Two Possible Caterpillar Complexes in Neotropical Danaine (: ) Author(s) :Keith R. Willmott, Marianne Elias, and Andrei Sourakov Source: Annals of the Entomological Society of America, 104(6):1108-1118. 2011. Published By: Entomological Society of America DOI: URL: http://www.bioone.org/doi/full/10.1603/AN10086

BioOne (www.bioone.org) is a a nonprofit, online aggregation of core research in the biological, ecological, and environmental sciences. BioOne provides a sustainable online platform for over 170 journals and books published by nonprofit societies, associations, museums, institutions, and presses. Your use of this PDF, the BioOne Web site, and all posted and associated content indicates your acceptance of BioOne’s Terms of Use, available at www.bioone.org/page/terms_of_use. Usage of BioOne content is strictly limited to personal, educational, and non-commercial use. Commercial inquiries or rights and permissions requests should be directed to the individual publisher as copyright holder.

BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers, academic institutions, research libraries, and research funders in the common goal of maximizing access to critical research. SPECIAL FEATURE Two Possible Caterpillar Mimicry Complexes in Neotropical Danaine Butterflies (Lepidoptera: Nymphalidae)

1,2 3 1 KEITH R. WILLMOTT, MARIANNE ELIAS, AND ANDREI SOURAKOV

Ann. Entomol. Soc. Am. 104(6): 1108Ð1118 (2011); DOI: http://dx.doi.org/10.1603/AN10086 ABSTRACT Caterpillar mimicry is surprisingly scarce, despite many examples of apparently de- fended, aposematic . Here, we describe two possible examples of caterpillar mimicry in two tribes of the Neotropical : and . The Þrst example, from the Caribbean island of Hispaniola, includes two subtribes of Danaini: plexippus (L.), Danaus gilippus (Cramer), Danaus cleophile (Godart) (Danaina), briarea (Godart), and Anetia jaegeri (Me´- ne´trie´s) (Itunina). The Þrst two widespread Danaus species have unusually dark phenotypes on Hispaniola, which we suggest are the result of mimicry with endemic Caribbean danaines. The second example, from the upper Amazon of eastern , involves four subtribes of Ithomiini: Forbestra olivencia (Bates) (Mechanitina), fluonia (Hewitson), Hypothyris semifulva (Salvin) (Napeogenina), amarilla Haensch (Ithomiina), anchiala (Hewitson), sex- maculata (Haensch) (Oleriina), and florula (Hewitson) (Godyridina). Hyposcada illin- issa (Hewitson) (Oleriina) is a possible additional member. This mimicry ring shows a color pattern known only from the upper Amazon, with the caterpillar having a yellow body and bright blue anterior and posterior segments, and this pattern has clearly evolved at least four times in the Ithomiini. We suggest that precise mimicry among caterpillars may be rarer than among adult butterßies because of a lack of sexual selection to drive the initial of bright colors in larvae. We also suggest that the evolution of warning colors in protected caterpillars is more difÞcult than in butterßies, because a novel, conspicuous caterpillar is less able to avoid capture than the more agile adult.

KEY WORDS , immature stages, larval mimicry, life history,

Predation pressure on caterpillars is evidently high, berg 1988). Such warning coloration implies the im- contributing to a diverse array of adaptive responses portance of vertebrate predators with keen color vi- (Dempster 1984). Caterpillar defenses include crypsis sion, believed principally to be birds, which are key or protective resemblance to inedible items such as selective agents on caterpillars (Holmes 1990, Hooks bird droppings (e.g., Futahashi and Fujiwara 2008), et al. 2003, Van Bael et al. 2007) and are even consid- chemical camoußage (e.g., Portugal and Trigo 2005), ered potential biological control agents of caterpillar development of false eyes to startle predators (e.g., pests (Sanz 2001, Mols and Visser 2002). Janzen et al. 2010), physical defenses including spines Under such a scenario, therefore, it is surprising that and hairs (e.g., DeVries 1987), behavioral strategies few cases of either Batesian (Bates 1862) or Mu¨ llerian (e.g., Gentry and Dyer 2002), spatial (e.g., Murphy (Mu¨ ller 1879) mimicry, phenomena so widespread 2004), and temporal avoidance (e.g., Stefanescu et al. among adult butterßies, have been reported among 2003), recruitment of ant defenders through mutual- caterpillars (Berenbaum 1995). Conspicuous warning isms (Pierce et al. 2002; Rico-Gray and Oliveira 2007), color patterns in larvae, such as a striped black-and- and chemical defense (Brower 1984, Trigo 2000, Opitz white body and a red head, or an orange body with and Mu¨ ller 2009). Shifts in dominant predators black projections, occur commonly throughout the throughout the life cycle (Dempster 1984) often lead Lepidoptera (Wagner 2005, Conner 2009, Janzen and to ontogenetically distinct defensive strategies Hallwachs 2009). Demonstrating that similar patterns (Nentwig 1985, Futahashi and Fujiwara 2008, Higgin- result from mimicry, however, can be difÞcult. For son and Ruxton 2010). Bright warning coloration, or example, although Berenbaum (1995) suggested that aposematism, often accompanied by gregariousness, larvae of Papilio polyxenes F. (Papilionidae) are mi- has also evolved in many species that are physically metic of Danaus plexippus (L.) (Nymphalidae), and/or supposedly chemically defended (Sille´n-Tull- Brower and Sime (1998) argue convincingly, on the basis of phylogenetic, biogeographic, and ecological 1 McGuire Center for Lepidoptera and , Florida Mu- evidence, that any resemblance is merely incidental. seum of Natural History, University of Florida, Gainesville, FL 32611. Other cases of caterpillar mimicry involve both puta- 2 Corresponding author, e-mail: kwillmott@ßmnh.uß.edu. 3 CNRS, UMR 7205, Muse´um National dÕHistoire Naturelle, 45 Rue tive Batesian and Mu¨ llerian systems and are listed in Buffon, CP50, 75005 Paris, France. Table 1.

0013-8746/11/1108Ð1118$04.00/0 ᭧ 2011 Entomological Society of America November 2011 WILLMOTT ET AL.: MIMICRY IN DANAINE CATERPILLARS 1109

Table 1. Proposed cases of caterpillar mimicry

Mimicry Model/co-mimic Mimic/co-mimic Reference Batesian Euphydryas phaeton (Drury) Nymphalidae Chlosyne harrisii (Scudder) Nymphalidae Bowers (1993) Batesian Papilio memnon L. Papilionidae erminea (Esper) Bowers (1993) Batesian Dendrolimus pini (L.) Lasiocampidae Hyloicus pinastri (L.) Sphingidae Pelzer (1992) Mu¨ llerian terpsichore (L.) Nymphalidae Acraea alicia (Sharpe) Nymphalidae Carpenter (1913) Mu¨ llerian Acraea alciope Hewitson Nymphalidae Acraea orestia Hewitson Nymphalidae Carpenter (1913) Mu¨ llerian Acraea macarista (Sharpe) Nymphalidae Acraea poggei Dewitz, Acraea arenaria Carpenter (1913) (Sharpe) Nymphalidae Mu¨ llerian melpomene (L.), Eueides tales (Cramer) Nymphalidae Brown and Holzinger (1973) Heliconius numata (Cramer) Nymphalidae Mu¨ llerian Eueides spp. Nymphalidae Josia spp. Notodontidae Brown and Holzinger (1973) Mu¨ llerian Meris alticola Hulst Geometridae Neoterpes graefiaria (Hulst) Geometridae Poole (1970) Mu¨ llerian Parnassius apollo (L.) Papilionidae Glomeris guttata Risso Diplopoda Deschamps-Cottin and Descimon (1996) Mu¨ llerian Heliconius hewitsoni Staudinger Nymphalidae Eueides vibilia (Godart) Nymphalidae Mallet and Longino (1982)

We suggest that multiple criteria, individually or in by predators that have sampled a protected indi- concert, might be used to support a hypothesis of vidual, ideally in natural surroundings. Thus, larvae caterpillar mimicry, including the following: are in some way unproÞtable (not necessarily un- palatable) and bear a color pattern that can be 1) Behavioral or physical traits that suggest that mod- learnt and avoided. els are protected, such as gregariousness, bright warning coloration, exposed feeding or observed Although convincing experimental evidence is dif- rejection by predators. Gregariousness is perhaps Þcult to obtain for a range of predators in natural the simplest trait to score and may beneÞt pro- conditions, fulÞllment of either criterion two or 4, tected prey by reducing the chances of an indi- especially in conjunction with other criteria or evi- vidual encountering inexperienced predators dence for host plant toxicity, makes a hypothesis of (Leimar et al. 1986) and by limiting attacks mimicry reasonable. through predator satiation (Sille´n-Tullberg and Larval mimicry is a potentially rewarding Þeld for Leimar 1988). Furthermore, gregariousness may research, but very few cases are yet reported that will increase the effectiveness of aposematic signals in permit rigorous testing. In particular, conÞnement of unpalatable prey (Gamberale and Tullberg 1996). our knowledge of larval mimicry to only a few species, 2) Phylogenetic evidence that one or more distinctive typically only a species pair in each case, hinders the warning signal elements are uniquely derived in kinds of comparative analyses that have recently be- putative mimics (and perhaps their descendants). gun to elucidate the evolution and ecology of adult For example, Brower and Sime (1998) point out butterßy mimicry rings (Mallet and Gilbert 1995, Bec- that putative mimetic pattern elements evolved in caloni 1997a, DeVries et al. 1999, Elias et al. 2008, Hill basal members of the Papilio machaon L. group that 2009). Here, we therefore describe two likely larval had probably never had contact with putative mimicry complexes among Neotropical members of model D. plexippus, arguing against mimicry. the tribes Danaini and Ithomiini (Nymphalidae, Da- 3) Ecological and biogeographic evidence that puta- nainae). Because these groups are some of the best tive mimics co-occur in space and time, at multiple studied nymphalids, integrating ecological, phyloge- scales, from microhabitat to geographic region. netic and biogeographic data should in future permit Demonstrating that species that mimic one an- novel tests of the larval mimicry hypothesis in these other (co-mimics) co-occur more often than non- two groups. mimics would strengthen any such observations. 4) Concordant geographic change in warning signal Materials and Methods among co-mimics and models, such as is ob- served in most mimicry rings involving adult Study Groups. The Danaini include Ϸ150 species of butterßies. Notable examples of the latter in- relatively large, colorful butterßies in tropical and clude the precisely matching zones sep- temperate regions throughout the world, with the arating co-mimetic subspecies of the Neotropi- highest species diversity in the oriental region (Ack- cal Heliconius erato (L.) and Heliconius ery and Vane-Wright 1984). Both immature stages melpomene (L.) (Sheppard et al. 1985), and the (Wiklund and Sille´n-Tullberg 1985) and adults geographically restricted mimetic female (Brower 1958, Chai and Srygley 1990, Pinheiro 1996) morphs of Brown in are unpalatable and warningly colored, and adults are (Ford 1936). Importantly, this kind of evidence extensively involved in both Mu¨ llerian and Batesian supports mimicry even among closely related mimicry rings (Ackery and Vane-Wright 1984). In the species that share the same ancestral warning Americas, the Danaini include only 13 species in three color pattern (Willmott and Mallet 2004). genera, Danaus Kluk, Doubleday, and Anetia 5) Experimental evidence that a warning color pat- Hu¨ bner (Lamas 2004). Adult mimicry is typically lim- tern reduces the frequency of attacks on co-mimics ited, with the exception of Lycorea, which is a member 1110 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 104, no. 6 of diverse mimicry complexes with ithomiines and Results heliconiines (e.g., Beccaloni 1997b). Danaini. Five danaine species were observed ßying The tribe Ithomiini, recently included in the Dan- Ϸ together in the study locality in Dominican Republic: ainae (Wahlberg et al. 2009), contains 360 species the widespread Danaus plexippus and Danaus gilippus distributed from sea level to 3,000 m from Mexico to (Cramer) and the Caribbean endemics Danaus cleo- southern and Argentina. Ithomiines are inhab- phile (Godart), Anetia jaegeri (Me´ne´trie´s), and Anetia itants of humid forest, where they are particularly briarea (Godart). All of these species, except A. bri- abundant in the understory. Adults of all species are area, are known or suspected to feed on A. nivea at the believed to be unpalatable (Brown 1984) and partic- study site. Larvae of D. plexippus (Fig. 1A), collected ipate in mimicry with other ithomiines or putatively as eggs from the latter host plant, had much broader unpalatable butterßies, especially the dark bands than typical individuals from continental (Chai and Srygley 1990), dominating these mimicry United States (Fig. 1B; Warren et al. 2010). D. gilippus rings in species diversity and abundance (Beccaloni larvae in the continental United States usually bear a 1997b). banded white, black, and yellow pattern (e.g., Fig. 1D; Although there are few published observations of Warren et al. 2010) similar to D. plexippus but also predation on unpalatable tropical butterßies in nature occur very rarely as a darker phenotype (e.g., Calhoun (e.g., Brower 1988), the primary predators driving the 1996). This darker phenotype was uniformly observed evolution of mimicry are believed to be insectivorous at the Dominican Republic study site (Fig. 1C). D. birds, being the only abundant predators with suitably cleophile, observed ovipositing on A. nivea during the developed color vision to explain precise mimicry study, is described as “velvety black . . . with narrow (Chai 1986, Langham 2004). bands of yellow lying at the junction of the segments” Field Methods. Danaine immature stages and adults (Brown and Heineman 1972, p. 94), a description were studied in 1995 by A.S. above Mata Grande in the closely matching the observed Dominican Republic Cordillera Central of Dominican Republic (Santiago larvae of D. plexippus and D. gilippus. Several larvae of Province, 1,000Ð1,200 m, 19Њ 18Ј N, 70Њ 54Ј W). Eggs A. jaegeri were reared from the same host plant. These were collected on nivea L. (Apocynaceae) larvae were described by Sourakov and Emmel (1996) and reared to adult in resealable plastic bags with fresh and bear multiple narrow black and white rings (Fig. leaves of the natural host plant provided every 2Ð3 d. 1F). Finally, adults of A. briarea also were found in the same habitat, although no immature stages were en- Ithomiini were studied in 2005 at Napo Wildlife countered. In the fourth instar, A. briarea larvae have Center (NWC) and surroundings in eastern Ecuador a banded black-and-white pattern similar to A. jaegeri (Orellana Province, 240Ð300 m, 0Њ 31.43Ј S,76Њ 23.41Ј (Fig. 1F) and the mainland A. thirza Geyer (Llorente W) by K.R.W., M.E., and colleagues (Elias et al. 2008). et al. 1993), which changes in the Þfth instar (Fig. 1E) The study site was topographically complex and cov- to a pattern that seems predominantly dark with more ered with wet lowland rain forest, mostly undisturbed dispersed white bands. except along the edges of the Rõ´o Napo and Rõ´o Based on these observations, we suggest that the An˜ angu. Supplementary information comes from fur- unusually dark phenotypes of the widespread D. ther studies in 2001Ð2002 and 2006 at cloud forest plexippus and D. gilippus in Dominican Republic result localities in eastern Ecuador (Sucumbõ´os and Zamora- from mimicry with the three Caribbean-endemic da- Chinchipe, 1,000Ð3,000 m) (Willmott and Mallet 2004; naines observed at the study site. The endemic dan- Willmott and Lamas 2006, 2008). Individuals were aines have similar banded patterns that are notably located by searching many individual plants in known darker than those of the Þrst two species in the con- ithomiine host plant families, Solanaceae, Apocyn- tinental United States, and we consider the following aceae and (Willmott and Freitas 2006; in evaluating whether this dark, banded coloration is Beccaloni et al. 2008). Eggs and larvae were collected a warning color pattern and mimetic: and reared as described above. Photographs were 1) The black larval coloration is distinctive against the taken of each instar, and cast larval skins, head cap- host plant leaves where these larvae feed, and sules, and specimens of larvae and pupae were pre- unpalatability of all species in the putative mimicry served, where possible, in locally available indus- ring is probable based on their close phylogenetic trial alcohol (ethanol) and are deposited in the relationship, shared host plant, and evidence that Florida Museum of Natural History, Gainesville, FL. D. plexippus larvae, at least, are unpalatable to birds Voucher host plant specimens were collected and (Brower et al. 1967, Wiklund and Sille´n-Tullberg are deposited in the Natural History Museum, Lon- 1985). don, United Kingdom; and the Museo Ecuatoriano 2) The dark larval coloration observed in Hispaniolan de Ciencias Naturales, Quito, Ecuador. Additional danaines is widespread throughout the Danaini, information on the height above ground, host plant also occurring, for example, in the mainland A. height, density, and microhabitat also were re- thirza (Llorente et al. 1993) as well as in other corded. Data on the systematics and distribution of members of Danaus (e.g., Bascombe et al. 1999). It Ithomiini come from research in progress by Will- is therefore not clear whether the dark color pat- mott and Lamas to revise several ithomiine genera terns of Hispaniolan danaines represent an ances- (Willmott and Lamas 2006, 2008). tral or recently derived trait. However, the Þfth November 2011 WILLMOTT ET AL.: MIMICRY IN DANAINE CATERPILLARS 1111

Fig. 1. (AÐF) Putative Danaini dark-banded caterpillar mimicry complex from Dominican Republic (A, C, E, and F), with geographically distinct phenotypes (B and D). (A) D. plexippus, Dominican Republic. (B) D. plexippus, Wisconsin, North America. (C) D. gilippus, Dominican Republic. (D) D. gilippus, Arizona, North America. (E) Anetia briarea, Þfth instar, Dominican Republic. (F) Anetia jaegeri, Þfth instar, Dominican Republic. (GÐK) Approximate distributions of the Þve Dominican Republic danaines recorded at the study site (black dot). Distribution data from Ackery and Vane-Wright (1984), Smith et al. (1994), and www.butterßiesandmoths.org. Caterpillar images B and D courtesy of Gene Hanson, image E courtesy of Linda S. Fink; butterßy images from Warren et al. (2010). (Online Þgure in color.)

instar of A. briarea (Fig. 1E) is even darker than except A. briarea also occur on Jamaica, we would earlier instars (Brower et al. 1992), which resemble expect Jamaican D. plexippus and D. gilippus to A. jaegeri (Fig. 1F) and A. thirza (Llorente et al. have a similar dark larval color pattern. Unfortu- 1993). The particularly dark pattern of the Þfth nately, we have been unable to Þnd any precise instar of A. briarea might thus represent a unique descriptions of the early stages of these species on derived trait which at least enhances its similarity Jamaica (Brown and Heineman 1972). The Þfth to D. cleophile (based on Brown and HeinemanÕs instar D. plexippus on Guadeloupe (Zagatti et al. (1972) description of the latter). 2010) and on Puerto Rico (Warren et al. 2010) is 3) All putative mimics occur together in time and also very similar to that observed on Hispaniola, space, with four of the danaine species feeding or but we have otherwise not observed this pheno- ovipositing on the same host plant species in the type outside the Caribbean. same area. 4) The distinctive color pattern of D. plexippus and D. Ithomiini. Immature stage data were obtained in gilippus on Dominican Republic are the only ex- the Þeld for 32 of the 57 ithomiine species recorded at amples of which we know where an aposematic NWC (Fig. 2; Supp Table 1 [online only]) and from larval color pattern seems to have varied geograph- publications for an additional nine species (Brown and ically to match putative co-mimics. Sympatric D. Freitas 1994, Hill and Tipan 2008). Of the 41 species for cleophile, A. jaegeri, and A. briarea also bear similar which some data were available, 19 seem to be mostly dark patterns, and these three species co-occur cryptic (Fig. 2), being entirely green, or olive-green to only on Hispaniola (Fig. 1). Given that all species buff, often with diffuse pale yellow lateral markings 1112 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 104, no. 6 Ml menophilus whereas in the remaining species it became intense eurimedia Tth harmonia only in later instars (fourth to Þfth). The late Þfth

Methona curvifascia Mn instar to prepupa of two otherwise cryptic species, confusa Hypothyris fluonia (Hewitson) and H. semifulva (Sal- psidii vin) Fig. 3E and G), developed similar but more sub-

Scada zibia Mechanitina Forbestra equicola dued coloration with the blue coloration replaced by Forbestra olivencia mazaeus bluish gray. We tentatively classiÞed Hyposcada illin- issa (Hewitson) as cryptic, even though this species also shows a similar but less intense coloration in the Mechanitis messenoides Ithomia salapia Ithom Þfth instar (Fig. 3J). Ithomia agnosia Evidence for the blue-yellow color pattern being Ithomia amarilla Hypothyris semifulva Napeogenina aposematic and mimetic comes from several anecdotal Hypothyris fluonia observations. Hypothyris euclea pharo Napeogenes inachia 1) Larvae of F. olivencia are moderately gregarious, occurring mostly in groups of two to four individ- Hyposcada illinissa uals (with clusters of up to 13 eggs observed; Hill Hyposcada anchiala Oleriina Hyposcada kena 2006; K.R.W. and M.E., personal observation), a Oleria gunilla trait typically associated with aposematism and Oleria sexmaculata presumably unpalatability (Sille´n-Tullberg 1988). lenea 2) The blue-yellow color pattern is highly distinctive, tutia Dircennina Key sulphurea being unknown among well-studied ithomiine cryptic sao communities in adjacent montane Ecuador (e.g., Pteronymia primula Willmott and Mallet 2004), as well as throughout aposematic, Pteronymia vestilla mimetic? zavaleta the remainder of the range of the Ithomiini. For Godyridina aposematic, nephele example, Brown and Freitas (1994) Þgure ithomi- mimetic Brevioleria arzalia lavinia ine larvae for many species from several countries, aposematic Pseudoscada florula especially southeastern Brazil, with no evidence of 5th instar Pseudoscada timna a similar pattern elsewhere. In eastern Ecuador, Fig. 2. Cladogram for Ithomiini caterpillar community the pattern has apparently evolved at least four studied at NWC, showing phylogenetic distribution of cryp- times, in the subtribes Mechanitina, Ithomiina, tic (green), aposematic, possibly mimetic (black) and apose- Oleriina, and Godyridina. In Forbestra Fox, F. equi- matic, mimetic (yellow-blue) caterpillar color patterns. Sub- cola (Cramer) also has aposematic, gregarious lar- tribal names abbreviated as follows: Ml, Melinaeina; Tth, vae, but these are black, white, and yellow (Bre´vi- Tithoreina; Mn, Methonina; Ithom, Ithomiina. Double cat- gnon 2003), whereas the immature stages of the erpillars indicate gregariousness. Cladogram is reduced from that in Elias et al. (2008) based on DNA sequence data. only other congener, F. proceris (Weymer), are (Online Þgure in color.) unknown. Mechanitis F., the sister to Forbe- stra (Willmott and Freitas 2006), also has gregar- ious and aposematic larvae but of different color (Fig. 3D, F, H, and M). Hypothyris mamercus (Hewit- pattern (Figs. 2 and 3B). The color pattern of son) (all taxonomic authorities for ithomiine names Ithomia amarilla (Fig. 3C) differs sharply from are in Supp Table 1 [online only]) was cryptic in early known congeners, which typically resemble instars, but because no data were available for the late Ithomia agnosia Hewitson (Fig. 3D), although Þfth instar (see below), it is omitted from Fig. 2. The Ithomia salapia Hewitson is also aposematic, being remaining 22 species, representing all tribes and 14 black with yellow rings (K.R.W. and M.E., personal genera (Fig. 2), seem to be warningly colored to some observations). The early Þfth instar color patterns extent, at least in the Þfth instar, with many exhibiting of H. fluonia (Fig. 3F) and H. semifulva (Fig. 3H) color patterns that include black, white, and yellow are typical of the genus, and it is possible that late bands or mosaics. Although mimicry might be in- instars of species in other regions also may resem- volved among some of the black-, white-, and yellow- ble the putative mimics in Fig. 3E and G, which patterned caterpillars, these patterns are geographi- would suggest that resemblance in this case could cally widespread and not sufÞciently similar to be purely incidental. Similarly, H. illinissa is not provide strong evidence of mimicry. By contrast, Þve strikingly different from the Þfth instar of the of the warningly colored species in four different sub- closely related montane Megoleria orestilla (He- tribes [Forbestra olivencia (Bates), Ithomia amarilla witson) (Willmott and Lamas 2008), and resem- Haensch, Hyposcada anchiala (Hewitson), Oleria sex- blance may perhaps be fortuitous. However, the maculata (Haensch), and Pseudoscada florula (Hewit- intense blue anterior and posterior segments of H. son)] (Figs. 2; 3A, C, I, L, and N) exhibited a partic- anchiala are not known in any other congener or ularly novel color pattern in which the body is a member of the subtribe, except the sympatric O. greenish yellow with bright blue anterior (T1-T2) and sexmaculata (Fig. 3L). Finally, P. florula differs posterior (A9-A10) segments. In F. olivencia, this col- dramatically from the sympatric congener Pseu- oration was present from at least the third instar, doscada timna (Hewitson), which has a color pat- November 2011 WILLMOTT ET AL.: MIMICRY IN DANAINE CATERPILLARS 1113

Fig. 3. Putative Ithomiini yellow-blue caterpillar mimicry complex from eastern Ecuador (left column), with nonmimetic phenotypes (right column). Range maps based on distribution data compiled by K.W. (unpublished data). All specimens from Napo Wildlife Center, 250 m, Napo, Ecuador, except K, from Reserva Arcoiris, 2,100 m, Zamora-Chinchipe, Ecuador. (A) Forbestra olivencia (Mechanitina). (B) Mechanitis messenoides (Mechanitina). (C) Ithomia amarilla (Ithomiina). (D) Ithomia agnosia (Ithomiina). (E) Hypothyris fluonia, late Þfth instar (Napeogenina). (F) Hypothyris fluonia, early Þfth instar. (G) Hypothyris semifulva, late Þfth instar (Napeogenina). (H) Hypothyris semifulva, early Þfth instar. (I) Hyposcada anchiala (Oleriina). (J) Hyposcada illinissa (Oleriina). (K) Megoleria orestilla (Oleriina). (L) Oleria sexmaculata (Oleriina). (M) Oleria onega (Oleriina). (N) Pseudoscada florula (Godyridina). (O) Pseudoscada timna (Godyridina). (Online Þgure in color.)

tern typical of its genus (Brown and Freitas 1994) lated but nonmimetic sympatric species also have and of the closely related genus Heming similar ranges (e.g., O. onega Hewitson) but more (Willmott and Freitas 2006). often represent species at the edge of a largely 3) All of the proposed mimics are sympatric, being submontane distribution (e.g., I. agnosia and P. recorded from a single site of Ϸ1km2, where all timna) (Fig. 3). occur in forest understory. The geographic ranges of the proposed mimics also broadly overlap in the Discussion upper Amazon basin (Fig. 3); although H. illinissa and H. anchiala also have Transandean records, P. Danaini. Little is known about the palatability of florula has disjunct subspecies in the Guianas and danaine larvae beyond studies with D. plexippus (e.g., southeastern Brazil, and H. fluonia also extends to Wiklund and Sille´n-Tullberg 1985), but the correla- the lower Amazon and Guianas. Some closely re- tion between warning coloration and typically toxic 1114 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 104, no. 6 asclepiad host plants among many Danaini suggests that most species are chemically defended. Four of the Þve species at the Dominican Republic study site fed on the same host plant and are probably Mu¨ llerian mimics, whereas the natural host plant of A. briarea is not known (Brower et al. 1992). The hypothesized shift to a darker warning color- ation in D. plexippus and D. gilippus may have resulted from the abundance of other dark danaine larvae on Hispaniola, which contains the highest species rich- ness for this tribe in the Caribbean. The Caribbean Fig. 4. Possible model for the Ithomiini yellow-blue cat- endemics D. cleophile, A. briarea, and A. jaegeri are erpillar mimicry complex: sawßy species (Hymenoptera: relatively common in upland forests (Smith et al. Symphyta) observed near study site at Ecuador, Sucumbõ´os, 1994), certainly no rarer than the widespread D. Garzacocha, by R. Hill. (Online Þgure in color.) plexippus and D. gilippus. The apparent occurrence of dark D. plexippus larvae on Guadeloupe (Zagatti et al. 2010) and Puerto Rico (Warren et al. 2010), which In all of the remaining putative mimics, the larvae lack endemic Caribbean danaines, suggests that Ca- are solitary, members of that largely have cryp- ribbean Danaus plexippus megalippe (Hu¨ bner) colo- tic larvae, and cryptic until the later instars. In Ithomia nized these islands from Hispaniola if the dark larval Hu¨ bner, all known larvae of species in the con- phenotype is indeed mimetic. Clearly, however, much taining I. amarilla (Mallarino et al. 2005) are cryptic more research on the distribution of larval phenotypes (Brown and Freitas 1994; Fig. 3D). In Hypothyris Hu¨ b- and butterßy phylogeography throughout the Carib- ner, H. euclea is also aposematic (Young 1977, Brown bean is needed. and Freitas 1994), but there is no evidence that it Given the widespread presence of dark color pat- forms a clade with the two putative mimics described terns throughout the Danaini, it is of course possible here, H. fluonia and H. semifulva (C. Arias, unpub- that the dark Caribbean phenotype in D. plexippus and lished data). Delaying warning coloration until later D. gilippus represents an ancestral state, and that there instars in palatable species should reduce the relative is selection for paler forms in the continental United abundance of Batesian mimics, because late instars are States. Without any obvious selective advantage for less abundant than early instars and thus result in pale continental coloration, however, and given that fewer attacks on all individuals by naõ¨ve predators. the pale form is much more widespread, we prefer the Notwithstanding the previous discussion, conspic- hypothesis of a shift in color pattern in the Caribbean. uous larval coloration is relatively common in the Ithomiini. The Ithomiini mimicry ring perhaps rep- derived tribe Dircennina, in Callithomia Bates, Cera- resents Batesian or mixed BatesianÐMu¨ llerian mim- tinia Hu¨ bner, and in all three Pteronymia Butler & icry, with F. olivencia the most obvious unpalatable Druce at the Ecuadorian study site (and elsewhere; model. All basal ithomiine clades (Tithoreina, Metho- Brown and Freitas 1994; K.R.W. and M.E., personal nina, Melinaeina, and Doubleday ϩ observations), although none of these species is in- Fox) have brightly colored larvae, whereas most spe- volved in the blue-yellow mimicry ring (Fig. 2). Most cies in the more derived subtribes Napeogenina, ithomiines, including all Dircennina, feed on So- Ithomiina, Oleriina, Dircennina, and Godyridina have lanaceae, a plant family with diverse secondary chem- cryptic larvae (Fig. 2) (Brown and Freitas 1994, Will- icals (Brown 1987), and the likelihood that larvae of at mott and Freitas 2006). Within the transitional least some of these species are chemically protected Mechanitina, both cryptic and bright color patterns seems high. For example, Dyer (1995) found that occur, but aposematism probably was present in the cryptic oto (Hewitson) (Ithomiini: ancestor of Mechanitis ϩ Forbestra because these sister Godyridina) caterpillars were unpalatable to ants. It is genera are both aposematic (Fig. 2). Widespread clear, therefore, that assays of larval palatability are bright larval coloration among basal Ithomiini clades needed, and the fact that larvae of F. olivencia were not suggests that their larvae are protected, and Methona signiÞcantly more abundant in the study site than Doubleday and Mechanitis, at least, are unpalatable to those of putative Batesian mimics (K.R.W. and M.E., chicks (Massuda and Trigo 2009). The pyrrolizidine unpublished data) suggests that at least some Mu¨ lle- alkaloids that protect adult Ithomiini (Brown 1984) rian mimicry is probably involved. Finally, additional are known only in Tithorea Doubleday larvae (Trigo likely models are the later instars of a sawßy species and Brown 1990, Trigo et al. 1996), but different com- (Hymenoptera: Symphyta) that was observed by R. pounds might well be involved in other genera, such Hill (personal communication) at Garzacocha, a site as the tropane alkaloids present in dÕAlmeida several kilometers from the study site (Fig. 4). Sawßy (Freitas et al. 1996) and as yet unknown compounds larvae have various behavioral and chemical defenses in Methona (Massuda and Trigo 2009). The gregari- (Boeve´ and Pasteels 1985, Vlieger et al. 2004) and are ousness and appearance of aposematic coloration in often gregarious and aposematic. According to Hill, early instars of F. olivencia (Hill 2006) are both also the sawßy larvae at Garzacocha rested and fed on the consistent with the hypothesis that its larvae are un- underside of fern leaves (Pteridophyta), in the same palatable. areas as larvae of F. olivencia. November 2011 WILLMOTT ET AL.: MIMICRY IN DANAINE CATERPILLARS 1115

One obvious question is why only some of the con- the range of the majority of supposedly mimetic spicuously colored caterpillars at NWC adopt the ithomiines in Fig. 3. blue-yellow pattern, whereas others have a geograph- Brower and Sime (1998) suggested that because ically widespread black, yellow, and white pattern many caterpillars are host plant speciÞc, host plants (Brown and Freitas 1994, Willmott and Freitas 2006). will form a part of the predatorÕs search image. Thus, Among adult ithomiines, mimicry rings are at least predators learn to avoid speciÞc combinations of partially segregated by microhabitat (DeVries et al. plantÐcaterpillar color pattern rather than the color 1999, Elias et al. 2008, Hill 2009), ßight height (Bec- pattern itself. Although this is an interesting and plau- caloni 1997a, Elias et al. 2008), and season (DeVries et sible hypothesis, it is potentially relevant only to Ba- al. 1999), perhaps partly because of host plant distri- tesian mimicry. If color pattern was a reliable indicator bution (Beccaloni 1997a, Willmott and Mallet 2004). of unpalatability, then host plant phenotype would Notably, in many of the previously proposed cases of cease to be an important cue for identiÞcation of caterpillar mimicry the models and mimics occur in edible prey. However, in the Ithomiina example we restricted habitats (e.g., Deschamps-Cottin and Des- describe here, which we argue is likely to represent cimon 1996) or on the same host plants (Poole 1970, , there is little obvious phenotypic Mallet and Longino 1982, Pelzer 1992). Larvae occu- similarity among the host plants. pying narrower niches are exposed to a limited frac- We suggest an additional hypothesis that focuses on tion of the total predator population and are therefore the initial evolution of aposematic coloration. Explain- perhaps more likely to evolve mimicry, because the ing the origin of aposematic coloration in protected chances of encountering educated predators is higher. species is a challenge because the beneÞts of warning At NWC, however, there is no immediately obvious coloration become signiÞcant only when the apose- similarity among the microhabitats of the host plants matic pattern is abundant (e.g., Mallet and Singer for the eight species (M.E. and K.R.W., unpublished 1987). In adult butterßies, novel color patterns prob- data). ably arise frequently through sexual selection, because However, it is notable that the geographic ranges of many butterßies are brightly colored even if not mi- two of the most remarkable mimics, I. amarilla and O. metic or aposematic. Such patterns also may inciden- sexmaculata, are narrowly conÞned to a small region in tally provide a crude resemblance to a protected spe- the western Amazon. P. florula, also a notable mimic, cies that can be reÞned by selection for mimicry has a disjunct range, with a population conÞned to a (Mallet and Singer 1987). Obviously, larvae lack the similar area in the western Amazon, and other popu- mechanism of sexual selection to provide an initial lations in the Guianas and southeastern Brazil. If, as we beneÞt for bright coloration to outweigh the disad- suggest below, mimicry evolves much more rarely in vantages of increased visibility to predators, and this larvae than in adults, then the most likely warning might limit the evolution of larval mimicry. The role color pattern for a caterpillar to display is that which of sexually selected color patterns in the evolution of is most common throughout its geographical range. mimicry could be examined by testing whether mim- Species with broad ranges are therefore most likely to icry in adult butterßies arises more frequently in adopt already widespread patterns, whereas those clades containing species that have bright coloration with restricted ranges are more likely than widespread for sexual signaling. Furthermore, we suggest that a species to adopt existing patterns that are unique to a novel aposematic pattern in a cryptic, but protected, small region. Thus, in caterpillars, shared host plants butterßy is more likely to become widespread if it among co-mimics (Poole 1970, Mallet and Longino evolves in the adult rather than larval stage. This is 1982, Pelzer 1992, and Danaini reported here) may because adult butterßies are more likely than cater- result in mimicry because of similarity of geographic pillars to be able to escape capture once identiÞed as range rather than similarity of microhabitat. a potential prey item. Several studies have experimen- Evolution of Caterpillar Mimicry. Four main hy- tally documented the selective value of mimicry in potheses have been suggested to account for the rarity adult butterßies (Mallet and Barton 1989, Kapan of known cases of mimicry among lepidopteran larvae 2001), and similar studies could be conducted on cat- (Berenbaum 1995, Brower and Sime 1998). Poor erpillars. Finally, J. Mallet (personal communication) knowledge of tropical lepidopteran immature stages suggested that number dependence and apparency must surely contribute to the paucity of documented might explain rarity of caterpillar mimicry. Because cases of caterpillar mimicry (Berenbaum 1995). Fur- caterpillars are smaller and move less than adult but- thermore, the tendency for caterpillar researchers terßies, they are more likely to escape detection by (ourselves included) to study taxonomically limited using crypsis, whereas the relative rarity of adults hosts or herbivores must lead to cases of cross-taxon means that they gain more from mimicry by reducing mimicry being frequently overlooked. For example, the fraction sacriÞced during predator learning. the blue-yellow ithomiine pattern discussed here is The example of Danaus described here is the only also very similar to that of a presumably highly toxic example of which we know where larval aposematic (Murphy et al. 2010) Amazonian stinging slug cater- coloration is geographically variable, matching local pillar (Limacodidae), whose photograph we discov- models. This rarity of precise local mimicry, such a ered by chance at http://www.onlineadventure.com/ striking phenomenon in adult butterßies, itself con- photos/1964. Although the full distribution of this tributes to why larval mimicry is so rarely docu- limacodid is unknown, it occurs at least partly within mented. 1116 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 104, no. 6

Much of this discussion is necessarily speculative Boeve´, J.-L., and J. M. Pasteels. 1985. Modes of defense in with the hope of stimulating further research. The nematine sawßy larvae. J. Chem. Ecol. 11: 1019Ð1036. putative ithomiine larval mimicry ring documented Bowers, M. D. 1993. Aposematic caterpillars: life-styles of here is one of the most diverse known, and the ex- the warningly colored and unpalatable, pp. 331Ð371. In cellent knowledge of ithomiine phylogeny, ecology, N. E. Stamp and T. M. Casey (eds.), Caterpillars. Chap- and distribution should permit future comparative man & Hall, New York. Bre´vignon, C. 2003. Inventaire des Ithomiinae de Guyane analyses to test more rigorously what ecological traits Franc¸aise (Lepidoptera, Nymphalidae). Lambillionea most favor the evolution of larval mimicry. 103: 41Ð58. Brower, A.V.Z., and K. R. Sime. 1998. A reconsideration of mimicry and aposematism in caterpillars of the Papilio Acknowledgments machaon group. J. Lepid. Soc. 52: 206Ð212. Brower, J.V.Z. 1958. Experimental studies of mimicry in We thank Rau´ l Aldaz, Ismael Aldas, Alexandre Toporov, some North American butterßies. Part I. The monarch, Julia Robinson-Willmott, Aniko Zo¨lei, Gabor Papp, Chris Danaus plexippus and the , Limenitis archippus Jiggins, and Carlos Giraldo Sanchez for help with rearing archippus. Evolution 12: 32Ð47. caterpillars in Ecuador. We thank the Ecuadorian Ministerio Brower, L. P. 1984. Chemical defence in butterßies, pp. del Ambiente and Museo Ecuatoriano de Ciencias Naturales 109Ð134, pl. 2: Þgs. AÐD. In R. I. Vane-Wright and P. R. for providing research permits and the Napo Wildlife Center Ackery (eds.), The biology of butterßies. Academic, Lon- and Fundacio´n Arcoiris for logistic support in the Þeld. don, United Kingdom. K.R.W. thanks the numerous museum curators and individ- Brower, L. P. 1988. Avian predation on the monarch but- uals who allowed access to their collections to record ithomi- terßy and its implications for mimicry theory. Am. Nat. ine distributional data, and Fraser Simpson, in particular, for 131: S4ÐS6. help with databasing. Finally, we thank Jim Mallet and three Brower, L. P., J.V.Z. Brower, and J. M. Corvino. 1967. Plant anonymous reviewers for helpful comments on the manu- poisons in a terrestrial food chain. Proc. Natl. Acad. Sci. script and Ryan Hill for information on the sawßy that U.S.A. 57: 893Ð898. he observed at Garzacocha and for permission to reproduce Brower, L. P., M. A. Ivie, L. S. Fink, J. R. Watts, and R. A. his photograph in Fig. 4. The Leverhulme Trust (K.R.W. and Moranz. 1992. Life history of Anetia briarea and its bear- M.E.), the National Science Foundation (DEB 0103746, ing on the evolutionary relationships of the Danainae 0639861; K.R.W.), and the National Geographic Society (Re- (Lepidoptera: Nymphalidae). Trop. Lepid. 3: 64Ð73. search and Exploration 5751-96 [K.R.W.] and 5717 [A.S.]) Brown, F. M., and B. Heineman. 1972. Jamaica and its but- supported Þeldwork in Ecuador and the Dominican Re- terßies. E. W. Classey Ltd., London, United Kingdom. public. Brown, K. S. 1984. Adult-obtained pyrrolizidine alkaloids defend ithomiine butterßies against a spider predator. Nature 309: 707Ð709. References Cited Brown, K. S. 1987. Chemistry at the Solanaceae/Ithomiinae Ackery, P. R., and R. I. Vane-Wright. 1984. Milkweed but- interface. Ann. Mo. Bot. Gard. 74: 359Ð397. terßies. Their cladistics and biology. Being an account of Brown, K. S., and A.V.L. Freitas. 1994. Juvenile stages of the natural history of the Danainae, a subfamily of the Ithomiinae: overview and systematics (Lepidoptera: Lepidoptera, Nymphalidae. British Museum (Natural Nymphalidae). Trop. Lepid. 5: 9Ð20. History), London, United Kingdom. Brown, K. S., and H. Holzinger. 1973. The heliconians of Bascombe, M. J., G. Johnston, and F. S. Bascombe. 1999. The Brazil (Lepidoptera, Nymphalidae). Part IV. Systematics butterßies of Hong Kong. Academic, London United and biology of Eueides tales Cramer, with description of Kingdom. a new subspecies from . Z. Arb. Entomol. 24: Bates, H. W. 1862. Contributions to an fauna of the 44Ð65. Amazon Valley. Lepidoptera: Heliconidae. Trans. Linn. Calhoun, J. V. 1996. Conquering soldiers: the successful in- Soc. Lond. 23: 495Ð566. vasion of Florida by (Lepidoptera: Beccaloni, G. W. 1997a. Vertical stratiÞcation of ithomiine Nymphalidae). Hol. Lepid. 3: 7Ð18. butterßy (Nymphalidae: Ithomiinae) mimicry complex- Carpenter, G.D.H. 1913. Synaposematic resemblance be- es: the relationship between adult ßight height and larval tween Acraeine larvae. Trans. Entomol. Soc. Lond. 1912: host-plant height. Biol. J. Linn. Soc. 62: 313Ð341. 702Ð705. Beccaloni, G. W. 1997b. Ecology, natural history and be- Chai, P. 1986. Field observations and feeding experiments haviour of ithomiine butterßies and their mimics in Ec- on the responses of rufous-tailed jacamars (Galbula ru- uador (Lepidoptera: Nymphalidae: Ithomiinae). Trop. ficauda) to free-ßying butterßies in a tropical rainforest. Lepid. 8: 103Ð124. Biol. J. Linn. Soc. 29: 161Ð189. Beccaloni, G. W., A. L. Viloria, S. K. Hall, and G. S. Robinson. Chai, P., and R. B. Srygley. 1990. Predation and the ßight, 2008. Catalogue of the hostplants of the Neotropical but- morphology, and temperature of Neotropical rain-forest terßies/Cata´logo de las Plantas Hue´sped de las Mariposas butterßies. Am. Nat. 135: 748Ð765. Neotropicales. m3m-Monografõ´as Tercer Milenio, Vol- Conner, W. E. (ed.). 2009. Tiger and woolly bears: ume 8. Sociedad Entomolo´gica Aragonesa (SEA)/Red behavior, ecology and evolution of the Arctiidae. Oxford Iberoamericana de Biogeografõ´a y Entomologõ´a Sis- University Press, Oxford, United Kingdom. tema´tica (RIBES)/Ciencia y Tecnologõ´a para el Desar- Dempster, J. P. 1984. The natural enemies of butterßies, pp. rollo (CYTED)/Natural History Museum, London, 97Ð104. In R. I. Vane-Wright and P. R. Ackery (eds.), The United Kingdom, (NHM)/Instituto Venezolano de In- biology of butterßies. Academic, London, United King- vestigaciones Cientõ´Þcas, Venezuela (IVIC), Zaragoza, dom. Spain. Deschamps-Cottin, M., and H. Descimon. 1996. A possible Berenbaum, M. R. 1995. Aposematism and mimicry in cat- case of mimicry in the larva of Parnassius apollo (L.) erpillars. J. Lepid. Soc. 49: 386Ð396. (Lepidoptera: Papilionidae). Linn. Belg. 15: 309Ð310. November 2011 WILLMOTT ET AL.: MIMICRY IN DANAINE CATERPILLARS 1117

DeVries, P. J. 1987. The butterßies of Costa Rica and their Leimar, O., M. Enquist, and B. Sille´n-Tullberg. 1986. Evo- natural history. Papilionidae, Pieridae, Nymphalidae. lutionary stability of aposematic coloration and prey un- Princeton University Press, Princeton, NJ. proÞtability: a theoretical analysis. Am. Nat. 128: 469Ð490. DeVries, P. J., R. Lande, and D. Murray. 1999. Associations Llorente, J. E., C. Pozo, and A. M. Luis. 1993. of co-mimetic ithomiine butterßies on small spatial and thirza (Lepidoptera: Nymphalidae): Su ciclo de vida y temporal scales in a Neotropical rainforest. Biol. J. Linn. distribucio´n. Publ. Esp. Mus. Zool. Univ. Nac. Aut. Me´xico Soc. 67: 73Ð85. 7: 63Ð87. Dyer, L. A. 1995. Tasty generalists and nasty specialists? Mallarino, R., E. Bermingham, K. R. Willmott, A. Whinnett, Antipredator mechanisms in tropical lepidopteran larvae. and C. Jiggins. 2005. Molecular systematics of the but- Ecology 76: 1483Ð1496. terßy genus Ithomia (Lepidoptera: Ithomiinae): a com- Elias, M., Z. Gompert, C. Jiggins, and K. R. Willmott. 2008. posite phylogenetic hypothesis based on seven genes. Mutualistic interactions drive ecological niche conver- Mol. Phylogenet. Evol. 34: 625Ð644. gence in a diverse butterßy community. PLoS Biol. 6: Mallet, J., and N. H. Barton. 1989. Strong natural selection 2642Ð2649. in a warning-colour hybrid zone. Evolution 43: 421Ð431. Ford, E. B. 1936. The genetics of Papilio dardanus Brown Mallet, J., and L. E. Gilbert. 1995. Why are there so many mimicry rings? Correlations between habitat, behaviour (Lep.). Trans. R. Entomol. Soc. Lond. 85: 435Ð466. and mimicry in Heliconius butterßies. Biol. J. Linn. Soc. 55: Freitas, A.V.L., J. R. Trigo, K. S. Brown, L. Witte, T. Hart- 159Ð180. mann, and L.E.S. Barata. 1996. Tropane and pyrroliz- Mallet, J., and J. T. Longino. 1982. Hostplant records and idine alkaloids in the ithomiines Placidula euryanassa and descriptions of juvenile stages for two rare species of Miraleria cymothoe (Lepidoptera: Nymphalidae). Che- Eueides (Nymphalidae). J. Lepid. Soc. 36: 136Ð142. moecology 7: 61Ð67. Mallet, J., and M. C. Singer. 1987. Individual selection, kin Futahashi, R., and H. Fujiwara. 2008. Juvenile hormone reg- selection, and the shifting balance in the evolution of ulates butterßy larval pattern switches. Science 319: 1061. warning colours: the evidence from butterßies. Biol. J. Gamberale, G., and B. S. Tullberg. 1996. Evidence for an Linn. Soc. 32: 337Ð350. increased aposematic effect through aggregation. Anim. Massuda, K. F., and J. R. Trigo. 2009. Chemical defence of Behav. 52: 597Ð601. the warningly coloured caterpillars of Gentry, G. L., and L. A. Dyer. 2002. On the conditional (Lepidoptera: Nymphalidae: Ithomiinae). Eur. J. Ento- nature of Neotropical caterpillar defenses against their mol. 106: 253Ð259. natural enemies. Ecology 83: 3108Ð3119. Mols, C.M.M., and M. E. Visser. 2002. Great tits can reduce Higginson, A. D., and G. D. Ruxton. 2010. Optimal defen- caterpillar damage in apple orchards. J. Appl. Ecol. 39: sive coloration strategies during the growth period of 888Ð899. prey. Evolution 64: 53Ð67. Mu¨ ller, F. 1879. Ituna and Thyridia; a remarkable case of Hill, R. I. 2006. Life history and biology of Forbestra oliven- mimicry in butterßies. Trans. Entomol. Soc. Lond. 1879: cia (Bates, 1862) (Nymphalidae, Ithomiinae). J. Lepid. xxÐxxix. Soc. 60: 203Ð210. Murphy, S. M. 2004. Enemy-free space maintains swallow- Hill, R. I. 2009. Habitat segregation among mimetic ithomi- tail butterßy host shift. Proc. Natl. Acad. Sci. U.S.A. 101: ine butterßies (Nymphalidae). Evol. Ecol. 24: 273Ð285. 18048Ð18052. Hill, R. I., and L. Tipan. 2008. Description of the immature Murphy, S. M., S. M. Leahy, L. S. Williams, and J. T. Lill. stages of confusa Butler, 1873 and Me- 2010. Stinging spines protect slug caterpillars (Limaco- thona curvifascia Weymer, 1883 (Nymphalidae, Ithomii- didae) from multiple generalist predators. Behav. Ecol. nae) from eastern Ecuador. J. Lepid. Soc. 62: 89Ð98. 21: 153Ð160. Holmes, R. T. 1990. Ecological and evolutionary impacts of Nentwig, W. 1985. A tropical caterpillar that mimics feces, bird predation on forest : an overview. Stud. Avian leaves and a snake (Lepidoptera: Oxytenidae: Oxytenis Biol. 13: 6Ð13. naemia). J. Res. Lepid. 24: 136Ð141. Hooks, C.R.R., R. R. Pandey, and M. W. Johnson. 2003. Im- Opitz, S.E.W., and C. Mu¨ ller. 2009. Plant chemistry and pact of avian and predation on lepidopteran insect sequestration. Chemoecology 19: 117Ð154. caterpillar densities and plant productivity in an ephem- Pelzer, A. 1992. Mimicry in the larva of the pine hawkmoth Hyloicus pinastri (Lepidoptera: Sphingidae)? Nota Lepi- eral agroecosystem. Ecol. Entomol. 28: 522Ð532. dopterologica 15: 148Ð151. Janzen, D. H., and W. Hallwachs. 2009. Dynamic database Pierce, N. E., M. F. Braby, A. Heath, D. J. Lohman, J. Mathew, for an inventory of the macrocaterpillar fauna, and its D. B. Rand, and M. A. Travassos. 2002. The ecology and food plants and parasitoids, of Area de Conservacion evolution of ant association in the Lycaenidae (Lepidop- Guanacaste (ACG), northwestern Costa Rica (nn-SRNP- tera). Annu. Rev. Entomol. 47: 733Ð771. nnnnn voucher codes). (http://janzen.sas.upenn.edu). Pinheiro, C.E.G. 1996. Palatability and escaping ability in Janzen, D., W. Hallwachs, and J. M. Burns. 2010. A tropical Neotropical butterßies: tests with wild kingbirds (Tyran- horde of counterfeit predator eyes. Proc. Natl. Acad. Sci. nus melancholicus, Tyrannidae). Biol. J. Linn. Soc. 59: U.S.A. 107: 11659Ð11665. 351Ð365. Kapan, D. 2001. Three-butterßy system provides a Þeld test Poole, R. W. 1970. Convergent evolution in the larvae of two of Mu¨ llerian mimicry. Nature 409: 338Ð340. Penstemon-feeding geometrids (Lepidoptera: Geometridae). Lamas, G. 2004. Nymphalidae. Danainae, pp. 171Ð172. In G. J. Kans. Entomol. Soc. 43: 292Ð297. Lamas (ed.), Checklist: Part 4A. Hesperioidea-Papilion- Portugal, A.H.A., and J. R. Trigo. 2005. Similarity of cutic- oidea. In J. B. Heppner (ed.), Atlas of Neotropical Lep- ular lipids between a caterpillar and its host plant: a way idoptera. Vol. 5A. Association for Tropical Lepidoptera; to make prey undetectable for predatory ants? J. Chem. ScientiÞc Publishers, Gainesville, FL. Ecol. 31: 2551Ð2561. Langham, G. M. 2004. Specialized avian predators repeat- Rico-Gray, V., and P. S. Oliveira. 2007. The ecology and edly attack novel color morphs of Heliconius butterßies. evolution of ant-plant interactions. The University of Chi- Evolution 58: 2783Ð2787. cago Press, Chicago, IL. 1118 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 104, no. 6

Sanz, J. J. 2001. Experimentally increased insectivorous bird Athalia rosae (Hymenoptera: Tenthredinidae), against density results in a reduction of caterpillar density and predation by lizards. Bull. Entomol. Res. 94: 283Ð289. leaf damage to Pyrenean oak. Ecol. Res. 16: 387Ð394. Wagner, D. L. 2005. Caterpillars of eastern North America: Sheppard, P. M., J.R.G. Turner, K. S. Brown, W. W. Benson, a guide to identiÞcation and natural history. Princeton and M. C. Singer. 1985. Genetics and the evolution of University Press, Princeton, NJ. Muellerian mimicry in Heliconius butterßies. Philos. Wahlberg, N., J. Leneveu, U. Kodandaramaiah, C. Pen˜ a, S. Trans. R. Soc. Lond. 308: 433Ð613. Nylin, A.V.L. Freitas, and A. Brower. 2009. Nymphalid Sille´n-Tullberg, B. 1988. Evolution of aposematic butterßy butterßies diversify following near demise at the Creta- larvae: a phylogenetic analysis. Evolution 42: 293Ð305. ceous/Tertiary boundary. Proc. R. Soc. B. doi: http:// Sille´n-Tullberg, B., and O. Leimar. 1988. The evolution of dx.doi.org/10.1098/rspb.2009.1303. gregariousness in distasteful insects as a defense against Warren, A. D., K. J. Davis, N. V. Grishin, J. P. Pelham, and predators. Am. Nat. 132: 723Ð734. E. M. Stangeland. 2010. Interactive listing of American Smith, D. S., L. D. Miller, and J. Y. Miller. 1994. The but- butterßies. (http://www.butterßiesofamerica.com). terßies of the West Indies and South Florida. Oxford Wiklund, C., and B. Sille´n-Tullberg. 1985. Why distasteful University Press, Oxford, United Kingdom. butterßies have aposematic larvae and adults, but cryptic Sourakov, A., and T. C. Emmel. 1996. Notes on the life pupae: evidence from predation experiments on the mon- history of Anetia jaegeri (Lepidoptera: Nymphalidae: Da- arch and the European swallowtail. Evolution 39: 1155Ð nainae) from Hispaniola. Trop. Lepid. 7: 155Ð159. 1158. Willmott, K. R., and A.V.L. Freitas. 2006. Higher-level phy- Stefanescu, C., B. Pintureau, H.-P. Tschorsnig, and J. Pujade- logeny of the Ithomiinae (Lepidoptera: Nymphalidae): Villar. 2003. The parasitoid complex of the butterßy classiÞcation, patterns of larval hostplant colonisation and Iphiclides podalirius feisthamelii (Lepidoptera: Papilion- diversiÞcation. Cladistics 22: 297Ð368. idae) in north-east Spain. J. Nat. Hist. 37: 379Ð396. Willmott, K. R., and G. Lamas. 2006. A phylogenetic reas- Trigo, J. R. 2000. The chemistry of antipredator defence by sessment of Hyalenna Forbes and Doubleday, secondary compounds in Neotropical Lepidoptera: facts, with a revision of Hyalenna (Lepidoptera: Nymphalidae: perspectives and caveats. J. Braz. Chem. Soc. 11: 551Ð561. Ithomiinae). Syst. Entomol. 31: 419Ð468. Trigo, J. R., and K. S. Brown. 1990. Variation of pyrrolizidine Willmott, K. R., and G. Lamas. 2008. A revision of the genus alkaloids in Ithomiinae: a comparative study between Megoleria (Lepidoptera, Nymphalidae, Ithomiinae). species feeding on Apocynaceae and Solanaceae. Che- Trop. Lepid. Res. 18: 46Ð59. moecology 1: 22Ð29. Willmott, K. R., and J. Mallet. 2004. Correlations between Trigo, J. R., K. S. Brown, L. Witte, T. Hartmann, L. Ernst, and adult mimicry and larval hostplants in ithomiine butter- L.E.S. Barata. 1996. Pyrrolizidine alkaloids: different ac- ßies. Proc. R. Soc. Lond. B (Biol. Lett. Suppl.). 271: S266Ð quisition and use patterns in Apocynaceae and So- S269. lanaceae feeding ithomiine butterßies (Lepidoptera: Young, A. M. 1977. Notes on the biology of Hypothyris eu- Nymphalidae). Biol. J. Linn. Soc. 58: 99Ð123. clea in Costa Rica. Pan-Pac. Entomol. 53: 104Ð113. Van Bael, S. A., P. Bichier, and R. Greenberg. 2007. Bird Zagatti, P., B. Lalanne-Cassou, and J. D. d’Aubigny. 2010. predation on insects reduces damage to the foliage of Catalogue of the Lepidoptera of the French Antilles. cocoa trees (Theobroma cacao) in western Panama. J. Institut National de la Recherche Agronomique. (http:// Trop. Ecol. 23: 715Ð719. www.inra.fr/papillon/indexeng.htm). Vlieger, L., P. M. Brakefield, and C. Mu¨ ller. 2004. Effec- tiveness of the defence mechanism of the turnip sawßy, Received 27 May 2010; accepted 21 October 2010.