VOLUME 27, NUMBER 3 175

LITERATURE CITED COMMON, I. F. B. 1964. Australian Butterflies. Jacaranda, Bristbane. 132 p. D'ABRERA, B. 1971. Butterflies of the Australian Region. Landsdowne, Melbourne. 415 p. DEGER, D. & R. EDEN. 1970. Collecting Australian Butterflies. Horwitz, North Sydney. 52 p. EHRLICH, P. R. 1958. The comparative morphology, phylogeny and higher clas­ sification of the butterflies (: Papilionoidea). Univ. Kansas Sci. Bull. 39: 305-370. Fox, R. M. 1956. A monograph of the Ithomiidae (Lepidoptera), Part 1. Bul!. Amer. Mus. Nat. Hist. lll: 1-76. HEMMING, F. 1967. The generic names of the butterflies and their type- (Lepidoptera: Rhopalocera). Bull. Brit. Museum (Nat. Hist.): Entomo!. Supp!. 9: 1-510. KOHIYAMA, K., T. TAKASE & T. FUJIOKA. 1971. Butterflies of Japan. Yama-to­ Keikoku, Tokyo. 204 p. (in Japanese). MILLER, L. D. 1968. The higher classification, phylogeny and zoogeography of the Satyridae (Lepidoptera). Mem. Amer. Entom. Soc. 24: 1-174. TITE, G. E. 1963. A revision of the Candalides and allied genera (Lepidop­ tera: Lycaenidae). Bull. Brit. Museum (Nat. Hist.): Entomol. 14: 199-260. YAGI, N. & N. KOYAMA. 1963. The Compound Eye of Lepidoptera: Approach from Organic Evolution. Shinkyo, Tokyo. 320 p.

STUDIES ON THE (NOCTUIDAE) OF SOUTHERN NEW ENGLAND. IV. A PRELIMINARY ANALYSIS OF BEAK-DAMAGED SPECIMENS, WITH DISCUSSION OF ANOMALY AS A POTENTIAL ANTI-PREDATOR FUNCTION OF HINDWING DIVERSITY

THEODORE D. SARGENT Department of Zoology, University of Massachusetts, Amherst, Massachusetts 01002

An intensive study of the biology of the Catocal,a is presently being conducted in southern New England (Sargent & Hessel, 1970; Kellogg & Sargent, 1972; Sargent, 1972a). During the course of these investigations, substantial numbers of beak-damaged specimens have been collected. The present study was undertaken in hopes that an analysis of such specimens might shed some light on various aspects of the predator-prey relationships involving and these moths.

Prior Studies Beak-damaged Lepidoptera have attracted considerable attention in the literature, though most prior studies have been concerned almost 176 JOURNAL OF THE LEPIDOPTERISTS' SOCIETY exclusively with butterfly examples, A number of interpretations of various damage patterns have been advanced, and some of these seem generally accepted, Thus, V-shaped tears in the wings, which suggest that the specimens involved have torn free from attacks, have been regarded as evidence of palatability in their bearers. On the other hand, disabling injuries (i.e. those sufficient to prevent flight) have been regarded as evidence that their bearers are unpalatable or otherwise noxious (e.g. Poulton, 1913). Unpalatability has also been advanced as an explanation for crisp beak-imprints on the wings, the presumption being that specimens exhibiting such marks have been purposely released by birds (e.g. Collenette & Talbot, 1928). This last explanation receives support from studies showing a relatively high frequency of crisp beak­ imprints on the wings of aposematic species (e.g. Carpenter, 1941) . Beak-damage patterns have also provided evidence that the small, eye­ spot markings on the margins of butterfly wings direct the attacks of birds (e.g. Marshall & Poulton, 1902; Swynnerton, 1926; Blest, 1957), and so function as deflective devices. Although beak-damage patterns in moths have received little attention, the prior work with butterflies encouraged the present study and raised hopes that some insight could be gained into predator-prey relationships involving the Catocala moths.

Anti-Predator Functions of Catocala Wings Catocala moths appear to rely upon the bark-like crypsis of their fore­ wings as their primary defense against . A number of studies reveal the extent to which this crypsis is enhanced by means of behavioral adaptations, e.g. the selection of appropriate backgrounds ( Sargent, 1966, 1968, 1969a, 1973; Keiper, 1968; Sargent & Keiper, 1969), and the adoption of appropriate resting attitudes (Sargent, 1969b). It seems likely that forewing variations, particularly the dramatic polymorphisms exhibited by many species, function to augment the effectiveness of crypsis by foiling predator tendencies to form "specific searching images" (Tinbergen, 1960) for particular cryptic patterns (Sargent, 1972b). Hindwing function in Catocala moths has not been extensively studied. It is generally assumed that these boldly patterned, and often colorful, structures are examples of "flash coloration," being revealed when crypsis fails to deter attack and a takes flight, only to be concealed when the moth again alights. Such "flash and cover" sequences are presumed to confuse predators as to the whereabouts of the moths (Cott, 1940; Ford, 1967). It is sometimes assumed that the flash of the hindwings is itself the important anti-predator device, functioning to startle predators and VOLUME 27, NUMBER 3 177

thereby adversely affecting the efficiency of their attacks. In this event, Catocala hindwings would have a function similar to that suggested for the large eye-spots of saturniid moths (Blest, 1957~ Coppinger, 1970). Catocala hindwings might also serve a deflective function. In this case, predators would direct their attacks toward these prominent structures, thereby being directed away from more vulnerable body parts. Catocala hindwing patterns would then be functionally related to the small eye-spots found along the margins of some butterfly wings, which are known to direct predator attacks (Poulton, 1890; Blest, 1957). The interspecific diversity found in hindwing colors and patterns in the genus Catocala has also been suggested as an anti-predator adapta­ tion (Sargent, 1969c). In this view, hindwing variation introduces the potential of novelty (unfamiliar stimuli) or anomaly (unexpected stimuli) as a startle mechanism into predator-prey relationships involving birds and several Catocala species. For example, a bird, after several suc­ cessive encounters with species possessing a particular hindwing pattern (e.g. yellow and black bands), might become habituated to that pattern; but this same bird, upon encountering a distinctly different hindwing pattern (e.g. pink and black bands), might be effectively startled. If the level of startle exhibited by a bird to a particular hindwing pattern were a function of the extent and recency of its experience with that pattern, then considerable advantage might accrue to an assemblage of rather similar cryptic species which evolved a variety of hindwing pat­ terns (see discussion of schizomimicry; Sargent, 1969c). While it appears likely that Catocala hindwings serve one or more of the preceding anti-predator functions, it is also quite possible that they play an important role in other aspects of the biology of these moths. The rather surprising lack of intraspecific hindwing variation in the genus suggests that these structures may function as releasers and anti­ hybridization devices in courtship and mating behaviors. Little is known of these behaviors, but in one species, C. relicta Walker, mating can oc­ cur under conditions of essentially complete darkness (Sargent, 1972a). However, until much more is known of Catocala courtship and mating behaviors, the question of a sexual role for the hindwings must remain open.

METHODS The primary aims of the present study were to describe and interpret the beak-damage patterns found on the wings of wild-caught Catocala moths. In order to achieve these aims, two procedures were followed: ( 1) recording the behavioral interactions between captive birds and Catocala moths in aviaries, and collecting those moths which escaped 178 JOURNAL OF THE LEPIDOPTERISTS' SOCIETY under these circumstances for later analysis of their beak-damage pat­ terns; and (2) collecting all Catocala specimens showing suspected beak­ damage from a large field sample being obtained in connection with other studies.

Aviary Study The Catocala used in this study were 50 fresh, undamaged specimens which were taken in a Robinson trap in Leverett, Massachusetts during July and August of 1971. These moths were released to birds in two aviaries (each 8 X 5 X 10 ft.) which were located on the roof of the Morrill Science Center on the University of Massachusetts (Amherst) campus. The aviaries were lined with fine-mesh hardware cloth which prevented escape of the moths. Each aviary housed seven blue jays (Cyanocitta cristata) which were 1-2 years old, and which had been hand-reared from approximately 14 days of age by Dr. Alan C. Kamil and his graduate students in the Psychology department. These birds had been tested in discrimination and learning-set experiments (e.g. Hunter & Kamil, 1971), but their prior experience with was limited to the mealworms (Tenehrio larvae) which were used as reinforcements in these experiments, and to various flies, etc. which frequented their aviaries. The birds were normally maintained on a free-feeding schedule of food (mynah pellets) and water, but prior to the release of Catocala into the aviaries, the food was withdrawn for a period of 16-24 hours. Under these circumstances, the moths were quickly attacked by the birds. The observations of moth~bird interactions included careful attention to such matters as; ( 1) whether a moth was attacked while flying or resting; (2) if attacked while resting, the direction from which attack occurred (rear, side, etc.); and (3) if escaping from attack, the duration of retention in the bird's beak (instantaneous, several seconds, etc.). Of the 50 Catocala specimens released to the birds, 23 were recovered for analysis of their beak-damage patterns. Every effort was made to recover each moth which escaped its initial attack, but some specimens suffered second or third attacks before they could be captured. The birds became progressively more efficient at taking the moths, losing 10 of the first 12 released, but successfully capturing each of the last 15 speci­ mens. Moths that were eaten by the birds were initially ingested wings and all, but after some expericnce thc birds invariably attempted to remove the wings by holding the moths in their feet and pecking with their beaks. None of the birds developed the clean, efficient shearing-off of the wings, without use of the feet, that I have witnessed in blue jays in the field. VOLUME 27, NUMBER 3 179

Field Sample A total of 2047 Catocala specimens of 33 species were recorded at lights (incandescent, black light fluorescent, and mercury vapor) and bait (brown sugar-beer mixture) in Leverett, Massachusetts during the summer of 1971. Of this total, 1623 specimens were carefully examined 'n glass jars (primarily for sexing, by means of the frenulum), and if suspected beak-damage was discovered during this examination, the specimen was retained for later comparison with specimens obtained in the aviary study. Of the 75 wild-caught specimens that were retained, 65 were later judged to exhibit clear beak-damage, with the remainder exhibiting -inflicted damage patterns.

RESULTS The behavioral observations in the aviaries, coupled with subsequent analyses of specimens, permitted a classification of beak-damage patterns, as follows: Type I. Attack: while moth in flight Characteristic damage: unilateral; tear from one wing only a. hindwing tear (Figs. 1a; 4a,b,c; 7a) h. forewing tear (Figs. 1b; 4d) Type II. Attack: while moth resting; grasp of beak not including costa of either forewing Clwracteristic damage: corresponding tears from ipsilateral forewing and hindwing a. forewing and hindwing tears overlapping when wings fully closed; unilateral (Figs. 2a; 5a,b) h. forewing and hindwing tears over­ lapping when wings partially closed; unilateral (Figs. 2b; 5c; 7b) c. forewing and hindwing tears over­ lapping on both sides when wings fully closed; bilateral (Figs. 2c; 5d) Type III. Attack: while moth resting; grasp of beak including costa of one forewing Characteristic damage: crisp beak-imprint on one forewing; tear from ipsilateral hindwing 180 JOURNAL OF THE LEPIDOPTERISTS' SOCIETY

a. apex of beak-imprint directed toward (but not across) inner mar- gin of forewing; unilateral (Figs. 3a; 6a,b; 7a,b) h. apex of beak-imprint directed toward (and across) costal margin of forewing; unilateral or bilateral (Figs. 3b; 6c) c. apex of beak-imprint directed toward (and across) inner margin of forewing; bilateral (Figs. 3c; 6d)

This classification of beak-damage patterns, based on specimens obtained in the aviary study, also proved adequate for classifying the field sample specimens. The number of examples of each damage pattern obtained in both the aviary and field samples is given in Table l. Note that a number of specimens (6 in the aviary sample, 8 in the field sample) exhibited two damage patterns (e.g. Fig. 7). Most of the field specimens were easily assigned to categories within the preceding classification, but occasional difficulties were encountered. For example, three specimens were apparently grasped from the rear while resting such that the apex of the bird's beak almost, but not quite, reached the forewing costa. The resulting damage, which technically had to be classified as Type IIa, included at least a portion of a crisp beak-imprint, and this is a characteristic of Type III damage. Another problem was posed by four specimens which exhibited smeared, rather than crisp, beak-imprints. However, since these specimens otherwise met the criteria of Type III damage, they were so classified. The only damage pattern which was not obtained in the field sample

TABLE 1. Distribution of damage patterns among Catocala in the aviary and field samples.

Number of Individual Examples Damage Patterns Aviary Sample Field Sample

Type I a 12 33 b 1 1 Type II a 5 16 b 2 3 c 1 2 Type III a 5 13 b 2 5 c 1 0 Totals 29 73 VOLUME 27, NUMBER 3 181 I.

a

b Fig. 1. Type I damage patterns. Diagrammatic representations of bird attacks on flying moths (left), and the resulting specimens (right). Damage is usually confined to one hindwing (a), rarely to one forewing (b). was Type IIIc. The attack which leads to this damage pattern (Fig. 3c) was the one that blue jays in the aviaries used almost exclusively after they had had some experience with Catocala. A moth grasped in this fashion rarely escaped, as its wings were securely held in place. How­ ever, on one occasion a bird apparently loosened its grip when attempting to transfer a moth from its beak to its feet, and the escaped specimen was recovered (Fig. 6d). Presumably such "carelessness" would be very rare in nature. 182 JOURNAL OF THE LEPIDOPTERISTS' SOCIETY n.

a

b

c Fig. 2. Type II damage patterns. Diagrammatic representations of bird attacks on resting moths (left), and the resulting specimens ( right) . Damage may be inflicted when the wings are fully closed (a) or partially spread (b), and while usually unilateral, may be bilateral (c) on occasion. VOLUME 27, NUMBER 3 183 m.

a

b

c Fig. 3. Type III damage patterns. Diagrammatic representations of bird attacks on resting moths (left), and the resulting specimens (right) . The beak-mark is usually confinecl to one forewing, inclucling an imprint of the apex (a) or not (b), but may extend across both forewings (c). 184 JOURNAL OF THE LEPIDOPTERISTS' SOCIETY

Fig. 4. Specimens exhibiting Type I damage patterns: A. C. antinympha, Ia (8 August 1970, Leverett, Mass., Robinson trap); R. C. l1niiuga, Ia (30 August 1971, Leverett, Mass., bait); C. C. palaeogama, Ia (28 August 1971, Leverett, Mass., Robinson trap); D. C. residua, Ib (8 July 1972, Fontana, No. Carolina, black light-specimen courtesy D. F. Schweitzer).

As mentioned previously, a number of specimens which were initially retained in the field sample were later judged to be bat-damaged rather than bird-damaged. On one occasion a moth was observed to escape a bat attack near the Robinson trap, and this specimen was immediately recovered (Fig. 8b). Other specimens exhibiting the same distinctively frazzled or tattered outer wing margins (e.g. Fig. 8a) were presumed to be bat-damaged as well.

TABLE 2. Distribution of damage patterns among field-caught Catocala of two hind wing types.

Number of Individual Examples Damage Patterns Total Hindwing Types II III Sample

Chromatic 28 14 10 1228 Achromatic 6 7 8 395 Percent Achromatic 17.6 33.3 44.4 24.3

Note: The achromatic hindwing totals include one specimen of C. relicta (banded, as opposed to uniformly black, upper hind wing surface) under damage patterns I and III, and there were 28 specimens of this species in the total sample. VOLUME 27, NUMBER 3 185

Fig. 5. Specimens exhibiting Type II damage patterns: A. C. ilia, lIa (5 August 1971, bait); B. C. epione, lIa (10 August 1971, black light); C. C. flebilis, IIb (4 September 1971, Robinson trap); D. C. retecta, IIc (12 September 1971, Robinson trap). All specimens from Leverett, Mass.

It seemed of some interest to ascertain whether the various types of beak-damage were distributed in the same way among moths having very different types of hindwings. Accordingly, the major beak-damage patterns were tabulated separately for moths having chromatic and achromatic hindwings (Table 2). (Chromatic hindwings include those characterized by color, this color generally providing a ground for prominent black bands on both the upper and lower wing surfaces. Achromatic hindwings include those characterized by the absence of color, with the upper wing surface generally solid black, and the lower wing surface banded with black and white. Both hindwing types may have a more-or-Iess prominent white fringe.) Analysis of the data in Table 2 revealed that significantly more speci­ mens with achromatic hindwings were found with Types II and III damage (escaping attacks while resting) than were found with Type I damage (escaping attacks while flying), or than were present in the total field sample (chi-square 2 X 2 contingency tests; P's < 0.05).

DISCUSSION In the present sample of wild-caught Catocala moths, 4% of the in­ dividuals exhibited clear evidence of at least one bird attack. These collected moths had successfully escaped their attacks, but it may be 186 JOURNAL OF THE LEPIDOPTERISTS' SOCIETY

Fig. 6. Specimens exhibiting Type III damage patterns: A. C. ultronia, IlIa (ex­ perimental moth 2); B. C. retecta, IlIa (experimental moth 6); C. C. ultronia, IlIb (11 August 1970, Robinson trap); D. C. concumbens, lIIe (experimental moth 10). All specimens from Leverett, Mass.

assumed that many other individuals were not so successful. Thus it appears that avian predation on Catocala moths is substantial. In such a circumstance, one would expect the prey to have some highly evolved defensive strategies. The bark-like crypsis of Catocal.a forewings, coupled with appropriate behaviors, seems a clear example of such a strategy. The anti-predator functions of Catocala hindwings have not been clearly established, but an analysis of the present beak-damaged specimens provides evidence for both deflective and startle functions. Nearly half of the beak-damaged specimens were apparently attacked while in flight, and all but one of these exhibited only hindwing damage. It appears that bird attacks are frequently directed towards these structures when Catocala moths are flying, and that such attacks may result in damage which is not highly detrimental to the moths (some specimens have been taken at lights which are missing virtually all of one hindwing). Thus Catocala hindwings apparently function on some occasions in a fashion similar to the colorful, but expendable, tails of certain lizards (Cott, 1940). The remaining half of the beak-damaged specimens were apparently attacked while resting. Escape from these attacks seemed to be of two sorts: (1) a pulling-free of the moth while being tightly gripped by a bird, resulting in a tearing of the wings around the region of beak con- VOLUME 27, NUMBER 3 187

Fig. 7. Specimens exhibiting two damage patterns: A. C. ultronia, IlIa on the left side, Ia on the right side (11 August 1971, black light); B. C. ultronia, IIb on the left side, IlIa on the right side (19 August 1971, Robinson trap). Both specimens from Leverett, Mass.

tact (Type II damage); and (2), a release of the moth after being tightly gripped by a bird, resulting in a clear beak-imprint along some of the lines of beak contact (Type III damage). Escapes of the first sort apparently occurred when the size, speed, and strength of the moth enabled it to break free from the grip of a bird. It is postulated that escapes of the second sort were at least in part a result of startle responses of the predators to the sudden appearance of the contralateral hindwing (i.e. opposite the side being gripped). The startle response is viewed as effecting a momentary relaxation of a bird's grip, this relaxation enabling the moth to make its escape, and leaving a crisp beak-imprint on some part of its wings. (It seems likely that some of the specimens exhibiting Type II damage could have startled their predators, but that tears normally occurred in the wings, rather than beak-imprints, because a bird's grip in such cases included only the more fragile por­ tions of the forewing. Thus, the normal swift escape reactions of the moths might have resulted in significant tearing of the wings before

Fig. 8. Specimens exhibiting bat-inflicted damage: A. C. ilia (8 August 1971, bait); B. C. habilis (18 September 1971, Robinson trap). Both specimens from Leverett, Mass. 188 JOURNAL OF THE LEPIDOPTERISTS' SOCIETY the birds relaxed their grip, with further tearing perhaps resulting from subsequent wear on the wings.) There appears to be a contradiction between the functions of deflection and startle for the same structures with respect to the same predators. However, the resolution of this apparent contradiction may lie in an understanding of the nature of startle in this situation. It is suggested that the degree of startle exhibited by a bird in any encounter with a Catocala hindwing pattern is not so closely related to that hindwing pattern per se, as it is to the degree of expectation of that pattern which the bird brings to the encounter on the basis of its past experience. Thus, anomaly, defined in terms of departure from expectation, is regarded as the critical factor in determining whether a bird will react by releasing a Catocala moth which suddenly displays a particular hindwing pattern. This instantaneous reaction might not interfere in any substantial way with the ability of a bird to attack a flying moth, when time to adjust to the appearance of the hindwings might be available. An "anomalous stimulus" (defined in terms of its departure from ex­ pectation and the momentary startle it elicits) would be clearly distin­ guishable from a "frightening stimulus" (which is usually defined in terms of its absolute properties and the innate rejection it elicits), and could often be distinguished from a "novel stimulus" (which is usually defined in terms of its unfamiliarity and the avoidance it elicits) (see Brower, 1971). Anomaly and novelty are obviously closely related, and although an anomalous stimulus need not be a novel stimulus (as these are here defined), the two phenomena may produce similar physiological effects in predators (e.g. high arousal (Coppinger, 1970), or an "orienting reflex" (Sokolov, 1960)). The primary evidence that Catocala hindwings function as anomalous stimuli is provided by the specimens which exhibit crisp beak-imprints on their wings (Type III damage). Such beak-marks on lepidopteran wings have long been regarded as evidence that the specimens involved were captured and subsequently released by birds. Since beak-marks of this sort are most often found on aposematic species, it has been assumed that release of the specimens followed predator recognition of some noxious quality (usually odor or taste) of the captured prey. However, Catocala moths, as far as is known, are entirely palatable (many species are readily eaten by cage-birds (pel's. obs.), and the work of Jones (1932) supports the assumption of their palatability). Why should these moths be released after capture by birds? The answer to this question may be related to some sort of startle response on the part of predators, and it is suggested that interspecific hindwing variation provides the key to understanding the situation. This VOLUME 27, NUMBER 3 189 hind wing variation introduces the potential of anomalous stimuli into predator-prey relationships involving these moths. Thus, a predator is seen as building up expectations regarding future hindwing patterns on the basis of its past experiences with these patterns, reacting more and more efficiently if these expectations are met, but inefficiently if they are countered. Inefficiency presumably results from some involun­ tary response (e.g. gaping) which is elicited by an unexpected stimulus, and which interferes with the completion of normal attack, allowing a moth to escape. Some indirect evidence for this view is provided by an analysis of the distribution of beak-damage patterns on moths having different hind wing types. The most obvious discontinuity among Catocala hind­ wings occurs between the chromatic and achromatic patterns. As in­ dividuals possessing chromatic hindwings are more common than those possessing achromatic hindwings (the latter comprising less than 250/0 of the Catocala taken in Leverett each year), it might be predicted that achromatic hind wings would be less often encountered, and therefore more often anomalous, than chromatic hindwings. Analysis of the data in Table 2 reveals that beak-damage patterns II and III are more com­ monly found on individuals possessing achromatic hindwings than would be expected on the basis of chance. This finding suggests that achromatic hindwings are particularly effective as startle devices, and anomaly may provide an explanation for this effectiveness. Predators exhibiting a tendency to react inefficiently to anomalous stimuli would provide strong selection pressures favoring diversity in their prey. Diversity would result in an increased number of potential predator expectations, and a corresponding increased number of po­ tential anomalous stimuli. Anomaly, within the limits of the advantage it provides, would then favor the origin and maintenance of considerable diversity in sympatric assemblages of closely related species. Hindwing diversity among moths in the genus Catocala may represent a response to pressures of this sort. Whenever one suggests that selective advantages result from inter­ specific hind wing diversity in Catocala moths, then the apparent problem posed by the lack of intraspecific hindwing diversity must be faced. If selection has favored diversity on the one hand, why has it opposed diversity on the other? This problem would be easily solved if the hind wings were involved in specific recognition; serving, for example, as releasers in mating behaviors and therefore as anti-hybridization devices. Thus far there is no evidence that the hind wings serve any such function, but the possibility must remain open until Catocala mating behaviors are thoroughly studied. 190 JOURNAL OF THE LEPIDOPTERISTS' SOCIETY

Another explanation for the lack of intraspecific hindwing variation may be related to the notion that parsimony would prevail with respect to the genetic bases of adaptive diversity in nature. Adaptive diversity in the case of Catocala hindwings is spread over an assemblage of species and seems to involve a large, but finite, number of categories, with each category being limited to a more-or-Iess fixed percentage of the total assemblage. The origin and maintenance of such diversity would seem to be more easily achieved if monomorphism, rather than polymorphism, characterized the species involved.

SUMMARY An attempt is made to describe and interpret beak-damage patterns found on the wings of Catocala moths. Two samples of beak-damaged specimens were studied: ( 1) 23 individuals recovered after being at­ tacked by blue jays in aviaries; and (2) 65 individuals retained from a large sample of these moths taken at bait and lights in the field. Analyses of these moths resulted in a classification of beak-damage pat­ terns into three major types: I (attacked while flying, tear from one wing); II (attacked while resting, tears from ipsilateral forewing and hindwing); and III (attacked while resting, beak-imprint on at least one wing). About half of the specimens in both the aviary and field samples were attacked while flying, the other half being attacked while resting. The damage patterns obtained provide evidence that Catocala hind­ wings serve both deflective and startle functions. The probable nature of the startle function is discussed in detail, and it is suggested that a sudden, unexpected display of hindwings results in startle (and con­ sequent unsuccessful completion of attack) in predators. The evidence for this view is provided by the crisp beak-imprints found on the wings of many Catocala specimens, and by the distribution of these beak­ imprints among moths having different types of hindwings. Anomaly (defined as departure of prey from predator expectation, with resultant startle in the predator) is proposed as a functional and adaptive basis for hindwing diversity in Catocala moths.

ACKNOWLEDGMENTS I thank Dr. Alan C. Kamil for providing the blue jays used in the aviary study, and for critically reading the manuscript. Dale F. Schweitzer kindly provided a number of Catocala specimens for study. My wife, Katherine, ably assisted in the preparation of the figures, and patiently helped in many other ways. VOLUME 27, NUMBER 3 191

LITERATURE CITED BLEST, A. D. 1957. The evolution of eyespot patterns in the Lepidoptera. Be­ haviour 11: 209-256. BROWER, L. P. 1971. Prey coloration and predator behavior, in Topics in the Study of Life: The BIO Source Book. Harper & Row, New York. p. 360-370. CARPENTER, G. D. H. 1941. The relative frequency of beak-marks on butterflies of different edibility to birds. Proc. Zoo I. Soc. London A Ill: 223-23l. COLLENETTE, C. L. & G. TALBOT. 1928. Observations on the bionomics of the Lepidoptera of Matto Grosso, Brazil. Trans. Entomol. Soc. London 76: 391- 416. COPPINGER, R. P. 1970. The effect of experience and novelty on avian feeding behavior with reference to the evolution of warning coloration in butterflies. II. Reactions of naive birds to novel insects. Amer. Nat. 104: 323-335. COTT, H. B. 1940. Adaptive Coloration in . Methuen, London. 508 p. FORD, E. B. 1967. Moths. 2nd. ed. Collins, London. 266 p. HUNTER, M. W. III & A. C. KAMIL. 1971. Object-discrimination learning set and hypothesis behavior in the northern bluejay (Cyanocitta cristata). Psych on. Sci. 22: 271-273. JONES, F. M. 1932. coloration and the relative acceptability of insects to birds. Trans. Roy. Entomol. Soc. London 82: 443-453. KEIPER, R. R. 1968. Field studies of Catocala behavior. J. Res. Lepid. 7: 113- 12l. KELLOGG, C. G. & T. D. SARGENT. 1972. Studies on the Catocala (Noctuidae) of southern New England. II. Compariwn of collecting procedures. J. Lepid. Soc. 26: 35-49. MARSHALL, G. A. K. & E. B. POULTO:--i. 1902. Five year's observations and ex­ periments (1896-1901) on the bionomics of South African insects, chiefly directed to the investigation of mimicry and warning colours. Trans. Entomol. Soc. London: 287-584. 15 pIs. POULTON, E. B. 1890. The Colours of Animals. Intern. Sci. Ser. 68, London. 360 p. 1913. Disabling and other injuries found in the Lepidoptera and their interpretation. Proc. Entomol. Soc. London: xix-xxii. SARGENT, T. D. 1966. Background selections of geometrid and noctuid moths. Science 154: 1674-1675. 1968. Cryptic moths: effects on background selections of painting the circumocular scales. Science 159: 100-101. 1969a. Behavioral adaptations of cryptic moths. II. Experimental studies on bark-like species. J. N. Y. Entomol. Soc. 77: 75-79. 1969b. Behavioral adaptations of cryptic moths. III. Resting attitudes of two bark-like species, Melanolophia canadaria and . Anim. Behav. 17: 670-672. 1969c. A suggestion regarding hindwing diversity among moths of the genus Catocala (Noctuidae). J. Lepid. Soc. 23: 261-264. 1972a. Studies on the Catocala of southern New England. III. Mating results with C. relicta Walker. J. Lepid. Soc. 26: 94-104. 1972b. Sketches of New England moths. 5. Polymorph isms. Man & Nature (September): 25. 1973. Behavioral adaptations of cryptic moths. VI. Further experimental studies on bark-like species. J. Lepid. Soc. 27: 8-12. --- & S. A. HESSEL. 1970. Studies on the Catocala (Noctuidae) of southern New England. I. Abundance and seasonal occurrence of the species, 1961- 1969. J. Lepid. Soc. 24: 105-117. 192 JOURNAL OF THE LEPIDOPTERISTS' SOCIETY

--- & R. R. KEIPER. 1969. Behavioral adaptations of cryptic moths. I. Preliminary studies on bark-like species. J. Lepid. Soc. 23: 1-9. SOKOLOV, E. N. 1960. Neuronal models and the orienting reflex, in M. A. B. Brazier, ed., The Central Nervous System and Behavior. Joseph Macy Jr. Found., New York. p. 187-276. SWYNNERTON, C. F. M. 1926. An investigation into the defences of butterflies of the genus Charaxes. PIDC. III Intern. Entomol. Congr., Zurich (1925) 2: 478-506. TINBERGEN, L. 1960. The natural control of insects in pinewoods. 1. Factors influencing the intensity of predation by songbirds. Arch. Neer!. Zool. 13: 265-343.

AN ORCHID ATTRACTANT FOR MONARCH BUTTERFLIES (DANAIDAE)

W. H. WAGNER, JR. Department of Botany, University of Michigan, Ann Arbor, Michigan 48104

In view of the recent spurt of research on insect-flower relationships, and in particular the studies of Dodson and his co-workers on biologically active compounds in orchid fragrances (Dodson et aI., 1969), it is especially interesting to discover an orchid which seems to have a practically "irresistible" attraction for monarch butterflies. The monarch butterfly is unquestionably the best known butterfly species in the United States, and its biology has been the subject of a great deal of attention (Urquhart, 1960). In the Great Lakes states, as elsewhere, the monarch is famous for its swarming behavior prepara­ tory to migration southward to the Gulf Coast and Mexico. Swarming is observed in the latter part of September and early October as a rule, although Moore (1960) reported a swarm in the middle of August, " ... thousands of individuals on Seul Choix Point on the north shore of Lake Michigan," and Urquhart points out that migration actually has its beginnings in July. During the first two weeks of autumn, monarch butterflies are fre­ quent everywhere in the vicinity of Ann Arbor in southeastern Michigan. Their behavior is languid, and they soar slowly across fields and along roadsides, feeding especially upon flowers of various species of asters (e.g. Aster azureus, A. ericoides, A. laevis, A. novae-angliae). In culti­ vated legume fields the monarchs visit red clover (Trifolium pratense) and alfalfa (Medicago sativa) primarily. In general, butterfly diversity is low at this time during most years-a few sulphurs (Colias philodice, C. eurytheme), some worn swallowtails (Papilio polyxenes especially),