Flower Choice in Ithomiinae James 1 How do Ithomiinae (Lepidoptera, Nymphalidae) Determine Which Flowers to Feed From? Jessie James Department of Ecology and Evolutionary Biology University of California, Santa Cruz EAP Tropical Biology and Conservation Program, Fall 2016 16 December 2016 ABSTRACT Glasswing butterflies, of the subfamily Ithomiinae, feed almost exclusively from plants that produce pyrrolizidine alkaloids as secondary metabolites. They are able to extract the alkaloids from the nectar and use them as pheromones, as well as for chemical defense. The most common of these alkaloid producing plants in the Monteverde zone is Santa Lucia, an Asteraceae. In order to determine what attracts Ithomiinae to these flowers, I presented them with three types of Santa Lucia flowers: unmodi fied fl owers, dried flowers, and red colored flowers. If the butterflies respond to visual cues, I predicted that they would visit the unmodified and dried flowers more frequently. If they are able to sense the contents/presence of the nectar, I predicted that they would visit the unmodified and colored flowers more often. If Ithomiinae recognize both of these factors, I predicted that they would show a preference for visiting the unmodified flowers. After four and a half days of observation, I found that I thomiinae visit unmodified flowers three times as oft en as dried or colored flowers. However, the length of each visit is highly variable, and does not correlate with visiting preference. This supports my prediction that Ithomiinae acknowledge both flower color, and nectar amount when selecting flowers. Omitting outliers, the average durations of the three presented flower types are not statistically significant, meaning that the butterflies visited each flower type for similar amounts of time. Cmo determinan las mariposas Ithominnae (Lepidoptera, Nymphalidae) de cules flores alimentarse? RESUMEN Las mariposas de alas transparentes, en la subfamilia Ithomiinae, comen casi exclusivamente nctar de flores que producen alcaloides pirrolizidnicos como metabolitos secundarios. Las mariposas pueden extraer estos alcaloides y los usan como feromonas y como defensa qumica. La ms comn de estas flores en Monteverde es un Ageratum sp. conocido popularmente como Santa Lucia. Para determinar qu atrae a las mariposas a stas flores, yo les present tres tipos de flores de Santa Lucia: flores sin modificar, flores secas, y flores teidas de color rojo. Si las mariposas responden a seales visuales, yo predec que ellas visitaran las flores sin modificar y fl ores teidas de rojo ms frecuentemente. Si las mariposas pueden sentir el contenido y presencia del nctar, yo predec que ellas preferiran las flores sin modificar y flores secas. Si las mariposas Ithomiinae reconocen ambos factores, yo pens que ellas preferiran solamente las flores sin modificar. Despus cuatro y medio das de observacin, encontr que las mariposas Ithomiinae visitan flores sin modificar tres veces ms que los otros tipos de flores, pero la duracin de cada visita es muy variable, y no tiene correlacin con
Flower Choice in Ithomiinae James 2 preferencia de visita. Esto apoya mi prediccin de que las mariposas Ithomiinae reconocen ambas seales visuales, y presencia del nctar al seleccionar flores. Si no considero algunos datos atpicos extremos que observ, la duraci n promedio de visita a los tres tipos de flores no son significativamente diferentes. Esto significa que las mariposas visitaron cada tipo de flor por periodos de tiempo similares Almost all animals have predators, and as such many have deve loped defenses to protect themselves. There are a variety of defense types each specially evolved to fit the unique needs and pressures of predator prey relationships. One common method of defense is camouflage. Camouflage entails an organism blending in with its surroundings, either by use of the materials around it, or the ap pearance of the animal itself. Animals can also be poisonous, containing chemicals rendering them distasteful or toxic to organisms that attempt to eat them. One type of butterfly commonly known as the Glasswing butterfly, utilizes both of these defenses. These butterflies are in the subfamily Ithomiinae within the family Nymphalidae, and are characterized by their transparent wings (DeVries, 1987). This transparency makes them nearly impossible to see in the forest understory where they reside, camouflaging them from predators. These clear wings are a result of a lack of colored scales on the majority of the wing. Instead, the wing membrane has microscopic nanopillars of random heights and location, which help to reduce reflectivity, allowing the butterflies to be even more difficult to see (Siddique, Gomard & Holscher, 2015). Ithomiinae also have chemical defenses that make them distasteful to predators, and in large amounts can cause liver deterioration (Diaz, 2015). However, unlike most butterf lies, Ithomiinae acquire this chemical defense in their adult form, rather than as larvae from their host plants. They rely on compounds known as pyrrolizidine alkaloids (PAs) that males obtain from the nectar of plants that produce them as secondary metab olites (DeVries, 1987). These plants include many Asteraceae, most Boraginaceae, and some Fabaceae. The most prevalent of these in the Monteverde zone is Santa Lucia, an Ageratum sp. (Asteraceae) which has compound inflorescences composed of light blue di sc flowers (Masters, 2000). Aside from aiding in defense, these alkaloids are also used by males to produce pheromones to attract mates This is important for Ithomiinae as their transparency, while useful for avoiding predators, also makes it difficult fo r them to find each other The PAs are transferred to females through the spermatophore during copulation, providing them with the same chemical defense as the males (DeVries, 1987). This relationship between Ithomiinae and PA producing plants i Given this fact, the male butterflies must have some way of identifying which flowers contain the chemicals. They may have knowledge of which flowers to visit based on visual cues, or could be able to sense the chemica l composition of the nectar. In this study I aim to answer the question, h ow do Ithomiine butterflies select which flowers to feed from? In order to test this, I will offer the butterflies three types of Santa Lucia flowers: unmodified flowers, dried flowe rs, and red colored flowers. If the butterflies are behaving based on visual cues, I believe they will visit the natural and dried flowers more often than the modified flowers. If they can sense the amount of PAs/nectar in flowers I believe they will vis it the natural and colored flowers more frequently than the dried flowers. If they take both of these factors into account, then I believe they will visit unmodified flowers most frequently. I also predict that the butterflies will spend the longest amount of time feeding on the natural and colored flowers, as opposed to the dried flowers, due to lack of nectar. Materials and Methods I conducted my research at the Jardin de Mariposas Monteverde, where they maintain a greenhouse exclusively for Ithomiine butterflies from the 28 November 2016 to the 2 December 2016 I
Flower Choice in Ithomiinae James 3 presented the butterflies with three flower treatments, unmodified flowers, dried flowers, and colored flowers. Unmodified flowers are natural Santa Lucia flowers t hat were not altered in any way. Flowers were dried by placing cut flowers in a beaker of water, and placing that in front of a heater, which was turned on for a total of six hours. Throughout th is process, water was dropped on the flowers every hour, to a llow them to be dried while minimizing the effect on their appearance. However, the drying treatment affected each flower differently, and resulted in variation of color and texture. Colored f lowers were dyed by placing them in a solution of water and red food coloring (3 drops for every 20ml of water ) for 2 days, while the flowers incorporated the color into their tissues. Both types of modified flowers were taped to skewer sticks and placed in two buckets containing Santa Lucia plants with unmodified flow ers. Within the buckets, the ratio of all three flower types was equal, approximately ten of each flower type in each. However, the greenhouse contained many Santa Lucia plants, causing the ratio of unmodified flowers to modified flowers to be much higher overall After placing the buckets containing all three flower types in the greenhouse, I observed butterfly visitation to the experimental flowers from 9:00 to 13:00 for four and a half days a total of 27 hours I classified visitation as a bu tterfly landing on a flower and feeding, for any length of time. I was able to determine if they were feeding by watching their proboscis; when feeding, they unfurl it and repeatedly probe the disc flowers to gather nectar. I recorded the time at which the y landed, which flower type they landed on, duration of feeding, general activity level of all butterflies, morphological characteristics of the butterfly, and any other notable observations. Every two days I replaced the colored and dried f lowers, to freshness. To analyze the visitation of different flower types, I utilized a Chi Squared Test, and to analyze the duration of visits I performed an ANOVA. As I was not allowed to handle butterflies at the garden, I do not know whether those visiting the flowers were male or female, nor can I identify them with full confidence. In addition to conducting this research at the Jardin de Mariposas, I created a small collection of Ithomiinae caught around the Estacin Biolgica M onteverde Ithomiine butterflies were captured in the native garden and frozen. I then pinned them, determinin g the sex based on the presence, or absence of androconial scales, or hair pencils, between the wings. Pinning them also allowed me to identify t heir species based on morphological characteristics such as color and hind wing venation. Results I recorded 30 visits over the course of my four and a half day study period. As stated previously, I was not able to handle the butterflies in the garden, and therefore had to rely on observations and photos to identify them. Not all species within the gre enhouse visited my flowers, however I categorized four species that did It should also be noted that there was not an equal ratio of butterfly types in the greenhouse. I was unable to determine the actual ratio, but it was clear that there was a greater number of Greta morgane oto than any other type.
Flower Choice in Ithomiinae James 4 Figure 1: Proportion of each butterfly species that visited flowers, not representative of all butterflies in greenhouse A) B) C) D) Figure 2: A) Greta morgane oto on a colored flower. B) Mechanitis polymnia isthmia. C) Ithomia heraldica D) Close up of androconial scales of a male I. heraldica 76.60 % 10 % 6.70 % 6.70 % Greta morgane oto Mechanitis polymnia isthmia Ithomia heraldica Unknown Tiger Stripe
Flower Choice in Ithomiinae James 5 164 365 886 0 200 400 600 800 1000 1200 1400 1600 1800 2000 Unmodified Dried Colored Average Duration in Seconds Flower Type Unmodified flowers were visited significantly more often than the dried or colored flower types Figure 3: The number of visits to each flower type was counted, regardless of duration of visit, over the course of the study The unmodified flowers were visited three times as often as the dried and colored flowers, which were visite d an equal amount. (n =30, X 2 =9.6 df=2, p <0.05 ) The duration of visits to colored flowers was significantly higher than the other two flower types However, there are three outliers, two visits of 2493 seconds and 1659 seconds to colored flowers, and a 1284 second visit to dried flowers Figure 4: The average duration of visits was calculated for each flower type. Including outliers, the difference between durations is statistically significant (n=30, F 2 27 =4.62, p=0.0188). When the outliers are removed from duration of visits the colored and dried average s decrease greatly These new values are both reasonable feeding durations for Ithomiinae T he differences 18 6 6 0 2 4 6 8 10 12 14 16 18 20 Unmodified Dried Colored Number of Visits Flower Type
Flower Choice in Ithomiinae James 6 15 1 5 3 0 2 4 6 8 10 12 14 16 Day 2 Day 3 Day 4 Day 5 Number of Visits Day 164 181 292 0 100 200 300 400 500 600 Unmodified Dried Colored Average Duration in Seconds Flower Type amongst all three averages are not statistically significant ; Ithomiinae fed from each flower type for a similar amount of time Figure 5: E xcluding outliers, there is not a significant difference between duration s (n=30, F 2, 27 =0.0392, p=0.9616) Butterfly visits per day declined sharply after the first full, six hour, day of observation. On days one and two, the butterflies visited all three flower types, but for every day after, days three through five, they only visited un modified flowers Figure 6 : The number of visits to all flower types was counted each day of the study. This graph excludes the first day of data collection as it was only three hours long, whereas all other days were six hours long. Butterflies visited flowers most frequently at 10:00. However, the number of visits does not directly represent the activity level of butterflies. At 14:00, for example, butterflies were often active with regards to flying, but never visite d the experimental flowers.
Flower Choice in Ithomiinae James 7 2 13 6 2 1 0 0 2 4 6 8 10 12 14 09:00 10:00 11:00 12:00 13:00 14:00 Number of Visits Time of Day Figure 7 : Number of times experimental flowers were visited each hour throughout the day (only includes data from full days of observation ) Ithomiinae are most active wit h regards to feeding at 10:00. Discussion I predicted that if Ithomiinae visited the unmodified and dried flowers more often than the colored flowers, then they most likely rely on visual cues to determine which flowers to feed from. If they visited the unmodified and colored flowers m ore often than dried flowers, then they might be able to detect differences in alkaloid or nectar concentrations. My data did not support either of these predictions, but rather my prediction that Ithomiinae consider both visual cues and presence of nectar when selecting flowers. Visitation Ithomiinae visited the unmodified flowers three times as often as the dried or colored flowers. The statistically significant difference between visitation s of the three flower types indicates that they do exhibit a pref erence for unmodified flowers, conceivably as a result of their appearance and amount of nectar. The fact that they visited the colored and dried flowers equally, suggests that the red color ce of which flowers to feed from. Nymphalidae butterflies have true color vision, meaning they can distinguish between different wavelengths of light. Which colors each species can discriminate between varies, but all have red green color vision (Sison M agnus & Briscoe, 2009). This means that Ithomiinae can recognize red, but still preferred to visit the unmodified light blue flowers, indicating that color has an impact on flower selection. The dried flowers, if containing a lower volume of nectar, may have been visited less often due to the butterflies being able to detect this difference Butterflies have chemical receptors in pits on the antennal club of the antennae (DeVries, 1987). These receptors are what allow the butterflies to identify chemicals in the air, such as the pheromones males use to attract mates and possibly the pyrrolizidine alkaloids within the nectar of the Santa Lucia However, the butterflies may also be responding to the change in appearance (either color or structure) of the dried flowers. Therefore, the inferred lack of
Flower Choice in Ithomiinae James 8 nectar in the dried flowers, in addition to visual stimuli are important factors when it comes to butterfly visitation. Duration The d ifferences b etween the duration of visits to the three flower types are significant wh en including the three outliers With the outliers, colored flowers have the highest average duration, 886 seconds. These unusual values, combined with a small sample size, are respo nsible for the very high averages of those flower types. The outliers from visits to colored flowers could be explained by the fact that the dying process took two days, and in that time (if the flower was still producing nectar) no nectar was depleted. Th erefore, the colored flowers may have had a higher volume of nectar. Another explanation could be that flowers with high amounts of nectar were picked by chance. Conversely, the flowers may have had a lower amount of nectar, and the butterflies were simply spending a prolonged amount of time searching for it. The dried flower outlier could be explained by the flower having an abnormally high amount of nectar when picked, or that the heater was ineffective in drying out the nectar. If the latter is true, and drying the flowers only altered the flowers appearance, then the unmodified flowers is a result of that visual change, and not a low amount of nectar. However, excluding the outliers, no flower type was visited significantly mo re than another. All three flower types had very large standard deviations with regards to duration which can be attributed to the small sample size, and outliers. Additionally, Ithomiine butterflies vary considerably in the amount of time they spend on each flower; the longest visit was 2493 seconds, and the shortest was 8 Furthermore, the rate of nectar intake depends on many factors. These include morphology of the butterfly, such as proboscis size, muscle strength, and sex (males feed faster than fem ales), as well as physical aspects of the nectar, such as sugar content and viscosity. These variables are not stagnant, they can be affected by things such as age of the butterfly, or weather. My data may have been affected by temperature, as higher tempe ratures increase viscosity and cause butterflies to be more active, while lower temperatures decrease viscosity and activity (Birke 1988). Number of Visits per Day The total number of visits was counted for every day of data collection, and the data from every full, six h our day was compiled in Figure 6 The number of visits dropped drastically from day two to day three, and stayed low for the rest of the study. The first day of data collection was only a half day (three hours, from 12:00 15:00) but in t hat time, six butterflies visited the flowers. This suggests that if I had been observing for a full six hours, the number would be comparable to that of day two. This downward trend can be explained by the hypothesis that the Ithomiinae learned that some of the flowers did not contain nectar, or did not resemble the flowers the y prefer, and therefore visited less frequently. This possibility is even more likely due to the fact that I kept the buckets containing the flowers in the greenhouse overnight, thus giving the Ithomiinae additional time to learn. Further supporting this hypothesis is the fact that after day two, all visits observed were made to unmodified flowers. Number of Visits by Time of Day The butterflies were most active at 10:00, with low activity levels with regards to feeding throughout the rest of the day. Butterflies can only fly when their body temperature is between 28 and 40 degrees celsius (Birke 1988). Since butte rflies are ectothermic, they rely on the surrounding air temperature to warm them. At 10:00, it is warm enough for them to start flying and feeding, but not so warm that they are in danger of overheating, as it is in the afternoon. This is consistent with my observations of Ithomiinae in the field, where they were most active between the hours of 10:00 and 12:00. Sources of Error These predictions and results are based on the assumption that drying the flowers changed the nectar concentration without drast ically altering their appearance, which may not be true. The Santa Lucia flowers had such low concentrations of nectar that it was impossible for me to measure if any was
Flower Choice in Ithomiinae James 9 removed by drying them. In addition to this, the color and texture of the flowers wer e modified. Some flowers remained a light blue while others developed a light brown tinge, and some stayed soft while others became slightly brittle. This change in appearance could account for the fewer number of visits to dried flowers. I also assumed that coloring the flowers would not have an effect on the nectar, but it is that must be considered is that the modified flowers were cut from their ori ginal plants and attached to sticks. Both the general appearance of these modified flower s and the fact that it was detached from a plant, therefore not actively producing nectar, may have affected visitation. Acknowledgements: I would like to thank Fede rico Chinchilla, my primary advisor, for having endless patience and minimizing my suffering throughout this project. I would also like to thank my secondary advisor Emiliana Triana, especially for her help with insect pinning. The Jardin de Mariposas Mont everde was incredibly gracious in allowing me to conduct my research in their facility, it was the perfect environment for my study and I am extremely grateful to Brinna and all other staff for their help. Thank you to the Estacin Biolgica for providing a research and study center during my project. Furthermore, I better host family. Lastly, thank you to all EAP staff and students for creating such a supp ortive and welcoming environment, I have thoroughly enjoyed every moment of this program. Literature Cited: Birke, L. 1988 Butterflies Make the Best Use of Sunshine. New Science 31. Boggs, C. L. 1988 Rates of Nectar Feeding in Butterfl ies: Effects of Sex, Age, Size, and Nectar Concentration. JSTOR DeVries, P. 1987 The Butterflies of Costa Rica and Their Natural History. Princeton, New Jersey: Princeton University Press. Di az, G. J. 2015 Toxicosis by Plant Alkaloids in Humans and Animals in Colombia National Center for Biotechnology Information Masters, A R. 2000 Transparent Butterflies. Monteverde. Siddique, R., Gomard, G., & Holscher, H. 2015 The Role of Random Nanostructures for the Omnidirectional Anti Reflection Properties of the Glasswing But terfly. Nature Sison Magnus M. P., & Briscoe, A D. 2009 Molecular and Physiological Innovations of Butterfly Eyes (PhD thesis). U niversity of California Irvi ne, Irvine, California.