family: Carabidae. In: Escarabajos de Costa Rica. C. Agu ilar, D. Avila ed. Insitituto Nacional de Bi odiversidad (INBio), Heredia, Costa Rica. pp 48 49. Stiling, P. D. 1987. The F requency of Density Dependence in Insect Host Parasitoid Systems. Ecology, Vol. 68, No. 4, pp 844 856. Thompson, J. N. 1983. Selec ti on Pressures on Phytophagous Insects Feeding on S mall Host Plants. Oikos 40: 438 444. Weiss, M. E. 2000. A Fly Larvae Directly Alters Floral Sex in C. solanifolius . In: Monteverde: Ecology and Conservation of a Tropical Cloud Forest. N.M. Nadkarni, N.T. Wheelwright, ed. Oxford University Press, New York, NY. pp. 278 279. Weiss, M. E. 1996. Pollen Feeding Fly Alters Floral Phenotypic Gender in Centropogon solanifolius (Campanulaceae). Biotropica 28 (4b): 770 773. Zuchowski, W. Centropogon s olanifolius , Fam ily: Campanulaceae. In: A G uide to Tropical Plants of Costa Rica. D. Featherstone ed. Distribudores Zona Tropical, S.A., Miami, FL. pp. 290. A P PENDIX Potential Further Parasitism of Centropogon Solanifolius by a Beetle of the Family Carabidae: 1) There were five flowers that had a lot of damage to their corolla , anther and stigma . I assume that this was done by a feeding beetle of the Carabidae family , because the beetles were found in three of the five flowers . They were small (8 10mm), flat and b lack . None of the five flowers were used in the statistical analyses of this study. There are app roximately 2200 species of Carabid beetles in Costa Rica. Most of the larvae are nocturnal and live among forest leaf litter. Their diet normally consists of other insect s but others are known to eat seeds, fruits and other vegetative products (Solis, 1999).