SIBLING SPECIES IN THE EURYDICE GROUP OF LETHE (LEPIDOPTERA: SATYRIDAE)

INTRODUCTION We have uncovered a pair o sibling species in the comparatively well-known butterfly auna o eastern North America.: the common Grass Nymph or Eyed Brown, Lethe eurydice o recent authors, is actually two species, which are extensively sympa.tric. Curiously, the distinctness o these two has been. known since at least 936, when W. D. Field discovered and characterized them as subspecies. He assigned names to them which we now know to be inapplicable. This was corrected in 947 by R. L. Chermock, who named the presumably more southern "subspecies" appalachia. Neither o these authors was aware that the "subspecies" are sympatric. The present investigation was rst suggested when one o us (Clench) ound both orms flying in the same area near Leesburg, Mercer Co., Pennsylvania in 966. The conspicuous habitat difference between them implied that two species might be involved. In 968 another o us (Shapiro) ound the same situation in western and central New York and (with Card) investigated the immature stages and biology o the insects. The results o this study are partly reported elsewhere (Shapiro and Card, 97o’). Independently o us, C. F. dos Passos and his correspondents simultaneously made the same discovery. Several o the conclusions contained in the resulting paper (dos Passos, 969) appear erroneous. Since the taxonomic situation is very complex, we here review the whole subject, nomenclatorially, morphologically, and distributionally. In brie, we recognize two species in this group, as ollows: (a) Lethe eurydice eurydice (Johansson), widely distributed rom Labrador to Great Slave Lake and south to Delaware and Illinois, occurring in open marshes and sedge meadows. (I b) Lethe eurydice fumosa (Leussler), scattered in small isolated colonies (many now extinct) in sedgy permanent marshes in the


INTRODUCTION
We have uncovered a pair o sibling species in the comparatively well-known butterfly auna o eastern North America.: the common Grass Nymph or Eyed Brown, Lethe eurydice o recent authors, is actually two species, which are extensively sympa.tric.
Curiously, the distinctness o these two has been. known since at least 936, when W. D. Field discovered and characterized them as subspecies. He assigned names to them which we now know to be inapplicable. This was corrected in 947 by R. L. Chermock, who named the presumably more southern "subspecies" appalachia.
Neither o these authors was aware that the "subspecies" are sympatric.
The present investigation was rst suggested when one o us (Clench) ound both orms flying in the same area near Leesburg, Mercer Co., Pennsylvania in 966. The conspicuous habitat difference between them implied that two species might be involved. In 968 another o us (Shapiro) ound the same situation in western and central New York and (with Card) investigated the immature stages and biology o the insects. The results o this study are partly reported elsewhere (Shapiro and Card, 97o').
Independently o us, C. F. dos Passos and his correspondents simultaneously made the same discovery. Several o the conclusions contained in the resulting paper (dos Passos, 969) appear erroneous.
Since the taxonomic situation is very complex, we here review the whole subject, nomenclatorially, morphologically, and distributionally.
In brie, we recognize two species in this group, as ollows: (a) Lethe eurydice eurydice (Johansson), widely distributed rom Labrador to Great Slave Lake and south to Delaware and Illinois, occurring in open marshes and sedge meadows.
(I b) Lethe eurydice fumosa (Leussler), scattered in small isolated colonies (many now extinct) in sedgy permanent marshes in the 1Department of Entomology and Limnology, Cornell University, Ithaca, N.Y. 14850. 2Section of Insects and Spiders, Carnegie Museum, Pittsburgh, Pa. 15213 *Manuscript received by the editor April 24, 1970. 1970] Card, 8haliro, Clench Lethe The descriptions of both eurydice Johansson and canthus Linnaeus are too scanty to restrict on internal evidence to either of the sympatric northeastern species, both of which occur at the type locality (Shapiro,97oa). If that locality (Philadelphia) is accurate, there can be no doubt that Johansson'.s description applies only to a member of this group, even though no mention is made of eyespots on the forewing above (an objection to this usage, raised by Harris,862 and Edwards,x897). No type of eurydice or canthus (which was proposed explicitly as a replacement name for eurydice and hence has the same type) exists in the British Museum (Natural History) or in the De Geer collection at the Naturhistoriska Riksmuseum, Stockholm.
When appalachia (see below) was described as the southern subspecies of eurydice, the latter name became firmly associated with the assumed "northern" subspecies whose color and pattern were contrasted with appalachia by Chermock. It seems desirable to Psyche [March stabilize the nomenclature by preserving this usage through a neotype designation. This removes the possibility that a specimen Chermock's appalachia might eventually be selected as neotype eurydice, leaving the amiliar "northern" insect's name in question.
The only Philadelphia specimens o eurydice auct. with ull data which we have ound were collected by one o us (Shapiro). Several o these were placed in the United States National Museum two years ago, and we desiginate one such specimen the neotype Papilio eurydice Johansson.
Neotype.--A male deposited in the U.S. National Museum bearing the manuscript label "eurydice of/Morris Arboretum/Phila. Co. Pa./29 June 967/A. M. Shapiro" (fig.  The "immaculatis" may have been by inerence rom the lack reerence to spots in the earlier descriptions, but it seems more likely that Fabricius was working rom some other insect he conused with the Linnean one. In 779 (Fabricius,78 ), improperly emending it to arganthe in synonymy.
(Arganth,e is not available as a replacement name because it was proposed in synonymy.) He repeated this usage in 787 and 793.
His own descriptions o "canthus'" do. not fit Cramer's figure well. Godart (I82I) recognized that three species were included in the Fabrician concept "'canthus'" and attempted to end the. conusion by redescribing the true canthus (translating Linnaeus), and naming two new entities, argulus and cantheus. Godart's argulus is a replacement name for the preoccupied argante and is the oldest valid name or this taxon. Cantheus is a renaming o the entity Fabricius first thought was canthus, theretoCore without a valid name. The identity oi: this animal cannot be determined it:, as appears, Fabrician specimens o "'canthus'" do not exist. (east shore Hudson's Bay); Mingan, Labrador; and Great Slave Lake, NWT. All of these are plotted on the map. The western distribution of e.urydice is unclear. Puckering and Post (196o) record it ;rom 'Cass, Cavalier, Dickey, Grand Forks, and Pembina Cos., North Dakota. These are entered on the map.
We have s.een no South Dakota records. However, Leussler (I938), who was well acquainted with L. e. fumosa, reported typical eurydice in Sioux Co., northwestern Nebraska. This record seems to require special confirmation, and has not been plotted.
It cannot be identified to species by the description, and the type has not been found; it is not in the Carnegie Museum, where dos Passos recorded it. We have placed boweri provisionally in the synonymy of eurydice because that species is considered more likely from the type locality, Port Hope, Ontario. However., a specimen of at515alachia with no ocelli on the forewings above, labeled "Bowie, Md./v-29-45/DDT experiment," is in the U.S. National Museum. At any rate, the name is clearly infrasubspecific and has no standing. Summary of Characters.--Lethe aPl)alachia differs from both subspecies of eurydice in being grayish or mousy brown abo.ve (blackish when fresh) and somewhat purplish or lilac-tinged beneath; the postmedial lines rounded, with only slight indentations. The male valve is less clearly 4-sided in lateral view, and the tegumen is dorsally flattened. The larval head capsule bears side-stripes not reaching below the bases of the horns. can be assigned to the correct species by color and pattern alone. The most useful characters separating the two speci,es are the ground color above and beneath, and the waviness of the postmedial line beneath. The two subspecies of eurydice differ most consistently in the relative sizes of the forewing ocelli. All of the observed differences are given in Table I. We have not seen a truly fresh specimen of L. e. fumosa. Leussler (I916) describes the ground color of fresh specimens as "a very dark smoky grey even a blackish appearance." This sounds very much like the color of newly emerged appalachia. Old specimens of the two are very different, 'however: fumosa males are an even, .somewhat purplish or reddish brown, while appalachia is grayish or mousy brown. Some female [umosa, particularly from Colorado, are nearly identical in color to nominate female eurydice, but the ground color of the males is nearly always distinctive. A few male eurydice from the northeast are dark purplish brown when fresh, and fade to an even dark reddish brown. Their spot-sizes are normal and they lack the fumosa tendencies to "high angledness" of the forewing and blind and rimless ocelli above. A specimen of this dark form of nominate eurydice is probably represented by Edwards' figure 5 (I897, pl. 26).
Specimens of the three taxa are .shown in figs. I-I2.
Male Genitalia. Chermock (I947) and dos Passos (969) reported no genitalic differences between L. e. eurydice and L. appalachia. However, we have found that they do differ slightly but significantly. The tegumen of appalachia is flattened dorsally, while that of eurydice (both subspecies) is rounded. The valves, of appalachia are shorter and narrower dorso-ventrally, and from the side appear less quadrilateral than those of the eurydice subspecies. The male genitalia of L. e. eurydice and L. e. fumosa are substantially similar, but differ from each other and from appalachia in the density and arrangement of setae on the valves.. See Table and figs. 13-18.
Female Genitalia.--There seem to be no useful characters here. Some minor differences in the sclerotization of the genital plate were found among all three taxa. Early Stages. The larvae of L. e. eurydice and L. appalachia from central New York differ consistently in the maculation and tubercles of the head capsule. In L. e. eurydice the red side stripes become darker below the bases of the horns, extending to the ocelli. The darker part of the stripe consists of small, heavily pigmented, regularly arranged tubercles on a less heavily pigmented ground. In L. appalachia the stripe does not extend below the. horn, and its lower end contains several large, pale, irregularly placed tubercles which contrast with the red ground (figs. 19, 2o).  Kornerup and Wanseher (1963) and Ridgway (1912 A usually small emergence of fresh eurydice occurs in some localities in New York, New Jersey and Pennsylvania in the first half of August, tour to five weeks after the principal emergence. Males of this late "brood" are frequently of the dark form noted above. It is very unlikely that these butterflies are descendants of those which emerged a month earlier. There may be a genetic basis for the emergence times; a bimodal emergence of Hyalolkora cecrolia (L.) (Saturniidae) was recently reported by Sternburg and Waldbauer (I969), with no genetic data. We do not believe the late eurydice are identical with [umosa, but the slight possibility exists that they represent another sibling species, unrecognizable in the adult except by its flight period and a .statistical color difference.
We have not obtained ova from these insects.
Food Plants.--Dos Passos (969) speculates that a food plant difference between L. eurydice and L. al)palachia is likely. However, our observations suggest that both are sedge-feeders and that neither is ,speciesor group-specific within Carex. Female aloloalachia occur near sedges in shrub swamp or forest habitats where observation is di(tlcult. One oviposition was seen in the field, on Carex Adult Behavior.--The most striking difference between L. e. eurydice and L. appalachia, and the one leading to the discovery of their sympatry, is their differential habitat selection (Shapiro and Card(:,97o). At the McLean Bogs Reserve, Tompkins Co., New York, these two species are frequently found flying within a few feet of each other, but do not mix. The preference of L. appalachia for shaded habitats often results in its association with L. p. anthedon upland or L. p. portlandia on the Coastal Plain. We have found L. e. eurydice only in relatively open sedge marshes or, rarely, in drier meadows; it never enters dense shrub swamp or woods. We have seen L. eurydice and L. p. antkedon in copula in their usual habitats, once each (3 p.m. and 3:3o p.m., respectively). DISCUSSION Although the term "sibling species" has been in the literature for nearly thirty years and the concept is even older, it still seems necessary to point out that excessive dependence on morphologicaI Psyche [March Fig. 22. Distribution of Lethe appalachia. differences can hinder the recognition ot? such biologically interesting species as those of the Lethe eurydice group. Despite abundant museum evidence of sympatry, these species went unrecognized for twenty years after Chermock (1947) was unable to find genitalic differences between them.
As usually happens with .sibling species, recognition on biological grounds has led to discovery ot morphological characters hitherto overlooked. These, however, are of a magnitude which would not be considered diagnostic of species in most groups ot? Lepidoptera. In t?act, the genitalia seem to be among the most conservative characters in Lethe. Chermock t?ound only very minor genitalic differences between L. portlandia and L. creola Skinner in the other American species group, and circumstantial evidence suggests that portlandia itseh is really a pair of (largely allopatric) sibling species. Many, Asiatic Lethe we have examined also show only slight differences among themselves and t?rom their close American relatives. We consider it likely that what we are calling Lethe eurydice fumosa may also prove .specifically distinct when its biology 97 --particularly the early stages--becomes better known. Similar cases recently uncovered in the Lepidoptera include the tortricid moths Archil)s argyrospilus and A. mortuanus, which differ only in sex attractant and in some characters of the last-instar larva (Roelofs and Comeau,I969), and the papilionid butterflies P'apilio zelicaon and P. gothica, said to differ consistently only in host-plant specificity but to behave as species in genetic tests (Remington, 1968). The Holomelina aurantiaca complex (_Arctiidae), often thought to consist of two species, actually includes at least ten, exceedingly similar in genitalic morphology, color and pattern, but differing in chromosome number (Card6,unpublished).
The seemingly inevitable problem with sympatric sibling pairs such as Lethe eurydice and appalachia is to account evolutio.narily for the "elegant" manner in which they coexist. The view that reproductive isolating mechanisms and ecological differences evolve in response to deleterious hybridization and competition in secondary sympatry (Brown and Wilson,956) is now very widely accepted. It was recently challenged by Ehrlich and Raven (I969), who proposed that isolating mechanisms usually develop during the genetic differentiation of allopatric populations under different selective regimes. This is in effect a reformulation of the view of most nineteenthand early twentienth-century evolutionists. Attempting to explain the ecological relationship of a given set of ,sibling species requires consideration of the following points" I. The apparent absence of ecological interaction (e.g., competition) or gene flow between presently sympatric populations does not rule out such events in the past, nor for competition, in the future.
Furthermore, intermittent large-scale gene flow between normally allopatric populations, associated with fluctuations in population sizes, has probably been an important component of speciation (Brown,I957). Such fluctuations could also result in episodes of competition between .otherwise non-competing species. 2. Biogeographical evidence may offer important clue.s to episodes of prior sympatry or allopatry in the evolution of species differences (cJ. Mengel,I964). 3. In the absence of evidence for character displacement, it cannot be assumed that biological differences which appear to prevent competition evolved in response to the adverse effects of competition.
The ecological differences among the American species of Lethe can be resolved into two parts" that involving the eurydice group alone and that concerning the eurydice and portlandia group.s. The Psyche [March species eurydice, a])palachia, and portlandia (in the broad sense, including anthedon) divide neatly into "non-competing" pairs" the tvo sedge feeders (eurydice and a)palackia) differ in habitat; the two woodland species (alalackia and l)ortlandia) differ in larval food plant. (Similarly, in sexual behavior, eurydice and al)palackia are essentially non-territorial; 1)ortlandia is strongly territorial.) Letke and the genera closely related to it are hypothesized to have originated in southeast A.sia (Miller,I968), a region with many forest-dwelling, grass-(mostly bamboo-) feeding representatives of both the eurydice and l)ortlandia groups. It seems reasonable that the ancestors of both these groups migrated to North America via the Bering land bridge in the Arcto-Tertiary forest, and were forced southward by the events of the. Pleistocene.
With the vast majority of the many Asian Letke feeding on grasses, the evolution o.f sedge feeding in North America by the ancestor of the eurydice group is a tempting hypothesis. Evolution of this trait by proto-eurydice independent of competition with proto-1)ortlandia or by character displacement in sympatry with it are both pos.sibilities. T. Shir)zu (pets. comm.) informs us that Letke marginalis Motschulsky, which seems to be a member of the eurydice group, feeds on non-bamboo grasses and on sedges in Japan, as does Kirinia elsaminondas Staudinger, formerly placed in Letke. Ninguta ("Letke") sckrenckii Menetries is an obligate sedge feeder.
On the other hand, the speciation of eurydice and a'Da[ackia may have occurred when one o.f the Pleistocene glaciations isolated some populations of proto-eurydice in prairie to the west of populations in the eastern Austral forests. Virtually the entire range of eurydice was glaciated, and the distribution is therefore of recent origin. The same can be said for the northern portions of the ranges o.f al)palackia and l)ortlandia, but the southern portions are characteristic of many organisms which presumably survived the Wisconsin (and earlier glaciations) in the southeast. The lack of recorded relict populations of eurydice south of Pennsylvania in the Appalachians, if not due to inadequate collecting, suggests that the species did not have a Wisconsin refugium in the forested Austral Zone of the southeast; its habitat preference and developmental rate support this interpretation.
(We do know of species of Hesperiidae, e.g. Eulkyes bimacula, with ranges and biologies substantially similar to L. eurydice, which have relict populations in the southeast; Shapiro, 97ob.) Its most probable refugium, then, was in glacial Transition Zone somewhere west of the Appalachians. The existence, of L. e. fumosa also. supports a prior western distribution for eurydice, but does not help in dating it. We have no grounds for estimating evolutionary rates in this group; all that can be said now with some confidence is that eurydice and appalachia were more likely allopatric than sympatric in the Wisconsin (and appalachia and portlandia more likely sympatric).
The critical evidence concerns character displacement. We have found no morphological character displacement in sympatric vs.
allopatric populations of Lethe eurydice and L. appalachia. One of us (Clench) believes he has observed behavioral character displacement between them in Pennsylvania, within the area of general sympatry; in certain localities where only one species occurs, it appears that the habitat selection is not so. rigorous as elsewhere. This needs additional study and quantification. Another geographic area also bears close investigation in this connection. Specimens of eurydice from southeastern New York (Orange, Rockland, and Westchester Counties) are somewhat anomalous, tending to vary in color and pattern (but not genitalia) toward appalachia. We have seen very few specimens from outside this small area which we would hesitate to classify to species by color and pattern. The area is completely surrounded by normal, sympatric, well-differentiated populations of both. It is thus critical to determine the ecology of these anomalous insects. Should appalachia be rare or absent, and e.ur'ydice occupying its niche at least in part, one would have a powerful a.rgumet for character displacement as the origin of the habitat difference.. (There is a chance of natural hybridization due to man's extensive disturbance of Lethe habitats in southeastern New York.) There are, then, two basic questions: Did the behavioral and food plant differences between the American eurydice and portlandia groups evolve independently, or largely as a. result of competition? Did the sharp habitat selection between eurydice and appalachia in close sympatry evolve in isolation, or was it intensified by behavioral character displacement ?
On the first point, any evolutionary scenario, will require much more comparative data on the Asiatic species than is readily available. Only a comprehensive revision, identifying the closest relatives of the American species and comparing their biologies, will alloxv e. convincing argument.
On the second point, field studies in areas of allopatry will be critical. It should be noted that while we suspect the eurydiceappalachia habitat difference may have evolved to prevent competition IOO Psyche [March for larval food, it also may function, and have evolved, as a reproductive isolating mechanism. Greater knowledge of mating behavior in this group (Shapiro and Card,97o), as well as more information on the southeastern New York populations, may be able to distinguish the correct hypothesis.
Our current .state of knowledge does not allow us to choose between independent evolution and character displacement in accounting for the differences between the American eurydice and portlandia groups. But character displacement is an attractive hypothesis for the eurydice-appalachia habitat selection difference.