Zoological Journal of the Linnean Society, 58: 297-308. With 1 figure June 1976 Adaptive radiation and convergence in subdivisions of the butterfly genus Heliconius (Lepidoptera: Nymphalidae) JOHN R . G . TURNER Department o j Ecology and Evolution, Division ofBiologica1 Sciences, State University o f N e w York, Stony Brook, New York 11 794, U.S.A.* Accepted f o r publication July I975 The process of adaptive radiation and convergence, usually regarded as a feature of macro-evolution, can be seen in the mimetic colour patterns of the butterflies within the confines of the South American genus Heliconius. This can be shown by dividing the genus into subgroups on the basis of adult, pupal and larval morphology: the theory that the mimicry between species results solely from close systematic relationships is thereby refuted, as members of the same morphological group can display widely divergent mimetic patterns, and conversely mutual mimics may belong t o several different morphological groups. Various forms of parallel and convergent evolution are thought t o account for the present pattern of mimicry; the process is known to start even before full speciation has taken place. A new subgenus ( N e m d a ) is created t o contain three atypical members of the genus. CONTENTS . . . . Introduction History . . . . . Subgenera of Heliconius Neruda subgen. nov. Discussion . . . . Summary . . . . . Acknowledgements . . Note added in proof . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297 298 299 301 303 307 307 307 307 INTRODUCTION The hypothesis that evolution in the long term, that is over times exceeding a few human generations, is governed by the same forces that are observable in experiments and by field observation, is not open to direct experimental test. But it would be open to more direct testing than is usually believed, if a group Contribution No. 109 from the Program in Ecology and Evolution at the State University of New York at Stony Brook. 297 298 J . R. G. TURNER of organisms could be found in which some aspect of meso- or macro-evolution had happened sufficiently rapidly to be seen within the confines of one genus, within which artificial hybridization of geographical races, if not of full species, was possible. A still more direct approach could be used if the characteristics under study were of a known function (or demonstrable lack of it). Muellerian mimicry in some South American butterflies fulfills these criteria. The function of the mimicry is known to be mutual protection of the mimetic species against their predators, because all are unpalatable and by sharing a common pattern they reduce the load of predation, necessary for educating and for reeducating predators, on all the species in the mimicry ring. The adaptive value of mimicry has been demonstrated in many experiments (Rettenmeyer, 1970 for review). Within the genus Heliconius the mimetic patterns of the butterflies have undergone such rapid evolution that they show the pattern of adaptive radiation of evolutionary lines, with the concomittant functional convergence of members of different lines, which is usually thought of as characteristic of evolution in the long term. The whole process of race formation, speciation and adaptive radiation can be observed within the boundaries of this one genus. Much is now known about the dynamics and genetics of the beginning of the process, the formation of divergent mimetic races in a single species (Sheppard, Turner, Brown & Benson, 1976; Turner, 1971, 1976). T o understand the process as a whole, especially above the species level, we must first have a good understanding of the systematics of this very difficult genus. In particular, it is necessary to establish the natural subdivisions of the genus, in the hope that these are, to some degree of approximation, groups of phyletically related species. This paper formally defines one rather distinct group, in the context of the process of adaptive radiation in the genus. HISTORY Mimicry between species within the same genus was not often noted by early students of batesian and muellerian mimicry: it would hardly do to support a controversial theory by citing resemblances that could be the result merely of common ancestry, and the most convincing examples of mimicry are those between different orders of insects, or between different kingdoms (resemblance of insects to twigs, or plants t o stones). But mimicry can in theory occur between related species: the criterion is not whether the resemblance results from convergence or from common ancestry, but whether or not it protects the mimetic species. In the genus Heliconius there is considerable muellerian mimicry between species. Long ago Eltringham (1917) realized that this showed a curious regularity: he divided the genus morphologically into two subgroups, and found that the colour pattern of each species in the first subgroup could be matched by a “mimic” in the second. Except for talking about “models” and “mimics”, when it is clear that, because all tested species in the genus are distasteful to birds (Brower, Brower & Collins, 1963), one has pairs of muellerian comimics, neither of which can be regarded as the model, Eltringham had correctly described the situation. But it is rather more complicated than he thought, as some of his “species” in each subgroup were in ADAPTIVE RADIATION I N HELICONIUS 299 fact races of m o very widespread polytypic species now called Heliconius inelpomene and H. erato, and more detailed taxonomic studies have made it apparent that the genus can be divided into more than two subgroups (Emsley, 1963 ; Turner, 1968). Eltringham’s conclusions were disputed by Beebe, Crane & Fleming (1960), who claimed from a systematic study of the early stages of the fourteen Trinidadian species that the convergent or parallel mimics were in fact rather closely related, and therefore resembled one another simply by common ancestry . SUBGENERA OF HEI.IC0NIUS The genus Heliconius as recognized by Michener (1942) and Emsley (1965) is a convenient unit for the general biologist. A good naturalist, according to Linnaeus, knows all his genera (Cain, 1958); the large genus Heliconius is recognizable to the butterfly enthusiast in the same way that the genus Morpho is recognizable. Further subdivision becomes necessary for the study of evolutionary patterns within the genus, and is also useful to the specialist, particularly in view of the great use now made of the genus for studies in many branches of biology (references in Turner, 1971; more recent studies include Brown & Mielke, 1972; Brown & Benson, 1974; Ehrlich & Gilbert, 1973). Small members of th genus, with short antennae, have long been placed in Eueides, regarded by some as a subgenus (Michener, 1942; Emsley, 1965; Turner, 1968), by others as a full genus (Brown & Mielke, 1972; Brown & Holzinger, 1973; Neustetter, 1929; Stichel & Riffarth, 1905); assignment to this group has been correct, except for the rather small Heliconius species H. ricini and H. demeter which have often been erroneously placed in Eueides, the latter under a bewildering variety of synonyms (Turner, 1966; Brown & Benson, 197 5 ) . Recent studies of internal anatomy, karyotypes, chemistry, life-history and behaviour have further defined the differences between these groups, summarized in Table 1. The remaining Heliconius can be split into a number of clusters of related species, at least one of which is rather distinct. I t was initially noticed that Heliconius aoede, while showing an adult morphology similar to that of other Heliconius, had a pupa which was unlike other members of the subgenus and rather similar to the pupae of Eueides, and a larva that was unique among the known Heliconius and Eueides in lacking scoli on the head (Turner, 1968). Further studies of this species by K. S. Brown, Jr indicate that while much of its adult morphology corresponds with Heliconius, its early stages tend much more to Eueides (Brown & Holzinger, 1973; Brown, 1973); de Lesse (1970) found the haploid chromosome number to be 23, which lies between the number 21, which is found in the great majority of the subgenus Heliconius, and of 3 1 and above, normal for Eueides (Suomalainen, Cook & Turner, 1972). Brown reports that the small white egg is laid in clusters, apparently on Dilkea, an aberrant member of the Passifloraceae; this and further habits of the adult are described by Brown (1973). On these grounds a new subgenus is here created for Heliconius aoede, there being no name of generic rank available. From their general appearance and the 3 00 J . R. G . TURNER Table 1. Characteristics differentiating three subgenera of Heliconius. Adapted from Brown & Holzinger (197 3)’ with additional information from Brown (1973) ~ Character Segments in female foretarsi Eueides Neruda Heliconius four five five Spermatheca duct narrow broad broad Signa of bursa copulatrix usually bent at a right angle or more not bent at a sharp angle usually bent at a right angle or less, or absent Antenna length compared with iforewing radius less less more Haploid chromosome number over 30 21-32 21 in almost all species (except the sapho group) 3-Hydroxykynurenine as yellow pigment absent present present Egg (usually) small, white small, white large, yellow Head pattern of mature larva black, or with stripes black black or yellow, rarely striped Scoli on head of larva present absent present Chrysalis hangs horizontally horizontally vertically Pollen-feeding (imago) absent, but feeds on bird-droppings absent present Eggs laid (usually) ventral surface of older leaves growing bud or young leaves growing bud or young leaves close similarity of their internal anatomy to aoede (Emsley, 1965) it is clear that H. metharme and H. godmani should be included in the subgenus, but as no studies have been published in their early stages, the diagnosis is based only on aoede. D’Almeida (1937) describes H. metharme as showing a curious nervous behaviour when in flight, rapidly fleeing the approach of strange objects. Recent work indicates that the haploid chromosome number in the three species ranges from 21 (most races of aoede) to 32 (Brown, Emmel & Soumalainen, unpubl.). Linnaeus created Heliconius as a subsection of Papilio, to contain delicate butterflies not suited as were the Danaids and Swallowtails, to being named after the combatants from the Iliad; they were mainly named after muses, graces and female deities (melpomene, egeria. vesta), Heliconius signifying dwellers on Mount Helicon, and the group originally contained Phoibos Apollon Musagetes himself in the butterfly now known as Parnassius apollo (Papilionidae). Later authors have continued the association with poetry and the arts by using names like lucretius, isolda, elsa, elvira, and aida. In continuance of this tradition, and recalling that the butterflies are South American, the new subgenus is named after the author of Alturas de Macchu Picchu. Veinte Poemas de Amor, Residencia en la Tierra etc. Senor Pablo ADAPTIVl? RADIATION I N HELICONICJS 301 Neruda graciously consented to the use of his name, although sadly he did not live to see this paper written. * Nemda subgen. nov. Heliconius butterflies characterized by a singular wingshape, particularly the broad triangular forewings, with very extensive friction patches in the male (Fig. 1) (females having wings of more normal shape for their genus), by antennae about half the length of the forewing, by a broad duct on the spermatheca and the absence of a sharp angle in the signa of the bursa copulatrix. 3-hydroxy-kynurenine is used as a yellow pigment, and the haploid chromosome number is 21 to 32. Figure 1. Silhouette of a male Heliconius aoede to show characteristic wing shape and unusually large friction patches. (Drawing by Joyce Roe.) The eggs are small, white and laid in clusters. The larva lacks scoli on the head, and the pupa is bowed at the rear end of the abdomen, so that it rests horizontally beneath its support; it lacks prominent head appendages, flanges on the body, antenna1 spines and gold spots; the abdominal spines are slender and arise from tubercles. Type species. Migonitis aoede Hubner, 1816 (Verz. bekannt. Schmett., 12). Needless to say, the gender of the new subgenus is masculine, despite its ending. The rest of genus Heliconius can be divided into subgroups recognizable by adult morphology and, in so far as this is known, by details in the morphology * No volveris del fondo de las rocas. N o volveris del tiempo subterrineo. No volveri N voz endurecida. N o volverin tus ojos taladrados . . . Per0 una permanencia de piedra y de palabra: la ciudad como un vaso se levant6 en las manos de todos, vivos, muertos, callados, sostenidos de tanta muerte, un muro, de tanta vida un golpe de petalos de piedra: la rosa permanente, la morada: este arrecife andino de colonias glaciales . . . Hablad por mis palabras y mi sangre. Alturas de Macchu Picchu 302 J. R. C. TURNER of the pupa. From what is known at present, the only one of these groups sufficiently distinct to be a candidate for subgeneric rank is the single species Heliconius doris, which has a pupa lacking the gold spots, flanges and well developed antenna1 spines of the other species, and which has the distinction of being the only species with a marked colour polymorphism as an adult, the only species with a marked morphological polymorphism as a pupa, and the only species, apart from H. (N.) metharme, to produce blue colour not by iridescence but by laying white scales over black. Its variable chromosome number lies between Neruda and Eueides (Suomalainen, Cook & Turner, 1972). The tentative separation of this species into a separate subgenus Laparus is maintained in this paper (Turner, 1968). In summary, the subgenera, which for the present purposes are adequately differentiated by Fig. 1 , Table 1 and the notes in the text, are: Heliconius Kluk, 1780 Type species Papilio charitonia Linnaeus, 1767 designated by Hemming, 1933 (Entomologist, 66: 223). Eueides Hiibner, 18 16 Type species Papilio isabella Cramer, 1781 through its subspecies Nereis fulva dianasa Hubner, 1816, designated by Scudder, 1875 (Proc. Am. Acad. Arts Sci., 10: 164). Laparus Billberg, 1820 Type species Papilio doris Linnaeus, 1771 designated by Hemming, 1934 (Entomologist, 67: 37). Neruda subgen. nov. Type species Migonitis aoede Hiibner, 1816 Although one of these names is frequently used as a separate genus, it is clear that some authorities will continue to treat the whole group as a single genus; it is important that in creating new names for species and races homonyms should not be made within Heliconius in the broad sense, as changes in the generic classification will make the nomenclature at the species level unstable. This situation already exists for the nine names listed in Table 2. Replacement names are not proposed, and should not be proposed until the infraspecific Table 2. Homonyms created by the fusion of Meliconius and Eueides (for citations see Neustetter, 1929). Attribution of subspecies t o species follows modern practice, not necessarily the original combination Subgenus Heliconius or Neruda Subgenus Eueides clysonymus apicalis J & K,1917' melpomene ecuadorensis Neustetter, 1908. erato fuliginosa Riffarth, 1907 numata f. gracilis Riffarth, 1907 burneyi huebneri Staudinger, 1896 aoede lucretius Weymer, 1890. melpomene f . Pluto Staudinger, 1896. melpomene f. riffarthi Stichel 1906 [melpomene x cydnol seitzi Neustetter, 1916 lampeto apicalis (Rober, 1927) isabella ecuadorensis (Strand, 1914) lampeto fuliginosa (Stichel, 1903). aliphera gracilis (Stichel, 1903). isabella huebneri (MCnCtrits, 1857)' tales f. lucretius (Zikan, 1937) eanes f. Pluto (Stichel, 1903) eanes f. riffarthi Stichel, 1903). isabella seitzi (Stichel, 1903). * Senior homonym ADAPTIVE RADIATION I N HELICONIUS 303 classification of the whole genus has been overhauled. For all names except possibly ecuadorensis or hrtebneri, at least one of each pair will certainly turn out t o be an infrasubspecific variety, an interracial hybrid, or an interspecies hybrid, and less confusion will ensue from keeping the names, appropriately sunk, than by replacing them. Although one of the two uses of ecuadorensis and hucbneri might require a replacement under strict application of the Rules, neither will ever be raised t o full species, and in practice there will be no chance of confusing them. Within the larger subgenera (Eueides and Hdiconius) it is possible t o recognize further clusters of species. I t is not recommended that these be named, even when the subgenera are given full generic rank, as there are too many of them for Latin names to be remembered by any but the most ardent specialist; they can conveniently be referred to by the name of the most prominent species (as erato-group, rnelpornene-group), or as in this paper by numbers. DISCUSSION Once the genus Heliconius has been divided into its subgenera and further subgroups, shown vertically in Table 3 , it is possible to demonstrate very clearly the pattern of adaptive radiation and convergence mentioned in the introduction by tabulating the colour patterns of the species (horizontally in Table 3). If morphology and colour pattern had followed the same evolutionary course and related species had similar colours because of common ancestry, then few cells of the table could be filled with the names of species. For instance all the blue species should be in a single morphological group. As it is reasonable to suppose that the mimetic colour patterns are much more likely to undergo convergent evolution, and more rapidly, than are morphological characters, this table shows that there has been considerable adaptive radiation of the colour patterns within each morphological group. Thus the three members of the subgenus Neruda, presumably derived from a single common ancestor, have radiated to join three mimicry rings (as groups of species with similar patterns are called), the red and yellow (radiate) ring, the blue and yellow ring, and one of the rings which mimic various Ithomiid butterflies. I t has been argued that this process of radiation and convergence starts at the earliest possible stage in evolution, during the formation of allopatric races (Sheppard e t al., 1976; Turner, 1976). The beginning of the process is shown by pairs of species that have widely divergent races within each species, each race of the one exactly matching a sympatric race of the other; H. melpomene and H. eruto are the best example of this, and their parallel variation is partly shown in Table 3 by their each appearing in two columns (for a colour illustration, see Turner, 1975). The force causing divergence of the races is held t o be considerable change in the faunas of the forest “islands” to which the species were confined during the last glaciation (Haffer, 1969), causing the species t o “switch” from their original pattern to a new one mimicking the most abundant or distasteful local species (Brown, Sheppard & Turner, 1974). As race formation can presubably lead t o species formation, it is held that the parallel series of species shown in Table 3 are the result of repeated cycles of 304 J . R. G . TURNER Table 3 . Classification of some Heliconius species by morphology anc Orange and yellow Red, sometimes (radiate)t yellow barst Subgroup Red (blotches) and yellow Blue and yellow procula (Venezuela) tales (Colombia) eanes (black) Eueides tales eanes (red) - Neruda aoede’ - - metharme Laparus doris (red)* - doris (red) doris (blue)* la erato‘ era to * hermathenu (Faro) hortense‘ ClysOnymUS*‘ leucadia Sara * sapho ( E . Andes) Ib demeter’ - ricini* [la me/pomene*‘ timareta (form) melpomene* timareta’ (form) timareta (form) Ilb elevatus hesc kei * - - 111 xanthocles* - - IV hurneyi. egeria astraea - wallacei* Heliconius ’ - Notes. This table includes, o u t of around 54 species of Heliconius (sensu Into) now recognized (Brown & Mielke, 1972), 31 species whose systematic position is well established, preferably with descriptions of the early stages (marked *) (Beebe, Crane & Fleming, 1960;Turner, 1968; Brown, 1972; Brown & Mielke, 1972; Young, 1973; Brown & Holzinger, 1973; Brown & Benson, 1975a, 1975b; Turner, unpubl. for H. atthis). Addition of the remaining species would n o doubt introduce subgroups intermediate between those shown, which are rather distinct, but would not alter the general conclusion. The pattern classes are also rather distinct, but d o embrace racial variation, usually in parallel, of the included species; this applies particularly t o those marked t. The “blue ane white” and “black and white” groups contain races with the white changed t o yellow, and some of the “blue and yellow” lack the blue iridescence. The groups marked $ are heterogeneous and include several different patterns. Where the same species appears in several pattern groups it has widespread races (or morphs) belonging t o all of them; locations are given only for very restricted races. glaciation during the quaternary, the earlier cycles having now led to full speciation. In that event, there are several patterns of evolution which could account for the arrangement of species in Table 3. First, parallel evolution. Two species, initially muellerian comimics either by common ancestry or convergence remain comimics during several cycles of divergent race formation and speciation, thus “budding” several pairs of mimetic species. There is genetical evidence that the parallel variation between Heliconius groups Ia and IIa has arisen in this way (Sheppard et al., 1976). Second, convergence. A species changes its pattern from one mimicry ring to another. This may produce empty cells in the table (an effect produced also by extinction), or cause several species to occupy the same celI, or cause a column (pattern) to disappear altogether. Brown & Benson (1976) describe it at the race level in H. hermathena and the non-mimetic species H, charitonia has apparently changed ADAPTIVE RADIATION IN HELICONIUS 305 colour pattern (revised and expanded from Turner, 197 1 , 1976) Blue and white Black and white Tiger ithomiinet Other ithomiine Black and or acrarineg yellow A isa bella lampeto procula ' pavana' vibilia2 - godnianih Black and yellow B$ - - - eleuchia eraro (Cauca) sapho antiochus sapho (Ocafia) - peruviana'*' - - - - cydno. - cydno' atthis. pachinus nutrererid" ' - luciana (Roraima) ethilla' numata nattereriQ hecale. - - luciana ' (Amazonas, Venezuela) ~ doris' (green) doris' (green) hewitsoni sara (Chiriqui) charitonia" hermathena*' - ' Mutually allopatric, but mostly sympatric with red doris. ""4Pairs of mimics within this heterogenous class of patterns. Mimic Actinore, and mimic Elzunia. 'All allopatric, except for charitonia and green doris. There are grounds for regarding this as the primitive pattern for groups I and I1 (see text). Life-history not published; known for two races to be identical with the "red sometimes yellow bars" races. 'Mr Gordon Small of Balboa tells me that within the rather confined distribution of hewitsoni and pachinus the green form of doris becomes much yellower, and is consequently an adequate mimic of the other two when in flight (also Benson, 1971). 'Both mimics of H. hecalesia and the associated ithomiines. Usually regarded as a subspecies of charitonia but shown by recent fieldwork by the author to be sympatric, apparently without hybrids, with typical charironia about 800 m u p the Andes above Santo Doming0 de 10s Colorados, Ecuador. ' to join an Ithomiid ring in western Ecuador, speciating in the process (H. peruuiunu). Third, reciprocal convergence. Suppose that the ancestral species of, say, Nerudu, had a red and yellow pattern, and that the ancestral species of some other group had a blue and yellow pattern. If Nerudu became the dominant species in one forest "island" it would cause the other to switch to the red and yellow ring, and if the reverse were the case in some other "island" Nerudu would there switch to the pattern of the dominant blue and yellow species. Speciation would produce parallel pairs of species. Again, there are no proven examples, but this or the effect before it must have taken place, as the patterns of warning coloured species are subject to strong normalizing selection (Benson, 1972) and can only change from one mimicry ring to another in this way (Sheppard et ul., 1976). Fourth, allopatric speciation without divergence of pattern. I t is to be expected that in the absence of selection to switch to another mimicry ring, or 22 3 06 J . R. C.. TURNER in the absence of the appropriate mutations, allopatric populations of a species will retain the same pattern, and while it is clear from melpomene and erato that the beginning of adaptive radiation may precede speciation, it is also extremely likely, unless alteration of the colour pattern is the only possible isolating mechanism in these butterflies (and the existence of sympatric and similarly coloured relatives refutes that suggestion), that speciation may proceed without the divergence of pattern. Thus hortense and clysonyrnus are allopatric relatives in group I, which should probably be regarded as good species, with basically the same pattern. If they became sympatric, they would have to be regarded as comimics by retention of an ancestral pattern. The closely similar sympatric sibling pair, melpomene and elevatus may be an example of this. Fifth, speciation with slight divergence, followed by reconvergence. If two mimicry rings are so similar that predators sometimes mistake one for the other, the rings will simply slowly converge to a common pattern (Sheppard et al., 1976). Thus allopatric speciation followed by sympatry could involve slight divergence of the patterns, which would then reconverge. The morphologically similar sympatric pair erato and demeter may be a case of this. Sixth, one might see non-mimetic resemblance between two not very closely related and allopatric species. This occurs in Table 3 in the last column (black and yellow) and is held by Brown (1972) and Sheppard et al. (1976) to show that the species have retained the pattern by common ancestry from the remote past (as non-mimetic patterns will evolve more slowly than mimetic ones) which means that this pattern is close to that of the common ancestor of groups I and 11, and possibly of the whole subgenus Heliconius. A similar consideration of the subgenus Eueides suggests that the ancestral pattern may have been of black and orange longitudinal stripes (not listed in Table 3), now exhibited by the largely allopatric species lybiu, emsleyi and libitina (for the last two, see Brown, 1975). Interestingly enough, this striping has a similar distribution on the wings to the yellow striping of the supposed ancestral pattern of the subgenus Heliconius, and is very similar to the orange striped pattern of the related non-mimetic species Dryudula phaetusu. Reconstructing the actual course of evolution will of course be extremely difficult, if not totally impossible, but it is now clear that the evolution of mimicry in the group follows a pattern which is neither exactly that proposed by Eltringham (19 17)-convergence or parallelism between two subsections of the genus-nor that of Beebe, Crane & Fleming (1960), who thought that the resemblances mainly resulted from close systematic relationships. As is often the case with controversies, the truth is more complicated and lies between the two extremes. There is in fact extensive parallelism and convergence, but it takes place between many more than two subdivisions, and plainly arises both from convergence and from the retention of ancestral patterns. In fact the most spectacular case of parallelism, that of melpomene, erato and their associated species, results from parallel evolution between two separate, speciating evolutionary lines, which may originally have been mimics by common ancestry; such a phenomenon fits neatly into neither of the conventional categories “convergence” and “taxonomic propinquity.” ADAPTIVE RADIATION IN HELZCONIUS 307 SUMMARY The genus Heliconius can be subdivided into around four rather distinct subgenera, one of which is named in this paper, and then into a number of other groups which have not been and should not be formally named. The groups are distinguished by the morphology of the adults and of the early stages, and by behaviour. Subdivision of the genus in this way shows that groups of related species have undergone wide adaptive radiation in their colour patterns, converging with members of other groups under selection for muellerian mimicry. Reciprocal convergence between two species in different parts of their range, and parallel evolution between two already mimetic species in the period in which they themselves speciate, will produce series of paired mimetic species such as are found in the genus. Speciation with little divergence of pattern, or reconvergence of pattern, will produce closely related mimetic species. Resemblance between non-mimetic species which are allopatric is likely to indicate that they have retained the pattern from a common ancestor. The beginning of the radiation of colour pattern can be seen in some species in the divergence of allopatric races; Heliconius thus presents the process of evolution from race formation to adaptive radiation, compressed within the confines of a single genus. ACKNOWLEDGEMENTS I am very grateful t o Mr N. D. Riley for searching for the existence of prior uses of the name Nerudu at the generic level, and to Dr K. S. Brown, Jr, Professor A. J. Cain and Professor P. M. Sheppard F.R.S. for criticism and for correspondence on topics related t o this paper. I gratefully acknowledge financial support in the form of seed-grants from U.S. Public Health Service Biomedical Sciences Support Grant 5 S05RR07067-08 (awarded t o the State University of New York at Stony Brook), the Graduate School of the State University of New York at Stony Brook, and the Research Foundation of the State University of New York, and in the form of a major grant from the National Science Foundation (number B039300). NOTE ADDED IN PROOF The first instar larvae of Heliconius (Neruda) aoede have now been observed from eggs obtained by Dr K. S. Brown at Leticia (Amazonas, Colombia). They have very long chaetae on the back (around three times the height of the head capsule) and are pale yellow with a brown-grey head, thus resembling the first instar larvae of Heliconius (Eueides) aliphera and differing considerably from those of the subgenus Heliconius (Beebe, Crane & Fleming, 1960). REFERENCES d'ALMEIDA, R. F., 1937. E x c u n i o scicntifica aos rios Cumina c Trombctas. Mems Insr. Oswaldo Cruz, 32: 235-98. BEEBE, W. CRANE, J. & FLEMING, H., 1960. A comparison of eggs, larvae and pupae in fourteen species of Heliconiine butterflies from Trinidad, W.I. Zoologicu, N. Y., 45: 1 1 1-54. BENSON, W. W., 1971. Evidence for the evolution of unpalatability through kin Selection in the Hcliconiinae (Lcpidoptcra). A m Nut., 105: 21 3-26. 308 J . R. G. TURNER BENSON, W. W., 1972. Natural selection for Miillerian mimicry in Heliconius erato in Costa Rica. Science, N. Y., 176: 936-9. BROWER, L. P., BROWER, J. VZ. & COLLINS, C. T., 1963. Experimental studies o n mimicry. 7. Relative palatability and Miillerian mimicry among Neotropical butterflies of the subfamily Heliconiidae. Zoologica, N. Y., 48: 65-84. BROWN, K. S. Jr. 1972. The heliconians of Brazil (Lepidoptera: Nymphalidae). Part 111. Ecology and biology of Heliconius nattereri. a key primitive species near extinction, and comments o n the evolutionary development of Heliconius and Eueides. Zoologica, N.Y., 5 7 : 41-69. BROWN, K. S., Jr, 1973. The heliconians of Brazil (Lepidoptera: Nymphalidae). Part V. Three new subspecies from Mato Grosso and Rondbnia. Bull. Allyn Mus., 13: 1-19. BROWN, K. S. Jr, 1975. Geographical patterns of evolution in Neotropical Lepidoptera. 11. Systematics and derivation of known and new Heliconiini (Nymphalidae, Nymphalinae). J. En?., Lond. ( B ) (in press). BROWN, K. S. J r & BENSON, W. W., 1974. Adaptive polymorphism associated with multiple rnullerian mimicry in Heliconius numata (Lepid. Nymph.). Biotropica, 6: 205-28. BROWN, K. S. J r & BENSON, W. W., 1975. The heliconians of Brazil (Lepidoptera, Nymphalidae). Part IV. Aspects of the biology and ecology of Heliconius demerer, with description of four new subspecies. Bull. Allyn Mus.. 26: 1-19. BROWN, K. S. J r & BENSON, W. W., 1976. The heliconians of Brazil (Lepidoptera Nymphalidae). Part V11. Evolution in modern Amazonian non-forest islands: Heliconius hermarhena. Biotropica (in press). BROWN, K. A. Jr, EMMEL, T. C. & SUOMALAINEN, E. Unpubl. Evolutionary significance of chromosome numbers in Neutropical Lepidoptera. Heliconiinae. (Manuscript). BROWN, K. S. J r & HOLZINGER. H., 1973. The heliconians of Brazil (Lepidoptera: Nymphalidae). Part IV. Systematics and biology of Eueides tales Cramer, with description of a new subspecies from Venezuela. Z. Arb. Os?. En?., 24: 44-65. BROWN, K S. J r & MIELKE, 0. H. H., 1972. The heliconians of Brazil (Lepidoptera: Nymphalidae). Part 11. Introduction and general comments, with a supplementary revision of the tribe. Zoologica, N. Y., 57: 1-40. BROWN, K. S . J r , SHEPPARD, P. M. & TURNER, J. R. G., 1974. Quaternary refugia in tropical America: evidence from race formation in Heliconius butterflies. Proc. R . SOC. Lond. (B), 187: 369-78. CAIN, A. J., 1958. Logic and memory in Linnaeus’s system of taxonomy. Proc. Linn. SOC.Lond., 16Y: 144-63. EHRLICH, P. R. & G!LBERT, L. E., 1973. Population structure and dynamics of the tropical butterfly Heliconius ethilla. Biotropica, 5: 69-82. ELTRINGHAM, H., 1917. On specific and mimetic relationships in the genus Heliconius, L. Trans. en?. SOC. Lond., I 9 1 6: 101-48. EMSLEY, M., 1963. A morphological study of imagine Heliconiinae (Lep.: Nymphalidae) with a consideration of the evolutionary relationships within the group. Zoologica, N. Y., 4 8 : 85-1 30. EMSLEY, M. G., 1965. Speciation in Heliconius (Lep., Nymphalidae): morphology and geographic distribution. Zoologica, N. Y., 50: 191-254. HAFFER, J., 1969. Speciation in Amazonian forest birds. Science, N. Y.. 165: 131-7. de LESSE. H., 1970. Formules chromosomiques de quelques Lhpidopteres Rhopaloceres de Guyane. Ann. SOC.en?. Fr. (N.S.), 6: 849-55. MICHENER, C. D., 1942. A generic revision of the Heliconiinae (Lepidoptera, Nymphalidae). A m . Mus. Novir., 1197: 1-8. NEUSTETTER, H., 1929. Nymphalidae: subfam. Heliconiinae. Lepid. Cat., 36: 1-1 36. RETTENMEYER, C. W., 1970. Insect mimicry. Ann. Rev. E n t , I S : 43-74. SHEPPARD, P. M., TURNER, J. R. G., BROWN, K. S. J r & BENSON, W. W., (1976. Genetics and the evolution of muellerian mimicry in Heliconius butterflies. (in preparation). STICHEL, H. & RIFFARTH, H., 1905. Heliconiidae. Tierreich, 22: xv i 290. SUOMALAINEN, E., COOK, L. M. & TURNER, J. R. G., 1972. Chromosome numbers of heliconiine butterflies from Trinidad, West Indies (Lepidoptera, Nymphalidae). Zoologica, N. Y., 56: 121-4. TURNER, J. R. G., 1966. A rare mimetic Heliconius (Lepidoptera: Nymphalidae). Proc. R. en?. SOC. Land (B), 35: 128-32. TURNER, J. R. G., 1968. Some new Heliconius pupae: their taxonomic and evolutionary significance in relation to mimicry (Lepidoptera. Nymphalidae). J. Zool., London., 155: 3 1 1-25. TURNER, J. R. G., 1971. Studies of Miillerian mimicry and its evolution in burnet moths and heliconid butterflies. In E. R. Creed (Ed.), Ecological generics and evolution: 224-60. Oxford: Blackwell. TURNER, J. R. G., 1975. A tale of two butterflies. Natural History, 84: 28-37. TURNER, J. R. G., 1976. Muellerian mimicry: classical ‘beanbag’ evolution and the role of ecological islands in adaptive race formation. Proc. Int. Con$ Popul. Genet, Ecol., Israel, 1975. London: Academic Press. YOUNG, A. M., 1973. Notes on the biology of the butterfly Heliconius cydno (Lepidoptera: Heliconiinae) in Costa Rica. Wasmann J. Biol., 31: 337-50.
© Copyright 2025 Paperzz