COMPARATIVE ECOLOGY AND MIMETIC RELATIONSHIPS OF ITHOMIINE BUTTERFLIES IN EASTERN ECUADOR By BOYCE ALEXANDER DRUMMOND III A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1976 UNIVERSITY OF FLORIDA 3 1262 08666 406 6 For Nancy, as she lays aside her net awhile to take up the caduceus ACKNOWLEDGMENTS It is my pleasure to thank the members of my committee, Drs. Thomas C. Emmel, Archie Carr, Clifford Johnson, and Thomas Walker, for the guidance and encouragement they have provided throughout my graduate career. I have profited greatly from their respective graduate courses and from the exposure to their divergent, but complementary, approaches to biology. I also thank Drs. John Ewel, Dana Griffin, and Jon Reiskind for helpful discussions and much useful information during the writing of this dissertation. For the countless ways in which they have assisted in all phases of the research reported here, I profess my deepest appreciation to Dr. Thomas Emmel, chairman of my committee, and Nancy Drummond, my wife and field assistant. Without the benefit of their help, many of the goals of this project could not have been accomplished. To Dr. Emmel, who first introduced me to tropical ecology and kindled my interest in the biology of the Lepidoptera, I am indebted for the constant personal, academic, and financial support he so graciously proffered. My wife, Nancy, whose great enthusiasm for our year of field work in Ecuador was matched only by her unflagging patience during the tedious year and a half that followed in Gainesville, assisted in the collection of specimens and population samples, handled most of the life-history rearings, and aided in the preparation and analysis of the data. For all this and much more I am thankfully beholden. 111 I am grateful to Dr. William G. D'Arcy, Missouri Botanical Garden, St. Louis, for identifying the solanaceous plants. At the Allyn Museum of Entomology, Sarasota, Florida, Dr. Lee Miller aided in the identification of several butterfly species and Jacqueline Miller kindly identified the Castniidae. I am indebted to Dr. Lincoln Brower, Amherst College, for loaning me a copy of Christine Papageorgis' dissertation, and to F. Martin Brown, Colorado Springs, Colorado, for loaning me one of the few extant copies of d' Almeida's "Melanges Lepidopteriques." I thank my wife, Nancy,' fbr translating the latter work and other papers from the French. Nancy Drummond also patiently typed and edited early drafts of the manuscript as well as most of the final copy, for which I am extremely grateful. Donna Gillis typed Tables 4 and 5 and the bibliography. The photography staff of the Office of Instructional Resources, University of Florida, prepared Figures 10 through 15. Nancy Drummond kindly drew Figures 33-35 and 46-47 and aided in the preparation of several others. The year of Ecuadorian field work was made possible only by the aid of a number of helpful persons and organizations in the United States and Ecuador. Dan Doyle of the Miami office of the Summer Institute of Linguistics severed yards of red tape at the Ecuadorian embassy to expedite the issuance of the required visas. Giovanna Holbrook of Holbrook Travel and the staff of JAARS (Jungle Aviation and Radio Service of SIL) kindly arranged the various transfers of our over- weight equipment and baggage to, from, and within Ecuador against improbable odds. Ruth and Steve Smith kindly stored our worldly possessions during our fourteen months in Ecuador. The personnel of IV the Ecuadorian Branch of the Summer Institute of Linguistics, especially John Lindskoog, Jonathon Johnson, and Don and Helen Johnson, facilitated our transportation in Ecuador and made available housing and supplies at Limoncocha. I am particularly grateful to Wayne and JoAnne Fitch, Jim and Kathie Yost, and Pat Kelley for many personal favors during the course of this research. Special thanks are due Mark and Phyllis Newell for allowing us to live in their spacious house at Limoncocha during the final three months of our stay. I thank Drs. Howard Weems and Robert Woodruff, Department of Entomology, Division of Plant Industry, Florida Department of Agriculture, for providing equipment and supplies. This research was supported in part by a Grant-in-Aid from Sigma Xi. TABLE OF CONTENTS 4 Page ACKNOWLEDGMENTS iii LIST OF TABLES . . ix LIST OF FIGURES xi ABSTRACT xv CHAPTER I INTRODUCTION 1 Study Area 3 Historical Background 4 Climatic Description 8 Physical Description 20 The Butterfly Community at Limoncocha, Ecuador .... 30 Composition 30 Microhabitat Distribution 42 The Butterfly Subfamily Ithomiinae 45 The Plant Family Solanaceae 59 II METHODS 71 4. Ill ADULT ECOLOGY OF THE ITHOMIINAE 77 Microhabitat Requirements 77 Activity Patterns 81 Perching Behavior 82 Predation and Parasitism 84 Food Resources and Feeding Behavior 87 Locating Food Sources 89 Flower Feeding 90 Coevolution of Ithomiines and White-flowered Plants 110 Ithomiine Attractants "•* vi CHAPTER Page Population Ecology 117 Spatial Heterogeneity 117 The Capture-Mark-Recapture Program at Site 4 121 * IV REPRODUCTIVE BIOLOGY OF THE ITHOMIINAE 148 Courtship 148 Function of Hairpencils in Male Ithomiines .... 150 Mate-locating Behavior of Male Butterflies .... 154 Mate- locating Behavior of Ithomiine Males .... 1*55 Species-specific Male Courtship Behaviors .... 164 Pursuit Behavior of Courting Males 181 Function of the Display Perch 185 Mating 186 Sperm Precedence 189 Multiple Matings 191 Carrying Pair Behavior 193 Oviposit ion 195 Oviposition Strategies 200 Modes of Oviposition Behavior 215 Foodplant Specificity 223 V COMPARATIVE LIFE HISTORIES OF THE ITHOMIINAE 225 Immature Stages 226 Eggs 230 Larvae 231 Pupae 235 Generation Time £ 237 Parasitism and Predation 256 Parasitism 258 Predation 262 VI THE ITHOMIINAE-SOLANACEAE INTERFACE 267 Plant Defenses and Larval Adaptations 268 Larval Foodplant Relationships of the Ithomiinae ... 270 Patterns of Foodplant Utilization 281 Foodplant Specificity of Limoncocha Ithomiines . . 286 Utilization of Larval Foodplants at Limoncocha .... 292 VII THE MIMETIC RELATIONSHIPS OF THE ITHOMIINAE 301 The Mimetic Subcomplexes at Limoncocha 302 Transparent Mimetic Complex 303 Tiger Mimetic Complex 315 vii Page CHAPTER ... 326 Polymorphic Mimetic Species ...•••■ Ithomiine Participation in Limoncocha ^ ^ Mimicry Complexes Evidence for Unpalatability in the Ithomiinae . . • • 331 . . 331 Circumstantial Evidence ^ Direct Evidence Some Mimetic Consequences of Ithomiine Ecology .... 335 ... 344 VIII CONCLUSIONS ' ' * 349 REFERENCES ... 361 BIOGRAPHICAL SKETCH vm LIST OF TABLES TABLE Page 1. Water Balance Calculation for Llmoncocha, Ecuador . 19 2. Comparison of the Butterfly Faunas of a Variety of Neotropical Locations 40 3. The Number of Genera and Species in the Eight Tribes of Ithomiinae 49 4. Species of Ithomiinae Collected at 'Llmoncocha, Ecuador ... 53 5. Species of Solanaceae Collected at Limoncocha, Ecuador ... 67 6. Nectar Sources of Ithomiine Butterflies 92 7. Summary of Collections of Ithomiine Butterflies Visiting Eupatorium I Flowers at Study Sites 1 and 3 in January 1974 108 8. Species Diversity of the Ithomiine Community at Study Site 4 132 9. Similarity of Species Composition Between Consecutive Samples of the Ithomiine Community at Study Site 4 (SjEfrenson's Quotient of Similarity) 136 10. Estimates of Longevity of Ithomiine Butterflies Based on Recapture Data at Study Site 4 144 11. Sex Ratio, Probability of Interspecific Encounter (PIE), and Probability of Intraspecif ic-heterosexual Encounter in the Ithomiine Community at Study Site 4 157 12. Postures of Male Ithomiines During Display Perch Courtship 172 13. Postures of Male Ithomiines During Patrol Perch Courtship 178 14. Mating Pairs of Ithomiines Observed at Limoncocha 188 15. Summary of Oviposition Parameters for Ithomiine Butterflies 202 16. Summary of Developmental Times of Ithomiine Butterflies 238 ix TABLE 17. Parasitism of Field-collected Ithomiine Eggs and Larvae 18. Larval Foodplant Records for the Ithomiidae 19. Numbers of Solanaceous Plants and Ithomiine Immatures at Site 4 20. Numbers of Solanaceous Plants and Ithomiine Immatures at Site 5 21. Numbers of Solanaceous Plants and Ithomiine Immatures at Site 1 22. Numbers of Solanaceous Plants and Ithomiine Immatures at Site 6 23. Summary of Foodplant and Juvenile Ithomiine Densities 24. Ithomiine Participation in Limoncocha Mimicry Complexes LIST OF FIGURES FIGURE Page 1. Map of East Ecuador (the Oriente) c 2. Map of Limoncocha, Ecuador .-. -, 3. Mean Monthly Temperatures, Mean Monthly Maximum Temperatures, and Mean Monthly Minimum Temperatures at Limoncocha, Ecuador 10 4. Annual Rainfall at Limoncocha, Ecuador: 1961-1974 11 5. Mean Monthly Rainfall at Limoncocha, Ecuador 13 6. Rainfall at Limoncocha, Ecuador: January 1974 to December 1974 13 7. Percent Rainy Days per Month and Mean Maximum Consecutive Rainless Days per Month at Limoncocha, Ecuador 14 8. Percent Days per Month with Greater Than 50% Cloud Cover at 1300 hours at Limoncocha, Ecuador 15 9. Mean Monthly Percent Relative Humidity at 1300 hours at Limoncocha, Ecuador 16 10. Near Beginning of Nature Trail (stem portion) in Secondary Forest 25 11. Nature Trail (loop portion) in Primary Forest, near Study Site 4 (100 m by 2 m transect) 25 12. Logging Trail, Study Site 3 27 13. Males of Scada batesi Haensch Feeding at Eupatorium I Flowers at Study Site 3 27 14. Logging Trail, Study Site 4 29 15. Forest Interior at Study Site 4 29 16. Number of Species in each Butterfly Family at Limoncocha, Ecuador 36 xi FIGURE 17. Number of Species in each Butterfly Family in Eastern Pernambuco, Brazil 18. Number of Species in each Butterfly Family in Eastern Espirito Santo and Southern Bahia, Brazil 19. Number of Species in each Butterfly Family at Jaru, Rondonia, Brazil 20. Relative Abundances of Species of Ithomiinae at Limoncocha, Ecuador 21. Ithomiine Feeding Activity at Heliotropium I Blossoms: June 26, 1974 22. Ithomiine Feeding Activity at Heiiotropium I Blossoms: June 29, 1974 23. Ithomiine Feeding Activity at Maranta I Blossoms: July 12, 1974 24. Ithomiine Feeding Activity at Eupatorium II Blossoms: July 12, 1974 25. Associations of Species of Ithomiines with Microhabitat Regions of the Nature Trail Observation Area 26. Distribution of Captured Ithomiine Individuals Among the 50 Subplots at Study Site 4 27. Summary Graph of the 26 Ithomiine Samples Taken at Study Site 4 28. Species and Individuals Present in the 26 Ithomiine Samples Taken at Study Site 4 29. Equitability (J1) and Quotient of Similarity (QS) Values for the 26 Ithomiine Samples Taken at Study Site 4 30. Regression of the Quotient of Similarity (QS) on Sample Size (N) 31. Age Class Distribution of the Ithomiine Community at Study Site 4 32. Courtship Activity of Male Ithomiines 33. Location and Erection of Hairpencils in Male Ithomiines . . xn FIGURE Page 34. Napeogenes apobsoleta in Display Perch Courtship Behavior 161 35. Mechanitis messenoides in Patrol Perch Courtship Behavior . ." 161 36. Perch Height versus Time of Day for all Courting Male Ithomiines 166 37. Perch Height versus Time of Day for Display-Perching Male Ithomiines in the Yellow Clear-wing Mimetic Subcomplex 167 38. Perch Height versus Time of Day for Display-Perching Male Ithomiines in the Tiger Mimetic Complex 168 39. Perch Height versus Time of Day for Patrol-Perching Male Ithomiines in the Tiger Mimetic Complex 169 40. Ethnograms of Eight Species of Display-Perching Ithomiine Males 176 41. Ethnograms of Three Species of Patrol-Perching Ithomiine Males 180 42. Mating Activity of Ithomiines 187 43. Oviposition Activity of Female Ithomiines 196 44. Summary of Ithomiine Reproductive Activities 199 45. Egg Volume versus Female Forewing Length 209 46. Modes of Ithomiine Oviposition Behavior 217 47. Illustrations of Some Eggs, Larvae, and Pupae of Ithomiines 229 48. Generation Times of 34 Species of Ithomiinae 245 49. Comparative Developmental Times Among Ithomiine Tribes . . . 248 50. Egg Volume versus Egg Development Time 251 51. Mature Larval Length versus Larval Development Time .... 251 52. Egg Volume versus Generation Time 253 53. Mature Larval Length versus Generation Time 253 Xlll FIGURE Page 54. Larval Foodplant Relationships of Ithomiine Tribes 284 55. Larval Foodplants of Limoncocha Ithomiinae 287 56. Numbers of Solanaceous Foodplants Utilized by Species of Ithomiinae at Limoncocha 290 57. Numbers of Ithomiinae Supported by Species of Solanaceae at Limoncocha 290 58. Transparent Mimetic Complex: Yellow Opaque Subcomplex and Orange-Tip Subcomplex 305 59. Transparent Mimetic Complex: White Subcomplex 307 60. Transparent Mimetic Complex: Yellow Clear-wing Subcomplex 309 61. Transparent Mimetic Complex: Large Clear-wing Subcomplex . 311 62. Tiger Mimetic Complex: Black, Yellow, and Orange Under story Subcomplex 317 63. Tiger Mimetic Complex: Yellow-Bar Canopy Subcomplex .... 319 64. Tiger Mimetic Complex: Yellow-Spot Canopy Subcomplex . . . 321 65. Tiger Mimetic Complex: Orange and Black Subcomplex .... 323 66. Abundances of Transparent Mimetic Subcomplexes at Study Site 4 338 4. 67. Abundances of Tiger Mimetic Subcomplexes at Study Site 4 340 68. Summary of the Abundances of the Transparent and Tiger Mimetic Complexes at Study Site 4 342 xiv Abstract of Dissertation Presented to the Graduate Council of the University of Florida in Partial Fulfillment of the Requirements for the degree of Doctor of Philosophy COMPARATIVE ECOLOGY AND MIMETIC RELATIONSHIPS OF ITHOMIINE BUTTERFLIES IN EASTERN ECUADOR By Boyce Alexander Drummond III August, 1976 Chairman: Thomas C. Emmel Major Department: Zoology Butterflies of the subfamily Ithomiinae (Lepidoptera, Ithomiidae) are among the most conspicuous and abundant flying insects of neotropical forests, yet they have received little attention from ecologists and their biology is not well known. A comprehensive ecological study of a rich (53 species) ithomiine community (in lowland rainforest at Limoncocha, on the Rio Napo , eastern Ecuador) was combined with an extensive search of the literature to characterize this subfamily in ecological terms. Ithomiines usually exhibit clumped distribution, even within seemingly 4. homogeneous forests, where their responses to gradients of light and rela- tive humidity result in transient concentrations that crest and ebb through time. Ithomiines are long-lived, with low daily rates of oviposition and adult-financed egg production. The early morning visits of ithomiine males to white flowers fit the syndrome of crepuscular pollination and may represent a coevolved relationship. Females feed on detritus from which they apparently obtain the nitrogen needed for continuous egg production. Courtship behavior is species-specific and involves two categories of mate-locating behavior by males, "display perching" and xv "patrol perching." The oviposition strategies of ithomiine females vary greatly and four modes of oviposition behavior are recognized. The adaptive significance of extended copulation in butterflies is discussed. Local populations of most ithomiine species appear to be narrowly specific in larval foodplant utilization, and the high diversity of the ithomiine community is probably due in large part to the high diversity of the larval foodplants (Solanaceae) in tropical forests. Larval development time depends on the mechanical and chemical characteristics of the foodplant, but the average generation time is about 28 to 30 days. Juvenile ithomiines display great interspecific variation in morphology in all life stages. The similarity of the previously unde- scribed larva of Melinaea menophilus (Ithomiidae) to larvae of other primitive ithomiines and to larvae of the Danaidae provides new evidence for the relationship between the two families. A nine-month censusing program at Limoncocha revealed that the ithomiine adult community appears to be quite stable in relative composi- tion of mimetic patterns, age distribution, species diversity (H'), and equitability (J1). Larval foodplants are greatly under-utilized at Limoncocha and intensive parasitism and predation of juvenile stages appear to be controlling the population sizes of most ithomiine species, although the overall abundance of the ithomiine community may be severely depressed by the indiscriminate adult mortality resulting from violent storms and periods of heavy rainfall. xvi CHAPTER I INTRODUCTION Community ecology may be defined as the study of the relationships among a set of coexisting interdependent populations (Price, 1975) . Early studies usually concentrated on communities defined largely by spatial constraints, e.g., a field, a rotting log, an algal mat, or any unit that contained several species. More recently, however, seg- ments of a community that contain biologically or evolutionarily closely related species have been employed as a means of studying the ecological relationships of a community. For example, Root (1967) has introduced the concept of the guild, a group of species (i.e., a set of populations) that exploits the same class of environmental resources in a similar manner. A guild, then, contains species that are ecologically similar, but may be taxonomically diverse. Another approach involves the study of a taxonomic segment of a community, or taxocene (Chodorowski, 1959; 4. Hutchinson, 1967), the members of which are likely to be of about the same size, to have similar life histories, and to compete over both evolutionary and ecological time (Deevey, 1969). This combination of characteristics suggests that comparative studies, both among the species within a taxocene and among taxocenes themselves, should be particularly useful in elucidating community structure. Although taxocenes can be defined at several levels, the broader the taxonomic category employed, the less similar the included species will be, and thus the more diffi- cult meaningful comparisons will become. The size and uniformity of the spatial or environmental dimension of a taxocene is determined by the organisms' size, mobility, and fidelity to particular microhabitats (Hurlbert, 1971). The present study is an attempt to better understand the butterfly community of a neotropical lowland rainforest through a comparative study of species of the subfamily Ithomiinae (Lepidoptera: Papilionoidea: Ithomiidae) , a group of butterflies that closely fits Deevey's descrip- tion of a taxocene. The ithomiine taxocene is attractive for such a study for several reasons. Although closely related and morphologically similar, ithomiines differ enough in Wing pattern, behavior, and repro- ductive strategy to permit meaningful comparative studies of species that compete over both evolutionary and ecological time. As a group, ithomiines are relatively abundant in neotropical rainforests, and the flight levels of most species make them accessible for detailed study. As adults, they are the most numerous participants in the extensive and widespread mimicry complexes of neotropical forests. As herbivorous immatures, they are known to specialize almost exclusively on a plant taxocene, the Solanaceae, that is notorious for the diversity of its 4 mechanical and chemical anti-herbivore devices. Such a relationship suggests that the study of the Ithomiinae-Solanaceae interface could provide extremely fertile ground in which to cultivate the theories of insect-plant coevolution. And yet, for all this, the Ithomiinae have received only scant at- tention from ecologists. In a preliminary study, Gilbert (1969) assessed their suitability for answering certain questions about community ecol- ogy. More recently, Young (1972, 1974a, b,c) has published life histories of four species of Ithomiinae from Cuesta Angel, a montane forest locality in Costa Rica, and thus has laid the foundation for a detailed comparative study there. Pliske (1975a, b,c; Pliske et al., 1976) has begun an investigation into the role of olfactory communication in the behavioral ecology of adult Ithomiinae. In addition, some ecological information may be gleaned from a few scattered observations in the early literature (e.g., Collenette and Talbot, 1928) and from some recent studies of mimicry complexes involving ithomiines (Poole, 1970; Papageor- gis, 1974; Brown and Benson, 1974) . In spite of these studies, however, the Ithomiinae remain poorly understood ecologically. Thus, in addition to using the Ithomiinae as a taxocene to approach butterfly community ecology, a main goal of this study is to increase the body of knowledge about the Ithomiinae, both to adequately character- ize the subfamily in ecological terms and to determine promising direc- tions for future research. Study Area The lowland tropical forests of eastern Ecuador (the Oriente) form the western extremity of that vast sea of tropical vegetation known as Amazonas, or the Amazon Basin. Drained primarily by the Napo and Pastaza Rivers and their numerous tributaries, the Oriente corresponds roughly to the western half of the largest and ecologically most diverse of the several lowland tropical refugia postulated to have harbored tropical forest floral and faunal elements during glacial periods of the Quater- nary. Named after the largest river in the Oriente, the Napo Refuge comprised mainly the lowlands of east Ecuador (prior to the political rearrangements of the 1942 Protocolo de Rio de Janeiro which ceded much of this area to Peru) from the Andes east to the upper Amazon (Haffer, 1974). The evolutionary significance of these Quaternary Refugia will be discussed later in this paper, but it can be noted now that the size and structural diversity of the. Napo Refuge was probably responsible in part for the high species richness that occurs in this region at the present. The study area for this research, Limoncocha, in Napo Province, Ecuador, is located within the boundaries of the Napo Refuge. Limoncocha is the field headquarters for the linguistic, education- al, and cultural work in eastern Ecuador of the Summer Institute of Lin- guistics (known in the United States as the Wycliffe Bible Translators), a nondenominational missionary organization that operates under contract with the Ministry of Education of the Ecuadorian government. Situated at 0° 24 » South latitude and 76° 38' West longitude, Limoncocha is lo- cated about 210 kilometers almost due east of Quito on the western edge of "Lemon Lake," an oxbow cut-off of the Rio Napo that lies two kilome- ters to the south. The Limoncocha area is bordered on the west by the Rio Jiveno (see map, Figure 1). Historical Background When the Summer Institute of Linguistics (S.I.L.) selected this site in 1957 as a logistic support base for its personnel working with isolated indigenous Indian tribes (Cofan, Secoya-Siona, Shuara=Jivaro, Waodani=Auca, Yumbo) in the Oriente, virgin forest covered the entire area and there were no Indians living in the vicinity, although the re- mains of pre-colonial villages have since been found on the lake's west- ern edge. Since those humble beginnings, when tools and supplies were flown in by amphibious plane, Limoncocha has grown into a sophisticated settlement, complete with electricity from a diesel^powered generator (operated 0700-2100 daily) and indoor plumbing. The twenty or so mis- sionary families that presently live and work at Limoncocha are supplied by weekly flights of the Institute's own D03 which lands on the 1100 meter airstrip completed in the mid-1960s (see map, Figure 2). The only other access to Limoncocha is by canoe from Coca, the terminus of the oil pipeline road built a decade ago by Shell-Texaco to exploit the con- siderable oil reserves of the Oriente (see map, Figure 1). During the 1960s, the rapid physical growth of the missionary base at Limoncocha was accompanied by the coalescence of the scattered low- land Quichua (=Yumbo) Indian families of the area into a village just south of the airstrip. As both the village and the missionary settle- ment grew, increasing demands were made on the surrounding forest, first for lumber and game, and then for cleared land for crops and pasturage. While the entire area within a ten kilometer radius has been heavily hunted and selectively logged, clearing of the land for yuca (=manioc) and banana fields had, until recently, been limited to the area between Limoncocha and the Napo, especially along the "Napo Road" and the lower Rio Jiveno (see map, Figure 2). More recently, however, the Ecuadorian government has initiated a colonization program that offers 50-hectare land grants to "colonistas" who settle in the Oriente, an effort designed to help alleviate the burgeoning overpopulation of the highlands by en- couraging eastward migration. The Limoncocha area has many attractions for these colonistas, including two rivers and one lake for fishing, availability of medical and dental care (two Registered Nurses reside at the missionary settlement), and Limoncocha' s growing importance as a transportation center. In addition to the Quito-based DC-3, the Insti- tute maintains two Heliocourier aircraft at Limoncocha for shuttling Figure 2. Map of Limoncocha, Ecuador, translators to and from tribal areas and for transporting Indians into Liraoncocha for various educational programs and for medical care. The synergistic effect of the government's colonization program and Limon- cocha's attractiveness as a settlement area presents an ever- increasing demand on the terrestrial and aquatic ecosystems of the area. At the present rate of growth there will be very few — if any — undisturbed for- est areas within a 15 kilometer radius of Limoncocha by 1980. Conse- quently, as the land is cleared or otherwise disturbed, Limoncocha's val- ue as a site for tropical scientific research will continue to diminish proportionately.* Climatic Description Weather records (rainfall, maximum and minimum temperature, rela- tive humidity and cloud cover) have been kept by S.I.L. and Ecuadorian military personnel at Limoncocha since 1961. Monthly and yearly means based on the first 14 years (1961-1974) of Limoncocha weather data fall well within the ranges of values given by Richards (1964) as characteris- tic of lowland tropical rain forests. Such forests are characterized by a high mean annual temperature with little seasonal variance, relatively heavy rainfall in all months of the year, high humidity, low wind veloc- ity except during storms, and frequent cloudiness. * Prior to this study, Limoncocha had hosted a number of visiting scien- tists, usually for visits of only one or two weeks. In addition, at least three long-term studies had been conducted at Limoncocha; one on soil com- position, one on bird foraging strategy, and a third on army ant behavior. Mean annual temperature at Limoncocha is 25. 2° C, with the range among monthly means less than 2°C. The mean annual maximum tempera- ture is 29.8°C and the mean annual minimum temperature is 20.6°C, a dif- ference of less than 10° C. The mean monthly maximum, mean monthly min- imum and mean monthly temperatures are graphed in Figure 3. The mean total annual rainfall is 3065 millimeters (range: 2600 to 3725 mm) . If total annual rainfall is plotted on a yearly basis from 1961 to 1974, the resulting histogram gives some indication that annual rainfall at Limoncocha is increasing (Figure 4) . Although this apparent increase may be only a portion of a much longer regional rainfall cycle, or merely chance variation, Emmel (1974) has suggested that the increase may be correlated with the "burning-of f " of the new oil wells opened to the northeast of Limoncocha in the early part of this decade. The par- ticulate matter lofted upward in the form of smoke from the burning wells may serve as nuclei for condensation and thus lead to an increase in precipitation. It is true that the mean annual rainfall for the years 1961-1970 (2935 mm) is considerably less than the same statistic for the period after 1970 (3390 mm), but the range of variation during the 14 years of records (1124 mm) is 2.5 times as great as the difference be- tween these two means (456 mm) . Several more years of data will be need- ed to determine if the mean annual rainfall since the burn-off began will stabilize around a new and significantly higher mean. Although Limoncocha lies almost directly on the equator, where cli- mate is theoretically aseasonal, its weather data tend to support Richards' (1964) statement that "there are probably no land surfaces within the tropics with a completely aseasonal rainfall." Limoncocha' s proximity to the equator insures that a substantial amount of rain falls 10 w Q > O H O O pv. w CM n CO 00 CM vO CM CM CM CM o CM CO — «- vO W 55 c o s c to > 3 rH CJ jr w 4-1 C m O cfl o C O to o CD CO ,J 0) u D 4-> cd to CO a d 0) CO H U at >. a. j= o z H o PL, w CO pi 5! 01 4J U ai CD c o o • U x o w I C 0) a) .£ aj u S o o •o c c o •5 4J C 4J o « S^ ai d ex o S 0) i*. u n) a) O o. >* m C ^ a OD a oi 0) rH U C M -H a) to Ph Pi 0) oo P* HINOH tf3 4J o ccj M 0) > o u 3 o CJ &•? o m c CO H J-i CO ■u CO 0) M O H O O W to 3 SI M W PL, o o o o oo o o O o o n o CN c o S-i S o -n M CO 0) 3 p, C) W en !>■ •i 01 CD O •C O ■u n a CJ 01 C a o Vj 6 CO •H CM j-J m ooex iv ^3aoo anoia %os hum savq ao aoviNaDsaa CO u 3 60 •H u o -o td 3 O w 16 o o > o to O o o c o e H O O P-. w to to H 3 O o o CO td I PC > Pi 5! Pi 4J c P-. pq W p-i C o S c cd 0) S o o o O 00 O o o o — i- o fN> M OOCT IV AliaiWflH 3AIIV13S IN3DH3 W OJ Ph w M O id cfl 3 CJ w CO .3 o o o s o 6 S-i o c o •H ■u CD rH U rH Cfl U CU CJ 3 ca rH CO m u CU CO H Pi o m CM 4J CO •• ^ co 4J •H •H O CJ a CO a. rH cO rt u o •H 13 P- rH o 0J u •H H Ph M • • O CU 3 cu o u N 3 4-) 0) CO UH •H •H O i-J 2 CD 0) 60 rH -3 .O •H cO h .-1 TJ •H rH CO O > a < H is O 2 o w > O is H O O w co O i=> erf S! w Ph CM VO • VO m O CM CO m CM m CM vO m CM CO CM CM -3- CM CM m CM CM m CM m CM CO m CM m CM CU u 3 ■u cfl 1-1 0) Ph g H 00 CM CM CO oo CM CM CO CM CM CM vO r-. CM o CO CO o CO oo CM en 00 vO rH CM 00 00 CM -a- CM 00 CM CM CM m CM as n o cO ■M o 0) u Ph m o o oo O m o m o o cm o o m o m ph m o CM c_) Py 00 rH rH o py m m rH py O rH rH py CM O rH c_) py m m CTl rH rH CM oo rH o m CM rH VO oo rH o CM CM rH CM oo rH o CM rH O vO rH o oo rH rH rH o 00 CN 00 oo o CO 3 rH P. U 3 CO u a> 4-> Cfl £2 0) 60 d cO X. CU u 3 •u CO •H O 2 o CO o CO cu rH .3 cfl cfl > cfl a o e CU TJ H 3 3 cu j-i CO 4J •H Cfl O 2 CM CM oo rH Ph vO 00 rH Ph CM oo rH u Ph O vO rH Ph rH Ph 00 00 o 3 3 Pi IH O cu J-> u 3 3 O 4-1 T3 w CO cu •H U CO O M M 6 3 Cfl CJ CU cfl cJ !>, O .3 VH O J-> O "H -H X CJ S -H IH 3 cu g -H T3 3 Z 0) CO •u 0) cu 60 •H T3 rH o w o 4-1 CU e CD o H 60 3 iH T3 U O O O CO cu CO 0 CO 3 o 3 a cfl o is 20 Physical Description Forest area Limoncocha is approximately 280 meters above sea level and almost 15 meters above the level of the lake and the two nearby rivers (Rio Napo and Rio Jiveno) . All other relief in the area is less than 10 meters. The soils are mostly laterites, covered in forested areas by a layer of humus and decaying organic matter several centimeters thick. Forest soils are always wet and forest trails are muddy year around. During the wetter periods, water may accumulate in poorly drained areas to depths ranging from 2 to 70 cm, with standing water remaining any- where from two days to two months. The canopy of the mature forest averages 40 to 50 meters in height, with emergents (e.g., Ceiba) occasionally reaching 70 meters. Undis- turbed areas have all the characteristic features associated with climax tropical rain forest: tall, smooth-barked trees with buttressed roots; abundant epiphytes, creepers, and climbers (lianas); a large saprophyte community; a relative openness of the forest understory and a correspon- ding density of the canopy; a monotony of leaf color (dark green) and leaf form (predominantly elliptical, with an acuminate tip or "drip tip"); a paucity of flowers in the understory; and a high species diversity (i.e., a great number of plant species present in a small area, each represented by only a few individuals) . In the Limoncocha forests palms (Palmae) are quite abundant, espe- cially those known locally as chonta (Iriartea sp.) and chontadura (Bactris gasipas H.B.K.). Natural clearings, from 0.2 to 1.0 or more hectares in size, are an unpredictable but not infrequent occurrence caused by the blow-down of trees. These temporary sunny clearings are 21 rapidly colonized by various fugitive plant species, especially by mem- bers of the genera Eupatorium CAsteraceae) , Ochroma CBombacaceae) , Heliotropium (Boraginaceae) , Cecropia (Moraceae) , Heliconia (Musaceae) , and S planum and Physalis (Solanaceae) . These same fugitive elements are « also common along the more open forest trails and along the edges of man-made clearings. Clearings In 1974, a total of approximately 83 hectares of land at Limoncocha had been cleared by S.I.L. and the local Quichuas. This total may be broken down as follows: airstrip clearing, 20 hectares; missionary set- tlement, 18 hectares; pasturage, 40 hectares; and Quichua village, 5 hec- tares (see map, Figure 2). As indicated on the map in Figure 2, there are several areas of secondary or heavily disturbed forest bordering the base, totaling approximately 66 hectares. Since 1974, a great many new clearings have been made along the Napo Road and on both sides of the Rio Jiveno. By early 1976, virtually all of the remaining lands in these areas had been claimed by colonistas (James Yost, personal communication), 4. and thus more clearings can be expected in the future. The open borders of the grassy airstrip are covered by Ipomoea (Convolvulaceae) and the exposed forest face by the introduced kudzu vine (Pueraria lobata) . The pasture blends gradually into the surround- ing disturbed forest. Only a few of the original forest trees are still standing in the settlement clearing, which has been heavily planted with fruit trees, including orange, grapefruit, and lemon (Citrus), avocado (Persea) , and banana (Musa) , and ornamentals, e.g., African oil palm (Elaeis guineensis Jacq.) and Hibiscus. 22 Trails There are several forest trails near Limoncocha. Those used in the course of this study are described below (also see map, Figure 2). Napo Road (NR) . This is a broad trail, ten meters wide, that runs south from Limoncocha for three kilometers to the Brazo of the Rio Napo, at a point just east of the mouth of the Rio Jiveno. Numerous side trails branch off to the east and west. Nature Trail (NT) . This trail runs west from the Power House for about 80 meters before it forks into two branches that later reunite to form a large closed loop. The 80-meter stem of the NT passes through second growth forest (see Figure 10) . The closed loop occurs in slightly disturbed primary forest (see Figure 11). This two meter -wide trail has been marked with numbered stakes corresponding to points of interest which are explained on a cassette tape (to be played by visitors while walking the trail) . Southeast of the NT loop are a series of anastomosing trails less than one meter wide which give access to the NT Study Area (Site 5). Hunting Trail (HT) . This trail begins at the western tip of the NT loop and continues in a broad west-northwest curve. Quite narrow, it is difficult to follow after the first two kilometers. Logging Trail (LT) . Originally a ten meter-wide trail that would accomodate a tractor for the first part of its length, this trail angles NW and roughly parallels the HT. Along it are study sites 1, 2, 3, and 4. In the past, logs were hauled by tractor down this trail to the saw- mill located in the Power House building. Greatly overgrown today, the LT fades to less than one meter in width beyond Site 4. A severe storm with localized twisting winds occurred in March 1974, and blew down 23 several very large trees and numerous smaller ones, between Sites 2 and 3. The resulting morass of tangled vines, shrubs, and fallen trees rendered this stretch of the LT nearly impassable. By June 1974, this devastated trail was so choked with new growth that it was never used again as a passage to Site 4. Cross-Cut Logging Trail (CCLT) . A five meter-wide trail first cut in early 1973, this north-south trail links the NT with the LT. After June 1974, it provided the quickest means of reaching Site 4. Because of its width, this trail receives much sunlight and thus is difficult to keep open. By May 1975, this trail, was completely blocked by vines and Heliconia plants and was no longer passable. Location of gtudy sites Intensive or long-term studies were conducted at six study-site localities. Site 1, Site 2, Site 3, and Site 4 were located at points along the Logging Trail. Site 5 and Site 6 were in the Nature Trail area (see map, Figure 2). During this study, Sites 1 and 3 were open sunny areas (see Figure 12) with large numbers of blooming Eupatorium plants that attracted many ithomiines (see Figure 13) . Site 2 was a much more shaded spot with several large Heliotropium plants in bloom. Site 4, bisected by the LT (see Figure 14) , was a one hectare rectangular plot staked out in rela- tively undisturbed, mature, but seasonally flooded forest (see Figure 15). Site 5 was a 0.5 hectare plot located in relatively undisturbed, mature, unf looded forest (similar to that shown in Figure 11) . Site 6 was a transect 100 meters long and 2 meters wide, located along the closed loop of the Nature Trail (see Figure 11) . -. o T3 ■H 3 • 4J +J tf*S (-1 C/J 4-1 O a (3x t-t . H ^ r^ QJ n s W CO 3 e o 4-1 •H o CO J-l -H !3 PU ■*— * (1) u 3 60 •H ft} M >s H Vi cfl a) -a H C 3 O ■U O cd 0) IS c/3 <4-l C O -H M^-v q C •H O C -H C 4J •H ^ 60 O 0) p. ■ PQ 4-1 •m 1 a) CO 4-1 u a) co o 2 — ' fe Q) VI 3 60 25 Figure 12. Logging Trail, Study Site 3. Figure 13. Males of Scada batesi Haensch Feeding at Eupatorium I Flowers at Study Site 3. 27 3 U o u QJ CO a) M o m 0) M 3 DO ■H Pn QJ 4J •H CO 3 4J CO M H 00 C •H to o M 3 t>0 •H ft. 29 30 The experiments and observations performed at these study sites are described in the methods section. The Butterfly Community at Limoncocha, Ecuador Composition Along with birds, butterflies are perhaps the most conspicuous of all neotropical animals. This is primarily the result of their diurnal habits, the often brilliant coloration of- their wings, and their rela- tive abundance compared to other animal groups in the same region. As is characteristic of many invertebrate taxa in the tropics (Elton, 1973), this abundance of butterflies results not from a few common species each with a large population, but from the presence of a great number of species most of which have a relatively small number of individuals (Ebert, 1969). In short, neotropical butterfly communities usually ex- hibit high species diversity (sensu Simpson, 1949) . The butterfly community of Limoncocha fits this pattern of diver- sity quite well. In over twelve months of collecting (two collectors) in the Limoncocha area, about 366 species of butterflies were taken (this does not include the skippers, Superfamily Hesperioidea, of which some 75 species were taken) . Based on the number of species represented by only one or two specimens in our collection, I estimate that there are at least another 75 to 100 species that occur in this area, and pos- sibly as many as 150 or more. Thus, the total number of butterfly species at Limoncocha (excluding skippers) is estimated to be between 400 and 500, with 460 as a reasonable figure. 31 The reasons for the incomplete sampling are several. First, only a small portion of the total time spent in the field was devoted to an active attempt to locate and collect new (i.e., previously untaken) species. Second, nearly all specimens collected were taken in the lower four meters of the forest or in and around clearings. Thus, those spe- cies that fly and perch in the mid to upper forest canopy are quite un- der-represented in the collection as they were caught only when they occasionally strayed into the lower levels of the forest or were attracted to flowering plants in clearings and along trails. Even without the above restrictions, however, the collection of all species of any insect group occurring in any one neotropical locality is exceedingly diffi- cult. Indeed, writing more than one hundred years ago on the low densi- ty of most insect populations on the upper Rio Solimoes in Brazil, Bates (1857) concluded that "one year of daily work is scarcely suf- ficient to get the majority of species in a district of two miles cir- cuit." Third, some species may be present in an area at a given time, then i disappear for a number of months or years, only to reappear at a later date. In this context, however, "disappear" probably has more than one meaning. It could mean that a species is always present in an area, but that its population density becomes so low at certain periods that its presence could only be detected by the collection of an extremely large sample. Or it may mean that the species population becomes lo- cally extinct at times and is re-established only through later immigra- tion. It is most likely that both meanings of the word describe the dynamics of some of the species at Limoncocha. 32 This same phenomenon of periodic occurrence has been observed in the Limoncocha avifauna. After eight months' study in 1971 and 1972, David Pearson (1972) compiled a species list of all birds occurring within a five kilometer radius of Limoncocha. The list numbered 3A4. In early 1974 an ornithologist visiting Limoncocha for ten days added over 30 species to the list. Another ornithologist presently studying at Limoncocha has now increased Pearson's original list by over 50 spe- cies. Some of these new birds are undoubtably seasonal migrants that Pearson might have seen had he been at Limoncocha for a full twelve month cycle, and some may be species that were present but so rare in 1971 that he never encountered them, but have since become more common. Still others probably represent species that simply were not present at the time of Pearson's study but have invaded the Limoncocha area since that time. Similar phenomena are probably influencing the butterfly community. Some of the species taken at Limoncocha only once are undoubtably moun- tain forms that apparently strayed or were blown eastward from higher elevations on the eastern slopes of the Andes (e.g., Athyrtis mechanitis salvini Srnka) . Also, during short trips to Limoncocha in 1972 and 1973, I have taken a few species that I never saw or collected again during the whole of 1974. Furthermore, some species at Limoncocha (e.g., Heli- copis interrupta) have been observed to occur in "blooms," synchronous emergences of large numbers of individuals at seemingly unrelated times of the year, probably triggered by some unique combination of climatic factors. Heinz Ebert (1969), after years of studies on the frequencies and distributions of butterfly populations in Brazil, concluded that 33 many species of butterflies, especially those of the most numerous fam- ilies (Hesperiidae, Lycaenidae, and Riodinidae) , have populations that are small in size, dispersed widely, and which are not constant in lo- cality, but "migrate" continuously within a great area of favorable habitat, such that local colonies are continually appearing and disap- pearing. Thus, just as the scattered distribution of rain forest trees resem- bles a spatial mosaic (Richards, 1964), so do the temporal vagarities of butterfly abundance resemble a phenological mosaic. Such variability in species composition must always be carefully considered when conduc- ting or evaluating research on tropical butterflies or other insects, and perhaps even birds and mammals. (This phenomenon has long been rec- ognized as characteristic of most tropical forest human populations.) Ross (1967) has observed that the systematics of the Lepidoptera, especially tropical forms, is in a relatively unstable state. While he was referring primarily to the generic level and below, it is also true that very few students of butterflies agree fully on the number and content of butterfly families. Ehrlich (1958) has published the most recent comprehensive review of the phylogeny and higher classification of the Papilionoidea (i.e., butterflies other than skippers), based on a comparison of a large number of morphological characters. In that paper he recognized only five families of butterflies, the Papilionidae, Pieridae, Nymphalidae, Libytheidae, and Lycaenidae. Several long-stand- ing families were downgraded to the level of subfamily or tribe, in- cluding the family Ithomiidae. In spite of this, however, many subsequent workers, especially those who have monographed family-level taxa (e.g., Miller, 1968), have rejected much of this downgrading and have chosen 34 to retain most of the families familiar to lepidopterists . Following the latter approach in this paper, I recognize nine families of butter- flies: the Papilionidae, Pieridae, and Libytheidae, plus four families formed by the separation of Ehrlich's Nymphalidae into the Ithomiidae, Danaidae, Satyridae, and Nymphalidae, and two from the separation of his Lycaenidae into the Riodinidae and Lycaenidae. The number of spe- cies collected at Limoncocha in each of these nine families and in the Hesperioidea (skippers) is presented in Figure 16. The number of species estimated to be present in each family (see discussion above) is also indicated in the same figure. Unfortunately, there are few published summaries of comparable data from other tropical forest localities. Most such faunal lists have been regional in scope, and thus cover a variety of habitat types and a sub- stantial altitudinal range. To provide a rough comparison, however, data from three Brazilian sites have been graphed in Figures 17, 18, and 19. Figures 17 (Eastern Espirito Santo and Southern Bahia; Brown, 1972) and 18 (Eastern Pernambuco; Ebert, 1969) represent regional totals for two areas located in the isolated strip of littoral forest along Brazil's i southeastern coast. Figure 19 (Jaru, Rondonia; Brown, 1976b) represents a specific locality in the southwestern Amazonian forest of Brazil. In general the littoral forests of southeastern Brazil are not so rich in species as are the lowland forests of the Amazon Basin. Indeed, the estimated number of species at the single locality of Limoncocha is greater than that recorded for the entire eastern portion of Pernambuco (Figure 18) and the single locality of Jaru (Figure 19) is estimated to have about the same number of species as do the extensive coastal forests of southern Bahia and Espirito Santo (Figure 17). i-H N M-l tO U U a) pq 4-1 4J •> 3 O PQ CJ 3 43 4= o B cfl cfl QJ C u 3 oj •H O. CO 3 a> w tH 0) cj 4-i 0) CO ex ca CO W M-l 3 O -H V4 >> CU T-t 43 iH | I 53 (*< I I CD 43 4J ID 0) 43 cd CJ 3 13 QJ QJ CO 4-1 CD U U O P. P< QJ 4-> O c •X3 T3 CO a a) j) •H 4-> & C8 co e tj C J-> 4-1 cu co u OJ OJ T3 »H CU CO rH • > 0) O /-^ Bl-H ON & cj r- cu M ot O, Q) r-t CO 43 J3 4-> O 3 S co o u u cu n rt cu 43 pq p]4)H^ a> n 3 60 Fn i-H a) 60 i u 43 -H 0 cu •U hJ 3 1 •U • 3 •H 4J 1-1 co 4-1 3 O CU • QJ CO M 13 4J "3 43 QJ 4„ "J CO 0) 4-) £: 3 O 4J CO a cj -H U CO QJ cd w 13 0) QJ •H cu 3 rH 4-1 O •^ •H iH CO QJ 3 a) O CJ Cu • •H 43 U CJ -H CO 4J O CO T3 a CO O 42 CO 3 t-H QJ 0) CJ cu •H cfl CO •h 3 iw -H 3 QJ cj o O CJ Vi o 1-1 cu e 0) cfl •H & 0.iH c a.43 4-> co .-J O CO -H QJ •H 4-1 T3 43 M-l 4-» 4-> iw O T3 O CO )-i O cfl O o 3 4-1 1-1 >% p< VI O y-i 0) iH a) •H o -3 43 -H 4«S 43 4-> QJ e e >-■ e !-l >-l 4-1 3 cfl CO 3 O QJ Cfl 2; f=4 o a a.43 B cu vi 3 00 36 3 M saioaas &o raaunN SaiOHdS JO H3ffWHN .-1 CO U-l • CU cu id M ■H X J3 s CU •H 4-1 4-1 N o -o 4J CO cu 4J cu 3 Vt 4-1 4-t • pq PQ tfl T3 o /-^ o CO cu ,n fi » •H 4-1 H \£> a cfl T3 CO r-t r~- cfl •H c I o o\ 0) c O •H •H 4J o .H a T3 CO CO M a •H c •a cu cu cj o a J3 [5 CO Pi CU CO g O CU cu ^3 Vi •H •t T3 •H 0 PQ O 3 CU O *»_x CU V-4 > CU CU Cu tfl CO a js T3 CO ^ & CO H at 4-1 14-1 4-1 X 14-1 >-t o BJ 4-1 o • o •H 4-1 a. 1-1 '■>, » u c cu CU iH cu cu V4 JO •H CO x> CO a M e cu ■u ^J cfl CO 3 Vj O ^ fH pq c & c o\ cu V4 3 GO •H Pn f*> rH <*-t g n o o CU 4-> u 4-> C 4-1 U flv 3 CO PQ 4.0 rH J3 4J tH O -H N Cfl H tfl tO w a -h Q) H J2 •H 4-1 3 43 o> O -H 4J rH 3 1J NO *• O H CO 4-1 ■9 "d u s e -a cu 3 to c J3 is tn to w 00 U 3 60 •H 38 NYM £ saioaas io uaannN H 1 1 h CM o CM h — i — i — i — ( — i — » — i — ^ saioaas ao HaaKM 39 Interestingly, the butterfly faunas of two coastal regions have a much greater percentage of Hesperioidea (32-36%) than do the Amazonian forest areas (25-26%) . The presence of extensive grassland and cerrado « areas in the coastal regions probably account for the greater relative number of skipper species in these areas; many skippers specialize on monocotyledonous hosts. A comparison of the two Amazonian forest localities reveals that the distribution of species among families is quite similar at Limon- cocha (Figure 16) and Jaru (Figure 19)., although the overall species richness of the Jaru site may be exceptional (with many years of field experience across South America, Brown, 1976b, has called it "...the most exceptional collecting area I have ever seen or imagined"). Thus, I conclude that Amazonian lowland forests may be characterized by a relative distribution of butterfly species among families similar to that shared by Limoncocha and Jaru. The families may be ranked accord- ing to the number of species they contain. Beginning with largest, the order is (1) Hesperiidae (Hesperioidea), (2 and 3) Riodinidae and Nym- phalidae (very nearly equal, but either family may outnumber the other at a given locality), (4) Lycaenidae, (5) Satyridae, (6) Ithomiidae, (7) Pieridae, (8) Papilionidae, (9) Danaidae, and (10) Libytheidae. The actual number of species present in each family at most lowland Amazonian sites probably varies roughly between the values given for the number of species actually caught at Limoncocha (Figure 16) and the number estimated to occur at Jaru (Figure 19) . A comparison of butterfly faunal lists from various areas in the Neotropics (Table 2) suggests that the Ithomiidae form from 4.5 to 12% of the number of species of Papilionoidea present in a given area. 40 co G o •H cfl o o cfl u •H ft O U 4J o 2 o >-. 4J CU CO > D-l o in a 3 cd M-l tt 4J 4-> 3 PQ 0) 43 O s o co •H M CO I CU rH .O CO H CO 0) co QJ x: u 3 cu u CO >> p. rH ■H 3 B •rl rd ft a 0) M > CU •H ft oo W CO 0) O u~i - o cy> n o r- fo £ rH °S" ? ^ S§5 8 - S O O rH jHJNCslsJ-CNrHvOvOr^cNr..,* ^£SS^SSJ2^SSS st Sf <* CM co <* co^- sj- <■ -3- m co co «3- r- CM csi O csi Ov0l0C3\0V10rHvO00CN CNbOrHCSIrHCSICM>^^> 4J CN B •H r- & C r-i a A= 3 o • •H rH o /-N CJ o B • CN •H « > 3 rH 1 o 0; o 4J •« • 3 3 rH TJ O o lo • • •H TJ 3 3 •H 3 5-1 • rH • vO cfl ^-\ 3 3 3 U A3 - 3 pq 3 > cfl rH • 3 eri •H cfl U B 3 3 3 %_• 4-1 ••> •H • 3 1 •— s J-l VD CJ rH 3 pq & o o 3 4-1 60 M LO A3 cfl •H CO O CO iJ B 4-1 CO p. cfl rH • VO rH > CO 3 O 3 3 H CO QJ a CO 3 r-» Tl cfl 4-> O TJ ••* U pq 3 O J_l j-i !3 • cfl • •* J-l o\ 3 « ■3 r4 3 CN rH T) QJ ^-^ 3 O •4H O 3 ^~\ rH TJ Cfl rH Cfl 4J 3 oo 3 1 3 3 A3 3 LO O <4-l O J-I 4-1 3 r-J rH H 3 N O U 3 % B r-J rH f>"> -H ID 9 CO * & CU ■3 •H O rH | 3 TJ 3 S 4-1 ps a. rH O 3 O A3 3 3 >4 m 3 CO O O 3 O •H Cfl H t> r! & rH W •H 3 O rH T) CO CO o TJ LO J-i 3 H 3 p. o PS O N— * rH K 3 in 3 rH 3 o a> o u rH •H O CO )-l T> rH O •» 3 § o 4-1 TJ & 3 pq cfl g CO -o O CO 3 60 3 •H 3 o O ^ 4-1 3 3 rH 13 O- 3 & PS 3 J-I rH I— 1 CO CU cfl 3 CO O M V4 3 rH O 3 o c 3 A3 CO J-J •H -T e cfl O PQ rJ pq cu Pm 3 O rH • •> cfl 4-1 rH A3 u a) U 4-1 3 o o 3 •H a. k 4J 0) O > J-i 3 •H CO CO cfl a 3 o u a a. 3 •H rH w V4 O cfl O O rH CO CO .« o CO 3 A3 3 3 O 3 -H a> 3 «— ' CO r-j MD .« & •H •H ~ •H » CO ••> 3 ex o o O O Cfl " 3 O CO 4J 4J 3 TJ •« o r4 •* 4-1 • *> » r. CJ rH I-i O ^ TJ H CO O CO 3 3 O t-i o •H N Cfl 4-1 3 a) QJ 4J QJ cfl 3 3* 4-1 3 r-j cfl ■— 1 4-1 m r4 O 3 4-1 •H CO A3 rO O rH • O rJ o u -> •H H 3 A3 3 00 H CO 3 cu £ 3 ^^ U O CTl U O CU 4-1 cfl 3 • n CN a CU ,rH CO TJ O SU O *T3 3 •4H o 3 PS T) •H Ph J-i o IH _ a 4-> CO S cfl - TJ 4-t IH § PS LH CO 3 4H a) O (-1 UH iH in e H u r3 3 3 O o CJ A3 CO rH CJ 3 ai 3 O 3 O u 3 Cu ■ 3 3 W 3 3 3 3 11 H • 3 TJ O 4-1 4-1 0) • 4j a) 4-> O 4-1 -rl W 4J ^— \ 4-1 CO v-' 4-1 J-I 4-i A3 4-1 {X r^k 3 3 D. CO cfl 3 *~\ cfl 3 cfl CO 3 4J s— " 3 CN 3 3 O cfl 3 o CT* iH 3 cfl -H 4-> tfl r^ 4J cfl 4-1 3 4J H 4-1 •• r^ 4J TJ g 4-1 4-1 4-1 rH 4-1 H VO 4-» r3 O CO 4J vJD CO 4-) CO o CO O s CO 3 -, LA •• 1 •• £3. • •cfl •* Q o 3 o •* rJ rJ 3 rJ -3 n rJ rJ r4 !-4 rH rJ 3 i-J o rJ M 4-1 1 II cj J-I CO M H M CO M cfl rH M TJ 3 M 3 1 M Cfl H J-i M O >-l M at en Nl CD IS] rH N u I N 3 O N B o CN t{ CN 4J CN 4-1 3 CN a) o o o O PS PS O 2 o < U K 3 o o pi r-i rJ pq -3- LO \o r^ oo o\ o 42 Samples from predominantly forested areas, the preferred habitat of ithomiines, have a higher percentage of Ithomiidae than do samples which include cerrado or grasslands. Likewise, the western and southern Ama- zonian forest areas have the highest percentages of ithomiines, no doubt because these areas, which correspond to the Napo and Rondonia Refuges (Brown, 1976a), were principal centers of geohistorical evolution in the subfamily (Brown, 1972) . The above discussion considers only the number of species present in a given area, and says nothing of the numbers of individuals. Unfor- tunately, estimates of absolute population size of neotropical butterfly populations in tropical forest areas are practically non-existent. The low density of most species' populations limits the use of the insect ecologist's favorite technique, the capture-mark-recapture sampling se- ries. Until alternative techniques of population size estimation are available, we must be content with estimates of relative abundance, al- though these are usually severely biased by the investigator's collecting technique, preference for certain taxa, and a tendency to concentrate on unusual or showy species. While techniques to determine the number of butterfly species present in tropical areas have been recently refined (Brown, 1972), little attention has been given to the estimation of abso- lute population sizes of tropical insects (but see Elton, 1973). Microhabitat Distribution Although all of the species of butterflies which occur at Limon- cocha are, by definition, inhabitants of lowland tropical rainforest, they differ widely in their preferred microhabitats . Species which are A3 characteristically found in open fields, and along roads and rivers may be called campestral, while those found predominantly in the forest can be termed sylvestral. Some species are sun-loving (heliophilic) , others are shade-loving (scotophilic) .* Various levels within mature forest, from the lower understory to the upper canopy, contain different, though overlapping, groupings of species. Of concern to the present work are those species which share the same microhabitats with members of the Ithomiidae, and especially those involved in mimetic relationships with ithdmiines. The characteristic habitat, or preferendum, of each butterfly family at Limoncocha may be briefly summarized as follows (all Limoncocha ithomiines and many mimetic species of other families are illustrated in Figures 58-65). Papilionidae. Most of the larger non-mimetic species are helio- philic, usually flying above the canopy, and thus are most often seen in clearings and along open trails (Papilio) or rivers (Graphium) . Mimetic species (Parides, Battus) appear to be rarer in number and are usually found in the mid to upper levels of the forest or along shaded trails. Pieridae. Most of the Pierinae (except the mimetic genera Archonias, Itaballia, and Perrhybris) and the Coliadinae (except for the fragile Leucidea brephos) are heliophilic and campestral, commonly flying along rivers and open trails or in clearings. The mimetic genera of the Pierinae are chiefly sylvestral, but often wander along the edges of clearings or shaded trails. The six Limoncocha species of Dismorphinae are all mimetic and rarely stray from the forest interior. D^. pinthaeus, I), theugenis, IK erythroe, and I), leuconoe are restricted to the lower levels of the understory, while ID. orise and IK amphione frequent the mid and upper canopy, respectively. 44 Ithomlidae. Most species are scotophilic and tightly restricted to the forest understory, although a few are found in the upper levels of the forest. Members of a few genera frequent the sunny edges of clear- ings (e.g., Ceratinia, Mechanitis) while others (e.g., Melinaea, Tithorea) often fly in the sunlit upper canopy. Danaidae. The monarch, Danaus plexippus, is campestral, especially common over pastures and large clearings. The mimetic I^ycorea ceres is sylvestral and heliophilic, visiting flowers of the upper canopy and forest edges. The clearwing mimic, Ituna iamirus, is a deep forest dweller. Satyridae. Most species are dark and cryptic in coloration, scoto- philic, and occupy only the lower levels of deep forest understory, al- though some Euptychia are found in second growth or grassy areas as well. The Brassolinae are crepuscular, often flying in banana groves or along open trails. Mimicry is rare. Nymphalidae. This large family is quite diverse both morphologically and behaviorally . Although most non-mimetic species are heliophilic to some extent and are often found in the sunlit upper canopy, along open trails and rivers, and in clearings, others, e.g. Nessaea regina, are sylvestral. Batesian mimics usually occupy the same microhabitats as their models, e.g., Phyciodes actinote is found in the forest interior, and Anaea (Consul) fabius is found in the upper canopy. Lycaenidae. This is a large family of generally quite small butter- flies, most of which are sylvestral, frequenting sunlit patches in the mid to upper levels of deep forest. Many species are rare, difficult to see, and harder to catch. Mimicry is uncommon. 45 Riodinidae. Another family of generally small butterflies, the metalmarks are quite similar ecologically to the Lycaenidae. Mimi- cry is not uncommon, but the mimetic relationships are for the most part poorly understood as they rarely involve classical models. Both species richness and heterogeneity of the Riodinidae is higher than that of the Lycaenidae at Limoncocha, and the range of wing coloration and patterning is astonishing. Hesperiidae (Hesperioidea) . In temperate areas skippers are charac- teristically campestral, associated with open fields and second growth areas, but in the tropics there are a great many sylvestral species. Some are brightly colored blue or red, but most are of a yellow, brown, or dull orange. Mimicry is rare. The mimetic associations present at Limoncocha, with special ref- erence to the Ithomiinae, will be discussed in Chapter 7. The Butterfly Subfamily Ithomiinae The Ithomiinae have been, and are, largely neglected by students of Exotic Lepidoptera, and one hears as the reason that they are less beautiful than other groups and more difficult. The first answer I will ask you to dismiss from your minds, for it cannot be a well-considered state- ment by any who make it. As for the latter — the one of difficulty — there is certainly more truth in it, but it is not nearly so great as is often averred, and what difficulty there is should stimulate us to fresh endeavours to over- come it . — W. J. Kaye (1914) The stimulation to fresh endeavours has been long in coming. During the first fifty years after Professor Kaye delivered his challenge before a meeting of the South London Entomological and Natural History 46 Society in 1913, the Ithomiinae were all but ignored except by a very few workers, mostly taxonomists. Only in the last ten years, and par- ticularly in the first half of the present decade, have ecological and behavioral studies been attempted. Since ithomiines are among the most numerous and conspicuous animals in a rapidly disappearing biome, the neotropical rainforest, such attention is long overdue. The butterfly family Ithomiidae comprises two geographically dis- junct subfamilies, the Tellervinae and the Ithomiinae. Containing only the single monotypic genus Tellervo, the Tellervinae are restricted to the insular belt of tropical forests from the Cape York peninsula of Australia north through New Guinea and the Bismark Archipelago, west to Celebes, and east to the Soloman Islands (Common and Waterhouse, 1972). The Ithomiinae are entirely neotropical, ranging from sea level to 3000 meters elevation in the tropical and subtropical portions of Central and South America. Only two species occur in the Antilles, one on Cuba and another on Jamaica and Hispanola (Fox, 1963). The Ithomiinae have had a nomadic taxonomic history (reviewed in Fox, 1956). Over the past two hundred years they have been shuffled among various family groups, including the Heliconiidae, Nymphalidae, and Danaidae. For most of this century both the Ithomiinae and Teller- vinae have been considered part of the Danaidae because that was the arrangement in Seitz's widely distributed The Macrolepidoptera of the World (Seitz, 1924). After considerable morphological and systematic study, Fox (1949, 1956) concluded that the Ithomiidae were worthy of familial rank and called into question their classical association with the Danaidae. He considered the Ithomiidae to be primitive among the 47 Nymphaloidea, and most closely related to the Satyridae. Gilbert and Ehrlich (1970), however, have reviewed the evidence of the larval, pupal and adult morphology, the foodplant relationships, and the behavioral characteristics of the Ithomiidae and Danaidae and concluded that the two share a much closer relationship than do the Ithomiidae and the Satyridae. Data collected in this study provide further evidence in support of the Ithomiidae-Danaidae relationship. Since the classic work of Richard Haensch (1903, 1905, 1909), the most active students of the Ithomiinae have' been R. Ferreira d' Almeida (see Brown, F., 1975, for a complete bibliography) in Brazil and W. T. M. Forbes (see Fox, 1956, for bibliography) and Richard M. Fox (see Brown, F., 1968, for a complete bibliography) in the United States. Fox began a monographic revision at the family level which was uncompleted at the time of his death in 1968. Of the eight tribes he recognized in the Ithomiinae (Fox, 1949), he completed the revisions of only the first four: Tithoreini (Fox, 1956), Melinaeini (Fox, 1960), Mechanitini (Fox, 1967) and Napeogenini (Fox and Real, 1971). Gerardo Lamas M. (1973) and Herman Real (personal communication) are presently working on the taxonomy of selected genera of the four remaining tribes (Ithomiini, Oleriini, Dircenini, and Godyridini) . The numbers of genera and species contained in the Ithomiinae are far from being decided with certainty. The recent trend in the evolu- tion of the taxonomic study of this subfamily has been a reduction in the number of recognized species. As the biological relationships of many allopatric forms become better known, many of these forms are com- bined to form wide-ranging, polytypic species, some with a great many 48 named subspecies. Simultaneously with the reduction in the number of species recognized, there has been a gradual increase in the number of recognized genera as the morphological relationships of greater numbers of species began to be more closely studied. For example, Haensch (1909*), in Seitz's The Macrolepidoptera of the World, listed 883 named forms, most of them treated by him as species, occurring among 33 genera. Since that time the number of recognized genera has been increased by about one-third and the number of recognized species reduced by about one-half. Presently, there are 49 genera recognized in the Ithomiinae and, in spite of the steady stream of descriptions of new species of ithomiines over the past five decades, the number of valid species is probably only about 450. In Table 3, the eight tribes of the Ithomiinae are listed, and the number of genera and species estimated to occur in each are presented. This table was prepared from the list of genera assigned to each tribe by Fox (1956) , modified by the subsequent description of new genera in the subfamily (Fox, 1967; Fox and Real, 1971; Brown and d' Almeida, 1970; 4. Brown et al., 1970; Lamas M., 1973). The numbers of species in the first four tribes are taken from Fox's Monograph. The numbers of species in the remaining four tribes were roughly estimated by applying a formula of reduction to the number of forms listed by Haensch (1909) . The * Although Volume V, The American Rhopalocera, of A. Seitz's The Macro- lepidoptera of the World was not published as a unit until 1924, various sections of this work were issued earlier as fascicles. The section on the Danaidae, which included the Ithomiinae, was written by R. Haensch and was first published in 1909. 49 Table 3. The Number of Genera and Species in the Eight Tribes of the Ithomiinae. Tribe Number of Genera Number of Species* Number of Named Forms Ratio Forms /Species Tithoreini 8 18 63 3.5 Melinaeini 1 17 55 3.2 Mechanitini 6 34 76 2.2 Napeogenini 7 104 236 2.3 Ithomiini 3 (28) 64 (2.3) Oleriini 4 (51) 117 (2.3) Dircennini 9 (139) 320 (2.3) Godyridini 11 (54) 125 (2.3) Totals 49 (445) 1056 Source: Data for number of genera from Fox, 1956; Fox, 1967; Fox and Real, 1971; Brown and d'Almeida, 1970; Brown et al., 1970- Lamas M., 1973. Data for number of species in tribes Tithoreini, Melinaeini, Mechanitini, and Napeogenini from Fox, 1956, 1960, 1967- Fox and Real, 1971. Data for number of species in the Ithomiini, Oleriini, Dir- cennini, and Godyridini from Haensch, 1909. Note: See discussion in text. * Numbers in parentheses are estimates. 50 formula was derived as follows. As treated by Fox (1967) the Mechanitini contain 76 named forms divided among 34 species, and the Napeogenini (Fox and Real, 1971) contain 236 named forms divided among 104 species. The ratio of forms to species for the two tribes is thus 2.2 and 2.3 respectively. Applying the figure 2.3 to the number of forms given by Haensch (1909) for each of the four tribes not treated by Fox, we arrive at an approximate number of species per tribe. (The ratios for the Tithoreini and Melinaeini are 3.5 and 3.2, but both of these tribes contain relatively primitive genera, each with a relatively few, wide- spread, polytypic species. The ratios of the Mechanitini and the Napeo- genini are more applicable for comparisons to the remaining four tribes.) This gives a total number of species for the Ithomiinae of approximately 445. Thus, the 53 species at Limoncocha represent a little more than 10% of the total number of ithomiine species, but only about 5% of the total number of named forms (which probably exceeds 1100 today) . Of the 49 genera in the subfamily, 24 are represented by species occurring at Limoncocha. Morphologically, members of the Ithomiinae may be characterized as follows. Adults are generally medium-sized butterflies, although wingspan within the subfamily varies from less than 3 cm to over 10 cm. The forewings are generally long and narrow, but not quite so pronounced in these characteristics as are the wings of the Heliconiinae. Wings may be either opaque (usually some combination of black, brown, orange or yellow) or translucent or transparent, in which case the wing scales are reduced to tiny hairs in the clear areas. Antennae are long (be- tween one half and two-thirds the length of the forewing) and thin, with 51 a slender, unsealed club. The abdomen is quite long and slender, pro- jecting well beyond the anal margin of the hindwing. All males possess an androconial scent patch located on the dorsal costal margin of the hindwing. The scent patch is accompanied by an androconial brush con- sisting of a cluster of hair-like scales attached along the posterior mar- gin of the scent patch. Both the scent patch and the androconial brush are normally covered by the overlapping forewing. The androconial scent patch and brush primitively extend the length of the discal cell. They may be separated, however, into a distal portion and a proximal portion so that some species have two androconial patches. Either the distal or proximal androconia may be absent in other species. This costal scent patch is diagnostic of the Ithomiinae (it is not present in the Tellervinae) and, because of its variation, is a useful taxonomic charac- ter at the generic level. With the exception of the genus Thyridia, fe- males of the Ithomiinae do not have the androconial patch and brush. The neuration of the hindwing, another generically useful taxonomic character, is nearly always different in the sexes, often markedly so. A discussion of the ecological requirements and characteristics of the adults, and descriptions of the early stages of the Ithomiinae will be presented in Chapters 3 and 5, respectively. A list of the 53 species of Ithomiinae collected at Limoncocha dur- ing this study is presented in Table 4. The species are arranged by tribe and a subjective estimate of the relative abundance of each at Limoncocha (based on thirteen months' observations) is given in the table. The number of species occurring in each of the five categories of abun- dance is presented graphically in Figure 20. It is obvious from this 52 graph that the majority of species at Limoncocha are uncommon or rare. In fact, only about 40% of the species are abundant or common. About 15% of the species are very rare (i.e., no more than 5 specimens were collected in 13 months) which suggests that, in an ecological sense, the effective number of species at Limoncocha may be somewhat less than 53. For example, a single male of Oleria tigilla was collected during a three week visit to Limoncocha in September 1972. This" species was therefore expected during the 1973-1974 study period, but, in fact, not a single Limoncocha specimen was taken during this time. During the course of this study it was possible to make a few brief collecting trips to other lowland forest areas in the Oriente. One- day collections were made at Dureno (275 m) , on the Rio Aguarico, and at Veracruz (900 m) , just east of Puyo, and several days were spent at Shell (1065 m) , on the Rio Pastaza, and at Tiwaeno (410 m) , on the Rio Tiwaeno (see map, Figure 1). These supplementary collections re- vealed that several of the species that fall into the "very rare" category at Limoncocha are much more common in other areas of eastern Ecuador (e.g., Oleria tigilla, was found at both Dureno and Tiwaeno). Such species are so indicated in Table 4. Thus it may be concluded that Limoncocha occurs at, or just beyond, the normal distribution of these species at this time. It must be remembered, however, that such species may have been breeding residents of the Limoncocha area in the past, and may be so again in the future (see discussion in previous section of this chapter). 53 60 Si" u ox Ix u no c c £3* n o •o CO 3 u X u o u c i •o 01 o u w z X 01 « 3 Id «J 01 • X u ^ O fci o « U-.IM o « 0) R B CTv c « 01 4J B B i-l O ^ u B o 01 rH O.B ■ o • u 0) M u II «t o. 01 • B o — 6 C •» B 0) 01 B 0) « .* 01 E •X 3 01 •* 5 a~i 4J ■ o *>H & 4J 8:* 8: o CM & 8: CM 8: CM o *>4 rH « CO •a « JC B E 1 to •o id s c iH «0 > 9 f-c •H en e 0) n X u 0 U a id id c E co O cj E 41 a fa X o s X 6 id > CO C a e •a o o E O 3 «-» 0) E as o •rf 3 3 -< 01 i-l X X en 3 E > 01 CO •a c O X X m en o tJ 3 B 00 i-l 01 id X E c o. o o o X CD E u CO 01 o b H B X o co E 01 a x x al X u CO X 01 o id ■o > u 3 > 0) ^01 3 o E O 3 01 ^ B Z E 2 <2 3 s CO a a (1 01 41 IS CO a E E E ■-4 t-t t-d 01 01 01 X 2: 33 >-. o u. 54 M e~ ■•+ E o 3 E h OX. c c * CI u c a •a e 3 < u CJ Vj x: oj o XI 3 6 O 3 >Z 3 C CJ ** XI XI -H a u o g — a 2 00 O oo*o (M 6* N .o O n B 41 tJ WCM u a f-t cu Z coZ ZZZZZZZZZZ ootnmmotrtmoo r^r»ovofnooo*i-*vom CMCMCMCIdCMCMCMCMCM -a- -a o oi vo in a* o •-» oo co CM CM CM CM o CM CM o o o o o •J- CM CM CM w-i !-• f-» a •^ i-4 c u o a a It s cn m TO a #-« 3 f* -a > > SN >i !« >. 01 0) 01 41 01 X. X £ X x: °< CK ^ Ol 0£ AJ u kj i_i w o1 o O 0 O a o O o o 01 Ol Ol OJ 01 a & a. a c- tx a. o a >, >N >i >i X Z •H U u 01 rH O. 0) 3 tn -o M a- C — a 1 O a. CM O a u r- C « D. 0) « v. 01 co 3 -a- o . i rH ■ I Z z z SB Z z z Z O -H •H 6 H o Z o Z vZ z5 o CO o CM O H CO CM o CM U1 CM CO o iH CO g. o m r-1 o o rH r-l CM oo CO CO r-l o r-l CO r-l CO in •3- m o >J- o o o o o O O o o o m ■n m m tn m m m in m •o a> 3 C 10 H a. c to 91 X Ei I >-l o c ID 4J -H o to u 3 -r-l CJ X 01 X x co u ■H en CO co C -r-l o 01 a. c a co oc X r-l a ■o PC -8 X 3 A CO r-l r-l a ■H n h CO CO M g 0) CO •a CO CO CO r-l •H -r-l •r-l z e E E r-l o O O r-l X X X rC 4-1 ij 4-1 o r-l M l-l 3 a X CO AJ o u W 09 -H -r-l C 5 In 01 u CO a X CO X n^ a ^ — - CO CO a| *-> i_> ■° o w •Hi r-l M COl CO cs c l-l ot o 0 CO r^| f— 1 u 01 §> o> rS 01 a. o o 01 o. X CO 3 r-l a u o g CM o o o o CM 9* cm eg U 0) 8 CM 8: o o* o CM 8 o •-4 f-H 8: o f-H f-i i— i Q. f-t »» rH t-l r- ■J- 1 f-H 00 • * 00 O £* % \ .. | t ft * •- c \ * X * fc CM \> CM CM CM CM CM CM CM CM vO m t-H tH r- 1 f-H r-H co 3 r-4 f-H f-H rH t-H 1 I 1 1 I 1 1 1 ■ ■ I I 1 2 Z z Z z z z z Z cd z z z z z CI • Wt m io cm «n O i-l en en cm cm o CM CM o un u ox oo c c eg oj OJ \> o o o o o en f-H o r^ f-H en cm o f-4 cn o 00 CM rH •* CM -M* c a •o c 3 x < a % u 0) U X Z o o •» f-H cn un o o O -3- O O en *o cn cn o o •o 3 C x u CO X u 0) C 01 CO X •O c a rH X X fH eg in H H o f-H a fH eg c B e c at o CJ M 1H 01 -H u a Cu o c c 3 T3 C 01 •o f-H 01 (X PS -9 Z t-t o ►H a. >• a 8 X o CO c 01 eg x eg ca r-V 01 "> C eg X X u CO C 01 eg X C "-^ o S -H 0> 4-t X eg c I c > a] e o o 01 X u n) B a! 57 •o oi 3 6 C o o 01 H o c s c eg o 6 01 2 01 01 N 6 -H o c ■H 00 u O eg u O V o u 41 u u o eg m o r-l 6 <0 V T3 U C 3 3 00 X> o CO eg «i • n c« . w CO bOT3 C C -i-l -r4 1-1 00 -O U 01 B 0 u a 01 o tH 00 1-1 r-l B H O-H 3 OTJ 09 01 - 0) S ■a h o eo <« -O eg 00 01 SsSs 01 t-l 0) U 01 |4 £uu U CO 00 C-H *-. 1-1 > O Ml 01 10 tj oi ao Vl 3 01 u V >^o U C a.i-1 o m C >,«« 6 0 eg o - 01 •o u e 6 U jC-rt eg eg |4 oo o £ eg 3 01 t> >h a. eg a. O CJ oo 01 eg U u Ssm O Vi SsiH B eg^-i ij c a. ai eg eg >■. u .-i jr f-t CO 3 3 u .-1 >>01 00 eg eg oi 01 93UIOI OrlOhll 3<2H,J o eg oi c 01 oo B » 01 O «J .. C c^. o eg c S •o o 01 u eg ■ oi §E O 01 >, B o U U XI O C eg 01 01 01 CO SI 5 >en:? eg -O a o n X-l ■o 0) u 3 eo eg 01 00 s 41 00 B T-l » 41 o «H C eo 0) S ■a o o u ■o 01 tl 3 CO CO 41 B oi o. o u 01 JO C ■ 58 20r •J rev- •' ;. _v •' • Is? ■ :'■' - ' '.$ > ■ ■ ^ - S A -i ■ i : 1 "? "* s CUR CATEGORY OF ABUNDANCE VR Figure 20. Relative Abundances of Species of Ithomiinae at Limoncocha, Ecuador. A = Abundant C = Common U = Uncommon R = Rare VR = Very rare 59 The Plant Family Solanaceae The Solanaceae, or nightshade family, comprises some 80 genera of diverse form (D'Arcy, 1973) and contains approximately 2300 species * (Cronquist, 1968). Although the family is strongly centered in tropical America, there are many species of a few genera in Africa, several iso- lated genera in Australia, and at least a few species in most parts of the world. Endemism is greatest in South America, where there are 38 endemic genera (Lawrence, 1951), of which 14 occur in the temperate portion (D'Arcy, 1973). The Solanaceae have long been included in the Tubiflorae, a large order of some 25 families containing a great number of species (Cronquist, 1968). Recent taxonomists usually subdivide this unwieldy taxon into three orders, the Polemoniales, the Lamiales, and the Scrophulariales. Depending on the cut of the taxonomist's knife, the Solanaceae are usually included in the Polemoniales (Lawrence, 1951; Cronquist, 1968) or the Scrophulariales (Takhtajan, 1969). Floras treating the Solanaceae of tropical America have been pub- lished for most of Central America, including Costa Rica (Standley and Morton, 1937), Guatemala (Gentry and Standley, 1974), and Panama (D'Arcy, 1973), but only one country in South America, Peru (MacBride, 1962; Correll, 1967), has been so treated. Unfortunately there is no Flora of Ecuador, the site of my research, and thus I have used the publications of D'Arcy (1973) and MacBride (1962) to arrive at generic determinations in the field. Dr. D'Arcy has been kind enough to supply the specific determinations of the Ecuadorian solanaceous specimens I have sent to him. There exist a few publications containing information on some of the cultivated species of Solanaceae from the Andean slopes of Columbia, 60 Ecuador, and Peru, most notably Lawrence (1960), Schultes and Romero- Castaneda (1962), Patino (1962), Jimenez (1966), Pacheco and Jimenez (1968), and Heiser (1968a, b, 1971). D'Arcy (1974) provides a bibliography of some of the more recent literature on Neotropical Solanaceae. The family is quite diverse morphologically, and includes herbs, shrubs, trees, lianas, and epiphytes. The trunks, petioles, and midribs of Solanaceous plants are often armed with acicular (needle-like) or recurved (hook-like) spines, the latter type being especially common on lianous forms. The leaves may be glabrous but more often are pubes- cent with simple, dendritic, stellate, scutellate, and/or glandular hairs. Stellate hairs in particular are well developed in the family and quite diverse in form. Some species have tough, coriaceous (leathery) leaves while others have leaves with a waxy cuticle. In short, the family employs a variety of mechanical defenses against herbivory that are developed to a greater or lesser degree among the various genera and species. In addition, secondary plant substances in the form of alkaloids are widely distributed in the Solanaceae (Henry, 1949; Raffauf, 1970; Willaman and Schubert, 1961) and there has been increasing recognition of their possible role as a chemical line of defense against herbivory (Fraenkel, 1959; Ehrlich and Raven, 1964). Many members of the Solanaceae are of considerable economic impor- tance to man. Indeed, Uphof (1968) lists 92 speceis from 21 genera as possessing value as agricultural crops, as sources of pharmaceutical chemicals, or as ornamental plants. The fascinating history of man's 61 encounters with many of these plants has been delightfully related by Charles Helser (1969) . Agricultural crops include the common tomato (Lycopersicon esculentum) , the cultivated and Andean potatoes (Solanum tuberosum) , peppers (Capsicum, several species) , tree-tomato (Cyphomandra crassicaulis) , naranj ilia or lulo (Solanum quitoense) , and eggplant (Solanum melongena) , to name a few. Approximately 151 different alkaloids had been isolated from solanaceous plants by 1968 (Raffauf, 1970), many of which have strong narcotic or other physiological properties. Accord- ingly, a number of solanaceous plants and their alkaloids have medicinal value and thus are economically important to the pharmaceutical industry. For example, belladonna (Atropa belladona) and Jimsen weed (Datura stra- monium) contain the alkaloids atropine and scopolamine, respectively. Other drug-producing solanaceous plants are grown for non-medicinal pur- poses, e.g., the various species of tobacco (Nicotiana) which are important cash crops in many parts of the world. Ornamentals come from many genera in the Solanaceae, especially Petunia, Salpiglossis, Schizanthus, Lycium, Solanum, Streptosolen, Cestrum, Datura, Solandra, Browallia, Nierembergia, and Brunf elsia (Lawrence, 1951). 4. Of interest to the present study is the fact that nearly all reported foodplants of the larvae of neotropical ithomiine butterflies (Ithomiinae) are in the Solanaceae. This familial-level correlation between the Ithomiinae and the Solanaceae, and the concomitant phytochemical ecol- ogy of alkaloid metabolism, have implications to insect-plant coevo- lution that will be discussed in Chapter 6. With the addition of the information contained in the present study, the number of solanaceous genera on which ithomiine larvae have been reported to feed is raised 62 to eleven. These eleven genera can be briefly characterized as follows (information from D'Arcy, 1973, and Uphof, 1968, unless otherwise noted). Brunfelsia. This is a tropical American genus of 25-30 non- herbacious species, mostly shrubs, with leaves simple, entire, coriaceous to chartaceous, and often glabrate. Several species, e.g., B. calycina var. floribunda, are widely cultivated as ornamentals. In Spanish they are usually known as "galan de noche." The Manaca Raintree, B^. hopeana, contains the very poisonous alkaloid manacine, which resembles strychnine in physiological effect. In Brazil, the dried root of this plant is used in the treatment of rheumatism and syphilis. Capsicum. The pepper genus contains perhaps a dozen species, most of which (and especially C_. annuum and C_. frutescens) are cultivated throughout the world for their culinary and ornamental fruits. Peppers are annual or short-lived perennial herbs, with leaves simple, entire or weakly toothed, glabrous or pubescent with simple, sometimes glandular hairs. Cestrum. This is a tropical American genus containing 150-250 species of unarmed shrubs or trees, with leaves simple, entire, mostly glabrous above and variously pubescent beneath, usually with simple or dendritic hairs. Several species are widely cultivated as ornamentals, and a few are used for medicinal purposes on a local basis. One species, C. lanatum, is placed in hens' nests by Mexicans to keep away parasitic arthropods. Cyphomandra. Occurring mainly in montane South America, Cypho- mandra contains about 50-60 species, some of which are used by man for food. The plants are unarmed shrubs, trees, or vines, leaves simple 63 or compound, entire or lobed, usually glabrous above and pubescent or puberulent below. The tree tomato, C_. crassicaulis (=C. betacea ) , has been cultivated for centuries by Peruvian Indians, and is presently grown for its edible fruit in Australia, New Zealand, and in some parts of South America. C_. hartwegii is occasionally cultivated for its edible fruit in Chile, Argentina, Colombia, and Panama, and is known in the latter country as "contra gallinazo" (against chicken-heartedness) Datura. This infamous genus contains about ten species of unarmed, ephemeral or perennial herbs or shrubs, sometimes with woody trunks, occasionally fetid, leaves simple and entire, repand or variously pin- nately lobed or toothed, pubescent with mostly simple, sometimes viscid, hairs. Species of Datura are found in warm temperature regions in both the New and Old Worlds, but Mexico, with eight native species, is prob- ably the center of speciation. The potent alkaloids found in this genus, such as atropine, hyoscyamine, and scopolamine, impart to these plants a variety of drug-related uses, ranging from the treatment of asthma to criminal poisoning. Alkaloids occur in several parts of the plants, but especially in the seeds and leaves. Juanulloa. This is a little-known genus of about ten species of unarmed shrubs, mostly growing as hemi-epiphytes high in the forest canopy. The leaves are simple, entire, and glabrous to tomentose with dendritic hairs. A few species are cultivated as novelties in temperate greenhouses. Lycianthes. Closely related to Solanum and often confused with it, Lycianthes contains about 200 species, mostly from tropical America. The genus includes herbs, shrubs, and vines, usually unarmed, the leaves 64 simple, entire or nearly so, glabrous or pubescent with various types of hairs. One or two species are occasionally grown as ornamentals. Lycopersicon. Also closely related to -Solatium," LycopersiCon is a genus of six species and several varieties centered in the coastal region of western South America. The plants are unarmed, sprawling, ephemeral or perennial herbs with pinnately lobed leaves, pubescent with glandular, aromatic hairs. The cultivated tomato, L. esculentum, is an important vegetable crop in many countries. Physalis. This genus contains about 90 species, with the greatest number in Mexico, some in Central and South America, and a few in the Old World. These unarmed plants are mostly herbs, rarely shrubs, with leaves simple, entire, shallowly toothed or lobed, pubescent with sim- ple, often glandular or viscid, rarely dendritic hairs. Commonly called ground cherries, several species are cultivated for their fruits which are eaten raw or cooked, or made into preserves (e.g., P_. ixocarpa) . P. alkekengi is cultivated both as an ornamental and for medicinal purposes. Solanum. Solanum is one of the largest genera in the plant kingdom. Over 3,500 species have been described, but probably only about 1,400 of these are sound (D'Arcy, 1974). A recent study (D'Arcy, 1972) rec- ognizes 7 subgenera and 50 sections. The morphological diversity of this genus is striking, and includes herbs, shrubs, trees, and lianas, with epiphytic, procumbent, and tuber-bearing species all represented. Plants may be armed or not, glabrous or pubescent with a variety of simple, branched, stellate, or peltate hairs, these often glandular, sometimes accompanied by bristles. Leaves may be simple or compound, 65 entire, toothed, or variously lobed, sometimes armed. Almost worldwide in distribution, the genus has its greatest number of species in tropical America, but is well represented in temperate America and in Africa. Hybridization and polyploidy are widespread (Heiser, 1969). There are a great many S planum species of economic importance to man, including agri- cultural crops, medicinal plants, ornamentals, and. noxious weeds. In terms of acreage, tonnage production, market value, and dietary impor- tance, Solanum tuberosum, the cultivated potato, is one of man's most important dicotyledonous crop plants. Other important food plants in this genus include the eggplant (S . melongena) , cultivated in various parts of the world, and the lulo, or naranjilla (S. quitoense) , of Colom- bia and Ecuador. Ornamentals include S^. wendlandii, S_. seaforthianum, S. pseudocapsicum, and many others. jS. torvum is cultivated in eastern Europe for its steroid alkaloids. The fruits of S. mammosum and other members of the section Acanthophora (subgenus Leptostemonum) are used in Ecuador and elsewhere for killing roaches, exterminating rats (MacBride, 1962), and for a variety of medicinal purposes. They are perhaps poi- i sonous to mammals, but their armament is sufficient to discourage most grazing animals . Wither ingia. This genus has only recently been clarified as a taxonomic entity (Hunziker, 1969) and includes about 18 species ranging from Mexico to northern South America and perhaps one species in the South Seas. These plants are unarmed, erect herbs or subshrubs, with leaves simple, entire or sinuate, dentate, membranaceous, with simple or dendritic hairs. They have no known economic importance to man. 66 The examples given above of local uses of solanaceous plants for medicinal and insecticidal purposes attest to the potent nature and widespread occurrence of alkaloids in the family. During the course of this study, forty-two different plants at Limoncocha were recognized as belonging to the Solanaceae. There are undoubtedly many more Solanaceae present in the area, especially trees and lianas of the upper forest canopy, but recognition of solanaceous plants not in flower or fruit, especially those occurring in the canopy, is quite difficult. A list of these forty-two species, along with the voucher number, estimated abundance, habit, and typical habitat of each one, is given in Table 5. Although there are eleven genera represented, nearly two-thirds (64.3%) of the forty-two species are in the genus Solanum. No other genus has over three species. The eleven genera, and the number of species in each, are: Brunf elsia (1), Capsicum (2), Cestrum (2), Cyphomandra (2), Jaltomata (1), Juanulloa (1), Lycianthes (2-3), Lycopersicon (1), Physalis (2), Solanum (27), Witheringia (1). Grouped by growth form, the Limoncocha species include eighteen 4. herbs, fourteen shrubs, two trees, and eight lianas. Two of the lianas are epiphytic. In regard to habitat, nineteen species are characteris- tically found in primary forest and an equal number are characteristically associated with disturbed areas or secondary growth. Four species are entirely domestic. Grouped according to light tolerance, nineteen are found growing in full or partial sun, fourteen occur in areas of open shade, and only nine are found in deeply shaded areas of the forest interior. 67 u o ■o 3 o Id ■ X u o u c o 6 •O 01 s ■0 0> «J o o 01 a- to •o OI w u u C a a B-o -H c aj 3 m xj u < Vi «l w x oi O -O 9 B O 3 > X -o •• • 01 «l AJ •o K -o 09 09 ex AJ 09 W X o »>. 00 X U 09 a B. AJ •- 01 01 Uj O >- 9 O V 5S OU4 « 60 ■ a C ■ C ■o c 09 09 A. >\ K oo m V 09 &N ■ o 60 M a o r-l r-l o. ci vi Xr-I C 01 • ■H « A. OI -4 ►» 09 A. o 60 CO -o-o to 09 •rl At 00 01 r4 01 >N •rl —4 ■H 09 O oi 01 c C - 09 3 OI U a. AJ AJ AJ-0 AJ 09 o « 09 01 > m m c c . u c ■ • (1 a. 60 r-l — AJ o e 4) 0) • oj o AJ A. AJ OJ o c (1 AJ 00 ox -rl e-u X 09 X 09 ■rA X 0) AJ M Ml Ml ■) oo M O A. >- 60 A. — * o. AJ AJ o « ■HX -J.-0 AJ 09 AJ •rl O AJ O JB X AJ — 1 AJ r-l C 01 X r-l AM X A. • - • eo oo ■o c » §" §11 60 c 60 60 01 >■< —4 —4 oi c >1 3 O •rl 3t> •rl T3 A. r-l • r-4 ■ -O 09 u 00 w 00 09 09 r-l r-l -O a) o c o c o 09 o 60 60 AM C 09 SI x AJ 3 T* 3-1 X rKAJ r-l -* c r-l O • 3 r-l B a ■o 09 AJ 00 AJ 09 1- (0 o-o a -h <0 OO 09 09 -rl 09 O »- B0 a 09 u •H C -4 A. -H 09r-l •H U 01 O. CJ t—I 0) r-l 01 C E 41 AJ O AJ CO AJ 01 -4 r-l aj a. r-l • ■1 -a r-4 B «-j B 01 t4"0 u u Al 0) v.- eo a r-l U r-l 01 § 3 O 3 O O.A. c 09 01 09 r-l «--a A. 3 09 AJ 09 09 c •rl B « c ■o c AJ 09 01 • AJ 01 •x €9 •rl •o •o C m 0) II HX r-l — ■ c 41 X Mi s u u r-l r-l - B-O 01 u X • 00 BO. a? « u 09 >> 2 CM • - ■ • -3-X 09 • • 1— 1 a •o •o 09 C 09 -rl U 09 -O -1 X o-o - 01 • 01 -rl TJ o e 01 — > 00 - AJ 41 B x BX Aa-O o c AJ 09 -X s a • "O £ u u O C AJ 01 A. B t A. o •> C OS 09 it t. X X A. 09 • • A- r-l OX AJ ■ X C XX 0 O X ■o X - •o X W AJ o c -1 O. a to AJ >, AJ !►> 09 - 01 ■> B 01 09 AJ >. AJ 01 r-l-rl ^ r-l -H EX X 09 -r-l X a. Ml >s.-i — •-4 • • -4 >« o SNP4 X > -o •» c -o-^ c -O 01 3X X 3 TJ -rl e o 00 >-. A3 A. •8 o eg o a o-o U 60 u o O B a o --• 0) 01 01 01 o o A. o o A4 o o X —1 01 01 o oi •rl SX XX XX St AJX St ajx 3 B ID r-l x-o •XX r-l ATI ATI •» CO -a •» M O | tJ AJ 09 CO M c -4 01 C o 01 CO a) 01 1-4 AJ V-* 3 U • ah Am 1 <) AJ ■a u c a 01 X a oi t4 A. •o -o c c a a e e o o X X a o >■ >■ CJ CJ >s H C 01 -1 > o) u o A- °H a. 09 01 i-» o X 1 --I AJ rH c o 3 ■ i-l C -rl f-« 09 C « 3 > •-> <~> r-l 68 TJ at ■ o 4-1 c eg a §■0 •H c u s CD X UJ < I-l u u X 1 CI X 3 b <) - > ii IS H 4J CO TJ X a. i oo •H c >s 13 oo CO M c o CO a. C •H C 01 01 1-t C o •a > CO r-t 01 -a h CO CO CO iH c C u c a » c CO 01 r-l r-l •M 0 3 , eo CO DO w-i •a co to f-4 14 .-i c tj c ■ o 01 u r-t >> —1 oi to r-t O a i-l r-l C 14 4J 4J 4-1 o. B) M a 4J >> w 01 4-1 CO t, U 3 C 01 a o L. L, c r-l V4 00 r-l o CO • CO c a. u to ox: CO to OJ (J eo C U <4-t Ql CO 3 o 1 c u Q.TJ O.T4 00r-l r-l CO T> • o - - 3 c • B • - M •■ CO >S T3-rt eg •i •w 4J r-1 4J u O o o -u JJ O CO 4J 00 U 01 CO i-l -0 4J a CO eg co X M 4-1 o X X • 4-1 01 X c S u 4J C X 01 0) 00 OO Of 00 60 CI r-l 00 -W •-4J rl CO 00 • r. r-l •- H •r-l OJ CO 1-4 1-1 r-l ai u •H r-t ■H 01 eg i-t OI O 01 o r-l TJ TJ r-t t-l i-l tj r-t r-t M •a >, CL CO r-l CO TJI4-I TJ 14-1 c c n tj c o C co CO U-l C 01 CL co c DO C 00 eg eg 3 O x oi 3-H 3 V« x o • 3 3 x c o c 3 c X >. X >» in O CO o CO 4J GO 4-1 CO CO CO TJ • - ra 3 4J -r4 CO t4 CO u CO U , CO h l-t § cl r-l CO C CO w-t 01 r-l c C tv-i t-l rt •o rt c r-l CO r^ to e r-l CJ > r-l e r-t at 01 00 co r-l CO CO o at *u r-l 01 r-l ai 01 i-l 01 -r< 3 u-t Q.TJ 3 o 3 O. CLTJ M 3 0) X 4-1 D.C 3 r-l 3H D-»- CL r-t U. O O CO b.TJ u. o O OI 4-1 u. C C/l to O co u. o u. o o a. O Ct. •o 00 41 c x 1 1 ■ 1-1 g ■o •o •o •o * c ■ u » c c » 01 c 00 co co 00 o oo 01 - 41 00 E 01 * c C U a •o c o O c ct 60 1-1 -r4 Xl -r4 4J 01 1-1 CO GO 'r-l eg a c TJ TJ tj ex •o co . eg • T> c eg -rl c C S>\ e 01 cj c •o ■o c 9 "O CO 01 3 M CO o o CO o o • CO co co CO X ox to -* c >4S C to •H «M - CO TJ tj 01 u 01 CO eg TJ u •o • CO 01 •- 01 -X - c tj >■. - O h O V4 » 01 >» O 01 T» X E£ B B <0 r-t B 4JX 4J X BX X 4-IX V ' o o h * -r-l u o. u r! B C in c CO 4-1 00 XI B > in - >, " >» u-l c •vl •- c CO to • c t» ilrt X -I ■ to o. X to a -h u r-i 1-. O 01 o >, i-l 01 •-I -o 3 4-1 3 t4 r-t It 01 3 >» c X X X) r-4 X 01 1-tX Wt > X r- X 3 o O o e O 01 o 0 X X CO X f0 O • •> X 4-1 >, 4-> >N u 3 4J 4-1 u • u u CO -r4 co at 4J >N CO eo >> •- r-l r-* o Q u c rH X r* 3 r-l eo — 4J ■ - ■~i .. a •- U ■ a c • • a) >> »s r> i-t o >»»< C Xi X X > Xt TJ XI 01 X 01 X l-l •o - TO - X > c TJ > CO L. 00 I-l en h U TJ r. a. h X O 60 O 00 l-t to eg o a •H a> -h 01 01 at o 01 O 01 01 01 c o c O C 01 01 •H o at ■J 3: .-i X X X 4j X B X M 3: 3 3 -i S t4 X X ►J 3X O r-l O CO •o B eg eg 01 e CO — 1 i4 | c. u o c X 01 to f-l B 3 o CO U-l 01 r-l 01 X -J >-t 01 a o o >- ■-) CO to c 4J 01 to u n-l CO 3 01 04 X C 3 to OJ X Pi X + o <44 X o 01 eg B 3 C CO r-t O V> o IV) o in 69 ■o e CO X >. u a * o a c U rrj 00 u ■a o C rH 0 -a CI T3' 0) i-l °> 6 •a- . a) 4-i ci ar c a co m > o •a > • - ij 01 CD •a E CO -4 X M w a. M M a. 01 >N T3 M to o X t-> to a) c oi two CCC O 3 o. o 01 E o •a x 60 C 60 c 3 C 3 CO 1+ CO U ^ to i-t 01 •a at u «-i c to- to 6TJ •h c u 3 n jo w< M 01 u x " O X 3 E a 3 3 C 01 u- E S" a O S7 u i JO 10 60 c •H •a c • 01 -o CI OI co jr a u c - CO E %4 X >> o-« 4J o i-J .- to J3 U 3 01 M -o X o co E 3. 6a c 1+ c 01 It CO CO • •a - 01 Bx o X > M CO 01 01 3 60 c T+ •a c 01 CO 01 x - cj B c CO CM U X a .n > m to 01 0) X X -C CO 00 01 •m u •-> o C 1M 3 01 >N u w~* CO to B i-t 1-1 u M u a. > co u to O-B 01 —I 01 m a a. 60 c CO • 01 CO eor-i 01 CO M — -u oi •o Ps to c X c to 3 CO C 01 TJ D- C O co H CO 01 0> i-t ■-t CI ■a , ■W C >H C C 3 3 to CO h 60 CO c 01 ■ > .c O u o s o •- u c oo 3 co -o a -> o CO u i-l 01 U CO ki CO T3 C 01 •o o to to 01 ■O CM CO O X CO >> o. c o oi c a. to o u •o c CO CO t-+ a. co 01 oi •a J- c 01 a. 01 w B - CJ Be ^i to C7i U O • X 60 0 a c U >. il-rl <-i •a •- t4 "CO 01 > J3 01 01 co MM u oi ci a h> X X co ' 60 B ■H •a c 9 u to a] • •a - 4) BX cj • CO 0) °i B O eg XX c U 60 etc 0) -H l-f X^f iJ I ■o c o o CO a S u o J= •- cj X c sa e MM M jr j3 co w C 3 ^ CO •o - c O 60 CO o c »< 3 -" m) •a oi 60 c i4 •o c 01 o CN-T3 0) O -C u U c .. co X M U X ix c X » tC M- a jr oo «^- .- •->■ CO C 00 3 C CO i-l H i-l- CO i-l- «l 3:r-l (M O ■a: c 01 0 CJ-O- U 01 j=; .- u. J3 C 3 * M M J=x to ar. ■x •o: •■ O 60 a c co O tn o o l-f O !-• r-f IS 3 e 3 c (0 c 3" a B CO Ml B 6 3 3 01 T) AJ to 3 CO c C C :oi •W o i-l (J IM CJ •r-* >M 01 3 ccr ol or o 1-1 E i s 3 i-t | R •-+ 3 o C CM to 1+ ' o 1-t 01 CO i M CJ 1 01 c B CO col t-c CJ CO r-1 a> M oi e 3 E 3 > c C c co CO CO t-l f-( f-l 0 a o t/l CO en 1-1 i g | B 3 1 § | B 3 3 c P C c c c C C a CO CO CO CO CO CO CO H i-H ft r-l r-t F-< r-* i-l O O o a a o o c l/l col col on) to en. CO col 70 V 3 C a H E 3 | c c a n r-i o o CO to B 8 3 3 C c a CO ai v< pm 3 4-1 CO c a> CO cu u Pi 00 vi cu 3 oo cu 3 3 4-J CO C cu CO 0) Vl Ph 2 O M H <: o CO CU •H rH MH VI CU •u 4-1 3 m cu 3 B o 43 4-1 >4H O 09 CU o u 3 o Vl CO 4-1 CJ 0) 2 vO cu CO H O CO Pi Pi w o s hJ O O H-J a pl. Q W H H w M > H N CO u CO (3, CO tO M N tO u pq cu CO cu o CO Vl cu 4-1 CO < a, CO tO o •H Pi Cfl 4-1 CO o cu cfl cu cfl CO <; N cfl VI PQ cu cfl cu CJ cfl Vi cu 4-> CO cfl o •H Pi to o cu cfl v >1 •H O. 43 J3 Vi cu D. a CU rH O o iH o Vi Vi 4J 4-1 CJ a 4-) to CO cfl iH cu B B Pu X y—\ CN 00 <■ O n CO i*» r^ >— ' Vi O 4-1 tfl <^ 3 W •rl Vl o 4J Cfl a 3 W Vi o 4-1 cfl cxj 3 w Vi o 4-1 cfl 3 a, to S •H Vi o 4-1 tfl a. 3 W a. to •H Vi o 4-) tfl ex ? M CO Vl o 4-1 cfl a 3 W P. 10 Vl o 4J cfl a 3 W O. CO N cfl Vl pq ai cfl CU CJ cfl Vi Vl o cfl 3 O w cu cfl CU o cfl 3 •H 60 Cfl Vl o pq cu 0) 0) cu iH cu 4-1 a 4-1 4-1 4-1 4J 4-1 4J 4-1 4-1 4J •H •H tH •H Vl •H •H •H •H -H •H •pl •iH -H J3 J3 J3 43 3 43 43 43 43 43 43 43 J3 j3 !3 s s Pi & S > 5 & 13 & & & o CO M a, co o Vl 4-1 O •H rH cu VO vO CTi rH 93 3 o w 1 w 3 4-1 0} 3 0) CO Q) !-l vO to CO o Pi -a 3 4-1 CO 3 a) CO a) U PH !>» "3 3 4-) CO 4-1 3 C5\ CTi r-~ •H rH rH cr> P-i ^ rH n "3 n H 4-1 3 o 4-1 * cu rl H -cu *3 a> CU 4«i 43 a 43 CO 3 rH 5 rH •H .3 •H H •H rH > o O C5 Ph vo co co o Pi T3 3 4-> CO 3 cu CO 01 u p-1 T3 3 4J CO C a) co ai U Ph CO cr> oo 3 3 O >> CO cr» 00 s 3 o 25 O M CJ o ►J a Ph o Pi o ►J o CJ Ph CO Pi o l-J cfl cfl cfl CO cfl o u CJ Cfl CJ CJ iH •H •H rH •rl •H J-l w 1-1 Pi Pi h Pi CU H M Pi Pi o o o O o 3 O o o ■3 CJ •3 T> ctf cfl •3 CO N CJ -a "3 ctf cfl cfl •H CO CO 4-1 4-1 cfl 4-1 CU •H cfl CO 4J 4-1 3 * 3 3 co CO 3 CO a X 3 3 CO CO o CU o CJ O o CJ o CU a w w o CJ CU cu cfl CU cd CO cu cu cu cfl cu CU CO CJ CO CU o o cu cfl cu cj CU cu cu CO 52 CJ i CJ cfl cfl cfl cfl 3 3 cfl o cfl 3 -» •H •H •H M u u rH O rH rH 43 x> 43 O o cfl CU H CO O 3 3 3 PQ PQ a u CU "3 a cfl 4-1 O PM Pi Pi Pi 4. Pi 01 e^» c^- cu 0) c 3 (U 0) 00 •H •H 3 •H •H •H cfl •3 00 tH T3 43 43 cfl cfl 43 43 •H JC 43 43 u at •H CU CU S S rH e s S P & & & o u iH >. U "8 •H ■u 3 o CJ VO 43 (0 H s H M W) M > Ph a CO cfl u 42 CO iH 001 e cfl 3 •H • •H 4J Pi a- M CO o o t-i >4H CO 4J a) 4-1 o 3 3 -H M cfl rH 3 M 1) O cfl w H a 43 3 3 CJ e M iH iH 43 3 cfl CO 3 cfl a T3 O- •a 0) cu •H e •H >4H 3 IH •H n •H 4-1 T3 4-1 3 3 3 CU 0) a) •3 -3 -3 •H •H •H 3 3. 3 3 W 3 CO 3 a o 4-1 Ou or CO CO C 0) cfl r-l •H CO rH | o CO C4H »-i iH (0 -3 3 CO oo P. . cfl (0 3 3, •H •H O CO W H 3 4-1 4-> o cfl o O 00 a 43 43 •H •H CJ CJ 4-1 3 >! ►l 3 ai CO CO < o Pi Ph a\ cn m r-. s— l cfl 4-1 4-1 43 3 3 3 rl cfl cfl 3 CU •H rH rH 43 rH Ph P •3 *3 •3 •3 01 CU cu a) •H •H •H •H «4H MH 14H 4-1 •H •H •H •H 4-1 4-1 4-1 4-1 3 3 3 3 0) 0) 0) cu T3 T3 •3 •3 •H •H •H iH 3 3 3, 3 3 3 3 3 94 these flowers and the Oleriini. The proximate basis for the attractiveness of these flowers may be a spectral sensitivity and preference for the color orange in the Oleriini. This is suggested by the fact that H. kezia and the three Oleria species listed above, but no other ithomiines, were frequently found feeding at the ripe orange fruits of a common species of Araceae. These fruits occur about 40 cm above the forest floor, approximately the same height as the orange flowers (7339). Napeogenes pharo, Godyris zavaleta, and Hyposcada kezia were occasionally seen at the small reddish flowers of an understory melastomaceous shrub (Melastomaceae) , and Melinaea menophilus was once observed feeding at the pendant red inflorescence of an unidentified liana. By far the majority of species, and individuals of Limoncocha ithomiines, however, predominantly visit the small white flowers classically associated with the Ithomiinae. There are five terrestrial plant species (there may be others in the canopy) at Limoncocha with small white flowers, all of which are frequently visited by ithomiines. Two of these are perennial woody shrubs known as heliotrope (Boraginaceae) that have small tubular flowers lin- early arranged in a pendant cincinnus. Both species attain a height of about 5 m and are found in open shade or partial sunlight in advanced second growth or along trails in primary forest. Heliotropium I (voucher number 7309) is quite common. Heliotropium II (7351) is a similar species, but the only two specimens known at Limoncocha were found on the base grounds and thus will not be considered in the following discus- sion. There are two species of composites (Asteraceae) , both in the genus Eupatorium. These are "annual" herbs with small, but numerous, 95 dense flower heads. Eupatorium I (7308), a fugitive species that rapidly colonizes sunny clearings, reaches 1.0 to 1.5 m in height and is often found in large stands of up to a few hundred individuals. Eupatorium II (7342) is found only in the open shade of the forest under story, espe- cially along sun-flecked trails. It normally reaches only 50 cm in height and rarely occurs in patches containing more than. a dozen individuals. The fifth species, Mar ant a I (7322) (Marantaceae) , is similar to Eupatorium II in habitat but attains a height of a little over one meter and may occur in large numbers along forest trails. The plant is highly branched and a single individual may simultaneously produce several dozen of the small white to pale lavender flowers. Thus, of the four plant species under consideration, only Eupatorium I is consistently found in open sunny areas that lie outside the normal microhabitat of ithomiines. Full day observations were made on the three forest species of adult food plants and the results are shown in Figures 21, 22, 23, and 24. Figures 21 and 22 present data from two observation days showing the number of ithomiine species and individuals visiting a patch of eight Heliotropium I plants on the Logging Trail at Site 2. Censuses were per- formed at 15-minute intervals from 0545 until 1830 h. A number of obser- vations may be made from these data. First, ithomiines begin feeding shortly after dawn, which involves flying and locating a food source under extremely low light intensity. Although the low visibility of the dusk-like early morning in the forest understory was further reduced by the presence of fog, the first ithomiines (male Godyris zavaleta) arrived each morning about 0605 h. Second, the peak of feeding activity occurs early in the day, from 0630 to about 0930 or 1000 h. Third, the feeding CO Si CU O CO a> o IN o • 4-1 CO u CO a) o > rH o « o t-H 13 3 CU 3 CX O CN u ■u 0) o 4-1 •H •H -H co a) « T3 4-1 3 CO 4-1 CO CU 3 3 hi "3 CO 4-1 4-1 5 3 •H cfl J! H 3 > CO 3 co •H cu •H 4-1 •3 T-l 4J T-4 a n) i-H U r-t < 6 3 ■4-1 rt 3 CX M-l M cu CO 5 u c H 3 3 3 ■H OJ •H rt •H •H -H -a S CO 3 CD (4 -a CO CO CO CO QJ CO •H a) M t-l M fn 3 4-1 > •H cu cu a) o ja •H U 5 & & a) •H 60 TJ 0) o o o 3 4J •H a CXrH r-l .H -H a) i-H •H CO 4-1 y-i M-i «H > g O 4-1 U 0) co J3 1 I 1 |( ID H O • 6 ■ us u 01 u 3 bO ft* 97 X RELATIVE HUMIDITY o o o o o\ oo CO o H to PS w o o w CO O Ph H 0o NI ' o CJ ■a c cd CM a) 33 co *3 4-» 3 CO 4-> CO 4J -U u <: 00 s •H TJ a> cu cu s B o u cu co o 0) 3 •3 co cd .3 r-i w cd co 3 -H •h cd cd 3 U "3 ■u 60 T3 •H 3 > -H <-i U r-\ 3 cd 3 i+-i p. m C 3 C •H -H -H CO CO CO CO 0) M U U 0) CU CU & & & o o o tt-H HH CO CM CM cu u 3 00 99 X RELATIVE HUMIDITY o o o o o\ oo o T- ° . o\ " l-l I ' ° I _ o oo ■ - rH ■ "" O B - ° I I"* I ■" i-i " o ._ o I vO P ■" r-l P "On ._ o m I ■- iH i B -■ o to ._ o •* o ■■ 1-1 H 00 ■■ o Pi ._ o en 3 ■• tH O , £ ■On T -- o ^ CM o "■ .H 1 w -o 1 -- ° CO 1-1 o " H Ph "• x w ■- o ._o J o ■• H [ * r? -■ O ._ o . o\ ■ o ■■ o o 00 - o -■ o _ o r~- ■■ o | - o ; . o vO - o o m o n n n 0o HI *4-l • o 0) rH to CO a) 4-1 x: o 4-1 4-1 13 CO 0) 4-> js •H 4-1 co •H •H & > • CO >> iH r~- 3 )-i CTv cO cO rH P. 3 #> M • CM O CD iH .H t-H U .H 0) >> *o /-N S rH 0) CD 3 >-i CO CO *1 CD > O -a co •H H *3 •» o cu a) en •H T3 i "3 PM 3 o s ^ rH w cO CJ w ai c o t-H CO *H .h •H 3 pq cd •H 3 ,3 a) H H Oh CD cO 00 o 4-1 3 e CO 4-> •H CO ^ c O .3 cO 3 g sc a) co 4-J 0) 4-1 tH cO .c o 3 4-1 4-1 <1) 0 3 1* o ex T) w •U s CO CO 3 ■H o co ,s H 3 > u a> CO cfl CO ■H 0) cu co ■H 4-> "3 X> •H 0) a) 0) o 0) x: ■H O & S s 0) •H O 13 &0T3 0) o o o c 4-1 CO 3 •H 3 3. U g >-i o « 1 f ?■ iTTl n O a> iH O M : J= co cu CO U 4-1 U3 3 -H ■ H o ■H CO ■ 6B 1111 u ro CN CI) M 3 00 101 Z RELATIVE HUMIDITY o o o o cr> oo \ I o in o n cn cn o H w oc! o o CO o w D0 NI *dH3I saioads ho sivnaiAioNi ao sdagwriN avNiHd^owsia 3VNIIK0H1I •u 3 O O CO o co CM C crj 3 CQ 6 o (0 CO o iH pq o CN T3 d) cu > o a 13 3 cfl o cd a 3 CO H 60 3 c 3 c0 4J O ■u •H co P*> ■3 3 4J CO 4-1 to CO w 0) I rH «4-4 ! M o u id C_) Pi Pi S5 o I 60 C •H 4-1 •H co 0+ •H o+ > CO 01 •H rH HH U cu t 4J •u 3 « CU c •H •H s o 2 ,3 4-1 H M-l St o r- ON CO iH 3 Q O Sn •H U 4J CO O 3 cu a rH tfl W rH »-J H o M CJ 3 co •H ■H O CO >n T3 )-i 3 co co r-^ CO W cu M iH U .O w CO cu H to St > c CO U m ffl cO CO CO CO CO on § a> CO •o •o cu a) N N tO td 13 60 CO CO 4J 4-1 CO CJ o tO tO 01 to tO CO tO to CO X> Xi ■H •H •H •H •H CO O o M U 6 B 6 •H *3 -3 CO CO >> >> o o o U 3 3 "3 •3 "0 T3 x: x: xi cu 0) 01 <0 CO o O 4-1 4-1 4-1 rH to to o O a O rH M M O Ph PU CO CO . TJ •o cu cu h U 3 3 -U 4.1 a. a co • to O TJ O 0) CU CU M CO M CO CO CD CO rH rH rH tfl 0) CO 3 u 3 -O • -O -H to • -a ■H > >>T3 c > -H to a) CO •H *0 T3 4-> •3 C O T3 C -H 60 a) 0) •H B rH r* "3 •H rH rJ T3 0) rH O CO 01 ^ Cu a B r« H B S-i CO cO rJ u tO B en a) cu E -a XI CM ■4-4 g £ 14-1 O O 9 3 o c G 4-1 M u c CD rH rH cu cu X cd cO X o g 4-J 4-1 B r-l 3 o o 3 0) SB H H 2 Pu II II II II II Q rZ ^ pi Pi z 2 Z • • B^ cu 4-1 o 2 1 109 heliotrope, also contain pyrrolizidine alkaloids and thus it is inter- esting that, of a morning sample of ithomiines visiting Heliotropium I, about 90% were males (N = 36 for 7 species) . Because of the low visita- tion rates, comparable collections from Maranta I and Eupatorium II are not available, but observations recorded throughout the research period show that both males and females were frequent visitors to these two understory herbs. Another interesting feature of these samples is the high percentage of recaptures of marked and released butterflies (25.3% of 616 released butterflies for all species combined). Recaptured butterflies, of course, represent individuals that visited the same flower feeding area at least twice, between which times they presumably returned to the surrounding forest for other activities and to pass the night. Some Scada batesi were captured up to four times during the study period at Site 3, with the first and last captures being up to 17 days apart. Whether this is evidence of an ability of ithomiines to learn the location of food sources (as is known for some species of Heliconius; Gilbert, 1975) or merely a testimony of the attractiveness of Eupatorium I to relatively sedentary butterflies is unknown. Although individuals of most species were recaptured only once, the number of multiple recaptures of Scada batesi suggests that the males of this species have a rather restricted home range, whether this be controlled by innate or learned behavior. (At Site 3, 15 S^. batesi were recaptured twice, 3 were recaptured three times, and 1 was recaptured four times.) An attempt was made to estimate the population sizes of S. batesi and G_. zavaleta at Sites 1 and 3 by analyzing the capture-mark-recapture 110 data by the methods of Jolly (1965) and Bailey (1951). Unfortunately, the recapture rates were inconsistent from day to day, ranging up to 90% but with several days of no recaptures. This was partly a result of sample size, for unlike the situation in the forest interior, heavy cloudiness and light rain do depress ithomiine activity at these exposed flower patches, and there were several study days with bad weather. Thus, the extreme unevenness (and occasional absence) of recapture rates precluded the accurate calculation of population sizes, even for the most abundant species, S^. batesi. (The standard error for population estimates of S^. batesi ranged from 55 to 98% of the estimated population sizes themselves) . One basic problem with conducting capture-mark-recap- ture experiments to estimate population size at flower visitation sites is the lack of random sampling of the population — as evidenced in this case by the heavily male-biased collection. At Sites 1 and 3 it was the Eupatorium flower patch that actually collected the sample; I merely processed it by removing the feeding butterflies from the flowers long enough to mark them. Coevolution of Ithomiines and White-flowered Plants Butterflies normally visit brightly colored flowers, especially red and yellow ones (Faegri and van der Pijl, 1966), and have good color vision, including ultraviolet and red (Swihart, 1972). Is there any significance to the fact that the primary nectar sources of ithomiines are small white flowers? The combined facts that (a) ithomiines locate their nectar sources in the dim light of early morning and late afternoon, and (b) that the flowers they visit are , white, fit the syndrome (sensu Faegri and van der Pijl, 1966) of Ill crepuscular pollination and suggest a mutualistic relationship. Crepus- cular ly pollinated flowers are usually white (Faegri and van der Pijl, 1966), and thus provide high reflectance of available light, facilitating ■ their location by visually oriented pollinators. Because of their frequent and prolonged visits to the flowers of Heliotropium I, Eupatorium I, and Eupatorium II and the relatively infrequent visits to these flowers by other butterflies and insects, it is reasonable to assume that the ithomiines are indeed pollinators (as opposed to mere visitors) and that these plants are specializing in attracting ithomiines. Because ithomiines are denizens of the deep shade of the forest understory, they are, in a sense, physiologically pre-adapted to flying in the dim light and cooler temperatures of the early morning and late afternoon, al- though this does not explain why their feeding behavior peaks at these times. Gilbert (1969) has suggested that since ithomiine courtship areas (forest understory) and nectar feeding areas (forest edges) are spatially separated, selection should force males to complete their nectar feeding before female emergence in the mornings (before 1000 h for ithomiines) . Early morning nectar feeding by males should reduce the time spent feeding during the peak courtship period, 1000-1400 h (presumably, also the period of maximum female sexual receptivity) . If there is less than enough nectar available to meet the needs of the male ithomiine community, then competition could force male ithomiines to seek out nectar sources as early as possible to insure the procurement of an adequate allotment before the courtship period begins. It seems unlikely that the emergence time of females plays a significant part in this scenario, however, be- cause it is doubtful that female ithomiines are sexually receptive on 112 the day of emergence (see Chapter 4). But a receptive female, whether or not she has previously mated (females are known to mate up to six times or more; see Chapter 4), is certainly a limited commodity and it would behoove any male to have established his courtship "display arena" before receptive females become available each day. Is nectar a limited resource to the male ithomiine community? The nectar sources available to ithomiines at Limoncocha fall into two basic categories: those that occur in shaded areas of primary or secondary forest understory, and those that lie beybnd the forest edge in open clearings. Clearly at Limoncocha Eupatorium II, Maranta I, and the orange- flowered plant (7339) are members of the first group, and Eupatorium I is the sole member in the second group. Heliotropium I, although occa- sionally found in the open, usually occurs along forest borders and in areas of disturbed forest. Since most of the plants at Limoncocha were located in shaded forest, Heliotropium I is included in group one. The shaded nectar sources (group one) are generally of low abundance, but appear to be spatially and phenologically constant over long periods of time. Thus, within the normal microhabitat of ithomiines (shaded under- story), nectar sources are locally stable, highly predictable, but of low density. In contrast, the Eupatorium I patches (group two) that occur in sunny clearings are phenologically and spatially unpredictable, but make available large quantities of nectar for short periods of time. It appears that the feeding behavior of male ithomiines at Limoncocha reflects the differences in nectar availability of these two groups. The location of the shaded nectar sources (group one) near the court- ship areas, and their relative stability in place and time, should make 113 them extremely important feeding areas to ithomiine males, particularly those of scotophilic species. The low abundance of these plants, however, indicates that they are insufficient to satisfy the nectar needs of all males in the surrounding area, especially if those males should attempt to feed simultaneously. Although a maximum of only 21% (26 of 124) of the inflorescences at the Heliotropium I sample area were visited by ithomiines at any one time (see Figures 21 and 22) , the small florets (6 to 8 mature florets per inflorescence) were apparently quickly depleted of nectar. After an initial flush of feeding, the effective number of nectar feeding stations (i.e., inflorescences with nectar-filled florets) may have been greatly reduced because the depleted florets passed through a recovery period. Thus, it seems likely that competition at group one plants for nectar procurement prior to courtship could occur, forcing males to visit these nectar sources as early as is physiologically pos- sible, as Gilbert (1969) has suggested. (As discussed above, an alterna- tive or additional explanation for early morning flower visitation by ithomiines is simply that they are avoiding the unfavorable environmental «. conditions — high temperatures and low humidities — that occur later in the day.) One would expect, then, that male ithomiines arriving too late to find adequate nectar at the forest flowers (group one) would leave the forest border to seek flowers in the sun. Interesting to this discussion is the fact that the clear-wing ithomiines were found at both groups of flowers while tiger ithomiines were abundant at group two flowers only, and were rarely seen at those of group one (shaded flowers) . Perhaps the tigers, with their protective aposematic coloration, their greater 114 vagility, and their greater tolerance for high light intensities (as evidenced by their higher flight levels in the forest) normally seek nectar sources in the sun (group two flowers). Conversely, the lower flying (and less light tolerant) clear-wings, with their basically cryptic coloration, apparently first seek out nectar sources in the understory shade (group one flowers), leaving the forest to search for group two nectar sources only when those of group one are depleted or too crowded. Returning to the suggestion made earlier that a mutualistic relation- ship may exist between certain white-flowered plants and ithomiine polli- nators, Pliske (1975b) has presented evidence that at least one plant, Eupatorium xestolepis Robinson in Venezuela, has specialized in attracting male Ithomiinae (and also male moths of the family Ctenuchidae) to pol- linate its flowers. The specialization involves ecological adaptations that allow the plant to subsist in the shady microhabitat of its pollina- tors and the production of a floral fragrance that may contain some of the volatile compounds typically released in dead tissue of PA producing plants that attract male Ithomiinae and Ctenuchidae. Apparently, Eupa- torium II at Limoncocha is ecologically quite similar to E. xestolepis, since both have diverged from the common sun-loving habit of most members of the genus (e.g., Eupatorium I) and both attract ithomiines but little else to their flowers. Other Eupatorium and Heliotropium plants are known to be pollinated largely by male ithomiines, danaines, and ctenuchids (Pliske, 1975b) suggesting that mutualistic relationships between male Lepidoptera and these plants may not be infrequent (Pliske, 1975a) . Male ithomiines are also known to be the primary pollinators of an orchid, Epidendrum panniculatum L. (van der Pijl and Dodson, 1966; Dodson, 115 unpublished obs. cited in Pliske, 1975b). Interestingly, this orchid has small white flowers and thus fits the syndrome of early morning ithomiine pollination, although Pliske (1975b) has suggested it may « attract ithomiines (and danaines) by mimicking the extra-floral attrac- tant of PA plants. Ithomiine Attractants A curious aspect of ithomiine feeding behavior, alluded to above, is the selective attraction of males to the dead plant tissues of Helio- tropium indicum Linn. (Boraginaceae) , known as Indian heliotrope, or fedegoso (Beebe, 1955). Beebe describes it as "a typical weed, wholly undistinguished, without intensive odor or color." And yet when uprooted, dried, and hung in favorable spots in the forest, it attracts great num- bers of ithomiine males, especially clear-wing species, as well as male Danaidae, and both sexes of the moth families Ctenuchidae and Arctiidae. Although Beebe (1955) found few visitors to the dried plant from other insect orders, Moss (1947) and Pliske (1975a) have recorded a great variety of other insects on the plants. The strong attraction of Ithomi- inae to dried fedegoso has been exploited by several neotropical collectors (Masters, 1968; Brown and d'Almeida, 1970; Brown, 1972). The evolutionary significance of the attraction of male Ithomiinae to dried H. indiCum is difficult to understand since the plant is a native of Asia and has presumably spread to the neotropical region only since the discovery and colonization of the Americas (Beebe, 1955). Also, the attractiveness becomes effective only on the death of the plant and after the consequent desiccation of the foliage (Beebe, 1955) . 116 Heliotropium is one of several genera in the Boraginaceae that, along with certain genera in the families Asteraceae, Leguminosae, and Apocynaceae, produce pyrrolizidine alkaloids (PA) (Bull et al., 1968). As with H. indicum, these other PA producing plants are, under certain conditions, highly attractive to various male Lepidoptera that feed on the surface of the plants (not at the flowers) for long periods of time (Pliske, 1975a). Pliske (1975a) has reviewed the published chemical evidence in light of his own extensive field observations and experiments, and the following scenario unfolds. Male Danaidae possess abdominal hair- pencils that disseminate an aphrodisiac pheromone during courtship. The aphrodisiac component of the pheromone is a dihydropyrrolizidine, pro- duced in the butterfly from chemical precursors acquired during males' visits to PA plants. Likewise, of the Ithomiinae so far analyzed (Ithomia, Pteronymia, Hymenitis) , the male hairpencils contain a lactone, derived from PA-side groups and apparently acquired during males' visits to PA-plants (Pliske, 1975b) . Pliske (1975c) has suggested that the lac- tone functions as a male territorial and recognition pheromone (see dis- cussion in Courtship section, Chapter 4) . Male ithomiines apparently locate PA-plants by olfactory detection of volatile "esterifying acids" that probably disassociate from the alkaloid ring nucleus in rotting plant tissue. These volatile acids elicit upwind orientation, landing, pro- boscis extension, and exploration of plant surfaces (Pliske, 1975a) . The intact alkaloids serve as phagostimulants inducing prolonged feeding (Pliske et.al., 1976). Beebe (1955) observed that the extended feeding at H. indicum exerted a noticeable effect on the reduction of the escape reaction of ithomiines, which "show an obvious reluctance to leave their 117 repast." A similar "reluctance" was noted in Ithomiinae visiting Eupatorium macrophyllum in British Guiana (Kaye, 1906). One other curious aspect of ithomiine feeding behavior seems « worthy of note. Rene Lichy (1944) recorded that he was "the stupified witness of a remarkable event" when, while collecting with friends in Venezuela, a great number of the clear-wing ithomiine, Hyalyf is (=Hypothyris) coeno, converged on the group, perching on their hats, hands, and faces. The three men were smoking, and, after a series of experiments with several lighted cigarettes, became convinced that both males and females of H. coeno were strongly attracted to the tobacco smoke. The butterflies whirled about the "sweet scrolls" of smoke and even extended their proboscises toward the burning tips before the heat drove them back. Lichy stated that he had since repeated the experiment at several other times and places with the same response from both males and females of H. coeno. In Lichy 's words: Was it not for them a reminiscence of their past, a remembrance of the time where they were still only larvae, when, under their primitive form, they ate wild tobacco or several neighboring Solanaceous plants. The odorous waves of the tobacco smoke seemed to them so familiar . Population Ecology Spatial Heterogeneity It has already been mentioned that ithomiines are not randomly distributed throughout the forest understory. Their vertical distribu- tion appears controlled by species-specific preferences for flight and 118 perch heights and their horizontal or spatial distribution is usually clumped, even within seemingly homogeneous forest. The reasons for the clumped distribution within homogeneous forest are not known. Al- though moisture and light intensity are probably important, other ex- ternal factors may also influence ithomiine distributions. The role of internal factors is just now coming under study. Poole (1970), in the study of a Venezuelan mimicry complex dominated by the Ithomiinae, men- tioned the possible existence of an "interspecific pheromone used to keep the various Mullerian members of [the mimicry] complex aggregated." W. Haber (pers. comm.) found that male scent scales of Costa Rican Ithomiinae attract males and females of many species and thus may account for aggregations of ithomiines. Apparently contradictory evidence has been published by Pliske (1975c), however, who found that the lactone component of an extract of male ithomiine scent scales is repellent both to con- specific males and males of other lactone-producing species. (For a discussion of scent-scale function and use, see Courtship section, Chapter 4.) My observations at Limoncocha, although they provide no evidence in favor of ithomiine aggregations being caused or facilitated by male Pheromones, would not be incompatible with such an interpretation. But the most likely explanation in my opinion is that the slowly changing constellation of vegetational and environmental characteristics of the forest understory occasionally produces a microhabitat particularly favorable for ithomiines, which then begin to congregate in the area. Pheromones are probably important in determining the nature and extent of activities within the aggregation, but probably do not cause its 119 formation. It should be noted that the "aggregations" at Limoncocha for the most part cover quite large areas (often several hectares in extent) and are not really comparable to the more tightly clustered groups described by Ross (1967) and Brown and d'Almeida (1970) that occur in drier or seasonal forest areas. Perhaps in these latter areas, where the location of a suitable microhabitat may mean the difference between survival and desiccation, there is stronger selection for a positive behavioral response to an area of pheromone concentration than in areas such as Limoncocha where vast, stretches of forest appear to be tolerable, if not preferable, habitat. Nevertheless, certain areas of the forest at Limoncocha appeared to consistently provide more favorable microhabitat for most species of ithomiines than other areas. One such area was just south of the Nature Trail at Study Site 5 (see Figure 2) , set aside early in the research period as an observational area. Within this area of several hectares, there were also regions where certain species of ithomiines could be found with reasonable consistency. For example, the orange-tip clear-wing, Hypoleria chrysodonia, was often seen flying or courting at the fork of the Nature Trail. On any given day this species might be encountered almost anywhere in the NT area, but if I wanted to find H. chrysodonia quickly, the first place I went was to the NT fork, for it was there that I was most likely to find individuals of this species. The same was true for several other small regions in the NT area, each one apparently selectively attractive to one or a small group of species. The most consistent of these associations between regions of the NT area and species of ithomiines are mapped in Figure 25 to provide some 120 o cd u 4-1 V cd 0) \ •£■ H \ o. cd - \ (X \ K> cd CX cd Jk-^ f\\ C 0) / \ \ Q) en cd i-H id I V | t>C -H •H cd 4-1 \ 1 I o M e > CO \ 1 / 0) >< o cd •H \ 1 1 p< TJ JS N M J ID o o 4J H, ^ ~M cd -H 4-4 o CO c o •H t>0 0) pc: ^3 cd .c o u CJ CO 0) c 6 o <4-4 o m a> ■H a a> ex CO CO c o cd cd •H > CJ U o «3- o\ St cr> 3 CM CO CO sr CO CO CM lO ■0 \m* n z r^ CM r>- CM r^ rH 0) o CM CO CO o CM 0> r- CO CO •H -J \- \0 rH CO CO IO rH -J M3 >* CM p M 01 p CO o CM IO «• CO O IO CM "■™ •H •H •P 4J M m rH rH CM CM T3 0) M 3 7 CO IO -3- t» 10 -a- 00 rH c o «J P. CO u \ -* a\ rH rH CM >4-l ^> o V r^ p O) CM CO o IO 00 CO CD 1 rH 05 0> rH o •H 4J 3 ro co rH rH CM •rl (X M 6 4J 3 (0 m * CNI to r-» CM * CO •H CM r» rH rH CM • CM 01 o> <* rH en IO rH m u 3 oo •rl rH vo rH rH CM fn 0) *J •H CO 123 their recaptures in the subplots. For reasons given below, only the second and third of these objectives were accomplished with quantifiable results. The numbers of butterflies caught in each subplot provide further evidence of the nonrandom distribution of ithomiines in seemingly homo- geneous forest. When the number of individuals (of all ithomiine species combined) captured 'in each subplot during the nine month sampling period is mapped (Figure 26) , the resulting dispersion of individuals on the grid is far from random. My subjective impression of Site 4 is that the subplots with the highest number of individuals (above 40) all tended to be relatively open between one and three meters above the ground. Subplots 1 through 25 all have low numbers because they were sampled on only the first 7 sampling dates instead of on all 26, as were subplots 26-50. After the cyclonic storm of March 27, 1974. it was no longer possible to collect in subplots 1-25 (see below) . A graph summarizing the results of the sampling program is presented in Figure 27. The number of individuals of the species present on each of the 26 sampling days is given in Figure 28. From March 5 to 9, a sample was collected, marked, and released each day, but by the end of this period it was obvious that the data would be insufficient to calculate population sizes. Out of 246 marked individuals released during those 5 days, only 23, or 9.3%, were recaptured by the sixth sampling date, March 18. As these 23 recaptures were distributed among 30 species, the recapture data for any one species was insignificant for population size analysis. These figures confirmed the impression that discrete populations of ithomiine species do not exist at Limoncocha. co >> •V 3 u co u CTS c M 1 SO •H CM CO CD C .C •H 4-> c co M-l CD ,-H O > CO •H 3 ,C MX) CO a, •H 01 to CO > •w M •H -H o o TJ - k< I H 1 1 »- -+ »— -+ h & & O if\ O IO ™n ki 1TVINIVH o o o O CO o 060 o «» ssioads ho STvnaiAioNi 10 sraswnN •ONiaOOTi XN3DH3d Figure 28. Species and Individuals Present in the 26 Ithomiine Samples Taken at Study Site 4. Ithomiine species are listed by voucher number in the same order as they appear in Table 4. The dates of the 26 samples may be obtained from Table 8. 127 ' 3TUDY SITE 4 SAMPLE NUMBER 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 2021 22 .23 2425 26 5037 1 1 2 2 5023 1 1 1 5024 1 1 3 2 5006 1 2 2 5 2 3 1 1 1 2 3 5 4 1 1 1 2 1 2 3 1 3 5019 1 1 1 2 3 2 2 2 3 5 3 2 2 3 4 3 5020 1 1 .2 1 1 1 1 2 1 1 5 2 3 3 5021 2 1 2 3 4 2 2 2 2 1 1 2 3 3 5 2 5 5022 2 1 3 2 3 3 2 2 5035 1 2 2 1 2 3 1 1 1 2 *- 1 1 1 1 5012 3 1 1 1 5013 1 2 1 3 1 2 2 5 1 2 3 5014 1 1 1 1 5042 1 1 _ 1 5015 1 1 1 5028 2 1 1 1 3 1 1 5026 1 2 1 3 1 2 1 1 1 1 1 3 1 4 5027 2 1 1 1 1 1 2 1 1 2 1 1 1 5040 1 1 1 1 1 5029 5 16 2 4 6 7 8 2 4 5 8 15 8 9 5 3 1 1 9 7 5 1 3 8 5016 3 6 3 6 7 7 16 12 7 4 5 1 3 1 2 2 1 5017 3 2 4 5 1 6 4 1 2 2 111 2 6 4 1 3 1 1 1 1 1 5018 1 5001 1 1 2 1 1 1 1 1 1 1 5010 1 1 3 4 2 1 1 2 2 1 2 1 4 1 1 2 1 1 1 2 1 2 5030 10 3 12 7 6 6 7 1 5 2 2 10 2 3 4 3 2 5 rJT 6 2 6 5011 2 1 1 2 1 1 4 1 1 1 5031 2 2 1 1 2 1 1 1 5003 2 4 2 1 3 5 2 1 2 t 1 5 5041 1 5050 1 1 1 5004 2 1 5034 1 1 1 2 2 2 2 1 1 1 5007 9 4 4 5 6 7 7 2 2 1 7 3 9 9 8 1 3 3 1 2 2 8 17 3 14 5008 1 1 1 3 6 4 4 4 1 1 1 3 6 1 5009 3 1 1 3 2 4 1 1 1 2 1 2 2 5039 1 2 1 1 1 1 1 1 2 1 1 5043 2 1 2 3 2 1 5032 3 3 1 1 5033 1 1 1 1 2 1 1 8 5 11 7 1 2 2 4 1 5036 - 1 1 1 2 3 1 2 1 3 128 The almost continuous expanse of lush forest in this area, despite the microhabitat heterogeneity described earlier, provide ithomiines with an almost unbroken macrohabitat. The clumped distributions of ithomiines within this macrohabitat do not represent discrete communities and populations, but are merely transient concentrations that crest and ebb erratically through time. The marking program was continued, however, in hopes that there would be sufficient recaptures over the remainder of the year to provide some information on longevity and home-range size. From this point on, a weekly or biweekly sample was taken to provide information on the relative frequencies of ithomiine species and to see if ithomiine abundance varied during the year. Catastrophic fluctuations A study of Figure 27 reveals that the one-hour samples normally contained approximately 40-60 individuals and about 20 species, with 74 the maximum number of individuals and 24 the maximum number of species collected during one sample. The periods of strong deviation from this generalized pattern occurred in late March to early April and in mid June to mid July. During these periods the number of individuals collected dropped to about 20 or below and the number of species was less than 10. As all one-hour samples were collected between 1000 and 1200 h on com- parably sunny days, and since the low values were consistent for two consecutive samples in the first period and for five consecutive samples in the second, I consider both of these to be actual minima of ithomiine abundance and not merely sample variation or the result of poor collecting weather. Thus, assuming that the pattern in Figure 27 is not the result of chance, what might the causes of this variation be? 129 The answer could come from one or both of two possible categories of events affecting population sizes: biotic or abiotic factors. During the year I had observed no noticeable changes in the number or quality t of ithomiine larval foodplants, nor in any other aspect of the Limoncocha vegetation. The scant data available to me on the predation and parasi- tism of adults and immature stages suggested no possibilities, so I examined the possible abiotic factors. The study area at Site 4 periodically flooded, in places up to 80 cm in depth, following periods of heavy rains that apparently brought much more water than the saturated soil and low relief of the area could accomodate. When these periods of flooding are plotted with the ithomiine 4 abundances, as they are at the bottom of Figure 27, it is clear that the first two flooding periods correlate closely with the two largest drops in ithomiine sample size, and that the third occurs at another, but much less dramatic, ithomiine minimum. A close examination of the 1974 rainfall pattern (Figure 6) revealed that in the month before each drop in ithomiine abundance there had been a fairly sharp rise in the amount of rainfall over the preceding month. To better understand the pattern of rainfall distribution, precipitation was grouped into six day totals, instead of monthly totals, and the results were graphed at the top of Figure 27. Now a correlation between rainfall distribution and ithomiine abundance became more easily recognized. Just prior to each flooding period of Site 4 was a six day period of exceptionally high rainfall (over 100 mm) , and the longest flooding periods (July and October) each had two such periods. Most dramatic of the three, however, was the first flooding period which literally began overnight after a 130 spectacular storm struck Liraoncocha on the night of March 27. In less than eight hours, 91 mm of rain fell while localized twisting winds felled dozens of trees in the clearings and forests of the Limoncocha area. The next morning, the effects of the storm's violence were everywhere, and the hike to Site 4 from the base, which normally took about 45 minutes, required more than 3 hours. At Site 4 itself I found the entire western half of the one hectare study plot devastated by the felling of two large trees which brought everything else down with them to form an impenetrable tangle of flooded vegetation. Just within the one hectare study plot (the eastern half was flooded but no trees had fallen) I found two dead snakes, a dead porcupine, and numerous dead insects, especially Orthoptera. I have no doubt that the destructive winds, torrential rains, and crushed vegetation took a heavy toll of ithomiine adults and immature stages, and that the low sample sizes of April 2, 9, and 22 reflect this mortality. Although less dramatic, similar storms and torrential rains fell in mid June, early July, and mid October, as my notes and the weather records (and the flooding at Site 4) indicate. It is my interpretation that these intense rain storms were the direct cause of the mortality that resulted in the decreases in ithomiine abun- dance observed at Site 4. The devastating effects of these rain storms are similar to those attributed to a late snowstorm in a subalpine plant community in Colorado that resulted in the extinction of at least one butterfly population (of the lycaenid, Glaucopsyche lygdamus Dbldy) and severely depressed the populations of many other insects, mammals, and birds (Ehrlich et al., 1972). Tropical rainstorms are also suspected to be a major source of mortality for two species of Parides (Papilionidae) during the wet season in Trinidad (Cook et al., 1971). 131 Community stability With the exception of the three depressions in ithomine abundance that appear correlated with storms and intense rainfall, the samples at Site 4 give the impression of a reasonably stable ithomiine community To provide some other measures of the stability of the community as measured by the similarity of successive samples, certain diversity indices were calculated. Table 8 contains the results of the Shannon- Weaver diversity index (H') and evenness values (J') calculated for each sample at Site 4. Formulae are given' at the end of the table and were taken from Poole (1974) . Calculations use natural logarithms (base e) • The utility of diversity indices, especially those based on information theory (such as H'), has been recently challenged (Hurlburt, 1971), but it is not my intent to imply that they have any intrinsic value as used here. However, they do provide a convenient method of tracking the distribution of individuals among species through time in the sequential samples from Site 4. H' is generally considered to be independent of sample size (Pielou, 1967) , although Peet (1974) has shown that, despite being sluggish to changes in sample size, it is not entirely independent of them. The values of H' at Site 4 vary from 1.476 to 2.951, but only 4 of them are less than 2.000 and 17 are over 2.500, indicating that the H' values are uniformly high. The only comparable figures for Ithomiinae known to me in the literature are those given by Gilbert (1969) for several localities in Costa Rica and for two localities in Brazil calcu- lated on the data of Collenette and Talbot (1928) . His figures for three 132 Table 8. Species Diversity of the Ithomiine Community at Study Site 4, SAMPLE NUMBER DATE N S H» H max J' 1 05MAR74 56 • 22 2.770 3.091 0.896 2 06MAR74 53 22 2.587 3.091 0.837 3 07MAR74 35 10 2.040 2.302 0.886 4 08MAR74 50 18 2.654 2.890 0.918 5 09MAR74 52 21 2.753 - 3.045 0.904 . 6 18MAR74 62 19 2.715 2.944 0.922 7 25MAR74 64 17 2.416 2.833 0.853 8 01APR74 22 9 . 11604 2.197 0.730+ 9 09APR74 20 6 1.609 1.791 0.898 10 22APR74 35 19 2.797 2.944 0.950 11 29APR74 64 24* 2.951* 3.178* 0.929 12 16MAY74 50 20 2.632 2.996 0.879 13 22MAY74 74* 22 2.709 3.091 0.876 14 30MAY74 55 17 2.534 2.833 0.894 15 04JUN74 53 19 2.664 2.944 0.905 16 14JUN74 20 11 2.194 2.398 0.915 17 21JUN74 20 13 2.415 2.565 0.942 18 28JUN74 19 9 2.004 2.197 0.912 19 05JUL74 7 5+ 1.476+ 1.609+ 0.917 20 11JUL74 6+ 5+ 1.562 1 . 609+ 0.971* 21 27JUL74 33 18 2.600 2.890 0.900 22 11SEP74 47 21 2.814 3.045 0.924 23 29SEP74 63 24* 2.947 3.178 0.927 24 120CT74 52 22 2.545 3.091 0.823 25 230CT74 37 21 2.864 3.045 0.941 26 12NOV74 59 18 2.570 2.890 0.889 Note: N = Number of individuals in sample. S = Number of species in sample. H' = Shannon-Weaver diversity index = -S(ni/N)ln(ni/N) H'max = Maximum value of H' (obtained if all nWN were equal) = ln(S) J' = Evenness = H' /H'max *Highest value in series. +Lowest value in series. 152 1.26 .81 50 1.13 .88 94 1.22 .97 133 wet forest sites in Costa Rica were S N H' J' La Selva (75 m) 17 Tilaran (600 m) 20 San Vito (1200 m) . 18 and for Collenette and Talbot's data (riparian areas in dry forest) Serragem (450 m) 13 1216 .87 .77 Tombardor (450 m) 9 378 .82- .86 Unfortunately, Gilbert did not specify tbe base of bis logarithms and thus, although I suspect he used log^g* perhaps it is best to restrict comparisons to the J' values. J' = H'/H'max and thus provides a measure of the equitability or evenness of the sample by filtering out the contribution of S (number of species) to the index H'. The J' values of the Site 4 samples vary between .730 and .971 with the majority above .900. Thus, most of the time the ithomiine individuals are over 90% as evenly spread among the species as possible. The equitability of the Limoncocha samples most resembles that of the richest area in Costa Rica, San Vito. Interestingly, Gilbert found a nearly linear increase in equitability with altitude, which he attributed to a corresponding in- crease in diversity in solanaceous larval foodplants with altitude in Costa Rica. It would be most interesting to sample a transect running up the eastern slopes of the Andes in Ecuador, measuring the diversity and equitability of the Ithomiinae and Solanaceae from the lowlands of the Oriente to the upper altitudinal limits of ithomiine distribution. The equitability values (J') for the Site 4 samples have been plotted in Figure 29. It is interesting that the lowest values occurred in the sample (April 2, 1974) immediately following the March 27 storm, suggesting 134 c r< 01 > ca O H ^ to 01 .-1 < a H O co *0{ u c •H ■H ^-s CO < _J J*. 3 4J a) < r-t s < co «z o ^ 4J e ai < 4J <- O 3 O" < T3 >- a * Q. CO 3 < ■U AJ < •H CO 3 C W CO 0 • <§ CM 5 0) : * 3 60 •H -t Pn so yo ,r jo arnvA 135 that perhaps a few species were less susceptible to the action of the storm than others. The greater survivorship of these species would result in their greater representation in the April 2 sample, thereby ■ lowering the value of J'. Another statistic calculated to compare the samples taken at Site 4 through time was S^frenson's Quotient of Similarity (QS) , also graphed in Figure 29. The formula (from Southwood, 1966) and actual figures are given in Table 9. The Quotient of Similarity compares the species composition of two samples in terms of .the number of species shared between them. By comparing Figure 29 with Figure 27, it is readily apparent that QS (Figure 29) varies in much the same way as does N, the sample size (Figure 27). To test this correlation, the regression of QS on N was calculated and the resulting regression line drawn in Figure 30. The correlation coefficient of the regression of QS on N (r = .874) was highly significant (P > .001, df = 23), indicating that about 87% of the variation of QS is accounted for by the size of the sample taken (N) . We have already seen that the strongest variations in N appear to be due to "catastrophic" environmental conditions. Thus, when the overall abundance of ithomiines is reduced, the species compositions of consecutive samples taken during this period are apt to be quite dif- ferent, even though the H' and J' values of consecutive samples may be very similar (e.g., samples 8-9, and samples 17-21 in Table 8 and Figure 29). In other words, small samples may be similar in number of species (S) and the evenness (J1) with which the individuals are distributed among the species, but the species composition of the two samples may be entirely different. This may lead to changes in the. nature of 136 Table 9. Similarity of Species Composition Between Consecutive Samples of the Ithomiine Community at Study Site 4. SAMPLE NUMBER DATE * N S j QS 1 05MAR74 56 22 2 06MAR74 53 22 3 07MAR74 35 10 4 08MAR74 50 18 5 09MAR74 52 , 21 6 18MAR74 62 19 7 25MAR74 64 17 8 01APR74 22 9 9 09APR74 20 6 10 22APR74 35 19 11 29APR74 64 24* 12 16MAY74 50 20 13 22MAY74 74* 22 14 30MAY74 55 17 15 04JUN74 53 19 16 14JUN74 20 11 17 21JUN74 20 13 18 28JUN74 19 9 19 05JUL74 7 5+ 20 11JUL74 6+ 5+ 21 27JUL74 33 18 22 11SEP74 47 21 23 29SEP74 63 24* 24 120CT74 52 22 25 230CT74 37 21 26 12N0V74 59 18 13 .591 7 .438 8 .571 13 .667 12 .600 10 .556 8 .615 4 .533 6 .480 17* .791* 14 .583 12 .571 14 .718 12 .667 9 .600 4 .333+ 6 .545 4 .571 2+ .400 2+ .348 13 .667 16 .711 14 .609 12 .558 12 .615 137 Table 9 - continued. Note: N = Number of individuals in sample. S = Number of species in sample. j = Number of species in common between two adjacent samples. QS = S^renson's Quotient of Similarity = 2j , where S^ and S2 are the number of species S^ + S2 in each of the two samples to be :. compared . *Highest value in series. ^Lowest value in series. 138 o o • A oo en ii a •a o\ oo * o o CO o CO CM W CO O O O o 00 o o o o o n o CM o H to w i-i M o EC H O CO Pi u PS H O N ■rl CO V i-l O. e CO C o CO 0) 8 CO >4-l o a o O* 0) 43 14-1 O C o •H 09 CO 0) 60 a) Pi cm (so) AIIHV7IWIS JO IN3II0nb o en 0) u 3 60 •H Pn 139 predator-prey relationships. For example, during low ithomiine abundance, a predator may find a sample of ithomiines (i.e., a set of ithomiines flying at a given time in a given place) to be quite different in species composition from one day to the next. If, as is quite likely (see Chapter 7) , mimetically similar species of ithomiines are distasteful to varying degrees (i.e., a palatability spectrum exists among species of Ithomiinae) , the type of "education" received by a naive predator could vary from day to day, depending on the species composition of the butterfly sample flying during feeding time. A variable education for a predator means a longer period (and more butterflies sampled) to successfully learn to associate an aposematic coloration with distastefulness. Here then may be another explanation for interspecific ithomiine aggregations during periods of low abundance in seemingly homogeneous wet forest — better predator education. If most ithomiines are distasteful but some are either palatable or only slightly unpalatable, it would be selectively advantageous for less palatable individuals to associate with other ithomiines sharing a similar mimetic pattern, if, >on the average, those other ithomiines were to a greater degree distasteful. Further evidence for the stability of the ithomiine community at Site 4 is provided by the measurements of wing condition taken on all individuals as described in Chapter 2. Although wing condition was scored on the basis of seven categories of increasing scale loss as a measure of relative age, the data were lumped into four categories of wing condition for ease of comparison. These four categories are teneral (F+) , fresh (F) , intermediate (I) , and worn (W) , and the percentage of individuals (of all species combined) falling into each age class on 14 0 each sample day Is presented in Figure 31. The first impression given by Figure 31 is of a relatively stable age class distribution for the ithomiine community as a whole, in spite of the fluctuations in community size that occurred during the sampling period (Figure 27) . The fluctuations about the nine month averages (10.2% F+, 52.6% F, 26.6% I, and 10.6% W) , illustrated on the far right of the graph, are not very great. I had expected a decrease in average wing condition to occur with each catastrophic drop in ithomiine abundance followed by an increase in the proportion of teneral individuals in the next samples as the community recovered. The proportion of worn individuals does not appear to have significantly increased following the storms, however, nor to have been particularly higher during the wettest periods. In fact, the highest proportions of worn individuals occur during peaks of ithomiine abundance (samples 6, 14, 24) suggesting that the increased flying activity that occurs during periods of favorable weather may lead to more rapid scale loss. The proportion of teneral individuals does appear to have been significantly reduced during the low abundance period from mid June to mid July (samples 16-20), however, and may be the result of their vulnerability to hard morning rains that occur soon after emergence. A heavy mortality of late instar larvae or pupae (e.g., by drowning or increased incidence of fungal infection) could also account for the reduced number of teneral individuals. Thus, the relatively even age distribution of the ithomiine community through time indicates that the stormy wet weather that depresses ithomiine abundance does so by causing mortalities among adult butterflies irrespective of age, al- though teneral individuals may be slightly more affected than those in other age classes. ■> 4J CO +J •H g o o PS e 5 OJ 43 <4-l O O •H 4-1 3 .£> •H >-. 4J (0 CO CO cd 0) 00 Pu M JS E9 CD n 3 (30 •H 142 oo r*. vo m -a- n cm SSV1D 33V NI STVnaiAICINI JO INHDSHoT 143 Longevity Although the overall recapture rate of released butterflies at Site 4 was small (33 recapture events involving 27 individuals out of a total 754 released butterflies, or about 4%), a few of the recaptures provide some insight into the longevity of ithomiines. All of the marked individuals recaptured at Site 4 are listed. in Table 10, along with their condition at each capture and the length of time between first and last capture. The number of days (D) between first and last capture can be divided by the number of new age class categories entered by the butterfly since it was marked (A). The quotient (D/A ) provides a crude estimate of the number of days spent in each age class category. Multi- plying this times seven (the number of age class categories) we arrive at an estimate of the life span of these individuals. Obviously, a butterfly first caught at the end of one age class and recaptured at the beginning of another age class will tend to overestimate longevity, and vice versa. Also, this method assumes that the amount of time spent in each age class is the same, which is probably not true. For example, Ehrlich and Gilbert (1973) had enough multiple recaptures of Heliconius ethilla on Trinidad to calibrate their measurement of age based on wing- scale loss. They obtained an approximate ratio of 1:3:1.3 for the rela- tive amounts of time spent in the age classes F:I:W. A similar ratio would not be unexpected for ithomiines at Limoncocha if enough data were available, but without it we must be content with the crude estimates given in Table 10. Oleria agarista, for which the most data are available, has life span estimates ranging from 63 to 182 days. The average life span for all eleven individuals of this species comes to nearly 70 days, 144 Table 10. Estimates of Longevity of Ithomiine Butterflies Based on Recapture Data at Study Site 4. 1 L= SPECIES # SEX Ri D Co Cx A d/a (D/A) (7) Aeria eurimedia 1 ? 2 35 I W 2 17.5 122.5 Forbestra truncata 6 rf 1 1 F F- 1 1 7 Godyris zavaleta 10 a 1 1 I I- 1 1 7 13 d 1 2 F- F- 0 >2 >14 25 ? 1 16 .. . I W 2 8 56 Hyposcada kezia 1 d 1 1 F F 0 >1 >7 3 9 1 1 I I 1 1 7 15 d 2 29 F F- 1 29 203 16 d 2 37 I W 2 18.5 129.5 Hypothyris fluona 1 d 1 3 F F 0 >1 >7 Ithomia amarilla 4 d 1 1 F- F- 0 >1 ?7 9 d 1 11 F F- 1 11 77 Napeogenes apobsoleta 1 9 1 19 I- I- 0 >19 >133 Napeogenes corena 3 9 1 2 F+ F- 2 1 7 Napeogenes pharo 23 9 2 18 F I 2 9 63 31 9 1 10 F F- 1 10 70 Oleria agarista 11 ? 1 1 F- F- 0 >1 >7 13 9 1 1 F F 0 >1 >7 14 9 1 11 I I- 1 11 77 18 ? 2 18 F I 2 9 63 23 d 1 2 F- F- 0 >2 >14 26 d 1 52 F I 2 26 182 ■ 27 d 1 1 I I 0 >1 >7 31 9 1 9 I- W 1 9 63 47 9 1 37 F I 2 18.5 129.5 52 9 2 50 F+ I 3 16.7 116.9 58 9 1 14 F F- 1 14 98 X ■ 9.9 69.5 Oleria kena 3 9 1 2 F F 0 >2 >14 145 Table 10 - continued. Note; § = Individual number of marked butterfly. R^ = Number of times the butterfly was recaptured. D = Number of days between initial capture and final recapture. C0 = Condition of wings at initial capture. Cx = Condition of wings at final recapture. £> = Number of units of change in wing condition between initial capture and final recapture (C0«-Cx)'. D/^ = Number of days spent in each category of wing condition. L = Estimated life span of the individual, assuming that an equal number of days was spent in all 7 categories of wing condition = (D/a) (7) . ' 146 or over two months. In fact, nearly all the individuals of all species that have long recapture intervals provide life span estimates of at least two months, and many are considerably longer. For example, a male (it 15) of Hyposcada kezia was first captured in fresh (F) condition, but, nearly a month (29 days) later, was recaptured in fresh-minus (F-) condi- tion. The life span estimate is thus 203 days, or almost 7 months. While this may be an exceptionally long life for an ithomiine, the average life span for most species would appear to be on the order of two to four months. Similar life spans have been reported by Gilbert (unpubl., but cited in Gilbert, 1972) for Ithomia pellucida and Hypothyris euclea in Trinidad, where marking studies indicated these species can live at least four months. The individual with the longest recorded life span in the wild during my study was a male (# 26) of Oleria agarista, first captured on March 8 and recaptured on April 29, 1974. Thus, the recorded life span was 52 days, but this individual had undoubtedly been alive for several days before initial capture (condition F) , and probably lived for quite a while after final release (condition I) . A life span measured in months instead of weeks or days is not un- r - expected in a tropical butterfly, especially for sylvestral species that inhabitat the relatively uniform environment of the tropical rain- forest understory. Indeed, Ehrlich and Gilbert (1973) have records of many individuals of Heliconius ethilla living as long as 140 days in the wild and one male, recaptured for the last time 162 days after marking, probably lived a total of 6 months or more. Long life spans have been found for other species of Heliconius (Turner, 1971; Benson, 1972), Marpesia berania (Benson and Emmel, 1973), and in the genus Morpho (Young, 1971). 147 A life span, however, is not the same as a life expectancy, the latter being greatly influenced by the frequency of catastrophic weather conditions, predation, and other events that place a butterfly's life « in jeopardy. Life expectancy is normally calculated from the survivor- ship term obtained in analyzing recapture data, but, once again, the low recapture rates in any one species population at Limoncocha precluded its calculation. Very few survivorship rates and life expectancy estimates are available for tropical butterflies. Cook et al. (1971) found that both Parides neophilus and P_. anchises had life expectancies between 5 and 10 days in Trinidad, although this was considerably shorter than the observed life spans (several weeks to months) of these same species in insectaries. Turner (1971), however, obtained a life expectancy for Heliconius erato in Trinidad of 50 to 100 days. This figure is substan- tially greater than that recorded for most temperate zone species, the life expectancies of which are generally less than two weeks (Scott, 1973) . By comparing the estimated life spans of ithomiines with those established for Heliconius, it can be inferred that most species of ithomiines probably have life expectancies measured in tens or perhaps even scores of days. CHAPTER IV REPRODUCTIVE BIOLOGY OF THE ITHOMIINAE Courtship Courtship activity of male ithomiine butterflies was observed in 26 of the 53 Limoncocha species, and 139 specific observations of court- ship initiation behavior by males are available for analysis. Although field observations were made during all hours of the day (between 0600 and 1900 ) , male courtship behavior in ithomiines was only observed be- ll! tween 0900 and 1700 h. Courtship activity increases rapidly after 1000 and reaches its peak between 1100 and 1300 h, trailing off gradually over the afternoon, but generally ceasing between 1600 and 1700 h (see Figure 32) . Similar peaks in courtship activity for male ithomiines have been recorded by Gilbert (1969) in Costa Rica (1030-1330 h) and 4. Pliske (1975c) in Venezuela (1000-1600 h) . The number of hours of field observations between 0900 and 1700 out- numbered the number of hours spent between 0600 and 0900, and after 1700, by a factor of three. 148 149 to 0 •H O 0) p. 0) \D CM co o u-1 O LO o m o m -3" -H (D o <: C o Q •H CO ■H i-H fn R o o X, O w u O s M CN M rH H 0) rH cd o IS o rH iw rH O O >» O 4J o •H rH > -rl 4J o U O < a\ o •H X. o 00 o •U 00 Vl o 3 O o o r-» • o o a o u \o 3 o tc ft, diHsinnoD do SNOixvA^asao ro ciaawnN 150 Function of Hairpenclls in Male Ithomlines Males of all species of ithomiines haver-ose or two tufts of andro- * conial scales arising from the dorsal surface of the hindwing. These elongated scales (usually 6 to 8 mm long, but ranging from A to 12 mm in length in Limoncocha species) are attached to the wing membrane in the anterior portion of the hindwing discal cell, just posterior to the radial sector (Rs) vein (see Figure 33a) . These androconial scales are grouped to form a compact bundle or brush, sometimes called a hairpencil (Pliske, 1975c), that normally lies flat against the hindwing, such that the free end points distally. At certain times during courtship (see below), however, these hairpencils may be erected to form a hemispherical splay, such as that shown in Figure 33b. Just anterior to the hairpencil is a region of specialized scales that lacks the normal coloration of the hindwing, having instead a highly reflectant sheen. I suspect that these scales secrete the scent that is disseminated by the hairpencils, * Curiously, females of some species of the genus Thyridia also have hair- pencils, anatomically positioned like those of the males. It is not known if this is an ancestral (Fox, 1940) or a derived (Fox, 1956) condition, although Lamas M. (1973), whose dissertation I have not seen, may have resolved the problem. At Limoncocha, females of Thyridia confusa psamathe were dimorphic for the presence or absence of hairpencils. Of eleven wild- caught females examined, four (36%) had hairpencils and seven did not. Three females reared from the egg stage did not have hairpencils. It is not known if Thyridia females use the hairpencils during courtship, but since all eleven wild-caught females were mated, there is no indication * that the presence or absence of hairpencils in female Thyridia affects their mating success. 151 Figure 33. Location and Erection of Hairpencils in Male Ithomiines (Napeogenes apobsoleta) . :. ::in. : a. Dorsal view of hindwing, showing hair- pencil tucked under costal fold of wing. b. Anterior view of hindwing, showing hairpencil in erected state. 152 but at present there is no evidence to support or discredit such a suggestion. In some species this region of specialized cells is quite distinct, as in the genus Ithomja in which it is located in an oval cell bounded by the radial sector (Rs) and subcostal (Sc + R^) veins. In other species the scent-producing scales (assuming for the moment that this is their function) are apparently more widely, distributed anterior to the Rs vein (e.g., in the genus Mechanitis) . Although the scent- producing function of these surface scales has not been demonstrated, the scent-dissemination function of the hairpencils has long been sus- pected by early students of the Ithomiinae (Muller, 1877; Kaye, 1914) and recently confirmed by Pliske (1975c). The scent released by the androconia of male ithomiines was originally thought to have a protective function. Longstaff (1912) described the scent of several Venezuelan clear-wing ithomiines as being "of a disagree- able character, recalling stables or pig-sties" and associated the odor with the hairpencils of the males. Muller (1877) observed that "the Ithomiae (sic) of the Amazon . . . are said to be90%) encounters between ithomiine individuals will be between members of different species. The Probability of Intra- specific Encounter is simply 1 - PIE, and these values are also expressed as percentages in Table 11. To estimate the probability that a male will encounter a conspecific female as the result of a random event, the average proportion of females (for all species combined) for each sample has been multiplied by the Probability of Intraspecif ic Encounter (1 - PIE) . The results range from 1.1 to 14.9%, but the majority of the values are less than 5%. The fact that each species has a characteristic flight level (although this level may be quite broadly defined) improves the 157 Table 11. Sex Ratio, Probability of Interspecific Encounter (PIE), and Probability of Intraspecif ic-heterosexual Encounter in the Ithomiine Community at Study Site 4. SAMPLE NUMBER DATE S dtf • %22 PIE [(^2) xlOO] 1-PIE [ U^2> xlOO] (1-PIE) x(%??) 1 05MAR74 22 26 30 53.6 93.5 6.5 3.5 2 06MAR74 22 27 26 49.1 89.1 10.9 5.4 3 07MAR74 10 18 17 48.6 85.2 14.8 7.2 4 08MAR74 18 30 20 40.0 86.0 14.0 5.6 5 09MAR74 21 26 26 50.0 . 93.5 6.5 3.3 6 18MAR74 19 33 29 46.8 93.7 6.3 2.9 7 25MAR74 17 22 42 65.6 89.4 10.6 7.0 8 01APR74 9 11 11 50.0 70.3+ 29.7* 14.9* 9 09APR74 6 5 15 75.0* 81.1 18.9 14.2 10 22APR74 19 24 11 31.4 95.9* 4.1+ 1.3 11 29APR74 24* 42 22 34.4 95.2 4.8 1.7 12 16MAY74 20 27 23 46.0 92.0 8.0 3.7 13 22MAY74 22 53 21 28.4 92.2 7.8 2.2 14 30MAY74 17 35 20 36.4 91.8 8.2 3.0 15 04JUN74 19 35 18 34.0 93.0 7.0 2.4 16 14JUN74 11 12 8 40.0 76.6 22.0 9.0 17 21JUN74 13 9 11 55.0 94.2. 5.8 3.2 18 28JUN74 9 13 6 31.6 89.1 10.9 3.4 19 05JUL74 5+ 6 1 14.3+ 85.8 14.2 2.0 20 11JUL74 5+ 3 3 50.0 93.4 6.6 3.3 21 27JUL74 18 25 8 24.2 93.5 6.5 1.6 22 11SEP74 21 20 27 57.4 94.6 5.4 3.1 23 29SEP74 24* 48 15 23.8 95.3 4.7 1.1+ 24 120CT74 22 28 24 46.2 87.5 12.5 5.8 25 230CT74 21 16 21 56.8 95.6 4.4 2.5 26 12N0V74 18 32 27 45.8 91.2 8.8 4.0 Totals 626 482 Overall Sex Ratio 130: :100 43.5 158 Table 11 - continued. Note: S = Number of species in sample. dtf = Number of. males of all species in sample. 9?= Number of females of all species in sample. PIE ■ Probability of Interspecific Encounter = &2* 1-PIE = Probability of Intraspecif ic Encounter = 1-A2. (1-PIE)(%$?) = Probability of Intraspecif ic-heterosexual Encounter (i.e., probability that a male of species A will encounter a female of species A). *Highest value in series. +Lowest value in series. 159 probability of intraspecif ic encounter slightly. Nevertheless, it remains evident that an ithomiine male relying solely on chance encounter to locate an appropriate mate will probably squander a lot of time and en- ergy in the process. It is not surprising, then, that most ithomiine males exhibit active, rather than passive, mate-locating behavior. In none of the 26 species of Limoncocha ithomiines for which observations are available did males use the "seeking" or "patrolling" technique for locating females. Instead, all observed mate-locating behavior involved the use of a perch by males, but not in the sense of simply "waiting" for passing females. Males use their perches in two very different ways, for which I propose the terms "display perching" and "patrol perching." Patrol perching may be considered a special case of "waiting" behavior, but display perching requires a third category of male mate-locating behavior that is perhaps best called "assembling." Both types of behavior have been recently but incompletely described in the literature (Gilbert, 1969; Pliske, 1975c). Display perching Gilbert (1969) was apparently the first to record display perching behavior in ithomiines, but unfortunately his observations have never been published. Pliske (1975c) has recently described such behavior for Pteronymia nubivaga and has observed similar behavior in eleven other Venezuelan ithomiine species. Like most other butterflies at rest, male and female ithomiines normally hold their wings together over their backs when perched on a leaf. In display perching, however, a male ithomiine holds his wings 160 apart in such a way that the hairpencils are exposed and erected (Figure 34) . Gilbert (1969) suggested that the function of this androconial display is to disseminate the male scent and in this way assemble females « much in the same way as female moths assemble males. Pliske (1975c) recorded that the wings are held at a 180° angle in Hypothyris and at a 60-120° angle for species of the other genera he observed (Pteronymia, Ithomia, Aeria, Hymenitis, Oleria, Prittwitzia, and Tithorea. ) Pliske noted that species in the genera Pteronymia, Hymenitis, Ithomia, and Prittwitzia generally perched in open sun where the understory was sparse, but that species of the genera Hypothyris, Oleria, and Aeria were more apt to be found displaying in shady areas. In general, Pliske found that a displaying male ithomiine would fly after many of the butterflies (not just conspecif ics) that entered its visual sphere, but that interactions with other species and conspecific males were quickly broken off after only a few seconds. Pursuits of conspecific females lasted much longer, though none were actually observed to lead to copulation. After most pursuits, a male would return to his original perch or an adjacent one. Patrol perching At Limoncocha, males of species of the genera Mechanitis, Forbestra, and Thyridia perch on the tips or edges of leaves at heights between one and three meters, usually positioned so that they face an open air space (i.e., an area free of vegetation) of at least one or more cubic meters (Figure 35). If another butterfly (or a falling leaf, or almost any other moving body) penetrates this air space, the perched male flies 161 Figure 34. Napeogenes apobsoleta in Display Perch Courtship Behavior. Figure 35. Mechanitis messenoides in Patrol Perch Courtship Behavior. 162 out and pursues it for an indeterminate distance, the length of which apparently depends on the type and strength of the stimuli provided by the pursued object. Pliske (1975c) also described the mate-locating behavior of two species of Mechanitis (isthmia and polymnia) and of Thyridia (=Methona) confusa. In each of these species the male perches on terminal twigs or leaves with its wings together over its back and, according to Pliske, flies toward any butterfly with its own color pattern. Once again, interspecific or conspecific male encounters were substantially shorter than those involving conspecific females. After most pursuits, males resumed a perch, but not necessarily in the same place from which they flew. Thus, according to Pliske1 s description, the mate-locating behavior of Mechanitis and Thyridia could be classified as waiting (sensu Magnus, 1963) Interperch flights My observations at Limoncocha, however, suggest that both "display perching" and "waiting" in ithomiines are more complex than Pliske' s descriptions indicate. In both types of behavior ». the perching portion of the behavior is consistently, and more or less regularly, punctuated by "interperch" flights of varying, but usually quite short, duration. The interperch flights of display perches are slow and fluttering and are made at about the same height as the perch. They usually circumscribe an area of about one square meter or less. Most often the butterfly flies a simple loop around such an area and returns to the original perch or one quite nearby at the same height. These flights are usually two to three seconds long and, after landing, the open-wing display is imme- diately reassumed. Occasionally these flights are slightly longer and 163 involve a back-~and-forth or figure-eight pattern. The fluttery aspect of these interperch flights is unlike any other ithomiine flight behavior, but most closely resembles the pre-oviposition flight characteristic of female ithomiines searching for foodplants. It seems likely that the purpose of this interperch flight is to waft the male scent emanating from the hairpencils into the air space near the perch. The regular pattern and limited extent of these flights serves to concentrate the scent over an area of about one square meter in a layer of air about 10-20 cm thick. (The actual size of the arc circumscribed tends to vary among species but is relatively consistent within any particular species) . By distributing his scent within such an air space over a period of several minutes to an hour or more, a male apparently produces an olfactory disc, or courtship arena, with a powerful center that radiates the aphrodisiac pheromone (sensu Butler, 1967) in all directions by diffusion. Presumably, a passing receptive female that detected the odor would respond by orienting so as to fly through the arena and thus into the male's sphere of vision, at which point he would initiate pursuit. Likewise, the species utilizing the "waiting" perches, or "patrol perches" as I have named them, also have an interperch flight component (patrol) to their behavior. Their interperch flights are stronger and less fluttery, but also circumscribe an area by a looping arc or by some crisscrossing pattern of flight. The area covered, however, is much larger than that of display perchers, and usually encompasses about 10 to 12 square meters. Most of the butterflies that enter this air space at approximately the perch level of the male will stimulate the male to pursuit. The function of the interperch flights is not well understood. 164 Most probably, they represent "seeking" or "patrolling" mate-locating behavior interspersed among longer stretches of "waiting" or "perching." If this is true, then the apellation "patrol perchers," implying both « types of behavior, is appropriate. On two occasions I caught a male J_. confusa after observing it for several minutes in patrol perching behavior. When captured, both males were redolent with the musty odor of vanilla, characteristic of this species. It may be, then, that patrol perchers also disseminate their scents during the patrol portion of their mate-locating behavior. A more likely possibility, however, is that both males had recently encountered a female, and that this stimulated the release of their scents (see below) which were still strong at the time of capture. Species-specific Male Courtship Behaviors At Limoncocha, display perching behavior in male ithomiines varies consistently from species to species. Pliske (1975c) noted that some species perch in shade while others generally perch in open sun where the understory is sparse. At Limoncocha, male ithomiines of different species not only show preferences in terms of lighting conditions (al- though none of the species perched in full sun) but also tend to consis- tently perch at particular levels of the forest understory, although these levels are probably influenced to some extent by light conditions. The fact that males of a given species display at a particular level in the forest is not unexpected since selection should favor males which perch at a level where conspecific females are most likely to be flying. As noted in the previous section, each species of ithomiine at Limoncocha 165 (for which enough observations are available) has a characteristic flight level in the forest. Although Papageorgis (1974) has shown that members of a particular mimetic subcomplex fly at a given stratum in the forest, it can be predicted that within a given subcomplex, selection will favor flight levels for each species such that the numbers of visually similar but non-conspecific females that fly past a displaying male will be minimized. In other words, we would expect the ithomiine members of a particular mimetic subcomplex to subdivide the flying space of that subcomplex. The preliminary evidence available suggests that this is indeed the case. In Figure 36, the perch heights of 123 courting male ithomiines are plotted against the time of day that each observation was made. The observations were made throughout the entire study period and repre- sent males from 24 species, including both display perchers and patrol perchers. The histogram in Figure 32, which includes all these observa- tions and more, has already revealed that the peak of courtship activity occurs between 1100 and 1200 h. The scatter diagram of Figure 36 further shows that courtship activity most frequently occurs between 1.5 and 2.5 m above the ground, resulting in a cluster of courtship activity in time and space. By replotting these data for particular mimetic sub- complexes, we can test the hypothesis that the species within a subcomplex are subdividing the available courtship space and time. In Figure 37, the display perch heights of males of four species of the yellow clear-wing mimetic subcomplex (see Chapter 7) are plotted for three of these species, and they do indeed seem to display at dif- ferent levels, although there is considerable overlap. Napeogenes pharo displays primarily from 1.7 m to 2.4 m, Godyris zavaleta from 1.0 to 166 n CM rH I • • • • • • # • • • • • • • •• • ••• • # * • • • •• + + vO ir» -d" ro SE313H Kl HDHad JO IHDI3H o o 00 • rH en a) c O •H o •H r^ g rH o J= 4J o M o NO QJ rH rH a) 2 o o bO m a r-l •H 4J h o 3 o O • o O •U OS 45 o 60 •H a> o PC o 00 42 o o u .£> .O .O ,0 o o o o O Oi H C1P1 vO i-H CM • CO CO o en i-H 4-1 u M CO 0) CO CO CO 4J t-H CD 4= CM o o • CO > cO U (X en H S° " CO 0) CO cO -- a N TD Q) T-t • CO cfl c OJ t ^ o -H "H M C . < M B o o >i o 0) v-l TJ -C Pu 0) o 4J CO ■U . o HH 5Z PN '•• • o o d • < o o • • • • o . - 8 • • • • • ■ i • • 1 1 1 1 i I 1 1 1 1 v£> in V* t-H 01 ■ O 1 o cd «tf iH .H D. • 0) X •H 4) o O rH o a. CO M e t-H O o <4-l CJ o f>s D o CO CO CN o rH CJ U-l -H o JJ O a) o 0) £ ■H e -H rH H OC O 0) c O D -H O CD 3 rH U Vj 0) CO > CTv ,c O OC 2 •H o 0) rH O £2 rH O 0) CO x: >* o CJ Vj HJ 0) fl o Pu u o r^ o ro o o o 00 168 co co m co co en J3 ,0 ,0 X> X> X OOOOOO < r-l CM ^ H O a! •H N 01 ,ai 4 -a to o CO o PJ u >i X- u o >, >. co cc i-> .c CD « CO c CD 6C O CD CD ^H O CO -O O P-l rt CO 0) C CD OC o CD cr O a ft u o S'HaiSW NI HD^3d 30 XH0I3H 0 CO 0 0) eo G r-i •H g O O O X. r^ -U t-l M 0) O iH O CO \o s rH 00 C O «H O X. IT) 0 .H O 1 O >> -* CO rH i-l CO O -H O C ro i-H n 0 O O >. CN cfl iH Q O O • O X rH CD CD rH e *-{ •H P. H e 0 0 0 CO u 0 3 r-* CO 0 l-i iH 01 OJ O > CD O £ o> •U -H O 00 X •H n O 0> CD O p: 00 00 •H O 0 H >-i CD O 01 X O ex, .u r^ O co O en O vO CD 0 u 3 60 169 BO CO CO CO CO to CD .O .O X> .O U3 .O X) OOOOOO 0 cn ,-h nco'Ovo vO iH m 0) 0) i-i -0 CO 0) to ■H 4-1 «0 -H rt -H CO 0 0 ■u co •H C D c kH n3 C J e CD CD 0 01 -H cO CO C CO 4-1 CO N CO D 60 w >> CO 0) H 0) -H .-1 e e 4-1 CO CO CO CO m CO -H -H tH •H •H U 4J 4-> 4J 4-1 4-) 4-1 -H -H -H -H ■H CD C C C C C QJ cO <0 CO CO at ,£> n J2 iH O M 01 0 Pj 0 1 -* .H i-H O M 4-1 O CO O P-l cn f-i U O iw 0 0 >^ CM tfl t-H Q CM O O • O X rH 0) 0) tH e h -H O. h e O 0 O CO CJ O 3 i-l CD CJ U -H 0) 4-1 O > CD O e EC & I w o Pi 12 O H to S W ►J O <: h EC O z w B o fn M O Ph H EC O o ps m e > W W v- < Ph EC Pi H CO m I m m r-. rH CN o CN r-. o O rH z O PC CO x CO CU o X CO CO 4-1 s 0) cO rH cfl C c0 O CO 01 •H U CO CO o — I 6 X X CO h o ^H 4-1 o 0) CO 3 3 0) ex 3 X CU i+-i CO CO •3 o, z w o o w 5! CO co CO CO CO CO ■H •H a> 0J 0) cu H U c 3 3 c >, rl a) cu cu cu X X 60 60 60 60 ■u 4J o O o o o o 0) cu cu cu a a, a ft p. ex >. >. CO CO CO CO EC EC z z z Z « iH CO ■H CO Vh cfl cfl u § CU cfl •u l-l z o EC H cfl CO •H •H | a 0 o X x •u 4J H M m 4J 4-1 x X •-N ^"N 60 60 ■ • •H •H CO CO rH rH X X CO CO cu o o T3 CU 14 u u rH .-H f^ M O Ps >N o cu X s_x ^w CO i-t O CU M X! CO CO X p. "O T3 X a X cu CO T3 rH r-l CO T3 CO -CU CU co co CO X 3 CU CX (X 3 0) iH CO (0 3 3 Cu u >-. CO CO 14 >^ 0 o o M >-. U CO X CO •H 6 g Xi CO rH cu X X a. X X 4J X -H 4-1 & 3 3 4-1 Cu a) cu 4J X 4J CU X CO -H •H X CO rH X 60 X X 60 T) T3 60 i-t 4J 4-1 60 •H 60 1 •H CU CU •H 1, -3 o o •H rH 1 •H 1 H £ s H 1 E z z IH CO 1 H cu cu cu CU CU CU CO o 0 o o o o o 4J o o o 0 0 O o O O O o O 4J O O O O O vO o CN r-H 00 a\ O 3 m ro CN <• IT) rH CN CN CM rH i-H CN ^« CN i-\ rH O O O o o o o o o o O O O f> o m CO CO o o o o o CO CO O CM <• m CN t-H CN CM CM m CM CM iH -3" CO o rH CM CO SO O O w o cfl CO CO -H •H 4-> "O N CO a) cu •H a ^i ^4 3 3 3 u CO 60 CU 3 TJ CO1 .* 0J 3 U 3 3 cfl CO •H •H •rH o M M u a 3 3 CD >. iH i-H < EC O O 173 13 CD 3 3 c o o CM rH X 3 H J ■fe h < o o z M z H w w t-H P-. H CO E § O >j < hJ PL. c/j CO Q ►J fn W M O to o o !S w w Pi o o o CO o Z O I o o Z. z K 12 I W O Pi Z O M PL. & Z W w w •J ? • w > M PC c_> p<; EH w O Ph H o Pd m g Pu PC CO w M w Pm 00 M Pi H in .3 CO 3 s-i X 4-1 .3 oo 00 3 o 5-1 4-1 CO 60 c O rl O o o in SO o CO CO u CO CX| CO \o \J3 w u s o u 3 4-1 Pli M Z Pi Q O O m 4-> ,3 00 •H i-H co CD rl o 0) 4-> X CX J3 42 03 CO CO •H 3 3 £ S-J >, u CD ^3 CO X .3 i-H 4-1 CX 4-1 rH X CO X t-l 00 oo 3 •H •H Ph H H 00 ■u 00 3 XI 3 o 00 O S-i •H J-i 4-1 rH 4-> CO CO CO 00 4-1 cu c XI a o 00 o S-i ■H 23 o o O O o\ m I m o o o o -* m o O oo o CO co oo CO m • • rH o rH CO •H C O •a CO CO o 3 4-1 (0 6 CD !^ •H rH M 4-1 CO X > o cfl CO *3 N CO CO -H O CO M co •H CD o M .-I ■3 >% O 3 ■a ex 01 o >. CO o 33 Ph u CD O >-l C UM LM 3 «H O 4-> on >, co )-i c cd O cfl -H J3 ex g oo -lj M • PX t— 1 CO •> rH •H CO e 13 3 x -u cu • 4-1 CO CO iH 4J rH 3 4-> O CO 3 3 O U U 3 cfl 4J N 3 co a 3 •H O CD o U o x cu cu N o co 4-1 J3 >-l •H X -rl CD 4-i a) V* X ,3 CO O 3 • cfl H 3 CO S X > CO cfl 3 S • O -H >i • CD CO cu 10 JJ X3 CD • 6 CD "33 > •W 3 CO O >H cu ec o X XI .Q 3 rs . CO ^3 3 . r^ cc 1 cfl Fn S 3 >o O 4-1 1- • 13 u*> J3 J o <■ >-l t> CU 4-1 CU 0) •H o rO O 3 cc 6 3 4J J3 X m 3 4J rH CO CO C 4-1 3 -J- O 3 a CO O cfl 4J 4J , O i- O X >> 3 3 >J -H U o" 13 ct CU •H t-l Cfl iH rl r-l U CO rH 5-1 CX 3 «rH CD O B • 3 3 rl 3 rH Cfl CC >s 3 O 4-1 3 CO 4-1 X rH £2 H -H •H 3 CX 00 4J C! O HS O M 13 3 3 3 O cfl -H Vj 3 •••HON o S >-i Vi cfl 3 00 4-J 4-i N iH 3 P O CU B CX 3 O G t1 U O 4J U -H 3 iH rl O 4J C2 UH CU 4-1 rl O O r3 cfl & O cfl ex 3 3 CX X X >. P. 4J •H A 4J H 01 CO •> o #\ 4f. #V >4-i J3 -l 4J X U CU 4-> CJ 3 4J 4J 4-1 14-4 CU !-i 3 O O O OO cfl 4J O O X CU ,Q C3\ C?\ CT> rH u 3 4-1 4-1 •H ,J3 00 •» A cfl cfl cfl cfl 13 3 cu o rH rH 3 CO -H > 4-1 3 3 00 00 00 OO •H 4J iH rH rH 3 3 3 3 14-1 C 4-1 CU rH rH •H -H -H -H CD O 3 cfl > 3 3 g g gg rH OrJrl /-^ S CD M CO XX iH CX CX >4H CM m 14-| *-^ 4J >s CU cu cu CU CU CU 3 CD iH H 3 3 cfl cfl cfl 3 X 00 CO rH 3 3 a a a c ■LJ 3 U i-\ cu cu CU CU 3 3 LH & r-4 4-1 iH 4J 4-1 4J 4-1 4-1 4J O "3 O 3 3 3 3 Z$$$ 3 > OJ >-i < <: >• •H 3 > 4J 13 X -H 3 O X 13 CU ■o to 01 > <-{ CN ■o •* in vo 3 cu m 0) fl fl DOT) X •H 3 CU ■■ i-l •^ »H 3 rH • 3 SB 00 > o 3 fe O (0 3 •H •H <— • ^3 -H cfl 13 4J 4-1 CU •H X 00 iH S (0 4-1 CO CO 3 E -H »-l o cfl •H CU CU CX 13 3 B cu s > cu o X S 3 rH < J-l CO 4J -H 3 O CD 3 <4-l 14-4 >4-l ^3 C 3 " O O 4-1 cfl 3 CD CU 4-1 4-> O H 4 ■z * + 174 observed. This male, perched in the center of a large Heliconia leaf, was rapidly fluttering his wings from a vertical position to some 10-20° above horizontal. The fluttering continued, with only an occasional split-second stop, for over five* minutes before it began to decrease in speed and amplitude and the stops inbetween gained in duration. After ten minutes the stops were much longer than the fluttering periods. Two minutes later the butterfly made a two-second looping flight, then returned to his original perch, and did not display or flutter again. Perhaps more interesting than the species differences in anatomical orientation are the species differences in flight and perch dynamics. Figure 40 presents a sequence of perches and interperch flights (ethno- gram) for a male of each of eight display-perching species. The dif- ferences in behavior of these eight species are quite pronounced. The perches and interperch flights of T. harmonia are of approximately equal length. (The relatively long interperch flights of T. harmonia are related to the type of scent dissemination that occurs during them. A male was observed on several occasions to use a 7 m perch on a particular tree at Site 4. After display perching on a favorite leaf, the male fluttered around the tree several times, weaving in and out of the lianous vegetation surrounding the trunk — apparently concentrating his scent — before returning to his original perch.) Hypothyris euclea has a series of short perches broken by relatively longer flights until a very long perch with wing pumping is made. The interperch flights of the remainder of the species vary in length but are shorter than the perches themselves. Not only does each species have a characteristic length of display perch and length of interperch flight, but the species also vary in the to 3 -a •s 60 p CO CO cu iH CO p e o 60 p .p o n at I >> CO iH P. CO o CO a) •H CJ cu a. CO .p •H W O CO § 60 o a 14-1 o p o cu CO X> o co 3 O 3 P •H 4-> P o o p 0) CO cu M p. cu n CO cu •H o a) a. CO n cu P. co CU P CU 3 60 p •H CO CO cO P. T3 CO u-t o 60 p 60 60 p c H -H o u cu o o CO 4J M— 1 •H J5 CO 60 p p •H O p. "H 6 +* 3 cO P. -U C 60 0) P -H •H H S O a o > !-( CU en o P 0) •H CO .o co o 3 CO O •H J3 P O >>iH H 4J CO ctf •M P P -H CU g 6 M 5 cu 6 *J W W wr o <* QJ I-t 3 60 •H 176 o "00 o * o • CM o u w CO o ON JZ o [] o CI A A A AAA Ttlfl A A 1 d cfl ■H CU C! rH O O g 3 01 n) -C to •A tfl U 0) >, >-l A o 4-1 J=. O i-> a ■H >, H PC o u n) D. to 0) c 0) o a) nj nj •H •H T3 N 0) a B A5 nj •rH C u 0) CD =) T3 X QJ n) o n) (0 CO •H •H o M )-i a at 0) >n r-i < W o 4-1 0) rH cfl > a) N CO •H Vj >^ 13 O a) rH o ?^ 33 C o T3 O CO >^ u x: a 177 presence and length of wing pumping behavior and in the presence and frequency of body rotations during the display. More detailed observa- tions are planned for the future to collect enough data on these and other species to quantify these apparently species-specific differences in behavior. Species-specific differences apparently occur in patrol perchers as well. Detailed observations are available for only three species, but all species observed in patrol perching behavior are listed in Table 13, which summarizes their perch postures. Ethnograms similar to those drawn for display-perchers have been constructed for three patrol per- chers and are presented in Figure 41. In general, it can be seen that the ratio of the length of the perch to the length of the interperch flight is much greater for patrol perchers than for display perchers (Figure 40) . This is to be expected since display perchers must fly relatively often in order to establish a pheromone-saturated display arena. Figure 41 shows that Mechanitis isthmia and M. messenoides are quite 4. similar in patrol perch behavior, although M. isthmia in general has more frequent interperch flights. Both Mechanitis species tend to assume patrol perches in open areas or along trails, while Thyridia confusa tends to perch in more dense understory. J_. confusa also has a character- istic wing flexing behavior in which the wings are lowered slowly until they form an angle of nearly 180° and then are quickly returned to their position over the back. Although Pliske (1975c) observed this same species occasionally perching with wings held open at a 60° angle at Rancho Grande, Venezuela, individuals at Limoncocha were never observed with their wings held open. The hairpencils do not appear to be exposed 178 x to 4-1 u 3 O o X a u 3 p-i o H 4-1 3 P-< 00 s •H 3 P en CU C g o 3 X UH o (» 3 u 3 4J (0 o Oh XI 3 H Px ^ o z "-1 2 H w W M P-i H CO H z o < P-l >j o § orf > H Pn w < H O P-i H PS a • CJ) O ^s O Pi h e > w w O <: p-i a CO w M U u PU w p ^ w M M Pi H o o CN T3 TD 3 a) 4J 4J 4J 4-> 4-> O O o o o o z Z S3 Z z z a) (D a> a) cu CU 3 3 3 3 3 3 o o o o o O z Z z Z Z z cu cu lO in a) Tj •H O 3 0) CO to QJ g co CO CO CO CO CO •rl •H •H •H •H U 4-1 4J 4J 4-1 4-1 4-1 •H •H •H iH ■H to 3 3 3 3 3 0) cfl CO CO CO CO X x x X X X u u o o CJ o o cu cu cu cu cu h S 2 2 S 2 •3 CU CO o a x cu o • Z CU 3 o z cu 3 o z o O O o CM O CM PS p o o X H a cu O H 3 <4-l Uh 3 -H O 4-> 00 t^ CO M 3 3 O Cfl •iH X & 6 00 >-i 4J • P*.-l CO n rH •H cfl 6 *3 cfl X 4J 3 • 4-1 CO CO i-H 4-1 M 3 4J O CO 3 3 O V-l O 3 3 4-1 N 3 3 CJ 3 •H O CU O M • O X QJ 3 N O CO 4-» X U •H X •H CU 4J 3 >-J X X CO O QJ • 3 H 3 CO ^5 X > CO cfl 3 5d • o •H rS • CU 3 3 to X X -3 QJ • 6 QJ T3 > -H CO cfl O rH CD 6 3 O X X X CO cfl E g 3 X 3 cfl ^J -a O rH I-l S ►« o o 4-1 •H a*3 ^III • -J T3 m X CO CU T3 CO O -3- U > CU 4-1 QJ 3 •H o X o 3 CO e 3 4-> X Jfl CO co 4-1 rH cfl 3 3 4J co Cfl •H CO cu <: 4-1 T3 U o TJ X < o O 4-1 CO >.m 3 to 4J X X 3 rH rH St CU 3 • 3 V. 3. o 3 X • 3 O >> o o CO u •H •H 4-1 3 3 4J .H -H •H cu 4J O •r-l X u 13 IH 4J 3. CO 0) X 00 3 U o C QJ cfl > •H ^ 3 •H QJ 4-1 ■i U rH H 3 iH u & 3 #i rH QJ O i • 3 QJ U 3 rH 3 to >, 3 CJ 4J QJ cfl 4-1 X H S w •H •H 3 CX oo 4-1 3 o i-l PC o 00 •3 CO 3 3 o 3 •H u 3 •V •H O N o B S u 3 3 00 4J 4-1 N •H 3 QJ S 3- 3 O 3 ■H H o 4J U •H 0) •H r< O 4-1 3 g to iH 3 4-1 Vj O O X cfl O 3 3. 3 QJ D-X X >> a- 4J •H *y% 4-1 rH 3 co n O »■ ■K •\ sj M-l X QJ o 4-1 P. to H 4J X o CJ VK : O 0) 0) 4-1 3 4-1 O o O o 4J 4-1 IM 3 )-l 3 O O O oo CO ■u o o X 3 X o> a-i c^ rH o 3 4J 4J •H X 00 W% #s 3 CO cfl 3 •n 3 QJ o rH rH 3 3 -H > 4-1 3 0) 00 00 CD 00 •H 4-1 •H rH rH 3 C 3 3 <4-l 3 4-1 cu iH rH •H •H ■H •H OJ O 3 3 > 3 cfl g £ g E3 ^4 O rH •H M Vj i-i i-j {-1 X* •-n QJ 4J 3 3 O O O O CO S 3 u 3 P* (X 4H I4H IH UH 4-1 S X iH w 4J >> 3 CU 0J QJ 0) 0) 3 Q) ^H U 3 3 cfl CO cfl 3 X 00 CO r-l 3 3 3 c 3 3 4-1 3 cu 3 >s 3 3 3 3 3 3 •H > U <-\ 3 QJ OJ QJ 0) 3 <4-l & H 4-> .-I 4J 4-) 4-1 4J 4-1 4-1 o T3 O 3 3 3 3 3 a 3 3 3 > QJ U «: < <8 > ■H 3 > 4J 13 X -H 3 O T3 3 X T) 00 3 cu 0) i-l > r-l CN rn -d- m vO 0) 3 3 00 T3 X •H 3 3 • • 4-1 CO rH • 00 > 3 O 3 & o CO 3 •H •H ^x 4-) •H 3 -3 3 4-1 •H X 00 t4 £5 CO 4J CO to 3 pbi •H w O CO •H 3 3 Cu -o S 6 cu o OJ X 5 > 55 b iH < U CO o IH 14-1 4J U-l •H 3 X 3 3 3 3 ma 0 o 4J 3 OJ QJ 3 4-1 4J X 3 o H < z * + • to 3 X) •H -a rH CO uh o 3 o • ■H CO ■U cu CO rH CO , c o rH -H TJ •H CO iH 0 3 P. O o CO ,3 3 U 4-1 3 M •H 4-» 3 60 3 c o CO «t •H o >. w ?*> J3 rH 3 rH CJ CO UH CU & U CD co is § o rH CM 1 4J 3 4-1 > •u o CO rH cu s a 3 O CO .fi & 3 U CU CO O O -u u 60- TJ -H CO p« 3 U 4-1 PL. cu •H CU * <0 rJ CO > co > 3 m CO !_i 60 V) O o CO CO CU 3 CU iH cu a. co •H CO 4-1 co •H .O 15 ^3 18 cu o CO o O > iH cu cu u CJ a. >^ >sj3 a ai Q) co ^Q X> 4J tH 00 CX J3 W u TJ 13 CM CO O cu a) cu O 3 CD p. ■U 4-1 4-> U-l CD cfl cfl X cO O M 0) •H -H a) -h .3 cu 4J 4-> rH .3 3 H s ■H iH <4-l O •H 3 3 rS"H «*H rH 60 •H iH H) M *-> O s rH CO CO m -h DO 00 60 60 4-) 3 CO O ,3 c 3 3 3 3 -H e CJ •H •H -H H 11 g CO 4-> u >, >. >, co e u M CU (U .-H <-\ l-l o cu 60 co O.U-! M-) MH CO 0 4-1 O , 1 3 A i > ) InJt" ,3 4-> CJ cO * o< Hi w w < < 1 M CU u 3 60 iH 180 ^ o 00 o m o o CM A A A o to O u w GO a O o CO ta cfl •H •H J C ■u d in 43 •rH u 0J a 'co a) T3 •H 01 o •a c *J a) •H CO c CO 03 cu 43 e a a) 2 cfl W nj 3 •rH U-l T3 c •H o ^ o >i .c H 181 during wing flexing, so the function of this behavior remains unknown. M. isthmia, M. messenoides, and M. mazaeus were also observed to flex their wings occasionally during patrol perching, but it was more fre- quent and regular in Thyridia confusa. Pursuit Behavior of Courting Males Up to this point the discussion of courtship has been concerned only with the mate- locating behavior of male ithomiines. The next step in the courtship sequence is the male's pursuit of the butterfly that — for whatever reason — has flown close by the male's perch and thereby stimulated him to pursuit. Most of the pursuit encounters I observed were between two different species and thus the male's pursuit was rapidly terminated. When initiating pursuit, both display perchers and patrol perchers fly directly and rapidly toward the potential mate (intruder) . Upon reaching the intruder, the courting male and the intruder usually "fly around each other" for a few seconds, after which the intruder resumes its normal flight and the male returns to his former perch (if the intruder is of another species) or continues his pursuit (if th& intruder is a conspecific female, or sometimes if it is a conspecific male). From this point, display perchers and patrol perchers appear to differ in their pursuit behavior. * A display perch male that continues the pursuit does so with a unique mode of flight that involves the rapid flutter of the wings through * Observations cover Aeria eurimedia, Godyris zavaleta, Hypoleria chryso- donia, Hyposcada kezia, Hypothyris euclea, H. f ulminans , Ithomia amarilla, I_. derasa, Napeogenes aethra, N. apobsoleta, N. duessa, N. pharo, Oleria agarista, 0_. kena, Pteronymia sparsa, Pseudoscada timna, and Tithorea harmonia. 182 a shallow arc. He generally flies 15 to 20 cm above the pursued butter- fly and about 5 cm behind it. The difference between the fluttery flight of the pursuing male and the normal flight of the pursued butterfly is quite dramatic at this stage, the more so because the pursuing male usually pulses his flight downward every few seconds, coming within 5 cm of the pursued butterfly. I assume the combination of the rapid flutter- ing of the wings and the pulsating nature of the flight propel the male's aphrodisiac pheromones down over the female. If the pursued butterfly is a male or a nonreceptive female, it generally continues flying in its original direction and may even increase the speed of its flight. A receptive female, however, lands on a leaf with her wings together over her back while the male continues his rapid and pulsating flight above her, gradually lowering himself toward her. Unfortunately, I have never seen an entire courtship carried to completion. Most of the females I have seen land at this point generally fly off again as the male lowers himself over her during the hovering flight. On two occasions, however, a Napeogenes pharo female remained perched while the male landed of the same leaf. As he walked up to her, curling his abdomen forward, she flew away and he resumed pursuit immediately. Apparently this se- quence may be repeated several times before copulation is effected. ' Pliske (1975c) reported somewhat different sequences for four species (Pteronymia nubivaga, P^ beebei, Aeria elodina and Ithomia iphianassa) of display-perching ithomiines in Venezuela. Males of these species fly above and behind females, chasing them with rapid, erratic flight that includes periodic darting pounces toward the females' s dorsum. Although no heterosexual encounters were observed to conclusion, Pliske did see the 183 termination of a male-male encounter between P_. nubivaga and Hymenitis andromica at close range. While P. nubivaga was chasing H. andromica, the latter male slowed his flight and canted his hindwings backward displaying the hairpencils, whereupon the pursuing P_. nubivaga male ceased the chase and returned to his perch. Pliske (1975c) has presented evidence that a lactone component of the male hairpencil pheromone serves as an allomone repellent to conspecific males and males of other lactone producing species. The lactone (derived from compounds obtained while feeding at PA plants) apparently facilitates male recognition of other males and allows males to terminate male-male intra- and inter- specific courtship pursuits, as described above for P_. nubivaga and H. andromica. Even so, physical contact between males may still occur. Pliske (1975c) recorded three instances of male-male contact for T_. confusa and one for Ithomia iphianassa. In these curiously violent encounters, a display perching male pounced upon a conspecific male, clutching the intruder with his legs, after which the pair fell to the ground and separated. 4 Observations at Limoncocha on pursuits by patrol perch males include only Mechanitis isthmia, M. lysimnia, and M. messenoides. Unfortunately, no pursuit encounters were observed for Thyridia confusa. A patrol perch male may maintain pursuit of a co-mimic of another species or of a conspecific male for 30 seconds or more before returning to a perch. For example, I caught the following three pairs of species flying within 10 m of each other within half an hour: Mechanitis messenoides and M. lysimnia, M. lysimnia and M. lysimnia, and M. lysimnia and Napeogenes apobsoleta. All six individuals were males. On several occasions I have 184 watched as many as ten tiger ithomiines mutually pursuing one another at heights of 5 to 8 meters in a sunspeckled open spot in the forest (Nature Trail) • Pursuits of conspecif ic females by males generally last longer and lead to the next stage in the courtship sequence. On two occasions I watched a courting pair of M. messenoides flying 2-3 m above the ground. In each case the male was flying about 30 cm above and 10 cm behind the female. The male's flight was not as agitated or as fluttery as was the flight observed in display perching males engaged in pursuit, but it did include the pulsing component that periodically brought the male close to the female. Pliske (1975c) has recorded similar observations on heterosexual encounters for M. isthmia veritabilis and M. polymnia doryssus in Venezuela, although the males he observed had their hairpencils exposed and erected and were flying just in front of the females instead of just behind them. This apparently resulted in an airstream (and, presumably, pheromones) passing down over the female from the male's beating wings. Instead of the pulsing movements I observed, Pliske reported that the males make 4. darting pounces onto the female's dorsum, as did the display perchers he observed. Pliske also found this same darting and pouncing behavior in males of Thyridia confusa psamathe (same as the Limoncocha subspecies) , which climaxed in the male clutching the female's head and thorax with his legs. Once this contact was made, the pair either fell to the ground, where copulation was effected, or else the male achieved copulation while the pair was still in the air. 185 Function of the Display Perch Pliske (1975c) has suggested that the lactone component of ithomiine hairpencil pheromone functions not only to terminate male-male pursuits as described above, but, when released by display-perching males, also serves as a territorial allomone that repels other males. Pliske reasons that by releasing such an allomone during a display perch, a male ithomiine reduces the probability that another male of any species will enter what I have called the courtship arena. The result, there- fore, is that since males are kept out, most of those individuals enter- ing the courtship arena will be females. As evidence for this interpretation, Pliske cited his findings that the attractiveness of dried Heliotropium indicum baits to male ithomi- ines was greatly reduced by placing the lactone component of the ithomiine hairpencils in competition with it. Even though this highly artificial experiment provides convincing evidence that the lactone can be repellent to conspecific and other males, there is no evidence that the lactone is being released during the display perch for the purpose of repelling other males. Indeed, Pliske admits he has never observed a male exhibit avoidance behavior upon approaching another male engaged in a display perch. It seems much more reasonable to me that a male ithomiine in a dis- play perch is releasing a female-attracting or aphrodisiac pheromone for the purpose of "enticing" passing conspecific females into his courtship arena. In this manner the male increases the probability that a butterfly that enters his arena will be a conspecific female — and not just any fe- male, as the male-repellant theory necessitates. Once while watching a 186 male Godyris zavaleta in a display perch, I saw three different G. zaya- leta females enter the male's courtship arena from three different compass directions, two of them apparently changing their course of flight to do so. (G. zavaleta is sexually dimorphic and the sexes are thus easy to distinguish). The male chased all three of them, but, as none of them landed, he returned to his original perch after each encounter. Although such observations imply the existence of a female-assembling function of male display perching, behavioral experiments are needed to determine the exact nature and purpose of this behavior. Perhaps the display-perching male is releasing both the lactone repellent and a female-assembling aphrodisiac simultaneously to maximize his probability of a conspecific heterosexual encounter. Mating Because mating butterflies tend to remain motionless, they are not often observed and information concerning the behavior of in copula pairs is scant (Miller and Clench, 1968). During the course of this study, a total of only 18 mated pairs representing 6 species of ithomiines was observed in the field. These observations, grouped according to time of day, are presented as a histogram in Figure 42. The specific time of day and perch height for each pair are given in Table 14. Based on this small sample, copulation appears to peak between 1200 and 1300 h, just after the peak occurrence of courtship activity at 1100- 1200 h (Figure 32). This peak in copulatory activity closely parallels the general pattern of mating times of butterflies of the temperate zone 187 m CD •H u a) a. a rH I o m o m o m o in -<* rH o o m rH o o -a- rH o >H o <: CO a rH (x, o • O CO O w CD CM S C rH r-l •H H •H S O o o .G rH ■U rH M o O o o >> rH iJ •H > o •H O •U <3\ O o <: 60 o c o •H 00 W o CO o o • r^ CM o 0) o u o 3 vo 60 o •H Pm NoixvinaoD ao SNOiiVA'aasao ao uaawnN 188 z o o z OP* oo § M H l-l ox> z, <->o W OH > zx HI >-i o UUDJ JO o o ■ M IV. H ►h a: E1 XOb a jc o •■3 c 41 • O.JJ OJC 00 • vl 00—1 C«J o •o i-IJC «l ■>-> j= oo c 41 111+ + I fi. • tn Ivi U, Sn (v. p« • HhH I* I HlhHHhH I ■b o- ■bo- S I U.MH I % OHOmHHHHH IA tni^lA o o o m in o o o m o mom m %j -» *•* m m o m »-• o r>nn o» ^> i, u, 0. b. O* * * * o m O «n o i-l h o a. NOC WH 01 « > ;» a c >a,a: UkUUbOOUh o H>> O O O O Z O Z Z cl rtNM > Z O 5 z •-» o -» .-• CM a: ft. < >H OO Z O oj m a 1 1 J-t O CO o> a) -J --J u u ■-J tJ C C a a x: jc J-J o h •^1 3 u 4J o a. « « U «J J= h u 41 4J u c a C ■w 01 It a •rl a 41 a- c rH ■H JH s O o M-l o o rH iH O > O -H O 4-1 i-i < o c o o tA o t-l ca o o o Q. 00 -H o > o o o • r^. CO o cu o u o 3 \o fc3 o •H 197 cover 22 species, occurred mostly in the afternoon and reveal a peak of activity between 1400 and 1500 h. Thus, the daily sequence of events in the reproductive activity of the ithomiine community parallels that which occurs in the life of an individual female; first courtship, followed by mating, and finally oviposition. This diel sequence is illustrated graphically in Figure 44, in which the number of observations of courtship, mating, and oviposition during each hour of the day is expressed as a percentage of the total number of observations of that behavior. From this graph it is seen that the peak periods of activity for courtship, mating, and oviposition behavior represent about 35, 40 and 25% respectively of the daily totals. The reasons for an afternoon peak in oviposition activity are not clear. Since the number of days that a female engages in mating is relatively small compared to the number of days over which she lays eggs, it seems unlikely that afternoon oviposition represents a delay caused by prior mating activity. Perhaps it represents a delay resulting from morning feeding activity. As mentioned earlier, female ithomiines derive i most of their nourishment from detritus-type food items (e.g., bird droppings, falling pieces of fruit, etc.) which usually become available in the morning as a result of early feeding activity by birds. Another possibility is that, for unknown reasons, the afternoon is a better, or safer, time for laying eggs. That the afternoon may be a safer time for oviposition is suggested by the behavior of females of the genus Mechanitis. Female M. isthmia, M. lysimnia, and M. messenoides predominately oviposite on solanaceous plants characteristic of sunny second growth areas (see CO * /-N QJ ^-s CO •H CO CU O CO iH CU tH CJ • ex a a) CO CO CU CX ^\0 N > «•-• %■• •H co ■u C to co U o c c < «H O O 4-1 «H -H a> CO 4J 4-1 > > to co •H M > > 4-1 -. CJ CO cu cu 3 ,£> CO CO T3 o ,a U3 O o o Vi o ex CO 00 vO cu Hrlst Pi II II II Q) c zzz •H •H g a J= o ■u CX -H M •H 4J J= -H tt-l CO 60 CO O 4J C O Vi -H CX ^ 3 4J tH Vi O CO > CO u £ o OD< -a- o ion H to—' 0) x> 5 f5 u U M o: u to u M O o U < < ■< uououuuuu u o u > > > O Q Q 0» ■ on o CO CM o o o o O t-< o •a CM r» to o X H I-I H ■H I C o CO V Id o S3 z M •J oi X o. o c OI B X a o c CI E t« . ■•< e *— . 01 c u o x £ X r-l 'oi <*4 C CO u u u AJ X 1- o JJ to 09 CO 3 1 -O 00 -W c C C to pa CO AJ t C 01 at ftl 01 u U. — 1 01 eg a a i-i 03 \) > X g X X X eg ~r* 3 01 •o *-^ to I-I O CO X CO o to T-l 01 1-1 It u -s >, £ •*4 -r4 00 CO 4J 3 C >-. u XJ U r-l >l N c i-i 3 3 3 e> 01 0) r-l CO CO •0 CJ SI •H > nj •9 S E (0 01 a CO ro 03 •r-l *f4 ■> tn «0 i_> —< ■H •*-! -r4 p* rr 3 — i —4 to e B b E 6 n 01 <^ a u x x X jC i-i •r-l 09 CO CO c u u 4-1 4J to in rj C e 3 CO CO 03 CO >- Sh CO S 01 E 01 E (4 •I a) o ^01 c c ■ R U r-l O O •o a 01 XI c 09 JC C CO x: to x: c IS x: c CO X 01 C CO x: o c CO X o a c a X 203 Q H «/>N l H i-i Ik OZ J O o< >-< MZ U >oo OHZ He* us. o wo S6.0 U oa, 10 PC 10 U H o « w Z 3N oun «z too ON E UH | •a w a c 01 5 •-t t-t r-t ft .-I •» CI o u a eiec u U U U U CIS u o cj w M to oo. M M Id M W « H w W o QO U U o o uu u o u o u o u o u o oo o > >s» >>>>>> > >> >>>>>> t-i no ■"-< •-< t-l r* — I o o o u o> mo cr> o •* CM x. u ■J) M c ■»< 3 <0 •< — ■ a at a •a c X id N-* ^-t > a 9 c U M 01 0) a. .O O. t4 O) 0 o> •H •U 01 01 r-t XI a o OI JC -o 4-1 ca c o a CO X £ «0 .c M o K OI as O •o c t- t-l 01 0) o 09 £ U. X i-l (1 a CO a B C (C u 01 ai 3 a a .0 aj to X. a a • m o al u 3 — * co CD -h| a C U > a a co| 0} c « a -r4 c pharo euryan B 01 i— i VI 3 o MH o m CO ■ CO CO CO CO -^ ■H ■H •H ■H 01 01 CO Vi I-l (J >J «J c c f-i ^ > >> > !» 01 01 3 XL X X X x IX ofl •o !-> i-> u AJ i_* o o -w o o o o o 01 01 (1 G, G D a a O. a CO >i >. >-, >- >■ CO CO f-f X X X xl X z Z a. c o 01 c ai CO X r o CO CO C9 CO t4 •H •^ •rl e B E e o O o o JC .e X. .c i-l *j is ij l-l M IH M 204 B % + MZ U >oo OMZ H W UJO-M 3 O CO O ou.ou IOO.W : o to wo ON £ Ot-i E 3 c g w r-l ■s CM •-! «» \^ *m* V-T ee o n: O M O • •» r-l m m «»> m» .-i **» —I r-l r-l rj > O r-l «l > I u o I > > I > > I II I 1 I I U II CJ i lUlOl I I O U l I > > • I I >>>>>>>> >> l> •>>>>> I I i > > SO M « * r-« U « • (-. c w a as t-< o i •o --1 c k "J > ai U ■J o o a o 2 H w I w «3 z z u o o kl r-^ co kl CO ^8 2 2 o u rl 4> I c 01 a t< Q. E r-l 0) 3C ki u I a c 3 W B3 a n n ki • at «j c -H JO o ■3 II .X o. c 1 1 c 01 to 205 ■ + M Z U >ou blOt-i 5 a oio Ofcou ZOMn bl U H o uiu Z 3N KOUC1 1*1 UN E OH | to w- •o V 3 e c o o a H in u M u u m a z < u pa t-i •"-" .a SO M O id e g CO M « M M CM r-l —l .-« — I | <-< I II | III ■ tfl CO I I < , III >>>>>o> > > > > > > S» > > > O > > « > > > o> CM oi a. a c at co X z a 55 a 8 a •»4 « u u a • r* Ol u 0 01 c a> ■r* X XI u V> o. :3 a 01 01 X fc CO c a (1 c •-1 r-» "D 3 a <0 0) 01 > > n a «9 01 N N ■H 01 CO x ca CO m u -v4 -w O > u u U AJ > >, 01 1-1 •0 T3 *-» O. 0 0 01 u 0 0 X CO co c o c o c o -I v. ^ ai a' s I £ •r? * E >■ X M 01 o o. >- X u o Q. X 3 o> x co o fH a 10 • X >« >« X « •0 s < 0 0 ■0 o-. Mr .-i 0 CO •0 c-. ©> CM c -0 • •H CO a o\ u t-l r-> u • g OF >» - • JO a. cw 0 X r-l a. 0. 1- 0 •rf 3 01 CO u. O u to M> r«. co o» 01 -C a •H r-t JO CM CO -3- 3 CM r-i .» >. D. a* CT> 3 S •-I r-< iH xi CO ■ ■3 a 0 •O -0 ■O *j c *r4 •H ■H c 0 01 OJ 01 01 H u B 01 ^* 1-1 rH 01 3 < < •< hi U p. O •O •0 TJ «-« cm «o •* m 206 o g • •M "O i-l 8 Of u -1 B • cMJ to 5 ■3 o. 1 u • • a « CJ X r-i «l m 00 CO -M Q 9 u 01 B CJ 00 , 4J 01 ■M o ■ u 01 -o to •M i-l ■ ox § eg B 41 01 o B V ox • •-* > CO l— I CO »n c •M •M to -»4 u 01 T3 00 U 01 • 3 ■w B C 01 •M > • eo t-l o CO 01 U M « E C a •o M a. ■M • • r-l Ot4 u X CM 01 *s ^-1 > > ■ ••4 4J o eg « » CO 01 .01 • • • u CO r-< 01 o o t-< rH B > V* — 1 — ■ • 00 .-1 >« u O , 01 01 • • • to -D 01 CO c M •M t-l o > t-l f« o C 3 •M cm 4J 01 B B .-I 01 01 01 • 0.0 o o ^^ O 5 c 01 • CM •M 0) — 1 u -t4 M CO CO ^^ m in t-l •) 8 8 m > i-l to 3 41 o *-> •M • •* O CO o x> eg a) CJ U CI CI CO •M tt B o o U e 00 eg 01 01 41 • *-s. o oo CD CM 01 •v at >- cm a •o u U **^ ■» o o o • CO o ~> oi M X U > X o. •I CO ^ 01 m > O o M 00 3 CM 4-1 o ^^ H a o -M C7> « a • ■ CO 0)t-4 ~-i ■ CM CO to —1 )E •-» > *•* H **\ a TJCM t-4 0) •M CO 01 B ei CO tg a 41 • a 01 O AJ u O CM -j u eg 01 ~J •M •»4 -M u u •M E y— V M 00 a o E 3 u ■ 41 f-l eg X CM CO B > a O a 8 E •o c CO 3 —1 M CM 1_! O M eg s to ci <~* ^4 r 0> C 3„ «M i-i O -M 01 CO 2 o E u o 8 ti -o -M-H 0J eg CM o c 3 41 o i-l to a oo a 44 — 1 X X > o 01 CU i-J M J U ~* o CM 01 -o CO V u 01 > o CO CO CO tg 01 M X — 1 t-4 X m o eg 01 u eg E CJ CO 41 3 •M at • • o. M O eg 4J ■J u eg c CM O CO C B CO Cu M **k a B o hi eg CM o M ■M -M ■»-^ • CO a 01 o u o £ O- CO § • >M M 3 i-l ^ X m W ■ o X o. CO ci U u I-l 3 CO CO 01 ^-* tg S-* to r-» M a c 3 01 CO o #-^ 3 CO ■o o *-* c M s^ 35 z M -rf eg XT) c o to c •-4 o eg M M a B B u .X v^ r-l CO eg «M -M 01 -M CI Jc: a o o 5 B 4-> a •-I eg «~l M u X 01 •*4 01 5 ^«* •M o 3 COX M a) M «J u 41 M 3 CM m CI ec c X -M 1-1 0) ■ CO M -o c -M a. 3 a- •M •*^ 3 a o J CM' M 1 -*M M c c 41 cm BO o •H u H r^ w 01 t-l CO r u eg O eg a > — i 41 CO i-i a a CJ %• 3 X *!** ■o 3 < o •I •o > t-i E a. > ■o AJ 01 O S X c t-4 o c 41 » • eg H a t-i B a • P -M AJ c C i-i O O- t-l CO X a 01 a 01 O B — < a U t-i «J 01 •M o o co •M CO -i to J oo X O •o t-i 01 CO eg U X o CO CO N -M •g 5 ■«-' B M Cm 60 X 3 i-l CO CO 01 03 O CO 5 ■o a a CO a a 5 c 01 i->X 01 oo 00 -~! 00 B M o -l CI •-i a 1-4 -T-* 1-t o (1 00 00 ^H 41 to •g E E a X a 00 CO - i-l eg « 01 CM i-i ■ M «M a •M u o ci E o 0) c n CM CO eg c o M IM CI b. CO ~n o t-i no c« >, eg CO M to CO c >> X C t-l 41 c 41 a u H B a t-l CJ u u 9 >N >^ 01 eg i-i 01 3 to A ■M • 3 ta .B 14 XI B o 01 CO eg CO X w) X CO > lb • O 'I ■ H r— i o U-i w* •-I i-l 01 i-l Pm • > > • • en X '->< 01 t-l — 1-1 3 o o • O » ■ •*4 > ■ 4J o < OS eg m • • CM eg O o 1-1 u ■ U > > ■J o ■ 01 1-1 I— »4J M CM >M CM eg u CO la) B a. > Cw > o >4 > J£ a. o CJ eg a CO 01 *--! to eg o o u «• a. o u c t» E 01 o> 01 t-< •o >s CO 0) M CO u B. ■ M CO 01 Q 01 f-l »-4 t-* E ^ u-i u P. B a. > a P. 1- u U ts u > CM i-l X o o a E 01 o X XI CO CJt-4 IM CM CU cu o a! 1-1 a a 41 a a M CJ o I— 1 z oi M I o O o en g CO •M (J -M li -r4 o a. B u CJ sa J 01 01 3 41 M a 13 to •o ►1 *4 3 e CO r-« > OJ 01 01 o -^ CO i-i VI E 41 U 01 o X B — i I o-e B u M CM UJ M H CV. to U O X Pa X to z o. X M r-i 01 C/0 *M u u M 3 eg 01 ZTJ f— ■ eg 3 3 3 CO CJ U CO o ■z at CO CO CO o U C hH 01 «ct f-4 C9 CO H-M ,_} r-4 r-t r-C eg »-^ *s ^ v_^ N-^ ^^ ^-y ^-* ^^ s-^ t-^ W 0) 1-1 P- U-! CO CO CO M •o a a 6 to c o O eg CO CO u 01 •c X o o 8 M U u C u 111 o 1 1 p. --4 01 O O o 01 01 3 o r-l U u- t-l ■o -a T3 > Q a C w > — r >-H eg B u ■M o O » u. CO •a ■o •o •v N u a 3 3 > u o pq r-l c c c C CO a 3 rl • to a. a £ «H c c -a M CO C 60-H •H C! •H o c to o e •H CU u J= -H X. 0( o M O t^ 4-1 »H CJ P..C at Vj •a •H 01 a* co -u r-i ■H o H S3 S 55 M o o o ■ D 4 0) o > 60 60 W CO ■vj" 0) Vi 3 60 •H FX4 210 is, however, one significant exception to the pattern in the Dircennini, 3 Pteronymia notilla (0.94 mm ), and in the Godyridini, Thyridia confusa 3 (3.08 mm ), and a lesser deviant in the Ithomiini, Ithomia amarilla 3 (0.46 mm )]. Second, in the Mechanitini, the four Mechanitis species and Forbestra truncata are relatively similar in egg size and wing length, while Xanthocleis psidii departs from the pattern with an uhex-~ pectedly small egg. Scada batesi, another member of the Mechanitini, is the smallest ithomiine represented, but nevertheless has an egg size larger than most members of the first four tribes discussed above. Third, the most interesting pattern is that shown by the Oleriini, which, when compared to the other tribes, has an extremely large egg for the size of the female. The Oleriini species show a sharp increase in egg volume with body size, suggesting that selection has tended to maximize egg size in this group. By contrast, there appears to be selection for a relatively small egg (independent of adult size) in species belonging to the first four tribes discussed (Napeogenini, Ithomiini, Dircennini, and Godyridini). The extreme departures in egg 4. size from the tribal means shown by T_. confusa, P_. notilla, and X. psidii probably represent adaptations to unique situations where selection has favored very different oviposition strategies. Until these species are better known biologically, the significances of their aberrant oviposition patterns will remain unknown. Unfortunately, data on the Melinaeini and Tithoreini are too meager to drawn any general conclusions, The reproductive effort (total number of eggs laid multiplied by egg size) of butterflies is nutritionally financed by feeding — either in the larval stage, the adult stage, or both. Labine (1966), in her 211 study of the nymphalid Euphydryas editha in California, found that females emerge as heavy-bodied adults with large numbers of eggs already mature in their oviducts, proving that the bulk of the repro- ' ductive effort in this species is financed by larval feeding. By contrast, females in the neotropical genus Heliconius rarely lay eggs during the first few days of life and probably have no mature eggs at the time of eclosion (Gilbert, 1972). In Heliconius, then, adult females must obtain most of the nutrients required to manufacture the eggs. The situation in ithomiines appears to be the same as that in Heliconius. Dissections of wild-caught adult females of several species showed that very fresh (F+) specimens were rarely mated and carried no * mature eggs (N = 8 ) . Further evidence for adult financing of egg pro- duction in ithomiines comes from reared virgin females that were fed sugar water after eclosion and then dissected at various ages. The following females showed no sign of egg production (time after emergence in parentheses): Ceratinia poecila (13 hours), Napeogenes pharo (1 day), Mechanitis isthmia (1 hour; 3 days), and M. lysimnia (1 day). Older reared females did show evidence of egg production. A seven-day-old The following wild-caught females were dissected and found to have no mature or immature eggs (only thin oviducts and yellow supporting tubules were present): 1 Godyris zavaleta (F+) , 1 Hypoleria chrysodonia (F+) , 2 Napeogenes apobsoleta (F+) , 1 |, pharo (F) , 1 Oleria agarista (F),~l Oleria lota (F+) > 1 Scada batesi (F) . G_. zavaleta, N. pharo , and S_. batesi each carried one spermatophore, but the remainder of these females were unmated. 212 Mechanitis lysimnia female had many tiny (less than 0.1 mm diameter) embryonic eggs in her oviducts, but no mature or immature eggs were present. Two reared M. messenoides females were kept alive for 21 days before dissection. One contained 10 mature and 9 immature eggs; the other carried 3 mature and 12 immature eggs. Both females also had numerous tiny embryonic eggs. These preliminary results indicate that ithomiine females emerge with no mature eggs in their oviducts and must pass through a period of at least several days of egg production before oviposition can begin. Although mating can occur before egg production has begun (as in the females of (J. zavaleta, ftf. pharo and S^. batesi cited above) , it appears that many females are not sexually receptive on the day of emergence and may not become receptive until after egg production is under way. Judging from the data on reared Mechanitis, it probably takes females two weeks to produce mature eggs if they are limited to a sugar-water diet as these were. A source of nitrogenous compounds is probably essential to the longevity and continued egg production of ithomiines just as has been shown to be true for Heliconius. Gilbert (1972) has demonstrated that many Heliconius females obtain amino acids and proteins by gathering and digesting pollen on their proboscises and use these nitrogenous compounds in the manufacture of eggs. Fe- males without access to pollen sources lay fewer eggs than those that can obtain pollen (Gilbert, 1972). Female ithomiines probably obtain many of the amino acids and proteins needed for their egg production from bird droppings and other nitrogen-rich detritus on which they feed 213 (and perhaps also from decomposing spermatophores; see above). If the two M. messenoides females I kept for 21 days on only sugar water had had access to fresh bird droppings as well, perhaps their rate of egg production would have significantly increased. Male ithomiines probably obtain the amino acids needed for a long life span from the nectar of the flowers they visit, for Baker and Baker (1973) have recently shown that flowers visited and pollinated by butterflies have nectars with high concentrations of several different amino acids. Returning to Table 15, it is clear that the majority of ithomiines lay their eggs singly, just as do the majority of all butterfly species (Labine, 1966). The major exception to this is the genus Mechanitis, most species of which lay clusters of eggs ranging in size from 3 or 4 to nearly 60. M. isthmia and M. mazaeus seldom lay egg clusters of less than 10 eggs, while M. lysimnia and M. messenoides rarely lay clusters containing over 10 eggs. The eggs laid by M. isthmia, M. mazaeus and M. messenoides are tightly clustered and evenly spaced, while those of M. lysimnia and the closely related Forbestra truncata are more widely spaced in a loose cluster. There are no other species at Limoncocha that lay egg clusters, but large egg clusters have been reported for Hypothyris euclea in Costa Rica (Gilbert, 1969), and for Placidula eurynassa and Episcada clausina in Brazil (d'Almeida, 1922). Although egg cluster size appears to be constant within a species population (as at Limoncocha) , some widespread species may have very different oviposition strategies at different localities. For example, Hypothyris euclea, an extremely widespread and usually 214 abundant ithomiine, lays large egg clusters in Costa Rica (subspecies leucania) but deposits eggs singly at Limoncocha (subspecies peruviana) . It is interesting that H. euclea appears to be much less common at Limoncocha than has been reported for Costa Rica (Gilbert, 1969), Trinidad (Barcant, 1971), and Brazil (Brown and Benson, 1974), although the egg cluster, size in the latter two areas is not known. Another case of subspecific populations differing in oviposition strategy is that of Mechanitis menapis, which in Costa Rica (subspecies satura) apparently lays clusters of eggs on the ventral surface of the foodplant leaves (Gilbert, 1969), but in western Ecuador (subspecies mantineus) lays single eggs on the dorsal surface of the foodplant leaves (Drummond, unpublished). Two other published reports suggest the occurrence of variation in oviposition behavior within a species, but these appear to * contain incorrect information. ra« = Egg clustering in ithomiines is clearly correlated with larval gregariousness, although a few species (such as Forbestra truncata) that lay small, loose clusters have non-gregarious larvae (Table 15). * ;Young (1972) mentioned that-M. isthmia lays egg clusters on the ven- tral surfaces of leaves in Costa Rica (subspecies isthmia) , but this is undoubtedly an erroneous statement since Gilbert (1969) clearly reports that the egg clusters of M. 1. isthmia are laid on the dorsal surface, just as they are at Limoncocha (subspecies eurydice) and in western Ecuador (subspecies chimbarazona) (Roy McDiarmid, color trans- parency) s Young mentions in the same paper that Godyris zavaleta caesiopicta in Costa Rica lays "loose clusters of eggs on the ventral leaf surface of Solanum hispidum," but, in his life history of this butterfly, Young (1974c) states that "the eggs are laid singly" and that the female he observed "never laid more than one egg on a leaf." 215 Egg clustering is apparently not closely correlated with the position of the eggs on the leaf, although most species that lay clusters do so on the dorsal leaf surface. With the exception of M. menapis satura in Costa Rica, all observed Mechanitis species deposit their eggs on the dorsal leaf surface, as does the closely related F_. truncata. The other two species in the Mechanitini for which records are available (S. batesi and X. psidii) lay single eggs on the ventral leaf surface. Outside the Mechanitini, only one ithomiine species, Hymenitis nero in Costa Rica, has been recorded to lay its eggs on the dorsal leaf surface (Young, 1972). H. nero is even more exceptional in that it is the only ithomiine in which the larvae are known to feed and rest on the dorsal surface of the foodplant leaves. Modes of Oviposition Behavior In all ithomiine species studied so far, females lay their eggs on either the upper or lower surfaces of a foodplant leaf, and rarely, if ever, on petioles, stems, fruits, or flowers of fhe larval foodplant. Eggs are generally laid on mature rather than young leaves (Young, 1972, 1974c; present study), suggesting that there may be chemical changes with age in foodplant leaves, rendering older leaves more attractive. The behavioral sequences leading to foodplant recognition by female ithomiines are not well understood, and the chemical and/or mechanical stimuli that elicit these behaviors are as yet unknown. In attempting to locate a larval foodplant, females characteristically fly slowly and evenly through the forest understory, weaving in and Figure 46. Modes of Oviposition Behavior. a. Mode A. Female lands on dorsal surface of leaf, lays eggs on dorsal surface of leaf. b. Mode B. Female lands on dorsal surface of leaf, lays eggs on ventral surface of leaf by curling abdomen under. y c. Mode C. Female lands on dorsal surface of leaf, then flies to and lands on ventral surface of leaf and lays eggs. d. Mode D. Female lands on ventral surface of leaf, lays eggs on ventral surface of leaf. 217 218 out of vegetation, but rarely stopping to perch. Once a foodplant has been located, presumably by olfaction, the female generally flies around or near the plant for some time (from several seconds to a few minutes) before landing on a leaf. My observations at Limoncocha have resulted in the identification of at least four distinct behavioral sequences (modes) leading to leaf selection and egg deposition. In Table 15 these have been designated Modes A, B, C, and D. These modes, illustrated in Figure 46, describe behavioral sequences that begin after the adult female has located the larval foodplant. Mode A results in eggs being laid on the dorsal leaf surface of the larval foodplant. Modes B, C, and D all lead to egg deposition on the ventral leaf surface. Other modes of oviposition behavior may exist in the Ithomiinae, but only the four described below were observed at Limoncocha. Mode A (Figure 46a; description based on Mechanitis isthmia) . Only Forbestra truncata and species in the genus Mechanitis were observed to use this mode of oviposition at Limoncocha. The female generally spends a relatively long time circling and "inspecting" the foodplant 4 before alighting, which may happen several times before an egg is laid. Some females may hover above a leaf for several seconds before descending to perch. After landing, usually near the center or toward the tip of the leaf, the female remains motionless for a moment or two, and sometimes for several minutes, before the abdomen is lowered to the leaf and the first egg is laid. If the species lays single eggs, as does M. menapis mantineus in western Ecuador, the female flies off immediately after raising the tip of her abdomen from the leaf. If several eggs 219 are laid, the female stays motionless for several seconds before lower- ing her abdomen again. In M. isthmia the time elapsing between the deposition of one egg and the next decreases as time passes. In species that lay tight egg clusters (M. isthmia, M. mazaeus, and M. mes- senoides) , the female palpates with the tip of her abdomen to find the last egg in a row so as to lay the next one adjacent to it. After completing a "row" (the width of which is apparently determined by the distance a stationary female can angle her abdomen in either direction from an anterior-posterior axis), the female moves forward slightly and lays another row, and the sequence continues until the egg cluster is complete. It is not known how "completeness" is determined, but it may simply be the result of the number of mature eggs present in the female's abdomen. Mode B (Figure 46b; description based on Godyris zavaleta) . As in Mode A, females using Mode B land on the dorsal surface of a leaf, but usually with much less time spent "inspecting" the plant in flight before landing. Shortly after landing, the abdomen is curled over the edge and up under the leaf, the female adjusting her position to maximize the dis- tance she can reach under the leaf with her abdomen. All species observed with this behavior at Limoncocha utilized foodplants with rela- tively small leaves, so that little body adjustment was needed. Oviposi- tion takes place along the side of the leaf and rarely at the tip or near the base. All species known to display this behavior lay single eggs and females fly from t-he plant almost immediately after the egg is laid. An interesting outcome of this behavior is that the eggs of a given species are nearly always found at the same distance from the leaf 220 margin. At Limoncocha, I became aware of this because I routinely drew the location on the leaf of each egg I collected, and very early made an association between the mode of oviposition (B) observed in « Godyris zavaleta and the distance of her eggs from the leaf margin. Therefore, when I found a consistent marginal distance in placement of eggs of Thyridia confusa and Xanthocleis psidii, I predicted that both species would use Mode B oviposition behavior. I was fortunate enough to subsequently observe oviposition by a Xanthocleis female and the prediction was fulfilled. Although I have never , observed ovi- position in T_. confusa, the prediction of Mode B is strongly supported by the consistent 10 mm distance of the eggs of this species from the foodplant leaf margin. Mode C (Figure 46c; description based on Ceratinia poecila) . Once again, the female lands on the dorsal leaf surface, usually after much flying about the foodplant, but with little or no prior inspection of the particular leaf on which she perches. Several leaves may be visited in quick succession before a suitable one is found. If the leaf is in- appropriate she stays but a second before flying on. If the leaf is correct she stays only a little longer (perhaps two seconds) before flying slightly upward in a hovering flight, and then down and around to the underside where she lands on the ventral surface. After landing, the abdomen is slowly raised to the leaf and a single egg is laid. Immediately following the release of the egg, she flies off to repeat the sequence, usually on another leaf of the same plant. Mode D (Figure 46d; description based on Hypothyris fulminans) . This is the only mode in which the female first lands on the underside 221 of the leaf, and she does so after hovering around, and especially under, the leaf for some time. Only one egg is laid, after which the female immediately flies off, usually to repeat the sequence on another * leaf of the same plant. Obviously, Mode A is used exclusively by species that lay eggs on the dorsal leaf surface. This includes most of those that lay eggs in clusters. Unfortunately, it is not known what mode is used by species that lay clusters on the ventral surfaces of leaves, but Mode B can probably be ruled out because of the physical difficulty of laying a great many eggs in a tight cluster on the underside of the leaf while curling the abdomen around from the dorsal surface. Of the remaining two observed modes, C and D, I suspect the latter, for reasons given below. Based on my observations at Limoncocha, Mode C is clearly the most inefficient. On numerous occasions I have watched a female of Ceratinia poecila or Ithomia agnosia land on the dorsal surface of a leaf of the correct foodplant, then leave the leaf to fly to the underside, only to land on a leaf of another, usually unrelated, plant and lay her egg there. The females that landed on the underside of an inappropriate leaf, after first perching on the dorsal surface of a correct leaf, always proceeded to lay an egg, never correcting their mistake. It seems apparent that the stimulus eliciting egg-laying behavior is received during the perch on the upperside of the leaf. Once the female leaves the upperside of the leaf, she will lay her egg on the next "underside" she encounters, even if it be an incorrect one that happens to block 222 her short flight to the underside of the correct leaf. This explains why I suspect that species which lay egg clusters on the underside of a leaf employ Mode D (or some other, as yet unobserved, mode), because the chance for error in Mode C appears too great to risk a large batch of eggs. Interestingly, Hypothyris euclea, which lays single eggs at Limoncocha (subspecies peruviana) , uses Mode C. Since the same species lays large egg clusters on the underside of the foodplant leaves in Costa Rica (subspecies leucania) , it would be nice to know if the subspecies there follows the prediction that Mode D should be used. Both Modes B and D should provide less chance for error than Mode C, but they may still allow mistakes to be made. It is not surprising then that the "safest" of the four Modes (A) is utilized by most of the clustering species (Mechanitini) . Since single egg deposition on the ventral surfaces of leaves is the most widespread condition in ithomiines, it is probably also the ancestral condition. If, for unknown reasons, selection should favor an oviposition strategy utilizing egg clusters in certain species of ithomiines, one would expect concomi- tant selection to minimize egg placement mistakes. Perhaps the switch to the dorsal surface of the foodplant leaf in the egg-clustering species of the Mechanitini is the result of just such a selective process. Gilbert (1969) suggested that the dorsal position of egg clusters of M. isthmia facilitated their relocation by the female if she were disturbed during oviposition, implying that there is a critical minimum cluster size. I could not elicite such relocation behavior in females of this species at Limoncocha (N = 2) , and since the egg clusters of M. 223 isthmia were quite variable in size, ranging from 10 to 59 eggs, there is no evidence for egg cluster relocation behavior at Limoncocha. « Foodplant Specificity The relationship of foodplant specificity to oviposition strategy in the Ithomiinae while undoubtably important, is unclear at present. At Limoncocha, most of the species in the genus Mechanitis have broad foodplant tolerances, and considerable overlap in foodplant use as well (see Chapter 6) . Most of the f oodplants utilized by Mechanitis are fugitive or weedy species that colonize open areas and are thus patchily distributed both spatially and temporally. Perhaps the strategy of egg clustering permits better exploitation of these isolated patches when they are found. For example, because of this patchy distribution, Mechanitis females may spend proportionally more time flying in search of f oodplants than species that utilize forest plants. As a result, energetic constraints may favor the deposition of eggs in clusters in most Mechanitis species. Also, if larval predators are more common in open areas than in forest situations, the larval gregariousness made possible by egg clustering might have a defensive function resulting in higher progeny survival rates than would be provided by single egg deposition (see Chapter 5). Wiklund (1974) has raised the question of how the suitability of various plant species as food for the larvae is interrelated with the oviposition preferences of the adults. Three possible alternative explan- ations were identified: (1) the larval food is somehow remembered by 224 the adult female and predisposes her to lay eggs on the same plant species as that upon which she fed as a larva (Hopkins' host selection principle), (2) the larval foodplant suitability and the adult oviposit- ing preferences are determined by the same gene complex, and (3) the larval foodplant suitability and the adult oviposition preferences are determined by different gene complexes. Based on his elegant experi- ments with Papilio machaon L. in Sweden and a review of the literature, Wiklund (1974) disproved the first possibility (at least for P. machaon) and concluded that the third possibility, separate genetic control of foodplant tolerance in larval feeding and female oviposition behavior, was the most likely. Part of the argument against the second possibil- ity (a single controlling gene complex) was based on numerous observa- tions of larvae successfully feeding on plants never oviposited upon by adults of the species, and of adults ovipositing on plants on which their larvae were unable to complete development. A few preliminary observations are available on the foodplant tolerances of ithomiine larvae (Chapter 6), but controlled experiments on larval feeding pre- 4. ferences and adult oviposition preferences of the Ithomiinae are clearly needed. CHAPTER V COMPARATIVE LIFE HISTORIES OF THE ITHOMIINAE The complete metamorphosis of a butterfly involves four life stages: egg, larva, pupa, and adult. The bias shown by lepidopterists in studying one of these stages to the near exclusion of the other three has led to the accumulation of a tremendous body of knowledge about the adult stage. By comparison, very little information is available concerning immature stages of butterflies. The preoccupation with adult butterflies has less to do with the importance of the imago as the stage of reproduction and genetic recombination than with the comparative ease with which adult butterflies may be collected, preserved, identified, and, of course, admired in display cabinets. The present study reflects this same bias, for immature stages of ithomiines are indeed difficult to work with, but I have attempted to reduce the imbalance somewhat by initiating a study of the life histories of the Ithomiinae at Limon- cocha. Although butterfly eggs and pupae are usually regarded merely as vessels of cellular reorganization and the larva as little more than an efficient feeding machine, all four stages carry a suite of adapta- tions that represent a series of selective compromises made in response to the matrix of environmental variables to which the organism has been exposed. The sum of these adaptive responses accumulated over 225 226 evolutionary time constitutes a life history strategy (Wilbur et al., 1974), similar in concept to the strategy of oviposition discussed earlier. The study of adaptations at the life history level has re- ceived a growing attention recently and has as its objectives: (1) to explain the diversity of observable life history patterns in terms of a minimum number of selective pressures, and (2) to identify those areas in need of more detailed examination and experimentation (Wilbur et al., 1974). Although knowledge of life histories' of ithomiines is minimal at present, it is already clear that the unit of study for ithomiine life history phenomena must be the local or subspecific population and not the species. The variability in oviposition strategies among conspecific subspecies of wide-ranging species has already been discussed. It will be seen below that foodplant specificity and the length of developmental time may also be quite different among conspecific subspecies. Perhaps the two most fruitful lines of research into ithomiine life history phenomena will be the comparative life history strategies of (1) sub- specific populations of wide-ranging species and (2) all the species present in a given area. The present research is an attempt to lay the foundation for a study of the latter type. Immature Stages The information available on immature stages of ithomiines comes from the studies of a very few researchers. Brief descriptions, and occasionally illustrations, of eggs, larvae, and pupae may be found in 227 several older works (Muller, 1886; Guppy, 1894; Haensch, 1909; Fountaine, 1913; Kaye, 1914; d'Almeida, 1938) and a few more recent ones (Costa Lima, 1950; Fox, 1967; Gilbert and Ehrlich, 1970; Rathcke and Poole, 1975; Brown, 1976b) . Reasonably complete life histories for 19 species have been published: 1 by Moreira (1881); 12 by d'Almeida (1922, 1938, 1944), one of which duplicates that of Moreira; 1 by Brown and d'Almeida (1970); and 5 by Young (1972, 1973, 1974a,b,c). The papers by the latter two authors illustrate all life stages photographically. In addition, A. M. Young and W. A. Haber (pers. comm.) have life histories of several Costa Rican Ithomiinae in manuscript. During the Ecuadorian research reported here, life histories of 18 species of Limoncocha ithomiines were obtained, as well as information on some of the life stages of 8 more species. These observations, however, cover less than half of the 53 Ithomiinae known from Limoncocha. Furthermore, the total number of detailed studies made to date represent less than 10% of the species in the subfamily. Clearly, much more work must be done before the first objective of life history studies, that of explaining life history 4. pattern diversity in terms of a minimum number of selective pressures, can be realized. The data summarized here, however, should do much toward meeting the second objective, that of identifying those areas in need of more detailed examination and experimentation. Descriptions of immature stages of Tellervo zoilus (Tellervinae) are given in Barrett and Burns (1951), Common (1964), McCubbin (1971), and Common and Waterhouse (1972), the latter work containing a photograph of a larva. Figure 47. Illustrations of Some Eggs, Larvae, and Pupae of Ithomiines. Eggs (17.5 x life size) a. Mechanitis messenoides C. & R. Felder b. Thyridia confusa psamathe Godman & Salvin c. Godyris zavaleta amaretta (Haensch) d. Ceratinia poecila poecila (Bates) e. Xanthocleis psidii ino (C. & R. Felder) Larvae (2 x life size) f . Mechanitis isthmia eurydice- Haensch g. Forbestra truncata juntana (Haensch) h. Ceratinia poecila poecila (Bates) i. Dircenna relata Butler & Druce (Drawn from Young, 1973) j . Melinaea menophilus menophilus (Hewitson) Pupae (2 x life size) k. Xanthocleis psidii ino (C. & R. Felder) 1. Ceratinia poecila poecila (Bates) m. Ithomia amarilla Haensch n. Melinaea menophilus menophilus (Hewitson) o. Thyridia confusa psamathe Godman & Salvin 229 r*lt 4it 0&4 f m n 230 The following discussion makes no attempt to present detailed accounts of the morphology of the early stages of the Ithomiinae but simply tries to give some indication of the range of morphological variability within the subfamily, and to discuss those aspects of the information available on immature stages most relevant to this analysis of ithomiine community ecology. Eg8s Egg size in terms of volume was presented earlier for many species in Table 15. The majority of ithomiine eggs are basically ovoid in shape, but the ratio of length to width varies considerably from a little over 2:1 for the spindle-shaped eggs of Mechanitis to about 1:1 for the nearly spherical eggs of Xanthocleis and several other genera (Figure 47). At Limoncocha, there are at least two exceptions to the ovoid shape, Godyris zavaleta (Figure 47c) and Heterosais edessa. Both species have eggs in the shape of a truncated cone. Eggs of most ithomiines are either white, pale yellow or cream in color. Because of their small size and ventral position on the leaf, yellow and cream-colored eggs are difficult to see. On the other hand, the large white eggs of Mechanitis , Forbestra, Oleria, and Thyridia are much more conspicuous, especially those of the first two genera, which are laid in clusters on the dorsal leaf surface. The eggs of all species examined have a series of vertical ridges, variable in number within predictable limits. In most species, these vertical ridges are traversed by faint horizontal ribbing, forming a grid of shallow depressions on the surface of the 231 egg. Eggs are cemented to the leaf surface during ovlposition in the usual way, but on foodplant leaves with a thick covering of stellate hairs, the eggs may be less securely anchored in the tomentum. The incubation period is usually five or six days, and rarely exceeds a week. Just prior to hatching, the eggs turn slightly darker, especially at the upper end where the mandibles undergo sclerotization. Most eggs kept in the laboratory, especially clusters of Mechanitis eggs, tended to hatch in the late afternoon or early evening. Such timing may be a protective adaptation that helps prevent larval desiccation, since most Mechanitis eggs are positioned dorsally on the leaf and the foodplants often grow in open sun. Larvae Upon hatching, larvae of most species consume from 50 to 100% of the egg casing before feeding on the leaf tissue. Most species also consume the larval skin following ecdysis. All ithomiine species reared to date have five larval instars, but the time spent in the larval 4. period varies considerably, from as short as 11 days to nearly a month (see Table 16). The first and last instars are generally the longest, the fifth including an active prepupal stage that lasts about a day. The beginning of the prepupal stage is marked by the termination of feeding which leads to a change in color at least partially due to the expulsion of the gut contents. The prepupa may wander a considerable distance from the foodplant before pupation begins. The behaviors of solitary and gregarious larvae differ substantially and will be treated separately below. 232 Solitary larvae, with one exception (Young, 1972), feed and rest on the underside of the foodplant leaves. Feeding may be initiated anywhere on the ventral surface, but the larva rarely begins at the edge of the leaf. The most common feeding pattern in solitary larvae is that of scraping away the leaf tissue from the underside, thereby leaving only the upper cuticle and major veins of the leaf intact. Most larval feeding appears to occur in the daylight hours. When not feed- ing, most larvae rest in a J-position, with the head curled around to face posteriorly. The wide-spread occurrence of this resting position among solitary ithomiine larvae of many species suggests that it may be a selectively advantageous behavior, possibly with a defensive func- tion. Such a resting position allows the head quick access to the mid- dorsal region of the body, where the larva's mandibles or regurgitated gut contents may provide a means of defense against the oviposition attempts of parasitic wasps and flies. Most solitary larvae appear to be cryptic in coloration and pattern, physical attributes which correlate well with their behavioral pref- erences for positions on the ventral surfaces of the foodplant leaves. The single known exception to the ventral position' of ithomiine larvae on leaves, Hymenitis nero in Costa Rica, achieves crypsis by curling the edges of its leaf (holding them in place with spun silk) to form a shallow cradle in which it rests (Young, 1972). In at least two species of ithomiine, Episcada clausina in Brazil (d'Almeida, 1922) and Godyris zavaleta caesiopicta in Costa Rica (Young, 1974c), late instar larvae live hidden between foodplant leaves bound together with silk thread. 233 Interestingly, the larvae of both these ithomiines have fine, nearly transparent skin, a trait shared with the larvae of some skippers (Hes- periidae) that also construct similar leaf shelters. Young (1974a, c) has described the behavior of solitary larvae of two species of ithomiines, which, when artificially placed with other, conspecific larvae on the same leaf, reacted with apparent agonistic behavior. This suggests that the solitary habit of these larvae, orig- inating in the single egg deposition by the female parent, may be main- tained by selection for agonistic behavior.' Gregarious larvae are synchronous in their feeding and resting behavior. Gregarious larvae feed at the edge of a leaf (on the under- side), usually lined up in rows facing outward. During resting periods, the larvae retreat en masse inward on the leaf to a silken mat constructed soon after hatching. The coloration of Mechanitis larvae, the only genus having gregarious larvae at Limoncocha, appears to be neither aposematic nor strongly cryptic. Their position on the undersides of leaves and their generally light color (dingy white tinged with yellow) combine to make them fairly inconspicuous, however. The gregarious larvae of M. isthmia, M. lysimnia, M. mazaeus, and M. messenoides all engage in a behavior that results in the expulsion of frass from the leaf occupied by the brood. Upon encountering a fecal pellet, a larva grasps it with its mandibles and raises its head a centi- meter or so above the leaf before releasing the pellet. Since the larvae live on the undersides of leaves, the raising of the head lowers the pellet relative to the leaf and the pellet falls free. Gregarious larvae in humid conditions are notoriously susceptible to fungal and viral 234 infections (e.g., Drummond et al, 1970), and thus by removing frass from the leaf surface before it accumulates, the larvae reduce the probability that infectious micro-organisms will become established « in the colony. Ithomiine larvae described to date appear to be of three basic morphological types (Figure 47). Larvae of the first type, characterized by eight pairs of stubby lateral protuberances, have been called "cog- wheel" caterpillars (Fox, 1967). The cogwheel type is apparently limited to the Mechanitini and may be characteristic of all members of the tribe, although the larvae of Sais and Scada have never been described. Larvae of the genus Mechanitis (Figure 47f) have relatively long, tapering lateral protuberances, while the larvae of the closely related genera Forbestra (Figure 47g) and Xanthocleis have rather stubby lateral pro- tuberances. The second type of larval morphology lacks any protuberances and is usually smooth, but often has a "corrugated" effect (Brown and d'Almeida, 1970), in which the apparent number of segments of the body is substan- ft tially greater than the actual number, an appearance caused by several indented annulations per segment. All of the larvae reared at Limon- cocha in tribes other than the Mechanitini and Melinaeini were basically of this type. For example, Figure 47h shows the basically smooth larvae of Ceratinia poecila, while Figure 47i illustrates a larva of Dircenna relata (Young, 1973) in which the setae are well-developed. The third morphological type of ithomiine larvae, one that is remarkably similar to larvae of the danaid genus Lycorea (see illustration in Guppy, 1904), bears a pair of very long flexible protuberances 235 (the "Scherndornen" of Muller, 1886) extending from the dorsal surface of the second thoracic segment. Although the only species reared at Limoncocha that has this larval type is Melinaea menophilus (Figure 47j), larvae of Tithorea harmonia meg'ara on Trinidad (Guppy, 1894; 1904), Tithorea tarricina duenna Bates in Costa Rica (Fountaine, 1913), and Aeria elara (and other Aeria species) in Brazil (Brown, 1976b) all have been reported to have similar larvae. The close morphological resem- blance between larvae of Tithorea and Lycorea has been cited by Gilbert and Ehrlich (1970) as evidence for the close relationship between the Ithomiinae and the Danainae. The discovery made during this work that a Melinaea species has the same larval type provides further evidence for such a taxonomic association. It is also of interest that the larva of Tellervo zoilus gelo (Tellervinae) (figured on Plate 18 in Common and Waterhouse, 1972) is of this larval type, although the flexible protuberances appear to be on the third thoracic segment instead of the second. The foodplants utilized by these species are also important in sorting out the taxonomic affinities of the Ithomiinae and will be discussed in Chapter 6\ Pupae The prepupae of most ithomiine species apparently leave the food- plant in their wandering search for a pupation site, for pupae are rarely found on the larval foodplant (Young, 1972; Brown and d'Almeida, 1970; present study). For example, Beutelspacher (1972) reported finding a pupa of Melinaea imitata Bates in a leaf of an epiphytic bromeliad in Veracruz, Mexico. This observation is of interest because larvae of 236 M. imitata, like those of M. menophilus at Limoncocha, probably feed on a species of Juanulloa, an epiphytic genus in the Solanaceae. The pre- pupa of M. imitata apparently wandered from its epiphytic foodplant and ended up in a bromeliad on the same or an adjacent tree. Further evi- dence for prepupal wandering is provided by Mechanitis larvae reared on foodplants growing in the large (3.7x3.7x1.8 m) screened insectory at Limoncocha; these invariably pupated on the roof of the enclosure or on the uppermost leaves of the tallest plants available. This circumstance suggests that these larvae seek relatively distant and high pupation sites in the wild. A few species, however, appear to occasionally utilize the foodplant as a pupation site. Young (1974b) has reported that pupae of Pteronymia notilla are often found suspended from the ventral midribs of foodplant leaves, and at least one species, Episcada clausina in Brazil, regularly pupates on the foodplant. As men- tioned earlier, the larvae of the latter species live between two leaves bound together with silk. It is within this larva-constructed shelter that pupation normally occurs (d'Almeida, 1922). I have observed a similar situation at Limoncocha where a field-collected fifth instar larva of Dircenna loreta bound the two curled edges of a food-leaf together with silk to form a shelter in which pupation occurred. Although I never found any pupae on their foodplants in the field at Limoncocha (see Chapter 6) , many pupae of Thyridia confusa were found on their larval foodplant in Puyo, Ecuador in August, 1974. In the gardens of the Hotel Turingia, a hedge row of the ornamental shrub, Brunf elsia calycina var. f loribunda, supported a large population of eggs, larvae, and pupae of this species. 237 Pupal morphology varies quite a bit in the Ithomiinae and some examples are given in Figure 47. The pupae of Xanthocleis (Figure 47k), Mechanitis, and Forbestra (Mechanitini) are all of a similar elongate shape and a brilliant metallic appearance of burnished silver when mature. Many of the compact pupae of species in the Dircennini (e.g., Figure 471) have gold and silver patches but are not uniformly metallic like those of the Mechanitini. Metallic markings in pupae of the other tribes appear to be rare. Both the yellow pupa of M. menophilus (Figure 47n) and the milky white pupa of Thyridia confusa (47o) are streaked with black markings when mature, but differ radically in shape. The Ithomiinae share with the Danaidae the trait of pupal rigidity (Muller, 1886), further advancing the argument in favor of a close relationship between the two groups (Gilbert and Ehrlich, 1970). Generation Time The available data on developmental times of ithomiine life stages are presented in Table 16. At least partial data are available for 42 species, and the total generation time has been computed for 32 species. Unfortunately, many of these generation times are only estimates, be- cause the eggs were collected in the field and the times given for egg development reflect only the time since harvesting and bear no predictable relationship to the time elapsed since oviposition. This is especially true for many of d' Almeida's (1922) life histories and also for several of mine. Such times are placed in brackets in the table. Another dif- ficulty in comparing the generation times given in Table 16 is the 238 S CM .-) ,-| _| u M o td o o u o u u W H H td Id M c_> o o o r4 u U U o f ■> td M Id td m td td Id Id U ■o 01 § . hi 01 01 c o X «M O 01 B 1 a. o r-l I 01 a S H g o ►J u _ e tn c T> — I 01 a. c B) -H 01 c o. ■ SBC < e» en ■ s «-» c 3 r- |4 *^ S^ 0 X eg r* eg ft o. u ■ CM *gr m in 1*1 CM CM CM CM co r» r*> r» r«> r«> r»» r- r«- r» r» i «o VI I i-l t-M cm cm en en co CM H H —I ft n n hi s n is in ci n n X u to Q u to 3 z i-e c o 2 X o o c 1 N c H eg eg b b a: >0 hi « eg eg eg 01 ft •H ■H 4-1 4J u 4J eg •*-! ■H sM X c C c eg eg eg a X X X •o CI CI CI a) 01 01 01 CI Z X £ CO 239 »-< H F- Z i O- O ►J •o s B C o o XI c» i —< «» r>. r^ c*> tn N-* S-* 01 14 r*. r» u o N-^ v^ ■ • • o u •^ a. D. m CO x> a XI m « 01 .c 01 m § s § 3 § u C 4J c a c c C C . H >. — 1 .c1 o JC o a. CO Q. m in CM 1 © r-i CM 1 CM ^s *^ O «7> O «W t-l «-■ «4 c. n .-. c s ^ r-l eL a 5 CO eo to v 01 X 3 § 4-> c c c c ■0 CO CO CO -•4 •-* •-I 1-1 u o o o >■ m in tfl -11 M S ©■» cf1 co r* oo I CO CM P". I CM r^ CM ■a p» r» a> CM «n •* ■» CM CM CM •» CM t-e CM • w-t CM •» •» tn z u O O Ul p» < z o a> o a • eo c ■H 3 N O «> O C r» tn .» « 4 4} 4 . o o. K o 01 a. C9 z -h a > CO <0| cq CO C to 0) d u o 01 a eg z C O 3 0) ■X. I XI X u a co C >H Ol CX a) co X ^ 4 CO O E a CO CO •9 •^ -r-i -r-( t4 H b r. a o o o O X X X. JC i_i u u 4J >-i HI Hi H< £ •0 o 240 a. 2 •J i i I CO o 01 Ml o CO c_ o: OS DC a u « o o OQ CO PQ u M (0 w c 1 Vj ^o I *-> a. It eg «l u a c «o »-• o to CM c-l a c co *-* o CO a ai u a c m o to o CM a o os o a: m M O M M CM 1 i 3 § Li c c c AJ a CO B CO i— < •-I .-I 01 • o o o <-> 'CO co CO o CM © CM ^ •J B 0 O. U •— • a) B J3 *-• « tl tl o a ^ S 1 ci c H c U c • i-» eg aj CO •-* CO «-■ CO r>t o 01 o «l O CO CJ CO U CO vo CM vo ■ o» CM CM CM I r-< r-i m> CM .» CM I O CO O CM co iq m < •H f* CM I in cm m ^^ < «l CO » co ^^ tl o> U CO u 2 N AJ 2 a *—* CO c a *+ 3 Q ^-^ CO •o c o AJ AJ vB 5 a) § c e> q-1 u Vi jr 5 t-i CO AJ l-i 01 01 CI *r4 o •*4 (X E i-^ CO Vi CI hi u » N«* CO AJ c 01 » CU CO cu 01 I-. 3 01 c II o 1— 1 X CO CJ co CO JO — 1 a. CO aj N^ 01 X • 3 u i-l 3 CO CO CO X Ol co Li CO CO c 01 — i CO o. f-H CO c 01 AJ i-i CQ o a -r-4 c CO •w f, t4 t4 M > O CO aj Ol Li AJ CO ^-1 -O 01 o CO 3 X 3 O a. J. c o i-t CO 01 c CO a a) a co 01 l-l CO •o l-l CJ -*4 CO CO CO CO A CO CO N •H — < — i •H 3 t-J -v4 c CJ CO CO c CO •o AJ •--I 6 I I £, 01 z ^4 -r4 c CO 3 c c C C eo 1-4 i-j AJ 01 CI AJ o o o o -^ z CO CO tl CO AJ u V- u lU z U u u "■J -H 01 01 01 01 f-« w 01 01 ~-l a M AJ AJ AJ JJ i £ u o o u P- 0- a. 0. cu X JC CJ «l a ai —i z i 8 a CO a) X ce 03 CO ^■^ AJ CJ *— - -w^ CO -*4 01 V p AJ a —1 f-H a) AJ o 01 01 C 01 -4 -C J= JO l-l CO p. a ■ 3 CO cu 01 01 X E co CO u c c 01 « <0 o. CO CD CO CO h AJ AJ CO CO o CU V CU 01 D. l-l f-« ■o -o o CO CO 01 01 eg > > CO CO CO CO en N M -H ■H ai CO CO j= CO CO CO CO CJ — < -" 1 o o >> 14 V4 l-l Li AJ >• > cu 01 T-l ■o •o AJ AJ ^H o o 01 01 M u o X X 241 b u-5k ZOH< Z u •J i3C H z < 3 C u C O o •s •* «-< CM »H a-4 C6 O (J •O c a I w CI p • «-4 « eg 3 « Ll c c {4 «M xj m • xj S a —i .-I co 01 o o 01 M O in CO 0 « e CM 5 01 C > —i m xa -o c c o c o u -^ a * I » 01 X 01 o a >• X .eg -o o Vet •» ^ 1^ l»- r» o\ ©> o* HUH § § 5 000 V 1 O "9 01 »-i t> ■O M .0 U3 ■ O O. 03 0) X) o o c 1 2 c o — I t-l XJ a V - 1 U CO O 0 01 * li c —I X 6 p eo 00 — 5 c c < O 3 3 L. O O T3 (O >• >• C • to- ol XJ X) XJ to • o II ft mi • u c xj 01 01 c 9 M 01 » B ■ 01 a. we o 600) t-l 01 01 w^ > O 01 • 9 1 » *J c o o c u xj a. -h »-» t-l ■ e o u o -J -* O c « • eg CO *-» •a >h g o CO (I 01 0) CL 01 O »-■ H w tg c 0 wi a 01 en 01 o U m m ■ XJ « 0) « c c > > V 01 01 01 B B • > B « • t-i 0 > m —> 01 01 t-i B — Sk X B u -< 0) * T9 01 X X CM OO 3 cm 0 t>s a at t> ■a C t-4 i-l 3 3 AJ - co - co m ■O -o -o ■xj c -* -rf c o o) 01 o) g B B 0) ■OX 0) XJ E - g 0) >s »3H CO XH t4 • O 0) CB/-s CSX EO TJtc 3—1 — 1 v B a >JZ — 1 eg — 1 u C O T> O — I U C e o -h c ■ ci -> cm ri ^ 1—1 o n C U eg a u eg X 3 U 0) v4 OI> ja eg il E C ^ C W c u u u X 01 01 u 3 a. 3 o- «o o -^ 01 ft u c - ."5 ~i eg ►) -< eg 2 S ,° ca ■ - 1- c o u h i-i ci 01 3 CO X 0) u (it 3 O 0 O u m teg c to eg >-. CO t-l lu o o II 3 01 M a u vi 01 eg a. W u eg Ot CO u eg o eg ci CJ -^ CO B U 3 in cu ■n u v^ co • B a eg eg a. > t-i ■ 01 _) co S 2 c > o - B. > o CI u •u Bt eg eg eg o -rf eg -o C 01 eg l» 3 01 U X o o xj in ^ o> n > eg - 14 C >-. eg eg o «-i co u u c CO eg o t-l *-l B JD BE kl 01 00 eg x • U u vac CO o 11 V 01 E E r. |« 0 >_• 3 •O m r* c eg ci O J 00 X eg •< eg u i-l c -«g -H AJ O h ■ n 01 eg • - 3 aa > ° - 2 : > o> > o o u eg h P* eg c- c CI 01 eg eg Ic — 1 O a -o —, XJ ci CKC 3 01 eg cl, x co > ci • B e > ci o eo e O B 5 I O XJ O. CO ■ Z M 3' eg e -< o -a «-» c (4 « r4 X eg u c O t< O H B 8 J H • At > « O X Vi u n> e o x o. u eg ^ Z P. a u o »h r-icMd -^cMco^m C C •C O TJ 01 ■rl-^ B * co— 1 0 B 01 eo X K O xj c 3 CL O 00 I 01 Li CI 1 c 5 o .tJ Li CO cc u O U ctf u L. o •o o 242 apparent lack of uniformity in rearing conditions. While the species at Limoncocha were reared under relatively uniform conditions, as were those in Costa Rica (Young, 1972, 1973, 1974a, b,c), little is known about the methods used by d'Almeida (1922, 1938, 1944) and other authors. The variation in larval development within some of the species reared by d'Almeida (1922) suggests to me (based on broods at Limoncocha that gave similar results) that the availability of fresh foodplant was not uniform throughout the rearing procedure. Such problems will, of course, tend to overestimate generation times by prolonging larval development. The generation times given for Limoncocha species in Table 16 are averages of the number of individuals reared on a single plant species. The variation in developmental times among individuals or broods of the same species reared on a particular foodplant was usually quite small. Only when the variation exceeded two days for any one stage within a single brood (as in Mechanitis isthmia on Solanum quitoe'nse) was this variability entered in the table. Broods that experienced a lack of food during the rearing period were not used in calculating the averages. A few Limoncocha species regularly utilize more than one foodplant and, for these, the developmental time required on each foodplant is presented in Table 16 along with the developmental time averaged over all the foodplants. The degree of accuracy of the generation times given in Table 16 depends on how many of the three immature stages for which accurate data was gathered. For 14 species, accurate measurements of all three immature stages (egg, larva, pupa) are available, resulting in very re- liable generation times for these species. For an additional 18 species 243 the data on the larval and pupal stages are accurate, but egg develop- mental times are not, since the eggs were collected in the field. Of these 18, however, the egg developmental times given for 14 species are within the range of values of the other species in the same tribe, and thus may be considered reasonable, albeit minimal, estimates of egg developmental times. The total generation times given for these 14 species, then, are probably quite close to the true values. Of the remaining 4 species, for which the egg development times listed are only one or two days (being the time between field collection and hatching) , reasonable estimates of total generation time may be made by selecting the minimum accurate egg development time from the same tribe. In two species the pupal stage was not accurately measured, although the 7 day minimum for Melinaea menophilus is probably a reasonable estimate. For Thyridia confusa, there is no information on pupal developmental time, but a minimum estimate of at least 8 days may be assumed, this being the shortest pupal developmental time recorded in the Godyridini. Ap- plication of these estimation procedures augments the complete data to i yield a total of 34 accurate measurements or reasonable estimates of gen- eration times covering 32 species of ithomiines (2 subspecies of M. lysimnia and of G_. zavaleta are included) . These data have been plotted graphically in Figure 48, in which the species have been grouped by tribal affinities. The range between the shortest and longest generation times is about 20 days, with Oleria agarista (Limoncocha) completing development in only 24 days, while Pteronymia euritea in Brazil and Thyridia confusa (Limoncocha) both require a little over 40 days (see Table 16). Thus, the range in HI •u to 1 4J CO CU id cu 4J • co co rH cu 3 -H CJ CJ rH CU to (X O CO 4-1 to 3 A CU CJ CO 0 cu 0 H 3 & 0 cu B U -H iJ co cu cu U 4J n) n) 3 O cr-H OT 13 3 • •H 0) • to co co 3 CU iH •H e 0 •H •h ,c § " & .3 3 co ■u 0 H •H 13 4J -H 14-J CO t-l O u 0 CU W to 3 0) cu •H bO . O /— N CD Ti 4-1 P. CU x W > cu u 4J vfr 01 cn to cu Xi cu M-l 0 CO 0 Nw' 4-> CO C co o> cu cu 1 co 6 CU -H H U 4J a 3 0) 3 0 u 0 •H •H 4-1 CO 4-1 to CU CO u iH M d> O CU 3 U 3 CU •H CU 0 U 60 00 *» cu u 3 00 245 o D o -a- DO O o» D O • n o o o o o in eg CO O M H O lO A A A A A A A A M 53 H W Pi o W H H H 2: M M 55 H O o w SI H 25 HI M o « H H M H 25 13 M M M 13 § P M H W Pd p4 o £ W prf O ►J M o O O o 246 length of total developmental time among species in the lowland forests of Limoncocha is as great as that recorded to date for all Ithomiinae so studied. Considering all species listed in Table 16, the range in * development times for the egg stage is 4 to 10 days, the range in larval development times is 11 to 28 days, and the range in pupal development times is about 6 to 12 days (although d' Almeida, 1922, recorded 15-16 days for Placidula euryanassa) . Although these data represent only a fraction of the species in the Ithomiinae, they provide enough information to make a tentative comparison of development times among tribes and among certain genera. In Figure 49 the data for the species in each tribe have been averaged to provide a mean developmental time for the egg, larval, and pupal stages of each tribe. The tribal averages for each of these stages have been added together to obtain a mean tribal generation time. Unfor- tunately, no data is available for the Tithoreini, and the Melinaeini is represented only by M. menophilus , but several species contributed to the means of the other tribes (see Figure 48) . Figure 49 reveals that the average total generation time for each of the seven tribes studied is just over 4 weeks, or approximately 30 days, with a range of 27 to 32 days. Since many of the Tithoreini species are relatively large butterflies and inhabit the cool mid to upper elevational slopes of the eastern Andes, it can be predicted that they will have compara- tively long generation times. The mean generation time for the Ti- thoreini will probably exceed that of the Godyridini, which has the longest generation time of the other seven tribes. u CD a § M Q CO Q) 00 248 o en o n m «o o s A A A A A A M M 1 55 M H M M 53 W o o W o Pi w -J o o PS A M 55 M Q M a o o 249 The comparison of the tribal averages of the three immature stages (Figure 49) shows that the shortest egg development times (4 and 4.3 days) are found in the Melinaeini and Ithomiini, and the longest (6.9 days) in the Dircennini. For the larval stage, the Mechanitini has the shortest developmental average (14.3 days) while the Ithomiini average (17.7 days) is just slightly longer than those of most of the remaining tribes. Average pupal development times are about 7-7.5 days in the Melinaeini, Mechanitini and Napeogenini, but range from 8 to 9.7 days among the Oleriini, Dircennini and Godyridini (9.7 days).- Thus, the range in mean generation times among all tribes is matched by the range in variation among each of the three life stages as well. No single stage accounts for most of the observed variation in generation time. To see if the variability in egg size and mature larval size (Table -16) is correlated with the development times of these stages, the graphs in Figures 50 and 51 were constructed. There seems to be no correlation between egg volume and egg development time (Figure 50) , and only a slight correlation between larval length at maturity and the length of larval development (Figure 51) . When the sizes of these two stages are plotted against total generation time (Figures 52 and 53) , a similar pattern results. Figure 52 shows that there is apparently no correla- ■--. - tion between egg volume and total developmental time. More likely, egg size is a response to the requirement for a newly hatched larva of a certain minimum size, perhaps related to the need for larval . == mandibles of a minimum size to successfully initiate feeding on that species' foodplant. Figure 53 shows a definite, but not very strong, 1 \ -. u : : Figure 50. Egg Volume versus Egg Development Time. Figure 51. Mature Larval Length versus Larval Development Time, 1.5T tn 1.0-- o o 251 o > o o w .5-- 8 a o -+- + 4 5 6 TIME IN DAYS 10 60 50-- A0-- o J30- s o n 10-- o o oo oo o o o °o o 8 oS°8 4- -+- + 10 15 20 25 TIME IN DAYS 30 35 40 Figure 52. Egg Volume versus Generation Time. Figure 53. Mature Larval Length versus Generation Time. 253 3.08 1.5t «n 55 M W -3 8 1.0- .5- O o o o ■+■ ■+- 10 15 20 25 TIME IN DAYS 30 35 40 45 o o 2 60T 50- 40-- 30-- 20- 10- ° o • o O O ooo o S 9 • o o • 8* • o° o e + -+- + 10 15 20 25 30 TIME IN DAYS 35 40 45 254 correlation between the size of the mature larva and the total genera- tion time. The fact that this correlation is more pronounced than that of Figure 51 (mature larval size and larval development time) is prob- ably due to a relatively long pupal development time in species with large mature larvae (Table 16). Three ithomiine species at Limoncocha were reared on several different larval f oodplants and the developmental ' times for these ^species (Mechanitis isthmia, M. lysimnia, and Oleria agarista) on each f oodplant are given in Table 16. All of these f oodplants were utilized in the wild by the ithomiine species which were reared on them, but some were more frequently used than others. The variability in developmental times resulting from feeding on different foodplants was manifested only in the larval stage. The developmental times of the eggs and pupae (with one possible exception in (). agarista) were species-specific regardless of the foodplant. In each of the two Mechanitis species, the most commonly used foodplant resulted in the shortest generation time. For M. isthmia, this was Solanum coconilla, which, however, is dimorphic for the presence or absence of acicular spines on the- stems, petioles, midrib, and most of the leaf veins. It was no surprise that larval development took slightly longer on the spined form (15 days) than on the unspined morph (14 days), but it is perhaps premature. to ascribe the differences in developmental time solely to an impediment to feeding caused by the presence of the spines since nothing is known -of the chemical defenses of the two morphs. Nevertheless, a similar correlation between mechanical defenses of foodplant leaves and longer larval development time occurred in M. lysimnia and 0. agarista and on two other foodplants used by M. isthmia, S^. quitoense and L. esculentum. 255 Solanum quitoense is not native to the Limoncocha area, but a few introduced plants grew on the missionary base. Although lacking spines and trichomes, the leaves and stems of S_. quitoense are densely pubes- cent with stellate hairs. Two egg clusters of M. isthmia were found on this species and reared to maturity. The larvae raised on JS. quitoense required 18 to 23 days to mature, however, instead. of the usual 14 re- quired on S. coconilla. The added development time probably reflects a detrimental effect on larval feeding ability, especially of the earlier instars, caused by the stellate pubescence, much of which was left un- eaten by later instars. Larval development also took longer on the cultivated tomato (L. esculentum) , on which M. isthmia females occasionally laid clusters of eggs (16 days required as compared to the normal 14 days). L. esculentum lacks spines and trichomes but is well endowed with glandular hairs. The principal foodplant used by M. lysimnia is Solanum sp. (7304), and the larval development time on this plant (12 days) was slightly shorter than on Solanum sp. (7334) (13 days). Both of these plants, however, gave considerably shorter larval development times than Solanum pectinatum (22 days). Solanum spp. 7304 and 7334 are soft-leaved, unarmed, glabrous plants. By contrast, S^. pectinatum has a dense cover- ing of stellate and glandular hairs on all vegetative parts of the plant and extremely sharp acicular spines on the stems, petioles, and leaf veins, making it one of the most well-armed solanaceous plants at Limon- cocha. These mechanical defenses were almost surely responsible in large part for the long larval development time on S_. pectinatum. 256 Oleria agarista utilizes four foodplants at Limoncocha. Although it is not possible to say which of these is most frequently used, Solanum sp. (7319) is the most abundant. Three of the plants are en- tirely unarmed and the leaves appear to be glabrous. These three gave similar larval development times of 13 to 14 days. The fourth plant (Solanum sp. , 7310), however, is heavily armed with recurved spines on the midrib, trichomes on the leaf veins, and is heavily tomentous with stellate hairs. Not unpredictably, this plant resulted in a larval development period of 17 days, .3 to 4 days longer than those of the other three foodplants. It appears, therefore, that even though a solanaceous plant may . be acceptable as a larval food source for ithomiines, the mechanical defenses employed by the plant can result in considerable lengthening of the larval development time, thereby prolonging the exposure of the immature stages to parasitism and predation. Chemical defenses, in the form of toxic alkaloids, may also significantly affect the developmental time of ithomiine larvae. Some of the "foodplants" on which ithomiine ft females occasionally oviposit may even prove lethal to the resulting larvae, depending on the efficacy of the plant's mechanical and chemical defenses, and the tolerances of the larvae (see Chapter 6). Parasitism and Predation The data presented in Chapter 4 indicated that the ithomiine community at Limoncocha (or at least that portion of it sampled at Site 4) achieves a relatively stable existence for most of the year as measured by the 257 relative abundances, diversity, and comparative sample sizes throughout the year. Catastrophic weather conditions appear to depress the overall abundance of ithomiines at various times during the year but this mor- tality appears to be rather equally distributed among age classes and species of the community. The recovery shown by the community to these acute depressions in abundance appears to return the community to ap- proximately its former level of diversity and abundance. Thus, the ithomiine community at Limoncocha gives the impression of compositional and numerical stability occasionally, and perhaps regularly, depressed by adverse weather conditions. There are no seasonal or other periodic population explosions of ithomiine species such as those recorded by Brown and Benson (1974) and Brown (1976b) in the seasonal forests of Brazil. In addition, the relatively even distribution of a female's reproductive effort over time, and the resulting overlapping of two or mote generations, certainly must contribute to the stability of species populations by dampening the effect of cyclic predation and parasitism so well known in the population biology of temperate insects. i The abundance of ithomiines at Limoncocha, although periodically and strongly affected by environmental conditions, does not appear to be regulated by them. Censuses of forest and second growth areas at Limoncocha reveal very low densities of ithomiine juvenile stages and - a great under-utilization of available f oodplant tissue (Chapter 6) ; Thus, it is highly unlikely that low availability of larval f oodplant is limiting the population of any Limoncocha ithomiine. It is possible that adult nectar sources, and possibly detritus food sources as well, may be somewhat limiting (Chapter 3) . As will be discussed below, there 258 appears to be a high juvenile mortality of ithomiines at Limoncocha, apparently the result of intense parasitism and predation by other insects. This high juvenile mortality, coupled with the low daily reproductive rate of most ithomiines, appears to be the controlling factor in setting the levels of abundance of Limoncocha Ithomiinae. Parasitism Based on field collections of immature stages at Limoncocha, it appears that parasitism of eggs makes a greater contribution to ithomiine juvenile mortality than does larval parasitism. This may be partly a result of the small sample size of field collected larvae, but one impor- tant factor may be a great reduction in the number of larvae available for parasitism caused by the prior action of egg parasites and larval predators. Table 17 summarizes the parasitism data for immature stages of Limoncocha ithomiines collected in the field. These data underesti- mate the rates of parasitism, however, because, by collecting these juvenile forms before they had completed their life stage, a sample was chosen that had been exposed to parasites for less than the normal time. Even so, the rates of egg parasitism appear to be high for many species, ranging up to 60% in Hypothyris fulminans and Ceratinia poecila, species for which relatively large samples were taken. Curiously, egg parasitism is not equally distributed among species, many of which (e.g., Napeogenes pharo) appear to be virtually free of parasitism. Such observations imply either that egg parasites (mostly species of braconid and trichogrammatid wasps) are very host-specific, or that the 259 PQ P-i CO CO iJ •3 3 cfl CO 60 60 W CU c •H •H 0 o u 0) U 0) o u •H Pn - Cr 0 CD •H ■P •H co cfl U cfl P-. OJ rH ■8 H to M H H SB H W CO PS (2 W < Ph Ph co CO pj Pi oo w W O O H Q W W CO fn pi O O o IT) CM J3 CM m ro vo CM o o o m CM CO vD o m CM o CJ O o CJ o o CJ vO •0- o rH , >i >> cu cu JS J= X 6C 60 4-J 4J 4-1 O O o O O cu 0) p P P p P >> K >> cfl CO EC PC « z S3 260 c/} Cfl p; pj u o w W CD O H eg w w c/3 o o c •H •u c o a t-H m vO CO m CO CM r-i CO CO O CM o CT> CM CO CO CO I/O CO CO CM m m tfl cfl rH •H rH (0 co •H CO o Vj tfl c to u bt a 0) m (0 T3 o as to « CO •H •H •H e S e o O o JZ J3 X. 4-1 4-1 4-1 H M M w o o 0) o H C w c_> Pi CO w >-< to w PS § to > ffl N CO •H U O O CO co CO CD T3 5 Pn Pi >J 2; o o u CO Pi o o w W p O H -a ■a) C •H 4-1 c o o CO CM CO oo VO CN sr co CO r-» r» . . rH i-\ O o CO co CO w M c c U o o W Ph CO CD CO a) CU •H •H § o CU a CU o p. ex CO CO T3 cu N •H 4-1 •H CO CO M cd M CU 4-1 CO 3 T3 CU N co cfl co p. cu N CO cd u cd a. cu )-i CU O o u CO > u cfl cd cfl c CO CU 00 cfl 00 CU I cfl iH iH i-l cfl iH iH M-4 cd o »4-l B^S o o o e^s r-i o 1 o CO iH CTi V— ' V-* rd CJ U U cu CU 4J ■u 4-1 4-1 cu CU iH ■H cu CU J3 .C 4-1 4J ^ J*. ,o ,o •a •a cu cu 4-1 4J cfl Cfl CJ a •H •H 13 T3 a 0 •H •H cu CU 1-1 l-i cfl cfl CO CO t3 u o CU o 4-> 1-1 CO x> 3 tH r-l CJ cd > 00 u 60 cfl W hJ 262 eggs of some ithomiine species are much more vulnerable than others. The rates of parasitism of egg clusters in the Mechanitini are lower than those for the single eggs for species in other tribes, but the actual number of parasitized eggs found in the Mechanitini exceeded the number found in other tribes, because nearly all of the eggs in a parasitized egg mass were parasitized. The parasite female is ob- viously very efficient in her oviposition behavior. Larval parasites were found in only two ithomiine species, Forbestra truncata and Mechanitis lysimnia. Two broods of F. truncata, one of four larvae and another of five larvae, were completely parasitized by a species of braconid wasp (Hymenoptera: Braconidae) . Each brood con- tained larvae of equal age (fourth instar; presumably from eggs laid by the same female) scattered over the five or six leaves of the food- plant. The fact that all larvae on the plant were parasitized suggests that the female parasite can either recognize the larval foodplant, which she subsequently searches thoroughly, or after a chance encounter with one larva she adjusts her searching behavior to include all the leaves in the vicinity of the one on which she encountered the first larva. Of four mature larvae of M. lysimnia collected in pairs on two foodplants, only one was parasitized. The parasite was a tachinid fly (Diptera: Tachinidae) and, unlike the braconids on F. truncata, emerged from the chrysalis rather than the mature larva. Predation Although circumstantial evidence indicates that predators take a heavy toll of ithomiine larvae, predation of immature stages was actually 263 observed on only three occasions. A second instar larva of Olerla agarista was once found (1630 h) impaled on the proboscis of a small (5 mm) yellow reduviid nymph (Hemiptera: Reduviidae) that was walking on the underside of a leaf of the larval foodplant (Solanum sp., 7319). On another occasion (1500 h) , several dozen small black Crematogaster ants (Hymenoptera: Formicidae) were observed crawling over a cluster of Mechanitis isthmia eggs on Solanum coconilla. A closer inspection revealed that several of the eggs had been badly damaged by the ants, many of which were carrying away bits pf egg material in their mandibles. A second example of egg predation occurred when a cluster of ten Mechanitis mazaeus eggs was discovered in the process of hatching in the late afternoon. At the time of discovery, six of the eggs had hatched but the larvae were not present on the plant and the empty egg shells had not been devoured. Since the larvae in the other four eggs of the cluster were just emerging, it is most likely that the first six larvae had been predated just before I arrived. No pupal predation was observed, but an emerging adult of Ceratinia poecila, while in the process of abandoning the pupal shell, fell prey to a spider that was inside the rearing cage in which the larva had been placed prior to pupation. Although an artificial situation, this incident illustrates the vulnerability of teneral adults during the period before their wings harden. (Freshly emerged adults can fly within 45 to 60 minutes after emergence, but, in the laboratory at least, they often remain suspended from the pupal shell for several hours before first taking flight) . Ants and other social insects are undoubtedly responsible for a great deal of larval predation. Small Solanopsis ants construct nests 264 among the fruits of Solarium coconilla and along the stems, leaves, fruits, and flowers of several other solanaceous plants, which they patrol incessantly at Limoncocha. The much larger Ectatoma and Paraponera 4 ants also patrol many solanaceous plants, and probably take many larvae. The swarm raids of Eciton burchelli and foraging columns of other army ants are frequent events of the Limoncocha understory. Social wasps are probably also important predators of ithomiine larvae and may account for the wholesale disappearance of broods of Mechanitis isthmia known to take place in just a few hours. For example, a Solanum coconilla plant cultivated behind our house at Limoncocha attracted the attention of numerous ovipositing M. isthmia females, and some of the egg clusters were kept under close observation. One cluster of 50 eggs hatched on the evening of the sixth day after oviposition. By noon of the seventh day, 48 of the original 50 were left. On the eighth day 45 larvae were present, on the ninth day 44, and by the afternoon of the tenth, all the larvae (second instar) were gone. Another egg cluster laid under similar circumstances suffered a similar fate. Of the 32 eggs originally 4. laid, only 24 were present on the fourth day and these all hatched on day seven. By noon of the eighth day, 20 larvae were left, on the ninth day only 4 larvae were present, all of which were gone by noon of the tenth day. The plant on which these larvae occurred was large and healthy and plenty of leaf material was available, so lack of food was not responsible for this disappearance. The disappearance of only one or two larvae at a time is probably due to solitary patrolling ants or assassin bugs, but the sudden disappearance of large numbers of larvae may have been the result of social wasps finding the brood and 265 recruiting other individuals. Swarm raids of army ants could presumably have the same effect but none occurred in the area during the observa- tions described above. The shelters constructed by late instar larvae of Hymenitis nero and Episcada clausina probably do not prevent predation by large ants, but may significantly reduce the probability of wasp predation, and perhaps of parasitism as well. The gregarious larvae of Mechanitis engage in synchronous head jerking behavior when disturbed, much as do the gregarious larvae of many other species (e.g., Drummond et al. , 1970). This behavior, combined with the expulsion of the larval gut contents may be effective in detering some insect predators and para- sites. The ability of larvae to suspend themselves below a leaf on a silk thread has been described for all instars of Hypothyris euclea (Gilbert, 1969) and for the first two instars of the larvae of Hymenitis nero (Young, 1972). This "drop-off" behavior presumably reduces preda- tion by ants, which elicite the behavior by walking on the dorsal leaf surface. The "drop-off" behavior does not occur when other insects walk on the leaf and it cannot be elicited mechanically (Gilbert, 1969; Young, 1972). "Drop-off" behavior or suspension by silk threads under other circumstances was never observed in any larvae of Limoncocha ithomiines. Eggs and young larvae may also occasionally fall prey to canni- balism by their own species. I once placed several large leaves of Solanum coconilla in a plastic bag containing a nine-day old brood of fourth instar M. isthmia larvae. One of the leaves had a cluster of 25 M. isthmia eggs on it. Within a few hours this leaf — and the eggs on 266 it — had been devoured by the M. isthmia larvae. As there was plenty of other leaf tissue left, the eggs were apparently eaten inadvertently as a consequence of being on the leaf chosen by the older larvae. Se- lection should favor oviposition behavior in female M. isthmia that minimizes the chances of such occurrences. This may be another reason for the preference shown by ovipositing M. isthmia females for well- isolated foodplants (see Chapter 6). CHAPTER VI THE ITHOMIINAE-SOLANACEAE INTERFACE In their review of the larval foodplant relationships of the butterflies, Ehrlich and Raven (1964) concluded that the plant-herbivore "interface" may be the major zone of interaction responsible for generat- ing terrestrial organic diversity. These authors proposed an evolu- tionary sequence progressing by stepwise adaptive responses between plants and herbivores. Recognizing that the adaptive responses of the species in one trophic level affected the fitness of the species in the other, they termed this process "coevolution. " Secondary plant compounds , such as the alkaloids occurring in the Solanaceae , were recognized as being of primary importance in this evolutionary sequence. Plant groups that developed, by mutation or recombination, effective chemical deterents to herbivores, evolved and diversified under the protection of this biochemical shield. Several major groups of plants have apparently evolved in this way, including the Solanaceae, which are not well represented among larval foodplants of the Lepidoptera (Ehrlich and Raven, 1964). The alkaloid barrier erected by the Solanaceae, however, was breached by early ithomiine stock, which subsequently en- tered a new "adaptive zone." Relatively free from the competition of other herbivores, and possessing the basic physiological ability to tolerate solanaceous alkaloids, this ancestral stock was able to diversify rapidly on the closely related species of the Solanaceae. 267 268 In addition to tolerating or detoxifying the potent alkaloids in solanaceous plants, it is generally assumed that many ithomiines actually sequester these poisonous compounds, which in turn make them unpalatable to vertebrate predators. Although the extensive participa- tion of the Ithomiinae in mimicry complexes provides circumstantial evidence for their distastefulness, little empirical evidence exists (see Chapter 7). Plant Defenses and Larval Adaptations The considerable mechanical and chemical defenses of the Solanaceae were briefly characterized in Chapter 1. Among mechanical devices, particular attention was given to spines, trichomes, and various glandu- lar and stellate hairs, although leaf toughness may also be important in affecting herbivory rates. Despite the demonstrated presence of a great number and variety of alkaloids in solanaceous plants, almost nothing is known about the effects of these compounds on the growth rates or viability of ithomiine or other lepidopteVous larvae. The phytochemical ecology of ithomiine butterflies and their solanaceous foodplants is completely unstudied, but promises to be one of the most exciting and rewarding avenues of investigation into the coevolution of insect-plant relationships. The data presented in Chapter 5 demonstrated that, among Limoncocha ithomiines that utilize more than one species of foodplant, larval develop- ment took considerably longer on solanaceous species that were armed (with spines and/or stellate or glandular pubescence) than on those 269 plants which lacked such mechanical defenses. Lack of chemical knowledge of these plants should temper the conclusion to be drawn from these data, but it seems quite likely that the mechanical devices of some solanaceous plants can have an a'dverse effect on ithomiine larval feeding efficiency. The question has been recently raised as to the possibility of "absolute" plant defenses that could prevent larval feeding completely. Gilbert (1971) described one such deterrent, hooked trichomes, on Passi- flora adenopoda. Species in the genus Passiflora serve as larval food- plants for Heliconius butterflies, but the widespread P. adenopoda is apparently immune to Heliconius herbivory as a result of the selective effect of the hooked trichomes. Gilbert (1971) suggests that the highly 'specific nature of these trichomes (they are not known to affect other groups of herbivores) indicates that they were evolved after the chemical defenses of the plant, which presumably act in a more general manner by excluding large classes of herbivores. Once most herbivores have been "filtered out" by chemical deterrents, Gilbert reasons, mechanical defenses specific for the remaining herbivores are* selected for. Rathcke and Poole (1975) recently suggested that the gregarious larvae of Mechanitis isthmia avoid the detrimental effects of spines on Solanum hirtum by spinning a silk webbing over the tops of the spines. Since silk production is nearly universal in lepidopterous larvae, they reasoned that the trichome-avoidance strategy used by M. isthmia should be available to Heliconius as well. It is doubtful that there are any "absolute" defenses or responses in the "coevolutionary race" between insects and plants. What may appear 270 as an "absolute" now probably represents only a "step" in the Ehrllch and Raven (196A) sequence. Only by studying all the factors that affect the relationship between ithomiine larvae (e.g., mandible strength, assimilation efficiencies, microsomal oxidases in the midgut and other detoxification systems, etc.) and their foodplants (mechanical and chemical characteristics of different parts of the plant at different ages, nutritional content of leaves, etc.) will a full understanding of the coevolutionary steps that have led to the present situation be gained. Larval Foodplant Relationships of the Ithomiinae An attempt was made during this study to collect all the published records of larval foodplants for the Ithomiinae. The results of this literature research, combined with the larval foodplant information ob- tained during my Ecuadorian field work, are presented in Table 18. All eight tribes of the Ithomiinae are represented by these records, which include at least preliminary information on 71 species and subspecies in 26 genera. Interpretations or conclusions drawn from the material in Table 18 must be made with caution, however, because, in the attempt to be all- inclusive, many vague and poorly documented records were included. Since some ithomiines are known to feed on Apocynaceae (see Table 18) and Gesneriaceae (W. A. Haber, pers. comm.), even records that merely state the family of the ithomiine foodplant have been included in Table 18. There is no way to check the accuracy of such records, however, since 271 o o r-tCM M N 00 - CN r-l o z < 5 e rs u U ft) s o o c 4J 4-1 £» O •H O »4 0U < < g. g. g. g. u u o o o o o o C o. a. a. < < < < o CO o to 0 a DC hi 3 § o J3 o v> a. s> o o. o Q» r-l SI o b 4-1 u a u Vi si x: >■ . •> 14 V o. « SI u o •a SI a: u « 00 hi ft. •a c a 3 Vi o •o H o u SI § r-1 o. •o o £ a) •J SI •J ••s ^"N 0 «« 01 3 3 -4 SI •r4 U 09 4 £> X X m h a t* SI (v. >-^ u N^ a) TO 3 a 3 3 r-l 1 r-l t4 O N •rl o N w M O u ft. UJ CD 3 z z l-l z > a: u rs o o > ri I SI SI r-l r-l ~-t r-l SI SI H H 3 z £ o rS o n 3 a SI a U « o ■rl ■o X a 4J > C r-l ■H a a. « a 01 C T-l •rl c U o <0 a fc a) *-> a o c a o c si e eg as SI SI l-l U o o X JC 4-1 J-» v4 -*4 H H u 01 M V. Z 4J M 0> H SI M Xt Z rl 3 o 2 u. 272 u o i-l o o CO 00 CO CO o CO •J s o ■J H rl H H rl i-4 CM -»»»•*-* e/> m >om»«o ri U CJ () o OS n: o: W w UJ u M o. u H N o d CJ > H w w a: to os a: pa to os of oa pa os tc a oi pa pa pa pa ac os pa oa OS os OS pa pa pa .J c 10 .-i o CO o > o o o • i-i O <-l £ H fl 4 u «*4 r^ r> c z 01 oa ■^ eo co i-i e «•> o X a > CO p-t 14 0) en c :eu c o o 4J o ~4 o 3 u cr tr u H A 9 JS c> C CO O CO > eo u *4 eo « ■4 o a aO a> 4J 14 o Xt 3 O u )4 a a. c . •» a •-4 eo u g 0) eo c 14 r*. CI ■ ►J 3 3 C c CO CO •-4 i-i o CO o CO |4 cu : o. o ■a u > Q. u <0 o « 14 U CO •rf C -4 CD CU CJ U JO C feu eo >4 1-1 out c o o CU •-4 40 c 4= CJ eg X V xeu -• X u I 3 C ■ pa ** « -M U CO ■•oj CO i-l 3 m x .-4 U O -H v_^ CO U. jO t CO ail in M 14 CO] 01 CO CO x i4 u i-4 CD W — 1 ca CO CO ffl *4 ■ri ^ k^ V^ X_* ^* %• "V^ N^- ee bs ac os otf es ei go oa co «o o > z CL. o o o R o 3 o C c a «-4 o § CO u • ■ XI m ; «i o u « >j 00 eg i~J *-v J o co a o c > o BUZ CJ D. O • •— « O 41 C .O .o a eg at Vj o — i t« •»« B M U O 3 B •" u 41 -O 3 X I § ■J eg eu c a | 41 r-i *j 0) •-I a C •*4 c J 41 c ■r* o o i-l -u » u v^ 01 n E 3 i 1 E 3 E 3 E 3 E 3 § I B 3 i I F c c c e C C c C c c C c c o eg <0 <0 a a CO eg a CO CO es CO eg i-1 o o o o o O o o o o o o o CO CO w CO CO CO CO CO co CO CO CO co -~ * eg c- u *-* TJ C <0 E O o eg c c 41 o .c o. CJ c o C a o w « * •H 41 9 X 41 X CO 3 en 01 3 C 01 •*-< CO 4J N c CO CO b b c o u • oo r-4 O r-t eg H co to 1-1 o IO CO to 00 en to 3 ■H V a CO CO N C CO 01 b B 41 •O C <~< J 41 3 pti ■8 u eu 3 oa en eg 3 BO t-t c< es cu 3 CJ en eu CO ■o >. U en O 41 •o u eg X I X s M 41 •J ^H eg T-l •^ •H en T CJ re; of u B CJ o > a* a o o c o a XI «4 hi T3 CO U. f*N CO « 1/1 3 hi CM O 3 CI X ■i-i at cl •u c hi O O CO eo 01 x J-i C a cj >■ ^ CO hi CO —I at «i x " i_i ot c XI a E ■-< o Cl hi >■ u ~) : 9C • eo •J c ai ffl fi Oj o hi «l O > jo a u 3 a a 6 «0 eg O hi U X 3 3 Cl 4-1 4J > CO co CJ O Q CM en c CD O CO OJ X 3 a a. .J «-> ►J ai XI c CO o CO •O 4J X CO a v^ > 3 •o o CQ CO CO -M •H C hi CO X CJ O. 3 CO 01 r-l —I 3 C CO •H U CJ u CO Cl CO z «H ■< » u -4 CO X a 0) eg M oi 01 i-l .-I *-« OJ O u CO 3 3 T3 01 at z u o o to re. m co eg ^t -r4 t4 hi hi hi >•>■>. x x x 4J U 4J o o o a ex a >■. >■ >. xl x x — 1 X w t> U X 01 3 ^^ 01 o •o CO JC C u. rH u o 01 Ol to 09 eg CO X U. *— ' C X -w CJ 01 0) CO a CO a eg c C o. c X hi 01 01 CO N»S 3 eg eg •JJ --I i-J >■ X > 03 a) eg • 3 c CO u eg o hi hi 3 •— i OJ 01 eg eg •H CO o. -O e CJ > CO " eg eg CO CO og C eg eg OJ C •H c p C f-l o 6 01 hi eg o 3 r-l hi eg >■ 3 r-l 3 o X hi OJ Cm W-l u a, 3 01 § m O c M eg eg CO r-l eg CO CO CO CO CO CO •w4 -i-l — 1 01 01 eg U hi hi c C r-l >~, > >< 01 01 3 X X X ex 01 •o i-l 4-1 4J o o •H o o o 01 01 CJ (X Cl o. a a eg >, >■ >, eg eg »-l X X X 2 Z re. o> eg X — i C CO 01 -~l hi Cn U a O <-* a a a CO t4 X. u « c X ml col Ol CO col S U\ co FJ 01 — < wl •ol -ol -o| O CO eg eg eg eg -^ t4 •rl t4 -r4 e e e B 6 o o o o o X X X X X i-l 4-1 4J i-l i-l 275 8 g o to ■ S CJN N N N U o 0 0 0*0 (4 M M M O BL os pi os pa a n c£ o£ O- CC p£ to O O O O 5 o o co to P. P. c c c c c u u a oj • ■ a O O r-l «-l r-l .-I r-l a. e. o o o o o < < to to co to en c c B C c c c c c c c a a a) - a 09 a . r-l r-i »-» r-l r-l r-l r-l r-c o o 0 O O o O o o O O CO CO m CO CO co CO CO co to CO vc. ex: w X o o > 0- o o o rv. tsl S 9 § § o u ■w -r< 8} » a a a o a o r-l u • g § -> B _ C u at 9 m co a> o VI ■H r- v4 o r-l U ■rl o c s ■M B o 3 *** 3 X •rl 01 «4 a t— I a r-v r-l « E O u OS O c cw c r-l tM > • «rl oj m «h ■s S • ~ 5 3 w rH ■ a. > U eo 01 oi -— m « ■ X. & s u c E 3 9 1 § u u a C c c c O -rl •rl a • o o o o : U) •Jl to to to CO u o X ■J •o o> > u. 5 U i-» u a c -n > o> < ■a eg ■ c 01 a o x: xr o •3 o, «x (I a i. 2 «-' m o. u • *o c 1 j: o rl 01 r"S TJ C r-l •rl ft) > rV V C U O CO «8 PS -0 •o t) 3 C « » o w «*-i X •a rH a) a r-l ri r-l V •H JZ J-J at a ai B °- •o •J o w Cl- io § B ■o s 5 cgi i-i ■ el a *j Ed H •h «1 « X1 a <0 a T-i •e4 B B f> o -C X. iJ i_) i-i I-I 01 X -H ai a B rl -H CO rl H 3 01 a « a) •rl -r«i vl Vl )-l ri 01 ■ •H a x: o CU 3 K 3 r-l r-l Ol Ji O -HI ■ tw J 01 o »«l c a «-» o CO u 3 0J E 3 CJ § 3 & 1-1 c c u en a a CO a t-i i-i 0> CO o o (_> o CO CO o to u c 41 CD 3 C t2 to 1 hi 41 3 I o hi o. P* »H s* hi 4-1 a a ■ a CO c C ■ iH I CO £ 3 43 co .-H a CJ •h H CO hi -o -H CO a •*4 UJ ±J M c CJ o U1 u p- en O CO l-l 1 SI to l-l IC x> l-l eg pi tr> H V c a c — i «- -J 3 3 3 C7 • U C7 V. CI li a o « a. c cJ u s> a u oo <■*> _o PI o r- u w CL T-t R tz E E V4 3 3 3 3 01 1- L. Vi L. .c 4J t-i u u u CO eo CB CO —1 ai 41 SI «J Z* CJ CJ u O •I C e ■ o 4J Vi ' CM e ° - u 8 a si CI a c a <-■ o co - u 3 vl g >■■ •H CI «— •■ l-l o ■o «M c «t-4 ■ 4J CL U ■ CO •-* i 1 § u ft Li 4-t JJ l> a> 09 O) 01 CJ 0) 01 u U l-l o X • #•* TJ -sr CM § en ■a r» ~* s^ u o *** •H § o o CO en a c ^^ > r» — i en .O *— ' u in o > «*i w. • t-i ~, r» a a eg a * — • CI »CI c CI «l SI eg eg a a CO eg •H — 1 «-i CI CI SI eg eg o u u CI wt i-i (O ■ a eg SI 0) c E c U-. «M «i eg eg eg C E c f-i «-4 «-l 3 3 3 o o o Vi {4 5 co CO CO CO « 3 U SI B =•§ X X CL CI CI ^^ ea •M CO C ■x. SI 01 •** 4J eg B eg X CO tJ pa ■^ LI ^ *m* U •-I a M •g SI u O CO l-l i-> O SI 01 VC •£ X u S9 a I SI E SI eg eg B ea CI B -O eg « eg O ea u LI CJ CO 01 SI 0) f-l 1-1 eg ■0 IS eg ■H SI > > IK id eg >■ ■) N N M T-I eg CO eo eo CO T* ■H -J 0 Li H Li Li >- >, >■ SI -a ■0 ■o i-i 0 0 O SI 0 ts CJ X LI SI > -a r-l X. eg o ^v co co S 1 *-. O eg] X B eg Li CI O Li >\ -H eg -H r-ll Li l> 5 Li -H 01 SI 5 X B 1 M X 5 CI X SI X E X "^ -H !N ^^ LI « s^ eg E c 01 CO 1-1 ■g O f-H eg w-l J3 Li eg 01 0 3 CI CI Li Li u B 0 O O SI 05 n 0) eg eg « eg iJ «u i-J Li u Vi HI •r-< t-t •H SI SI SI 3 r c c rH l-l f-l 0) CJ O 0 O -H 1 6) >• cv| a S CI CI X X X X X xl a- 5 co •0 B 278 g o 00 O o in m VD y£) vXJ pi OS 0l CQ PQ CQ a: c* CQ U M O tal •V V 3 c co 2 5 c c c e e e c e c o «i co c > ^ — • « ai E E *w ■ *4 C CO Vi 3 --I O. : : to co i- 3 a •«H CO •*4 «0 co 01 3 t-4 01 «M C •u c 3 00 •-I § u o Vl pa to pa r~ c ^^ « to « CQ § • ~J X T3 01 00 £ >-• 01 ID VI O. c ■^1 -M Vl CO «M T-t U-l u ■ c ^ 01 c X CO *-l •-• ~4 o ^ to !»■ ** o o 0» O- 00 «j ~j *» ^i OS H CO CO N N N N h CM - w-t <-l«£rtJOOOOteN OX C • COi-lr-ti-tr-lr-l —I — I - CO - 01 ,-t X -I V. U X. ..... - X JC •* «J U - 00 00 oo eo 00 tl3B«l^£CCCCCC >sUCVi>-IU0.33333 C0UC0O3C00IOOOOO •er^cooo.-ifNio.j-i/isor- 3 O J= u 01 c •3 X 3 3 •O •o CO CO o O v> u -o •o W 01 CO CO O o B c cs CO 0 « 00 00 CO CO to 4J VI b) CO CO w u tl 01 u Vi Vl 0. 3 3 CO 4J 4J C7> — I ^4 Oft C* CM iJ • .-I IO X CO • » • i-l wt c* UJ —• -3 <* tt a a co r-< e» o ^ B E - o t<» 4 • co oi >£ — I -H • JJ C «0 . JC •-• r-< •JtJv>*i^o\vi4-H333 UUOb]tv,tkiOUUUU o u « « « N ch o <-^ cm c*) ^ «n jo cn as x? 3oim^ cooi-t >«Q.O\C*C>iO Vi\OnO CO ^ CO C X — I oi c CO BO •OG^»-l-*r-.30">ON J3 r-l » 33 CT.03— i— li_ivO ■ - co eo os ^ o»-eoa •O'O'O'O - OOOf-4 CO0I0I0ICUCCC • 0JgEE6oj(i)a]o303pCC ">e^>-i-vooo VlV.--- C0f«.-H^4-HVlVlVl O-OTJ-O-OcQcaaacQCQCOBacQ r-cc>4cn*oriONOr.cocT%Or-«rM 41 -J 9 •H • O. CO x*o J « • jj W-4 o a >«M s c a-H X-O o • .1 ■ « • ■***■* B a 1J c § M X «4 4) c o jj »-l > CO c O to *i Q 01 -j o 1 oo r-l 4J B C e o oo ai • c "* tl -C > ■ o -< •H ■ ■ 60 41 «l c 41 -H • 00 0) « »J JJ • . -J5 f c o u Owl eg •HJ1 « ^•r-l c > >r-l • > » « •H 41 41 • > o 4) o u h HH >-■ • — 1 > >. U O eg 41 41 B • 4) a-4 •-I 03 v. •rlc-O X >r* • ox JJ « JJ B BOO* Ji I- X B.4I eg eg .—* JJ en o • U «l 4J O'-'f-l (/> eg Oilo • *-!.»■». 0J u •H tC i"l> r"* 41 O V -rl c Bi ■ ■ ■ eg 41 - « eo x> •-. 1- • — i jj o O-n c« b 8 C ■ ■ X cfl W« C r* > eg o eg u4 a -J r-l O W og 3 rs co tir-i S o co o B ex — 1 3 0) 1/5 eg ss tJ JJ 3 CO DO (V. jj co eg C JJ Vi E U r-l O ■ C CO -r< CQ CO CO o • Ji 0) EX-rl 3 jc in a. - o 41 -rl c C JJ CO 41 w> ^—, r>U E o C • cc CO C encu t. o c a C •H OI ,►£- 41 eo eg « o. ento >UX eg to • > eg rH a) o --4 VIOI» JJ 2 ►« o-r< o .* co C u • u OO ^^13" to « — ' oo o l-I — 1-J eg E Cj e o o •H r- o < 3 41 JJ UJt • _o Oi J . "• ~* a > UX 01 V> B ■ o •r« ai B« u t-i 41 t> 41 JJ S eg CD c hinujl > 2 C E -H at q.3 cr o»-» 3 ■~4^ 3 « to • 01 a) o-> c • o-w ^o u SK u br> u Xr-lt-t — i C — 1 i co 41 CS JJ 41 • ll JJ a m v «-- 3 Jj^xr eg u (J ti o c OI -O 41 a 1^ p. > -o 9 to o 01 . . tg •O- JJ oo --H r-l CO r-l B C eg B • t> •H Ov •k i-o oi x c o c CO a ■ 41 eg tg o O-rt tg C «M • IJ •«4 a r-l ra — i eg • V- u Jol J OCX vgo in .-r< 0) NiH eg 3 bCt. N «< o eg eg a < eg CO u • ■o JJ o > « o*acou i4 41 r-l •SO 3 * V c c iH UO • u u-o • OrJ O — 1 • CO r-l r^ • u qj es to eg o O^-v O •g c E 01 E c es eg X ■ U •*>•»•) u • l-i -o — 1 tg x X a CI a -r. fl-rl JJ O 41 tl C ■ o et (0 r-i O-H O oo i-l IB /-s X -rl en l. r-l 41 41 U O-rJ O IS u C 41 CO-H OH X c C • CD r-v 4-> T3 C «i 3 to -O - O — JX X 0) — h - 3 a o. C 3 -rt o B-. Ur-r~ 3 V. X a C 13 00 41 r-» tl • h. -u to 1 :°t H r-l o a o> «l « •>] U(X. - CO P CO — 41 Q. g .S . :l o w o X 1- —• JJ t-H CO 3 O WiJ - • • o rl 4-1 41 eg oo OOUO wo - JJ V- 14 . I< . > rJ O - c . 41 O 9 > •» tl COM) C c m c CO -* eo tV, >P, > O - u > O > ■o •O 3 CO 0) >s w-i ■ • CO CO o O tti4 a eg ai —4 o o u ■ -d O 1-1 O c Cto to O •a •OTJ >i W -< -H <0 <0 CO i—i c eg oi >-< to u Cu > u oo ll eg eg II u c c c !->>M cc co t-: V* o a Vj jj P- eg cw o > eg 0-> 41 P. tJ 1 JZ < -rl <0 3 O 41 JJ .< r-l r-< eg JJ - ■W v/U JJ tp to tg oi eg eo Dm u ti m O •o r-l 01 o CO tn to B n-4 ci S ■-4 » 41 O -J V- — 1 O P. C N c tl c OH >vr-l 01 c C 3 4> U u es in t- co "O CO tg -O BTJI C ■rJ tg tg 3 to x a X> X -r4 01 o B IJM to co )-i eg-ri eg ffl C 41 JJ 41 O O W. C H C E OX JJ 3 C CO o> 1-1 4J JJ B 01 4) eg a. ci C <0 C 0.-J O c au o a. 10 >N c c eo c -H 3 3 M JJ 09 3 O 4> -J to o — •* —4 3 41 3 oi eg i-i to —i to a 3 to •Jr-t 0) 0) o e CO -rl D.JJ u no a o 01 01 r-l r* O* Hrjn-jinu . .— 1 HNri«jinio HNrt i-i r-l —t r-ltN 01 v. >- N-r Sr.S^ W 0) u •o ^^ ^^ . ox 01 •o x » o v> u 01 u 1-1 01 eg O « 3 Id e X CS 41 ■a 41 tg C ~r4 0) o o -H u4 to c ■a r-l V4 X c 5j •-< OS h ?• ^< e to rS 41 u C9 i-t o c a r-l o r-l 3 tO •o a 3 C o 3 X 5 |4 •H • •o tg O c -rl 3 N co co 4 U 14 JJ eg eg to 3 -rl C CO 41 u 41 i-> CO eg eg X 3 3 Vi tl 1-1 3 c B JJ 3 u o 3 O rH 41 41 3 I* Vi 41 1 a e U ■■ < to U u to u z a. to t-l r> > C oi c B oo CO 41 41 nnnon pre rec i-i ►J 3 < as •a 00 u 3 O to u 00 a. to OS fr- g 12 X o o a o .►4 fr-i * + 280 no specific plant names are given. It may be tempting to entomologists to simply list the family name normally associated with the Ithomiinae, rather than determine the true identity of the foodplant by consulting a botanist or an appropriate botanical text. In one publication (Costa Lima, 1936) foodplants of Brazilian ithomiines were listed by common names (in Portuguese) rather than by their Latin counterparts. By comparing these common names with those given in other papers that also included the scientific names, a few of these plants have been identified. In all cases where only common names of plants were given, these have been entered in Table 18, followed by my ad hoc determination of the Latin name in brackets. Another source of error in these data may be the mis-association of ithomiines with foodplants. For example, Gilbert (1969) presented 18 associations between Costa Rican ithomiines and solanaceous plants, but because of the limited field time available, he was able to verify only 5 of these by rearing the immatures to the adult stage. The remainder of the associations were made by comparing the size and shape of eggs found on solanaceous plants with eggs dissected from female ithomiines. Although Gilbert claimed it was possible in most cases to match the eggs with great certainty, the associations remain unverified by actual rearings. Many of the Limoncocha Solanaceae submitted to Dr. D'Arcy for deter- mination have not yet been completely identified, owing to the scarcity of Ecuadorian Solanaceae in museum collections and the primitive state of the taxonomy of Amazonian Solanaceae. As a result, many species have so far been identified only to genus, although for some, the nearest 281 relative or affinity can be given in Tables 5 and 18. In addition, the plant names given by some authors (e.g., Rathcke and Poole, 1975; Gilbert, 1969; d'Almeida, 1922) were apparently not verified by botanists, t and in view of the many mis-identifications and lengthy synonymies in the history of the Solanaceae (D'Arcy, 1973), further caution is needed when using these names. Patterns of Foodplant Utilization Not surprisingly, perhaps, many of the foodplant records in Table 18, especially those from Brazil, are of common garden plants, often grown as crops (e.g. , Lycopersicon esculentum and Cyphomandra crassicaulis) or ornamentals (e.g., Brunfelsia hopeana, ji. calycina, Datura arborea) . Undoubtedly, most of these occur in a semi-wild state as well, especially in clearings and along roadsides, areas frequently visited by collectors. Only in the three studies that have focused on the foodplant relation- ships of a forest community of ithomiines (Gilbert, 1969; Young, 1972, 1973, 1974a, b,c; present study) have non-cultivated Solanaceae been well-represented in the larval foodplant records. The Brazilian records, if considered alone, would probably present an extremely biased picture of foodplant utilization of the Ithomiinae by overemphasizing the ithomiine species with broad foodplant tolerances and high vagility (e.g., Mechanitis lysimnia, Prittwitzia hymenaea , Thyridia themisto) . A mean- ingful understanding of the foodplant relationships of the Ithomiinae and of the revolutionary pathways that led to the present relationships await more detailed studies of natural communities. Despite these difficulties, however, it is of interest to attempt to summarize the foodplant data presently available for the Ithomiinae. 282 The tiered histogram in Figure 54 provides a broad overview of these data at the tribal level. Ithomiinae have been reported feeding on eleven solanaceous genera, but, not unexpectedly, over half of the foodplant records involve species of the extremely large genus Solanum. The scarcity of foodplant records in some tribes (e.g., Tithoreini and Melinaeini) makes generalizations or tribal comparisons difficult, but it is apparent that the Mechanitini and Dircennini display the broadest patterns of foodplant utilization. These two tribes have been reported from six and five genera of Solanaceae, respectively, including at least four subgenera of Solanum. In the Mechanitini this pattern is mirrored by the broad foodplant utilization of many Mechanitis species. Fox (1967) has pointed out the extreme intraspecif ic variability in color pattern of many species in the genus Mechanitis and suggested that these "plastic species" have apparently developed genetic traits that enable them to evolve rapidly and to take advantage of the ecological rearrangement of tropical forests that attended the cyclic glaciation of the northern hemisphere. As the tropical forest refugia expanded in the interglacial periods, one consequence of the renewed sympatry of these partially differentiated species and subspecies would be a "genetic revolution" (Mayr, 1954) resulting from population hybrid- izations.- The recombinations resulting from such hybridizations could alter the gene complexes controlling foodplant specificity and result in broader foodplant tolerances in both ovipositing females and larvae. Broad foodplant tolerances could be selectively advantageous to individuals in populations expanding their range beyond that of their original food- plants. co • a co •h a j= a CO cu C 00 o •H CO w p CO O iH .x> S-l r-1 -H H 4-1 (-1 Vj -W 0) 0) C 4-) CU •H ■u C •H 3 -H g JO iH O rf3 ■u O -C M •U H -H IH to O ,0 CM e o 0] 3 a. B G •H o J3 QJ -H CO J2 -U c: 4-1 Cd o c •H CO -H 4J cu .n CD +j e iH cd o a) o o Pi •H T3 J= •u C O c •H td cd CU iH ■u O. >4-4 M T) cu o o i-i 4-1 o b-i ■u t3 cd Q) iH 4J CO 0) M > r-H O J* Cd O. cd O 0) J CA iJ • -a- m a> V-J 3 60 Pn 284 C C c [ [ [ [ 1 C : C • [ [ t [ [ 1 * c c - • c c [ [ [ [ L_ [ [ [ [ [ c 1 C 1 c 1 1 1 t 1 1 t 1 1 1 c 1 It 1 1 t 1 1 1 1 Mill X) in m m 10 IO in in WITHERINGIA Unknown Solanum o Potatoe Pachystemonum Leptostemonum Brevantherium Bassovia PHYSALIS LYCOPERSICON LYCIANTHES JUANULLOA DATURA CYPHOMANDRA CESTRUM CAPSICUM BRUNFELSIA APOCYNACEAE M Z 53 z z w o o w 51 § w o M z pi z M a M g Q O o 285 At present, the Tithoreini have been reported feeding only on Apocynaceae and the Melinaeini only on Juanulloa. The other six ithomiine tribes have all been reported on at least three solanaceous genera. Only the genus S planum has been reported among the foodplants of all six of these tribes, although Cestrum has been reported for four of them. Undoubtedly, however, the eleven genera reported here and also many others in the Solanaceae will become better represented among the ithomiine tribes as the foodplants of more species are discovered. It was possible to place most of the identified Solanum species in the appropriate subgenus. So far, six of the seven subgenera of this large genus have been reported as ithomiine foodplants. Somewhat surprisingly, the subgenus Leptostemonum, containing many heavily spined or otherwise armed species (D'Arcy, 1973), is the best represented. Many of these are weedy species, however, and are thus more likely to be encountered by entomologists, which may explain their strong repre- - sentation among reported foodplants. The only published records of ithomiines feeding of non-solanaceous plants are those for Tithorea (Tithoreini) and Aeria (Oleriini) feeding on Apocynaceae, although the records for Aeria are poorly documented. At least one species of Apocynaceae (Parsonia velutina) has been reported as a foodplant of Tellervo zoilus (Tellervinae) in Australia (see Table 18) raising the possibility that the original foodplants of primitive Ithomi- idae were in the Apocynaceae and that the switch to the Solanaceae oc- curred later in the evolution of the group. The close relationship between the Ithomiidae and the Danaidae (Gilbert and Ehrlich, 1970) discussed 286 earlier is further supported by the fact that Tellervo has the same food- plant affinities as the Danaidae, Asclepiadaceae and Apocynaceae. The importance of the Apocynaceae in the repertoire of larval foodplants of primitive Ithomiinae (namely, Tithorea) has been recently discussed by Edgar et al. (1974). These authors suggested that the ancestral larval foodplants of the Ithomiidae (presumably Apocynaceae) contained both pyrrolizidine alkaloids (PA) , which provided male ithomi- ines with the precursors for sex pheromone manufacture (see Chapter 4) , and cardiac glycosides, which rendered both sexes unpalatable. After the switch to the Solanaceae, a family that lacks PAs, adult male ithomiines began visiting flowers of PA-containing plants (Asteraceae, Boraginaceae) to obtain the compounds needed for the manufacture of sex pheromones essential to successful courtship. Presumably, the toxic alkaloids in the Solanaceae allowed the continuance of the unpalat- ability previously provided by the Apocynaceous cardenolides . This is a tenuous theory, however, erected by analogy to a coevolutionary se- quence proposed for the Danaidae and their larval foodplants , and because of the established presence of a dihydropyrrolizine in Urechites karwinsky , the larval foodplant of Tithorea harmonia salvadores in El Salvador. Foodplant Specificity of Limoncocha Ithomiines The foodplant relationships of the ithomiines studied at Limoncocha are presented diagrammatically in Figure 55. Of the 53 Ithomiinae recorded from Limoncocha, foodplants were found for 27, although two of these ithomiines remain unidentified since they were not raised through to the adult stage. One of the five species of Mechanitis occasionally 287 Melinaea menophilus Forbestra truncata Mechanitis isthmia Mechanitis lysimnia Mechanitis mazaeus Mechanitis messenoides Scada batesi Xanthocleis psidii Hypothyris euclea Hypothyris fluona Hypothyris fulminans Napeogenes corena Napeogenes pharo Ithomia agnosia Ithomia amarilla Ithomia derasa Oleria agarista Ceratinia poecila Pteronymia sparsa X-l (5051) Godyris zavaleta Heterosais edessa Hypoleria orolina Pseudoscada timna Thyridia confusa Mechanitis sp. Unknown ithomiine Unknown ithomiine • Juanulloa sp. (7354) Lycopersicon esculentum Cyphomandra hartwegii Solanum anceps Solanum aff . antillarum Solanum bicolor Solanum coconilla Solanum evolvulifolium Solanum aff. lancaeifolium Solanum aff. nudum Solanum pectinatum Solanum quitoe'nse Solanum aff. schlechtendalilianum Solanum Solanum Solanum Solanum Solanum Solanum Solanum Solanum sp. sp. sp. sp. sp. sp. sp. sp. (7304) (7311) (7319) (7327) (7331) (7333) (7334) (7336) Lycianthes aff. howardiana (7307) Lycianthes sp. (7325 = 7307?) Lycianthes aff. maxonii Physalis angulata Physalis pubescens Cestrum (7338) Cestrum (l) sp. (7324) Solanaceae sp. (7353) Brunfelsia sp. (7330) Solanum sp. (7313) Solanum aff. intermedium Witheringia riparia Figure 55. Larval Foodplants of Limoncocha Ithomiinae. 288 laid small clusters of eggs on Solanum sp. 7313, but, as explained below, the larvae were unable to feed on the leaves of this plant and thus it is not known which Mechanitis species laid the eggs. (Mechanitis eggs are not consistently different enough in size or shape to allow their identification.) Of the 42 species of solanaceous plants found at Limoncocha, 33 were utilized by at least one ithomiine species. As stated earlier, however, many more species of Solanaceae undoubtably exist at Limoncocha and thus it is not possible to say what percentage of the solanaceous flora at Limoncocha is utilized by the Ithomiinae. In ad- dition, the fact that no early stages were found on ten of the solanaceous plants observed at Limoncocha (despite repeated searches of most of these) does not eliminate the possibility that they serve as ithomiine foodplants. The low densities of immature stages of most ithomiine species (see next section) make their discovery difficult. For exam- ple, Wither ingia riparia was searched repeatedly for a period of over four months before eggs and larvae were found on the leaves (unfortunately, these immatures did not survive to maturity) . Based on the information gathered to date at Limoncocha, the great majority of ithomiines in this community lay their eggs and complete their life cycle on a single foodplant (Figure 56). In other words, the ithomiine community at Limoncocha exhibits high foodplant specificity. Only five species of ithomiines (19%) were found to utilize more than one foodplant (Figure 56) , and three of these were in the genus Mechanitis (Figure 55) . Likewise, the degree of overlap in foodplant utilization by ithomiines is quite low, with 85% of the solanaceous hosts supporting only one species of ithomiine (Figure 57) . gure 56. Numbers of Solanaceous Foodplants Utilized by Species of Ithomiinae at Limoncocha. igure 57. Numbers of Ithomiinae Supported by Species of Solanaceae at Limoncocha. 30T 3 25 M O H PH CO Pi w 20 - M M § 15 o 10 5- 290 N - 27 i — I CZZ) 12 3 A 5 NUMBER OF SOLANACEOUS FOODPLANTS UTILIZED CO 30t 25- • a o 2 20- u o o 15+ to 10-- Exi O pi M N - 33 I 1 12 3 4 5 NUMBER OF ITHOMIINES UTILIZING THE SAME F00DPLANT 291 Similar patterns of high foodplant specificity have been found in the ithomiine communities studied in Costa Rica by Gilbert (1969) at La Selva (5 of 5 species utilized one foodplant) and San Vito (9 of 11 species utilized one foodplant) and by Young (1972, 1974a, b,c) at Cuesta Angel (3 of 4 species utilized one foodplant). Such high foodplant specificity exhibited by Ithomiinae in Costa Rica and Ecuador suggests that the diversity of the family Solanaceae has influenced the diversity of the Ithomiinae by providing a vast array of potential foodplants on which ithomiines have specialized. Although members of the Solanaceae share a basic chemical similarity, the alkaloid content of each species is probably qualitatively and quan- titatively unique. Likewise, the Ithomiinae share a physiological abil- ity to metabolize solanaceous tissues, but the specific chemical composi- tion of the larval foodplant probably requires a corresponding enzymatic specificity on the part -of the "butterfly utilizing it. As selection favors increased efficiency on a particular foodplant (e.g., due to the advantage of reducing generation time to reduce rates of juvenile para- sitism and predation) , the increasing physiological specialization on the part of the butterfly gradually narrows its foodplant tolerance. The exact nature and degree of this specialization and its consequences have not been investigated for any ithomiine, although such studies are clearly needed. A few preliminary observations on the larval tolerances of Mechanitis isthmia were made during this study, however. Briefly summarized, the M. isthmia feeding experiments led to two conclusions. First, freshly hatched (naive) M. isthmia larvae will feed and develop on any foodplant normally oviposited on by females of this 292 species, regardless of the foodplant chosen for oviposition by their own female parent. Second, if the larvae first feed on one foodplant and are later transferred to another, they will rarely feed on the second plant and almost never complete development. Thus the broad larval tolerances to foodplants exhibited by M. isthmia are characteristic only of naive larvae. Freshly hatched M. isthmia larvae were also placed on two solanaceous plants fed upon by other Limoncocha Mechanitis but not used by M. isthmia. Sixteen M. isthmia larvae placed on Solanum sp. 7333 (the foodplant of M. mazaeus) fed and developed normally through second instar (lack of available food ended the experiment here) . Fifty-two M. isthmia larvae (in three groups) placed on Solanum pectinatum (a foodplant of M. lysimnia and M. messenoides) failed to complete the first instar, apparently un- able to cope with the plant's dense stellate and glandular pubescence in which the larvae appeared to founder. This paralleled the experience of newly hatched larvae of an unknown Mechanitis species that laid two egg clusters on Solanum sp. 7313, on which the larvae were unable to complete the first instar, apparently because of the trichomes and dense stellate pubescence. Thus, it appears that mechanical, as well as chem- ical, defenses of plants require specialized adaptations on the part of the larvae that are able to successfully utilize them as foodplants. Utilization of Larval Foodplants at Limoncocha The great under -utilization of solanaceous leaf tissue by Limoncocha ithomiines was alluded to in Chapter 5, and the low densities of immature 293 stages were attributed to high levels of parasitism (especially of eggs) and predation (especially of larvae) acting on relatively low daily rates of ithomiine reproduction. « To quantify the low densities of juvenile ithomiines, and to pro- vide some estimates of the abundances of solanaceous foodplants, vege- tational surveys were conducted in areas of both primary and secondary forests as described in Chapter 2. The results of these surveys are given in Tables 19 through 23. The surveys at the two .25 hectare pri- mary forest plots, Site 4 (Table 19) and Site 5 (Table 20), will be considered first (see map, Figure 2). The number of solanaceous plants present at Site 4 (Table 19) was exceedingly low, there being only 16 individuals of 4 species, for a density of 64 plants per hectare. By contrast, Site 5 (Table 20) had 8 species and 224 individual plants, for a density of 896 plants per hectare. The most probable reason for the vegetational difference in the two sites was the periodic flooding that occurred at Site 4 (see Chapter 3). Indeed, six of the plants at Site 4 (two Lycianthes, two Solanum sp. 7319, and two Solanum sp. 7327) were growing not from the ground but were rooted in the decaying trunks of fallen trees so that their root systems, located at least 80 cm above the ground, were thus free from flooding. All the remaining plants were growing in the subplots that experienced little or no flooding. The forest of the NT observation area (Site 5) does not flood, however, and the solanaceous plants at this site were- greater in number and were distributed through- out the plot. 294 0) 4J GO 0) J-l c •H g c to to s 0) 3 o O s to .H o CO CO K CU o\ J3 tO H Pt4 H O 53 w < O Bi w hI < W W pu, « pq > <3 . Pi w w CO m CO oo i-H m co r^ o\ CTi r~- tN m CO m in CM oo d o m \o o r>» CO o CO co /-s ^"\ r-» r~ Oi r~- ^-^ **■• .H CM CO CO • • 1^. r-» p. a. N—' ■*— co CO cfl to CX a. •H 01 CO to co X, rH 4J C2 e 0) a 3 3 IH tO c C C ■H CO CO 3 O .h -H M >! o o pq hJ CO CO CO 0) o CI) co CD u 3 CD C •a c cd oa •u § a, CO o CD O cfl C cfl rH O CO «M o CO U CD 1 o tN at r-l J3 CO H o H pel O M w re o o CO o o o o o O o CO oo U0 m CN o CN ON CO CM CO "0 oo CM CM CTv CM 00 .h vO CD s_^ CJ CO CO a tH CO CO i-H 4J O O w H 296 CD •H CO co 0) M 3 •U c0 0) C c co to 3 o 0) o to c (0 o CO CD M CD 1 CN 0) .H .O CO H o ss w o H 53 U H M 53 in c\i oo m o o\ • • o .-t o CM SO vO CM vO 00 VO m CN r-- vO .H O CO CM 00 00 vO O o o Cvj o 00 vO o en o. co CO 0) 1-1 c CO •H o CO i-l iH en •H P, c a) O o CJ G O rt o c CO O CO d cfl o CO m ^ P. • • co p- a w 03 m S s 3 3 CO 3 3 .H CO cO to r-i rH •u O O o CO CO H 297 SO co 4J co CO 3 (U 3 s o r3 c CO 0) 4-1 3 to Ph (0 3 O a) a to 3 CO rH O CO CO u x> CM CN 00 oo -3- CM /— \ 3 CO 0) Vj 0) 0) P3 o o 60 cu 4-> M cfl rC cfl u T3 3 CO o X, 3. >, r-. o CO a. CO CO 0) rtS 4-> 3 CO •H CJ ►J m o O rH CO CO co co rH co a\ CO SO CM CO CM co CO CO IT) so o co co sr CO CO CO r^ r^ r^ co g 3 CO rH O CO co P-. ca 3. CO s c CO rH O CO a CO ex CO § 3 cO rH o CO 3« co p. CO ex CO Pu CO 3 cO rH o CO 3. CO § 3 CO rH o CO u CO CXI cO ■H cT •H M cu CO cu •H u cu 3. co oo to 4-1 o H 298 03 c Q s M a; -a CO CM a) x> & o oo o o m o IT) -d- CO CM o o m o o CO 00 oo . o o Sf o m *■ J CO so so en CM CM o o o -3- CM CM O lO CM CM o- sO o O sO o\ to lO oo r-- o CO oo to oo -3- lO iH SO ID 0 a a) 4-> ■u ■u 4-> •H •H 1-1 ■H CO CO CO CO 299 There were no juvenile ithomiines present at Site 4 on the day of the census, although larvae of Thyridia confusa had been collected there (on Brunfelsia sp. 7330) on previous occasions. At Site 5, the density of immatures was 48 eggs and 32 larvae per hectare (these figures being quadruple the amounts actually found on the .25 hectare survey plot). The density per plant was about .09, but the number of immatures per leaf was .007. The two most abundant plants, Solanum sp. 7327 and Witheringia riparia, are quite similar morphologically and occurred together in large stands in -favorable- spots of the forest. Nevertheless, the densities of ithomiine immatures on these plants were still quite low. The two 100 m by 2 m transects made along the Logging Trail (in full sun) at Site 1 (Table 21) and along the Nature Trail (in open shade) at Site 6 (Table 22) revealed much higher densities of ithomiine immatures. At Site 1, most of these immatures were Mechanitis isthmia eggs (61 distributed among three clusters) which were found in an extensive stand of the foodplant (S. coconilla) that dominated the area. A much greater diversity of solanaceous plants was found in the shaded Site 6 transect, although, as was Site 5, this area was dominated by Solanum sp. 7327 and Witheringia riparia. The densities of ithomiine immatures at the two secondary forest sites were 4300 eggs and 800 larvae per hectare at Site 1, and 2650 eggs and 800 larvae per hectare at Site 6. The distribution of these juvenile ithomiines, however, was strongly clumped both among and within species. Table 23 summarizes the densities of the foodplants and -the ^-ithomiine immatures at each of the four survey sites in terms Nbf " the number of individuals per hectare. 300 The fact that only three clusters of M. isthmia eggs were found in the large stand of S^. coconilla (68 plants) at Site 1 may be the result of a tendency, mentioned earlier, of M. isthmia females to seek out isolated foodplants for ovipbsition. An explanation offered for this behavior in Chapter 5 was the possibility that egg cannibalism by older larvae might select for females which lay eggs on well-isolated plants where other larvae might less likely be present. Ant predation may be another factor favoring this ovipositing behavior. Nearly all plants in the S_. coconilla stand at Site 1 supported large colonies of Solanopsis ants that nested among the clusters of fruit. Although ant densities were not quantified, such colonies on S. coconilla appeared to be smaller in size and less frequent in occurrence on single, isolated plants than on plants in large stands. Thus, it appears possible that the egg and early instar larval predation (Chapter 5) associated with large ant colonies on clustered S^. coconilla may have been a factor selecting for searching behavior in ovipositing M. isthmia females that favors isolated instead of clustered foodplants. CHAPTER VII « THE MIMETIC RELATIONSHIPS OF THE ITHOMIINAE There are hardly any species of Ithomiines that are not roughly, but very often most exactly, copied by Ithomiine species of other genera. — W. J. Kaye (1914) Kaye's succinct observation readily explains why the Ithomiinae have been able to simultaneously produce headaches in taxonomists and delight in students of mimicry. The extensive participation of ithomi- ines in mimicry complexes has long been recognized and, indeed, the original formulations of Batesian Mimicry (Bates, 1862) and Mullerian Mimicry (Muller, 1879), were based on neotropical mimetic complexes involving ithomiines. The literature on mimicry in butterflies is voluminous and has been summarized several times, most recently by Papageorgis (1974). Mimicry theory and considerations of the mimicry complexes entered by all members of the Ithomiinae are beyond the scope of this chapter. The purposes of the following discussion are (1) to characterize the mimicry complexes at Limoncocha in which ithomiines participate, (2) to review the evidence available concerning the palatability of ithomiines, and (3) to examine the consequences to mimicry of some aspects of ithomiine community ecology. 301 302 The Mimetic Subcomplexes at Llmoncocha As stated earlier (Chapter 1), ithomiines may be generally clas- « sified as belonging to one of two classes of wing pattern and coloration- those with part or most of the wings transparent or apparently so (Transparent Complex) and those with opaque wings, usually with some combination of black, orange, and yellow coloration (Tiger Complex). These two mimetic complexes, the Transparent Complex and the Tiger Complex, are the largest of the five mimetic associations occurring at Limoncocha and are the only two in which ithomiines participate. The other three mimicry complexes are the Orange Complex (red-orange wings with black borders) , the Blue Complex (blue basal areas on black wings, with yellow forewing markings), and the Red Complex (red basal areas on black wings, with yellow forewing markings). These three com- plexes consist largely of species of Heliconiinae, but include some Batesian mimics (mostly butterflies) and some Mullerian mimics (mostly diurnal moths) . All five complexes coexist throughout most of Amazonian South America and were recently studied in Peru by Papageorgis (1974, 1975), whose terminology for these complexes is followed here. Papageorgis found that each of these mimicry complexes occupied a different vertical stratum of the forest at her study sites in Peru. At the site most comparable to Limoncocha (Rio Llullapichis) , the com- plexes were segregated in the following manner: the Transparent Complex flew between ground level and 2 m, the Tiger Complex flew between 1.5 and 12 m, the Red Complex flew between 5 and 17 m, the Blue Complex flew between 15 and 25 m, and the Orange Complex flew above the canopy. 303 The degree of overlap between adjacent groups was in all cases rather low. My observations on the relative flying heights of the mimicry complexes at Limoncocha, although not quantified, agree fairly well with those of Papageorgis (1974)*. For example, the flying height separa- tion between the transparent and tiger ithomiines was previously dis- cussed in Chapter 3. Papageorgis' (1974) study was primarily addressed to the apparent contradiction posed by the presence of several mimicry complexes at a single locality. Classical Mullerian theory predicts that all protected species (and thus Batesian mimics as well) should converge on a single warning pattern. Papageorgis observed that the light conditions within the forest provide a varying background against which a butterfly will be seen by a predator. She concluded that the different colorations of the mimicry complexes in the different strata of the forest perform a dual purpose — warning coloration when the but- terfly is perched and crypsis during flight. The cryptic effect of the clear-wing ithomiine pattern during flight was discussed in Chapter 3, but Papageorgis1 arguments are less convincing for the Tiger Complex. Since ithomiines participate in only two mimicry complexes at Limoncocha, the remainder of this chapter will be restricted to the Transparent Complex and the Tiger Complex, both of which may be further subdivided into subcomplexes. ..... Transparent Mimetic Complex There are 49 species and subspecies of Lepidoptera in the Limoncocha Transparent Complex (Figures 58-61), which comprises five families of butterflies and three families of moths. By comparison Papageorgis (1974) J3 CO 1 o ft co CU O O o a. B o CJ rH j3 cu 3 43 co C ft a) •h > H tH I 00 a) co co C 01 cfl B s-s •o cu 13 43 cfl -u 4-1 O s a. a 6 o O -H o -u 43 CO 3 rl co 3 00 CU iH 3 HH cr C co o ft o o a) O H 01 . >h QJ N •H .. CO M Ol 01 rH 4H a -h 0 rH o u cu u O CO CO s o S 4J W CO CO •W M 3 C -U rH 01 CO rH U 3 iH CO rH ftrH 0) CO iH 43 C 4J CO rH M rH IM h <; o oo cu rl 3 60 8 rH a. s o CJ 43 3 w a •H H I CU 60 c CO M O 2 o 43 3 CO 0) 3 cr co o o rH rH 0) O) o M «J a) CO ■H 43 & r) O i CO W < CO M pq ft CO •H c o x> o co >i u i cj| cu rl 0) rH o a. I1 CO s O cO K pq fn 0+ w • Q P.HH CO pi 43 W 3 M CO ft X) 01 rJ 3 01 CO 43 CO 4-J 0+ rl cu TJ rH 0) /-N ft w rH £ u o 33 H H tO •H CO M r-^ 0) o < o •H r! 60 •H c 01 i o 42 W 0 , PC CO s CO CO -H 43 U O rl CU /-N HH W m ■< ft Q O M ft w H Pi CO H 0) b < •u Z CO 43 M 43 CJ M CO s CO C3 o X) 0) 33 CO CO H a n H CO ■ — * u CO CU c CO 60 rl 3 rH CO >i 33 Of OS s— » o w o <; ■>* o rH M ■*" Pm o co o 3 l-H Q) ctj CO W Ph 60 . w CO cu u CO rQ to -a cfl o CO "D g S3 X H o M Cfl g C o OJ K cO H 33 rH o+ a X o CO Cm QJ a CO O) •H oo CO o c 0) cu ft CO cO r»« S CO o 3 CO CJ to 0) Pi s o H ai oo] 3 OJ J3 CO •H 43 ft rJ O ■b o >. 43 a a CO CO /-> •H C pq XI cu •H 00 CO dl I 3 CU 4= 4J «4-l o c o •H V • cd X n al 3 rH t>0 (X -H g M-l o 3 o O 43 o 3 w 01 43 0) H 4-1 •H J§ CU N 1-1 • • CO X 3 CU iH M-l D.-H e rH o u CO • U CO CJ CO 3 •H o 4-1 CO -H CU 3 4J 6 O CO •H •H H s 44 4-1 CO CO 4-1 U 3 s 4J rH 0) CO rH h 3 T-i CO rH CXrH CU CO •h 4= s 4-1 CO rH u rH v»_t H < O • CM CU u 3 00 <0 CO CO CU CU co m •H >*-"' CO Cfl cu o rH u CU CU .3 4-1 P. <1J CU ffi 3 o+ CO w cu < »-> 55 CO M ~ M B cfl 3 •H i-l cu CM to B <-i CO 42 4J ■3 p. © 44 CO }j d) H M W (1) O J ao ^ s-' P-i W S^' M CO cu 4J X) CO rH ^N 3 •H a) w rH ^1 Fn <; iH 3 . 2: ■H 00 oi m 00 cfl HI •H v8 4-1 3 o •H CJ 33 3 Vj ^ H •H CU M }-i H 3 N— ' 0) O 4-1 rH CO O •H V4 3 00 3 ■bw M M 3 rH M » 55 /-^ M Cfl CO M •h in s moo 3 m 33 rH w H CO 3 > 3 /-v 3.f~ co o o 43 (14004 o 3 rH •^ H ^ rH v-' H S E rH 0) o M o M CO 3 ^«/ \mS •H e 44 Q O rH 3 d) CO 4C a ■b CJ ^-N ^-N 3 n w a) o -' 3 . 1 S 3 >, H 307 \ CO CO T3 3 o a. co CD u u o o 5 o rH •H 00 CO CD e Cfl 3 X a) a) .fi rH 4-1 ex g <4-l o o u rQ c 3 o to •H 4-1 OO CO g 1-1 •H 3 1 00 •H M <4H cO s 0) o i-l CJ C_> CD S X o H rH rH H a) CD M 1 rH M 3 CO S >< «o a ss CO o 1-1 cu T3 •H y-\ CX ■b w 0) 42 O CO c 3 0) 0) CO o a M •H Q IS H H S I 4= CJ co W 2: s CD cO a h a H CO •H CO U 4-1 >> 4-> T3 CD O U CJ a to CO c •H c o rH CO CO ■b X. y^S u M CO < 3 S3 a> M cO M a g ^-^ O iH rH cfl W •iH \-^ oo c CO 3 •H rH rH a -H a 4-1 S rO 3 CO o+ O CO W CO •H CO U 4-1 >s 4J T3 CD O IH Cfl CO e CO 52 o a H CO CO rH CO 3 CO rH ^-n CO O ^N u •-N •H W cO CO w CO ow V4 O «aj rJ 4J <; (X . 0) /^TJ <: r4 53 3 OS M a T3 «« S -h o ex a H 4J o a CO U •H >.^^ i-l o CO 00 3 >> 3 Q CD CJ U •H TJ fa 309 .3 o 4-1 CO ■a o (X oa cu u u o O a <— V & a) o CO <-\ tH 3 .H 0) n •H ^1 -H QJ e U c CO 4J cu tH CO > I-) •H «0 ^w* CO C 3 CO CO 4J 3 a) M SJ B •H § 1 tH . 0) aJ .u tH P. 14-1 e o o O c .fi o 3 -H CO 4-1 a) 00 ^ C 3 •H 00 S -H 1 M-l M C CO O a> o tH C_> 0) rC CU H oo M CO • -j a) N •H • • CO o cd iH U HI • co CO u co 3 in s-> O CO C M-l w •H O c <4» <: 4-1 CO -H o z ai C u o C M So cd CO M CO 6 S S 4J 4J •H -6 5 CO CO -3 o ss ■U M 9 •H cj H C 4J tH U M 0) CO tH >* 01 ^— i H 3 -H J= j= CO iH H •U D-rH O) CO CO -H J2 0 C 4J 3 CO iH co U H H-l a H <: O CO B CD O w o+ Q M M 0) M i 53 cO H ^ 00 U CO 01 -a -o -H c ^-s o r-t •— s o O w -H o < -H § o h-l co Pi 4= c w o •H Pm CO CO ^-^ CO S >, CO Q CO 4-4 s CU > •H Ml CM O a o* 2 < o Q <■ M •H Ch w o O MH C Pi •H W PC o o 1-f -^v X) M •H 0) CO T3 y^ ex r-t w CU < CO fe z; -H M .-1 Pi M s O ^J o O U i 4J •w M c ^— ' CO o X c o J3 tH cO a> H CO tH ^ 1-4 ^-s o 00 W - - CO 0) Q -H ^3 M ^ C « CH a) w V4 CO M O O PU E Pi *-" CO •H CO Q 4J cO U 00 •H C a> T3 V 3 oo 311 312 found 33 species in the Transparent Complex at Rio Llullapichis, the richest of her three Peruvian study sites. At least five relatively discrete subcomplexes occur at Limoncocha, including one (Orange-Tip Subcomplex) that is endemic to the region. The other four are similar in composition to those in Peru (at Rio Llullapichis) illustrated by Papageorgis (1975). Transparency in the wings of butterflies and other Lepidoptera has been achieved in various ways, involving both structural and pigmental modifications of the wing scales (Poulton, 1898; Kaye, 1905; Punnett, 1915). The scales of ithomiines are basically of two types, one long and narrow and the other broad and short, placed alternately on the wing membrane. In the transparent areas of the wings of most ithomiines, the long narrow scales have become no more than tiny hairs standing up away from the wing membrane. Likewise, the broader scales are greatly reduced in size, and altered from a fan shape to the appearance of the letter Y (Kaye, 1914). The following descriptions should be read while consulting the appropriate figures. Yellow Opaque Subcomplex (Figure 58). The seven species in this subcomplex have light yellow wings outlined with black, and a strong black forewing bar. The name of this subcomplex may be slightly mis- leading, since its ithomiine members have translucent and even partially transparent wings. The diffuse yellow pigment of the wings, however, produces an appearance of opacity, and indeed, the Bates ian members of the subcomplex are truly opaque. Two species, Scada quotidiana and Dismorphia theugenis are exceedingly rare (only one specimen of each taken at Limoncocha) and probably occur only occasionally at Limoncocha. 313 Scada quotidiana is more common south and west of Limoncocha (e.g., Tiwaeno; see map, Figure 1). Both Scada quotidiana and S^ batesi are dimorphic for the presence or absence of the hindwing crossbar (shown on one of the specimens of j>. batesi in Figure 58). The hindwing bar is always present in a third species, S^. ethica, that occurs along the foothills of the Andes in Ecuador (Fox, 1967). Scada batesi, extending from central Colombia to southern Peru, enjoys the greatest geographical range of the three species, and one would predict that the hindwing bar would be more common in the zone of. sympatry with S^. ethica than elsewhere, as this would enhance the effect of the Mullerian mimicry. Since Limoncocha occurs at the northern edge of this zone of sympatry, it may be of interest to record the frequency of the hindwing bar in this area. The proportion of individuals with the hindwing bar in the all-male samples collected on Eupatorium I flowers at Study Site 1 and 2 was 9.18% (N = 425). I predict that the frequency of the hindwing bar should decrease northward from Limoncocha (where S. ethica does not occur), and increase southward from Limoncocha to just beyond the ft Peruvian border, which marks the extent of the range of the monomorphic S^. ethica. Orange-Tip Subcomplex (Figure 58) . Geographically, this is perhaps the most restricted of the South American mimicry subcomplexes in which ithomiines participate. Found along the Rio Napo and Rio Putamayo drainages of east Ecuador, southern Colombia, and northeastern Peru into the western Amazonas region of Brazil, this subcomplex includes a number of other species not represented at Limoncocha (Haensch, 1909). As the name implies, the light-colored patch near the tip of the forewing is 314 orange, of a deep fulvous hue. The transparent wings are outlined with a dark brown band and the wing veins, even of the two of the Batesian mimics (Dismorphia erythroe and Ithomeis corena) , are dark brown. White Subcomplex (Figure 59) . In terms of the number of species involved (17), this is the most extensive of the transparent subcom- plexes at Limoncocha, and could actually be subdivided further, K For example, the four ithomiines in the righthand column of Figure 59 differ from the others in having larger transparent areas of the wing. They are, in fact, very similar to the Orange-Tip subcomplex except that, instead of a patch of orange near the forewing tip, they have a streak of white scaling. The remaining species in this complex (three lefthand columns of Figure 59) have broader wing borders, often with a trace of orange that may be slightly expanded to form a very small patch near the forewing tip. In addition, the transparent area of the forewing is broken up into a number of small hyaline patches, and the hindwing is suffused with white scaling that obscures most of the veins. Like the Yellow Opaque and Orange-Tip Subcomplexes, the numbers of the White Subcomplex are generally found within one meter of the forest floor. Yellow Clear-wing Subcomplex (Figure 60) . The transparent wings of the ten ithomiine members of this subcomplex are lightly tinted with yellow scaling and are bordered with dark brown. All species have a partial or complete forewing bar and two, Callithomia epidero and Dygoris dircenna (not pictured), have a hindwing bar across the end of the discal cell. As mentioned earlier, Godyris zavaleta is sexually dimorphic, the female with dark scaling at the bases of the wings and brighter yellow scaling in the transparent areas. The only non-ithomiine 315 member of the subcomplex, an unidentified Phyciodes, was very rare (only one specimen taken). The Yellow Clear-wing Subcomplex, occupying a slightly higher flight stratum than the first three subcomplexes, « generally flies one to two meters above the forest floor. Large Clear-wing Subcomplex (Figure 61) . The large size and bold black wing borders of these species make this subcomplex distinctive in appearance. The hyaline areas of the wing are tinted a lignt yellow, and are crossed by two or three bars on the forewing and one on the hindwing. The close pattern mimicry exhibited within this subgroup even extends to the enlarged yellow club at the tip of the antennae. Three of the species in this group are exceedingly rare at Limoncocha. Only one individual each of Dismorphia orise and I tuna lamirus, and two of Castnia linus, were collected during the year-long study. By comparison, the two pericopid moths and the two ithomiines were relatively common. The flight level occupied by this sub-complex (1.5 to 3m above the ground) is above those of the other groups in the Transparent Complex. Tiger Mimetic Complex There are 42 species and subspecies of Lepidoptera in the Limoncocha Tiger Complex (Figures 62-65) , which comprises four families of butter- flies and two families of moths. At the Rio Llullapichis site studied by Papageorgis (1974), the Tiger Complex contained 31 species. The Tiger Complex does not subdivide as neatly as does the Transparent Complex, although four overlapping patterns may be distinguished at Limoncocha on the basis of flight level and the distribution of the , yellow and black pigment . to X ■u o •M CO XI 3 o a. co 01 U u o a g o .-h cu JO • K C d) cu rH > D. -H e 60 o > M 0) o .3 •u 4J CO h >4-l 0) o X) c c P3 o •H A! 4J U «0 to u 1-1 3 pq 60 •H x> <4H c c n) o CJ 0) 60 CU c .3 to H M O • A 01 s N o ■H rH CO 4 4-1 X CU • 0) M CO a) a o ■H ■b /*S rH /- — sa cd CO q U 3 M cu 2: CO CO M •w a a 4-1 c o JC •H n u rJ pej cfl %— t N_^ rH <0 4-> CO o 0) o Cm1 CO *w^ •H c tfl -H rH 4-J •H CO CJ *-l a) a) o c_> p. co u , CO PC ac a H 4-1 CO M O -H *w* a > PC 3 0) CM rH CO s a. o 04- g co •H CO 5 c 4J •H c_> o •H CO r4 1 CJ 4J 4J 4J co w •H CO CO CO c — ' t-I rH U-l CJ M h <: o CU 2 3 cu • CN V0 CU M 3 60 CO 4-> 0) rH O ^^ CO w rd «i o CH-J5 3* M CO ai m >>2 CO CU C CO O cu M 60 N-/ o a) a CO S5 to .-> CO 3 CO O ^-s CU CO w 3 4J CU W g 2 o PC H CO PC M s^ 4-1 CO cfl CU 3 43 cfl r4 4-1 O 3 h 3 ■n 3 cu CO 3 O . 3 rH OH O CO cu •H CO M PC ►i v-' JC 4.) tfl o 3 PV r4 ^ CU PC rO SI CO 3 tO 3 •H e rH 3 o CO 3 cu CO w , r4 4C 3 4-1 4J o tO a co PC 2 O PC H cu o JC u u •H X S B o JC "bw < z • M a m CO 2 43 o 3 PC CO H N w cu •H C>0 co a> eg .c 4J MH • O aj c rH o O. iH B 4J o td cj u J3 3 3 (30 co •H M-l ^ c o. O o o s cd a) u .3 H u td « • 1 a) U N o •H rH to rH 0) a) t« 4-1 •H rH • • i CU • Vj to iH td C P. o S CO tH o 3 4J o O td •H IH o 4J 4-1 •H td co 4J U 3 ID 4J rH | CO rH 3 «H a rH ■H Q) u ■H .3 0) 4J 60 rH H . CO .3 3 4-1 •H o ,£> D- CU >> o rc B < M M s o M H .3 a. o d a) e c o CO ,"-\ 4J M & 2! 0J M K M s o S3 H rd to 0) 3 td rH s •H •H X! ^H a cu o S 3 CU B CO a) •a -H o c 0) CO CO CO E 0) T) H CU fa n3 U CO •H co 4J CU •H -a 3 •H td o X. 3 u CU 0J CO S CO CU B i M l-l § O u H cd ■u cd 4-t J-J o ■H > ^s •~N •H CM W .O O j vi- M ed rH Ph (J) »»^ O a u rl M CO CU « •H AS H ft rH P-i o cd •*— * o & •H rl CU Ph -b 3 •H > td rH 4-1 td •~\ CO CO w CU < 60 ^ !s cu M .- 1- 1 td td S a) B O cd •o m 3 O H •rl O M rH 0) td s 4-1 CO CU 60 0) to CU •a •H O a cu CO CO CO 3 CU 4-1 B td o CO 3 •H rH 4-J rH •H td 3 J3 cd ,s M o td cu CU S 3 w < 13 m H T3 "b /~x cd co •H cu <-i T) cd •iH 3 O CO ■H •H C O o rQ c 3 o O CJ td cd •H •H C 3 o 4J c CO o cd o o w <: o T3 M 3 M td f3 JH H 4J CO CO m cu m ■ri X X co M 2 4-1 o •H w o •H m n H SB 3 3 H R CO H cd cu M JS 3 M ^ cd %** 4J a *w^ o N O )j 01 cd ft cu S E >> E 33 cd S 319 3 o •u to ■a o 0) O* o. c to o M 0) •H cu u •C i-H ^ u CU U W o B 3 -i c i o (11 CO CO > -H cu •H « ' a bO CO ai § to 3 01 .3 4J • M-l s ° rH 3 cu V 5 a. o 3 l e >h o > 4 o w •H 0 J o to .3 r- H ^-n x> m CX 4. J W 3 3 g : 3 >M-I to 0 3 Pi a c ■H f : w o o J2 n J M 3 O a c X PM to m a ) ^ c_> cu o V ( J3 B c 4-> H to o ) o •h a p. P X to • I a> & N O -H iH CO rH cu cu Jw H-l •H tH • • X 0) • ai M co O H It C o. o a S CO -H •H 5 c *j tO > U O CO •H iH •H >-. •c CO ^^N U -U 4J o co w •H (0 CO *-» H 3 1 ^3 *5 a) u h tO M B OH •H 3 -H XI 3 M to S S >H <0 B O tH 0) CU •a sa ^i -H J3 h O H CD 4J o O M M.H XI ^/ •H iH «W 4-1 •H CO CO H •H . a «* CU SO .3 CU M 3 00 o 01 s o CU CO XI CU .H •-v u CU w QJ pn •< o Q OS M CO <: cu u og £j o U Q o N-/ >> w -J cu rH CO CM o 3 O /-x CO M w a ■d z CU M S o H nj 33 CU (0 CO C! •H •H 3 r-t o (U cu g cO B c •H > rH CO CO rH co o U tfl ^-N -H w 3 c <; cr tO 2 0) B M x> i-i CO o s u o o ■U w pc w H cu CO M J3 0) N—' u XI o •H f"H o rH O CJ •H 3 CT cu 321 4J co x: CO *3 3 o ft co 01 u u o o & o iH (1) ,0 e CO > •H 00 co 0 e § 0) J3 4-1 IH O • CO o tH -H ft 4J 6 cd O V4 cj 3 ^3 60 3 -H CO M-l 3 a: o o o co tH CO (Q A H ■o c cd • CO CO N 60 -rl a co a) H CO O M-t •H •H • • CO Vi • CO i-H CO 3 p. o 6 CO •h 5 3 4-1 o o cO •H K CJ 4J 4J •H CO CO 4J U 3 CO 4J r-t g CO .H •3 3 T-l S H .H QJ U -H J= 0) 4-1 60 .H c o CO /-n 4-> W i-i H V< H v_- * >% CO ,3 CO 4-1 S-l a u o •^^ o iH o o >^ u a 3 O a \0 CO u 3 60 323 324 Yellow, Orange, and Black Understory Subcomplex (Figure 62). The members of this complex, which includes most of the smaller tiger species, are found primarily in the forest understory, usually from « about 1.5 or 2 m to 7 m. By contrast, the members of the remaining three tiger subcomplexes are rarely found below 4 or 5 m and usually occur in the mid to upper canopy. With sixteen species, this is the largest of the four tiger complexes but also the least consistent in pattern. In general, the hindwing and base of the forewing are orange with black spots or lines, while the distal half of the forewing is black with yellow markings. In Figure 62, those species in which the hindwing black bar tends toward enlargement to form a black patch represent transitional forms that also have characteristics of the Yellow-Bar Canopy Subcomplex (Figure 63) . These transitional forms include the closely matched species, Napeogenes duessa and N. apobsoleta (center of Figure 62) and the Chlosyne Batesian mimic (lower right of Figure 62). Two other species included in Figure 62, Hyposcada kezia and Stalactis calliope, are similar to each other (both 4. lack any yellow markings and have white spots in the black forewing tips) but do not closely match the other members of the subcomplex. Interestingly, H. kezia is consistently found at heights of 1 to 2 m, considerably below most of the other members of the subcomplex, and- the three specimens of j^. calliope collected during the study were also taken at this height. Three members of this subcomplex, Mechanitis isthmia, M. lysimnia, and Ceratinia poecila are regular visitors to forest edges and clearings but the remainder of the species rarely leave the forest interior. 325 Yellow-Bar Canopy Subcomplex (Figure 63) . This subcomplex is characterized by a large black patch on the hindwing and a strong yellow median bar on the forewing. The forewing tip is black and the lighter areas of the hindwing and the base of the forewing are orange. Once again, a few species are transitional. For example, Melinaea egesta has an extra yellow bar at the tip of the forewing making it look similar to the Yellow-Spot Subcomplex (Figure 64) hut has the black hindwing patch characteristic of the Yellow-Bar Subcomplex. The rather tattered Castnia cononia was the only specimen taken of this species. Yellow-Spot Canopy Subcomplex (Figure 64). The members of this subcomplex are similar in size and flying range (upper canopy) to those of the Yellow-Bar Subcomplex, but lack the large black disc on the hindwing, and have broken or spotted bars in the distal half of the forewing instead of a single intact yellow bar. Dismorphia amphione is a member of this group and illustrates the incredible ability of this pierid genus of Batesian mimics to mimic not only the divergent color patterns of different mimetic complexes (see also*' Dismorphia species in Figures 58, 59, 60, and 61) but also the flight characteristics of its models. Orange and Black Subcomplex (Figure 65) . As the name implies, there is no yellow coloring in this group, only orange and black, Otherwise, it is quite similar to the Yellow-Bar Subcomplex, since the groups share the large black hindwing disc and the black forewing tip. Two butterflies included in Figure 65, Lycorea pasinuntia and Athyrtis mechanitis are probably not residents at Limoncocha, as their lack of 326 conformity to any of the tiger subcomplexes suggests. Both specimens illustrated are the only individuals of the two species collected at Limoncocha, and probably represent strays from outlying populations to the southwest (see Chapter 1*) . Like the Yellow-Bar and Yellow- Spot Subcomplexes, the Orange and Black Subcomplex inhabits the upper canopy, although the Mechanitis species are frequently encountered in the under story. Polymorphic Mimetic Species With the exception of the sexual dimorphism of Godyris zavaleta (Figure 60) and the hindwing bar polymorphism of Scada batesi (Figure 58) , members of the Transparent Complex exhibit very few clear-cut pattern polymorphisms. The Tiger Complex, however, has several, involving both Batesian and Mullerian mimic species. The familiar case of sex-limited mimetic polymorphism of a Batesian mimic is exhibited by the pierid, Perrhybris pyrrha, in which the female (Figure 63) is mimetic but the male (white with black markings) is not. Minor pattern differences occur between the sexes (both of which are mimetic) of Phyciodes pelonia (Figure 62) and of Dismorphia amphione (Figure 64) . Two subspecif ic forms of Anaea fabius, a Batesian mimic in the Nymphalidae, occur at Limoncocha. These two forms clearly illustrate the pattern differences described earlier that exist between the Yellow-Bar Canopy Subcomplex (subspecies quadridenta- tus, Figure 63) and the Yellow-Spot Canopy Subcomplex (subspecies tithoreides, Figure 64). 327 Among Mullerian mimics, the most dramatic polymorphism is that of the three forms of Heliconius numata, each of which closely mimics a different species of Melinaea (H. n, near euphrasius mimics M. menophilus, « Figure 63; H. n. near superioris mimics M. maeonis, Figure 64; and H. n. aristione mimics M. mothone, Figure 65) . Similar mimetic pairs of Melinaea species and II. numata mimics in Brazil have been described by Brown and Benson (1974) , and their theory of the adaptive polymorphism of Heliconius numata will be discussed in the last section of this chapter. Among tiger ithomiines, there are two polymorphic species whose forms participate in different mimetic subcomplexes. Both Mechanitis mazaeus and M. messenoides have an all orange and black form (Figure 65) and a form with a yellow median bar on the f orewing (Figure 63) . Mullerian Mimicry theory, of course, predicts that pattern-stabilizing selection should keep all Mullerian species monomorphic and converged on a single ~ warning pattern. Papageorgis' (1974) theory as to why several mimetic complexes exist at a single locality has already been discussed, but the cases just described refer to polymorphic species participating in different subcomplexes of the same (Tiger) mimetic complex. Why do these species not conform to theory at Limoncocha? The answer apparently lies in the geographic distribution of mimetic patterns, and the relationship of this distribution to the Quaternary tropical forest refugia mentioned earlier, a subject of growing interest (Brown et al., 1974; Brown, 1976a). Anaea fabius (Batesian mimic), Mechanitis mazaeus, M. messenoides, and Heliconius numata (Mullerian mimics) are all widespread species with numerous subspecies, each with 328 a different mimetic pattern depending of the geographic location and the historical events that occurred at the sites of the nearest neo- tropical refugia. Limoncocha appears to lie at a confluence of adjacent mimetic patterns. Along the eastern foothills of the Andes in Ecuador, Peru, and Bolivia, the Orange and Black Subcomplex is well developed, but in upper Amazonia (between Manaus and Iquitos) tiger patterns with much yellow mottling of the f orewing occur (Moulton, 1909) . Thus in the western Amazon basin there is an increasing representation of orange and black to the west and south and an increasing representation of yellow and loss of black to the north and east (Moulton, 1909; Papa- georgis, 1975). The most recent map of proposed Quaternary Refugia for butterflies (Brown, 1976a) places Limoncocha near the western edge of the large Napo Refuge, but under the influence of the Abitagua Refuge to the west and the Loreto Refuge to the southeast (in northwest Peru). Thus, many of the polymorphic forms present at Limoncocha may actually result from the overlapping of parapatric subspecific popula- tions, e.g., the three "subspecies" of Mechanitis messenoides (Table 4; Figures 63 and 65) , and the three "subspecies" of Heliconius numata (Figures 63, 64, and 65) present at Limoncocha. The fact that the forms of these two species appear to segregate in the Limoncocha populations with few intermediates suggests that the patterns of these subspecies are under the genetic control of a single gene (supergene) . By contrast, the two forms of Mechanitis mazaeus shown in Figures 63 and 65 represent the ends of a continuum of pattern variation that occurs in the Limoncocha populations of this species. 329 Ithomiine Participation in Limoncocha Mimicry Complexes Table 24 summarizes the taxonomic representation of the 53 ithomiine species in the 9 mimetic subcomplexes at Limoncocha. The species are equally divided between the Transparent and Tiger Complexes, but, within the former, most species and genera occur in either the White or Yellow Clear-wing Subcomplexes. In the Tiger Complex, the great majority of the genera and species are found in the Yellow, Orange, and Black Understory Subcomplex and the Yellow-Bar Canopy Subcomplex. While species in a particular genus may participate in as many as three different mimetic subcomplexes, all of these tend to be in one or the other of the two main complexes. Indeed, only 2 genera, Napeogenes and Callithomia, of the 25 genera at Limoncocha have species occurring in both the Transparent and Tiger Complexes. Of the remaining 23 genera, 8 are exclusively transparent and 15 are exclusively tiger in pattern. Half of the eight tribes include both transparent and tiger species. Two tribes, Ithomiini and Godyridini, are exclusively transparent, and two others, Tithoreini and Melinaeini, are exclu- sively tiger in pattern. It must be remembered that these comparisons and restrictions are applicable only at Limoncocha, and may not reflect the true extent of the participation of these genera and tribes in mimicry complexes throughout South America. For example, there are several clear-wing species in the Tribe Tithoreini (Fox, 1956) . 330 Table 24. Ithomiine Part icipation in Limoncocha Mimicry Complexes • NUMBER OF SPECIES OR FORMS PER MIMETIC SUBCOMPLEX # OF * TRANSPARENT COMPLEX TIGER COMPLEX SUBCOM- PLEXES/ GENUS YO OT W YC LC TOTAL YOB YBC YSC OB TOTAL GENUS Athyrtis (1) Tithorea (1) 1 1 1 1 1 1 Melinaea (A) 2 1 1 A 3 Forbestra (2) ^ Mechanitis (5) Scada (2) Xanthocleis (1) 2 1 1 1 1 2 • 1 3 3 2 8 2 3 1 1 Hypothyris (5) Napeogenes (5) 1 1 2 3 3 2 5 3 2 3 Ithomia (3) 1 2 3 2 Aeria (1) Hyposcada (1) Oleria (6) 1 6 1 6 1 1 1 1 1 Ceratinia (1) Callithomia (2) Dircenna (1) Pteronymia (1) X (1) 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 Dygoris (1) Godyris (1) Heterosais (1) Hypoleria (3) Mcclungia (1) Pseudoscada (2) Thyridia (1) 2 1 1 1 1 1 1 1 1 1 1 1 3 1 2 1 V 1 1 1 2 1 2 1 Totals (genera) 2 3 5 9 2 17 7 3 3 3 10 - Totals (species) 3 A 10 10 2 28 12 7 3 5 27 Note: YO: OT: W: YC: LC: Yellow Orange White Yellow Large Opaque -Tip Clear -wing Clear-wing YOB: YBC: YSC: OB: Yellow, Orange, and Black Under story Yellow-Bar Canopy Yellow-Spot Canopy Orange and Black Genera are grouped by tribe, in the same order as in Table A. Numbers in parentheses indicate the number of species present at Limoncocha in each genus. 331 Evidence for Unpalatabllity In the Ithomiinae Although members of the Ithomiinae have long been considered distasteful (Bates, 1862; Muller, 1879; Kaye, 1914), the evidence for their unpalatability is almost entirely circumstantial. Kaye (1914) reviewed some of this circumstantial evidence and concluded that the sum total of proof by inference was large enough to warrant considering the Ithomiinae as a protected group. He noted further, however, that the degree of distastefulness probably varies greatly among different genera. Circumstantial Evidence Abundance. Bates (1862) and numerous later observers (e.g., Collenette and Talbot, 1928) have reported that the Ithomiinae are nearly always exceedingly numerous and occur where insectivorous birds abound. This is certainly true in the forest understory at Limoncocha, where the Ithomiinae are very abundant in comparison with other butter- fly groups, and insectivorous birds are common (Pearson, 1972). Slow flight. Part of the seeming abundance of ithomiines may be more apparent than real, owing to their slow, conspicuous flight. It is well known that butterflies frequenting open areas (e.g., Pieridae, Nymphalidae) often have strong and rapid flight, which is apparently their only means of defense against birds if attacked in flight. Because of the much slower flight of ithomiines (and other protected groups, such as the Heliconiinae, Acraeinae, and Danainae) , it has been assumed that most rely on distastefulness as a defense (Kaye, 1914). 332 Wing marks. Kaye (1914) argued that notches taken out of the wings occurred less often in the Ithomiinae as compared to the Nymphalidae and Satyridae, suggesting that, despite their slow conspicuous flight, ithomiines are less susceptible to attack. The beak mark frequencies (beak marks are not the same as notches) presented by Collenette and Talbot (1928) were used by Carpenter (1941) to demonstrate that ithomiines are more often released by birds after capture, presumably because of the butterflies' distastefulness (Chapter 3). Tenacity of life. Longstaff (1908) characterized the Ithomiinae as being "tenacious of life," referring to their relative success (as compared to other butterfly groups) at surviving rough handling by collectors, and, presumably, by other predators. Strong tenacity of life is considered a trait of distasteful butterflies (Kaye, 1914), but the character also seems to be correlated with body size. The larger ithomiines at Limoncocha (Yellow-Bar and Yellow-Spot Canopy Subcomplexes, Orange and Black Subcomplex, and Large Clear-wing Subcomplex) are much hardier than the smaller members of the subfamily. Coloration. In addition to slow flight and tenacity of life, another character traditionally associated with distasteful butterflies is a similar coloration on. both the dorsal and ventral surfaces of the wings (Kaye, 1914), a trait held by most ithomiines. Papageorgis (1974), however, concluded that the patterns of members of the Transparent and Tiger Mimicry Complexes are not aposematic per se, but provide a warning stimulus only through their great repetition, and perform a second func- tion by conferring flight crypticity. 333 Strong scent. Although Longstaff (1908, 1912) has referred to the strong and occasionally unpleasant scent of some ithomiines, most of the species at Limoncocha emitted a faint but neutral or slightly pleasant odor. The possibility that sexual pheromones may have originated as protective odors was considered in Chapter 4. Feigning death. As Gilbert and Ehrlich (1970) noted, death feign- ing is a behavioral trait exhibited by many Ithomiinae and some dis- tasteful Danaidae (e.g., Lycorea ceres) . The death feigning of an in copula pair of Thyridia confusa was described in Chapter 4, and similar instances of individuals remaining motionless in the net after capture occurred for most species at Limoncocha, but was more common among the larger tiger -patterned species. Death-feigning upon release after marking (placing the individual on a leaf) was much less common than playing dead in the net. The shock of an encounter with another object (e.g., the wall of an insect net or the beak of a bird) while flying apparently provides the stimulus initiating this behavior. Models. Since the time of Bates, many lepidopterists have commented on the great number of species from several generally palatable families that have converged on the patterns of the Ithomiinae. This phenomenon is well illustrated in the mimicry subcomplexes at Limoncocha (Figures 58-65) in which species from the Pieridae, Nymphalidae, and Riodinidae — all presumably palatable — have converged in pattern and behavior to mimic ithomiines. It is presumed that these palatable (Batesian) mimics would only derive benefit from such a convergence if the Ithomiinae Were unpalatable and thus protected from predation. 334 Foodplants. The early generalization that the larval foodplants of ithomiines belong to the Solanaceae (Muller, 1879), interpreted against the background of the longstanding notoriety concerning the poisonous qualities of this plant family (Heiser, 1969), reinforced the belief that ithomiines are unpalatable and suggested that ithomiines derive their distastefulness from these plants during the larval stage. Despite the vast numbers of alkaloids known to occur in solanaceous plants (Chapter 1) and the sophisticated biochemical techniques available to identify these compounds, I know of only one published report of an alkaloid analysis being performed on any life stage of an ithomiine. Rothschild (1973) reported that an adult of Mechanitis polymnia polymnia (L.) was found to be lacking solanaceous alkaloids, although the larval foodplant and locality of capture of the analyzed specimens were both unknown. Chemical analyses of all life stages of ithomiine butterflies and of their larval foodplants are clearly needed. It appears then, that the circumstantial evidence presently available concerning the unpalatability of the Ithomiinae is ambiguous, but, in the aggregate, tends to support Kaye's (1914) conclusion that ithomiines are a protected group. Direct Evidence Feeding experiments. Although often cited as direct evidence for the unpalatability of the Ithomiinae to vertebrate predators, Brower and Brower's (1964) study in Trinidad tested the palatability of only one species of ithomiine, the clear-wing It horn ia drymo pellucida, on 335 only one species of bird, the North American Blue Jay (Cyanocitta cristata bromia Oberholser) , imported from Massachusetts. As Fox (1967) has noted, these experiments tell us nothing about the feeding habits of the potential avian predators of ithomiines nor of the rela- tive palatabilities of ithomiines belonging to different mimetic com- plexes and subcomplexes. These experiments do, however, strongly suggest that at least some ithomiines may be unpalatable to some birds at some times. To determine the extent of this unpalatability among taxonomic and mimetic groups of ithomiines will require carefully controlled feeding experiments involving natural potential predators. If performed in conjunction with the chemical experiments described above,: such experiments should lead to a clearer understanding of the various roles played by ithomiines in the Batesian-Mullerian milieu. . ^ Some Mimetic Consequences of Ithomiine Ecology The diversity and equitability values calculated on the Site 4 ithomiine samples (Chapter 3) revealed that the ithomiine community at Limoncocha appears to be relatively stable in character through time, and only occasionally depressed in abundance by adverse weather conditions. During the periods of low abundance, however, the similarity in species composition between successive samples (as measured by the Quotient of Similarity) dropped, suggesting that, if ithomiines are indeed variable in palatability (see above), different assemblages of : individuals during these periods may provide naive predators with different learning experiences regarding the palatability of the color 336 pattern, depending on the species composition of the assemblage (Chapter 3) . The question then arises as to the stability of the relative abundances of the mimicry complexes and subcomplexes at Limoncocha. To determine if the relative abundances of the various mimetic subcomplexes vary during the year, the Site 4 data were re-analyzed on the basis of color pattern. The results are given in Figures 66- through 68. In Figure 66 the number of ithomiine individuals present in each. of the subcomplexes of the Transparent Mimetic Complex on the 26 sample days at Site 4 is graphed (mimetic individuals of other lepidopteran families are indicated separately) . Likewise, the number of ithomiine individuals present in each of the subcomplexes of the Tiger Complex is given in Figure 67 . These data are summarized in Figure 68, which presents the total number of individuals and species in the Transparent and Tiger Complexes present at Site 4 during the sampling period. These figures show that all nine subgroups of both mimetic com- plexes varied in approximately the same manner throughout the year, paralleling the variation in total ithomiine abundance as given in Figure 27. The one possible exception to this pattern is the slight increase in abundance of the Yellow Opaque Subcomplex in early April (Figure 66a). From the point of view of a predator, then, the relative frequencies of the various mimetic complexes vary but little, even though the variation in overall ithomiine abundance may be substantial. Brown and Benson (1974) have shown that the ithomiine pockets in southeastern Brazil are spatially and temporally unstable, and that the Figure 66. Abundances of Transparent Mimetic Subcomplexes at Study Site 4. a. Yellow Opaque Subcomplex b. Orange-Tip Subcomplex c. White Subcomplex d. Yellow Clear-wing Subcomplex e. Large Clear-wing Subcomplex . 338 MAR APR MAY JUN JUL H k 1 i—i 1 1 SEP 1 OCT I NOV 1 Figure 67. Abundances of Tiger Mimetic Subcomplexes at Study Site 4, a. Black, Yellow, and Orange Under story Subcomplex b. Yellow-Bar Canopy Subcomplex c. Yellow-Spot Canopy Subcomplex. d. Orange and Black Subcomplex 340 30 30 3 20i § M > 10 - i£^ O £ 30~ H 20- • 10 A&- 30 20 10 i*U* 1 t .I K^l> H 1 — I — h — I — » » JUN JUL NOV I MAR APR MAY Figure 68. Summary of the Abundances of the Transparent and Tiger Mimetic Complexes at Study Site 4. a. Transparent Complex • — • individuals i species b. Tiger Complex • — • individuals species 342 343 great differences in the relative abundances of the species in these pockets over time result in corresponding changes in the relative abundances of the mimetic color patterns as well. Polymorphic populations of Heliconius numata occur in the same areas, although this species is much less abundant than its demographically unstable ithomiine co- mimics. In an attempt to explain the polymorphism of Heliconius numata, these authors proposed that this distasteful Mullerian mimic is acting numerically in a pseudo-Batesian fashion. In other words, the various H. numata morphs are maintained in the population because the changing character of the dominant mimetic pattern (of the more abundant ithomiine co-mimics) favors different H_. numata morphs at different times of the year. Such an explanation for the presence of the three morphs of Heliconius numata at Limoncocha appears unlikely, however, since the relative abundances, of the three mimetic subcomplexes in which they participate appear to be fairly stable through time (Figure 67b, c, and d) . (Unfortunately, the sample sizes for these three mimetic subcomplexes are small because the members of these subcomplexes spend most of their time in the mid to upper canopy, while the samples themselves were collected only in the lower four meters of the understory.) It seems more probable that the polymorphism of H. numata at Limoncocha is T:he result of a confluence of different mimetic patterns, each one favored in an adjacent geographical area. CHAPTER VIII t « CONCLUSIONS Butterflies of the subfamily Ithomiinae are among the most con- spicuous and abundant flying insects of the lowland tropical forests of eastern Ecuador. They are sensitive to microhabitat differences within seemingly homogeneous forest, apparently responding to gradients of light and relative humidity, resulting in a clumped distribution of adults. Capture-mark-recapture experiments reveal that ithomiines appear to have much larger home ranges than previously believed and that the clumped distributions of ithomiines within the continuously forested lowlands of the upper Amazon basin do not represent discrete communities and populations, but merely transient concentrations that crest and ebb erratically through time. Ithomiines, active from dawn until dusk, are able to fly and i navigate in near darkness. Feeding activity occurs primarily in the early morning and late afternoon. The early morning visitation to white flowers by males fits the syndrome of crepuscular pollination, and adaptations of the butterflies and some of the adult foodplants suggest a coevolved relationship. Females forage within the forest understory for bits of detritus that may provide the nitrogenous compounds needed for egg production and/or longevity. Both males feeding at flowers and females feeding at detritus appear to act as 344 345 stimuli attracting other ithomiines to the food sources, behavior which may aid the exploitation of food sources patchily distributed in space and time. Reproductive activities of ithomiines occur mostly during the middle of the day. Courting males are most active in late morning, and mating activity peaks between 1200-1300 h. Oviposition occurs through- out the afternoon and may occur as late as 1800 h, but peaks between 1400-1500 h. Male ithomiines face the difficulty of locating a mate among a large number of similarly colored species and exhibit two types of mate-locating behavior. "Display perching" males hold their wings open at a species-specific angle and display the hindwing hair- pencils. Short interperch flights between perches tend to concentrate the males' aphrodisiac pheromone in a broad disc that presumably lures females into the male's sphere of vision, at which point male pursuit is initiated. The sequence of perches and interperch flights and the postures of displaying males indicate that this courtship behavior is under species-specific genetic control. "Patrol perching" males perch with their wings together and make periodic interperch flights that patrol an area of several square meters. Almost any moving body within the patrolled area will elicite pursuit behavior by the male. As in most butterflies, copulation probably lasts several hours, and it is proposed that the long length of copulation in butterflies is an adaptive behavior of the male to insure that sperm precedence, rather than sperm displacement, occurs if the female mates again soon (females are known to mate up to 6 times) . Oviposition strategies among ithomiines vary greatly with regard to egg volume, onset of oviposition, clutch size, 346 and probably in daily fecundity. Egg manufacture is financed by the adult stage, not the larva. Four modes of oviposition behavior, each defining the sequence of events after the location of the larval food- plant and before the laying of the eggs, are described. It is proposed that the switch from laying single eggs (ancestral condition) to lay- ing egg clusters was accompanied by selection favoring oviposition on the dorsal leaf surface to minimize mistakes in egg placement. Juvenile ithomiines display great variation in morphology in all life stages. Of particular importance is the discovery that the larva of Melinaea menophilus bears a strong resemblance to larvae of the primitive ithomiine genus Tithorea, and to larvae of Lycorea and other Danaidae, providing further evidence for the close relationship between the Ithomiidae and Danaidae. The generation time for most ithomiines is approximately 28 to 30 days. Larval development takes considerably longer on foodplants with well-armed leaves than on unarmed plants. Chemical differences in foodplants probably also affect the length of the larval feeding period. Larval feeding on the ventral surfaces of leaves is almost universal among ithomiines. Parasitism and predation take an extremely high toll of juvenile stages of ithomiines. The combined effects of this high juvenile mortality and the low daily reproductive rate of most adult females result in an extreme under -utilization of foodplant tissue. Although parasitism and predation of juvenile stages, and not foodplant availability, appear to regulate the size of ithomiine populations, adult mortality from severe storms and periods of heavy rainfall can severely depress the abundances of most ithomiine 347 species. In spite of these periodic depressions in abundance, however, the ithomiine community at Limoncocha (as measured by the samples collected at Site 4) appears to be quite stable in relative composition i of mimetic patterns, age distribution, species diversity (H1), and equitability (J'). This high species diversity results in part from the large number of species (53) present in the Limoncocha ithomiine taxocene. The sympatric occurrence of a large number of closely related ithomiine species at Limoncocha is probably a reflection of the large number of potential larval foodplants in the Solanaceae available in the same area. Local populations of most ithomiine species appear to be highly specific in foodplant utilization, rarely developing on more than one foodplant. The foodplants utilized by subspecific populations of widely distributed species, however, may be different, resulting in broad foodplant associ- ations at the species level. The primary exceptions to foodplant specificity within populations appear to be wide-ranging, vagile, species that develop on fugitive, or second-growth, Solanaceae. 4 The participation of ithomiines in mimicry complexes is extensive. The ithomiine species at Limoncocha are approximately equally divided between the Transparent Complex (with 5 subcomplexes) and the Tiger Complex (with 4 subcomplexes) . Most ithomiines are probably unpalat- able to some degree, but it is likely that, as a group, there exists a broad palatability spectrum among ithomiine species. There are many promising areas for future research in ithomiine ecology, and several of these have been mentioned in the text. The most interesting areas in my opinion include (1) the role of chemical 348 and mechanical defenses of larval foodplants in determining the food- plant utilization of the Ithomiinae, (2) the foodplant tolerances of larvae and ovipositing females, (3) the contributions of nectar and « detritus food sources to longevity and fecundity in ithomiines, (4) the significance of multiple matings of ithomiine females and the mechanics of sperm transfer from copulation to oviposition, (5) be- havioral studies of male courtship behavior with experiments designed to delimit the function of the male hairpencils and pheromones, (6) palatability experiments offering adult and immature ithomiines to natural potential predators to determine the nature of the palatability spectrum, and (7) chemical analyses of ithomiines and their foodplants to determine the basis for any observed ithomiine unpalatability and to define the phytochemical ecology of the Ithomiinae-Solanaceae interface. REFERENCES Allee, W. C. 1926. Measurement of environmental factors in the tropical rain forest of 'Panama. Ecology 7: 273-301. d' Almeida, R. F. 1922. Melanges lepidopterologiques. Etudes sur les lepidopteres du Bresil. Berlin. 226 pp. (Mimeo) d'Almeida, R. F. 1938. Estudo sobre tres generos de sub-familia Ithomiinae. Mem. Inst. Oswaldo Cruz. 33: 381-394. d'Almeida, R. F. 1944. Estudos biologicos sobre alguns Lepidopteros do Brasil. Arq. Zool. Est. Sao Paulo. 4: 33-72. d'Almeida, R. F. 1945. Revisao do genero Xanthocleis Boisd., 1870. Arq. Zool. Est. Sao Paulo. 4: 97-112. Arms, K. , P. Feeny, and R. C. Lederhouse. 1974. Sodium: stimulus for puddling behavior by tiger swallowtail butterflies, Papilio glaucus. Science 185: 372-374. Bailey, N. T. J. 1951. On estimating the size of mobile populations from recapture data. Biometrika 38: 293-306. Baker, H. G. , and I. Baker. 1973. Amino-acids in nectar and their evolutionary significance. Nature 241: 543-545. Barcant, M. 1970. Butterflies of Trinidad and Tobago. Collins, London. 314 pp. Barrett, C, and A. N. Burns. 1951. Butterflies of Australia and New Guinea. N. H. Seward Pty. LTD. Melbourne1. 187 pp. Bates, H. W. 1857. Proceedings of natural-history collectors in foreign countries. Zoologist 15: 5659. Bates, H. W. 1862. Contributions to an insect fauna of the Amazon Valley. Lepidoptera: Heliconidae. Trans. Linn. Soc. 23: 495-566. Bates, H. W. 1863. The Naturalist of the River Amazons. Univ. of California Press, Berkeley. 1962 reprint. 465 pp. Beebe, W. 1955. Two little-known selective insect attractants. Zoologica 40: 27-32. Benson, W. W. 1967. Studies on the ecology of Ithomiinae and Heli- coniinae in Costa Rica. Mimeo Research Report. OTS, Ciudad Universitaria, Costa Rica. Feb. -March. 349 350 Benson, W. W. 1972. Natural selection for Mullerian mimicry in Heliconius erato in Costa Rica. Science 176: 936-939. Benson, W. W. , and T. C. Emmel. 1973. Demography of gregariously roosting populations of the Nymphaline butterfly Marpesia berania in Costa Rica. Ecology 54: 326-335. Beutelspacher, C. R. 1972. Some observations on the Lepidoptera of bromeliads. J. Lep. Soc. 26: 133-137. Biezanko, C. M. 1960a. Danaidae et Ithomiidae da Zona Sueste do Rio Grande do Sul. Arquivos de Entomologia (Inst. Agron. do Sul, Escola de Agron. Eliseu Maciel, Pelotas, Rio Grande do Sul, Brasil). III. Serie A: 1-6. Biezanko, C. M. 1960b. Danaidae et Ithomiidae da Zona Missioneira do Rio Grande do Sul. Arquivos de Entomologia (Inst. Agron. do Sul, Escola de Agron. Eliseu Maciel, Pelotas, Rio Grande do Sul, Brasil). III. Serie B: 1-6. Biezanko, C. M. de, A. Ruffinelli, and C. S. Carbonell. 1957. Lepi- doptera del Uruguay; Lista anotada de especies. Revista de la Facultad de Agronomia, Montevideo, Uruguay, No. 46, 152 pp. Birch, M. C. 1974. Aphrodisiac pheromones in insects. In M. C. Birch, ed. , Pheromones . North-Holland, Amsterdam, pp. 115-134. Brower, L. P. 1963. The evolution of sex-limited mimicry in butter- flies. Proc 16th Intern. Cong. Zool. 4: 173-179. Brower L. P., and J. V. Z. Brower. 1964. Birds, butterflies, arid, plant poisons: A study in ecological chemistry. Zoologica 49: 137-159. Brown, F. M. 1968. Richard Middleton Fox. J. Lep. Soc. 22: 192-195. Brown, F. M. 1975. An annotated entomological bibliography of Romualdo Ferreira d'Almeida (1891-1969). J. Lep. Soc. 29: 40-51. Brown, F. M. , and B. Heineman. 1972. Jamaica and Its Butterflies. E. W. Classey Ltd., London. 478 pp. Brown, K. S. 1972. Maximizing daily butterfly counts. J. Lep. Soc. 26: 183-196. Brown, K. S. 1973. A portfolio of neotropical Lepidopterology. Private pamphlet printed by the author. Rio de Janeiro, Brazil. 28 pp. Brown, K. S. 1976a. Geographical patterns of evolution in Neotropical Lepidoptera. Systematics and derivation of known and new Heliconiini (Nymphalidae: Nymphalinae) . J. Entomology, Series B, 44: 201-242. 351 Brown, K. S. 1976b. The 1975 Field Season Summary, Zone 10, South America. News of the Lep. Soc, 1976, No. 2 (15 May 1976). Brown, K. S., and R. F. D'Almeida. 1970. The Ithomiinae of Brazil (Lepidoptera: Nymphalidae) . II. A new genus and species of Ithomiinae, with comments on the tribe Dircennini d 'Almeida. Trans. Amer. Ent. Soc. 96: 1-17. Brown, K. S. , and W. W. Benson. 1974. Adaptive polymorphism associated with multiple Mullerian mimicry in Heliconius numata (Lepid. Nymph.). Biotropica 6: 205-228. Brown, K. S., 0. H. H. Miekle, and H. Ebert. 1970. 0s Ithomiinae do Brasil. I. Prittwitzia g. n. para Ithomyia hymenaea Prittwitz e suas subspecies. Rev. Brasil Biol. 30: 269-273. Brown, K. S. , P. M. Sheppard, and J. R. G. Turner. 1974. Quaternary refugia in tropical America: evidence from race formation in Heliconius butterflies. Proc. Roy. Soc. Lond. (B) 187: 369-378. Bull, L. B., C. J. Culvenor, and A. T. Dick. 1968. The Pyrrolizidine Alkaloids. North-Holland, Amsterdam. 293 pp. Butler, C. G. 1967. Insect pheromones. Biol. Rev. 42: 42-87. Carpenter, G. D. H. 1941. The relative frequency of beak-marks on butterflies of different edibility to birds. Proc. Zool. Soc. Lond. (A) 111: 223-230. Chodorowski, A. 1959. Ecological differentiation of turbellarians in-Harsz-Lake. Polsk. Archwm. Hydrobiol. 6: 33-73. Collenette, C. L. and G. Talbot. 1928. Observations on the bionomics of the Lepidoptera of Mato Grosso, Brazil. Trans. Roy. Ent. Soc. Lond. 76: 391-416. Common, I. F. B. 1964. Australian Butterflies. Jacaranda Press. 131 pp. Common, I. F. B. , and D. F. Waterhouse. 1972. Butterflies of Australia. Angus and Robertson, Sydney. 498 pp. Cook, L. M. , K. Frank, and L. P. Brower. 1971. Experiments on the demography of tropical butterflies. I. Survival rate and density in two species of Parides. Biotropica 3: 17-20. Correll, D. S. 1967. Solanaceae (Juss.) Lindl. Key to Section Neolycopersicon and Section Tuberarium. Iji "Flora of Peru," Publication 1015, Bot. Ser. , Field Mus. of Nat. Hist., Vol. 13, Part V-B, No. 2: 271-458. Costa-Lima, A. da. 1936. Terceiro Catalogo dos Insectos que Vivem nas Plants do Brasil. Rio de Janeiro. 469 pp. 352 Costa-Lima, A. da. 1950. Insetos do Brasil. Tomo 6, Lepidopteros, Parte 2. Escola Nacional de Agonomia, Serie Didatica, No. 8. 420 pp. Crane, J. 1955. Iraaginal behaviour of a Trinidad butterfly, Heliconius erato hydara Hewitson, with special reference to the social use of colour. Zoologica 40: 167-196. « Cronquist, A. 1968. The Evolution and Classification of Flowering Plants. Houghton Mifflin Co., Boston. 396 pp. D'Abrera, B. 1971. Butterflies of the Australian Region. Lansdowne, Melbourne. 415 pp. D'Arcy, W. G. 1972. Solanaceae studies II: typification of subdivisions of S planum. Ann. Missouri Bot . Gard. 59: 262-278. D'Arcy, W. G. 1973 (1974). Flora of Panama. Part IX. Family 170. Solanaceae. Ann. Missouri Bot. Gard. 60: 573-780. D'Arcy, W. G. 1974. S planum and its close relatives in Florida. Ann. Missouri Bot. Gard. 61: 819-867. Deevey, E. S. 1969. Coaxing history to conduct experiments. Bioscience 19: 40-43. Downes, J. A. 1973. Lepidoptera feeding at puddle-margins, dung and carrion. J. Lep. Soc. 27: 89-99. Drummond, B._ A. , III, G. L. Bush, and T. C. Emmel. 1970. The biology and laboratory culture of Chlosyne lacinia Geyer (Nymphalidae) . J. Lep. Soc. 14: 135-142. Ebert, H. 1969. On the frequency of butterflies in Eastern Brazil, with a list of the butterfly fauna of Pocos de Caldas, Minas Gerais. J. Lep. Soc. 23, Supplement 3: 48 pp. Edgar, J. A., C. C. J. Culvenor, and T. E. Pliske. 1974. Co-evolution of danaid butterflies and their host-plants. Nature 250: 646-648. Edmunds, M. 1974. Significance of beak marks on butterfly wings. Oikos 25: 117-118. Ehrlich, P. R. 1958. The comparative morphology, phylogeny and higher classification of the butterflies (Lepidoptera: Papilionoidea) . Univ. of Kansas Sci. Bull. 39: 305-370. Ehrlich, P. R. , D. E. Breedlove, P. F. Brussard, and M. A. Sharp. 1972. Weather and the "regulation" of subalpine populations. Ecology 35: 243-247. Ehrlich, P. R. , and S. E. Davidson. 1960. Techniques for capture- recapture studies of Lepidoptera populations. J. Lep. Soc. 14: 227-229. 353 Ehrlich, P. R. , and L. E. Gilbert. 1973. Population structure and dynamics of the tropical butterfly Heliconius ethilla. Biotropica 5: 69-82. Ehrlich, P. R. , and P. H. Raven. 1964. Butterflies and plants: a study in coevolution. Evolution 18: 586-608. Elton, C. S. 1973. The structure of invertebrate populations inside neotropical rain forest. J. Anim. Ecol. 42: 55-104. Emmel, T. C. 1974. Rainfall in the Amazon Basin (Letter). Science 183: 257-258. Faegri, K. , and L. van der Pijl. 1966. The Principles of Pollination Ecology. Pergamon Press, Oxford. 248 pp. Fountaine, M. E. 1913. Five months butterfly collecting in Costa Rica in the summer of 1911. Entomologist' 46: 189-194, 214-219. Fox, R. M. 1940. A generic review of the Ithomiinae (Lepidoptera: Nymphalidae) . Trans. Amer. Ent. Soc. 66: 161-207. Fox, R. M. 1949. The evolution and systematics of the Ithomiidae (Lepidoptera). Univ. of Pitt. Bull 45: 1-12. Fox, R. M. 1956. A monograph of the Ithomiidae (Lepidoptera), Part I. Bull. Amer. Mus. Nat. Hist. Ill: 76 pp. Fox, R. M. 1960. A monograph of the Ithomiidae (Lepidoptera), Part II. Trans. Amer. Ent. Soc. 86: 109-171. Fox, R. M. 1963 (1964) . Affinities and distribution of Antillean Ithomiidae. J. Res. Lep. 2: 173-184. Fox, R. M. 1967. A monograph of the Ithomiidae (Lepidoptera), Part III. Tribe Mechanitini Fox. Amer. Ent. Soc, Mem. 22: 200 pp. Fox, R. M. , and H. G. Real. 1971. A monograph of the Ithomiidae (Lepidoptera), Part IV. The tribe Napeogenini Fox. Amer. Ent. Inst., Mem. 15: 368 pp. Fraenkel, G. S.- 1959. The raisan de'etre of secondary plant substances. Science 129: 1466-1470. Gentry, J. L. , Jr., and P. Standley. 1974. Solanaceae. In "Flora of Guatemala." Fieldiana: Bot. 24: 1-151. -Gilbert* L. E. 1966. The effects of meterological factors on butterfly flight activity. Mimeo Research Report. OTS, Ciudad Universitaria, Costa Rica. July-Aug. Gilbert, L. E. 1969. Some aspects of the ecology and community structure of ithomid butterflies in Costa Rica. Mimeo Research Report. OTS, Ciudad Universitaria, Costa Rica. July-Aug. 354 Gilbert, L. E. 1971. Butterfly-plant coevolution: Has Passif lora adenopoda won the selectional race with Heliconiine butterflies? Science 172: 585-586. Gilbert, L. E. 1972. Pollen feeding and reproductive biology of Heliconius butterflies. Proc. Nat. Acad. Sci. 69: 1403-1407. 4 Gilbert, L. E. 1975. Ecological consequences of a coevolved mutualism between butterflies and plants. In Coevolution of Animals and Plants, L. E. and P. Raven (eds.), pp. 210-240, Univ. Texas Press, Austin. 246 pp. Gilbert, L. E. v and P. R. Ehrlich. 1970. The affinities of the Ithomiinae and the Satyrinae (Nymphalidae) . J. Lep. Soc. 24: 297-300. Gundlach, J. 1881. Contribucion a la entomologia Cubana. Havana. Guppy, J. L. 1894. Notes on Trinidad butterflies. J. Trinidad Field Nat. Club 2: 170-174. Guppy, J. L. 1904. Notes on the habits and early stages of some Trinidad butterflies. Trans. Ent. Soc. London 52: 225-228. Haensch, R. 1903. Kurzer Bericht uber die entomolg. Ergebnisse meiner Ecuador-Reise. Die Ithomiinen meiner Ecuador-Reise. Berliner Ent. Zeitschr. 48: 149-214. Haensch, R. 1905. Neue sudamerikanische Ithomiinae meiner Sammlung. Berliner Ent. Zeitschr. 50: 142-182. Haensch, R. 1909. Familie: Danaidae. In Seitz, A., Die Gross- Schmetterlinge der Erde. Stuttgart, Vol. 5, pp. 113-171. pis. 31-41. Haffer, Jurgen. 1974. Avian speciation in tropical South America. Pub. Nuttall Ornith. Club., No. 14. 390 pp. Heiser, C. B. 1968a. The prehistoric cultivated plants of Colombia with particular reference to the lulo (Solanum quitoense) . XXXVII Congreso Internacional de Americanistas, Republica Argentina, 1966, Actas y Memorias, Vol. 2: 535-545. Heiser, C. B. 1968b. Some Ecuadorian and Columbian solanums with edible fruits. Ciencia y Naturaleza (Quito) 11: 3-9. Heiser, C. B. 1969. Nightshades: The Paradoxical Plants. W. H. Freeman and Co., San Francisco. 200 pp. Heiser, C. B. 1971 (1972). Notes on some species of Solanum (Sect. Leptostemonum) in Latin America. Baileya 18: 59-65. Henry, T. A. 1949. (4th ed.) The Plant Alkaloids. The Blaikst'on Co., New York. 804 pp. 355 Holdridge, L. R. 1967. Life Zone Ecology. (Rev. Ed.). Tropical Science Center, San Jose, Costa Rica. 206 pp. Holdridge, L. R. , W. C. Grenke, W. H. Hatheway, T. Liang, J. A. Tosi, Jr. 1971. Forest Environments in Tropical Life Zones. A Pilot Study. Pergamon Press, Oxford. 747 pp. « Hunziker, A. T. 1969. Estudios sobre Solanaceae 5. Kurziana 5: 101-179. Hurlbert, S. H. 1971. The nonconcept of species diversity: A critique and alternative parameters. Ecology 52: 577-586. Hutchinson, E. H. 1967. A Treatise on Limnology. Vol. II. John Wiley and Sons, New York. 1115 pp. ~~ Jimenez, J. A. 1966. La naranjilla. Ministerio de Agricultura y Ganaderia. Puyo, Ecuador. Jolly, G. M. 1965. Explicit estimates from capture data with both death and immigration — stochastic model. Biometrika 52: 225-247. Kaye, W. J. 1905. Transparency in the wings of Lepidoptera. The Ent. Rec. and J. of Var. 17: 83-86, 115-121. Kaye, W. J. 1906. Notes on the dominant Mullerian group of butterflies from the Potaro district of British Guiana. Trans. Ent. Soc. London 53: 411-435. Kaye, W. J. 1914. _The Ithomiinae, Proc. So. Lond. Ent. and Nat. Hist. Soc. 1913-1914: 38-48. Kaye, W. J. 1921. A catalogue of the Trinidad Lepidoptera Rhopalocera. Mem. Dept. Agr. Trinidad and Tobago 2: 1-163. Klots, A. B. 1951. A Field Guide to the Butterflies of North America, East of the Great Plains. Houghton Mifflin Co., Boston. 349 pp. Labine, P. A. 1966. The population biology of the butterfly, Euphydryas editha. IV. Sperm precedence — A preliminary report. Evolution 20: 580-586. Lamas Muller, G. 1973. Taxonomia e evolucao dos generos Ituna Doubleday (Danainae) e Paititia, gen. n. , Thyridia Hiibner e Methona Doubleday (Ithomiinae) (Lepidoptera, Nymphalidae) . D.Sc. Thesis to the Departamento do Zoologia, Instituto de Biociencias, Universidade de Sao Paulo, pp. vii, 1-225. Lawrence, G. H. M. 1951. Taxonomy of Vascular Plants. The Macmillan Company, New York. 823 pp. Lawrence, G. H. M. 1960. The cultivated species of S planum. Baileya 8: 21-35. 356 Lichy, R. 1944. Documents pour servir a l'etude des lepidopteres du Venezuela. Bol. Acad. Cien. Fis. Matem. Nat. 24: 13-27. Longstaff, G. B. 1908. Bionomlc notes on butterflies. Trans. Ent. Soc. Lond. 1908: 607-662. Longstaff, G. B. 1912. Butterfly-Hunting in Many Lands. Longmans, Green, and Co. , London. 729 pp. MacBride, J. F. 1962. Solanaceae (Juss.) Lindl. In "Flora of Peru." Publication 951, Bot. Ser. , Field Mus. Nat. Hist., Vol. 13, Part V-B, No. 1: 1-267. Magnus, D. B. E. 1963. Sex limited mimicry II — Visual selection in the mate choice of butterflies. Proc. 16th Intern. Cong. Zool. 4: 179-183. Manski, M. J. 1960. Foodplants of some Queensland Lepidoptera. Qd. Nat. 16: 68-73. Masters, J. H. 1968. Collecting Ithomiidae with heliotrope. J. Lep. Soc. 22: 108-110. Mayr, E. 1954. Change of genetic environment and evolution. In J. Huxley, A. C. Hardy, and E. B. Ford, eds., Evolution as a Process. Allen and Unwin, London, pp. 157-180. McCubbin, C. 1971. Australian Butterflies. Thomas Nelson, Melbourne. 206 pp. Miller, L. D. 1968. The higher classification, phylogeny and zoo- geography of the Satyridae (Lepidoptera). Amer. Ent. Soc, Mem. 24: 174 pp. Miller, L. D. , and H. K. Clench. 1968. Some aspects of mating behavior in butterflies. J. Lep. Soc. 22: 125-132. Moreira, N. 1881. Metamorphose de uma Heliconia. Arch. Mus. Rio de Janeiro 4: 1-13. Moss, A. M. 1947. Notes on the Syntomidae of Para, with special . reference to wasp mimicry and fedegoso, Heliotropium indicum (Boraginaceae) , as an attractant. Entomologist 80: 30-35. Moulton, J. C. 1909. On some of the principal mimetic (Mullerian) combinations of tropical American butterflies. Trans. Ent. Soc. Lond. 56: 585-606. Muller, F. 1877. On hair-tufts, felted patches, and similar structures on the wings of male lepidoptera. Jenaische Zeitschrift fur Naturwissenschaft, XI (New Series, IV.): 99-114. Trans, in Longstaff, G. B. 1912. Butterfly-Hunting in Many Lands. Longmans, Green, and Co., London, pp. 604-619. 357 Muller, F. 1879. Ituna and Thyrldia: a remarkable case of mimicry in butterflies. Proc. Roy. Ent. Soc. Lond. 1879: xx-xxix. Trans, by R. Meldola from "Kosmos," May, 1879, p. 100. Muller, W. 1886. Sudamerikanische Nymphalidenraupen. Zool. Jahrb. 1: i-x, 1-225. Pacheco, R. , and J. Jimenez. ca. 1968. El cultivo do la naranjilla en el Ecuador. Ministerio de Agricultura y Ganaderia. Quito, Ecuador. Papageorgis, C. 1974. The adaptive significance of wing coloration in neotropical butterflies. Ph.D. dissertation, Princeton University. 196 pp. Papageorgis, C. 1975. Mimicry in neotropical butterflies. Amer. Sci. 63: 522-532. Patino, V. M. 1962. Edible fruits of Solanum in South American historic and geographic references. Bot. Mus. Leaf. (Harvard U.) 19: 215-234. Pearson, D. L. 1972. The birds of Limoncocha. Instituto Linguistico de Verano, Quito. 13 pp. Peet, R. K. 1974. The measurement of species diversity. Ann. Rev. Ecol. and Syst. 5: 285-307. Pielouj-E., C. 1967. The use of information theory in the study of the diversity of biological populations. Proc. Berkeley Symp. Math. Stat. Prob., 5th, 4: 163-177. Pliske, T. E. 1975a. Attraction of Lepidoptera to plants containing pyrrolizidine alkaloids. Environ. Ent. 4: 455-473. Pliske, T. E. 1975b. Pollination of pyrrolizidine alkaloid-containing plants by male Lepidoptera. Environ. Ent. 4: 474-479. Pliske, T. E. 1975c. Courtship behavior and use of chemical communi- cation by males of certain species of Ithomiine butterflies (Nymphalidae : Lepidoptera). Ann. Ent. Soc. Amer. 68: 935-942. Pliske, T. E., J. A. Edgar, and C. C. J. Culvenor. 1976. The chemical basis of attraction of ithomine butterflies to plants containing pyrolizidine alkaloids. J. Chem. Ecol. (in press). Poole* R. W. 1970. —Habitat preferences of some species of a Mullerian- mimicry complex in Northern Venezuela, and their effects on evolution of mimic-wing pattern. J. N. Y. Ent. Soc. 78: 121-129. Poole, R. W. 1974. An Introduction to Quantitative Ecology. McGraw- Hill, Series in Population Biology, New York. 532 pp. 358 Poulton, E. B. 1898. Essays on evolution. J. Linn. Soc. (1898): 558-612. Poulton, E. B. 1907. The significance of some secondary sexual characters in butterflies. Proc. Roy. Ent. Soc. Lond. (1907): xl-xliii. Poulton, E. B. 1931. The gregarious sleeping habits of a heliconian and ithomiine butterfly in Trinidad, observed by P. Lechmere Guppy. Proc. Ent. Soc. Lond. 6: 68-69. Poulton, E. B. 1933. The gregarious resting habits of dananine butter- flies in Australia: also of heliconine and ithomiine butterflies in Tropical America. Proc. Ent. Soc. Lond. 7: 64-67. Price, P. W. 1975. Insect Ecology. John Wiley and Sons, New York. 514 pp. - • Punnett, R. C. 1915. Mimicry in Butterflies. Cambridge Univ. Press, Cambridge. 188 pp. Raffauf, R. F. 1970. A Handbook of Alkaloids and Alkaloid-Containing Plants. Wiley- Interscience, New York. Unpaginated. Rathcke, B. J., and R. W. Poole. 1975. Revolutionary race continues: Butterfly larval adaptation to plant trichomes. Science 187: 175- 176. Retnakaran, A. 1974. The mechanisms of sperm precedence in the spruce budworm, Choristoneura fumiferana (Lepidoptera: Tortricidae) . Can. Ent. 106: 1189-1194. Richards, P. W. 1964. (2nd ed.) The Tropical Rain Forest. Cambridge Univ. Press, London. 450 pp. Root, R. B. 1967. The niche exploitation pattern of the blue-gray gnatcatcher. Ecol. Monogr. 37: 317-350. Ross, G. N. 1967. A distributional study of the butterflies of the Sierra de Tuxtla in Veracruz, Mexico. Ph.D. dissertation, Louisiana State University. 266 pp. Rothschild, M. 1973. Secondary plant substances and warning colouration in insects. In Insect/Plant Relationships. H. F. Van Emden (ed.), pp. 59-83, Blackwell Scientific Publications, Oxford. 215 pp. Sehultes, R. E. , and.R. Romero-Castaneda. 1962. Edible fruits of Solanum in Colombia. Bot. Mus. Leaf. (Harvard U.) 19: 235-286. Scott, J. A. 1972 (1973). Mating of butterflies. J. Res. Lep. 11: 99-127. Scott, J. A. 1973 (1974). Lifespan of butterflies. J. Res. Lep. 12: 225-230. 359 Seitz, A. (ed.) 1924. The Macrolepidoptera of the World. Vol. 5. The American Rhopalocera. Alfred Kernan Verlag, Stuttgart. 1139 pp. Sepp, J. 1828-1855. Surinaamsche Vlinders. 3 vols, (usually bound as one) . Amsterdam. Shields, 0. 1967. Hilltopping: an ecological study of summit con- gregation behavior of butterflies on a southern California hill. J. Res. Lep. 6: 69-178. Shields, 0., and J. F. Emmel. 1973. A review of carrying pair behavior and mating times in butterflies. J. Res. Lep. 12: 25-64. Silberglied, R. E. 1976. Communication in the Lepidoptera. In Animal Communication (2nd ed.), T. A. Sebeok, ed. Indiana Univ. Press, Bloomington. (in press) . Simpson, E. H. 1949. Measurement of diversity. Nature 163: 688. Southwood, T. R. E. 1966. Ecological Methods, With Particular Reference to the Study of Insect Populations. Methuen and Co., London. 391 pp. Standley, P. C, and C. V. Morton. 1937. Solanaceae. In "Flora of Costa Rica." Field Mus. Nat. Hist. Bot. 18: 1035-1099. Swihart, S. L. 1972. The neural basis of colour vision in the butter- fly, Heliconius erato. J. Insect Physiol. 18: 1015-1025. Takhtajan, A. 1969. Flowering Plants: Origin and Dispersal. Smith. Inst. Press, Washington, D.C. 310 pp. Turner, J. R. G. 1971. Experiments on the demography of tropical butterflies. II. Longevity and home-range behaviour in Heliconius erato. Biotropica 3: 21-31. Turner, J. R. G. 1975. Communal roosting in relation to warning colour in two Heliconiine butterflies (Nymphalidae) . J. Lep. Soc. 29: 221-226. Uphof, J. C. T. 1968 (2ad.ed.) Dictionary of Economic Plants. J-.... ^, ,. Cramer, Lehre, Germany. 591 pp. van der Pijl, L. , and C. H. Dodson. 1966. Orchid Flowers: Their Pollination and Evolution. Univ. Miami Press, Coral Gables. 214 pp. Vane-Wright, R. I. 1972- Pre-courtship activity and a new scent QEganr v > in butterflies. Nature 239: 338-340. Wiklund, C. 1974. Oviposition preferences in Papilio machaon in relation to the host plants of the larvae. Ent. exp. appl. 17: 189-198. 360 Wilbur, H. M. , D. W. Tinkle, and J. P. Collins. 1974. Environmental certainty, trophic level, and resource, availability in life history evolution. Amer. Nat. 108: 805-817. Willaman, J. J., and B. G. Schubert, and Their Contained Alkaloids Bull. No. 1234. Govt, 1961. Alkaloid Bearing Plants U.S.D.A. Agric. Res. Serv. Tech. Pointing Office, Washington, D.C. 287 pp. Wynne-Edwards, V. C. 1962. Animal Dispersion in Relation to Social Behaviour. Hafner Publ. Co., New York. 653 pp. Young, A. M. 1971. Notes on gregarious roosting in tropical butter- - flies of the genus Morpho. J. Lep. Soc. 25: 223-234. Young, A. M. 1972. On the life cycle and natural history of Hymenitis nero (Lepidoptera: Ithomiinae) in Costa Rica. Psyche 79: 284-294, Young, A. M. 1973. The life cycle of Dircenna relata (Ithomiidae) in Costa Rica. J. Lep. Soc. 27: 258-267. Young, A. M. 1974a. A natural historical account of Oleria zelica pagasa (Lepidoptera: Nymphalidae: Ithomiinae) in a Costa Rican Mountain Rain Forest. Studies Neotrop. Fauna 9: 123-140. Young, A. M. 1974b. Notes on the biology of Pteronymia notilla (Ithomiidae) in a Costa Rican mountain forest. J. Lep. Soc. 28: 257-268. Young, A. M. 1974c. On the biology of Godyris zavaleta caesiopicta (Lepidoptera: Nymphalidae: Ithomiinae). Ent. News 85: 227-238. BIOGRAPHICAL SKETCH Boyce Alexander Drummond III was born 28 February 1946 in Denison, Texas. He received his primary and secondary education in the public schools of Arkadelphia, Arkansas. From 1964 until 1967 he attended Henderson State College, from which he graduated with a Bachelor of Science Cum Laude in biology. In 1967 he began graduate study in the Department of Zoology of the University of Texas at Austin, where he held a position of Research Scientist Assistant. After serving in the United States Army as a member of the Fourth U.S. Army Band in San Antonio, Texas, from 1968 until 1971, he graduated with a Master of Arts in zoology from the University of Texas at Austin in 1971. He entered the Department of Zoology of the University of Florida in the fall of 1971 and has since pursued studies leading to the degree of Doctor of Philosophy. During this time he has held positions in the Department of Zoology as a graduate fellow, a graduate teaching assistant, and a graduate research assistant. Mr. Drummond is married to the former Nancy Jean Price of Fort Smith, Arkansas, who is presently studying for the degree of Doctor of Medicine in the University of Florida's College of Medicine. He is a Research Associate of the Florida State Collection of Arthropods, Florida Department of Agriculture, and a member of the American Association for the Advancement of Science, the Society for the Study of Evolution, the Entomological Society of America, the Lepidopterists' Society, Sigma Xi, and Alpha Chi. 361 I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Thomas C. Emmel, Chairman Professor and Chairman of Zoology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Archie F. Carr Graduate Research Professor of Zoology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. F. Clifford Johnson Professor of Zoology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Thomas J. Walkt Professor of Entomology This dissertation was submitted to the Graduate Faculty of the Department of Zoology in the College of Arts and Sciences and to, the Graduate Council, and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. August, 1976 Dean, Graduate School