••.'• - . . . ,• ' •.••,, , .■ . . ^ C •. ' -. .-'i -h


:^£ > -.3

>-f§vv.y:''- “•-:'w.'./^--,fic

'■':■■■ :


>■•*. •^.w.% . ■•' -•

.'v. ril'VvV'.v;,'.'

1-- •;>•• .


■'AyJl , -







Family 1.— NOTODONTIDvE.






•• *.




I. Introduction.

II. Hints on the mode of Evolution of the Bristles, Spines, and Tubercles of Kotodontian and other Caterpillars.

III. On Certain Points in the external Anatomy of Bom-

bycine Larvae.

IV. On the Incongruence between the Larval and Adult

Characters of Notodontians.

V. Inheritance of Characters acquired during the Life- time of Lepidopterous Larvoa.


VI. Geographical Distribution of the American Noto- doutidie.

VII. Phytogeny of the Lopidoptera.

VIII. Attempt at anew Classification of the Lepidoptera.

IX. A rational Xomenclature of the Veins of the Wings of Insects, especially the Lepidoptera.

X. Systematic Revision of the Notodontidje, with spec- ial Reference to their Transformations.


For some years past the writer lias been collec ing materials for a general accouni, systematic and developmental, of our North American Bombycine moths. The leading object or motif of the essay has been to collect materials for working oat the origin of the laryal forms of the higlicr Lepidoptera,

The attempt has been made, so far as material and opportunity have allowed, to describe m as detailed a \y£Lj as possible the transformations of our Bombycine moths, in the light of the rrcent very suggestiv-e and stimulating work of Weismann, entitled Studies in the Theories of Descent (1882). Until within a few years the majority of descriptions of caterpillars have been prepared simply for the purpose of identification, or for taxonomical uses, and without reference to the philosophic or general zoological significance of tliese changes. The transformations of some of the European Sphiugidie have been very carefully worked out by Weismann, and also by Poulton, but it is believed that the life histories of the lower, more generalized families usually referred to the Bombyces, especially of the Notodontidm, Ceratocampidm, Satiirniidie, Hemileu- cidm, Cochliopodidm, and Lasiocampidfe, will bring out still more striking and valuable results, inasmuch as they, or forms near them now extinct, are believed to be closely similar to the stem forms from which many of the higher Lepidoptera have probably been evolved.

The aim therefore in such studies should be

1. To treat the larvm as tliongh they were adult, independent animals, and to work out their specific and generic as well as family characters.

2. To trace the origin of mimetic and protective characters, and to ascertain the time of larval life when they are assumed, involving

3. The history of the development of the more specialized setae (hairs), spines, tubercles, lines, spots, and other markings.^

^Besides the work of Weismann, compare also the suggestive papers of E. B. Poulton, in Transactions of the Entomological Society of Loudon, 1884-1888, and my papers: Proceedings of the Boston Society of Natural History, xxiv-v, 1890-91.




4. To obtain facts regarding the ontogeny of our native species and genera wbicli, when added to what we know of the life histories of Europea]), Asiatic, and Soiitb American Bombyces, may lead to at least a partial comprehension of the i)hylogeny of the higher Lei^idoptera, viz, those above tbe so-called Microlei>idoptera.

The transformations of the Bombycine moths are especially noteworthy and useful for the purposes Ave have indicated, since the group is rich in stem forms, because of its probable geological antiquity, and because of the remarkable and significant differences i)resented by the larvic of many of the groups in the numerous successive stages of their larval life, these stages being characterized by distinctive and highly modified shapes, colot^s, markings, and armatimes. These peculiarities, signalizing nearly each stage, were, we believe, evolved in direct response to the changes in their environment, in their mode of life, or to changes in their food plants, and the necessity of being protected through unconscious mimicry from the assaults of insects and reptilian and avian enemies.

The transformations also afford the clearest i:)ossible evidence of the action of AAhat Darwin calls “inheritance at corresponding periods of life,” and which Ilmckel has tersely designated as “homochronic heredity.”

This fact, moreover, of inheritance at corresponding periods of life throws light on the problem so much under discussion at the present day of the transmission of characters acquired at different epochs during the life of the individual. We have devoted a section to a discussion of this question, or rather to a review of some of the facts which strongly suggest the truth of this principle.

The characters, so unexpected and striking, as for those worked out in Heierocampa hiunilata^ H. guttivitta, and ohliqiia, for example, as well as numerous other of the ^otodontiaus and allied families, are idainly enough useless to the insect in the pupa or imago condition, and have evidently been inherited as the result of impressions or stimuli received from without at different periods in the life of the caterpillar alone.

Such cases occur in many other Arthropods, especially in the barnacles, and in the Decai^oda,. as well as in the parasitic worms, but the causes can nearly as well be investigated in these insects, which are so accessible.

Another series of problems is opened up by a study of the mouth-parts of the Bombyces and of their venation, which disclose facts intimately bearing on the genealogy of the Lepidoptera.

, In no other Lepidoptera has the agency of use and disuse, particularly the latter, been more marked. While the mandibles are present in certain of the Tinema and Fgralklinaj they have totally disappeared from the so-called Macrolepidoptera, or higher and less generalized and primi- tive groups. In the Bombyces, particularly the Saturnians, the maxilhe, owing to disuse, have undergone great reduction, with complete loss of their original function. In another direction, i. e., in the veins of the wings, there has been a reduction in their number, and this is correlated with their loss of power of taking food, the great but weak wings of these colossal moths being of no use in seeking for food, which they do not need; as, unlike the swift visitors of dowei\s, the butterflies, Sphinges, and l!loctuids, they are too feeble of flight to sip the nectar of flowers, or too short lived to need any nourishment.

The geograj»hi(;al distribution of the Bombyces also tends to confirm the view that they are an ancient and generalized group, and to this subject we have given special attention.

In the systematic portion of the work I have endeavored to arrange the families, genera, and even the species, in accordance with the probable phylogeny of the group. I have begun my account of the entire superfiuuily with what I regard as the most primitive family. The seven subfamilies of i^otodoiitiaus easily fall into this arrangement; it is not difficult to lAerceive that the Gluphisiinm and Dataninm are the most generalized, and that the Cerurinm are the most spec- ialized, whether we study the larvm or imagines, though much the clearest light of course is thrown upon the subject by thelarvm. It is less easy to indicate the true succession of the genera, though the way is made very plain in the subfamily of IIeterocampiua\

The proper sequence of the species in a large genus is always difficult to make out. It is obvious, however, that the old, unphilosophic method of designating such and such a species as the type of a genus, and then arranging all the others under it, is a thoughtless procedure. Usually



the tyi^e species is the most modified, that most uulike its congeners, 'uuless, as is often the case, it happens to have been the first one of its genus to have been discovei'ed and described.

We have thought it better and more i)hilosophical to begin with that species whose larva is the most simple and generalized, and then arrange in their natural order those whose larva? are more and more si>ecialized or modified, as regards the number and variety of their markings, or the comiilexity of their armature. In the genus Ichthyura, for exam]»le, the larva of I. apicalis (van) is the most simple and generalized, not having the high tul)ercles and bright varied mark- ings of 7. hiclusa and albosigma. 1 have therefore supposed this to have been the first species to have evolved, and this decision is supported by the wide distribution of the species and the rather large number of varieties and subvarieties into which the form has been broken up.

In the case of the imago, that species which has plain wings without complicated bars and spots is more primitive than those with more coini)lex niurkings.

This course may at times lend to error and uncertainty, and involve more or less hypothesis or guesswork, but the simple attempt will lead to a more careful scrutiny of the larval character- istics, and to a profounder, more thorough, and better knowledge of the biology of the genus, and that of course is the aim in such work. Of course the systematic part of this or any other work of the sort is a necessary i)relimiuar3^ to all other higher endeavors to a complete history of the group from a morphological and biological point of view.

On this account it is, we think, a great pity that some of the compilers of our check lists of Lepidoptera and other insects, and of our zoological text-books and other works of the sort, still persist to cater to the tastes, rather than true needs, of amateurs and collectors by beginning at the wrong end, i. e., with the “highest’- forms rather than with the “lower” or more primitive. Such lists and works would have a far higher educational value and lead to much better mental training if such compilers could have had some knowledge of the immense impetus given to the science and the new waj' of dealing witli s^xstematic zoology which has resulted from the labors of Darwin, Fritz IMiiller, Weismann, and others.

Ill describing cateriiillars, particularly those of the Bonibyces, I have been particular to dis- tinguish between tlie three thoracic and the ten abdominal segments, because the former usually differ from tlie abdominal segments in the number, arrangement, and relative size of the tuhcrcles, warts, and other markings. The warts or tubercles also are grouped into dorsal, subdorsal, and supraspiracnlar rows (though this latter may in some cases be the subdorsal row), aud an infraspiracular row or series.

In order to obtain further material to finish and to perfect this monograph of the Bombyces, the author would like to obtain from collectors and students in all parts of the country’-, especially in the Southern, 'Western, and Pacific States, the egg, larvie, or moths, in order to fill up gai>s, as well as to afford material for illustration.

Should anyone rear any of these Bombyces, with a view to publication, I should be greatly obliged for alcoholic specimens of the eggs aud difierent larval stages, which might be sent after such descriptions were published.^ Such specimens would be carefully" kept and returned. It will only be by such cooperation that wo shall arrive at a fair knowledge of the transformations of this extensive group.

This monograph could uot have been i>repared without generous aid from friends aud cor- respondents, as well as from those in charge of the several museums mentioned below, whose hearty cooperation I now acknowledge.

I am specially indebted to Prof. G. Y. Eiley for the opportunity of freely examining from time to time his extensive collections, so rich in preserved larvin, both blown and alcoholic, the result of years of labor while residing in Illinois, St. Louis, and in Washington, D. G. After presenting them to the United States Natiou<al Museum, he has continued to allow me to examine the Bom- bj^ces, and loaned me specimens of larvie as w'ell as moths for study and illustration. He has also permitted the use of numerous colored sketches, made by himself or his assistants under his

^ It is earnestly hoped that anyone receivin’^ this memoir will kindly reciprocate by sending the eggs aud larv£e of any Bombyciiie moths not herein described, packed in tin boxes, to tbo author, at Providence, R. I., or during July and August, at Brunswick, Me. We still lack the eggs and young larva? of Ellida, Lophodonta, Drymonia, and Xotodoiita.



'direction, and has j^enerously turned over to me all his notes on transfonnatioiis, geographical distribution, etc., his contributions very much enhancing the value of this work. .

I am also indebted to the authorities of the American Museum of Natnralllistory, New York, for the opportunity of examining the types of the late Mr. IJemy Edwards, and a few types of Mr, Grote. Other material and types in the Museum of Comparative Zoology at Cambridge, Mass., of the Boston Society of Natural History, particularly the Harris collection, and the collection of the American Entomological Society at Philadelphia, have been examined, and to the authorities in charge I am specially indebted. I should also acknowledge the frequent aid rendered by Mr. Henry Edwards before his death, and the labors of those who have in former years done much pioneer work in collecting and describing the Boinbyces, especially of my friend, Mr. Aug. E. Grote, now of Bremen, Germany.

Mrs. Annie Trumbull Slosson, of New Y"ork, has-generously given me valuable material, and given me free access to her collection, and in this and other ways laid me under special obligations.

Mr. H. G. Dyar and Mr. B. Neumogeu have freely shown me their imi)ortant collections, and generously loaned specimens for illustration and study. Mr. Dyar has in a number of ways rendered most efficient aid, and has my hearty thanks. We have together made a number of 'Comparisons, and thus arrived at results which otherwise would have been less certain.

Dr. J. A. Lintner, State Entomologist of New York, has opened his collection to me, and loaned me several colored drawings of larvie.

From Eev. E. D. Hulst, of Brooklyn, I have received by exchange many specimens.

Dr. E. Thaxter has permitted me to examine his very valuable collection of larvre, now in the Cambridge Museum, and Professor French, of Carboiidale, 111., has also kindly heli)ed me. I am much indebted to Miss Emily L. Morton, of Newburg, N. Y., for eggs, larva*, and the use of several colored drawings of Dataiia larvm, etc., and for notes on their habits. To Miss Caroline E. Soule also I am under obligations for a line colored sketch of Nerice hifleniata-.

I am also indebted to the following entomologists who have aided me with larva*, eggs, moths, local lists, etc.: Mr. O. S. Westcott, Chicago, 111.; iMr, Tallant, Columbus, Ohio; Mr. Graef, Brooklyn, N. Y.; Mr. Trevor Kincaid, Olympia, Wash.; Mrs. Feruald, Amherst, Mass.; Mr. Charles Palm, of New York; Mr. William Beutenmueller, in charge of the collection of insects in the American Museum of Natural History, New York, and to others whose aid is acknowledged in the course of the work.

To Mr. Joseph Bridgham, who has made the drawings of the larval stages, I am under special ■obligations. Besides the work of drawing, lie has secured many of the larvie, and sliown the utmost pleasure iu aiding me to the extent of his ability. It is to be hoped that the work of the lithographer will bring out the delicacy of color and fidelity in drawing of the artist.

I have also had ten drawings of WalkePs tyjies iu the British Museum, made by Mr. H. Knight, of London, with the permission of Dr. A. Guenther, suiierinteudcut of the zoological department, to whom my hearty thinks are due; also for his courtesy iu allowing me, with the kind aid of Mr. A, G. Butler, assistant in entomology, to examine some of Walker's types.

I have also had copied iu the plates a number of excellent colored drawings of caterpillars, made by the late Maj. John Eatton Le Conte, which were loaned me for such a juirpose by his son, Dr. John Lawrence Le Conte, a few years before his death. They were made iu Georgia, presuma- bly at Sans Souci, on the Ogeechee Elver, about 16 miles south of Savannah.^

Brown Unta^ersity, Providence, E. I.

^See Sciitldor’s biograpliical skotcli of .1. L. Le Conte, Traus. A.iuer. Eiit. Soc., Aug., 1884, p. 9.





It is not improbable that, as a rule, all caterpillarvS at first lived on grasses, herbaceous and low-growing plants generally, and that gradually they began to climb trees, as the latter became developed, and in time became adapted to an arboreal station. As is well known, no deciduous trees or flowering plants aj^peared in such numbers as to form genuine foi'ests before the Cretaceous period, and about that time in geological history began to appear the kinds of insects which visit flowers and trees that blossom.

The species of the great lepidopterous family !N'octnid:e, of which we have in the United States alone over a thousand species, are, as a rule, low feeders. Certain species of Mamestra and of Agrotis, ordinarily feeding on grasses and low herbs, will however, especially early in the spring, ascend trees and shrubs of different kinds and temporarily feed upon the buds; and in summer a species of Nlamestra will ascend currant bushes in the night and cut off the young, fresh shoots.

In the group of forms represented by Catocala, Homoptera, and Pheocyina we have true tree inhabiting caterpillars, and, like the Notodontiaus and dendricolous Geometrids, their bodies differ remarkably from those of the low feeders, being variously spotted and mottled with shades of brown and ash, to assimilate them to the color of the bark of the tree they rest upon, and are, besides, provided with dorsal and lateral humps and warts, to further assimilate them, in outline as well as in color, to the knots and leaf-scales on the smaller branches and on the twigs among which they feed. And then there is the small group of Noctuo-bombyces, represented hy species of Apatela, Platycerura, Raphia, Charadra, and their allies, which closely ‘‘mimic” the hairy, lieuciled, or spiny arboreal Bombyces.^ It should, however, be observed that this is scarcely a case of mimicry, but rather of adaptation; the iiresence of hairs, pencils, spines, and bristles being apparently due to the caterpillars having changed their environment from herbs to trees, and being subjected to the same conditions as the Bombyces themselves.^

In the exclusively low feeding caterpillars of certain groups of butterflies the body is usually smooth and adorned with lines and spots, while the general feeders and many arboreal forms are often variously spiued and tuberculated, yet many S])ined caterpillars of butterflies feed on low herbs.-* The Sphingidm in part feed on low plants and in part on trees, and they do not, except as regards the caudal horn, exemplify our thesis.

‘This section is reprinted with some .alterations from an article in the Proceedings of the Boston Society' of Natural History, xxir, 1890, pp. 482-515, 556-559.

2 Of 34 species of North America Noctiio-homhyces, whose transformations are known, all except 1 feed uimn trees. (See Edwards's catalogue.)

®It is hardly necessary for ns to express onr entire disagreement with the view of Mr. A. G. Butler, that these Noctuidic are really Notodontians, or in any way allied to tliem. It seems to ns that the characters w’hich lie uses.to remove them from the Noctnidm are superficial and adaptive. Nearly twenty-five years ago 1 satisfied myself, after an examination of the denuded head and wings, that the Noctuo-bombycos were true Noctuidie, and did not depart essentially from the tyincal genera.

■‘While many, tliougU not all, butterfly larvfp, as shown by Scudder and W. H. Edwards, have spine-like gland- ular hairs in the flrst stage, which may in some cases persist into one or two later stages, the body in many species, especially in those which are not general feeders, but select low-growing, herbaceous plants, becomes smooth and ornamented with stripes or spots. However, as a rule, butterfly larva* can not be divided, as the Bombyces, etc., into high and low feeders; yet from Scudder’s “Classiliedlist of food plants of American butterflies'' (Psycho, 1889) the following facts and conclusions may be stated:

I/esperida'. Out of 45 species enumerated, all but 6 feed on lierba and especially on grasses, and those which feed on tall shrubs or trees, such as J^pargyreus lityrus and 5 species of Thanaos, stand at the head of the group, which, as everybody knows, is the lowest family of butterflies and nearest related to the moths.

rajnUonida’. Of the 6 species enuiueratod, 3 feed ou trees as well as shrnbs and herbs; 1 of these, however (P. cresjihontefi), feeds ou trees alone. None of this family are hairy or spined when mature, except P. philenor, wdth its peculiar flexible, spike- like growths.

Pierina:. Of 10 species, all feed on herbs, rarely ou low shrubs, and none arc armed with hairs, bristles, or spines. The other two groups {LycwnUlw and Xyinphalidw) are general feeders, occurring indifferently on herbs, vines, and trees, except the striking case of the 8 Satyrinie, which feed exclusively on grasses and herbs (E. ^yortlandia, however, sometimes freipieuting the Celtis). The very spiny Argyunis larvcc feed ou Viola. It should also be noted



Of the great group of GeometridiB many kinds are arboreal (Dendrogeometrids), and in sucli cases are almost invariably tuberciilated in manifold ways. We know of no liairy or tufted caterpillars of this group or of any family below tbeiUj with the exception of the Pterophoridie,

The arboreal Pyralidie, Tortricid:e, and Tineida* live in such concealment, between leaves, or in buds, or as miners, that they differ little in their surroundings from the low-feeding forms, and are thus scarcely ever tubercidated or spiny; in fact, we can not recall one of these groups which are s^. The Pterophoruhe are, to be sure, spiny, but they are low feeders, and their peculiar excretory setai (the Driiseuharcheu or giandular hairs of Zeller^) are similar, as Diminock has observed, to the glandular or long hairs of plants; Miss Murtfeldt adding that “there is a very close imitation in the dermal clothing of the larvie [of LeioptUus Hericidactylus] to that of the young leaves of Veruoiiia, on which the spring and early summer broods feed.” (Psyche iii, 390, 1882.)

Eeturning to the Bombyces, all the IS'otodoutians, without any exception, known to us have trees as their principal, if not exclusive, food plants. Thus, of the 37 species of this group whose larval forms ai'e known, and whicli are enumerated in Mr. H. Edwards’s “Bibliographical catalogue of the described transformations of North American Lepidoptera,” together with an additional species (lehtJujura sirigosa) omitted from the catalogue, all are known to feed on trees, unless we except Datana mojor^ which feeds on Andromeda. It is noteworthy that the only si)ecies found thus fill' on a herbaceous plant is the caterpillar of Apatelodes torrefacta^ which Harris found on the burdock, though usually it is an arboreal insect. This apparently omnivorous feeder resembles the species of Halesidota, all of which occur more commonly on trees than on herbs, and thus differs markedly from the majority of the Lithosiaus and Arctians, unless we except the Nolidte, Now the larva of Apatelodes is hairy, the long, white hairs having scattered among them black ones, with more or less black pencils, thus resembling the peculiar yellowish or white caterpillars of Halesidota, with their black tufts and i^eucils. Similar forms are some of the arboreal, hairy Noctuidm, as Gharaclra deridem. It seems evident that the resemblance to each other in such different groups is the result simply of adaptation, brought about by two factors, the i)rimary one being a change from a low- feeding to an arboreal station, and consequent isolation or segregation, and the secondary one being natural selection, the latter farther tending to pre- serve the specific form.

It will be seen b}’’ the following review that the North American Bombyces in general, with the excexition of the Aixtians and Litliosiaus, live on trees, and this will in general apply to the Old World species. In the group of Lasiocampuhe, represented by Tolype,' Artace, Heterocampa, Gastropacha, and Clisiocampa, the station is an arboreal one, none being known to feed on herbaceous idants. x\ll the Geratocamx>idie, all the neniileucidm and Attaci, the Platyptericidiu,- all the Cochliopodidic (Limacodes), including both the naked and spiny genera, as well as the Psychidic, live exclusively on trees. Of our North American Liparidie, all are arboreal in station, except the Galiforniau Orgyia vetustay which lives on the luijiue. Finally we come to the Arctians- and Lithosiaus, whose hairy, or rather setose, larvm in general feed on herbaceous plants and sometimes on trees, being in many cases omnivorous, wliile those of the Nolidm and Nycteolidm whose history is known, are arboreal.

Of the Zygicnidm, including the Gteuuchida}, the sx)ecies are low feeders, living on lichens, - grasses, and other low plants, or upon vines. The Dioptid genus Phryganidia feeds on the oak. Of the Agaristidm, some are low feeders, EiiSGirrliopteriis gloveri feeding on Portulaca, while the majority prefer vines (Vitis, etc.). As to the boring habits. of the Hepialidm and Cossidm, which we now consider as independent groups, related to the Tineina, rather than belonging to the superfamily Bombyces, these seem to be the result of early adaptation.

An examination of the food plants of the British species of Bombyces, taken from Stainton’s- Manual of British butterflies and moths (1857), gives the same results for the Old World, as will be seen by the following statements:

that mauy moths, Notoilontiaus among them, Trhich in the Northern States feed on trees alone, in the Gulf States^ according to Abbott, feed on shrubs, vines, and low plants, as well as trees.

In rei>ly to an inquiry, Mr. W. H. Edwards kindly writes me: ‘‘1 do not think that the butterfly larvte which, live on trees are under more favorable conditions than low feeders as to healthiness or ease of rearing/’

^Revision der Pterophorideu. Liuniea entom., 1852, vi, 358. Mentioned by Diminock.



Wolulw, Of the 3 Britisli species, 2 feed on tlie oak and 1 on tlie liawtliorn and sloe.

Lijparidw. Of the 12 species, all feed on trees and shrubs, except Lcvlia ccoiosa, ^\diich lives on reeds and other water plants. It is tufted.

Kotodonikhv, Of 24 species, 1 ( DiJoha ewruleocepliala, which is smooth, with no protuberances) feeds on the hawthorn and other plants.

Fkiti/pierwidw.—Of the 0 species, 5 feed on trees and 1 on a shrub.

Endromidw. The single sptcies is arboreal.

Psychidw, The 2 species, whose larval habits were known, feed on trees and shrubs.

CochUopodida\ The 2 species feed on trees.

Saturniida\ The single British species feeds on the heather, a shrubby plant.

Lasiocampida\ Of 11 species, 5 feed on trees, the others on shrubs and herbs.

NocUio homhyces, All the British species are reported as ‘‘living on trees and shrubs quite exposed.'^

Bomhycoid(i\ All the species of Acronycta live on trees and shrubs.

Inflxtence of a change from low to high feeding plants^ L e., from Uring on an herhaceons to an -arboreal ntation, It appears, then, that the more typical Bombyces, such as the Ceratocanipida?, Ilemileucidic, Attaci, Notodontians, Cochliopodidm, and Liparida?, are arboreal in their station, their bodies being variously protected by spines, spinulated tubercles, hairs, or tufts. The grouj) is indeed i)articnlarly distinguished for the manifold modifications undergone by what are morpho- logically setic, and it is an interesting inquiry whether the great development of these spines and liairs may not have originally resulted from some change in environment, such as that from low-feeding to high-feeding or arboreal habits.

It may be objected that the setm and spines were originally due to the stimulus arising from the attacks of parasitic insects, such as ichneumons and Tachime, or that, as hairy caterpillars are not usually devoured by birds, these hairs and spines have originated through natural stdec- tiou, and are danger signals, indicating to birds that the Avearers of such hirsute and bristling armature are inedible. But Avhile the final purpose or ultimate use of such an arinatiire may serve the useful purpose of protection, and while natural selection may have been the leading secondary factor in the iireservation of varietal and specific forms of hairy and spiny caterpillars, this does not satisfactorily account for the initial causes of the growth of tubercles, spines, etc.

If spines and hairs form hedge-like guards against the attacks of parasitic insects, why are they not devehqied as well in the great multitude of low feeders as in the less numerous high feeders ? It may be said, however, that Euprepia caja is more subject to the attacks of ichneumons than almost any other larvae. (A, G. Butler in Ann. and Mag. Nat. Hist., 1891.) Everyone knows how efficacious any hairs or bristles are in deterring ichneumons and Tachinm from ovipositing on caterpillars, and it is well known that naked or slightly piliferous larvm are more subject to their attacks than those which are densely hairy or spiuose.

The eniclfornt type of (arva\ In eudeavmring to account for the origin of the tubercles and spines, as well as the hairs of caterpillars, let us glance at the probable causes of the origin of the caterpillar form, and of the more primary colors and markings of the skin.

It was Fritz Miiller who, in his Fiir Darwin (18G4), maintained that “the so-called complete metamorphosis of insects, in which these animals quit the eggs as grubs or caterpillars, and afterwards become quiescent pupie, incapable of feeding, was not inherited from the primitive ancestor of all insects, but acquired at a later period.” *

In 1869 Dr, F. Brauer^ divided the larvm of insects into two groups, the campodca form and raupen form, and in 1871^-1873 we adopted these suggestive views, giving the name of cruciform to the larvje of weevils and other coleopterous larvie of cylindrical form, as well as to the larvic of Diptera, Lepidoptera, and Hymenoptera, all of which are the result of adaptation, being derivatives of the primary campodea type of larvm. BraueFs views on these two types of larvte were also adopted by Sir John Lubbock in his Origin and IMetamoiqdioses of Insects, 1873.

^ Facts aud Argumouts for Darwin, with additions by the author. Translated from the German by VV. S. Dallas, y. h. S., Loudon, 1869.

^Betrachtungeu iiber die V^orwaudlung der lusekteu im Siune der Descendenztheorie. V^orh. K. K. Zool. bot. Oes. Wien, 1869.

^Embryology of Chry802)a. American Naturalist, Sejit., 1871.



While the origin of the eruciform larvje of the Ceraiuoycidiej Curculionidie, Seolytidie, aud other wood-boring and seed-inhabiting and burrowing Coleopterous larva3 in general, is plainly attributable to adaptation to changed inodes of life, as contrasted with the habits of roving,, carnivorous, cam pod ei form larvae, it is not so easy to account for the origin of the higher metabolous orders of Diptera, Lepidoptera, and Hymenoptera, whose larvie are all more or less eruciform. We are farced to adojit the supposition that they have independently originated from groups either belonging to the ^5’euroi)tera (in the modern sense) or to some allied but extinct group.

Restricting ourselves to the Lepidoptera; as is well known the Lepidoptera are now by some- believed to have descended from the Trichoptera or from forms allied to that group. We should, however, iirefer the view that the Lepidoptera, Trichoptera, and Mecoptera had a common origin fiom some earlier, extinct group. The similarity of the imagines of certain of the lower Tineina and certaiu of the smaller Trichoptera is certainly very marked, the most significant feature being- the fact that the mandibles in the two groups are either absent or minute aud rudimentary.

We have attempted, however,^ to show that the larvm of the Panorpidie, judging fromBrauer’s figures aud descriptions, are much nearer in shape and ornamentation to caterpillars than to case Avorms, Hence, it seems to us probable that the ancestral or stem form of the Lepidoptera. was probably a now extinct group, somewhat intermediate between the Mecoptera (Panor])idie) and the Trichoptera.

The prmitivG caterpillar, We Avould suggest that the earliest type of Lepidopterous larva Avas allied to some Tineoid which lived not only on land but on Ioav herbage, not being a miner or sack-bearer, as these are evidently secondary adaptive forms. It is evident, when Ave take into account the remarkable changes in form of certain mining Tineoid larvm described and figured by Chambers 2 and by Dimmock,^ that the flattened, footless, or nearly apodous mining larvie of the earlier stages are the result of adaptation to their burrowing habits. The generalized or primitive form of the first caterx)illar Avas, then, like that of Tineoid larvie in general, aud Avas an external feeder rather than a miner. The body of this forerunner or ancestor of our present caterpillars- (which may have lived late in Carboniferous times, just before the appearance of flowering plants and deciduous trees) was most probably cylindrical, long, aud slender. Like the Panorpid larvie, the thoracic and abdominal legs had already becom differentiated, and it differed from the larvie of Panorpids in the plautie of the abdominal legs being provided Avith j)erhaps tAvo pairs of crochets, thus adapting them for creeping with security over the surface of leaves and along twigs and branches. The x)rothoracic or cervical shield Avas present, as this is axiparently a primitive feature, often reaxApearing in the Noctuidie, aud sometimes in the Bombycina, and ahvays lAresent in the boring larvie of the Hexnalida; and the Cossidte.

As tactile hairs, defensive or locomotive setie, aud si)iues of manifold shax^es occur in AA'orms, often arising from fleshy Avarts or tubercles, it is reasonable to assume that the xnliferous Avarts of lepidoiAterous larvm are a direct heirloom of those of the vermian ancestors of the iusects. In our primith’-e caterpillar, then, the x>iliferous Avarts Avere i»resent, eventually becoming arranged as they now are in ordinary Tineoid, Tortricid, Pyralid, Geometrid, and Xoetuid larvie.

Origin of the green color of caterpillars. The cuticle may at first, as in that of caseworms and Panorind larvie, ha\^e been colorless or horn colored. But soon after habitually feeding in the direct sunlight on green leaves, the chlorophyll thus introduced into the digestive s^^stem and into the blood and the hypodermal tissues would cause the cuticle to become green. AfterAAUirds, by further adaptation aud by heredity this color would become the hue in general common to- caterpillars. Moreover, some of the immediate descendants of our jwimithm cateri)illars were probably lighter in hue than others; this was probably due to the fact that the lighter-colored ones, fed on the x^fil^-g'reen underside of the lea\^es, this difference becoming transmitted by heredity.

Third Report U. S. Entomological Commission. Genealogy of the Hexapoda, pp. 297-299, 1883. Also American Naturalist, Sept., 1883,932-945.

2 American Entomologist, iii, 1880, 255-262; Psyche, ii, 81, 137-227; iii, 63, 135, 147; iv, 71. Refers to the larvm- of the •‘Gracilaridie^' aud “Lithocolletidie^^ together Avith Phyllocnistis.

3 Psyche, iii, Aug., 1880, 99-103.

•* See the important and quite conclusive footnote by Professor Meldola on j). 310 of AA’^eismaim’s StuiUes in the- Theory of Descent. Vol. i (‘^I have already given reasons for suspecting that the color of green caterpillars may he: due to the jiresence of chlorophyll in their tissues, Proc. Zool. Soc., 1873, 159. R.



Origin of the lines, As Weismann lias shown, tlie primitive markings of caterpillars were lines anti loiigitmliual bands, the spots appearing from interruptions or what may be called the serial atrophy of the lines or bands. It is not difficult to account for the origin of the dorsal line, as this would naturally be due to the presence of the heart underneath. This dorsal line is, for example, wanting in the freshly hatched larvie of Spilosoma virginica and Hgphantria iextor^ but after the first molt of S, virginica there is a slight, diffuse dorsal line of no decided color, though after the second ecdysis it is decidedly whitish, or at least much jialer than the surrounding dorsal region. In pale catei'iiillars the dorsal line may be darker. In the first stages of the two moths in question there are no lines or bands j only the luliferous warts. Whether the subdorsal or the spiracular lines were the first to origiuale is uncertain, but probably, from Avhat Weismann has concluded from his studies of the Sphingidte, the subdorsal arose first. In the second stage of' Spilosoma virginica the subdorsal lines are reddish lines extending between the two subdorsal rows of alternating subdursal piliferous warts, the line becoming more decided, however, in the third stage of this species, there being as yet no signs of a spiracular or of any lateral line. In the freshly hatched larva of JI. textor^ however, what maybe the first beginnings of the subdorsal line are elongated brownish linear spots inclosing the subdorsal row of larger piliferous dots, but not reaching the sutures between the segments. These patches, however, do not in the second stage unite to form continuous lines, but two rows of decided black elongated spots inclosing the black i>iliferous tubercles. In the freshly hatched larva of Edema alhifrons each of the two subdorsal lines is a row of elongated black spots connected on the three thoracic segments, but separated by the sutures along the abdominal segments.

The si)iracular line is seen in the same larva of the same stage to be a yellowish baud inclosing the spiracles, and there seems to be a tendency in some, if not many, larvie for the spiracles to be inclosed and connected by a parti-colored or bright line, and for this to have a darker (as in Edema} or lighter edging. Why the spiracles themselves are so apt, as in Bombyces and Sphinges, to be- inclosed by a dark or cousi>icuous line remains to be explained.

To return to the subdorsal lines in the pale-reddish larva of Datana, probably D, integerrima^. these lines before the first molt are also inclosed by the two rows of subdorsal j^iliferous si)ots, and in both the first and second stages there are pale spiracular lines, which appear to be contem- poraneous with the subdorsal line. In the third stage a new dark-red line is interpolated between the subdorsal and spiracular. In the fourth stage the spiracular line has disappeared, and there is a supra and an infra* spiracular pale line on the now brown, dark skin of the caterpillar. Seen fi’om above there are four pale lilac lines, but after molting two of them disappear, and in the last stage there are only two subdorsal lines to be seen, if iny colored drawings, very carefully made by Mr. Brigham, are correct. We thus see that after the subdorsal and spiracular lines are formed, others are rapidly introduced and some may as rapidly vanish, as necessary features of' certain stages which, when they become useless are discarded.

The admirable and most suggestive work of Weismann has placed on a sound basis the theory of the origin of the lines, bands, and spots of the Sphingidie. The additional notes by Professor Meldola and the beautiful researches of Mr. Poultoii have abided to the strength of the arguments of Weismann. The lines, bars, stripes, spots, and other colorational markings of caterpillars, by which they mimic the colors and shadows of leaves, stems, etc., have evidently been in the first ])lace induced by the nature of the food (chlorophyll), by the effects produced by light and shade, by adaptation to the form of the edge of the leaf, as in the serrated back of certain Notodontians, by adaptation to the colors of different leaves and to the stems, often reddish, shades of greens, yellows, reds, and browns being as common in the cuticle of caterpillars as on the surface or cuticle of the leaves and their stems or in the bark of the twigs and bi'anches. We (and jirobably others) have obsei'ved that the x^eculiar brown spots and patches of certain Notodontians do not appear until, late in larval life, and also late in the summer or early in the autumn contemporaneous with the appearance of dead and sere blotches in the leaves themselves.

Now, to say that these wonderful adaptations and marked changes in the markings of cater- l)illars are due to ‘‘natural selection,” and to let the matter rest there, is quite unsatisfactory. Natural selection may account for the elaboration of these larval forms with their markings after- they have once appeared, but we Avant to discover, if possible, the original causes of such orna-



mentation, i. e., the priniiiry factors concerned in tlieir evolution. Weisnuinn in his earlier work repeatedly asserts that these changes are due to the direct action of external conditions together with natural selection. AVithin a few years ])ast many naturalists have returned to a more profound study of the causes of variation along some of the lines vaguely pointed out hy Lamarck.* It is noteworthy that Darwin changed his view's somewhat in his A'ariation of Animals and Plants under Domestication, and laid more stress on the intiueuce of the surroundings than in his Origin of Species.

Neither Weisinann nor other authors, however, so far as we know, haA^e formally discussed the i^robable mode of origin of humps, horns, tubercles, sx)ines, and such outgrowths in larva3. They are so marked and so manifold in their variations in form, and so manifestly related, and in fact have so evidently been directly developed by adaptation to changes in the habits of the Notodontian caterpillars and tree-feeding larvai in general that this group affords favorable material for a study of the general iwoblem.

Spines and ])rickles in animals, like those of idants, serve to protect the organism from external attack, and also to strengthen the shell or skin; they are adaptive structures, and have evidently arismi in response to external stimuli, either those of a general or of a cosmical nature, or those resulting from the attacks of animals. It is almost an axiomatic truth that a change of habit ill the organism precedes or induces a change of striictiiic.

What has caused the enlargement and specialization of certain of the piliferous warts ? As remarked by Sir James Paget, Constant extrapressiire on a part always appears to produce atrophy and absorption; occasional iiressnre may, and usually does, produce hypertroiJiy and thickening. All the thickenings of the cuticle are the consequences of occasional pressure, as the pressure of shoes in occasional AA'alking, of tools occasionally used with the hand, and the like, for it seems a necessary condition for hypertrophy, in most parts, that they should enjoy intervals in which their nutrition may go on actively.’^ (See I^ectures on Surgical Pathology, I, p. 89, (pioted