1The role of cyanide in oviposition by Heliconius (Nymphalidae: Heliconiinae) on Passiflora (Passifloraceae) Zach Sheff Department of Biology, Indiana University ABSTRACT Vines of the Passifloraceae contain cyanide that deters most potential herbivores but not Heliconius spp. (Nymphalidae: Heliconiinae). Desp ite their ability to cope with cyanide in Passifloraceae, Heliconius spp may preferentially oviposit on leaves with lower cyanid e concentrations, as young larval stages may still be subject to toxicity. Leaf samples both with and without eggs were collected from Passiflora spp. from the Monteverde Butterfly Garden and were assayed for cy anide concentration. Th ere was no significant relationship between cyanide concentration and ov iposition (mean concentration with eggs = 6.04 + 9.41 (g/mg), mean con centration = 11.13 + 18.84 (g/mg); t = 1.15, p = 0.26, df = 42). Thus, ovipositing Heliconius are either indiscriminate in their choice of leaves or are selecting leaves based upon other traits that are more important to determining egg survivorship or early larval development. RESUMEN Los bejucos en Passifloraceae contiene n cianuro que ahuyenta a la mayor a de los herbvoros potenciales pero no a Heliconius spp. (Nymphalidae: Heliconiinae). A pesar de su habilidad para enfrentar cianuro en Passifloraceae, Heliconius spp. puede de manera preferencial ovi positar en hojas con concentraciones ms bajas de cianuro; como jvenes etapas larvales pued en ser todava susceptibles a la toxicidad. Hojas con y sin huevos fueron recolectadas de Passiflora spp.. en el Jardn de Mariposas de Monteverde y fueron probadas para determinar la concentracin del ci anuro. No hubo relacin significativa entre la concentracin de cianuro y oviposicion (la concentracin media con huevos = 6,04 + 9,41 (g/mg), la concentracin media sin huevos = 11,13 + 18,84 (g/mg)). As, Heliconius no discrimina en su eleccin de hojas o escoge las hojas basadas en otros rasgos que son ms importantes a la supervivencia del huevo o el desarrollo temprano larval. INTRODUCTION Heliconius butterflies (Nymphalidae: Heliconiinae) range from the southern United States through South America, reaching th eir highest density in the Amazon Basin (DeVries 1987). Eggs are laid exclusively on vines in the fa mily Passifloraceae, its larval host plant, commonly on leaves or meristems near the tip of a shoot (Benson et al. 1975; Williams and Gilbert 1981). Almost all ge neralist herbivores are deterred from Passifloraceae because it contains noxious cyanogenic glycosides (Benson et al. 1975). The specialization by Heliconius larvae on Passifloraceae, due to their ability to cope with the increased cyanide c oncentration, has resulted in a series of coevolutionary chemical, physical, and behavi oral adaptations in both Heliconius spp and Passifloracae (Gilbert 1971; Benson et al 1975; Smiley 1978; Williams and Gilbert 1981). The plant
2 has evolved more specific defenses against heliconiine oviposition to reduce herbivory: glands on leaves that mimic eggs, extraflora l nectaries to attract predacious ants, and irregular leaf shape to disgui se the plant; to whic h the butterflies have responded in turn (Gilbert 1971; Benson et al. 1975). As a result, there exists a coevolution between these two organisms. More specifically I am i nvestigating whether the female shows a preference for leaves with lower cyanide con centrations. The toxic effect states that compounds such as cyanogenic glycosides dete r herbivory primarily through direct harm to organisms which ingest the compounds (Adsersen and Adsersen 1993). However, under this model the time span between the consumption of the toxin and feeling its negative effects is too long fo r the herbivore to learn to avoid the plant, assuming it survives. Alternatively, Adsersen and Adse rsen (1993) also propose that the toxic effect could function in concert with a repe llent effect which deters animals from even initiating herbivory, thus reducing damage to both organisms and allowing for coevolution and, ultimately, coexistence. The latter seems more appropriate for the instance of Heliconius and Passifloraceae because the Heliconius have chemoreceptors positioned on their forelegs and often inspect potential ovipos ition sites thoroughly before actually laying any eggs (Benson et al. 1975; Williams and Gilbert 1981). Females may select a specific leaf for oviposition based on several different criteria (Freitas et al. 1999). These include, looking fo r the presence of other eggs or egg-mimic glands, testing the presence of cyanide, choosi ng leaves that are younger and softer, or choosing plants that ar e growing more vigorously (Freitas et al. 1999). However, the choice to oviposit on a leaf must also account for circumstantial factors such as the presence of competition or para sites. If there is higher competition then Heliconius may choose leaves with more cyanide to avoid other herbivores; if there are parasites present then Heliconius may choose leaves with more cyanide to avoid parasites that are unable to withstand la rge amounts of cyanide. On the other hand, larvae may be particularly vulnerable to the effects of cyanide, which would lead to oviposition on leaves with lower cyanide concentration. Furt hermore, the amount of cyanide in a leaf is regulated by several factors including leaf age (Copp and Davenport 1978; Futuyma 1983; Woodman and Fernandes 1991). Clearly, the relationship between Heliconius and Passifloraceae is a complex one deserving a closer look. I expect that, due to the vulnerability of Heliconius larvae, the butterflies will oviposit more frequently on leaves that are younger and have a lower cyanide concentration. The younger leaves should be easier for the larvae to eat because they have not yet toughened. Additionally, lower cyanide content is not as strenuous on the newly hatched larvae. The purpose of this st udy is to determine what, if any, connection exists between the cyanide concen tration of leaves and oviposition. METHODS Study Site I performed this experiment at the Montever de Butterfly Garden in Monteverde, Costa Rica. Data collection was carried out from October 25th to November 14th during the fall of 2007. Leaf samples were coll ected from several vines in Heliconius garden and one vine in the Large Garden. The species of Passiflraceae used was Passiflora biflora The
3 species of Heliconius present in the gardens were H. sapho, H. melpomene H. sara, and H. doris Leaf Collection and Cyanide Determination I selected an equal number of leaves with eggs and without. Leaves were gathered from two individuals of Passiflora biflora in the Heliconius garden and one individual in the Large Garden Eggs were all from the genus Heliconius, but were not identified to species. All leaves were of similar age; none had toughened beyond the initial stage of leaf development and they were still unfol ding or had just unfolded. Leaves that exhibited herbivory were excluded. Leaves we re selected from new growth shoots of the plant within 15 cm of the tip. I weighed each leaf sample, composed of the entire sample leaf, and then placed it in separate vials, wh ich were approximately five mL in size, to determine the concentration of cyanide. I cr ushed the leaf samples inside the vials and then added three drops of toluene. After this a strip of Watman No. 1 filter paper, one cm by five cm, was soaked in sodium picrate solution by dipping the paper into a small amount, approximately three mL, poured into a Petri dish. Once the filter paper had been dipped in the solution, it was gently passed agai nst the edge of the Petri dish to remove any excess sodium picrate. The saturated paper was then suspended five mm above the leaf sample in each vial and secured with the lid. Then each sample was left in a dry box at approximately 33 C for two hours. The filter paper begins yellow and in the presence of cyanide turns purple. It was removed and soaked in five mL of distilled water for 30 seconds. The samples were placed in a Sequoia-Turner spectrophotometer, which had been zeroed using a cuvette cont aining five mL of distilled water that had had a strip of filter paper saturated with sodium picratebut not exposed to cyanidesoaked in it for 30 seconds, and the transmittance was recorded at a wavelength of 540 nm. Once all of the samples had been analyzed the concentr ation of cyanide was determined using a standard transmittance curve obtained from a previous student (Morris 2005); however, I also created my own standard curve (see Re sults). The concentration is given in micrograms of cyanide per mg of leaf. Standard Curve The standard curve for cyanide concentra tion was created usi ng a serial dilution technique. I m ade an ini tial concentration by adding 2000 mg of potassium cyanide to two mL of distilled water. This created a concentration of 1000 mg/mL. A serial dilution was performed by using a micropipette to put 0.2 mL of the original solution into 1.8 mL of distilled water to reduce each successive concentration by a factor of ten. In this manner concentrations of 1000 mg/mL, 100 mg/mL, ten mg/mL, one mg/mL, 0.1 mg/mL, and 0.01 mg/mL were created. On ce each concentration had been made approximately 100 L of each was placed in th e same vials as were used for the leaf samples along with three drops of toluene. Then one cm wide by six cm long strips of Watman No. 1 filter paper were soaked in s odium picrate and secured five mm above the cyanide solution following the same procedur e from above. The samples were then placed in a dry box at approximately 33 C for two hours as before. After removing the samples from the dry box, the strips of filter paper were each soaked in five mL of
4 distilled water for 30 seconds. The spectrophotometer was zeroed, using the same technique as above, and set to a wavelength of 540 nm. The transmittance for each five mL sample was recorded. Once the transmittances were recorded they were entered into an Excel spreadsheet with the corresponding co ncentrations and a curve was constructed plotting transmittance against concentration. RESULTS No trend between leaf cyanide concentr ation and oviposition was found in this experiment. The mean cyanide concentration for leaves that had e ggs present was 6.04 + 9.41 (g/mg), and the mean concentration for leaves that did not have eggs was 11.13 + 18.84 (g/mg). While the mean cyanide concentration for leaves on which Heliconius did oviposit was lower than for those which did not contain eggs, the standard deviations were too large to extrapolate a reliable trend (t -test, t = 1.15, p = 0.26, df = 42; Fig. 1). 0 5 10 15 20 25 30 35 noyes Eggs Present FIGURE 1. Mean cyanide concentration (g/mg) for Passiflora leaves with and leaves without Heliconius eggs. Standard deviation bars ar e shown. Mean value for leaves without eggs is 11.13 + 18.84 (g/mg), and mean value fo r leaves with eggs is 6.04 + 9.41 (g/mg). The total number of samples collected was 44, consisting of 23 samples containing Heliconius eggs and 21 samples that did not. However, when an analysis of covari ance was performed there was statistical significance for the entire model (ANCOVA, model: F = 2.95, p = 0.04, df = 43). This was due to a significant trend for the covariate, leaf weight, and for the combined effects of cyanide concentration and leaf weight (ANCOVA, leaf weight: t = 2.09, p = 0.04, n = 44; combined: t = 1.98, p = 0.05, n = 44). The relationship derived from the combined effect of cyanide concentration and weight was that leaves with a higher cyanide concentration and larger size were less likely to contain eggs (Regression, R2 = 0.53, p =
5 0.0002, n = 21; Fig. 2), while no reliable trend was present for leaves that did contain eggs (Regression, R2 < 0.0001, p = 0.9047, n = 23; Fig. 3). Even though there was generally little variation in cyanide content, leaves that were heavier, and by inference older, showed a higher concentration of cy anide than lighter, younger leaves (Regression, R2 = 0.16, p = 0.0068, n = 44; Fig. 3). Additional Results The standard curve that I constructed was not used to analyze any of the data because I did not have high confidence in its accuracy. This is due to inadequate equipm ent in the lab to accurately create the small meas urements needed in the serial dilution. 0 10 20 30 40 50 60 70 80 00.050.10.150.20.25 Weight (g) With Eggs Without Eggs FIGURE 2. Weight (g) is plotted against cy anide concentration (g/mg) and trendlines are compared from linear regressions fo r leaves with eggs (y = -1.2296x + 6.1007, R2 < 0.0001, p = 0.9047, n = 23) and leaves w ithout eggs (y = 331.24x 9.8628, R2 = 0.53, p = 0.0002, n = 21).
6 0 10 20 30 40 50 60 70 80 00.050.10.150.20.25 Leaf Weight (g) FIGURE 3. A linear regression of all Passiflora leaf samples compar ing leaf weight (g) to cyanide concentration (g/mg). (y = 127.21x + 1.1567, R2 = 0.1615, p = 0.0068, n = 44) DISCUSSION My initial expectation that Heliconius butterflies would preferenti ally oviposit on leaves with lower cyanide concentration and on younge r leaves is partially supported by the results from this experiment. There is no si gnificant relationship be tween cyanide levels of leaves that contain eggs and those that do not. While the mean cyanide concentration for leaves lacking eggs is higher than the mean for leaves with eggs (11.13 + 18.84 g/mg compared to 6.04 + 9.41 g/mg), the standard deviation is much too large to extrapolate a reliable trend (F ig. 1). However, there was a trend for leaves that did not contain eggs to generally have higher cyanid e concentration and highe r weight (Fig. 2). The reason that this trend di d not correspond to a trend of leaves containing eggs having less cyanide and lower weight could be that the correlation came primarily from one parameter, but not both. This means that the weight, or age, of the leaf could be more important to the butterfly than the cyanide concentration. A lternatively, this could mean that Heliconius larvae can cope with tougher leaves or higher cyanide content, but suffer when presented with both de terrents simultaneously. On the other hand, there is evidence that other adaptations by both Passiflora and Heliconius could be more important than the chemical defense afforded by cyanogenic glycosides (Gilbert 1971; Benson et al. 1975; Smiley 1978; Williams and Gilbert 1981). Gilbert (1971), citing the eff ectiveness of hook-shaped tric homes at trapping and killing Heliconius larvae, suggested that physical adap tations by Passifloraceae represent a stronger deterrent for herbivory than its se condary compounds. Past experiments have shown that Heliconius have keen vision facilitating a strong image recognition system for locating potential leaves for ovi position. Not only has this pr ecipitated further evolution in Heliconius possessing the most advanced nervous system of any genus in Lepidoptera,
7 but it has also encouraged further adaptation in Passifloraceae, variable leaf shapes (Benson et al. 1975). There is also evid ence that leaf age contribut es largely to choice of oviposition site (Benson et al. 1975; Copp and Dave nport 1978; Freitas et al 1999). Taking into account the possibility that cyan ide defense is not as effective as other defenses utilized by Passiflora and the fact that st udies have shown that Heliconius oviposition is influenced by leaf age, it seems plausible that the cyanide content alone is not a determining factor for ovi position choice. Leaf age c ould have more implications to Heliconius larvae than just the toughness of the leaf. The physical defenses described by Gilbert (1971) may not be fully effective until the leaf has had sufficient time to develop. This may provide an even str onger advantage to ovipositing on young leaves. In Passifloraceae, younger leaves contain a greater amount of cyanide than older ones, possibly because they are of higher value to the plant than older ones and are particularly prone to herbivory because they have not yet toughened (Futuyma 1983; Woodman and Fernandes 1991). However, the leaves in my study exhibit the opposite trend, actually becoming more toxic as they gr ow older (Fig. 3). This could be due to the fact that all of the leaves included in my study were of a young age, some weighing as little as 0.005 g, and the smallest had not been given enough time to become maximally toxic. In these extremely young leaves that ge nerally show low cyanid e content, the most critical criteria could be t oughness. Ultimately, the role of cyanide is complex as are the interactions between Passifloraceae and Heliconius. A butterflys choice of oviposition site seems to be governed more by physical ch aracteristics of the le af rather than the chemical composition of that leaf. However, more research into this area needs to be done to ascertain the specific role(s ) of cyanide in th is interaction. Future Studies Future studies in this area should preferably include Passiflora spp. bred for higher and lower cyan ide content to ensure a wider variation in cyanide concentration from which to draw trends. This strategy might provide greater insight into the role the cyanide plays in interactions with Heliconius by magnifying the natural grad ient of cyanide experienced by the butterflies My study suffered from a low varia tion in cyanide content that did not allow me to draw any concrete conclusions a bout the role of cyanid e in oviposition, and I feel that a wider range of concentrations would allow for more defini tive answers in this area. I would also recommend conducting a study that examines more closely the connection between leaf age, cyanide concentration, and frequency of oviposition to determine if cyanide or leaf age is truly th e more integral component. This study could be conducted by using two groups of Passiflora : one which contains plants of low cyanide variation and one which contains plan ts of higher cyanide variation. Both groups would be offered to Heliconius, separately and in isolati on, and this would allow for a separation of whether Heliconius prefer leaves with more or less cyanide, and leaves that are tougher. ACKNOWLEDGMENTS
8 Many thanks to Alan Masters for helping me devise and execute an interesting and dynamic project. I also thank Karen Masters for her statistical genius and ability to turn one figure into thousands, as well as, the st aff of the Monteverde Butterfly Garden, especially Zach and Eli for answering my overly specific and often superfluous questions. Finally, thanks to Taegan Mc Mahon for proofreading my paper at midnight and Pablo Allen for helping me with the statistics. LITERATURE CITED Adsersen, Anne, Henning Adsersen. 1993. Cyanogenic Plants in the Galapagos Islands: Ecological and Evolutionary Aspects. Oikos. 67(3): 511-520. Benson, Woodruff W., Keith S. Brown, Lawrence E. Gilb ert. 1975. Coevolution of Plants and Herbivores: Passion Flower Butterflies. Evolution. 29(4): 659-680. Copp, Newton H., Demorest Davenport. 1978. Agraulis and Passiflora I. Control of Specificity. Biological Bulletin. 155(1): 98-112. DeVries, Phillip J. 1987. The Butterflies of Costa Rica and Their Natural History. Princeton University Press, Princeton, New Jersey, USA. Ehrlich, Paul R., Peter H. Raven. 1964. Butterflies an d Plants: A Study of Coevolution. Evolution. 18(4): 586-608. Fraenkel, Gottfried S. 1959. The Raison dtre of Secondary Plant Substances. Science. 129(3361): 1466-1470. Freitas, Andre V.L., Inara R. Leal, Sirayama O. Ferreira. 1999. Selection of Oviposition Sites by a Lepidopteran Community of a Trophic Forest in Southeastern Brazil. Biotropica. 31(2): 372-375. Futuyma, Douglas J. 1983. Evolutionary Inter actions Among Herbivorous Insects and Plants. In: Coevolution, Douglas J. Futuyma and Montgomery Slatkin, ed. Sinauer Associates, Inc., Sunderland, MA, pp. 207-231. Gilbert, Lawrence E. 1971. Bu tterfly-Plant Coevolution: Has Passiflora adenopoda Won the Selection Race with Heliconiine Butterflies? Science. 172(3983): 585-586. Morris, Melissa. 2005 Herbivore-Induced Defenses in Passiflora biflora CIEE, Spring 2005. Smiley, John. 1978. Plant Chemistry and the Evolution of Host Specificity: New Evidence from Heliconius and Passiflora Science. 201(4357): 745-747. Williams, Kathy S., Lawrence E. Gilbert. 1981. Insects as Selective Agents on Plant Vegetative Morphology: Egg Mimicry Reduces Egg Laying by Butterflies. Science. 212(4493): 467-469. Woodman, Robert L., G. Wilson Fernandes. 1991. Differential Mechanical Defense: Herbivory, Evapotransporation, and Leaf Hairs. Oikos. 60(1): 11-19.
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 M39-00155
El papel de cianuro en la oviposicin de Heliconius (Nymphalidae: Heliconiinae) sobre Passiflora (Passifloraceae)
The role of cyanide in oviposition by Heliconius (Nymphalidae: Heliconiinae) on Passiflora (Passifloraceae)
Vines of the Passifloraceae contain cyanide that deters most potential herbivores but not Heliconius spp.
(Nymphalidae: Heliconiinae). Despite their ability to cope with cyanide in Passifloraceae, Heliconius spp
may preferentially oviposit on leaves with lower cyanide concentrations, as young larval stages may still be
subject to toxicity. Leaf samples both with and without eggs were collected from Passiflora spp. from the
Monteverde Butterfly Garden and were assayed for cyanide concentration. There was no significant
relationship between cyanide concentration and oviposition (mean concentration with eggs = 6.04 + 9.41
(g/mg), mean concentration = 11.13 + 18.84 (g/mg); t = 1.15, p = 0.26, df = 42). Thus, ovipositing
Heliconius are either indiscriminate in their choice of leaves or are selecting leaves based upon other traits that are more important to determining egg survivorship or early larval development.
Los bejucos en Passifloraceae contienen cianuro que ahuyenta a la mayora de los herbvoros potenciales, pero no a Heliconius spp (Nymphalidae: Heliconiinae). A pesar de su habilidad para enfrentar al cianuro en Passifloraceae, Heliconius spp puede de manera preferencial poner huevos en las hojas con concentraciones mas bajas de cianuro; como las etapas larvales jvenes pueden ser todava susceptibles a la toxicidad. Se recolectaron muestras de hojas con y sin huevos de Passiflora spp del jardn de Mariposas en Monteverde y fueron analizadas para la concentracin de cianuro.
Text in English.
Tropical Ecology 2007
Ecologa Tropical 2007
t Monteverde Institute : Tropical Ecology