1 Associative color learning and age in Heliconius butterflies Emily Hollenbeck Department of Biology, University of Pennsylvania ABSTRACT Flower c olor is an important attractant and ma ny pollinators show distinct color preference s . This study determine s if color preference and its lability differ between experienced and naÃ¯ve Heliconius but terflies. Butterflies were offered nectar in Lantana camara inflorescences, which naturally have flowers of two colors (yellow and red) on a single inflorescence. Butt erflies visited yellow flowers more frequently, with 76% of visits to yellow flowers when both flower colors offered nectar. When nectar was offered only by red flowers, yellow preference decreased significantly over time. Newly eclosed butterflies offered inflorescences where only red flowers rewarded showed a 56% yellow visitation rate over two days, 22% less than for butterflies previously offered inflorescences where both flowers rewarded . These results suggest that yellow preference in Heliconius is bo th strong and innate , but can be weakened by experience. Further, it appears that naÃ¯ve butterflies are more labile, allowing them to track changes in rewards more quickly. This reward associated learning may help optimize foraging success by enabling indi viduals to adapt behaviorally to different environment s . RESUMEN El color de las flores es un atractivo importante y varios polinizadores muestran distintas preferencias. Este estudio determina si la preferencia por color y su labilidad difieren entre mar iposas con experiencia e ingenuas del gÃ©nero Heliconius . Se les ofreciÃ³ nÃ©ctar a las mariposas en inflorescencias de Lantana camara , que naturalmente tiene flores de dos colores (rojo y amarillo) en una misma inflorescencia. Las mariposas visitan flores amarillas con mayor frecuencia, con un 76% de visitas a flores amarillas cuando ambos colores fueron ofrecidos con nÃ©ctar. Cuando e l nÃ©ctar se ofreciÃ³ solo en las flores rojas, la preferencia por las flores amarillas decreciÃ³ significativamente con el tiempo. Mariposas reciÃ©n eclosionadas a las que se les ofreciÃ³ nÃ©ctar solamente en flores rojas mostraron un 56% de visitaciÃ³n en dos dÃas, 22% menos que las mariposas a las que se les ofreciÃ³ con recompensa en ambos colores. Estos resultados sugieren que la preferencia por el color amarillo en Heliconius es tanto fuerte como innata, pero puede ser disminuida por la experiencia. AdemÃ¡s , parece ser que las mariposas ingenuas son mÃ¡s lÃ¡biles, permitiÃ©ndoles determinar los cambios en la recompensa mÃ¡s rÃ¡pido. Este comportamiento asociado a la recompensa puede ayudar a optimizar el Ã©xito de forrajeo permitiÃ©ndole a los individuos adaptarse a diferentes ambientes. INTRODUCTION Animal mediated pollination occurs when plants reward animals to move pol len to conspecific stigmas (Faegri & Van der Pijl 1971, Proctor and Yeo 1973; both in Keister et al. 1984 ) . To improve conspecific transfer of po llen, many plants have evolved specific morphologies, phonologies and rewards to attract a specific grou p of potential pollinators (Kiester et al . 1984, Aigner 2006 ). Further, animals have coevolved with plants to be more effect ive at extracting rewards , causing them to be restricted in the flowers they visit (Waser et al. 1996 ). Such specialization reduces pollen waste and stigmatic fouling as pollinators are more likely to move between flowers of the same species, and reduces time and energy expenditure by pollinators in their se arch for rewarding flowers (Keister et al. 1984 ). Thus, floral adaptations
2 such as size, shape, color, and presentation of reward and pollen help ensure maximum pollination success and pollinator foraging efficiency (Kevan & Baker 1983). Color is an important floral characteristic that can selec t for certain pollinators (Kevan & Baker 1983) and is an important component of pollinator syndromes: suites of traits that assure a certain subset of potential pollinators (Bawa 1990 ). Many butterflies, for example, prefer bright colors , such as yellow s (Kevan & Baker 1983) while most hummingbirds preferred ( Schatz 1990, Aigner 2006 ) . Thus, pollinators can mo re quickly recognize flow ers that allow them to extract appropriate rewards more easily ( Kevan & Baker 1983, Weiss 1995) . An overly narrow use of color cues may constrain the pollinator, however, as they may miss reward opportunities (Waser et al. 1996). In the case of floral color, it may be advantageous for pollinators to sample and learn which colors are currently most rewarding, rather than be constrained to innate color preferences that would overly limit flower choice. Some butterflies are able to learn which flowers are most rewarding by their floral color ( Swihart & Swihart 1970 , Swihart 1971, Weiss 1991). While butterflies may have innate color preferences, they sometimes switch preferences between flower colors depending on presentation of a consist ent reward. For example, H. charitonius butterflies showed an increased preference for a given flower color if they were raised with rewarding flowers of only that color (Swihart & Swihart 1970, Swihart 1971). Further, another heliconiine, Agraulis vanillae , switched color preference from yellow to orange to avoid toxic nectars (Masters 1991). Butterflies favor yellow flowers on the bicolor inflorescences of Lantana camara ( Swihart & Swihart 1970, Weiss 1991). In L. camara , yellow flowers offer more nectar while red flowers are largely emptied of nectar (Weiss 1991). This study sought to determine whether preference for yellow flowers in H eliconius butterflies is learned or innate, and if it can be reversed throug h altered reward cues . It differs from previous studies ( Masters 1991, Weiss 1991, Weiss 1997) in altering reward and investigating change in preference over time, and with the inclusion of newly eclosed butterflies to determine if there is a naÃ¯ve preference . It was predicted that older butterflies may prefer yellow flowers because they reinforce preference with a better reward. NaÃ¯ve butterflies, on the other hand, would have no reason to select for yellow flowers over red if both offered equal reward, unless there were an innate preference for yellow. All butterflies should eventually learn to favor red flowers if only red flowers contain nectar. MATERIALS AND METHODS Study organisms Heliconius charitonius butterflies were raised in the Monteverde Butterfly Garden , Monteverde, Costa Rica. To feed adult butterflies, the garden fills both yellow and red flowers on L . camara inflorescences with 20% sucrose solution . Therefore, the H. charitonius used for this study were previously exposed to equal rewards in both yellow and red flowers. L. camara natural infloresences naturally have a bull s eye pattern, with older red flowers forming a ring around younger, yellow flowers (Weiss 1991). Mature inflorescences normally contain around 30 flowers. For this study, inflorescence color was controlled by 1. selecting inflorescences t h at were entirely or almost entirely one color , and 2. removing a small number of flowers of the opposite color to create monochromatic inflorescences. These were then offered to butterflies to assess color preferences independently of flower placement or inflorescence siz e.
3 Study site and conditions The study site was the Heliconius room (Room 2) of the Monteverde Butterfly Garden in s in a 5 x 6 array, each approximately 15 cm apart (Figure 1A) . Each station was a small vase of water containing a single L. camara inflorescence of only one color, either yellow or red, with 10 20 flowers (usually around 15) per inflorescence (Figure 1B) . Y ellow/red stations alternated throughout the grid to yield 15 of each color. Stations were mounted on posts 1 1.5 m tall. Nectar was in serted into the flowers with a syringe twice each morning, once at the beginning of the observation interval and once at the end. Most observation intervals fell betw een 9:00am and 11:00am. Figure 1A&B. Grid array of L. camara flowers in the Heliconius room of the Monteverde butterfly garden. A. Full grid of 5 x 6 feeding stations each with one L. camara infloresence, either yellow or red, mounted on poles 1 1.5m tall. Colors alternated throughout the grid. . B. Yellow (left) and red (right) L. camara inflorescences in vases, before placement on poles. Part 1: Experienced Butterflies, Nectar in All Flowers For t he first 13 days (April 11 24 , 2011), all flowers contained nectar. Observations were taken on ten of those days (discounting April 15, 18, 23, and 24) using H. charitonius that were previously released in the garden and had been feeding at least several days , and up to several months , on L. camara with nectar added to all flowers of bicolored inflorescences . Each observation period lasted 30 minutes, and most days contained two with less than ten minutes between periods . Exceptions were 4/16 and 4/11 wit h one observation period , and 4/17 and 4/22 with three . During each period , the total number of visits by H. charitonius butterflies to each color inflorescence was recorded. A visit was defined as a butterfly landing on a n inflorescence , pausing long enough to cease fluttering it s wings, and attempting (regardless of success) to probe the flowers. Part 2: Experienced Butterflies, Nectar Only in Red Flowers For the next seven days (April 25 May 1), nectar was inserted into red flowers onl y. Observations were taken on six of those days (discounti period , three 30 min observation intervals were taken each day, usually consecutive or with less A B
4 than ten minutes between intervals. Again, total number of visits by H. charitonius to each flo wer color was recorded. On April 26 and 27, in addition to visit number, visit duration was recorded for all visits to both flower colors for one 30 min interval each day. Visits were timed from the time a butterfly ceased fluttering its wings after landin g to the moment it flew away. Part 3: NaÃ¯ve Butterflies, Nectar Only in Red Flowers On April 30, 13 newly eclosed H. charitonius were marked on the abdomen and released, and their visits were recorded separately for two days: 4/30 5/1. RESULTS Total number of visits in each 30 minute interval ranged fro m 8 68 with an average of 33 visits. Because individual observation periods differed markedly in total number of visits, rather than comparing number of flower visits, the percent visits to yell ow was calculated for each period. The data reported are the mean percent s yellow, where n = the number of observation periods for that particular treatment and the mean is the average of the percent visits to yellow flowers for all pertinent observation p eriods, regardless of total number of visits per observation period. Part 1: Experienced Butterflies, Nectar in All Flowers When all flowers contained nectar, experienced butterflies displayed a preference for yellow flowers (paired t test, DF = 20, N = 21, t = 6.51, p < 0.0001) with a mean percent of visits to yellow flowers per observation period Â± 1SD of 76.4% Â± 15.9% (Figure 2) . O ne outlying observation period had 25% yellow visitation rate, and t hat interval comprised of only eight visits total , the lowest count ob served. Because of small sample size for this observation period, the unusually low relative visits to yellow flowers may be spurious. If this observation period is ex cluded, the mean percent of visits to yellow flowers per observation peri od is 79.0% Â± 10.9% . Since the trends hold even including the outlier, it will remain included for further discussion. Part 2: Experienced Butterflies, Nectar Only in Red Flowers When only red flowers contained nectar, an overall preference for yellow still existed (paired t test, DF = 17, N = 18, t ratio = 2.42, p = 0.02), but with a lower mean percent yellow visitation rate per observation period of 60.2% Â± 19.3% (Figure 2) . Yellow preference when all flowers contained nectar was 16% higher than when just red flowers contained nectar , and this difference was significant (unpaired t test, DF = 1, F = 8.09, p = 0.007 ; Figure 2 ).
5 Figure 2. Mean proportion of visitation to yellow L. camara inflorescence s , out of an array of equal number s yellow and red, by H. charitonius across all 30 minute observation intervals. N ectar was present in all flowers for 13 days (n = 21) followed by red flowers only for seven days (n = 18) . Yellow was preferred for both treatments, but yellow preference was hi gher with nectar in all flowers . Asterisks over bars represent significant differences. Error bars represent 1 SD. Change in yellow preference over time Yellow preference over the first 11 days with nectar in all flowers showed no significant change over time (Regression test, R 2 = 0.168, N = 9, p = 0.177), whereas over the following seven days with nectar in red flowers only, yellow preference decreased over the week (Regression test, R 2 = 0.853, N =6, p = 0.0005; Figure 3) . Figure 3. Yellow preference of experienced H. charitonius butterflies , measured by percent of visits which occurred to yellow L. camara flowers , remained constant over time while all flowers had nectar for 13 days, with nine days of data collection ( A ) and decreased while only red flowers had nectar for seven days, with six days of data collection ( B ). Equations and R 2 values refer to the trendline for yellow visitation rate over time. 0 20 40 60 80 100 All Red % visits to yellow flowers Nectar conditions y = 0.817x + 80.77 RÂ² = 0.168 20 40 60 80 100 0 5 10 15 % visits to yellow flowers Day of treatment: all nectar y = 7.149x + 87.57 RÂ² = 0.853 20 40 60 80 100 0 2 4 6 8 % visits to yellow flowers Day of treatment: red nectar * ** B A
6 The average yellow preference during the red treatment (Figure 2) is not an accurate representation of butterfly behavior since visitation preference changed significantly over time (Figure 3B). Yellow visitation was 78.2% in the first two days of the red treatment and dropped to 40.6% over the last two days . Visit duration On the first three days experienced butterflies were exposed to the red treatment , which all had higher yellow visitation rates than the mean for that treatment ( 71%, 84%, and 70%, respectively; Figure 3), butterfly behavior was notably different . During two 30 minute intervals on the second and third days of red treatment, experienced butterfly visits to red flowers were significantly longer than yellow visits (Wilcoxon sign rank test, p < 0.0001; Figure 4). They repeatedly visited many yellow flowers in a ro w for a very short period of time (mean 12.3 seconds, Wilcoxon Score Mean 31. 3, n = 57), but remain ed for a long time on red flowers (mean 93.4 seconds, Wilcoxon Score Mean 58.3, n = 17), although with less frequency (Figure 3). Figure 4. Mean visit duration by H. charitonius to yellow and red L. camara flowers during two 30 minute intervals, one each on 4/26 and 4/27, the second and third days of experienced butterfly exposure to red only reward . Asterisk indicates that visits to red flowers were significantly longer. Error b ars represent 1SE. Numbers above bars indicate sample size. Part 3: NaÃ¯ve butterflies, Nectar Only in Red Flowers The naÃ¯ve butterflies released on April 30 were exposed to con ditions of nectar in red flowers only for two days, for a total of six observation intervals. The total number of naÃ¯ve visits during that time was 66, and 37 of those visits to yellow flowers. The overall yellow visitation rate of 56.0% was not a signific ant preference for yellow (Chi square test, 2 = 0.48, p > 0.05) , even though it was within 4% of the mean yellow visitation rate of experienced butterflies during the red treatment , which was significant (figure 5 ) . Howeve r, over only the first two days of the red treatment for experienced butterflies, yellow visitation rate was 78.2% , 0 20 40 60 80 100 120 140 160 Yellow Red visit duration (sec) N = 17 * N = 56
7 different from both the naÃ¯ve yellow preference (Chi square test, 2 = 40.0, p < 0.05 ) and total experienced yellow preference during red treatment (Chi square test, 2 = 22.2, p < 0.05; Figure 5). Figure 5 . Mean proportion of visitation by H. charitonius to yellow L. camara inflorescence s , when only red flowers contained nectar. Yellow preference was not significant for naÃ¯ve butterflies over two days (left bar) and was significant for experienced (Exp.) butterflies across all days of red treatment (middle) and the first two days of experienced butterflies in red treatment (right). Asterisks represent significant differences. Percent above bars is the mean, and number inside bars is the number of 30 min observation periods included in each mean. DISCUSSION Color preference of experienced butterflies Sin ce H. charitonius preferred yellow flowers even when all flower colors were equally rewarding, the preference for yellow is probably innate, an adaptation which optimizes foraging success by favoring flowers which most efficiently reward butterflies. Preference for yellow gradually changed to a preference for red in experienced butterflies when only red flowers rewarded. A lthough the six data points for this second treatment yielded a significant preference for yellow, I believe that given more time under the red treatment the visitation preference would have stabilized below 50% as seen in the final two days. I attribute this to butterflies learning to associate the color red with exclusive reward . The learning process was quantified by the significantly shorter visits to y ellow flowers than red during the initial days of the second treatment, when mean visitation rate still significantly favored yellow. Butterflies were observed moving fairly quickly from one yellow inflorescence to another in search of nectar, and leaving each inflorescence after probing the flowers and finding no reward. When they finally tried a red inflorescence, upon discovering the nectar reward they would sit on the inflorescence for a long time to feed, and then often leave the array without visiting any more red flowers, presumably satisfied. 56.1% 60.3% 78.2% 0 10 20 30 40 50 60 70 80 90 100 NaÃ¯ve: two days Exp: Total Exp: First two days % visits to yellow flowers ** * *** N = 6 N = 18 N = 6
8 An innate color preference can be overridden by reward and associative learning. The ability to learn and shift foraging behavior based on experience also optimizes foraging success, by enabling individuals to more efficiently extract a reward across a variety of environments. This balance between innate and learned preferences demonstrates how coevolution can provide enough specialization to create mutually beneficial interactions whi le allowing generalist flexibility which buffers the ins tability of high specialization (Waser et al. 1996). Color preference of naÃ¯ve butterflies NaÃ¯ve butterflies showed no significant yellow preference even though only red flowers were rewarding. This c ould be explained by a strong innate preference for yellow, mediated by the immediate effects of the opposite reward distribution, or because two days was not long enough for them to learn to favor red, and so before learning they visited the two colors ar bitrarily. An innate yellow pre ference seems most likely because the experienced butterflies also demonstrated a yellow preference even though they were in an environment where reward did not differ between colors. Since naÃ¯ve butterflies were only exposed to red only reward for two days, it makes biological sense to compare their yellow preference with that of experienced butterflies for only the first two days of the experienced red reward trial. The experienced yellow preference was significantly higher than naÃ¯ve for those two days, and the naÃ¯ve butterflies showed no statistical color preference, so I conclude that naÃ¯ve butterflies were able to respond more readily to an alternate reward. If butterflies can learn more quickly when they are young, they demonstrate some phenotypic plasticity which enables them to adjust their behavior to the environment into which they were born. ACKNOWLEDGEMEN TS Muchas gracias to every one at the Monteverde Butterfly Garden, and especially Mauricio Vargas, Alan Vargas, and Julio Garro Alan Masters, who both inspired my interest in Heliconius and advised me throughout. Finally thanks to Roxanne Reiter for invaluable d iscussion and peer review. LITERATURE CITED Aigner, P.A. 2006. The Evolution of Specialized Floral Phenotypes in a Fine grained Pollination Environment. In Plant Pollinator Interactions: From Generalization to Specialization, Was er N.M. and J. Ollerton, (Ed.) . The University of Chicago Press, London. pp. 23 45. Bawa K.S. 1990. Plant pollinator interactions, Sexual systems and pollen flow Commentary. In Reproductive Ecology of Tropical Forest Plants, K.S. Bawa and M. Hadley (Ed.). UNESCO, Paris. pp. 65 68. Brown K.S. Jr. 1981. The Biology of Heliconius and Related Genera. Annual Review of Entomology 26:427 456. Burger W.C. 1981. Why Are There So Many Kinds of Flowering Plants? Bioscience 31:572+577 581. Cardoso, M.Z. 2001. Patterns of Pollen Collection and Flower Visitation by Heliconius Butterflies in Southeastern Mexico. Journal of Tropical Ecology 17:763 768. Ehrlich, P.R. & Gilbert, L.E. 1973. Population Structure and Dynamics of the Tropical Butterfly Heliconius ethilla . Biotropica 5:69 82.
9 Faegri, K. and L. Van der Pijl. 1971. Principles of pollination ecology. Pergamon, Oxford. Gilbert, L.E. 1972. Pollen Feeding and Reproductive Biology of Heliconius Butterflies. PNAS 69 : 1403 1407. Jordano, P. 1987. Patterns of Mutualistic Interactions in Pollination and Seed Dispersal: Connectance, Dependance Asymmetries, and Coevolution. The American Naturalist 129:657 677. Kevan P.G. and H.G. Baker. 1983. Insects as Flower Visitors and Pollinators. Annual Review of Entomology 28:407 453. Kiester A.R., R. Lande, D.W. Schemske. 1983. Models of Coevolution and Speciation in Plants and Their Pollinators. The American Naturalist 124:220 243. Masters A.R. 1991. Dual Role of Pyrrolizidine Alkaloids in Nectar. Journal of Chemical Ecology 17:195 205. Murawski, D.A. 1987. Floral resource variation, pollinator response, and potential pollen flow in Psiguria warscewiczii . Ecology 68:1272 1282. Murawski, D.A. & Gilbert, L.E. 1986. Pollen Flow in Psiguria warscewiczii : A Comparison of Heliconius Butterflies and Hummingbirds. Oecologia 68:161 167. Proctor, M. and P. Yeo. 1973 . The pollination of flowers. Collins, London. Schatz, G.E. 1990. Some aspects of pollination biology in Central American forests. In Reproductive Ecology of Tropical Forest Plants, K.S. Bawa and M. Hadley (Ed.). UNESCO, Paris. pp. 69 84. Swihart C.A. & Swihart S.L. 1970. Colour selection and learned feeding preferences in the butterfly, Heliconius charitonius Linn. Animal Behavior 18:60 64. Swihart C.A. 1971. Colour discrimination by the butterfly Heliconius charitonius linn. Animal Behavior 19:156 164. Tiffney, B. H. 1984. Seed size, dispersal syndromes and the rise of the angiosperms: evidence and hypothesis. Annals of the Missouri Botanica l Garden 71:551 576. Waser N.M., L. Chittka, M.V. Price, N.M. Williams, J. Ollerton. 1996. Generalization in pollination systems, and why it matters. Ecology 77: 1043 1060. Weiss M.R. 1991. Floral colour changes as cues for pollinators. Nature 354:227 229. -------------1995. Floral Color Change: A Widespread Functional Convergence. American Journal of Botany 82:167 185. -------------1997. Innate colour preferences and flexible colour learning in the pipevine swallowtail. Animal Behavior 53:1043 1052.
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 M39-00434
Associative color learning and age in Heliconius butterflies
Flower color is an important attractant and many pollinators show distinct color preferences. This study determines
if color preference and its lability differ between experienced and nave Heliconius butterflies. Butterflies were
offered nectar in Lantana camara inflorescences, which naturally have flowers of two colors (yellow and red) on a
single inflorescence. Butterflies visited yellow flowers more frequently, with 76% of visits to yellow flowers when
both flower colors offered nectar. When nectar was offered only by red flowers, yellow preference decreased
significantly over time. Newly enclosed butterflies, offered inflorescences where only red flowers rewarded, showed a
56% yellow visitation rate over two days, 22% less than for butterflies previously offered inflorescences where both
flowers were rewarded. These results suggest that yellow preference in Heliconius is both strong and innate, but can be
weakened by experience. Further, it appears that nave butterflies are more labile, allowing them to track changes in
rewards more quickly. This reward-associated learning may help optimize foraging success by enabling individuals to adapt behaviorally to different environments.
Text in English.
Butterflies--Feeds and feeding
Costa Rica--Puntarenas--Monteverde Zone--Monteverde
Tropical Ecology Spring 2011
Monteverde Butterfly Garden
t Monteverde Institute : Tropical Ecology