USF Libraries

Passiflora (Passifloraceae) defenses against Heliconius cydno (Nymphalidae: Heliconiinae) oviposition

MISSING IMAGE

Material Information

Title:
Passiflora (Passifloraceae) defenses against Heliconius cydno (Nymphalidae: Heliconiinae) oviposition
Translated Title:
Las defensas de Passiflora (Passifloraceae) contra la oviposición de Heliconius cydno (Nymphalidae: Heliconiinae) ( )
Physical Description:
Book
Language:
English
Creator:
Khuc, Kim
Publication Date:

Subjects

Subjects / Keywords:
Passiflora   ( lcsh )
Nymphalidae   ( lcsh )
Butterflies   ( lcsh )
Passiflora
Nymphalidae
Mariposas
Tropical Ecology 2009
Oviposition
Ecología Tropical 2009
Oviposición
Genre:
Reports   ( lcsh )
Reports

Notes

Abstract:
The purpose of this study was to determine the relative effectiveness of Passiflora defenses: elevated levels of cyanide, egg mimics, and variable leaf shapes, against oviposition from Heliconius cydno. The effectiveness of these defenses would reveal H. cydno’s preferences for oviposition sites and the primary criteria it uses when evaluating oviposition sites. Forty-two Passiflora oerstedii vines from the Monteverde Butterfly Garden, in Costa Rica, were divided into three treatments that added either cyanide, false eggs, or changed the shape of young leaves. Another twelve Passiflora coccinea vines were used in a second experiment that added false eggs to the young leaves, tendrils, and meristems. The results show trends that the various treatments deterred oviposition. Out of the 11 total eggs laid on leaves in the cyanide treatment, 82% of the total number of eggs was on the control leaves, then 9% on the methanol leaves and 9% on the methanol-cyanide leaves, indicating that the chemical cues from methanol discourage oviposition by H. cydno (!2 = 11.64, df = 2, p = 0.003). For the vines with false eggs, out of a total of 19 eggs, 68% of the eggs were on the control leaves (! 2 = 2.58, df = 1, p = 0.11). Lastly, the control leaves of the modified leaf shape experiment had 73% of the total number of eggs, 15, while the modified leaves had 27% (! 2 = 3.23, df = 1, p = 0.07). These trends show that H. cydno likely relies on chemical cues more than visual cues when choosing oviposition sites. In addition, the lack of evidence for larval cannibalism in H. cydno and its willingness to oviposit on multiple Passiflora species demonstrate that egg mimics and variable leaf shapes are less significant factors in oviposition decisions.
Abstract:
El propósito de este estudio fue determinar la efectividad de las defensas de Passiflora: altos niveles de cianuro, mimetismo de huevos y las diferentes formas de las hojas; contra la oviposición de Heliconius cydno.
Language:
Text in English.
General Note:
Born Digital

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
usfldc doi - M39-00153
usfldc handle - m39.153
System ID:
SFS0001329:00001


This item is only available as the following downloads:


Full Text
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 M39-00153
040
FHM
0 041
eng
049
FHmm
1 100
Khuc, Kim
242
Las defensas de Passiflora (Passifloraceae) contra la oviposicin de Heliconius cydno (Nymphalidae: Heliconiinae)
245
Passiflora (Passifloraceae) defenses against Heliconius cydno (Nymphalidae: Heliconiinae) oviposition
260
c 2009-05
500
Born Digital
3 520
The purpose of this study was to determine the relative effectiveness of Passiflora defenses: elevated levels of cyanide, egg mimics, and variable leaf shapes, against oviposition from Heliconius cydno. The effectiveness of these defenses would reveal H. cydnos preferences for oviposition sites and the primary criteria it uses when
evaluating oviposition sites. Forty-two Passiflora oerstedii vines from the Monteverde Butterfly Garden, in Costa Rica, were divided into three treatments that added either cyanide, false eggs, or changed the shape of young leaves. Another twelve Passiflora coccinea vines were used in a second experiment that added false eggs to the young leaves, tendrils, and meristems. The results show trends that the various treatments deterred oviposition. Out of the
11 total eggs laid on leaves in the cyanide treatment, 82% of the total number of eggs was on the control leaves, then 9% on the methanol leaves and 9% on the methanol-cyanide leaves, indicating that the chemical cues from methanol discourage oviposition by H. cydno (!2 = 11.64, df = 2, p = 0.003). For the vines with false eggs, out of a total of 19 eggs, 68% of the eggs were on the control leaves (! 2 = 2.58, df = 1, p = 0.11). Lastly, the control leaves of the
modified leaf shape experiment had 73% of the total number of eggs, 15, while the modified leaves had 27% (! 2 = 3.23, df = 1, p = 0.07). These trends show that H. cydno likely relies on chemical cues more than visual cues when choosing oviposition sites. In addition, the lack of evidence for larval cannibalism in H. cydno and its willingness to oviposit on multiple Passiflora species demonstrate that egg mimics and variable leaf shapes are less significant factors in oviposition decisions.
El propsito de este estudio fue determinar la efectividad de las defensas de Passiflora: altos niveles de cianuro, mimetismo de huevos y las diferentes formas de las hojas; contra la oviposicin de Heliconius cydno.
546
Text in English.
650
Passiflora
Nymphalidae
Butterflies
4
Passiflora
Nymphalidae
Mariposas
653
Tropical Ecology 2009
Oviposition
Ecologa Tropical 2009
Oviposicin
655
Reports
720
CIEE
773
t Monteverde Institute : Tropical Ecology
856
u http://digital.lib.usf.edu/?m39.153



PAGE 1

Passiflora (Passifloraceae) defenses against Heliconius cydno (Nymphalidae: Heliconiinae) oviposition Kim Khuc Department of Integrative Biology, University of California, Berkeley ABSTRACT The purpose of this study was to determine the relative effect iveness of Passiflora defenses: elevated levels of cyanide, egg mimics, and variable leaf shapes, against oviposition from Heliconius cydno The effectiveness of these defenses would reveal H. cydno's preferences for oviposition sites and the primary crit eria it uses when evaluating oviposition sites. Forty two Passiflora oerstedii vines from the Monteverde Butterfly Garden, in Costa Rica, were divided into three treatments that added either cyanide, false eggs, or changed the shape of young leaves. Anot her twelve Passiflora coccinea vines were used in a second experiment that added false eggs to the young leaves, tendrils, and meristems. The results show trends that the various treatments deterred oviposition. Out of the 11 total eggs laid on leaves in the cyanide treatment, 82% of the total number of eggs was on the control leaves, then 9% on the methanol leaves and 9% on the methanol cyanide leaves, indicating that the chemical cues from methanol discourage oviposition by H. cydno ( 2 = 11.64, df = 2, p = 0.003). For the vines with false eggs, out of a total of 19 eggs, 68% of the eggs were on the control leaves ( 2 = 2.58, df = 1, p = 0.11). Lastly, the control leaves of the modified leaf shape experiment had 73% of the total number of eggs, 15, wh ile the modified leaves had 27% ( 2 = 3.23, df = 1, p = 0.07). These trends show that H. cydno likely relies on chemical cues more than visual cues when choosing oviposition sites. In addition, the lack of evidence for larval cannibalism in H. cydno and its willingness to oviposit on multiple Passiflora species demonstrate that egg mimics and variable leaf shapes are less significant factors in oviposition decisions. RESUMEN El prop—sito de este studio fue determinar la efectividad de las defensas de Passiflora : altos niveles de cianuro, mimetismo de huevos y diferentes formas de hoja; contra la oviposici—n de Heliconius cydno. La efectividad de estas defensas se demuestra por los sitios de oviposici—n y el criterio que las mariposas toman para elegir los sitios de oviposici—n. Cuarenta y dos lianas de Passiflora oersedii del Jard’n de Mariposas de Monteverde fueron divididas en tres tratamientos en los que se agregaron cianuro, huevos falsos, o se cambio la forma de la hoja. Otras doce plantas de la especie P. cocc’nea fueron usadas en la segunda parte del experimento en los que se agrego huevos falsos a las hojas falsas, zarcillos y meristemas. Se muestra una tendencia a que varios tratamientos evitan la oviposici—n. De los once tratamientos en lo s que se pusieron huevos, 82% fueron en las plantas que se pusieron huevos falsos, 9% en las hojas con metanol, 9% en los hojas con metanol cianuro, indicando que las pistas qu’micas con metanol ahuyentan la oviposici—n por parte de H. cydno ( 2 = 11.64, d f = 2, p = 0.003). De las lianas con huevos falsos, de un total de 19, 68% de los huevos se depositaron en las hojas control ( 2 = 2.58, df = 1, p = 0.11). Por œltimo las hojas control de la forma modificada experimentaron un 73% del nœmero de huevos, 15 mientras que las hojas modificadas un 27% ( 2 = 3.23, df = 1, p = 0.07). Estas tendencias muestran que H. cydno utiliza m‡s se–ales qu’micas que visuales cuando eligen los sitios de oviposici—n. Adem‡s, la falta de evidencia de canibalismo larval en esta especie y la capacidad de ovipositar en varias especies de Passiflora demuestra que el mimestismo de huevos y la forma variable de las hojas son menos significativas a la hora de elegir los sitios de oviposici—n.

PAGE 2

INTRODUCTION Coevolution is the ph enomenon in which two or more species evolve traits in response to another species reciprocally (Futuyma and Slatkin 1983). The co evolutionary arms race between Heliconiine butterflies and Passifloraceae vines provides an excellent example of a series of coevolutionary responses (Gilbert, 1983). In spite of the many defenses that Passifloraceae species have evolved to inhibit herbivory, species of Heliconiine still find ways to recognize suitable Passiflora sp. on which to lay their eggs. Many members of Passifloraceae have evolved toxic chemical compounds such as saponins, alkaloids, and phenolic and cyanogenic glycosides, which serve to harm herbivores and thus discourage them from eating these plants (Gibbs 1974, in Benson et al. 1975). Certain spe cies have evolved toughened mature leaves to deter herbivory and sharp modified hooked leaf hairs to puncture eggs (Gilbert 1971). Since many members of Heliconiine have evolved to lay their eggs on tendril tips, some Passiflora have small tendrils on the meristem, causing eggs to easily fall off the plant, or lose the tendrils and take the eggs with them (Benson et al. 1975). In addition, some Passiflora species have evolved glands on the leaves, stipules, or tendril tips that mimic the eggs of specific s pecies of Heliconiine (Benson et al. 1975). Some Passiflora species' larvae may be cannibalistic (Benson et al. 1975). Thus, these egg mimics may discourage oviposition when the adult females lay eggs individually (Benson et al. 1975, Gilbert 1975). Mor phologically, many Passifloraceae species have diverged greatly in leaf shape since Heliconiine species rely on vision to search for oviposition sites. Some species have even evolved to mimic leaf shapes of other groups of common tropical plants in an atte mpt to escape detection (Benson et al. 1975). While the Passiflora vines have evolved elaborate defenses specifically against Heliconiine oviposition, Heliconiines have also evolved traits to counter these defenses and to continue using these plants. T he well developed vision and learning ability of these butterflies has aided in the evolution of strategies to overcome Passiflora defenses (Benson et al. 1975). They use chemoreceptors on their antennae, proboscis, and forelimbs to examine the suitabilit y of the host plants for oviposition, including evaluating the authenticity of eggs already present on the plant. Even though chemical compounds synthesized by Passiflora spp. are meant to be deterrents, Heliconiines instead use them as recognition symbol s, and the larvae and adults can use them to protect themselves against predators (Brower and Brower 1964 in Benson et al. 1975). Previous studies have examined the independent impacts of different plant defenses against oviposition. When given a choic e, various Heliconiine species, including Heliconius sara fulgidas, Heliconius charithonia charithonia, and Dryas iulia monerata, prefer to oviposit on leaves with higher concentrations of cyanide (Burkholder 2008). There are several possible reasons for this preference. The larvae could sequester the cyanide and use it for their own defense against predators or remove the nitrogen to simultaneously deactivate the compound and gain a limiting nutrient (Engler et al. 2000). In addition, since other herbiv ores cannot digest or dislike the bitter taste of cyanide, the eggs and larvae benefit from reduced herbivory on the plant and the reduced competition (Engler et al. 2000). The presence of egg mimics on Passiflora leaves reduces the frequency of ovipositi on on plant structures. The more closely the leaf glands mimic the Heliconiine eggs, the more effective they are at deterring oviposition on leaves, because adult females would not want to lose their eggs to cannibalism (Biermaier 2008).

PAGE 3

While the indiv idual effect of each plant defense is known, their relative effectiveness is not. This project aims to examine the relative effectiveness of cyanide, egg mimics, and leaf shape variability at deterring oviposition. Determining the relative effectiveness o f these three plant defenses will elucidate how Heliconiine butterflies prioritize which factors are important in their oviposition site, and which sense they primarily use to help them decide where to lay their eggs. They are expected to use chemorecepti on to smell or taste the cyanide on the leaves, both chemoreception and vision to detect the egg mimics, and vision to identify leaf shape (Gilbert and Singer 1975). Secondly, since H. cydno is known to lay eggs on tendrils and meristems in addition to l eaves, this study also asks if changes in the leaves will affect oviposition preference on the tendrils. In relation to that, a next logical step in the coevolutionary arms race between H. cydno and its host plants is for the host plants to evolve egg mim ics on extra foliar structures, such as tendril tips and meristems. I predicted that vines with false eggs on the leaves, tendril tips, and meristems will deter oviposition. METHODS Study One: Relative Effectiveness of Three Plant Defenses This study was conducted at the Monteverde Butterfly Garden in Monteverde, Costa Rica for 16 days between April 18 th and May 7 th 2009. Passiflora oerstedii and H. cydno were studied in a garden enclosed in a mesh material that allows light, precipitation, and wind t o penetrate. There were approximately 25 other butterfly species in the garden, but their eggs were laid on different host plants, looked different and/or were laid in groups. The garden represents a moist forest edge habitat at 1,300 meters above sea le vel. H. cydno lays small yellow eggs singly on the tendrils and leaves of a variety of Passiflora host plants (Smiley 1978; DeVries, 1987). It is important to use a butterfly species that lays eggs singly in order to test the hypothesis that single egg l aying butterflies will be deterred from oviposition by eggs and egg mimics already present on leaves. P. oerstedii naturally does not have egg mimics, but it does have extra floral nectaries meant to attract ants to defend the plant against herbivory. A total of 42 P. oerstedii were used in this study. Each vine was randomly divided into the three treatments: addition of cyanide, addition of egg mimics, or change in leaf shape. For each treatment leaf and corresponding tendril, the adjacent leaf and its corresponding tendril (either one above or below) were identified as the control. The relative positions of treatment leaves and tendrils compared to control leaves and tendrils alternated. For instance, if on one plant the first fully opened leaf an d tendril from the meristem were given the treatment, and the second leaf and tendril were the controls, on the next plant the first fully opened leaf and tendril from the meristem were the controls and the second leaf and tendril received the treatment. Vines were monitored daily for the presence of eggs, and the number of eggs per leaf and tendril were recorded and removed. The eggs were collected daily at the same time. As the plants grew, older and tougher leaves and tendrils were retired and new lea ves and tendrils were added. Fourteen P. oerstedii vines were given artificially elevated levels of cyanide. The six youngest fully opened leaves of each vine were used because H. cydno primarily lays eggs on the youngest leaves. There were two replicat es of the control, methanol treated, and methanol cyanide treated leaf per vine. Since young leaves have the highest cyanide concentrations, I finely cut up 8.92 grams of new leaf material (of P. oerstedii ) and mixed it in 100 ml of

PAGE 4

methanol. Then I pain ted the leaf extract onto a set of 28 leaves using a small paintbrush. From here on, this treatment is called the cyanide methanol treatment. A second treatment with only methanol was painted onto a second set of 28 leaves in order to account for the eff ects of adding any solution to the leaves. The treated leaves were marked on the underside of the leaves. On day eight, I did a second round during which I added leaf extract and methanol only onto the new fully opened leaves of each treated vine. I use d 2.33 g of new leaf material in 25 ml of methanol for the cyanide methanol treatment. To add artificial egg mimics to leaves, I used yellow paint pens to paint 1 2 small eggs in random spots on each treated leaf on fourteen vines (Figure 1). The four yo ungest fully opened leaves of each vine were used in the experiment: two served as the controls and two had false eggs. On the last set of fourteen vines, I cut treatment leaves so that they had two lobes instead of one, thus changing their leaf shape (Fi gure 2). The four youngest fully opened leaves of each vine were included in the experiment: two controls and two modified. The average new surface area was 92% of the original. FIGURE 1 (left) An example leaf with a H. cydno painted false egg. FIG URE 2 (right) An example leaf with modified shape. Study Two : Egg Mimics on Vines Twelve Passiflora coccinea vines and the same H. cydno population (in the same garden) at the Monteverde Butterfly Garden were used for this study for nine days between A pril 26 th and May 7 th 2009. P. coccinea naturally has extra floral nectarines and tending ants. I painted yellow egg mimics using a paint marker on the three youngest fully opened leaves and tendril tips, and the developing leaf cluster on six of the vin es. Each plant structure received one egg mimic. The other six served as the control. Each vine was randomly chosen as either the control or treatment. Additional Observations Every few days for a total of eight days, I put 2 3 recently collected H. cydno eggs into a container with 4 5 con specific young larvae ranging in size from 0.5 2.5 cm long. In sum, I put in eight eggs into the larvae container. In addition, the larvae were placed on a young P. oerstedii vine kept in water. I checked for th e presence or absence of the eggs in the larvae container daily. Since it usually takes about 8 10 days for eggs to hatch, if they were absent, they were considered to have been eaten or pushed off the leaf by the larvae. Data Analysis For daily coun ts of eggs on leaves and tendrils, the Wilcoxon sign rank test was used. For total counts of eggs per treatment and control on the leaves of the first study and the vines of the

PAGE 5

second study, a chi squared test was used. To analyze the total counts on th e tendrils, I used the multinomial test for the cyanide treatment and the binomial test for the egg mimic and modified leaf shape treatments. RESULTS Study One There were approximately 10 15 H. cydno individuals present in the garden. Unfortunately less oviposition by H. cydno occurred than expected, despite relatively favorable weather conditions and availability of desired host plants. Little courtship and mating were observed and no new individuals of H. cydno were introduced into the garden du ring the study period. In all, cyanide methanol treated, methanol treated and control leaves had a total of 11 eggs. For the cyanide treatment, the control leaves had 82% of the total number of eggs while the methanol and the methanol cyanide leaves eac h had the same percentage of eggs, 9% (Figure 3). For the total number of eggs, there was a statistically significant difference between the control, and methanol and methanol cyanide treatments ( 2 = 11.64, df = 2, p = 0.003), but no difference was detected between the leaves with methanol and methanol cyanide. The control leaves had an average of 0.56 eggs per day (SD = 1.75; Figure 4). The average number of eggs per day on the methanol and met hanol cyanide leaves were the same, 0.06 (SD = 0.25; Figure 4). Thus the effect of methanol on the leaves cannot be eliminated, and the effect of increased levels of cyanide on the leaves cannot be tested (signed rank = 3, df = 15, p = 0.88 between the c ontrol and methanol leaves, and signed rank = 0, df = 15, p = 0.50 between the methanol and methanol cyanide leaves). For the vines with egg mimics, 68% of a total of 19 eggs was laid on the control leaves, and 32% of the eggs were laid on the leaves wi th false eggs ( 2 = 2.58, df = 1, p = 0.11; Figure 3). On a daily average, 0.81 eggs were laid on the control leaves, while an average of .38 eggs were laid on the experimental leaves, but this difference was not statistically significant (SD = 1.17 and 1. 25, respectively, signed rank = 7.5, df = 15, p = 0.86; Figure 4). For the modified leaf shape experiment, the control leaves had 73% of a total of 15 eggs while the leaves with a modified shape had 27% of the eggs ( 2 = 3.23, df = 1, p = 0.07; Figure 3 ). The control leaves had an average of 0.69 eggs per day, and the treatment leaves had an average of 0.25 vines per day (SD = 1.74 and 0.44 respectively, signed rank = 6, df = 15, p = 0.72; Figure 4). Although these differences between the control and treatment leaves in the total number of eggs and average daily number of eggs were not statistically significant, the p value was close to 0.05, meaning that it is very likely that a trend exists.

PAGE 6

FIGURE 3 Methanol, leaf extr act (here denoted as cyanide), egg mimics, and modified leaf shape on P. oerstedii negatively affect oviposition by H. cydno on those leaves. However, for the cyanide treatment, since there is no difference between the methanol leaves and the leaves with methanol and cyanide, then the effect of methanol cannot be ruled out. FIGURE 4 This shows trends between the control and treatment leaves of the three treatments showing that the cyanide, egg mimics, and

PAGE 7

modified leaf shapes inhibit ovipo sition by H. cydno on P. oerstedii, but do not completely prevent it. For the corresponding tendrils, of the vines in the cyanide treatment, the control tendrils had 68% of a total of nine eggs, while the methanol tendrils had 22%, and the tendrils wit h methanol and cyanide had 11% ( 2 = 4.67, df = 2, p = 0.10; Figure 5). As for the daily averages, the control tendrils had 0.75 eggs, the methanol tendrils had 0.25 eggs, and the methanol cyanide tendrils had 0.125 eggs (SD = 0.88, 0.35, and 0.46 respectively; Figure 6). The control tendrils of the egg mimic experiment had 75% of a total of 16 eggs and the treatment tendrils had 25% of the eggs ( 2 = 4.00, df = 1, p = 0.05; Figure 5). The control tendrils averaged 1.5 eggs per day while the experimental tendrils averaged 0.5 eggs pe r day (SD = 2.07 and 0.53 respectively; Figure 6). For the variable leaf shape experiment, the control tendrils had 86% of a total of seven eggs while the treatment tendrils had 14% of the eggs ( 2 = 3.57, df = 1, p = 0.06; Figure 5). Lastly, the daily average for the control tendrils was 0.75 eggs and the daily average for the modified leaf shape tendrils was 0.125 eggs (SD = 0.89 and 0.35 respectively; Figure 6). FIGURE 5 The tendrils of P. oerstedii show strong trends that the three t reatments: cyanide, egg mimics, and different leaf shape on leaves deter oviposition by H. cydno on tendrils.

PAGE 8

FIGURE 6 This graph shows that there are significant trends between the controls and treatments of each treatment g roup of the P. oerstedii tendrils: cyanide, egg mimics, and different leaf shape. The three treatments deterred oviposition by H. cydno but at similar levels of effectiveness. Study Two Of the total 38 eggs that were laid on these vines, the control vines had 58% while the experimental vines had 42% ( 2 = 0.94, df = 1, p = 0.33; Figure 7). There was an average of 2.44 eggs per day on the control vines and an average of 1.78 eggs per day on the experimental vines (SD = 1.01 and 1.78 respectively; signed rank = 1.5, df = 8, p = 0.54; Figure 8). Becau se the p values are so high, only a weak trend that the vines with egg mimics deterred oviposition existed.

PAGE 9

FIGURE 7 There is a trend showing that the presence of egg mimics on the leaves, tendrils, and meristems on P. coccin ea somewhat deter oviposition by H. cydno FIGURE 8 The addition of egg mimics on the leaves, tendrils, and meristems to P. coccinea vines weakly inhibits oviposition by H. cydno Additional Observations

PAGE 10

The larvae did not eat the recent ly collected eggs. Every day after I put the new eggs into the container, I found them in their original positions. The 4 5 larvae were observed eating only plant material. At the end of the eight day study period, the eggs were present and remained unha tched. In addition, while both P. oerstedii and P. coccinea have extra floral nectaries, only P. coccinea was observed to have ants tending the vines. DISCUSSION The first study shows that H. cydno responds to chemical cues from the cyanide treatment m ore than visual cues from the egg mimics and variable leaf shape experiments when evaluating oviposition sites. The statistically significant difference between the control, and the two sets of treated leaves, indicates that the H. cydno butterflies are a ble to detect the addition of methanol. Given that there is no visual difference between the control and methanol leaves after the methanol dried, it is deduced that they use chemoreception to recognize the methanol. Therefore, the smell or taste of the methanol is a strong deterrent against H. cydno oviposition. Next, the ratios of the numbers of eggs on the control to treatment leaves of the egg mimic and variable leaf shape experiments of the first study are very similar (Figures 3 and 4), thus in thi s study, it is not possible to determine which of these two leaf characteristics is a better deterrent against H. cydno oviposition nor which host plant characteristic, egg mimics or leaf shape, H. cydno relies on more when choosing oviposition sites. T he results from the eggs laid on tendrils support the results from the corresponding leaves: that the three treatments deter H. cydno from oviposition. Again, the negligible difference between the methanol and the methanol cyanide leaves of the cyanide tr eatment indicates that the effect of elevated cyanide levels only is not testable. Many species of Heliconius lay on new plant growth and tendril tips to avoid predaceous ants that tend the vine (Turner 1981). However, once the larvae are too large, they will eventually have to move down to the adjacent leaf to feed, thus leaf features for these Passiflora spp. are still an important factor in discouraging oviposition. Although these eggs are laid on tendrils, the adult females are still deterred from la ying eggs near leaves that have methanol, methanol and cyanide, egg mimics, or a different leaf shape. A comparison of the cyanide treatment versus the egg mimic and variable leaf shape treatments show that H. cydno likely relies more heavily on chemical cues than visual cues when choosing oviposition sites. The results from the second study on P. coccinea support this hypothesis. False eggs on leaves, tendril tips, and meristems only mildly deterred oviposition from those vines. While a different leaf shape and the presence of false eggs somewhat reduce oviposition, they do not completely stop it. Possibly, the high leaf shape variation among Passifloraceae species has led H. cydno to not heavily rely on leaf shape as an indicator of suitable host pla nts. Unlike many other Heliconius species, H. cydno can lay eggs on a variety of Passiflora host plants, each of which has a different leaf shape. Examples include Passiflora vitifolia, Passiflora biflora, Passiflora oerstedii, Passiflora coccinea, and P assiflora cyanea whose leaves have one to three lobes (Corrales 1996; DeVries 1987). Another study on H. cydno showed that although the presence of egg mimics increases the search time for good oviposition sites and reduces oviposition frequency, they do not completely prevent it from occurring on plant structures with egg mimics (Williams and Gilbert 1981). Perhaps in this species, the presence of eggs on leaves or other plant structures are not as important because the larvae are not highly cannibalist ic. My observations of the non cannibalistic larvae support this hypothesis.

PAGE 11

Ovipositing females that either cannot detect the presence of false eggs or can successfully determine the difference between real and false eggs are rewarded, because their lar vae hatch to find themselves on healthy leaves and plant material. In the future, a longer study with a larger population of ovipositing H. cydno would provide more conclusive evidence for H. cydno's preference for oviposition sites and the criteria it p rimarily uses to evaluate oviposition sites. Furthermore, Passifloraceae species have evolved many other defenses, including extra floral nectaries and trichomes that are also meant to deter oviposition. Testing the relative effectiveness of these defens es in addition to the ones previously studied would add to our understanding of how Heliconiine species respond to their host plants. It would also be interesting to determine if the mere presence of extra floral nectaries on a Passiflora species but abse nce of tending ants changes Heliconius oviposition behavior in comparison to a Passiflora species that has both extra floral nectaries and tending ants. While this study did not directly answer the initial question of which plant defense is most effective at deterring oviposition, it did reveal several conclusions. H. cydno uses its chemoreception more than vision when choosing oviposition sites. This species's willingness to oviposit on multiple host plants and the absence of larval cannibalism relaxes the selection for this species to use leaf shape and presence of eggs as important indicators for desirable oviposition sites. Furthermore, leaf characteristics affect which tendrils H. cydno chooses for oviposition. Since this species relies more on chem oreception than vision, this has important implications for its coevolutionary arms race with Passifloraceae plants. Plant species that evolve chemical indications of undesirable oviposition sites are likely to be more effective at deterring oviposition f rom H. cydno than species that evolve visual cues. Changes in leaf characteristics will also affect Heliconiine species such as H. cydno that evolved to lay their eggs on tendrils. ACKNOWLEDGMENTS I am grateful to the staff at the Monteverde Butterfl y Garden for permitting me to conduct my study there and for their willingness to answer my questions, especially Eli, Ana Luisa, Marvin, and Julio. I would also like to thank Yi men Araya for his help in identifying plants and performing statistical anal yses. Lastly, I am deeply appreciative to Dr. Alan Masters for his assistance in designing my project and for helping me throughout the research process. LITERATURE CITED B ENSON W.W., K.S. B ROWN AND L.E. G ILBERT 1975. Coevolution of plants and herb ivores: passion flower butterflies. Evolution 29 (4): 659 680. B IERMAIER N. 2008. Oviposition on the host plant, Passiflora biflora (Passifloraceae), by Heliconiine, in the presence of egg mimics. In: CIEE Spring 2008: Tropical Ecology and Conservation. CIEE. Council on International Educational Exchange, Monteverde, Costa Rica, pp. 178 187. B ROWER L.P. AND J. V .Z. B ROWER 1964. Birds, butterflies, and plant poisons: A study in ecological chemistry. Zoologica 49: 137 159. In Benson, W.W., K.S. Brown, and L.E. Gilbert. 1975. Coevolution of plants and herbivores: passion flower butterflies. Evolution 29 (4): 659 680. B URKHOLDER P. 2008. Preferential oviposition by Heliconiinae (Nymphalidae) butterflies on Passiflora biflora (Passifloraceae) leaves with hig her cyanide concentrations. In: CIEE

PAGE 12

Spring 2008: Tropical Ecology and Conservation. CIEE. Council on International Educational Exchange, Monteverde, Costa Rica, pp. 188 199. C ORRALES J ORGE F 1996. Las mariposas Heliconius de Costa Rica. Instituto Naci onal de Biodiversidad. Heredia, Costa Rica. p. 20. D E V RIES P.J. 1987 The butterflies of Costa Rica and their natural history. Princeton University Press, Princeton, New Jersey. pp. 186 195. E NGLER H.S., K.C. S PENCER AND L.E. G ILBERT 2000. Insect meta bolism: preventing cyanide release from leaves. Nature. 406: 144 145. F UTUYMA D.J., AND M. S LATKIN 1983. Introduction. In D.J. Futuyma and M. Slatkin (Eds.). Coevolution, p. 1. Sinauer Associates, Inc., Sunderland, Massachusetts. G IBBS R.D 1974. Chemot axonomy of flowering plants. McGill Queen's University Press, London. In Benson, W.W., K.S. Brown, and L.E. Gilbert. 1975. Coevolution of plants and herbivores: passion flower butterflies. Evolution 29 (4): 659 680. G ILBERT L.E 1971. Butterfly plant coev olution: Has Passiflora adenopoda won the selection race with Heliconiine butterflies? Science. 172 (3983): 585 586. G ILBERT L.E 1975. Ecological consequences of a coevolved mechanism between butterflies and plants. In L.E. Gilbert and P.H. Raven (Eds.) Coevolution of animals and plants. p. 210 240. University of Texas Press, Austin, Texas. G ILBERT L.E. AND M.C. S INGER 1975. Butterfly Ecology. Annual review of ecology and systematics. 6: 365 397. G ILBERT L.E 1983. Coevolution and mimicry. Coevoluti on. In D.J. Futuyma and M. Slatkin (Eds.). p. 263 281. Sinauer Associates, Inc., Sutherland, Massachusetts. S MILEY J 1987. Plant Chemistry and the Evolution of Host Specificity: New Evidence from Heliconius and Passiflora Science. 201(4357): 745 747. T URNER J.R.G 1981. Adaptation and evolution in Heliconius : A defense of neo Darwinism. Annual review ecological systems 12: 99 121 W ILLIAMS K.S. AND L.E. G ILBERT 1981. Insects as selective agents on plant vegetative morphology: Egg mimicry reduces egg la ying by butterflies. Science 212(4493): 467 469.