Studies in the Theory of Descent (Volumes 1 and 2)

by August Weismann

STUDIES IN THE THEORY OF DESCENT.

---

Part II.
ON THE FINAL CAUSES OF TRANSFORMATION.

---

I.
THE ORIGIN OF THE MARKINGS OF CATERPILLARS.

---

INTRODUCTION.

The general idea which has instigated the researches described in the present essay has already been expressed in the Preface, where it has also been explained why the markings of caterpillars, and especially those of the Sphinx-larvæ, were chosen for testing this idea.

The task presented itself in the following form:—In order to test the idea referred to, it must be investigated whether all the forms of marking which occur in the Sphinx-larvæ can or cannot be traced to known transforming factors.

That natural selection produces a large number of characters can be as little doubted as that many varying external influences can bring about changes in an organism by direct action. That these two transforming factors, together with their correlatively induced changes, are competent to produce all characters, howsoever insignificant, has indeed been truly asserted, but has never yet been proved. The solution of the problem, however, appeared to me to depend particularly on this point. We are now no longer concerned in proving that a changing environment reacts upon the organism—this has already been shown—but we have to deal with the question whether every change is the result of the action of the environment upon the organism. Were it possible to trace all the forms of markings which occur, to one of the known factors of species transformation, it could be thus shown that here at least an “innate power of development” was of no effect; were this not possible, i.e. did there remain residual markings which could not be explained, then the notion of an “innate principle of development” could not be at once entirely discountenanced.

The attempt to solve this problem should commence by the acquisition of a morphological groundwork, so that the phyletic development of the markings might by this means be represented as far as possible. It cannot be stated with certainty, primâ facie, whether some form of development conformable to law is here to be found, but it soon becomes manifest that such is certainly the case in a great measure. In all species the young caterpillars are differently marked to the adults, and in many the markings change with each of the five stages of growth indicated by the four ecdyses, this gradational transformation of the markings being a “development” in the true sense of the word, i.e., an origination of the complex from the simple, the development of characters from those previously in existence, and never an inconstant, unconnected series of per saltum changes. This development of the markings in individuals very well reveals their phyletic development, since there can be no doubt but that we have here preserved to us in the ontogeny, as I shall establish more fully further on, a very slightly altered picture of the phyletic development. The latter can have been but slightly “falsified” in these cases, although it is indeed considerably abbreviated, and that in very different degrees; to the greatest extent in those species which are most advanced in their phyletic development, and to the least extent in those which are less advanced. From this the value of being able to compare a large number of species with respect to their ontogeny will appear. Unfortunately, however, this has only been possible to a very limited extent.

The youngest larval stages are those which are of the most importance for revealing the phyletic development, because they make us acquainted with the markings of the progenitors of the existing species. For these investigations it is therefore in the first place necessary to obtain fertile eggs. Female Sphingidæ, however, do not generally lay eggs in confinement,[66] or at most only a very small number. In the case of many species (Deilephila Galii, D. Lineata, D. Vespertilio, D. Hippophaës) I have for this reason unfortunately been unable to observe the entire development, and such observations would in all probability have given especially valuable information.

I was certainly successful in finding the young larvæ of some of the above as well as of other species on their food-plants, but even in the most favourable instances only individuals of the second stage and generally older. When, however, notwithstanding this imperfection of the materials, and in spite of the important gaps thus inevitably caused in these series of observations, it has nevertheless been possible to form a picture, on the whole tolerably complete, of the phyletic development of the Sphinx-markings, this well indicates what a fertile field is offered by the investigation of this subject, and will, I trust, furnish an inducement to others, not only to fill up the various gaps in the small family of the Sphingidæ, but also to treat other Lepidopterous families in a similar manner. Such an investigation of the Papilionidæ appears to me to be especially desirable; not only of the few European but also of the American and Indian species. We know practically nothing, of the youngest stages of the Papilio larvæ from this point of view. No entomological work gives any description of the form and marking of the newly hatched larvæ, even in the case of our commonest species (Papilio Machaon and P. Podalirius), and I believe that I do not go too far when I assert that up to the present time nobody has observed them at this early stage.[67] When, however, we consider that in these young caterpillars we have preserved to us the parent-form, extinct for centuries, of the existing species of Papilio, it must assuredly be of the greatest interest to become accurately acquainted with them, to compare them with the earliest stages of allied species, and to follow the gradual divergence of the succeeding stages in different directions, thus forming a picture of the phyletic development of an evolving group. In the course of such observations numerous collateral results would doubtless come out. Investigations of this kind, whether conducted on this or on any other group, would, above all, show the true systematic affinities of the forms, i.e., their genealogical affinities, and that in a better way than could be shown by the morphology of the perfect insects or the adult caterpillars alone. If I am diffident in founding these conclusions upon the development of the Sphinx-markings treated of in the present essay, this arises entirely from a knowledge of the imperfections in the basis of facts. If however, through the united labours of many investigators, the individual development of all the species of Sphingidæ now existing should at some future period be clearly laid before us, we should then not only have arrived at a knowledge of the relative ages of the different species, genera and families, but we should also arrive at an explanation of the nature of their affinities.

It is erroneous to assert that Classification has only to take form-relationship into consideration; that it should and can be nothing else than the expression of form-relationship. The latter is certainly our only measure of blood-relationship, but those who maintain the assertion that form- and blood-relationship are by no means always synonymous, are undoubtedly correct. I shall in a future essay adduce facts which leave no doubt on this point, and which prove at the same time that modern systematists—especially in the order Lepidoptera—have always endeavoured—although quite unconsciously—to make the blood-relationship the basis of their classification. For this reason alone, larvæ and pupæ would have an important bearing upon the establishment of systematic groups, although certainly in a manner frequently irregular.

It must be admitted that so long as we are able to compare the species of one group with those of another in one form only, we are often unable to ascertain the blood-relationship.[68] In such cases we can only determine the latter from the form-relationship, and as these are not always parallel, any conclusion based on a single form must be very unsound. If, for instance, butterflies emerged from the egg directly, without passing through any larval stage, a comparison of their resemblances of form would alone be of systematic value; we should unite them into groups on the ground of these resemblances only, and the formation of these groups would then much depend upon the weight assigned to this or that character. We might thus fall into error, not only through a different valuation of characters but still more because two species of near blood-relationship frequently differ from one another in form to a greater extent than from other species. We should have no warrant that our conception of the form-relationship expressed the genealogical connection of the species. But it would be quite different if every species presented itself in two or three different forms. If in two species or genera the butterflies as well as the larvæ and pupæ exhibited the same degree of form-relationship, the probability that this expressed also the blood-relationship would then be exceedingly great. Now this agreement certainly does not always occur, and when these different stages are related in form in unequal degrees, the problem then is to decide which of these relationships expresses the genealogy. This decision may be difficult to arrive at in single cases, since the caterpillar may diverge in form from the next blood-related species to a greater extent than the butterfly, or, conversely, the butterfly may diverge more widely from its nearest blood-related species than the caterpillar.

For such cases there remains the developmental history of the caterpillar, which will almost always furnish us to a certain extent with information respecting the true genealogical relationship of the forms, because it always reveals a portion of the phyletic (ancestral) development of the species. If we see two species of butterflies quite dissimilar in form of wing and other characters, we should be inclined, in spite of many points of agreement between them, to place them in entirely different genera. But should we then find that not only did their adult larvæ agree in every detail of marking, but also that the entire phyletic development of these markings, as revealed by the ontogeny of the larvæ, had taken precisely the same course in both species, we should certainly conclude that they possessed a near blood-relationship, and should place them close together in the same genus. Such an instance is afforded by the two Hawk-moths, Chærocampa Elpenor and C. Porcellus, as will appear in the course of these investigations. These two species were placed by Walker in different genera, the form-relationship of the imagines being thus correctly represented, since Porcellus (imago), is indeed more nearly related in form to the species of the genus Pergesa, Walker, than to those of the genus Chærocampa.[69] Nevertheless, these species must remain in the same genus, as no other arrangement expresses their degree of blood-relationship.

An intimate knowledge of the development-stages of caterpillars thus offers, even from a systematic point of view, an invaluable means of judging the degree of blood-relationship, and from this standpoint we must regard the study of the caterpillar as of more importance than that of the perfect insect. Certainly all groups would not be so rich in information as the Sphingidæ, or, as I am inclined to believe, the Papilionidæ, since all families of caterpillars do not possess such a marked and diversified pattern, nor do they present such a varied and characteristic bodily form. The representation of the true, i.e., the blood-relationship, and through this the formation of natural groups with any completeness, can certainly only be looked for when we are intimately acquainted with the different stages of development of the larvæ of numerous species in every group, from their emergence from the egg to their period of pupation. The genealogical relationship of many forms at present of doubtful systematic position would then be made clear. This investigation, however, could not be the work of a single individual; not only because the materials for observation are too great, but, above all, because they are spread over too wide a field. It is not sufficient to study the European types only—we should endeavour to learn as much as possible of the Lepidoptera of the whole world. But such observations can only be made on the spot. Why should it not be possible to trace the development from the egg, even under a tropical sky, and to devote to breeding and observing, a portion of that time which is generally spent in mere collecting? I may perhaps be able to convince some of the many excellent and careful observers among entomologists, that beyond the necessary and valuable search for new forms, there is another field which may be successfully worked, viz., the precise investigation of the development of known species.

The first portion of the present essay consists of the determination of this development in those species of Sphingidæ which have been accessible to me. Seven genera are successively treated of, some completely, and others only in some of their stages; and thus I have sought to present a picture of the course of development of the markings in each genus, by comparing the species with each other, and with allied forms in cases where the young stages were unknown. In this portion, as far as possible, the facts only have been given, the working up of the latter into general conclusions upon the development of marking being reserved for the second portion. A complete separation of facts from generalizations could not, however, be carried out; it appeared convenient to close the account of each genus with a summary of the results obtained from the various species.

After having established that the markings of the Sphinx-caterpillars had undergone an extremely gradual phyletic development, conformable to law, in certain fixed directions, it appeared desirable to investigate the causes of the first appearance of these markings, as well as of their subsequent development. The question as to the biological significance of marking here presented itself in the first place for solution, and the third section is devoted to this subject. If it is maintained that marking is of no importance to the life of the insect, or that it is so only exceptionally, and that it is in reality, as it appears to be, a character of purely morphological, i.e., physiological, insignificance, then its striking phylogenetic development conformable to law cannot be explained by any of the known factors of species transformation, and we should have to assume the action of an innate transforming power. In the present investigations, this subject in particular has been extensively treated of, and not only the markings of Sphinx-caterpillars, but also those of caterpillars in general, have been taken into consideration. The results arrived at are indeed quite opposed to this assumption—marking is shown to be a character of extreme importance to the life of the species, and the admission of a phyletic vital force must, at least from the present point of view, be excluded. This leads to the fifth section, in which I have attempted to test certain objections to the admission of a “phyletic vital force.” The sixth section finally gives a summary of the results obtained.

I may now add a few explanations which are necessary for understanding the subsequent descriptions. It was found impossible to avoid the introduction of some new technicalities for describing the various elements of larval markings, especially as the latter had to be treated of scientifically. I have therefore chosen the simplest and most obvious designations, all of which have already been employed by various authors, but not in any rigorously defined sense. I understand by the “dorsal line” that which runs down the middle of the back; the lines above and below the spiracles will be respectively distinguished as the “supra-” and “infra-spiracular” lines, and the line between the dorsal and spiracular as the “subdorsal line.” The distinction between “ring-spots” and “eye-spots” will be made manifest in the course of the investigation. A glance at any of the existing descriptions of larvæ will show how necessary it was to introduce a precise terminology. Even when the latter is exact as far as it goes, the want of precise expressions not only makes the descriptions unnecessarily long, but it also considerably increases the difficulty of comparing one species with another, since we can never be sure whether the same designation applies to the same homologous character. For instance, when the larva of Chærocampa Elpenor is said to have “a light longitudinal line on the sides of the thoracic segments,” this statement is indeed correct; but it is not apparent whether the line is above or below, and consequently it does not appear whether it is the equivalent of the longitudinal line “on the sides” of the segments in other species. If, however, it is said that this line is “subdorsal on the thoracic segments, and on the eleventh abdominal segment,” it is thereby indicated that we have here a residue of the same marking which is found completely developed in many other Sphinx-larvæ, and indeed in the young stages of this same species. The mode of describing caterpillars hitherto in vogue is in fact unscientific; the descriptions have not been made with a view to determining the morphology of the larvæ, but simply to meet the practical want of being able to readily identify any species that may be found: even for this purpose, however, it would have been better to have employed a more precise mode of description.

---

I.
Ontogeny and Morphology of Sphinx-Markings.

THE GENUS CHÆROCAMPA, DUPONCHEL.

Although by no means in favour of the excessive subdivision of genera, I am of opinion that Ochsenheimer’s genus Deilephila has been correctly separated by Duponchel into the two genera Chærocampa and Deilephila, sensû strictiori. Such a division may appear but little necessary if we examine the perfect insects only; but the developmental history of the caterpillars shows that there is a wide division between the two groups of species, these groups however being branches of one stem.

Chærocampa Elpenor, Linn.

Some captured females laid single eggs sparsely on grass, wood, and especially on the tarlatan with which the breeding-cage was covered. The eggs are nearly spherical, but somewhat compressed, of a grass-green colour, a little lighter, and somewhat larger (1.2 millim.) than those of Deilephila Euphorbiæ. During the development of the embryo the eggs first became yellowish-green, and finally yellowish.

First Stage.

The young caterpillars are four millimeters in length, and immediately after hatching are not green, but of a yellowish-white opalescent colour, the large and somewhat curved caudal horn being black. The caterpillars were so transparent that under a low magnifying power the nervous, tracheal, and alimentary systems could be beautifully seen. As soon as the larvæ began to feed (on Epilobium parviflorum) they became green in consequence of the food appearing through the skin, but the latter also gradually acquired a dark green colour ([Pl. IV]., Fig. 17). All the specimens (some twenty in number) were exactly alike, and showed no trace of marking.

Second Stage.

The first ecdysis occurred after 5–6 days, the length of the caterpillars being from nine to ten millimeters. After this first moult they appeared of a shining green, the horn, which was black during the first stage, becoming a little red at the base, while a fine white subdorsal line extended from the horn to the head ([Fig. 18]). The head and legs were green; the divisions between the segments appeared as fine light rings, and the entire upper surface of the segments was also crossed by fine transverse rings, as was also the case in the first stage.

At the beginning of the present stage no trace of the eye-spots could be detected; but a few days after the first moult it was observed that the white subdorsal line was no longer straight on the fourth and fifth segments, but had become curved upwards into two small crescents. The latter soon stood out more strongly, owing to the filling up of their concavities with darker green. These are the first rudiments of the eye-spots ([Figs. 19 and 30]). A very fine white line now connected the spiracles (infra-spiracular line), and could be traced from the last segment to the head. This line takes no further part in the subsequent development of the markings, but disappears in the following stage. The blood-red colour of the base of the black caudal horn is retained till the fifth stage, and then also disappears.

Before the second moult, which occurs after another period of 5–6 days, the caterpillars, which were about 1.3 centimeters in length, had assumed their characteristic tapering, slug-like form. I did not notice that the larvæ at this stage possessed the power of withdrawing the three foremost segments into the two succeeding ones, as is so frequently to be observed in the adults; neither were these two segments so strikingly enlarged as they are at an earlier period.

Third Stage.

After the second ecdysis the marking and colouring only undergo change with respect to the eye-spots. The concavities of the crescent-shaped portions of the subdorsal line become black,[70] the remainder of this line at the same time losing much of its whiteness, and thus becoming less distinct, whilst the crescents assume the appearance of small eye-spots (Fig. 20). During this stage the curved, crescent-formed portions become prepared for complete separation from the remainder of the subdorsal line; and just before the third moult the eye-spots become sharply defined both in front and behind, whilst the black ground-colour curves upwards, and the white spots gradually become lenticular and commence to enlarge ([Fig. 21]).

Fourth Stage.

The third moult takes place after another interval of 5–6 days, the eye-spots then becoming very prominent. The white nucleus of the front spot is kidney-shaped, and that of the hind spot egg-shaped; whilst the black ground-colour extends as a slender border upwards along the sides of the spots, but does not completely surround them till towards the end of the present stage ([Fig. 21]). The central portion of the white spots at the same time becomes of a peculiar violet-brown colour inclining to yellow above, the peripheral region alone remaining pure white.

Of the subdorsal line only traces are now to be recognized, and these are retained, with almost unchanged intensity, sometimes into the last stage, remaining with the greatest persistence on the three front and on the penultimate segments, whilst on those containing the eye-spots, i.e., the fourth and fifth, not a trace remains. At the present stage the peculiar mingling of colours becomes apparent over the whole of the upper surface; the green is no longer uniform, but a mixture of short and gently sinuous, dark green striations on a lighter ground now appear. On the sides of the caterpillar these stripes, which are at first indistinct, but become more strongly pronounced in the next stage, are arranged obliquely on the spiracles, with the lower portions directed forwards.

Fifth Stage.

The fourth moult occurs 7–8 days after the third, the caterpillar being 4–5 centimeters in length. Whilst all the specimens hitherto observed were with one exception light green, they now mostly changed their colour and became dark brown. In one case only did the brown colour appear in the previous (fourth) stage. The striations previously mentioned appear as dull and interrupted dirty yellow streaks, the same dirty yellow colour showing itself continuously on the sides of the four front segments. Of the subdorsal line only a distinct trace is now to be seen on the eleventh and on the three front segments, whilst on the third segment the formation of another eye-spot commences to be plainly perceptible by a local deposition of black ([Fig. 23]). This third spot does not, however, become completely developed, either in this or in the last stage, but the subdorsal line remains continuous on the three front segments. Among other changes at this stage, there occurs a considerable shortening of the caudal horn, which at the same time loses its beautiful black and red colours and becomes brownish.

The two large eye-spots have now nearly attained complete development. The kidney-shaped white spot has become entirely surrounded by black; and on the brown, red, and yellow tints present in this spot during the last stage, a nearly black spot has been developed—the pupil of the eye ([Fig. 33]). In order to establish a definite terminology for the different portions of the eye-spot, I shall designate the pupil as the “nucleus,” the light ground on which the pupil stands as the “mirror,” and the black ground which surrounds the mirror as the “ground-area.”

In this fifth stage the larva attains a length of six centimeters, after which the fifth moult takes place, the caterpillar becoming ready for pupation in the sixth stage. No striking changes of colouring or marking occur after the present stage, but only certain unimportant alterations, which are, however, of the greatest theoretical interest.

Sixth Stage.

In this stage the eye-like appearance of the spots on the front segments becomes still more distinct than in the fifth stage; at the same time these spots repeat themselves on all the other segments from the fifth to the eleventh, although certainly without pupils, and appearing only as diffused, deep black spots, of the morphological significance of which, however, there cannot be the least doubt. They are situated in precisely the same positions on the 5–11 segments as those on the third and fourth—near the front, and above and below the subdorsal line. A feeble indication of the latter can often be recognized ([Fig. 23]).

In all dark brown specimens the repeated spots can only be detected in a favourable light, and after acquiring an intimate knowledge of the caterpillar; but in light brown and green specimens they appear very sharply defined.

There is one other new character which I have never observed at an earlier period than the sixth stage, viz. the small dots which appear in pairs near the posterior edge of segments 5–11. These dots cannot have been developed from the subdorsal line, as they are situated higher than the latter. Their colour varies according to the ground-colour of the caterpillar, but it is always lighter, being light green in green specimens, dull yellow in those that are light brown, and grey in the blackish-brown caterpillars. These “dorsal spots,” as I shall term them, are chiefly of interest because they are present in Chærocampa Porcellus, in which species they appear one stage earlier than in C. Elpenor.

Chærocampa Porcellus, Linn.

Females captured on the wing, laid in the breeding-cage single eggs of a light green colour, spheroidal in form, and very similar to those of C. Elpenor.

First Stage.

The caterpillars on first hatching measure 3.5 millimeters in length, and are of a uniform light green colour, with a fine white transverse line on the posterior edge of each segment, precisely similar to that which appears in the second stage of C. Elpenor. They resemble the latter species still further in showing a fine white subdorsal line, which can easily be recognized by the naked eye ([Fig. 24]). Although the adult larva is distinguished from all the other known species of Chærocampa by the absence of a caudal horn, a distinct but very small one is nevertheless present at this first stage, and is indeed retained throughout the entire course of development, but does not increase further in size, and thus gradually becomes so small in proportion to the size of the caterpillar that it may be entirely overlooked.

The first moult takes place after 4–5 days.

Second Stage.

The blue-green coloration remains unchanged; but a somewhat darker green dorsal line becomes apparent down the middle of the back (the dorsal vessel?), and the subdorsal line now becomes very broad and pure white, being much more conspicuous than in any stage of C. Elpenor ([Fig. 25]). The tapering of the three front segments occurs at this stage, and oblique, dark green striations on a lighter ground stand out distinctly on the spiracles. As with C. Elpenor, the first traces of the future eye-spots appear during the second stage; not in the present case as a curvature of the subdorsal line, but as a spot-like widening of the latter, of a brighter white than the somewhat greenish colour of the remainder of the line.

Third Stage.

After the second moult, the formation of the dark “ground-area” of the eye-spots commences by the appearance of a little brown on the under edge of the foremost of the white spots, this coloration gradually increasing in extent and in depth. At the same time both spots become more sharply distinguishable from the subdorsal line, which becomes constantly greener ([Fig. 27]). The brown colour soon grows round the white of the front eye-spot, which becomes so far perfected; whilst the completion of the hind spot is effected slowly afterwards. The formation of the eye-spots does not therefore proceed any more rapidly in this species than in C. Elpenor.

At the end of the present stage the length of the caterpillar is about four centimeters; the ground colour is still sea-green; the subdorsal line is much diminished, completely fading away at its lower edge, but remaining sharply defined above, against the green ground-colour ([Fig. 26]).

Fourth Stage.

After the third moult all the caterpillars (5) became brown, this change occurring therefore one stage earlier than is generally the case with C. Elpenor. In single instances the brown colour appeared in the third stage. The subdorsal line had disappeared from all the segments but the three first and the last. The eye-spots now rapidly attained complete development; they contained a black pupil, and gave the insect a truly repulsive appearance when, on being threatened by danger, it drew in the front segments, and expanded the fourth ([Fig. 28]). The eye-spots of the fifth segment are much less developed than in C. Elpenor; they remain small, and are not readily detected. On the other hand, there now appear on all the segments with the exception of the last, just as in the sixth stage of C. Elpenor, distinct rudiments of eye-spots, which present the appearance of irregular, roundish, black spots on the front borders of the segments, at the height of the former subdorsal line. In this latter region the black pigment is disposed as a longitudinal streak, and to this a median line is added, the whole forming a marking which perhaps makes the caterpillar appear still more alarming to its foes. This marking is, however, only to be distinctly recognized on the three first segments. The “dorsal spots” mentioned in the case of C. Elpenor then appear very distinctly on segments 5–11.

The caterpillars continued to feed for eleven days after the third moult, at the end of which period the fourth moult took place, but without the occurrence of any change of marking. The larvæ then buried themselves, the complete development having taken 28–29 days.

The development of the Porcellus caterpillar was twice followed; in 1869 in twelve, and in 1874 in five specimens. In no case did I obtain caterpillars which remained green throughout the entire course of development, although this colour is stated in the books to occur occasionally in these larvæ; neither have I been able to find any figure of an adult green specimen, so that it must in the meantime be admitted that such specimens, if they occur at all, are exceptional instances.[71] The theoretical bearing of this admission will appear later on.

Results of the development of Chærocampa Elpenor and C. Porcellus; comparison of these with the other known species of Chærocampa.

The first stage of Elpenor shows that the most remote ancestor of the genus possessed no kind of marking, but was uniformly green. At a later period, the white longitudinal stripe which I have designated the “subdorsal line” made its appearance, and at a still later period this line vanished, with the exception of a few more or less distinct remnants, whilst, at the same time, from certain portions of it, the eye-spots of the fourth and fifth segments became developed. After the perfecting of the eye-spots, weak repetitions of the latter appeared as black spots on all the segments except the last.

In Porcellus the caterpillar emerges from the egg with the subdorsal line, the first stage of Elpenor being omitted. From this fact we may venture to conclude that Porcellus is the younger species, or, what comes to the same thing, that it has further advanced in development. The whole subsequent history of Porcellus agrees with this view, its course of development being essentially but a repetition of the phenomena displayed by Elpenor, and differing only in one point, viz. that all new characters make their appearance one stage earlier than in the latter species. This is the case with the transformation of the green into a brown ground-colour; with the repetition of the eye-spots on the remaining segments in the form of suffused black spots; and with the appearance of the light “dorsal spots.” Only the eye-spots themselves appear, and the snout-like tapering of the front segments occurs in the same stage as in Elpenor, i.e. the second.

From these data alone, we may venture to infer the occurrence of four chief stages in the phyletic development of the genus. The first stage was simply green, without any marking; the second showed a subdorsal line; the third, eye-spots on the third and fourth segments; and the fourth stage showed a repetition of the eye-spots, although but rudimentary, on all the remaining segments with the exception of the twelfth.

Now if we compare the other known species of Chærocampa larvæ with the above, we shall arrive at the interesting conclusion that all these species can be arranged in three groups, which correspond exactly with the three last phyletic stages as just deduced from the ontogeny of C. Elpenor and Porcellus.

Of the genus Chærocampa,[72] over fifty species have been described,[73] of which the larvæ of only fifteen are known in the form which they possess at the last ontogenetic stage.

Group 1.—I can furnish but little information with respect to this group. The first species with which I became acquainted was Chærocampa Syriaca,[74] of which I saw two blown caterpillars in Staudinger’s collection, and which I have figured in [Pl. IV]., Fig. 29. The larva is green, and has the short oblique stripes over the legs common to so many species of Chærocampa, the only marking besides these being a simple white subdorsal line, without any trace of eye-spots. This species exactly corresponds therefore with the second ontogenetic stage of C. Elpenor and Porcellus. The account of the species, both in the larval and perfect state, is unfortunately so imperfect, that we cannot with certainty infer the age of the two caterpillars from their size. If the moth were of the same size as Elpenor, then the caterpillar figured, having a length of 5.3 centimeters, would not be in the last but in the penultimate stage, and it remains doubtful whether it may not acquire eye-spots in the last stage.

That species exist, however, which in their last stage correspond to the second stage of Elpenor, is shown by two of the forms belonging to Walker’s genus Darapsa, which was founded on the characters of the imagines only. Ten species of this genus are given in Gray’s catalogue, the adult larva of two of these being known through the excellent figures of Abbot and Smith.[75] These two caterpillars possess the characteristic tapering form in a very marked degree; one is figured in the attitude so often assumed by our species of Chærocampa on the approach of danger, the three front segments being withdrawn into the fourth. (Fig. 34, [Pl. IV]., is copied from this Plate). There are no eye-spots either in D. Myron or D. Chœrilus,[76] but only a broad white subdorsal line; underneath which, and to a certain extent proceeding from it, there are oblique white stripes, precisely similar to those which meet the subdorsal line in the third stage of C. Porcellus.[77]

Group 2.—This group contains numerous species which, like our native C. Elpenor and Porcellus, show eye-spots on the fourth and fifth segments, whilst these markings are absent, or at most only present in traces, on the remainder. To this section there belong, besides the two species mentioned, five others, viz. in Europe, C. Celerio and Alecto (not certainly known?);[78] in India, C. Nessus, Drury, and Lucasii, Boisduval;[79] and an unnamed species from Port Natal.

In the species belonging to this group the subdorsal line may be more or less retained. Thus, C. Celerio, according to Hübner’s figure, has a broad yellow line extending from the horn to the sixth segment, whilst it is completely absent on the three front segments. In the unnamed species from Port Natal[80] the subdorsal line extends to the front edge of the fifth segment, and on the fourth segment only is there a perfect eye-spot, whilst on the succeeding segments traces of such markings can be recognized as dark spots similar to those in Elpenor and Porcellus. The transition to the third group is through another unnamed species from Mozambique,[81] in which rather large eye-spots have become developed on the fourth and fifth segments and these are followed by a subdorsal line, which only appears distinctly at certain places. On this broken subdorsal line, and not completely separated from it, there are small, roundish eye-spots, situated near the front edge of each segment; these being, therefore, a somewhat more perfect repetition of the front eye-spots.[82]

Group 3.—In the species of this group the eye-spots are repeated on all the segments. I am acquainted with seven such Chærocampa larvæ, of which C. Bisecta, Horsfield,[83] shows some affinity to the foregoing group, since the eye-spots on segments 6–11 have not yet attained full perfection. In C. Odenlandiæ, Fabr.,[84] and in C. Alecto from India,[85] the eye-spots appear to be perfectly alike on all the segments; whilst in C. Acteus, Cram.,[86] and in the North American C. Tersa[87] ([Pl. IV]., Fig. 35) they are smaller on the other segments than on the fourth; and in C. Celerio, Linn., from India,[88] the size of the spots diminishes from the head to the tail.

In this group also the subdorsal line is retained in a very variable degree. In some species it appears to have completely vanished (C. Acteus, Celerio); in others it is present as a light stripe extending along all the segments (C. Alecto); whilst in others it is retained as a broad white stripe, which extends only to the fourth segment (C. Tersa, [Fig. 35]). In species possessing eye-spots, the subdorsal line is thus a very variable character. It is, however, an interesting fact that even in the present group, which has made the greatest step forward, the subdorsal line is of general occurrence, because the eye-spots in all these species may have almost a similar development to those of Elpenor and Porcellus. The ontogeny of the tropical species would alone give a definite reply on this point, but unfortunately we are not acquainted with any of the young forms, so that we can but presume that some of them at least would show only in the first stage the simple subdorsal line without eye-spots; that in the second stage the primary pairs of eye-spots would be formed on the fourth and fifth segments, whilst the transference of these spots to the remaining segments would take place in the last stage.

The foregoing assumption is based immediately on the ontogeny of Elpenor and Porcellus; it is supported by the considerable size attained by the eye-spots in many species of the third group, and would receive additional confirmation by observations on the Indian C. Celerio, supposing that Horsfield’s statements do not arise from a confusion of species. This skilful observer, who was the first to breed systematically a large number of tropical larvæ, has given a figure of the Indian caterpillar of C. Celerio, according to which this species possesses eye-spots on all the segments from the fourth to the tenth. The European form of this same species has eye-spots only on segments four and five, a fact which does not appear to have been known to Horsfield, as no mention of it is made in his notice of the Indian species. If the caterpillar figured is really that of Celerio, which I consider to be by no means improbable, not only is it thus shown that in the species of the third group the ocelli on the hind segments have a secondary origin through a repetition of the primary ones of the front segments, but we can also establish that the same species in two different regions may arrive at two different phyletic stages.

If, finally, we sum up the facts taught by the ontogeny of the two German species, and the adult forms of the other species, we can form therefrom a tolerably complete picture of the course of development of the genus Chærocampa. Of the four phyletic stages indicated by the ontogeny of Elpenor and Porcellus, three still form the terminus of the development of existing species. The great differences among the caterpillars of this genus can be very simply explained on the view that they stand at different levels of phyletic development; some species having remained far behind (Group 1), others having advanced further (Group 2), and others having reached the highest point of development (Group 3). The fact that the species of the third group are only tropical accords well with this view, since many facts prove that phyletic development proceeds more rapidly in the tropics than in temperate climates.

The striking markings of the Chærocampa larvæ may, in brief, be stated to originate from a local transformation of two portions of the subdorsal line into eye-spots, and the subsequent transference of these two primary ocelli to the other segments. The eye-spots always originate on segments four and five, and from these the transference mostly occurs backwards, although in certain cases it takes place at the same time forwards. Herein, i.e. in the origin of the eye-spots, there lies a great distinction between the genus Chærocampa and the genus Deilephila, with which it was formerly associated, and in which the origin of a very similar kind of marking can be traced to quite another source.

The Genus Deilephila, Ochsenheimer.

I am acquainted with the caterpillars of nine European and one North American species, these differing in marking to such a wonderful extent that they appear to offer at first sight but little hope of being able to trace them to a common form. These ten species can be separated, according to their markings, into five groups, which I will briefly define before entering upon their ontogeny.

The first group consists of three species, and comprises the commonest and most widely-ranging of all the European species, Deilephila Euphorbiæ, as well as D. Dahlii from Sardinia and Corsica, and D. Nicæa, a species of very restricted range, which appears to occur only in one small district on the French coast of the Mediterranean. These three species agree in marking to the extent of their possessing in the adult form two rows of ring-spots on each side, whilst the subdorsal line is completely absent.

The second group, consisting also of three species, shows a great resemblance to Euphorbiæ, but has only one row of ring-spots. It contains D. Vespertilio, D. Galii, and the Algerian D. Mauritanica.

For the third group I only know one representative, D. Livornica, Esp., which possesses a single row of ring-spots connected by a subdorsal line.

Another group is composed of D. Zygophylli, which occurs on the shores of the Caspian Sea, and the North American D. Lineata; these species possessing a strongly marked subdorsal line, associated with more or less distinct ring-spots, which I shall designate as “open rings,” because their black border does not intersect the subdorsal line, but has the form of an arch above and below it.

In the last group, represented by D. Hippophaës, which occurs at the foot of the Alps (Wallis), and southward as far as Andalusia, there is only a broad subdorsal line, generally without any trace of a row of spots.

The important differences of marking displayed by these five groups are not in any way accidental, but they represent different stages of phyletic development; or, in other words, the five groups are of different ages, the first (Euphorbiæ, &c.) being the youngest, and the last (Hippophaës) the oldest of the genus.

According to their phyletic age, the groups follow each other in inverse order, the first being Hippophaës, the second that of Zygophylli, the third that of Livornica, the fourth that of Galii, and the fifth and youngest that of Euphorbiæ. Only in this last am I acquainted with the complete development of one species, for which reason I commence with this group, thus proceeding from the youngest to the oldest forms, instead of taking the more natural course from the simplest and oldest to the youngest and most complicated.

Deilephila Euphorbiæ, Linn.

Some captured females were at once placed in an enclosure about the size of a small sitting-room. It was evident that they did not feel quite at home under these conditions, frequently beating their heads and wings against the tarlatan, but some of them nevertheless laid eggs at the base of the leaves of Euphorbia Cyparissias. The eggs much resemble those of Chærocampa Elpenor, being spheroidal in form, but rather smaller, and of a somewhat darker green. They were laid in small clusters composed sometimes of as many as seven, the single eggs being placed near together, but never touching, and seldom at the point of the leaf, but generally near the end of a twig, where young shoots are in close proximity. During the embryonic development the eggs become coloured, first yellow and partly blackish, and finally completely black.

First Stage.

The young caterpillars (Fig. 37, [Pl. V].) immediately after hatching measure four millimeters in length; they are at first rather light, but in the course of half-an-hour they are seen by the naked eye to become of a deep velvety black; later, on increasing in size, they again become paler, appearing of a greenish-black, and subsequently blackish-green. On further increasing in size ([Fig. 38]), they are blackish-green, with the horn, head, legs, and a crescent-shaped chitinous plate on the back of the prothorax black. There are also on the last segment a double and two single black chitinous plates. Of the later marking of the caterpillar there is scarcely anything present. The spiracles appear as white spots, and on each segment there are a number (mostly ten) of small warts, each of which emits a single bristle.

When the young larvæ have attained a length of seven millimeters they are olive-green, and do not contrast so brilliantly with the green of the Euphorbia leaves as before; neither do they as yet possess any markings.

Second Stage.

The first ecdysis occurs after five days, and with this there appears quite suddenly a very complicated pattern. The ground-colour is now a light yellowish-green ([Fig. 39]), and on each of the twelve segments, near the front border, there is a pure white round spot in the middle of a large black transverse spot. I shall designate these, in accordance with the nomenclature employed for Chærocampa, as the white “mirrors” on black “ground-areas,” both together constituting “ring-spots,” as distinguished from “eye-spots” proper, in which a “nucleus,” the pupil of the eye, is also added. In many, but not in all specimens, very distinct traces of a subdorsal line can be seen as a light whitish stripe connecting the white spots. The horn, the thoracic and prolegs, and some spots on the head, are black.

The caterpillars remain unaltered till after four days, when, having a length of 17 millimeters, the second moult takes place, bringing with it changes quite as great as those which occurred with the first.

Third Stage.

The caterpillar now assumes the shagreened appearance which it possesses in the adult state. Small white warts are arranged in rows from the dorsal to the spiracular line, and again underneath this line on the abdominal legs. These dots are not only of value as a character for differentiating the genera Deilephila and Chærocampa, but they also play a part in the peculiar spot-marking which will be shown later on. The ground-colour of the caterpillar is now light green ([Fig. 40]), replaced by black on certain parts. From the black “ground-area” of the ring-spots, two black triangles extend towards the posterior borders of the segments, but usually without reaching them.

The ring-spots are not essentially changed, although it may be observed that in most specimens the shagreen-dots under each ring-spot are somewhat larger, and stand closer together than in other places. In the following stage they become fused into a second white “mirror,” so that two ring-spots stand one above the other, their black ground-areas meeting. The formation of the second ring-spot sometimes takes place in the present stage ([Fig. 42]).

The subdorsal line has now completely vanished, whilst the spiracular line[89] appears as a broad stripe above the legs. The horn is yellow with a black point, and the black spots on the head have increased in size.

Fourth Stage.

The third moult, which again occurs after four days, is not accompanied by such important changes. The green ground-colour has now completely disappeared, and is replaced by a dull black. The caterpillars are now, as also in the previous stage, extremely variable. Thus, for example, a triangular patch of the green ground-colour may be retained on the posterior edge of the segments ([Fig. 41]), those specimens which possess this character generally having their markings retarded in development, as shown by the absence of the second “mirror” of the ring-spots.

In [Fig. 41] the shagreen-dots from which this second “mirror” is subsequently formed, are distinctly larger than the others, and on the eleventh segment two of them have already coalesced.

Fifth Stage.

After another period of four days, the fourth moult takes place. The marking remains the same, but the colours become more vivid; the brick-red of the head, horn, dorsal line and legs, changing into a fiery red. The spiracular line, formerly green alternating with yellow, generally becomes resolved into a row of reddish-yellow spots. Ten days later the caterpillar (8.5 centimeters in length), ceases to feed, and prepares for pupation.

In this last stage also there is great variability of colour, but although each particular character is subject to fluctuation, the individuals of the same brood show but little variation among themselves.[90] Thus, the dorsal line is sometimes black, and sometimes red, or again, this colour interrupted with black, so that only small red spots mark its course. The head may be entirely red, or this colour mixed with black. On the under side of the caterpillar, red generally predominates, but in some specimens this is replaced by black. The ground-colour is also variable, being generally a shining brownish-black, but sometimes dull coaly black. The shagreen-dots are sometimes white and sometimes yellow, and the “mirrors” of the ring-spots are also often yellowish.

The most interesting variation, however, appears to me to be the following:—In many specimens from Kaiserstuhl (Breisgau), the red was unusually vivid, and was not limited to the ordinary places, but occupied also the triangles on the posterior edges of the segments ([Fig. 44]), which are green in the third and fourth stages ([Fig. 42]). This variety has also been figured by Hübner. In one individual ([Fig. 43]), the under ring-spots were wanting, whilst the upper ones possessed a beautiful red nucleus fading away anteriorly, and showing the first step in the formation of a complete eye-spot.

I cannot positively assert that a fifth moult occurs in the last ten days, although I am very doubtful whether this is the case. It is certain, however, that some time before pupation, and whilst the larva is still feeding, the striking colours fade out, and become replaced chiefly by black.

The ontogeny of this species is obviously but a very incomplete representation of its phyletic development. This is at once apparent from the large gap between the first and second stages. It is not possible that a row of ring-spots can have arisen suddenly; in all probability they have been developed from a subdorsal line, which in Euphorbiæ is now only indicated in the second stage by a faint line. This conjecture is raised to a certainty when we call in the aid of the remaining species of Deilephila.

Deilephila Nicæa, De Prunner.

I only know this species from blown larvæ in Staudinger’s collection, and Duponchel’s figure, of which Fig. 51, [Pl. VI]. is a copy. The adult insect possesses two perfectly separated rows of ring-spots. Duponchel figures also two younger stages, of which the youngest is probably the third stage. The larva is 18 millimeters in length, of a leaf-green colour, and shows no trace of a subdorsal line, but possesses the two rows of ring-spots, which only differ from those of the succeeding stages in the green colour of the “mirror.”

Deilephila Dahlii, Treitschke.

I am familiar with numerous specimens in various stages, collected in Sardinia by Dr. Staudinger, and preserved by inflation.

The first stage is blackish, and shows no kind of marking; thus agreeing with the corresponding stage of Euphorbiæ. The second stage is unfortunately not represented in Staudinger’s collection.

The third stage shows a row of ring-spots, which are, however, connected by a very distinct and sharply defined subdorsal line. In the fourth stage a second row of (under) ring-spots is added, whilst the subdorsal line generally at the same time disappears.

The caterpillar remains unchanged during the fifth stage, when it shows a great resemblance in marking to Euphorbiæ; neither does it appear to differ essentially from this species in colour, so far as can be judged from preserved specimens and single figures (in Duponchel and Hübner). I have, moreover, seen several larvæ in the last stage, and the subdorsal could be distinctly recognized as a broad light stripe.

Of the four groups, the second (that of Galii), appears to me to be of but very little importance, as I shall now proceed to show from the development of D. Vespertilio.

Deilephila Vespertilio, Fabricius.

Hitherto I have unfortunately been unable to obtain fertile eggs of this species, so that I can say nothing about the first stage. The latter would have been of interest, not only because of the marking, but also because of the presence of a residual caudal horn.

I am likewise only acquainted with the end of the second stage, having found, at the end of June 1873, a single caterpillar on Epilobium Rosmarinifolium, just previous to its second ecdysis. In the case of such young caterpillars, however, the new characters which appear in the succeeding stage are generally perceptible through the transparent chitinous skin at the end of the preceding stage, so that the markings of the insect are thus caused to change. The caterpillar found was about 16 millimeters long, and of a beautiful smooth and shining grass-green ([Fig. 13]). A broad white subdorsal line extended from the first to the penultimate segment, from which the horn was completely absent. On close inspection the first traces of the ring-spots could be detected near the anterior edge of each segment as feeble, round, yellow, ill-defined spots, situated on the subdorsal line itself ([Fig. 13]). On the first segment only there is no spot, and here no ring-spot is afterwards formed. Besides these markings, there was only to be seen a yellowish-white spiracular line.

This solitary specimen unfortunately buried itself before the moult for which it had prepared itself had occurred; but this ecdysis is associated with a very important transformation. This statement is founded on a blown specimen in Staudinger’s collection; it is only 18 millimeters in length, but already shows the later grey colouring in place of the beautiful green. In this, the third stage, the broad white subdorsal line bears on each segment a red spot enclosed between black crescents above and below ([Fig. 49 A]). In the fourth stage, during which I have seen many living caterpillars, the subdorsal line is still distinctly present in some individuals ([Fig. 14]), but the spots (“mirrors”) are now completely surrounded by a narrow black ring (“ground-area”), which sharply separates them from the subdorsal line ([Fig. 49 B]). In the fifth stage this ring becomes a somewhat irregularly formed black “ground-area,” whilst the subdorsal line completely vanishes ([Figs. 51 and 49 C]). The mirrors are white, but generally have a reddish nucleus, which obviously corresponds to the primary yellow spots from which the whole development of the ring-spots originates. This character is, however, sometimes absent; and many other variations also occur in the earlier stages, all of which can be easily explained as cases of arrested, or retarded development. Thus, the subdorsal line often disappears earlier, and is only present in the fourth stage as a feeble light stripe.

Deilephila Galii, Fabricius.

The markings of this species appear to be developed in a precisely similar manner to those of D. Vespertilio. The adult larva, as in the last species, shows no trace of a subdorsal line. A row of large black spots, each having an irregular round, yellowish-white nucleus, is situated on an olive-green, blackish-brown, brown, or dirty yellow ground. I have, unfortunately, also in this case been unable to procure fertile eggs. There is, however, one figure of a caterpillar, 2.5 centimeters long, by Hübner, which is of a light green colour, and has five longitudinal lines; one dorsal, two subdorsal, and a spiracular line. The subdorsal is white, and bears in the place of the ring-spots small red dots, whilst the line itself is bordered with black where the red spots are situated. Hübner has probably figured the third stage, so that we may venture to conclude that in the second stage there is a subdorsal line either quite free from spots, or only showing such feeble rudiments as are to be seen in the second stage of Vespertilio.

I found two specimens in the fourth stage in the Upper Engadine. One of these ([Fig. 45]) was already of a dark, blackish-green ground-colour[91] with a broad, greenish-white subdorsal line sharply defined throughout its entire length, and containing ring-spots of a sulphur-yellow with an orange-red nucleus; the black “ground-area” did not encroach upon the subdorsal line, but was confined to two faint crescents situated above and below the “mirror.” Only the two foremost “mirrors” (on the second and third segments) were without nuclei.

The remaining peculiarities of coloration are shown in the figure. I may here only point out the shagreening present on the sides and a portion of the under surface.

The specimen figured was 3.3 centimeters long; a second example measured 2.8 centimeters in length, and was essentially similar, but showed that a considerable amount of variability must prevail at this stage of development. It was pitchy black, with a very indistinct subdorsal line and a few ring-spots, the “mirrors” of which were also sulphur-yellow, with the orange-red nucleus. The shagreening was quite as strong as in the first specimen, the dots being yellow instead of white. It is specially to be observed, because of its important theoretical bearing, that in this larva the ring-spots were absent on the three front segments, and on the fourth only, a faint indication of one could be perceived. In the caterpillar figured the ring-spots increase also in distinctness from the tail to the head.

Fifth Stage.

The two specimens just mentioned, after moulting, acquired the well-known markings of the adult caterpillar already briefly described above. The fifth is the last stage.

The larva is known to occur in several variations, Rösel having figured it in three forms; light green, olive-green, and dirty yellow. It has not been since considered worth the trouble to attend to the subject of caterpillar coloration. Thus, Wilde,[92] in his well-known work, takes no notice of Rösel’s observation, but simply describes the caterpillar of Galii as “blackish olive-green.”

Having had an opportunity of observing twenty-five adult specimens of this somewhat scarce species at one time, I am able to state that it is not in this instance di- or polymorphism, but a case presenting a great degree of variability, with which we have to deal. There are not several sharply-defined types of coloration; but the extremes are connected by numerous intermediate forms. The extreme forms, however, certainly preponderate.

I have never met with Rösel’s light green form; neither was there a dark green specimen among the twenty-five mentioned, and I only know this variety from single individuals, found at a former period. Among the twenty-five caterpillars; all gradations of colour occurred, from pitchy black to light clay-yellow, and even to an almost whitish-yellow; some were brownish-black, others of a beautiful chestnut-brown, and others yellowish brown, dark clay-yellow, or brownish-red. Out of twenty-one specimens of which the ground-colours were noted, there were nine black, nine clay-yellow, and three brown; each of the three groups again showing various minor modifications of colour. The other colours also varied somewhat. Thus, the “mirrors” were sometimes white, sometimes strong yellow, and occasionally they also contained a reddish nucleus.

The variations in the shagreening were especially interesting, inasmuch as these appeared to have a striking connection with the general colouring of the caterpillar. Black specimens seldom show such sparse shagreening as that represented in [Pl. V]., Fig. 46, but are generally thickly scattered with large shagreen-dots right up to the dorsal line (Fig. 47, [Pl. VI].), then strikingly resembling the adult larva of D. Euphorbiæ. The light ochreous-yellow individuals, on the other hand, were sometimes entirely without shagreening (Fig. 48, [Pl. VI].), being smooth, and much resembling the light ochreous-yellow or yellowish-red caterpillar of D. Nicæa (Fig. 51, [Pl. VI].). I have never seen a caterpillar of Galii which showed traces of the subdorsal line in the last stage, nor have I ever met with one which possessed a second row of “mirror” spots; so that retrogression or a sudden advance in development does not appear to occur.

Of the North African D. Mauritanica, which likewise belongs to the Galii group, I have not been able to obtain specimens or figures of the younger stages. The adult caterpillar is very similar to that of Euphorbiæ, but differs in the absence of the second row of ring-spots. For this reason it must be regarded as a retarded form at an older stage of phyletic development.

I now proceed to the Livornica group.

Deilephila Livornica, Esper.

This, the only European species here to be considered, possesses almost the same markings as Galii in its fourth stage, i.e., a subdorsal line with interpolated ring-spots. The species is known to be rare, and I have not been able to obtain living specimens, but I have examined several blown larvæ, all of which agree in having the ring-spots sharply distinct from the whitish subdorsal line, so that the latter is thereby interrupted. Figures of the adult larva are given in the works of Hübner, Boisduval, and Duponchel. In most specimens the ground-colour is brown, although Boisduval[93] also figures a light green specimen; from which it may be inferred, from analogy with Galii and Vespertilio, that the first stages are green. In Dr. Staudinger’s collection there is a young larva, probably in the fourth stage, the ground-colour of which is light ash-grey. The dorsal and subdorsal lines are white, the latter showing in the positions where the ring-spots subsequently appear, small white “mirrors” with red nuclei, exactly corresponding to the stage of Vespertilio represented in Fig. 49 A, [Pl. VI]. The “mirrors” are nothing more than dilatations of the subdorsal line, which is not therefore interrupted by them. The black “ground-area” does not surround the “mirrors” completely, but borders them only above and below, and is much more strongly developed above, extending in this direction to the dorsal line.

The fourth group comprises the two species D. Lineata, Fabr., and D. Zygophylli, Ochs., the former being the North American representative of our D. Livornica, but differing in remaining permanently at the fourth stage of this last species. I am acquainted with D. Lineata only through the figure of the adult larva given by Abbot and Smith, which figure, judging from the position and form of the spots, I am compelled to believe is not quite correct, notwithstanding the excellence of the other illustrations. The ground-colour of the caterpillar is green; the subdorsal yellow, bordered with black, slightly curved, arched lines, which nowhere interrupt its continuity. This North American species appears therefore to be an older form than our Livornica.

Deilephila Zygophylli, Ochsenheimer.

This species, which is the next allied form to D. Lineata, is an inhabitant of Southern Russia. I have seen four specimens of the caterpillar in Dr. Staudinger’s collection, three of which are certainly in the last ontogenetic stage. The ground-colour appears ash-grey, ash-brown, or blackish with whitish granulations. A broad white subdorsal line extends to the base of the black caudal horn, this line in one specimen appearing at first sight not to possess a trace of spot rudiments ([Fig. 50]). On closer investigation, however, there could be observed, in the same position where the ring-spots stand in the other species of Deilephila, small black crescents above and below the subdorsal line. In other specimens the white subdorsal line had also become expanded in these positions into distinct spots; indeed, in one individual light white mirror-spots, bordered above and below by black crescents, stood on the subdorsal line ([Fig. 50 A]).

It is thus in this distinguishing character that the caterpillar is extremely variable, and we may suppose either that this species is now in a state of transition to a higher stage of phyletic development, or else that the ring-spots were formerly more strongly developed, and are now degenerating. The developmental history of the larva could alone decide which of these two views is correct. There would be no difficulty in procuring materials for this purpose if one of the numerous and zealous Russian naturalists would take up the subject.

Deilephila Hippophaës, Esper.

This is the only representative of the fifth and oldest group known to me. The moth resembles D. Euphorbiæ to the extent of being sometimes confounded with it, a circumstance which is made the more remarkable by the fact that the caterpillars are so completely different.

The adult larva of this local moth has been made known by the figures, more or less exact, in the works of Hübner, Boisduval, and Duponchel. Wilde also gives a description of it, although from a foreign source. I will not here delay myself by criticizing the different descriptions and figures; they are partly correct, partly inexact, and sometimes altogether erroneous; they were of no avail for the question which here primarily concerns us, and new observation had to be undertaken.

I have been able to compare altogether about forty caterpillars, thirty-five of which were living. All these specimens possessed nearly the same greyish-green ground-colour, and most of them had exactly the simple marking as represented, for instance, in Hübner’s figure, i.e., a rather broad greenish-white subdorsal line, somewhat faded at the edges, and without a trace of spots on any of the segments with the exception of the eleventh, on which there was a yellowish, black-bordered mirror-spot, with a broad, diffused, vivid orange-red nucleus. Specimens also occur, and by no means uncommonly, in which no other markings are to be seen than those mentioned; there were nine among twenty-eight examples compared from this point of view.

In many other individuals of this species small red spots appear on the subdorsal line, exactly in the positions where the ring-spots are situated in the other species of the genus ([Fig. 60]), so that these spots are thus repetitions of the single ring-spot—a fact which must appear of the greatest interest in connection with the development of the markings throughout the whole genus. But this is not all, for again in other specimens, these red spots stand on a large yellow “mirror,” and in one individual ([Fig. 59]), they had become developed into well-formed ring-spots through the addition of a black border. We have thus presented to us in one and the same stage of a species, the complete development of ring-spots from a subdorsal line.

These facts acquire a still greater interest, as showing how new elements of marking are produced. The spots on the subdorsal line decrease from the posterior to the anterior segments, so that they must undoubtedly be regarded as a repetition or transference of the ring-spot previously developed on the eleventh segment. I will now proceed to furnish proofs in support of this statement.

I have never met with any specimens having ring-spots on all the segments—in the most prominent instances these spots were present on segments 10–5. This was the case in three out of the twenty-eight caterpillars minutely examined. On all these segments, however, the ring-spots were not equally developed, but increased in perfection from the posterior towards the anterior segments. In the larva represented in [Fig. 59] for example, there is a completely developed ring-spot on segment 10, which, although possessing but a feeble black “ground-area,” is still distinctly bordered; on segment 9 this border is less sharp, and not so dark, and it is still less sharp and much lighter on segments 8 and 7, whilst it has completely disappeared from segment 6, the yellow “mirror” having at the same time lost in size. On segment 5, only two small contiguous reddish spots, the first rudiments of the nucleus,[94] can be recognized on close inspection.

Specimens in which the spots extend from the eleventh to the seventh segment are of more frequent occurrence, five having been found among the twenty-eight. In these the spots diminish anteriorly in size, perfection, and intensity of colour. Still more frequently (in eleven specimens) are the ring-spots or their rudiments restricted to the tenth and ninth segments, the spot on the latter being without exception less developed than that on the former segment.

An anteriorly progressing formation of ring-spots thus undoubtedly occurs, the spots generally diminishing in perfection very suddenly towards the front segments; and specimens, such as that represented in Fig. 60, [Pl. VII]., in which traces of ring-spots are to be seen on all the segments from the tenth to the fifth, are of rare occurrence.

From what elements of marking are these secondary ring-spots resulting from transference developed? They do not, as in the case of the primary eye-spots of the Chærocampinæ, originate in the separation of one portion of the subdorsal line, and the subsequent formation of this detached spot into a “mirror;” but they arise from the formation of a nucleus, first one and then two of the shagreen-dots on the subdorsal line acquiring a yellowish or reddish colour (Fig. 61, [Pl. VII]., segments 6 and 7). The ground on which these two spots are situated then becomes yellow (Fig. 61, [Pl. VII]., segment 8), and a more or less distinct black border, having the form of two small crescents, is afterwards formed. At a later period these two crescents and also the two primary nuclei coalesce, producing a ring-spot which, as in Fig. 61, [Pl. VII]., segment 9, can be distinctly resolved into two portions.

It certainly cannot be denied that these facts may also be theoretically interpreted in a reverse sense. We might interpret the phenomena in this case, as also in that of D. Zygophylli, as a gradual disappearance from the front towards the hind segments of ring-spots formerly present, a view which could only be refuted by the ontogeny of the species. I have not been fortunate enough to procure eggs of D. Hippophaës, so that the younger stages are unknown to me. Among my caterpillars, however, there were two in the fourth stage of development, but these did not show ring-spots on all the segments, as we should expect on the above view; on the contrary, no trace of such spots could be seen on any of the segments with the exception of the eleventh, on which there was a ring-spot less perfectly developed than in the last stage.

In this fourth stage the larva of D. Hippophaës is of a lighter green ([Fig. 58]), the subdorsal yellowish with sharp boundaries, and the infra-spiracular line pure white, as in the next stage. The shagreening is present, but none of the shagreen-dots are red or reddish, and no trace of a ring-spot can be detected on the subdorsal line with the exception of that on the eleventh segment. In this last position this line is somewhat widened, and a long, diffused, rose-red spot can there be recognized upon it ([Fig. 58 A]). The black “ground-area” present in the fifth stage is as yet absent, and the spot is not so sharply separated anteriorly from the subdorsal line as it becomes later.

From these observations we might venture to expect that in the third stage of Hippophaës, the subdorsal line would also be free from this spot on the eleventh segment, and it is possible that in the second stage this line is itself absent.

The Genus Deilephila: Summary of Facts and Conclusions.

Regarding only the adult larvæ of the species of Deilephila, these represent in their five groups, five stages in the phyletic development of the genus; but if we also take into consideration the developmental history, two more stages must be added, viz., that in which the caterpillar possesses no particular marking, as was found to be the case in the first stage of the development of D. Euphorbiæ and D. Dahlii; and a second stage with a subdorsal line, but without any ring-spot formations. Seven stages of phyletic development must therefore be distinguished.

Stage 1.—No species with entire absence of marking in the adult form now occurs.

Stage 2.—A subdorsal, accompanied by a spiracular line, extends from the caudal horn to the first segment. This also no longer forms the final stage of the ontogeny, but is, however, undoubtedly retained in the second stage of several species (D. Vespertilio, Livornica, Lineata, and perhaps also Galii).

Stage 3.—The subdorsal line bears a ring-spot on the penultimate segment; the other markings as in the last stage. D. Hippophaës only belongs to this stage, a small number of specimens, however, showing a transition to the following stage by the transference of ring-spots from the posterior to the anterior segments.

Stage 4.—Open ring-spots appear on the subdorsal line on all the segments from the eleventh to the first. D. Zygophylli and the North American D. Lineata belong here.

Stage 5.—Closed ring-spots are situated on the subdorsal line. Of the known species, only D. Livornica concludes its development at this phyletic stage.

Stage 6.—A single row of ring-spots replaces the subdorsal line. D. Galii, Vespertilio, and Mauritanica represent this stage at the conclusion of their ontogeny.[95]

Stage 7.—A double row of ring-spots. Only D. Dahlii, Euphorbiæ, and Nicæa attain to this highest stage of Deilephila marking, the two first species in the fourth stage, and Nicæa in the third stage of its ontogeny.

Although our knowledge of the history of the development of the individual species is still so fragmentary, we may conclude with certainty that the development of the markings has been uniform throughout—that it has proceeded in the same manner in all species. All the species appear to be making for the same goal, and the question thus arises whether there may not be an innate force urging their phyletic development. The rigorous examination of this conception must be reserved for a later section. Here, as we are only occupied essentially in establishing facts, it must be remarked that retrogression has never been observed. The young larval forms of a species never show the markings of a later phyletic stage than the older larval forms; the development takes the same course in all species, only making a greater advance in the same direction in some than in others.

Thus, Nicæa and Euphorbiæ have advanced to the seventh phyletic stage, Zygophylli and Hippophaës only to the third, and some specimens of Zygophylli to the fourth. But at whatever phyletic stage the ontogeny of a species may terminate, the young larval stages always display the older phyletic stages. Thus, Galii in its last ontogenetic stage reaches the sixth phyletic stage; in its penultimate stage it reaches the fifth phyletic stage; and in its third stage; the fourth phyletic stage is represented, so that little imagination is required to anticipate that in the second stage the third or second phyletic stage would be pictured.

If we tabulate the development of the various species, indicating the ontogenetic stages by Arabic numerals, and the stages of the phylogeny which are reached in each stage of the ontogeny by Roman numerals, we obtain a useful synopsis of the series of developments, and, at the same time, it shows how many gaps still remain to be filled up in order to complete our knowledge even of this small group of species.

Table of Development of the Species of Deilephila.

Deilephila.	Ontogeny Stage 1.	Ontogeny Stage 2.	Ontogeny Stage 3.	Ontogeny Stage 4.	Ontogeny Stage 5.
1. Hippophaës	?	?	?	III.	III.-IV.
2. Zygophylli	?	?	?	?	III.-IV.
3. Lineata	?	?	?	?	IV.
4. Livornica	?	?	?	IV.	V.
5. Galii	?	?	IV.	V.	VI.
6. Vespertilio	?	II. (?)	IV.	V.	VI.
7. Mauritanica	?	?	?	?	VI.
8. Dahlii	I.	?	VI.	VII.	VII.
9. Euphorbiæ	I.	V.	VI.	VII.	VII.
10. Nicæa	?	?	VII.	VII.	VII.

From this very incomplete table we perceive that, in certain instances, the stages can be represented as a continuous series of phyletic steps, as in the case of D. Galii; that in others certain steps may be omitted, as with D. Euphorbiæ, in which grade I. of stage 1 is immediately followed by grade V. in stage 2. In reality the gap caused by this omission is still greater than would appear, as grade V. is only indicated, and not actually reached, the subdorsal not being present as a sharply-defined line, but only as a faint stripe. The suppression of phyletic steps increases with the advancement in phyletic development. The higher the step to which a species finally attains, the greater is the tendency of the initial stages to be compressed, or omitted altogether.

From what has thus far been seen with respect to the development of D. Hippophaës, there may be drawn what to me appears to be a very important conclusion, viz. that the ring-spots of Deilephila first originated on the segment bearing the caudal horn, and were then gradually transferred as secondary spots to the preceding segments. Complete certainty would be given to this conclusion by a knowledge of the young forms of other phyletically retarded species, especially those of the American D. Lineata, and perhaps also those of Zygophylli and Livornica. The few observations on the development of D. Galii already recorded give support to this view, since the absence of ring-spots on the three front segments in the young caterpillar (one instance), or their less perfect formation on these segments (second instance), indicates a forward transference of the spots.

If the foregoing view be accepted, there follows from it a fundamental difference between the development of the genera Chærocampa and Deilephila. In the former the formation of the eye-spots proceeds from a subdorsal line, but they first appear on two of the front segments, and are then transferred to the posterior segments. In Deilephila, on the other hand, a single ring-spot is formed on the penultimate segment bearing the caudal horn, and this is repeated on the anterior segments by secondary transference. With respect to the origination of the ring-spot also, there is a distinction between this genus and Chærocampa, inasmuch as the first step towards the eye-formation in the latter consists in the separation of a curved portion of the subdorsal line, whilst in Deilephila the nuclear spot first seems to originate and the separation of the mirror-spot from the subdorsal line appears to occur secondarily. It is difficult here to draw further conclusions, since the first appearance of the primary ring-spot has not yet been observed, and no more certain inference respecting the history of the formation of the primary ring-spots can be drawn from the manner in which the secondary ring-spots are formed. Because in Hippophaës the formation of the secondary ring-spots begins with the red coloration of one or two shagreen-dots, it does not follow that the primary spot on the eleventh segment also originated in this manner; and this is not without importance when we are concerned with the causes which underlie the formation of ring-spots. In Chærocampa also, the formation of the primary eye-spots appears to differ from that of the secondary—in the latter the black “ground-area” first appearing, and in the former the “mirror-spot.” The secondary eye-spots certainly remain rudimentary in this last genus, so that the evidence in support of this conclusion is thus much weakened; but it must be admitted that we are here on ground still too uncertain to admit of wider conclusions being based thereon.

As a final result of the investigation, we may advance the opinion that the existing species of the genus Deilephila have reached five different phyletic stages, and that their very different external appearance is explained by their different phyletic ages; the appearance from these caterpillars of moths so extremely similar, can otherwise be scarcely understood.

It may appear almost unnecessary to bring forward additional proofs in support of this interpretation of the facts, but in a field where the data are so scanty, no argument which can be drawn from them should be considered as superfluous. The variations which occasionally occur in the larvæ, however, to a certain extent furnish a proof of the correctness of the theoretical interpretation offered.

When, in the ontogeny of these species, we actually see before us a series of stages of phyletic development, we must admit that ordinary reversion may occur, causing an adult caterpillar to show the characters of the young. Forms reverting to an earlier phyletic stage must, on the whole, occur but seldom, as this stage is removed further back in the ontogeny. Thus, indications of the subdorsal line must occur but rarely in the adult larvæ of Euphorbiæ, and still less frequently in Nicæa, whilst they must be expected to be of more common occurrence in Vespertilio, and also, as has already been seen, in Dahlii. In this last species, as also in Vespertilio, the completely-developed subdorsal line is still present in the third stage, whilst it is possessed by Euphorbiæ only in the second stage, and then in a rudimentary condition.

The state of affairs may in fact be thus described: Among several hundred adult larvæ of Dahlii found in Sardinia by Dr. Staudinger, there were some which did not actually possess a distinct subdorsal line, but in place thereof, and as its last indication, a feeble light stripe. One of Dr. Staudinger’s caterpillars showed also a distinct line between the closed eye-spots. In the last stage of Vespertilio this line appears still more frequently, whilst in Euphorbiæ it is extremely rare, and when present it only appears as a faint indication. This is the case with one of the specimens figured in Hübner’s work as an “aberration,” and also with one in Dr. Staudinger’s collection. Of Nicæa I have at most seen only eight specimens, none of which showed any trace of the long-vanished subdorsal line.

It must be expected that any ontogenetic stage would most readily revert to the preceding phyletic stage, so that characters present in the preceding stage are consequently those which would most commonly arise by reversion. This postulate of the theory also finds confirmation in the facts. Caterpillars which, when full grown, belong to the seventh phyletic stage, e.g. D. Euphorbiæ, not unfrequently show variations corresponding to the sixth stage, i.e. only one instead of two rows of ring-spots—the upper and first-appearing series. On the other hand, forms reverting to the fifth phyletic stage (ring-spots with connecting subdorsal line) occur but very rarely. I have never met with such cases in adult living caterpillars of D. Euphorbiæ, although in one instance such a larva was found in the fourth ontogenetic stage; but the strikingly dark, brownish subdorsal line which connected the otherwise perfectly developed ring-spots, completely disappeared in the fifth stage of the ontogeny. Those larvæ which, in the adult state, belong to the sixth phyletic stage, not unfrequently show the characters of the fifth stage more or less developed, as, for example, D. Vespertilio.[96]

THE GENUS SMERINTHUS, LATREILLE.

The caterpillars of this genus are very similar in appearance, and all possess extremely simple markings. The occurrence of numerous stages of development of these markings is thus excluded, and the study of the ontogeny therefore promised to furnish less information concerning the phyletic development of the genus than in the case of the preceding genera. This investigation has nevertheless also yielded interesting results, and the facts here recorded will be found of value in likewise throwing light on the causes which have produced the markings of caterpillars.

I shall commence, as in former cases, with the developmental history. I have easily been able to obtain fertile eggs of all the species of Smerinthus known to me. Impregnated females laid large numbers of eggs in confinement, and also bred females of the commoner species can readily be made to copulate, when pinned, and exposed in a suitable place in the open air. A male soon appears under these circumstances, and copulation is effected as readily as though the insect were not fastened in the way indicated.

Smerinthus Tiliæ, Linn.[97]

The light green eggs are nearly spherical, and after fourteen days (beginning of July) the young larvæ emerge. These are also of a light green colour, and are conspicuous for the great length of the caudal horn, which is nearly half as long as the body. This horn is likewise of a light green at first, but becomes dark violet in the course of an hour. No trace of any markings can be detected at this stage.

As soon as the caterpillars are hatched they commence to nibble the empty egg shells; then they run about with great activity, and after several hours take up their position on the largest vein on the under side of the lime leaves, where they remain for a long period. In this situation they have the same form and colour as the leaf-vein, and are very difficult to discover, which would not be the case if they reposed obliquely or transversely to the vein. In about 4–5 days the caterpillars undergo their first moult, and enter upon the second stage. On each side of the segments 11–4, there now appear seven oblique whitish stripes on a somewhat darker green ground; these slope in the direction of the caudal horn. Owing to the transparency of the skin, a dark green dorsal line appears in the position of the underlying dorsal vessel, the green contents of the alimentary canal being distinctly visible through the absence of adipose matter in the tissues. The larvæ possess also a fine whitish subdorsal line, which extends from the horn to the head. The horn at this stage becomes black with a yellowish red base.

In the third stage, which occurs after six or seven days, the oblique stripes appear darker, and the subdorsal line disappears.

Fourth Stage.

After another period of 4–5 days the third moult takes place, and there now commences a dimorphism which will perhaps be better designated as variability, since the two extremes are connected by transitional forms. The majority of the larvæ have, as in the preceding stage, pure white oblique stripes, but many of them possess a blood-red spot on the anterior side of the stripes, this spot showing all gradations in size and depth of colour between maximum development and a mere trace. Special interest attaches to these spots, as they are the first rudiments of the coloured border of the oblique stripes which occurs in so many Sphinx caterpillars.

In the fifth stage—the last of the larval development—the red spots become more strongly pronounced. Among eighty caterpillars from one brood there were about twenty without any red whilst the remainder were ornamented with more or less vivid blood-red spots, often large and irregular in form. In some specimens the spots had become drawn out into lines,[98] forming a coloured edge to the oblique white stripes, similar to that possessed by the larva of Sphinx Ligustri. The caterpillar is thus represented in many figures, but generally the coloured stripe is made too regular, as in reality it is always irregularly defined above, and never so sharp and even as in Sphinx Ligustri. The character is here obviously not yet perfected, but is still in a state of development.

Smerinthus Populi, Linn.

From green spherical eggs there emerged larvæ 6.5 millimeters in length without any markings. They were of a light greenish-white, the large head and long caudal horn being of the same colour. The posterior boundary of the segments appears as a light shining ring ([Pl. VI]. Fig. 55).

The characteristic markings of the genus appear on the following day without the occurrence of any moult: seven oblique white stripes arise from near the dorsal line, and extend along the sides in a direction parallel to that of the horn. On the three front segments they are represented only by three small white spots ([Fig. 56]). The caterpillar likewise possesses a marking of which the adult species of the genus retain only a trace, viz., a well-developed, pure white subdorsal line, which is crossed by the six anterior oblique stripes, and uniting with the upper part of the seventh extends to the caudal horn.

I long believed that the markings described were first acquired in the second stage, as I was possessed with the generally accepted idea that the changes of form and colour in insects could only occur at the period of ecdysis. I at first thought that the moult had escaped my notice, and I was only undeceived by close observation of individual specimens.

Second Stage.

The first moult took place after five days, the larvæ being 1.4 centimeters in length. Only unimportant changes of marking are connected therewith. The subdorsal line loses much in thickness and definition, and the first and last of the oblique stripes become considerably broader than the intermediate ones ([Fig. 57]). The green ground colour and also the stripes acquire a yellowish hue.

On the other hand, there occur changes in form. The head, which was at first rounded, becomes of the characteristic triangular shape, with the apex upwards, common to all the species of the genus, and at the same time acquires two white lines, which unite above at the apex of the angle. The shagreening of the skin now also takes place, and the red spot at the base of the horn is formed.

There appears to be at this stage a general tendency for the suffusion of red, the thoracic legs also becoming of this colour.

Third Stage.

The second moult occurs after six or eight days, the marking only changing to the extent of the subdorsal line becoming still more indistinct. This line can now only be distinctly recognized on the three front segments in a few individuals, whilst in the majority it is completely absent. Sometimes the ferruginous red spots on the oblique stripes now appear, but this character is not completely developed till the fifth stage. Out of about ninety bred specimens in which I followed the entire development, only one possessed such spots, and these were situated on both sides of the sixth segment.

Fourth Stage.

The third moult, which takes place after another period of six days, is not associated with any change of marking.

In this stage also I observed in one specimen (not the one just mentioned) the ferruginous spots, and again only on the sixth segment. On account of the theoretical conclusions which may be drawn from this localization of the spots—supposing it to be of general occurrence—it becomes of importance to institute observations with different broods, so as to investigate their first appearance, frequency, and local limitation. It appears to me very probable that, with respect to frequency and time of appearance, there would be great differences, since, in the last stage, it is just this character which shows a great variability. It would be more remarkable if it should be established that the first appearance of the spots was always limited to a certain segment; and there would then be a great analogy with the first appearance of the eye-spots in Chærocampa and the ring-spots in Deilephila.

Fifth Stage.

The adult caterpillar does not differ in marking to any considerable extent from the preceding stages. The first and last stripes do not appear larger than the intermediate ones, as the latter now increase in size. Many specimens were entirely without red spots; in others they were present, but were small and inconspicuous, whilst in others again there were two spots, one above the other, of a vivid ferruginous red, these coalescing in some cases, and thus forming one spot of a considerable size. I have never seen these spots formed into a regular, linear, coloured border to the white oblique stripes—as occasionally happens in Tiliæ—either in living specimens, blown larvæ, or in figures.

Smerinthus Ocellatus, Linn.

The green eggs much resemble those of Populi, as also do the newly hatched caterpillars, which, as in the case of this last species, are entirely without markings. As with Populi, the markings are formed in the course of the first stage, and are distinctly visible before the first moult. The long caudal horn is of a red colour.

After two to three days the caterpillars moult, their length then being one centimeter; the seven beautiful oblique white stripes, and the fine white subdorsal line, are more strongly pronounced, the latter becoming broader in front. They differ from Populi in having the oblique stripes united in the dorsal line.

The second moult occurs after another three days, and brings no important change; only the fine subdorsal line becoming somewhat fainter. Neither is the third moult, which takes place four days later, associated with the appearance of any essentially new character. The oblique stripes remain as before, but their upper portions now stand on a somewhat darker green ground-colour, whilst the subdorsal line vanishes, leaving distinct traces only on the three or four front segments.

The fourth moult follows after a period of seven days, and my bred larvæ underwent scarcely any alteration in marking. Only small differences in coloration became perceptible in the head and horn, these changing to bluish. Specimens occur, although but rarely, which show in this last stage red spots in the vicinity of the oblique stripes, just in the same manner as with Populi, in which species, however, they occur more commonly. I only once found an adult larva of Ocellatus possessing reddish-brown spots above and below the oblique stripes,[99] exactly as in one of the specimens figured by Rösel.[100]

In this stage also there remains almost always on the three to six front segments, a more or less distinct residue of the subdorsal, which extends backwards from the head as a whitish line intersecting the foremost oblique stripes. (Fig. 70, [Pl. VII].)

Results of the Developmental History of Smerinthus Tiliæ, Populi and Ocellatus.

From the meagre materials furnished by these three obviously nearly related species, we may at least conclude that, with respect to marking, three stages of development can be distinguished:—(1) Simple (green) coloration without marking; (2) subdorsal lines crossed by seven pairs of oblique stripes; (3) more or less complete absence of the subdorsal lines, the oblique stripes remaining, and showing a tendency to become edged with a red border.

Which of the three species is the oldest I will not attempt to decide. If we might venture to form any conclusion from the frequency of the red spots, Tiliæ would be the youngest, i.e., the species which has made the farthest advance. But this does not agree with the fact that the oblique stripes appear somewhat later in this species. Both these distinctions are, however, too unimportant to enable us to build certain conclusions on them. Neither does a comparison of the adult larvæ with other species of Smerinthus furnish any further information of importance.

Of the genus Smerinthus, Latr., thirty species were catalogued by Gray,[101] of which I am only acquainted with the larvæ of eight (five European, and three North American). None of these in the last stage possess a complete subdorsal line together with oblique stripes. Neither, on the other hand, do any of them show a more advanced stage of development in having the red spots constantly formed into coloured border-stripes. We must therefore admit that they have all reached nearly the same stage of phyletic development. On turning to the doubtfully placed genus Calymnia, Boisduval, which is represented in Gray by only one species, figured by Westwood[102] as a Smerinthus, we first meet with an older stage of development of the genus.

The adult caterpillar of C. Panopus, from the East Indies, possesses, in addition to the oblique stripes, a completely developed subdorsal line,[103] and thus corresponds to the first stage of S. Populi. This species may possibly retain in its ontogeny a stage in which the oblique stripes are also absent, whilst the subdorsal line is present. From the early disappearance of the subdorsal line in the species of Smerinthus, we may venture to conclude that this character appeared at an early stage of the phylogeny, whilst the oblique stripes represent a secondary form of marking, as shall be further established subsequently.[104]

THE GENUS MACROGLOSSA, OCHSENHEIMER.

The adult larvæ of five species are known, and to these I can now add a sixth. In Gray the genus contains twenty-six species.[105] I cannot find any figures or descriptions of the young stages of these caterpillars, and I have myself only observed the complete ontogeny of one species.

By placing a captured female M. Stellatarum in a capacious breeding-cage, in the open air, I was enabled to procure eggs. The moth hovered about over the flowers, and laid its small, grass-green, spherical eggs (partly when on the wing), singly, on the leaves, buds, and stalks of Galium Mollugo. Altogether 130 were obtained in three days.[106]

First Stage.

After about eight days the caterpillars emerge. They are only two millimeters in length, and are at first yellowish, but soon become green, set with small single bristles, and they possess a short greenish caudal horn, which afterwards becomes black. The head is greenish-yellow. The young larvæ are entirely destitute of marking. ([Pl. III]., Fig. 1).

Second Stage.

The first moult takes place after four days, the caterpillar now acquiring the marking which it essentially retains to pupation.

Fine white subdorsal and spiracular lines appear, and at the same time a dark green dorsal line, which, however, does not arise from the deposition of pigment, as is generally the case, but from a division in the folds of the fatty tissue along this position. (Fig. 2, [Pl. III].)

The colour is now dirty green in all specimens, the skin being finely shagreened.

Third Stage.

The second moult, occurring after another period of four days, does not bring any change of marking, the colour only becoming somewhat darker. Length, twelve millimeters.

Fourth Stage.

The third moult (after another four days) likewise brings only a change of colouring, which is of such a nature that the caterpillar becomes dimorphic. At the same time that peculiar roughening of the skin takes place which, in the case of Chærocampa, was designated as “shagreening.” The colour is now light grass-green in some specimens, and dark green in others; in these last the subdorsal line is edged above with dark brown, and the spiracles are also of this colour. Length, seventeen millimeters.

Fifth Stage.

Four days later, after the fourth ecdysis, the dimorphism becomes a polymorphism. Five chief types can be distinguished:—

Variety I.—Light green (Fig. 7, [Pl. III].); dorsal line, blackish-green, strongly marked; subdorsal line broad, pure white, edged above with dark green; spiracular line, chrome-yellow; horn, black, with yellow tip and blue sides. Spiracles, blackish-brown, with narrow yellow border; legs, and extremities of prolegs, vermilion-red.

Variety II.—Blackish-brown (Fig. 6, [Pl. III].); head and prothorax, yellowish-brown; markings the same as above.

Variety III.—Blackish-green or greenish-black (Figs. 10 and 11, [Pl. III].); subdorsal line with blackish-green border above, gradually passing into a light green ground-colour; spiracular line, chrome-yellow; head and prothorax, greenish-yellow.

Variety IV.—Light green (Figs. 4 and 12, [Pl. III].); dorsal line quite feeble; subdorsal broad, only faintly edged with dark green; subspiracular line, faint yellowish; head and prothorax, green.

Variety V.—Brownish-violet (Fig. 8, [Pl. III].); the black dorsal line on a reddish ground either narrow or broad.

From these five varieties we see that the different types do not stand immediately next to one another; they are, in fact, connected by numerous transitional forms, the ground-colour varying greatly, being dark or light, yellowish or bluish. (Compare [Figs. 4], 5, 7, and 12.) The markings remain the same in all, but may be of very different intensities. The dorsal line is often only very feebly indicated, and the subdorsal line is frequently but faintly edged; the latter is also sometimes deep black above and bordered rather darkly beneath, the sides then being of a dark green, often with blackish dots on the yellow spiracular line (Fig. 5, [Pl. III].), this likewise being frequently edged with black. Only the horn and legs are alike in all forms. The green ground-colour passes into blackish-green, greenish or brownish-black, and again, from reddish-brown to lilac ([Fig. 3]), this last being the rarest colour.

The designation “polymorphism” may here appear very inapplicable, since we have no sharply distinct forms, but five very variable ground-colours connected by numerous intermediate modes of coloration. Should, however, the term “variability” be suggested, I am in possession of an observation which tends to show that the different colours have to a certain extent become fixed. I found a brown caterpillar, the five front segments of which were light green on the left side, and the fifth segment brown and green mixed (Fig. 9, [Pl. III].). Such parti-coloration can evidently only appear where we have contending characters which cannot become combined; just as in the case of hermaphrodite bees, where one half of a segment is male and the other half female, the two characters never becoming fused so as to produce a truly intermediate form.[107] From this observation, I conclude that some of the chief varieties of Stellatarum have already become so far removed from one another that they must be regarded as intermediate fixed forms, the colours of which no longer become fused together when they occur in one individual, but are developed in adjacent regions. Other facts agree with this conclusion. Thus, among the 140 adult larvæ which I bred from the batch of eggs above mentioned, the transition forms were much in the minority. There were forty-nine green and sixty-three brown caterpillars, whilst only twenty-eight were more or less transitional.

On these grounds I designate the phenomenon as “polymorphism,” although it may not yet have reached, as such, its sharpest limits. This would be brought about by the elimination of the intermediate forms.[108]

Immediately before pupation, all the caterpillars, both green and brown, acquire a lilac coloration. The fifth stage lasts seven days, and the whole larval development twenty-three days, the period from the deposition of the eggs to the appearance of the moth being only thirty-one days.

I have treated of the polymorphism of Stellatarum in detail, not only because it has hitherto remained unknown, and an analysis of such cases has been completely ignored,[109] but more particularly because, it appears to me, that important conclusions can be drawn therefrom. Moreover, such an extreme multiplicity of forms is interesting, since, so far as I know, polymorphism to this extent has not been observed in any insect.

The theoretical bearing of this polymorphism will be treated of subsequently. It is not in any way connected with a more advanced development of the markings, since M. Stellatarum shows in this respect a very low state of development. This species displays only two stages:—(1), complete absence of all markings; and (2), a simple subdorsal, with dorsal and spiracular lines. We must therefore admit that the phyletic development of the markings has for a long time remained at a standstill, or, what expresses the same thing, that the marking which the adult larva now possesses is extremely old.

In order to complete my observations on M. Stellatarum, I now add some remarks on the pupa, the colour variations of which it appeared of importance to investigate, owing to the extraordinary variability of the caterpillar. The pupa varies but very slightly; the ochreous yellow ground-colour sometimes passes into reddish, and sometimes into greenish; the rather complicated blackish-brown marking of streaky lines is very constant, especially on the wing portions, being at most only more or less strongly pronounced. The minute colour variations of the pupa therefore have no connection with the colour of the caterpillar, both green and brown larvæ furnishing sometimes reddish-yellow and sometimes greenish-yellow pupæ.

The comparison of M. Stellatarum with the other known species of the genus, brings scarcely any addition to our knowledge of the phyletic development. Thus, the two European species of which the caterpillars are known, viz. M. Fuciformis and Bombyliformis,[110] show essentially the same markings as Stellatarum, the chief element being a well-developed subdorsal line. The Indian M. Gilia, Herrich-Schäf., possesses also this line,[111] and, together with the East Indian M. Corythus, Walk.,[112] has oblique stripes in addition; the stripes do not, however, cross this line, but commence underneath it, and probably originated at a later period than the subdorsal line. Should this be the case, we must regard M. Corythus as representing a later phyletic stage. According to Duponchel’s figures, in both M. Fuciformis and Bombyliformis small oblique stripes (red) occur near the spiracles, but these have nothing to do with the oblique stripes of M. Gilia just mentioned, as they run in a contrary direction. Of the two European species, I have only seen the living caterpillar of Fuciformis, and this possessed no oblique stripes.

To these five species I am now enabled to add a sixth, viz. Macroglossa Croatica,[113] a species inhabiting Asia Minor and Eastern Europe, of which a specimen and notice were kindly forwarded to me by Dr. Staudinger. The adult caterpillar much resembles that of M. Stellatarum in form and marking, but the subdorsal line appears much less distinctly defined, and the dorsal and spiracular lines seem to be entirely absent. The colour is generally green, but varies to red, and the subdorsal is more distinct and sharper in the young than in the adult larva. The markings of this species do not therefore in any way surpass those of Stellatarum, but are, on the contrary, much simpler.[114]

THE GENUS PTEROGON, BOISD.[115]

Although I am acquainted with only a small portion of the developmental history of a single species of this genus, I will here proceed to record this fragment, since, taken in connection with two other species, it appears to me sufficient to determine, at least broadly, the direction of development which this genus has taken.

Pterogon Œnotheræ, Fabr.

The adult larva, as made known by many, and for the most part good figures, has very complicated markings, which do not seem derivable from any of the elements of marking in the Sphingidæ hitherto considered. I was therefore much surprised at finding a young caterpillar of this species, only twelve millimeters in length, of a light green colour, without any trace of the subsequent latticed marking, and with a broad white subdorsal line extending along all the twelve segments. ([Pl. VII]., Fig. 63). Judging from the size and subsequent development, this caterpillar was probably in the third stage.

The same colouring and marking remained during the following (fourth) stage; but in the position occupied by the caudal horn in other Sphingidæ, there could now be observed the rudiment of a future ocellus in the form of a round yellowish spot ([Pl. VII]., Fig. 64). The subdorsal line disappears suddenly in the fifth stage, when the larva becomes dark green (rarely) or blackish-brown; the latticed marking and the small oblique stripes are also acquired, together with the beautifully developed eye-spots, consisting of a yellow mirror with black nucleus and ground-area ([Pl. VII]., Fig. 65).

The North American Pterogon Gauræ and P. Abboti[116] also show markings precisely similar to those of this European species in the adult state; but in the two former the markings are of special interest as indicating the manner in which the primary Sphinx-marking has become transformed into that of the apparently totally different adult P. Œnotheræ. P. Gauræ is green, with a complicated latticed marking, which closer observation shows to arise from the dorsal line being resolved into small black dots, whilst the subdorsal line is broken up into black, white-bordered triangles. This caterpillar therefore gives fresh support to the remarkable phenomenon that the animals as well as the plants of North America are phyletically older than the European fauna and flora, a view which also appeared similarly confirmed by Deilephila Lineata, the representative form of D. Livornica. In entire accordance with this is the fact that the larva of P. Gauræ is without the eye-spot on the eleventh segment, and instead thereof still shows the original although small caudal horn. The perfect insect also resembles our P. Œnotheræ in colour and marking, but not in the form of the wings.

That the caterpillars of the genus Pterogon originally possessed the caudal horn we learn from P. Gorgoniades, Hübn.,[117] a species now inhabiting south-east Russia, and for a knowledge of which I am indebted to Dr. Staudinger’s collection. There are in this about eight blown specimens, from 3.7 to 3.9 centimeters in length, which show a marking, sometimes on a red and sometimes on a green ground, which unites this species with the young form of P. Œnotheræ, viz., a broad white subdorsal line, extending from the small caudal horn to the head. In addition to this, however, the caterpillar possesses an extraordinarily broad white red-bordered infra-spiracular line, a fine white dorsal stripe, and a similar line between the subdorsal and spiracular, i.e. a supra-spiracular line.

The caterpillars in Staudinger’s collection, notwithstanding their small size, all belong to the last stage, as the moth itself does not measure more than 2.6 centimeters in expanse, and is therefore among the smallest of the known Sphingidæ. This species has therefore in the adult condition a marking very similar to that of Œnotheræ when young—it bears to Œnotheræ the same relationship that Deilephila Hippophaës does to D. Euphorbiæ, only in the present case the interval between the two species is greater. Gorgoniades is obviously a phyletically older species, as we perceive from the marking and from the possession of a horn. We certainly do not yet know whether Œnotheræ possesses a horn in its earliest stages, although in all probability it does so; in any case the ancestor of Œnotheræ had a horn, since the closely allied P. Gauræ now possesses one.

We thus see that also in the genus Pterogon the marking of the caterpillars commences with a longitudinal line formed from the subdorsal; an infra-spiracular or also a supra-spiracular line (Gorgoniades) being added. A latticed marking is developed from the linear marking by the breaking up of the latter into spots or small patches, which finally (in Œnotheræ) become completely independent, their connection with the linear marking being no longer directly perceptible.

THE GENUS SPHINX, LINN.

Of this genus (in the narrow sense employed by Gray) I have only been able, in spite of all trouble, to obtain fertile eggs of one species. The females cannot be induced to lay in confinement, and eggs can only be obtained by chance.

I long searched in vain the literature of this subject for some account of the young stages of these caterpillars, and at length found, in a note to Rösel’s work, an observation of Kleemann’s on the young forms of Sphinx Ligustri, which, although far from complete, throws light on certain points.

From a female of S. Ligustri Kleemann obtained 400 fertile eggs. The caterpillars on emerging are “at first entirely light yellowish-green, but become greener after feeding on the fresh leaves;” the horn is also at first light green, and then becomes “darker.” The young larvæ spin webs, by which they fasten themselves to the leaves of their food-plant (this, so far as I know, has not been observed in any species of Sphingidæ). They moult four times, the border round the head and the purple stripes appearing after the third moult, these stripes “having previously been entirely white.” The ecdyses follow at intervals of about six days, increasing to about ten days after the fourth moult.[118]

From this short account we gather that in the third stage the marking consists of seven oblique white stripes, which acquire coloured edges in the fourth stage, a fact which I have myself frequently observed. On the most important point Kleemann’s observations unfortunately give no information—the presence or absence of a subdorsal line in the youngest stages. That he does not mention this character, can in no way be considered as a proof of its actual absence. I am rather inclined to believe that it is present in the first, and perhaps also in the second stage. There occur, however, species of the genus Sphinx (sensû strictiori) which possess a subdorsal line when young, as I think may be certainly inferred from the fact that the remains of such a line are present in the adult larva of S. Convolvuli.

This conclusion becomes still more certain on comparing the markings with those of a nearly allied genus; without such comparison the separation of the genus Macrosila, Boisd., from Sphinx is scarcely justifiable. If to these two genera we add Dolba, Walk., and Acherontia, Ochs., we must be principally struck with the great similarity in the markings, which often reaches to such an extent that the differences between two species consist entirely in small shades of colour, while the divergence of the moths is far greater.

Of the genera mentioned, I am acquainted altogether with fourteen species of caterpillars:—Macrosila Hasdrubal, Rustica,[119] and Cingulata;[119] Sphinx Convolvuli, Ligustri, Carolina,[119] Quinquemaculata,[119] Drupiferarum,[119] Kalmiæ,[119] and Gordius;[119] Dolba Hylæus;[119] Acherontia Atropos, Styx,[120] and Satanas.[120] With one exception all these caterpillars possess oblique stripes of the nature of those of the Smerinthus larvæ, and most of them are without any trace of a subdorsal line; one species—the North American M. Cingulata—has a completely developed subdorsal; and the typical European species, S. Convolvuli, has a rudimentary subdorsal line. The ground-colour in most of these species is of the same green as that of the leaves of their food-plants; some are brown, i.e. earth-coloured, and in these the markings do not appear so prominently; others again possess very striking colours (A. Atropos), the oblique stripes in these cases being very vivid. Only M. Hasdrubal[121] separates itself completely from this system of classification, since this species is deep black with narrow yellow rings, the horn and last segment being red.

The large and most striking caterpillar of M. Hasdrubal is the same which Wallace has made use of for his theory of the brilliant colours of caterpillars. The explanation of the origin of this widely divergent mode of marking could only be furnished by the ontogeny, in which one or another of the older phyletic stages will certainly have been preserved.

Strictly speaking the same should be said of the other species—nevertheless their comparison with the so similarly marked Smerinthinæ, together with the circumstance that in certain species a subdorsal line can be traced, makes it appear correct to suppose that here also the subdorsal was the primary marking, this line being subsequently entirely replaced by the oblique stripes. The Sphinginæ would therefore be a younger group than the Smerinthinæ, a conclusion which is borne out by the fact that in the former the oblique stripes have reached a higher development, being always of two, and sometimes even of three colours (S. Drupiferarum, white, red, black), whilst in the species of Smerinthus they only occasionally possess uniformly coloured borders.

THE GENUS ANCERYX, BOISD.

Although this genus is not admitted into most of the European catalogues—the solitary European species representing it being referred to the genus Sphinx, Linn.[122]—its separation from Sphinx appears to me to be justified, not because of the striking differences presented by the moths, but because the caterpillars, judging from the little we know of them, likewise show a similar degree of difference.

I have frequently succeeded in obtaining fertile eggs of Anceryx Pinastri and I will now give the developmental history of this caterpillar, which has already been figured with great accuracy in Ratzeburg’s excellent work on forest insects. Rösel was acquainted with the fact that the “pine moth” laid its eggs singly on the needles of the pine in June and July, and he described them as “yellowish, shining, oval, and of the size of a millet seed.”

On emerging, the caterpillars are six millimeters in length, of a light yellow colour, the head shining black with a yellow clypeus. The caudal horn, which is forked at the tip, is also at first yellowish, but soon becomes black. No particular marking is as yet present, but a reddish stripe extends along the region of the dorsal vessel, and the course of the spiracles is also marked by an orange-red line. (Fig. 53, A & B, [Pl. VI].)

As soon as the young larvæ are filled with food they acquire a greenish streak. The first moult occurs after four days, and immediately after this there is still an absence of distinct markings, with the exception of a greenish-white spiracular line. In the course of some hours, however, the original light green ground-colour becomes darker, and at the same time a sharp, greenish-white subdorsal line appears, together with a parallel line extending above the spiracles, which, in Pterogon Gorgoniades, has already been designated as the “supra-spiracular.” The dorsal line is absent: the head is light green, with two narrow blackish-brown lines surrounding the clypeus; the horn and thoracic legs are black; claspers, reddish green; length, twelve to thirteen millimeters. ([Fig. 54].)

Third Stage.

After another period of four days the second moult occurs, neither colour nor marking being thereby affected. Only the horn, now no longer forked, becomes brownish with a black tip. The young caterpillars are now, as before, admirably adapted to the pine needles, on which they feed by day, and from which they can only be distinguished with difficulty.

Fourth Stage.

The third moult also brings no essential change. The ground-colour and marking remain the same, only the spiracles, which were formerly dull yellowish, are now of a vivid brick-red. The horn becomes yellowish-red at the base.

Fifth Stage.

The marking is only completely changed in the fifth and last stage. A broad reddish-brown dorsal line replaces the subdorsal, more or less completely. The supra-spiracular line also becomes broken up into numerous short lengths, whilst the green ground-colour in some specimens becomes more or less replaced by a brownish shade extending from the back to the sides. Horn, black; the upper part of the first segment with a corneous plate, similar to that of the Deilephila larvæ.

This stage is very variable, as shown by the figures in various works. The variations arise on the one hand from the struggle between the green ground-colour and the reddish-brown extending from above, and, on the other hand, from a more or less complete disappearance of the associated longitudinal lines. The latter are sometimes completely retained, this being the case in a caterpillar figured by Hübner (Sphinges, III., Legitimæ C, b), where both the subdorsal and supra-spiracular lines are continuous from segment 11 to segment 1, an instance which may perhaps be regarded as a reversion to the primary form.

The entire change of the marking from the fourth to the fifth stage depends upon the fact that the young larvæ resemble the needles of the pine, whilst the adults are adapted to the branches. I shall return to this later.

The ontogeny of A. Pinastri makes us acquainted with three different forms of marking: (1) simple coloration without marking; (2) a marking composed of three pairs of parallel longitudinal lines; (3) a complicated marking, arising from the breaking up of the last and the addition of a darker dorsal line.

Of the fourteen species placed by Gray in the genus Anceryx, I find, in addition to the one described, notices of only two caterpillars:—

A. Coniferarum,[123] a North American species, lives on Pinus Palustris, and was figured by Abbot and Smith. Colour and marking very similar to A. Pinastri.

A. Ello, Linn.,[124] according to the authority of Mérian, is described by Clemens[125] as dark brown, “with a white dorsal line, and irregular white spots on the sides.” It lives on a “species of Psidium or Guava.”

Most of the species of Anceryx appear to live on Coniferæ, to which they show a general and decided adaptation. In the absence of decisive information, I partly infer this from the names, as Anceryx Juniperi (Africa). It has long been known that in our A. Pinastri the mixture of brown and fir-green, interspersed with conspicuous irregular light yellowish and white spots, causes the adult larva to present a very perfect adaptation to its environment. Of this caterpillar Rösel states:—“After eating it remains motionless, and is then difficult to see, because it is of the same colour as its food, since its brown dorsal line has almost the colour of the pine twigs; and who is not familiar with the fact that beneath the green needles there is also much yellow to be found?”

This adaptation to the needles and twigs obviously explains why this caterpillar in the adult condition is so far removed from those of the genus Sphinx, while the moths are so nearly related that they were only separated as a distinct genus when we became acquainted with a large number of species.

---

II.
Conclusions from Phylogeny.

The considerations previously set forth are entirely based on Fritz Müller’s and Haeckel’s view, that the development of the individual presents the ancestral history in nuce, the ontogeny being a condensed recapitulation of the phylogeny.

Although this law is generally true—all recent investigations on development having given it fresh confirmation—it must not be forgotten that this “recapitulation” is not only considerably abbreviated, but may also be “falsified,” so that a searching examination into each particular case is very desirable.

The question thus arises, in the first place, as to whether the markings of caterpillars, so distinct at the different stages of growth, are actually to be regarded as residual markings inherited from the parent-form; or whether their differences do not depend upon the fact that the caterpillar, in the course of growth, is exposed to different external conditions of life, to which it has adapted itself by assuming a different guise.

The former is undoubtedly the case. It can by no means be denied that the conditions of life in young caterpillars are sometimes different to those of the adults. It will, in fact, be shown later on, that in certain cases the assumption of a new guise at an advanced age actually depends upon adaptation to new conditions of life; but as a rule, the external conditions remain very similar during the development of the larva, as follows from the fact that a change of food-plant never takes place.[126] We should therefore rather expect a complete similarity of marking throughout the entire larval period, instead of the great differences which we actually observe.

Different circumstances appear to me to show that the markings of young larvæ are only exceptionally due to a new adaptation, but that as a rule they depend upon heredity. In the first place, there is the fact that closely allied species, exposed to precisely similar external conditions, as, for instance, Chærocampa Elpenor and Porcellus, possess exactly the same markings when young, these markings nevertheless appearing at different stages of growth. Thus, the subdorsal line first appears in Elpenor in the second stage, whilst in Porcellus it is present during the first stage. If this line were acquired by the young larva for adapting it at this age to special conditions of life, it should appear in both species at the same stage. Since this is not the case, we may conclude that it is only an inherited character derived from the adult ancestor of the two species, and now relegated to the young stages, being (so to speak), pushed further back in one species than in the other.

But the strongest, and, as it appears to me, the most convincing proof of the purely phyletic significance of the young larval markings, is to be found in the striking regularity with which these are developed in a similar manner in all allied species, howsoever different may be their external conditions of life. In all the species of the Chærocampa group (the genera Chærocampa and Deilephila) the marking—no matter how different this may be in later stages—arises from the simple subdorsal line. This occurs even in species which live on the most diverse plants, and in which the markings can be of no biological importance as long as the larvæ are so small as to be only visible through a lens, and where there can be no possible imitation of leaf-stalks or veins, the leaves and caterpillars being so very distinct.

Moreover, when in the Macroglossinæ (the genera Macroglossa, Pterogon, and Thyreus) we see precisely the same simple marking (the subdorsal) line retained throughout all the stages in two genera, whilst in the Smerinthinæ this line vanishes at a very early stage, and in the Sphinginæ is only present in traces, we can give but one explanation of these facts. We have here a fragmentary series representing the phyletic development of the Sphinx-markings, which latter have arisen from one original plan—the simple subdorsal line—and have then undergone further development in various directions. As this subsequent development advanced, the older phyletic stages would always be relegated to younger ontogenetic stages, until finally they would be but feebly represented even in the youngest stage (D. Euphorbiæ), or else entirely eliminated (most of the species of the genus Sphinx). I believe that no other sufficient explanation of these facts can be adduced. Granting that the correctness of the above views can no longer be doubted, we may now take up the certain position that the ontogeny of larval markings reveals their phylogeny, more or less completely, according to the number of phyletic stages omitted, or, in some exceptional cases, falsified. In other words, the ontogeny of larval markings is a more or less condensed and occasionally falsified recapitulation of the phylogeny.

Considering this to be established, we have next to deal with the uniformity of the developmental phenomena, from which we may then attempt to trace out the inciting causes underlying this development.

The law, or, perhaps better, the line of direction followed by the development, is essentially the following:—

1. The development commences with a state of simplicity, and advances gradually to one of complexity.

2. New characters first make their appearance in the last stage of the ontogeny.

3. Such characters then become gradually carried back to the earlier ontogenetic stages, thus displacing the older characters, until the latter disappear completely.

The first of these laws appears almost self-evident. Whenever we speak of development, we conceive a progression from the simple to the complex. This result therefore does nothing but confirm the observation, that we have actually here before us a development in the true sense of the word, and not simply a succession of different independent conditions.

The two following laws, on the other hand, lay claim to a greater importance. They are not now enunciated for the first time, but were deduced some years ago by Würtemberger[127] from a study of the ammonites. In this case also the new characters predominate in the later periods of life, and are then transferred back to the younger ontogenetic stages in the course of phyletic development. “The change in the character of the shell in ammonites, first makes itself conspicuous in the last chamber; but in the succeeding generations this change continually recedes towards the beginning of the spiral chambers, until it prevails throughout the greater part of the convolutions.”

In the same sense must also be conceived the case which Neumayr and Paul have recently made known respecting certain forms of Melanopsis from the West Sclavonian Paludina bed. In M. Recurrens the last convolutions of the shell are smooth, this being a new character; the small upper convolutions, however, are delicately ribbed, as is also the case with the last convolution of the immediate progenitor. The embryonic convolutions again are smooth, and the author believes (on other grounds) that the more remote progenitor possessed a smooth shell.

In this case therefore, and in that of the ammonites, every shell to a certain extent proclaims the ancestral history of the species; in one and the same shell we find different phyletic stages brought into proximity. The markings of caterpillars do not offer similar facilities; nevertheless I believe that by their means we are led somewhat further, and are able to enter more deeply into the causes underlying the processes of transformation, because we can here observe the living creature, and are thus enabled to study its life-history with more precision than is possible with a fossil species.

When, in 1873, I received Würtemberger’s memoir, I was not only struck with the agreement of his chief results with those which I had arrived at by the study of larval markings, but I was almost as much astonished at the great difference in the interpretation of the facts. The latter indicate the gradual backward transference of a new character from the latest to the earlier ontogenetic stages. Without further confirmation Würtemberger assumes that it is to a certain extent self-evident that the force producing this backward transference is the same as that which, according to his view, first called forth the character in question in the last stage, viz., natural selection. “Variations acquired at an advanced age of the organism may, when advantageous, be inherited by the succeeding generations, in such a manner that they always appear a little earlier than in the preceding generations.”

It is certainly theoretically conceivable that a newly acquired character, when also advantageous to the earlier stages, might be gradually transferred to these stages, since in this case those individuals in which this character appeared earliest would have the greatest chance of surviving. In the case of the development of larval markings, however, there are facts which appear to me to show that such backward transference of a new character is, in a certain measure, independent of the principle of utility, and that it must therefore be referred to another cause—to the innate law of growth which rules every organism.

When, in the larva of C. Elpenor, we perceive that the two eye-spots which are first formed on the fourth and fifth segments appear subsequently on the other segments as faint traces of no biological value whatever, we cannot explain this phenomenon by natural selection. We should rather say that in segmented animals there is a tendency for similar characters to be repeated on all the segments; and this simply amounts to the statement, that an innate law of growth is necessary for the repetition of such newly acquired characters.

The existence of such a law of growth, acting independently of natural selection, may therefore be considered as established, and indeed cannot be disputed (Darwin’s “correlation of growth”). In the present case it appears to me that an innate law of this kind, determining the backward transference of new characters, is deducible from the instances already quoted in another sense, viz., from the fact that in many cases characters which are decidedly advantageous to the adult are transferred to the younger stages, where they are at most of but indifferent value, and can certainly be of no direct advantage. This is the case with the oblique stripes of Smerinthus, which, in the adult larvæ, resemble the leaf ribs, as will be shown more fully later on, and, in conjunction with the green coloration, cause these caterpillars to be very difficult of detection on their food-plants. The insects are easily overlooked, and can only be distinctly recognized on close inspection.

Now these oblique stripes appear, in all the Smerinthus caterpillars known to me, in the second, and sometimes even in the first stage, i.e. in larvæ of from 0.7 to 1 centimeter in length. The stripes are here much closer together than the ribs of any of the leaves of either willow, poplar, or lime, and can therefore have no resemblance to these leaves. The young caterpillars are certainly not rendered more conspicuous by the oblique stripes, since they can only be recognized on close inspection. It is for this reason that the stripes have not been eliminated by natural selection.

The remarkable phenomenon of the backward transference of newly acquired characters may therefore be formulated as follows:—Changes which have arisen in the later ontogenetic stages have a tendency to be transferred back to the younger stages in the course of phyletic development.

The facts of development already recorded furnish numerous proofs that this transference occurs gradually, and step by step, taking the same course as that which led to the first establishment of the new character in the final ontogenetic stage.

Did this law not obtain, the ontogeny would lose much of the interest which it now possesses for us. It would then be no longer possible, from the ontogenetic course of development of an organ or of a character, to draw a conclusion as to its phylogeny. If, for instance, the eye-spots of the Chærocampa larvæ, which must have been acquired at a late age, were transferred back to the younger ontogenetic stages in the course of phyletic development, as eye-spots already perfected, and not showing their rudimentary commencement as indentations of the subdorsal line, the phenomenon would then give us no information as to the manner of their formation.

It is well known to all who have studied the developmental history of any group of animals, that no organ, or no character, however complex, appears suddenly in the ontogeny; whereas, on the other hand, it appears certain that new, or more advanced, but simpler characters, predominate in the last stage of development. We are thus led to the following modification of the foregoing conclusion:—Newly acquired characters undergo, as a whole, backward transference, by which means they are to a certain extent displaced from the final ontogenetic stage by characters which appear later.

This must be a purely mechanical process, depending on that innate law of growth, the action of which we may observe without being able to explain fully. Under certain conditions the operation of this law may be prevented by natural selection. Thus, for instance, if the young caterpillars of Anceryx Pinastri have not acquired the characteristic marking of the adults, it is probably because they are better protected by their resemblance to the green pine-needles than they would be if they possessed the pattern of the larger caterpillars in their last stage.

The backward transference of newly acquired characters may also possibly be accelerated when these characters are advantageous to the younger stages; but this transference takes place quite independently of any advantage if the characters are of indifferent value, being then entirely brought about by innate laws of growth.

That new characters actually predominate in the last stage of the ontogeny, may also be demonstrated from the markings of caterpillars. It is, of course, not hereby implied, that throughout the whole animal kingdom new characters can only appear in the last ontogenetic stage. Haeckel is quite correct in maintaining that the power of adaptation of an organism is not restricted to any particular period. Under certain circumstances transformations may occur at any period of development; and it is precisely insects undergoing metamorphosis that prove this point, since their larvæ differ so widely from their imagines that the earlier stages may be completely disguised. It is here only signified that, with respect to the development of caterpillars, new characters first appear in the adult. The complexity of the markings, which so frequently increases with the age of the caterpillar, can scarcely bear any other interpretation than that the new characters were always acquired in the last stage of the ontogeny. In certain cases we are able, although with some uncertainty, to catch Nature in the act of adding a new character.

I am disposed to regard the blood-red or rust-red spots which occur in the last stage of the three species of Smerinthus larvæ in the neighbourhood of the oblique stripes as a case in point. It has already been shown that these red spots must be regarded as the first rudiments of the linear coloured edges which reach complete development in the genus Sphinx. In some specimens of Smerinthus Tiliæ the spots coalesce so as to form an irregular coloured edge to the oblique stripes. In S. Populi they occur in many individuals, but remain always in the spot stage; whilst S. Ocellatus is but seldom, and S. Quercus appears never to be spotted.

The spots both of S. Tiliæ and Populi certainly do not show themselves exclusively in the fifth (last) stage, but also in the fourth, and sometimes in Populi even as early as the third stage, from which we might be disposed to conclude that the new character did not first appear in the last stage. But the majority of the spotted individuals first acquire their spots in the fifth stage, and only a minority in the fourth; so that their occasional earlier appearance must be ascribed to the backward transference of a character acquired in the fifth stage. Moreover, the fourth and fifth stages of the caterpillars are closely analogous both in size, mode of life, and marking, and are therefore analogous with reference to the environment, so that it is to be expected that new characters, when depending on adaptation, would be rapidly transferred from the fifth stage to the fourth.[128] We should thus have a case of the acceleration by natural selection, of processes determined by innate causes. Why changes should predominate in the last stage, is a question closely connected with that of the causes of larval markings in general, and may therefore be investigated later. But if we here assume in anticipation that all new markings depend on adaptation to the conditions of life, and arise through natural selection, it will not be difficult to draw the conclusion that such new characters must prevail in the last stage. There are two conditions favouring this view; the size of the insect, and the longer duration of the last stage. As long as the caterpillar is so small as to be entirely covered by a leaf, it only requires a good adaptation in colour in order to be completely hidden; independently of which, it is also possible that many of its foes do not consider it worth attacking at this stage. The last stage, moreover, is of considerably longer duration than any of the four preceding ones; in Deilephila Euphorbiæ this stage lasts for ten days, whilst the remaining stages have a duration of four days; in Sphinx Ligustri the last stage also extends over ten days, and the others over six days.

In its last stage, therefore, a caterpillar is for a longer period exposed to the danger of being discovered by its foes; and since, at the same time, its enemies become more numerous, and its increased size makes it more easy of detection, it is readily conceivable that a change in the conditions of life, such, for instance, as removal to a new food-plant, would bring about the adaptation of the adult larva as its chief result.

I shall next proceed to show how far the assumption here made—that all markings depend on natural selection—is correct.

---

III.
Biological Value of Marking in General.

Having now described the development of larval markings, so far as possible from their external phenomena, and having traced therefrom the underlying law of development, I may next proceed to the main problem—the attempt to discover those deeper inciting causes which have produced marking in general.

The same two contingencies here present themselves as those which relate to organic life as a whole; either the remarkably complex and apparently incomprehensible characters to which we give the name of markings owe their origin to the direct and indirect gradual action of the changing conditions of life, or else they arise from causes entirely innate in the organism itself, i.e. from a phyletic vital force. I have already stated in the Introduction why the markings of caterpillars appear to me such particularly favourable characters for deciding this question, or, more precisely, why these characters, above any others, appear to me to render such decision more easily possible; repetition is here therefore unnecessary.

The whole of the present investigation had not been planned when I joined with those who, from the first, admitted the omnipotence of natural selection as an article of faith or scientific axiom. A question which can only be solved by the inductive method cannot possibly be regarded as settled, nor can further evidence be considered unnecessary, because the first proofs favour the principle. The admission of a mysteriously working phyletic power appears very unsatisfactory to those who are striving after knowledge; the existence of this power, however, is not to be considered as disproved, because hundreds of characters can be referred to the action of natural selection, and many others to that of the direct action of the conditions of life. If the development of the organic world is to be considered as absolutely dependent on the influence of the environment, not only should we be able here and there to select at pleasure characters which appeared the most accessible for elucidating this point, but it becomes in the first place necessary to attempt to completely refer all characters belonging to any particular group of phenomena, however small this group might be, to known transforming factors. We should then see whether this were possible, or whether there would remain residual phenomena not explicable by known principles and compelling us to admit the existence of a force of development innate in the organism. In any case the “phyletic vital force” can only be got rid of by a process of elimination—by proving that all the characters generally occurring throughout the group of phenomena in question, must be attributed to other causes, and that consequently nothing remains for the action of the supposed phyletic vital force, which would in this manner be negatived, since we cannot infer the presence of a force if the latter exerts no action whatever.

I shall here attempt such an investigation of the group of phenomena displayed by larval markings, with special reference to those of the Sphingidæ. The alternatives upon which we have to decide are the following:—Are the markings of caterpillars purely morphological characters, produced entirely by internal causes? or, are they simply the response of the organism to external influences?

The solution of these questions will be arrived at by seeking to refer all the markings present to one of the known transforming factors, and the success or failure of this attempt will give the required decision. The first question to be attacked is obviously this,—whether the Sphinx-markings are actually, as they appear at first sight, purely morphological characters. If it can be shown that all these markings were originally of biological value, they must be attributed to the action of natural selection.

Did I here at once proceed to establish the biological value of larval markings—and especially of those of the Sphingidæ—so as to arrive in this manner at a conclusion as to their dependence upon natural selection, it would be impossible to avoid the consideration of the total coloration of the caterpillars, since the marking frequently consists only of a local strengthening of the colour, and cannot be comprehended without coming to this understanding. The action of the markings also often appears to be opposed to that of the colouring, making the caterpillar again conspicuous; so that the two factors must necessarily be considered together. I shall therefore commence the investigation with colour in general, and then proceed to treat of marking.

---

IV.
Biological Value of Colour.

The general prevalence of protective colouring among caterpillars has already been so frequently treated of that it is not here my intention to recall particular instances. In order to judge of the effect of marking, however, it will be well to bear in mind that these insects, being generally defenceless and thus requiring protection, have acquired the most diverse means of rendering themselves in some measure secure from their foes.

The sharp spines which occur on the caterpillars of many butterflies (Vanessa, Melitæa, Argynnis), and the hairs on those of many moths, serve for protective purposes. Among other means of protection—although in a different sense—we have in all the species of the great family of the Papilionidæ the strikingly coloured (yellowish red) odour-emitting tentacles concealed near the head, and suddenly protruded for terrifying foes; and likewise the forked horn at the tail of the caterpillars of the genus of moths Harpyia, the tentacles of which can be suddenly protruded in a similar manner. Adaptive colours and forms combined with certain habits[129] are, however, much more common than defensive weapons. Thus, the caterpillars of the Noctuæ belonging to the genus Catocala and its allies, feed only at night on the green leaves of various forest-trees; by day they rest in crevices of the bark on the tree trunk, which they resemble so perfectly in the colour of their peculiar glossy dull grey or brownish skin beset with small humps, that only sharp eyes can detect them, even when we are familiar with their habits.[130]

The striking resemblance of many moths to splinters of wood is well known, and to this is added a habit which helps their disguise, viz., that of remaining stiff and motionless on the approach of danger, just like a splinter projecting from the branch.[131] Among the moths coming under this category may be mentioned Cucullia Verbasci, and particularly those of the genus Xylina, which, when at rest, closely resemble a broken splinter of wood in the colour and marking of their fore wings, and when touched, have a habit of drawing in their legs and falling without opening their wings as though dead.

That simple adaptive colouring prevails widely among caterpillars is shown by the large number of green species.[132] It may be fairly said that all caterpillars which possess no other means of protection or defence are adaptively coloured. These facts are now well known; so also is the explanation of the varied and striking colours of many caterpillars given by Wallace.[133] There is, however, novelty in the proof contained in the foregoing descriptions of larval development, as to the manner in which the di- and polymorphism of caterpillars can be explained from the external phenomena which they present, these phenomena being well adapted for showing the great importance of protective colouring to the larvæ. We have here presented a double adaptation, although not quite of the nature of that which I formerly admitted on hypothetical grounds.[134] In the first place, from the developmental history there results the conclusion that all Sphinx-larvæ which, in the adult state, are di- or polymorphic, are unicolorous when young. Thus, the caterpillars of Chærocampa Elpenor all remain green till the fourth stage, when they mostly become light or dark brown, and only very seldom retain their green colour. Chærocampa Porcellus behaves in a precisely similar manner; as also does Pterogon Œnotheræ, which inhabits the same localities, and is found on the same food-plant, but is not very closely related to the Chærocampa. In this species also (P. Œnotheræ) the brown is more common than the green form in the adult state, both varieties showing a complicated marking. The young larvæ possess only a light green colour, and a pure white subdorsal line as the only marking; they are so well adapted to the leaves of their food-plants, Epilobium Hirsutum, and E. Rosmarinifolium, that they can only be detected with great difficulty. After the third moult they become brown, and can be easily seen when at rest on their food-plant.

Now in all known caterpillars brown colours are adaptive, sometimes causing a resemblance to the soil, and at others to dead leaves or branches. As soon, therefore, as the caterpillars have attained a considerable size, they remain concealed by day.[135] The truth of this observation not only appears from various entomological notes, but I have frequently convinced myself of its accuracy. I well remember from the earliest times that C. Elpenor, especially when the larva is adult, always rests by day among the dead branches and leaves of its shrub-like food-plant, Epilobium Hirsutum; and even when this species lives on the low-growing Epilobium Parviflorum, it conceals itself by day on the ground, among the tangled leaves and branches. I have observed that Sphinx Convolvuli has a precisely similar habit, for which reason it is difficult to obtain, even in localities where it occurs very commonly.

In the neighbourhood of Basle I once found at mid-day a brown caterpillar of Pterogon Œnotheræ on an isolated dead branch of Epilobium Rosmarinifolium, and I was informed by Herr Riggenbach-Stähelin—a collector of great experience who accompanied me—that these caterpillars always rest (by day) on withered plants as soon as they become brown, but before this change they are only to be found on green plants.

Thus, it cannot well be doubted that the change of colour is associated with a change in the habits of life, and the question arises as to which has been the primary change.

If the view here entertained, that the later brown coloration is adaptive, be correct, the species must have first acquired the habit of concealing itself by day on the ground and among dead herbage, before the original green colour could have been changed into brown by natural selection. This must represent the actual facts of the case.

Nearly allied species which at an advanced age are not dimorphic, but are darkly coloured in all individuals, are especially calculated to throw some light on this point. For instance, the caterpillar of Deilephila Vespertilio, which comes under this denomination, is light green when young, and rests both by day and night on the leaves of the plant on which it feeds. As soon as it acquires its dark colour—after the third moult—it changes its habits, concealing itself by day on the ground and feeding only by night. For this reason collectors prefer seeking for it in the evening, or with a lantern by night.

The most instructive case, however, is that of Deilephila Hippophaës, in which no change of colour is associated with age, the caterpillar, throughout its whole life, remaining of a greyish green, which exactly matches the colour of the leaves of its food-plant, Hippophae Rhamnoides. Nevertheless this species also possesses the habit of feeding only at night as soon as it has attained to a considerable size, hiding itself by day at the root of its food-plant. Collectors expressly state that this larva can scarcely be found by day, and recommend that it should be sought for at night with a lantern.

From the foregoing facts and considerations it may fairly be concluded, that the habit of hiding by day, possessed by these and other allied caterpillars, was acquired when they resembled the leaves in colour, and that the adaptation to the colour of the soil, or dead foliage and withered branches, ensued as a secondary consequence.

But why have these caterpillars acquired such a habit, since they appear to be perfectly protected by their resemblance in colour to the green leaves? The answer to this question is easily given when we consider in which species this habit generally occurs.

Does the habit prevail only among the species of the one genus Deilephila, and in all the species of this genus? This is by no means the case, since, on the one hand, many species of Deilephila, such as D. Euphorbiæ, Galii, Nicæa, and Dahlii, do not possess the habit, and, on the other hand, it occurs in species of other genera, such as Macroglossa Stellatarum, Sphinx Convolvuli, and Acherontia Atropos.

The habit in question must therefore be the result of certain external conditions of life common to all those species which rest by day. The mode of life common to them all is that they do not live on trees with large leaves or with thick foliage, but on low plants or small-leaved shrubs, such as the Sea Buckthorn.[136] I believe I do not err when I attribute the habit possessed by the adult larvæ, of concealing themselves by day, to the fact that the green colour is protective only so long as they are small—or, more precisely speaking, as long as their size does not considerably exceed that of a leaf or twig of their food-plant. When they become considerably larger, they must become conspicuous in spite of their adaptive colour, so that it would then be advantageous for them to conceal themselves by day, and to feed only by night. This habit they have acquired, and still observe, even when the secondary adaptation to the colour of the soil, &c., has not been brought about. We learn this from D. Hippophaës, which remains green throughout its whole larval existence; and no less from the green forms of the adult larvæ of Sphinx Convolvuli, Chærocampa Elpenor, and Porcellus, all of which conceal themselves by day in the same manner as their brown allies.

It may be objected that there are Sphinx-larvæ—instances of which I have myself adduced—which live on low small-leaved plants, and which nevertheless do not hide themselves by day. This is the case with the spurge-feeding D. Euphorbiæ, so common in many parts of Germany. This caterpillar must, however, be classed with those which, on account of their distastefulness, or for other reasons to be subsequently considered, are rejected by birds and other larger foes, and which, as Wallace has shown, derive advantage from being coloured as vividly as possible. I shall return to this subject later, when treating of the biological value of special markings.

On the other hand, it is readily conceivable that, from the conditions of life of caterpillars living on trees or shrubs with dense foliage, the habit of resting by day and descending from the tree for concealment would not have been acquired. Such larvæ are sufficiently protected by their green colour among the large and numerous leaves; and I shall have occasion to show subsequently that their markings increase this protective resemblance.

The di- or polymorphism of the larvæ of the Sphingidæ does not therefore depend upon a contemporaneous double adaptation, but upon the replacement of an old protective colour by a new and better one, and therefore upon a successive double adaptation. The adult caterpillars of C. Elpenor are not sometimes brown and sometimes green because some individuals have become adapted to leaves and others to the soil, but because the anciently inherited green has not yet been completely replaced by the newly acquired brown coloration, some individuals still retaining the old green colour.

When, in another place,[137] I formerly stated “that a species can become adapted in this or that manner to given conditions of life, and that by no means can only one best adapted form be allowed for each species,” this statement is theoretically correct speaking generally, but not in its application to the present class of cases. A comparison with one another of those caterpillars which repose by day, distinctly shows that they all possess a tendency to abandon the green and assume a dull colour, but that this process of replacement has advanced further in some species than in others. It will not be without interest to follow this operation in some detailed cases, since we may thus obtain an insight into the processes by which polymorphism has arisen, as well as into the connection between this phenomenon and simple variability.

In D. Hippophaës the process has either not yet commenced, or is as yet in its first rudiments. If we may trust the statements of authors, together with the ordinary green form there occurs, rarely, a silver-grey variety, which may be regarded as the beginning of a process of colour substitution. Among thirty-five living specimens of this scarce species which I was able to procure, the grey form did not occur, neither have I found it in collections.

In Macroglossa Stellatarum we see the transforming process in full operation. A large number of individuals (about thirty-five per cent.) are still green; the number of dark-coloured individuals reaches forty-six per cent., these, therefore, preponderating; whilst between the two extremes there are about nineteen per cent. of transition forms, showing all possible shades between light green and dark blackish-brown or brownish-violet, and even, in solitary individuals, pure violet (See Figs. 3–12, [Pl. III].). The relatively small number of the intermediate forms, taken in connection with the fact that all the 140 specimens employed in my investigation were obtained from one female, leads to the conclusion that these forms owe their existence to cross-breeding. It would be superfluous to attempt to prove this last conclusion with reference to the before-mentioned case, in which a caterpillar was streaked with brown and green (Fig. 9, [Pl. III].).

The process of transformation, as already mentioned, advances in such a manner that the intermediate forms diminish relatively to the dark individuals. This is found to be the case with Sphinx Convolvuli, and almost to the same extent with Chærocampa Elpenor, in both of which species the green caterpillars are the rarest.[138] Forms truly intermediate in colour between green and brown no longer occur, but apparently only different shades of light and dark brown, passing into brownish-black.

The process has again made a further advance in Chærocampa Porcellus and Celerio as well as in Pterogon Œnotheræ. In all these species the green form occurs,[139] but so rarely that very few collectors have seen it. The brown form has therefore in these cases nearly become the predominant type, and the solitary green specimens which occasionally occur, may be regarded as reversions to an older phyletic stage.

Deilephila Livornica appears to have reached a similar stage, but the caterpillar of this species has been so imperfectly observed, that it is difficult to determine, even approximately, the relative proportion of the brown to the green individuals. I have only seen one of the latter in Dr. Staudinger’s collection (Compare Fig. 62, [Pl. VII].).

In Deilephila Vespertilio, Euphorbiæ, Dahlii, Mauritanica, Nicæa, and Galii, the green form has completely disappeared. The blackish olive-green colour shown by many caterpillars of the two last species, can be considered as a faint retention of the light green colour which they formerly possessed, and which they both show at the present time in their young stages.

Beginning with the appearance of single darker individuals, we pass on in the first place to a greater variability of colouring, and from this, by the greater diminution of the intermediate forms, to polymorphism; the complete extermination of these forms ending in dimorphism. The whole process of transformation has been thus effected:—As the new colouring always prevailed over the old, the latter was at length completely displaced, and the caterpillars, which were at first simply variable, became polymorphic and then dimorphic, finally returning to monomorphism.

We thus see the process of transformation still going on, and no doubt can arise as to its inciting causes. When a character can with certainty be ascribed to adaptation, we can explain its origin in no other way than by the action of natural selection. If, as I believe, it can and has been shown, not only that caterpillars in general possess adaptive colours, but that these colours can change during the lifetime of one and the same species, in correspondence with external conditions, we must certainly gain a very high conception of the power which natural selection exerts on this group of living forms.[140]

---

V.
Biological Value of Special Markings.

The following questions now present themselves: Have the markings of caterpillars any biological value, or are they in a measure only sports of nature? Can they be considered as partially or entirely the result of natural selection, or has this agency had no share in their production?

The problem here offers itself more distinctly than in any other group of living forms, because it presents an alternative without a third possibility. In other words, if it is not possible to show that larval markings have a distinct biological significance, there remains only for their explanation the assumption of a phyletic force, since the direct action of the environment is insufficient to account for such regularity of development throughout a series of forms. The explanation by sexual selection is excluded ab initio, since we are here concerned with larvæ, and not with reproductive forms.[141]

The biological significance of marking—if such significance it possess—will be most easily investigated by examining whether species with similar markings have any conditions of life in common which would permit of any possible inference as to the significance of the markings.

Among the Sphingidæ we find four chief forms of marking; (1) complete absence of all marking; (2) longitudinal stripes; either a simple subdorsal or this together with a spiracular and dorsal line; (3) oblique stripes; (4) eye-spots and ring-spots, single, paired, or in complete rows.

Now if we consider in which species these four kinds of marking are of general occurrence, not only in the small group of the Sphingidæ but in the whole order Lepidoptera, we shall arrive at the following results:—

1. Complete absence of marking, so common in the larvæ of other insects, such as the Coleoptera, is but seldom found among Lepidopterous caterpillars.

To this category belong all the species of Sesiidæ (the genera Sesia, Trochilia, Sciapteron, Bembecia, &c.), the larvæ of which, without exception, are of a whitish or yellowish colour, and live partly in the wood of trees and shrubs and partly in the shoots of herbaceous plants. Subterranean larvæ also, living at the roots of plants, such as Hepialus Humuli at the roots of hop, and H. Lupulinus at those of Triticum Repens, possess neither colour nor marking. These, like the foregoing, are yellowish-white, evidently because they are deprived of the influence of light.[142] The larvæ of certain small moths, such as Tortrix Arbutana and Pomonana, which live in fruit, and many case-bearing Tineina, are likewise without marking and devoid of bright colour, being generally whitish. Many of the small caterpillars which feed exteriorly are also—so far as my experience extends—without definite markings, these being among the most minute, such as the greenish leaf-mining species of Nepticula. It is among the larger species that we first meet with longitudinal and oblique stripes. Eye-spots do not occur in any of these larvæ, a circumstance of the greatest importance for the biological significance of this character, as will be shown subsequently. The small size of the caterpillars cannot be the sole cause of the absence of such eye-spots, since in young Smerinthus caterpillars one centimeter long, the oblique stripes are beautifully developed, and the larvæ of many of the smaller moths considerably exceed this size. The surface of these caterpillars therefore, i.e., the field on which markings are displayed, is not absolutely too small for the development of such a character.

Besides the larvæ of the Micro-lepidoptera and of those species living in the dark, there is also a complete absence of marking in the young stages of many caterpillars. Thus, all the Sphingidæ of which I have been able to observe the development, show no markings immediately after emergence from the egg; in many they appear very soon, even before the first moult, and, in other species, after this period.

2. The second category of markings, longitudinal stripes, is very widely distributed among the most diverse families. This character is found among the larvæ of butterflies, Sphingidæ, Noctuæ, Micro-lepidoptera, &c., but in all these groups it is absent in many species. This last fact is opposed to the view that this character is purely morphological, and leads to the supposition that it may have a biological value, being of service for the preservation of the individual, and therefore of the species.

I find that such marking is of service, stripes extending longitudinally along the upper surface of the caterpillar generally making the latter less conspicuous. This, of course, does not hold good under all circumstances, since there are many species with very striking colours which possess longitudinal stripes. Let us consider, however, a case of adaptive colouring, such as a green caterpillar, which, on this account only, is difficult to see, since it accords with the colour of the plant on which it lives. If it is a small caterpillar, i.e., if its length and thickness do not considerably exceed that of the parts of its food-plant, it can scarcely be better concealed—stripes would hardly confer any special advantage unless the parts of the plant were also striped. But the case is quite different if the caterpillar is considerably larger than the parts of the plant (leaves, stalks, &c.). The most perfect adaptive colouring would not now prevent it from standing out conspicuously as a larger body, among the surrounding parts of the plants. It must be distinctly advantageous therefore to such a caterpillar to be striped, since these markings to a certain extent divide the large body into several longitudinal portions—they no longer permit it to be seen as a whole, and thus act more effectively than mere assimilative colouring in causing it to escape detection. This protection would be the more efficacious if the stripes resembled the parts of the plant in colour and size, such, for instance, as the lines of light and shadow produced by stalks or by long and sharp-edged leaves.

If this view be correct, we should expect longitudinal stripes to be absent in the smallest caterpillars, and to be present more especially in those species which live on plants with their parts similarly disposed, i.e., on plants with numerous thin, closely-growing stalks and grass-like leaves, or on plants with needle-shaped leaves.

It has already been mentioned that the smallest species are devoid of longitudinal striping. The larvæ of the Micro-lepidoptera show no such marking, even when they do not live in the dark, but feed either on the surface or in superficial galleries of the leaves (Nepticula, &c.), in which they must be exposed to almost as much light as when living on the surface. The fact that the subdorsal line sometimes appears in very young Sphinx-larvæ is explained, as has already been shown, by the gradual backward transference of adaptational characters acquired in the last stage of development.

It can easily be demonstrated that longitudinally striped caterpillars mostly live on plants, of which the general appearance gives the impression of a striped arrangement. We have only to consider in connection with their mode of life, any large group of adaptively coloured species marked in this manner. Thus, among the butterflies, nearly all the Satyrinæ possess larvæ conspicuously striped—a fact which is readily explicable, because all these caterpillars live on grasses. This is the case with the genera Melanargia, Erebia, Satyrus, Pararge, Epinephele, and Cænonympha, no species of which, so far as the larvæ are known, is without longitudinal stripes, and all of which feed on grasses. It is interesting that here also, as in certain Sphingidæ, some species are brown, i.e., adapted to the soil, whilst the majority are green, and are therefore adapted to living grass. Just as in the case of the Sphingidæ also, the brown species conceal themselves by day on the earth, whilst some of the green species have likewise acquired this habit. I have already shown how this habit originates from the increasing size of the growing larva, which would otherwise become too conspicuous, in spite of adaptive colour and marking. A beautiful confirmation of this view is found in the circumstance that only the largest species of Satyrus, such as S. Proserpinus, Hermione, Phædrus, &c., possess brown caterpillars. I should not be surprised if a more exact investigation of these species, which have hitherto been but seldom observed, revealed in some cases a dimorphism similar to that of the Sphingidæ; and I believe that I may venture to predict that the young stages of all these brown larvæ—at present quite unknown—are, as in the last-named group, green.

Besides the Satyrinæ, most of the larvæ of the Pierinæ and Hesperidæ possess longitudinal stripes, which are generally less strongly pronounced than in the former subfamily. Some of the Pierinæ live on Cruciferæ, of which the narrow leaves and thin leaf- and flower-stalks present nothing but a linear arrangement; other species of this group, however, feed on Leguminosæ (Lathyrus, Lotus, Coronilla, Vicia), and some few on broad-leaved bushes (Rhamnus). This last fact may appear to be opposed to the theory; but light lateral stripes, such for example, as those possessed by Gonepteryx Rhamni, can never be disadvantageous, and may be of use, even on large leaves, so that if we consider them as an inherited character, there is no reason for natural selection to eliminate them. In the case of caterpillars living on vetch, clover, and other Leguminosæ, it must not be forgotten that, although their food-plants do not present any longitudinal arrangement of parts, they always grow among grasses, the species feeding on such plants always resting between grass stems, and very frequently on the grass itself, so that they can have no better protective marking than longitudinal stripes. The striping of the Hesperidæ larvæ, which partly feed on grasses but mostly on species of Leguminosæ, can be explained in a similar manner.

It is not here my intention to go through all the groups of Lepidoptera in this manner. The instances adduced are quite sufficient to prove that longitudinal stripes occur wherever we should expect to find them, and that they really possess the biological significance which I have ascribed to them. That these markings are occasionally converted into an adaptive imitation of certain special parts of a plant, is shown by the larvæ of many moths, such for example as Chesias Spartiata, which lives on broom (Spartium Scoparium), its longitudinal stripes deceptively resembling the sharp edges of the stems of this plant.[143]

3. Oblique striping. Can the lilac and white oblique stripes on the sides of a large green caterpillar, such as those of Sphinx Ligustri; or the red and white, or white, black, and red stripes of Smerinthus Tiliæ and Sphinx Drupiferarum respectively, be of any possible use? Have we not here just one of those cases which clearly prove that such a character is purely morphological, and worthless for the preservation of the individual? Does not Nature occasionally sport with purposeless forms and colours; or, as it has often been poetically expressed, does she not here give play to the wealth of her phantasy?

At first sight this indeed appears to be the case. We might almost doubt the adaptive importance of the green ground-colour on finding coloured stripes added thereto, and thus—as one might suppose—abolishing the beneficial action of this ground-colour, by making the insect strikingly conspicuous. But this view would be decidedly incorrect, since oblique stripes are of just the same importance as longitudinal stripes. The former serve to render the caterpillar difficult of detection, by making it resemble, as far as possible, a leaf; they are imitations of the leaf-veins.

Nobody who is in the habit of searching for caterpillars will doubt that, in cases where the oblique stripes are simply white or greenish-white, it is extremely difficult to see the insect on its food-plant, e.g. S. Ocellatus on Salix; not only because it possesses the colour of the leaves, but no less because its large body does not present an unbroken green surface, which would bring it into strong contrast with the leaves, and thus arrest the attention. In the case of the species named, the coloured area of the body is divided by oblique parallel stripes, just in the same manner as a willow leaf. In such instances of course we have not presented to us any special imitation of a leaf with all its details—there is not a perfect resemblance of the insect to a leaf, but only an arrangement of lines and interspaces which does not greatly differ from the division of a leaf by its ribs.

That this view is correct is shown by the occurrence of this form of marking. It is on the whole rare, being found, besides in many Sphingidæ, in isolated cases in various families, but is always confined to those larvæ which live on ribbed leaves, and never occurring in species which feed on grasses or on trees with needle-shaped leaves. This has already been shown with respect to the Sphingidæ, in which the oblique stripes are only completely developed in the subfamilies Smerinthinæ and Sphinginæ. The species of Smerinthus all live on trees such as willows, poplars, lime, oak, &c., and all possess oblique stripes. The genus Anceryx also belongs to the Sphinginæ, and these caterpillars, as far as known, live on trees with needle-shaped leaves. The moths of this last genus are very closely allied to the species of Sphinx, not only in form and colour, but also in many details of marking. The larvæ are however different, this distinction arising entirely from their adaptation to needle-shaped leaves, the Sphinx caterpillars being adapted to ordinary foliage. The species of Anceryx, as has been already shown, are brown mixed with green, and never possess even a trace of the oblique stripes, but have a latticed marking, consisting of many interrupted lines, which very effectively serves to conceal them among the needles and brown bark of the Coniferæ.

Of the Sphinginæ living on plants with ordinary foliage, not a single species is without oblique stripes. I am acquainted with ten species of caterpillars and their respective food-plants, viz. Sphinx Carolina, Convolvuli, Quinquemaculata, Prini, Drupiferarum, Ligustri; Macrosila Rustica and Cingulata; Dolba Hylæus and Acherontia Atropos.

Besides among the Sphinginæ, oblique stripes occur in the larvæ of certain butterflies, viz. Apatura Iris, Ilia, and Clytie, all of which live on forest trees (aspen and willows), and are excellently adapted to the leaves by their green colour. In addition to these, I am acquainted with the larvæ of some few moths, viz. of Aglia Tau and Endromis Versicolora, both of which also live on forest trees.

Oblique stripes also occasionally occur in the smaller caterpillars of Noctuæ, Geometræ, and even in those of certain Pyrales, in all of which they are shorter and differently arranged. In these cases also, my theory of adaptation holds good, but it would take us too far if I attempted to go more closely into them. I will here only mention the extraordinary adaptation shown by the caterpillar of Eriopus Pteridis. This little Noctuid lives on Pteris Aquilina; it possesses the same green colour as this fern, and has double oblique white stripes crossing at a sharp angle on each segment, these resembling the lines of sori of the fern-frond so closely, that the insect is very difficult to perceive.

After all these illustrations it can no longer remain doubtful that the oblique stripes of the Sphingidæ are adaptive. But how are the coloured edges bordering these stripes in so many species to be explained?

I must confess that I long doubted the possibility of being able to ascribe any biological value to this character, which appeared to me only conspicuous, and not protective. Cases may actually occur in which the brightly coloured edges of the oblique stripes make the caterpillar conspicuous—just in the same manner as any marking may bring about a conspicuous appearance by presenting a striking contrast of colour. I am acquainted with no such instance, however. As a rule, in all well-adapted caterpillars, considering their colour in its totality, this is certainly not the case. The coloured edges, on the contrary, enhance the deceptive appearance by representing the oblique shadows cast by the ribs on the under-side of the leaf; all these caterpillars rest underneath the leaves, and never on the upper surface.

This explanation may, perhaps, at first sight appear far-fetched, but if the experiment be made of observing a caterpillar of Sphinx Ligustri on its food-plant, not immediately before one’s eyes in a room, but at a distance as under natural conditions, it will be found that the violet edges do not stand out brightly, but show a colour very similar to that of the shadows playing about the leaves. The coloured edges, in fact, produce a more effective breaking up of the large green surface of the caterpillar’s body, than whitish stripes alone. Of course if the insect was placed on a bare twig in the sun, it would be easily visible at a distance; the larva never rests in such a position, however, but always in the deep shadow of the leaves, in which situation the coloured edges produce their peculiar effect. It may be objected that the oblique white stripes, standing simply on a dark green ground-colour, would produce the same effect, and that my explanation therefore leaves the bright colouring of these edges still unaccounted for. I certainly cannot say why in Sphinx Ligustri these edges are lilac, and in S. Drupiferarum, S. Prini, and Dolba Hylæus red, nor why they are black and green in Macrosila Rustica, and blue in Acherontia Atropos. If we knew exactly on what plants these caterpillars fed originally, we might perhaps indulge in comparing with an artistic eye the shadows playing about their leaves, seeing in one case more red, and in another more blue or violet. The coloured stripes of the Sphingidæ must be regarded as the single strokes of a great master on the countenance of a human portrait. Looked into closely, we see red, blue, or even green spots and strokes; but all these colours, conspicuous when close, disappear on retreating, a general effect of colour being then produced, which cannot be precisely described by words.

Quite in accordance with this explanation, we see caterpillars with the brightest coloured stripes concealing themselves in the earth by day, and betaking themselves to their food-plants only in the dusk of the evening or dawn of morning and even during the night; i.e. in a light so faint that feeble colours would produce scarcely any effect. The bright blue of Acherontia Atropos, for example, would give the impression of oblique shadows without any distinctive colour.

It is precisely the case of this last caterpillar, which formerly appeared to me to present insurmountable difficulties to the explanation of the coloured stripes by adaptation, and I believed that this insect would have to be classed with those species which are brightly coloured because they are distasteful, and are avoided by birds. But although we have no experiments on this point, I must reject this view. Unfortunately, we know scarcely anything of the ontogeny of this caterpillar; but we know at least that the young larvæ (stage four) are greener than the more purely yellow ones of the fifth stage (which, however, are also frequently green), and we know further that some adults are of a dark brownish-grey, without any striking colours. From analogy with the dimorphism of the species of Chærocampa and Sphinx, fully considered previously, it must therefore be concluded that in this case also, a new process of adaptation has commenced—that the caterpillar is becoming adapted to the soil in and on which it conceals itself by day.[144] An insect which acquires undoubted protective colours cannot, however, be classed with those which possess an immunity from hostile attacks.

That the coloured edges are correctly explained as imitations of the oblique shadows of the leaf-ribs, may also be proved from another point of view. Let us assume, for the sake of argument, that these coloured stripes are not adaptive, and that they have not been produced by natural selection, but by a hypothetical phyletic force. We should then expect to see them appear at some period in the course of the phyletic development—perhaps at first only in solitary individuals, then in several, and finally in all; but we certainly could not expect that at first single, irregular, coloured spots should arise in the neighbourhood of the oblique white stripes—that these spots should then multiply, and fusing together, should adhere to the white stripes, so as to form an irregular spot-like edge, which finally becomes formed into a straight, uniformly broad stripe. The phyletic development of the coloured edges takes place, however, in such a manner, the species of Smerinthus, as has already been established, showing this with particular distinctness. In S. Tiliæ the course of development can be followed till the somewhat irregular red border is formed. In the species of Sphinx this border has become completely linear. It is very possible that the ontogeny of S. Ligustri or Drupiferarum would reveal the whole process, although it may also be possible that owing to the contraction of the development, much of the phylogeny is already lost.

I have now arrived at the consideration of the last kind of marking which occurs in the Sphingidæ, viz.:—

4. Eye-spots and Ring-spots.—These markings, besides among the Sphingidæ, are found only in a very few caterpillars, such as certain tropical Papilionidæ and Noctuæ. I know nothing of the conditions of life and habits of these species, however, and without such knowledge it is impossible to arrive at a complete explanation.

With Darwin, I take an eye-spot to be “a spot within a ring of another colour, like the pupil within the iris,” but to this central spot “concentric zones” maybe added. In the Chærocampa larvæ and in Pterogon Œnotheræ, in which complete ocelli occur, there are always three zones—a central spot, the pupil, or, as I have called it, the “nucleus;” then a light zone, the “mirror;” and, surrounding this again, a dark zone (generally black), the “ground-area.”

As ring-spots I will consider those ocelli which are without the nucleus (pupil), and which are not therefore, strictly speaking, deceptive imitations of an eye, but present a conspicuous light spot surrounded by a dark zone.

Between these two kinds of markings there is, however, no sharp boundary, and morphologically they can scarcely be separated. Species with ring-spots sometimes have nuclei, and ocellated larvæ in some cases possess only a pale spot instead of a dark pupil. I deal here with the two kinds separately, because it happens that they appear in two distinct genera, in each of which they have their special developmental history. Ring-spots originate in a different position, and in another manner than eye-spots; but it must not, on this account, be assumed without further inquiry, that they are called into existence by the same causes; they must rather be investigated separately, from their origin.

Eye-spots are possessed by the genera Chærocampa and Pterogon; ring-spots by the genus Deilephila. In accordance with the data furnished by the above-given developmental histories, the origination of these markings in the two genera may be thus represented:—

In the genera named, eye-spots and ring-spots are formed by the transformation of single portions of the subdorsal line.

In Chærocampa the primary ocelli originate on the fourth and fifth segments by the detachment of a curved portion of the subdorsal, this fragment becoming the “mirror,” and acquiring a dark encircling zone (“ground-area”). The nucleus (pupil) is added subsequently.

In Deilephila we learn from the development of D. Hippophaës, that the primary annulus arises on the segment bearing the caudal horn (the eleventh) by the deposition of a red spot on the white subdorsal line, which is somewhat enlarged in this region. The formation of a dark “ground-area” subsequently occurs, and with this, at first the partial, and then the complete, detachment of the mirror-spot from the subdorsal line takes place.

In both genera the spots arise at first locally on one or two segments, from which they are transferred to the others as a secondary character. In Chærocampa this transference is chiefly backwards, in Deilephila invariably forwards.

We have now to inquire whether complete eye-spots—such as those of the Chærocampa larvæ—have any significance at all, and whether they are of biological importance. It is clear at starting, that these spots do not belong to that class of markings which make their possessors more difficult of detection; they have rather the opposite effect.

We might thus be disposed to class ocellated caterpillars with those “brightly coloured” species which, like the Heliconinæ and Danainæ among butterflies, possess a disgusting taste, and which to a certain extent bear the signal of their distastefulness in their brilliant colours. But even if I had not found by experiment that our native Chærocampa larvæ were devoured by birds and lizards, and that they are not therefore distasteful to these insect persecutors, from the circumstance that these caterpillars are all protectively coloured, it could have been inferred that they do not belong to this category. It has been found that all adaptively coloured caterpillars are eaten, and one and the same species cannot possibly be at the same time inconspicuously (adaptively) and conspicuously coloured; the one condition excludes the other.

What other significance can eye-spots possess than that of making the insects conspicuous? Had we to deal with sexually mature forms, we should, in the first place, think of the action of sexual selection, and should regard these spots as objects of taste, like the ocelli on the feathers of the peacock and argus-pheasant. But we are here concerned with larvæ, and sexual selection is thus excluded.

The eye-spots must therefore possess some other significance, or else they are of no importance at all to the life of the insect, and are purely “morphological characters;” in which case, supposing this could be proved, they would owe their existence exclusively to forces innate in the organism itself—a view which very closely approaches the admission of a phyletic vital force.

I am of opinion, however, that eye-spots certainly possess a biological value as a means of terrifying—they belong to that numerous class of characters which occur in the most diverse groups of animals, and which serve the purpose of making their possessors appear as alarming as possible.

The caterpillars of the Sphingidæ are known to behave themselves in different manners when attacked. Some species, such, for instance, as Sphinx Ligustri and Smerinthus Ocellatus, on the approach of danger assume the so-called Sphinx attitude; if they are then actually seized, they dash themselves madly to right and left, by this means not only attempting to get free, but also to terrify their persecutor. This habit frequently succeeds with men, and more especially with women and children; perhaps more easily in these cases than with their experienced foes, birds.

The ocellated Chærocampa larvæ behave differently. They remain quiet on being attacked, and do not put on a Sphinx-like attitude, but only withdraw the head and three small front segments into the large fourth segment, which thus becomes much swollen, and is on this account taken for the head of the insect by the inexperienced.[145] Now the large eye-spots are situated on the fourth segment, and it does not require much imagination to see in such a caterpillar an alarming monster with fiery eyes, especially if we consider the size which it must appear to an enemy such as a lizard or small bird. [Fig. 28] represents the larva of C. Porcellus in an attitude of defence, although but imperfectly, since the front segments can be still more withdrawn.

These facts and considerations do not, however, amount to scientific demonstration, and I therefore made a series of experiments, in order to determine whether these caterpillars did actually frighten small birds. The first experiment proved but little satisfactory. A jay, which had been domesticated for years, to which I threw a caterpillar of Chærocampa Elpenor, did not give the insect any time for manœuvring, but killed it immediately by a strong blow with its bill. This bird had been tame for years, and was in the habit of pecking at everything thrown to him. Perhaps a wild jay (Garrulus Glandarius) would have treated the insect differently, but it is hardly possible that such a large and courageous bird would have much respect for our native caterpillars. I now turned to wild birds. A large brown Elpenor larva was placed in the food-trough of an open fowl-house from which the fowls had been removed. A flock of sparrows and chaffinches (Fringilla Domestica and Cœlebs) soon flew down from the neighbouring trees, and alighted near the trough to pick up stray food in their usual manner. One bird soon flew on to the edge of the trough, and was just about to hop into it when it caught sight of the caterpillar, and stood jerking its head from side to side, but did not venture to enter. Another bird soon came, and behaved in a precisely similar manner; then a third, and a fourth; others settled on the perch over the trough, and a flock of ten or twelve were finally perched around. They all stretched their heads and looked into the trough, but none flew into it.

I now made the reverse experiment, by removing the caterpillar and allowing the birds again to assemble, when they hopped briskly into the trough.

I often repeated this experiment, and always with the same result. Once it could be plainly seen that it was really fear and not mere curiosity that the birds showed towards the caterpillar. The latter was outside the trough amongst scattered grains of food, so that from one side it was concealed by the trough. A sparrow flew down obliquely from above, so that at first it could not see the caterpillar, close to which it alighted. The instant it caught sight of the insect, however, it turned in evident fright and flew away.

Of course these experiments do not prove that the larger insectivorous birds are also afraid of these caterpillars. Although I have not been able to experiment with such birds, I can certainly prove that even fowls have a strong dislike to these insects. I frequently placed a large Elpenor larva in the poultry yard, where it was soon discovered, and a fowl would run hastily towards it, but would draw back its head just when about to give a blow with the bill, as soon as it saw the caterpillar closely. The bird would now run round the larva irresolutely in a circle—the insect in the meantime assuming its terrifying attitude—and stretching out its head would make ten or twenty attempts to deal a blow with its bill, drawing back again each time. All the cocks and hens acted in a similar manner, and it was often five or ten minutes before one particularly courageous bird would give the first peck, which would soon be followed by a second and third, till the caterpillar, appearing palatable, would finally be swallowed.

These experiments were always made in the presence of several persons, in order to guard myself against too subjective an interpretation of the phenomena; but they all invariably considered the conduct of the birds to be as I have here represented it.[146]

If it be admitted that the ocelli of caterpillars are thus means of exciting terror, the difficulty of their occurring in protectively coloured species at once vanishes. They do not diminish the advantage of the adaptive colouring, because they do not make the caterpillars conspicuous, or at least any more easily visible at a distance, excepting when the insects have assumed their attitude of alarm. But these markings are of use when, in spite of protective colouring, the larva is attacked by an enemy. The eye-spots accordingly serve the caterpillar as a second means of defence, which is resorted to when the protective colouring has failed.

By this it must not be understood that the ocelli of the Chærocampa larvæ invariably possess only this, and no other significance for the life of the insect. Every pattern can be conceived to render its possessor in the highest degree conspicuous by strongly contrasted and brilliant colouring, so that it might be anticipated that perfect eye-spots in certain unpalatable species would lose their original meaning, and instead of serving for terrifying become mere signals of distastefulness. This is perhaps the case with Chærocampa Tersa ([Fig. 35]), the numerous eye-spots of which make the insect easily visible. Without experimenting on this point, however, no certain conclusion can be ventured upon, and it may be equally possible that in this case the variegated ocelli with bright red nuclei resemble the blossoms of the food-plant (Spermacoce Hyssopifolia).[147] I here mention this possibility only in order to show how an inherited form of marking, even when as well-defined and complicated as in the present case, may, under certain circumstances, be turned in quite another direction by natural selection, for the benefit of its possessor. Just in the same manner one and the same organ, such, for instance, as the limb of a crustacean, may, in the course of phyletic development, perform very different functions—first serving for locomotion, then for respiration, then for reproduction or oviposition, and finally for the acquisition of food.

I now proceed to the consideration of the biological value of incomplete eye-spots, or, as I have termed them, ring-spots. Are these also means of terrifying, or are they only signals of distastefulness?

I must at the outset acknowledge that on this point I am able to offer but a very undecided explanation. The decision is only to be arrived at by experiments conducted with each separate species upon which one desires to pronounce judgment. It is not here legitimate to draw analogical inferences, and to apply one case to all, since it is not only possible, but very probable, that the biological significance of ring-spots changes in different species. Nothing but a large series of experiments could completely establish this. Unfortunately I have hitherto failed in obtaining materials for this purpose. I would have deferred the publication of this essay for a year, could I have foreseen with certainty that such materials would have been forthcoming in sufficient quantity during the following summer; but this unfortunately depends very much upon chance, and I believed that a preliminary conclusion would be preferable to uncertainty. Perhaps some entomologist to whom materials are more easily accessible, may, by continuing these experiments, accomplish this object.

The experiments hitherto made by other observers, are not sufficient for deciding the question under consideration. Weir,[148] as is well known, showed that certain brightly coloured and conspicuous larvæ were refused by insectivorous birds; and Butler[149] proved the same for lizards and frogs. These experiments are unfortunately so briefly described, that in no case is the species experimented with mentioned by name, so that we do not know whether there were any Sphinx caterpillars among them.[150] I have likewise experimented in this direction with lizards, in order to convince myself of the truth of the statement that (1) there are caterpillars which are not eaten on account of their taste, and (2) that such larvæ possess bright colours. I obtained positive, and on the whole, very decided results. Thus, the common orange and blue striped caterpillars of Bombyx Neustria enjoyed complete immunity from the attacks of lizards, whilst those of the nearly allied Eriogaster Lanestris and L. Pini were devoured, although not exactly relished. That the hairiness is not the cause of their being unpalatable, is shown by the fact that L. Pini is much more hairy than B. Neustria. The very conspicuous yellow and black ringed caterpillar of Euchelia Jacobææ gave also most decided results. I frequently placed this insect in a cage with Lacerta Viridis, but they would never even notice them, and I often saw the caterpillars crawl over the body, or even the head of the lizards, without being snapped at. On every occasion the larvæ remained for several days with the lizards without one being ever missed. The reptiles behaved in a precisely similar manner with respect to the moth of E. Jacobææ, not one of which was ever touched by them. The yellow and black longitudinally striped caterpillars of Pygæra Bucephala were also avoided, and so were the brightly coloured larvæ of the large cabbage white (Pieris Brassicæ), which when crushed give a disagreeable odour. This last property clearly shows why lizards reject this species as distasteful. Both caterpillar and butterfly possess a blood of a strong yellow colour and oily consistency, in which, however, I could not detect such a decided smell as is emitted by that of the Heliconinæ and Danainæ.[151]

I next made the experiment of placing before a lizard a caterpillar as much as possible like that of E. Jacobææ. Half grown larvæ of Bombyx Rubi likewise possess golden yellow (but narrower) transverse rings on a dark ground, and they are much more hairy than those of E. Jacobææ. The lizard first applied its tongue to this caterpillar and then withdrew it, so that I believed it would also be avoided; nevertheless it was subsequently eaten. The caterpillars of Saturnia Carpini were similarly devoured in spite of their bristly hairs, and likewise cuspidate larvæ (Dicranura Vinula), notwithstanding their extraordinary appearance and their forked caudal horn.[152] These lizards were by no means epicures, but consumed large numbers of earth-worms, slugs, and great caterpillars, and once a specimen of the large and powerfully biting Orthopteron, Decticus Verucivorus. Creatures which possessed a strongly repugnant odour were, however, always rejected, this being the case with the strongly smelling beetle, Chrysomela Populi, as also with the stinking centipede, Iulus Terrestris, whilst the inodorous Lithobius Forficatus was greedily eaten. I will call particular attention to these last facts, because they favour the supposition that with rejected caterpillars a disgusting odour—although perhaps not always perceptible by us—is the cause of their being unpalatable.

Striking colours are of course only signals of distastefulness, and the experiment with Bombyx Rubi shows that the lizards were from the first prejudiced against such larvæ, the prejudice only being overcome on actually trying the specimen offered. A subsequent observation which I made after arriving at this conclusion, is most noteworthy. After the lizard had learnt by experience that there might be not only distasteful caterpillars (E. Jacobææ), but also palatable ones banded with black and yellow (B. Rubi), it sometimes tasted the Jacobææ larvæ, as if to convince itself that the insect was actually as it appeared to be, viz., unpalatable!

A striking appearance combined with a very perceptible and penetrating odour is occasionally to be met with, as in the caterpillar of the common Swallow-tail, Papilio Machaon. I have never seen a lizard make the slightest attempt to attack this species. I once placed two large specimens of this caterpillar in the lizard vivarium, where they remained for five days, and finally pupated unharmed on the side of the case.

I have recorded these experiments, although they do not thus far relate to Sphinx-caterpillars, with the markings of which we are here primarily concerned, because it appeared to me in the first place necessary to establish by my own experiments that signals of distastefulness did occur in caterpillars.

I now come to my unfortunately very meagre experience with Deilephila larvæ, with only two species of which have I been able to experiment, viz., D. Galii and Euphorbiæ.

The first of these was constantly rejected. Two large caterpillars, one of the black and the other of the yellow variety, were left for twelve hours in the lizard vivarium, without being either examined or touched. It thus appears that D. Galii is a distasteful morsel to lizards; and the habits of the caterpillar are quite in accordance with this, since it does not conceal itself, but rests fully exposed by day on a stem, so that it can scarcely escape being detected. It is almost as conspicuous as D. Euphorbiæ.

I was much surprised to find, however, that this last species was not rejected by lizards. On placing a large caterpillar, six to seven centimeters long, in the vivarium, the lizard immediately commenced to watch it, and as soon as it began to crawl about, seized it by the head, and, after shaking it violently, commenced to swallow it. In spite of its vigorous twisting and turning, the insect gradually began to disappear, amidst repeated shakings; and in less than five minutes was completely swallowed.[153] With regard to lizards, therefore, the prominent ring-spots of this larva are not effective as a means of alarm, nor are they considered as a sign of distastefulness.

Unfortunately I have not hitherto been able to make any experiments with birds. It would be rash to conclude from the experience with lizards that ring-spots were of no biological value. There is scarcely any one means of protection which can render its possessor secure against all its foes. The venom of the most poisonous snakes does not protect them from the attack of the secretary bird (Serpentarius Secretarius) and serpent eagle (Spilornis Cheela); and the adder, as is well known, is devoured by hedgehogs without hesitation. It must therefore be admitted that many species which are protected by distastefulness, may possess certain foes against which this quality is of no avail. Thus, it cannot be said that brightly coloured caterpillars, which are not eaten by birds and lizards, are also spared by ichneumons. It is readily conceivable therefore, that the larva of D. Euphorbiæ may not be unpalatable to lizards, because they swallow it whole; whilst it is perhaps distasteful to birds, because they must hack and tear in order to swallow it.

From these considerations it still appears most probable to me that D. Euphorbiæ, and the nearly allied D. Dahlii and Mauritanica, bear conspicuous ring-spots as signs of their being unpalatable to the majority of their foes. The fact that these species feed on poisonous Euphorbiaceæ, combined with their habit of exposing themselves openly by day, so as to be easily seen at a distance, may perhaps give support to this view. As these insects are not protectively coloured, this habit would long ago have led to their extermination; instead of this, however, we find that in all situations favourable to their conditions of life they are among the commonest of the Sphingidæ.

Thus, D. Euphorbiæ occurs in large numbers both in South and North Germany (Berlin); and Dr. Staudinger informs me that in Sardinia the larvæ of D. Dahlii were brought to him by baskets full.

But if the conspicuous ring-spots (combined of course with the other bright colours) may be regarded as signals of distastefulness in many species of Deilephila, this by no means excludes the possibility that in some species these markings play another part, and are effective as a means of alarm. It even appears conceivable to me that in one and the same caterpillar they may play both parts against different foes, and it would certainly be of interest to confirm or refute this supposition by experiment.

In the light yellow variety of the caterpillar of D. Galii the ring-spots may serve as means of alarm, and still more so in that of D. Nicæa, the resemblance of which to a snake has struck earlier observers.[154]

In those species of Deilephila which conceal themselves by day, the ring-spots cannot be considered as signals of distastefulness, and they must therefore have some other meaning. As examples of this class may be mentioned D. Vespertilio, which is protectively coloured both in the young and in the adult stages; and likewise D. Hippophaës, in which this habit of concealment is associated with adaptive colouring. In the case of the first-named species, it appears possible that the numerous large ring-spots may serve to alarm small foes, but the truth of this supposition could only be decided by experiment. In D. Hippophaës, on the other hand, such an interpretation must be at once rejected, since most individuals possess but a single ring-spot, which shows no resemblance whatever to an eye.

I long sought in vain for the meaning of this ring-spot, the discovery of which would in this particular case be of the greatest value, because we have here obviously the commencement of the whole development of ring-spots before us—the initial stage from which the marking of all the other species of Deilephila has proceeded.

I believe that I have now found the correct answer to this riddle, but unfortunately at a period of the year when I am unable to prove it experimentally. I consider that the ring-spots are crude imitations of the berries of the food-plant. The latter are orange-red, and exactly of the same colour as the spots; the agreement in colour between the latter and the berries is quite as close as that between the leaves and the general colouring of the caterpillar. I know of no species which more closely resembles the colour of the leaves of its food-plant, the dark upper side and light under side corresponding in the leaves and caterpillars. The colour of the Hippophae is not an ordinary green, but a grey-green, which shade also occurs, although certainly but rarely, in the larvæ. I may expressly state that I have repeatedly shown to people as many as six to eight of the large caterpillars on one buckthorn branch, without their being able at once to detect them. It is not therefore mere supposition, but a fact, that this species is protected by its general colouring. At first the orange-red spots appear rather to diminish this protection—at least when the insects are placed on young shoots bearing no berries. But since at the same time when the berries become red (end of July and the beginning of August) the caterpillars are in their last stage of development (i.e. possess red spots), it appears extremely probable that these spots are vague representations of the berries. For the same reason that these caterpillars have acquired the habit of feeding only at dusk and during the morning twilight, or at night, and of concealing themselves by day, it must be advantageous for them to have the surface of their large bodies not only divided by white stripes, but also interrupted in yet another manner. How could this be better effected than by two spots which, in colour and position, represent the grouping of the red berries on the branches? When feeding, the insect always rests with the hind segments on a branch, the front segments only being more or less raised and held parallel to the leaves; the red spots thus always appear on the stem, where the berries are likewise situated. It might indeed be almost supposed that the small progress which the formation of secondary ring-spots on the other segments has made up to the present time, is explicable by the fact that such berry-like spots on other portions of the caterpillar would be rather injurious than useful.

It may, however, be asked how an imitation of red berries, which are eaten by birds just as much as other berries, can be advantageous to a caterpillar, since by this means it would rather attract the attention of its enemies?

Two answers can be given to this. In the first place, the berries are so numerous on every plant that there is but a very small chance of the smaller and less conspicuous berry-spots catching the eye of a bird before the true berries; and, secondly, the latter, although beginning to turn red when the caterpillars are feeding, do not completely ripen till the autumn, when the leaves are shed, and the yellowish-red clusters of berries can be seen at a distance. The caterpillar, however, pupates long before this time.

I have considered this case in such detail because it appears to me of special importance. It is the only instance which teaches us that the rows of ring-spots of the Deilephila larvæ proceed from one original pair—the only instance which permits of the whole course of development being traced to its origin. Were it possible to arrive at the causes of the formation of these spots, their original or primary significance would thereby be made clear.

I will now briefly summarise the results of the investigation of the biological value of the Deilephila ring-spots.

In the known species of the genus now existing these spots have different meanings.

In some species (certainly in Galii, and probably in Euphorbiæ and Mauritanica) the conspicuous ring-spots serve as signals of distastefulness for certain enemies (not for all).

In a second group of species they serve as a means of alarm, like the eye-spots of the Chærocampa larvæ (Nicæa? light form of Galii?).

Finally, in a third group, of which I can at present only cite Hippophaës, they act as an adaptive resemblance to a portion of a plant, and enhance the efficacy of the protective colouring.

5. Subordinate Markings.—If, from the foregoing considerations, it appears that the three chief elements of the Sphinx-markings—longitudinal and oblique stripes, and spot formations—are not purely morphological characters, but have a very decided significance with respect to their possessors, there should be no difficulty in referring the whole of the markings of the Sphingidæ to the action of natural selection, supposing that these three kinds of marking were the only ones which actually occurred.

In various species, however, there appear other patterns, which I have comprised under the term “subordinate markings,” some of which I will select, for the purpose of showing the reasons which permit of their being thus designated.

I ascribe to this category, for example, that fine network of dark longitudinal streaks which often extends over the whole upper side of the caterpillar, and which is termed the “reticulation.” This character is found chiefly in the adult larvæ of Chærocampa, being most strongly pronounced in the brown varieties: it occurs also in Deilephila Vespertilio, Pterogon Œnotheræ, and Sphinx Convolvuli. As far as I know, it is only associated with adaptive colours, and indeed occurs only in those caterpillars which rest periodically at the base of their food-plants among the dead leaves and branches. I do not consider this reticulation to be a distinct imitation, but only as one of the various means of breaking up the large uniform surface of the caterpillar so as to make it present inequalities, and thus render it less conspicuous. There can be no doubt as to the dependence of this character upon natural selection.

There is, however, a second group of markings, which must be referred to another origin. To this group, for instance, belong those light dots in Chærocampa Porcellus and Elpenor which have been termed “dorsal spots.” I know of no other explanation for these than that they are the necessary results of other new formations, and depend on correlation (Darwin), or, as I may express it, they are the result of the action of the law governing the organization of these species.

As long as we are confined to the mere supposition that the character in question may be the outward expression of an innate law of growth, it is permissible to attempt to show that a quite similar formation in another species depends upon such a law.

Many of the dark specimens of Sphinx Convolvuli show whitish dots on segments six to eleven, one being situated on the front edge of each of these segments, at the height of the completely vanished subdorsal line ([Fig. 52]). These spots vary much in size, lightness, and sharpness of definition. Now it might be difficult to attribute any biological significance to this character, but its origin becomes clear on examining light specimens in which the oblique white stripes are distinct on the sides and the subdorsal line is retained at least on the five or six anterior segments. It can then be seen that the spots are located at the points of intersection of the subdorsal and the oblique stripes (Fig. 16, [Pl. III].), and they can accordingly be explained by the tendency to the deposition of light pigment being twice as great in these positions as in other portions of the two systems of light lines. Light spots are thus formed when the lines which cross at these points are partially or completely extinct throughout their remaining course.

A marking is therefore produced in this case by a purely innate law of growth—by the superposition of two ancient characters now rudimentary. Many other unimportant details of marking must be regarded as having been produced in a similar manner, although it may not be possible to prove this with respect to every minute spot and stripe. The majority of “subordinate markings” depend on the commingling of inherited, but now meaningless, characters with newly acquired ones.

It would be quite erroneous to attribute to natural selection only those characters which can be demonstrated to still possess a biological value in the species possessing them. They may be equally due to heredity. Thus, it is quite possible that the faint and inconspicuous ring-spots of Deilephila Vespertilio are now valueless to the life of the species—they may be derived from an ancestral form, and have not been eliminated by natural selection simply because they are harmless. I only mention this as a hypothetical case.

In the case of markings of the second class, i.e. oblique stripes, a transference to later phyletic stages can be demonstrated, although the stripes thereby lose their original biological value. Thus, the Chærocampa larvæ, when they were green throughout their whole life and adapted to the leaves, appear to have all possessed light oblique stripes in imitation of the leaf-ribs. All the species of the older type of colouring and marking, such as Chærocampa Syriaca ([Fig. 29]) and Darapsa Chœrilus ([Fig. 34]), and also the light green young forms of C. Elpenor ([Fig. 20]), and Porcellus ([Figs. 25 and 26]), show these oblique stripes. In these last species the foliage imitation is abandoned at a later stage, and a dark brown, or blackish-brown, ground-colour acquired. Nevertheless the oblique stripes do not disappear, but show themselves—in the fourth stage especially, and sometimes in the fifth—as distinct dirty yellow stripes, although not so sharply defined as in the earlier stages. These persistent stripes, in accordance with their small biological value, are very variable, since they are only useful in so far as they help to break up the large surface presented by the caterpillar, and are of no value as imitations of surrounding objects.

The oblique stripes of Sphinx Convolvuli offer a precisely similar case; and it may be safely predicted that the young forms of this species would possess sharply defined light oblique stripes, since more or less distinct remnants of these markings occur in all the adult larvæ, and especially in the green form. The entire pattern of this caterpillar depends essentially on the commingling of characters persisting from an earlier period, i.e. of residues of the subdorsal and oblique stripes, both these markings being extraordinarily variable. The black reticulation was added to the ground-colour as a new means of adaptation, this character appearing only in the phyletically younger brown form, and being entirely absent, or only faintly indicated, in the older green variety.

---

VI.
Objections to a Phyletic Vital Force.

It has been shown in the previous section that the three elements composing the markings of the Sphinx-larvæ originally possessed a distinct significance with respect to the life of the species, and that they were by this means called into existence. It has likewise been shown, that in most of the species which possess these characters at the present time they still have a decided, although sometimes a different use, for their possessors, so that from this point of view no objection can be raised to their being considered as having arisen by natural selection.

On looking at the phenomenon as a whole, however, certain instances occur which appear quite irreconcilable with this view.

The most formidable objection is offered by the genus Deilephila. The row of ring-spots which nearly all the existing species have more or less developed, has arisen from a simple subdorsal line. It would not, therefore, be surprising if a species were discovered which possessed this line without any ring-spots as its only marking. If D. Hippophaës were thus marked, there would be no objection to the theoretical assumption that this[155] was the ancestor of the other species. It would then be said that ring-spots were first developed in a later species by natural selection, and that they had been transmitted to all succeeding and younger species.

Certain individuals of D. Hippophaës, however, possess small ring-spots, some of which are well developed on several segments. In this species the row of ring-spots is therefore comprised in the development. The remaining species, which are much younger phyletically than Hippophaës, could not have inherited their ring-spots from the latter, since this species itself only possesses them occasionally, and, so to speak, in a tentative manner. The spots would therefore appear to have arisen spontaneously in this species, and independently of those in the other species. But if this were the case, how should we be able to prove that in the other species also the ring-spots did not arise independently; and if, moreover, a large number of species showed the same character without its being referable to inheritance from a common ancestor, how could this be otherwise explained than as the result of a force innate in these species and producing similar variations? But this is nothing but Askenasy’s “fixed direction of variation”—i.e., a phyletic vital force.

The only escape from this difficulty is perhaps to be found in proving that D. Hippophaës formerly possessed ring-spots, and that these have been subsequently either partially or completely lost, so that their occasional appearance in this species would therefore depend upon reversion. The ontogeny, however, teaches us that this is not the case, since the young caterpillar does not possess a greater number of more distinct ring-spots, but wants them altogether with the exception of a red spot on the eleventh segment, which is, however, much fainter than in the last stage.

This last-mentioned fact contains the solution of the problem. The premises from which this reasoning set out were all incorrect—the one red spot on the eleventh segment is likewise a ring-spot, and indeed the most important one of all, being primary, or the first to come into existence. Now all specimens, without exception, possess this first ring-spot, which is useful, and has therefore been called forth by natural selection; it is not inherited, but newly acquired by this species; at least, if the explanation of these spots which I have previously offered is correct.

The primary pair of spots may have been transferred from this to later species by heredity; and since, in all segmented animals there is a tendency for the peculiarities of one segment to be repeated on the others, this repetition must have occurred with greater frequency and more completely in the later species—the more so if the process were favoured by natural selection, i.e. if the row of ring-spots which originated in this manner could in any way be turned to the use of the species.

In Hippophaës itself there must also be a tendency to the formation of secondary ring-spots, and indeed in a number of specimens we actually see series of such ring-spots, the latter being present in varying numbers, and in very different states of development. The fact that the ring-spots have not become a constant and well-developed character, is simply explained by the circumstance that as such they would have endangered the existence of the species.

In this case there is therefore no necessity for assuming a phyletic vital force. The ring-spots of the genus Deilephila rather furnish us with an excellent explanation of a fact which might otherwise have been adduced in support of a phyletic vital force, viz., the strict uniformity in the development of larval markings.

Before I had been led to the discovery, by the study of the marking and development of Hippophaës, that the spots of the genus Deilephila originated on one segment only, from which they were transferred secondarily to the others, this astonishing regularity appeared to me an incomprehensible problem, which could only be solved by assuming a phyletic vital force. If it be attempted, for the ten species here considered, to construct a genealogical tree based on the supposition that it is the rows of spots which have been inherited in cases where they occur, and not the mere tendency to their production by the transference of the one originally inherited primary spot to the remaining segments, the attempt will fail. The greater number of the species would have to be arranged in one row, since one species always bears a perfected form of marking, which appears in the young stages of the following species. But it is very improbable that nine different species, derived directly the one from the other, would contemporaneously survive.[156] One species, D. Vespertilio, could not be inserted at all in the genealogical tree, since it wants one character which occurs in all the other species, viz., the caudal horn, which is absent even in the third stage, and must therefore have been lost at a very early period of the phyletic development, so that we may consider it to be on this account genetically allied to the oldest known form. But the markings of this larva pass through precisely the same stages of development as do those of the other species. Now if the ring-spots were inherited as such, the existence of a hornless species with ring-spots would be an insoluble riddle, and would favour the admission of parallel developmental series, which again could be scarcely otherwise explained than by a “fixed direction of variation.” We have here one of that class of cases which the supporters of a phyletic vital force have already so often made use of in support of their view.

The explanation of such a case—i.e. its reference to known causes of species transformation—is never easy, and is indeed impossible without a precise knowledge of the ontogeny of many species, as well as of the original significance of the characters in question. In the case of the Deilephila larvæ, however, such knowledge is still wanting. It is true that they present us with parallel developmental series, but these do not depend on an unknown phyletic force—the parallelism can be referred to the action of the imperfectly known laws of growth innate in segmented organisms. Because the characters of one segment have a tendency to repeat themselves on the others, from one parent-form possessing ring-spots on one segment only, there may have proceeded several developmental series, all of which developed rows of such spots independently of each other.

From these considerations we may venture to construct the following genealogical tree:—

Possible genealogical tree of the Genus Deilephila

The circles indicate the phyletic stages IV.-VIII.; the eighth is only reached by Nicæa, and is distinguished from the seventh chiefly by the ontogeny, in the third stage of which the seventh phyletic stage is reached, whilst in Euphorbiæ and Dahlii this stage is reached in the fourth ontogenetic stage. The phyletic stages indicated by queries are extinct, and only known through the ontogeny of existing species. It must be understood that this pedigree expresses only the ideal and not the actual relations of the species to one another. Thus, it is possible that Hippophaës is not the parent-form, but an unknown or extinct species, which must, however, have possessed the same marking, and so on.

Four parallel series here proceed from the parent-form Hippophaës; there may have been five, or possibly only three, but the incomplete state of our knowledge of the ontogeny does not permit of any certain conclusion. For the point under consideration this is, however, quite immaterial. The distance from the central point (the parent-form) indicates the grade of phyletic development which the respective species have at present reached.

There is another case which is no less instructive, because it reveals, although in a somewhat different manner, the action of a law of growth innate in the organism itself, but which can nevertheless by no means be regarded as equivalent to a phyletic vital force. I refer to the coloured edges of the oblique stripes which occur in most of the species of the genus Sphinx. It has already been insisted upon in a previous section, that the mode in which this character originates negatives the assumption of a phyletic force, because these coloured edges are gradually built up out of irregularly scattered spots. There is no occasion for a “developmental force” to grope in the dark; if such a power exists, we should expect that it would add new characters to old ones with the precision of a master workman.

If, however, the coloured edges certainly depend on natural selection, this agency causing the scattered spots to coalesce and become linear, we have here the proof that such spots first arose in a precisely similar manner in several species, quite independently of one another—that, in fact, a “fixed direction of variation” in a certain sense exists.

In three species of Smerinthus-larvæ, red spots appear towards the end of the ontogeny; in S. Populi and Ocellatus in only a minority of individuals, and always separate (not coalescent), and in S. Tiliæ in a majority of specimens, the spots frequently becoming fused into one large, single, longish marking. These three species cannot have inherited the spots from a common ancestor, since they are absent in the younger ontogenetic stages, or occur only exceptionally, becoming larger and more numerous in the last stage; they obviously form a character which must be considered as a case of “anticipated development.”

How is it then that three species vary independently of each other in an analogous manner? I know of no other answer to this question than that similar variations must necessarily arise from similar physical constitutions—or, otherwise expressed, the three species have inherited from an unknown parent species, devoid of spots, not this last character itself, but a physical constitution, having a tendency to the formation of red spots on the skin.[157] The case offers many analogies to that of the colour varieties of Lacerta Muralis, to which Eimer[158] briefly calls attention in his interesting communications on the blue lizard of the Faraglioni Rocks at Capri. The South Italian lizards, although having differently formed skulls, show the same brilliantly coloured varieties as those of North Italy; and Eimer believes that these parallel variations in widely separated localities, some of which have long been isolated, must be referred to a tendency towards fixed directions of variation innate in the constitution of the species.

I long ago insisted[159] that it should not be forgotten that natural selection is, in the first place, dependent upon the variations which an organism offers to this agency, and that, although the number of possible variations may be very great for each species, yet this number is by no means to be considered as literally infinite. For every species there may be impossible variations. For this reason I am of opinion that the physical nature of each species is of no less importance in the production of new characters than natural selection, which must always, in the first place, operate upon the results of this physical nature, i.e. upon the variations presented, and can thus call new ones into existence.

It requires but a slight alteration of the definition to make out of this “restricted” or “limited variability,” which is the necessary consequence of the physical nature of each species, a “fixed direction of variation” in the sense of a phyletic vital force. Instead of—the Smerinthus-larvæ show a tendency to produce red spots on the skin, it is only necessary to say—these larvæ tend to produce red borders to the oblique stripes. The latter statement would, however, be incorrect, since the red borders first arose by the coalescence of red spots through the action of natural selection. It is not even correct to say that all the species of Smerinthus show this tendency to produce spots, since this character does not seem to occur either in S. Quercus or S. Tremulæ.

The distinction between the two modes of conception will become clear if we ask, as an example, whether those Chærocampa-larvæ which do not at present possess eye-spots will subsequently acquire these markings, supposing that they maintain their existence on the earth for a sufficient period?

The supporters of a “fixed direction of variation” would answer this question in the affirmative. Ocelli constitute a character which occurs in nearly all the species of the group—they are the goal towards which the phyletic force is urging, and which must sooner or later be reached by each member of the group. On the other hand, I cannot express so decidedly my own opinion, viz., that such complicated characters as the many-coloured oblique stripes or eye-spots are never the results of purely internal forces, but always arise by the action of natural selection, i.e. by the combination of such minute and simple variations as may present themselves. It may be replied that the formation of eye-spots in those species which are at present devoid of them, cannot indeed be considered impossible, but that they would only appear if the constitution of these species had a tendency to give rise to the production of darker spots on the edge of the subdorsal line, and if at the same time, the possession of eye-spots would be of use to the caterpillar under its special conditions of life.

The condition of affairs would be quite different if we were simply concerned with the transference of a character from one segment on which it was already present, to the remaining segments. The transference would, in this case, result from causes purely innate in the organism—from the action of laws of equilibration or of growth (correlation), and the external conditions of life would play only a negative part, since they might prevent the complete reproduction of a character, such, for example, as eye-spots, on all the segments, in cases where it was disadvantageous to the species. The fact that our species of Chærocampa have only faint indications, and not a completely-developed eye-spot, on the remaining segments, may perhaps be explained in this manner. It is conceivable that the two pairs of ocelli on the front segments are more effective as a means of alarm than if the insects were provided with two long rows of such markings; but nothing can be stated with certainty on this point until experiments have been made with caterpillars having rows of eye-spots.

The question raised above—whether the species of Chærocampa at present devoid of eye-spots are to be expected to acquire this character in the course of their further phyletic development—brings with it another point, which cannot be here passed over.

If the utility of the four kinds of markings in their perfected form is demonstrated, their origination through natural selection is not, strictly speaking, thereby proved. It must also be shown that the first rudiments of these characters were also of use to their possessors. The question as to the utility of the “initial stages” of useful characters must here be set at rest.

In the case of markings such as longitudinal and oblique stripes, it is quite evident that the initial stages of these simple characters do not differ greatly from the perfected marking, but this is certainly not the case with eye- and ring-spots. The most light is thrown upon this question by the latter, because a species which has remained at the initial stage of the formation of ring-spots here presents itself for examination, viz. Deilephila Hippophaës.

I have attempted to show that the orange-red spots, which, as a rule, adorn only the eleventh segment, enhance the adaptive colouring of this caterpillar by their resemblance to the berries of the sea-buckthorn, whilst the general surface resembles the leaves in colour. If this be admitted, the origination of these spots by natural selection offers no difficulty, since a smaller spot, or one of a fainter red, must also be of some use to its possessor.

This case is of importance, as showing that a “change of function” may occur in markings, just as it does in certain organs among the most diverse species of animals, in the course of phyletic development. The spots which in Hippophaës are imitations of red berries, in species which have further advanced phyletically play quite another part—they serve as means of alarm, or signals of distastefulness.

It appears to me very improbable, however, that the perfect ocelli of the Chærocampa-larvæ have also undergone such a “functional change” (Dohrn). I rather believe that the first rudiments of these markings produced the same effect as that which they now exercise, viz., terror. We are certainly not so favourably circumstanced in this case in knowing a species which shows the initial steps of this character in its last stage of life; but in the initial steps which the second stage of certain species present, we see preserved the form under which the eye-spots first appeared in the phylogeny, and from this we are enabled to judge with some certainty of the effect which they must have produced at the time.

In the ontogeny of C. Elpenor and Porcellus we see that a small curvature of the subdorsal line first arises, the concavity of which becomes filled with darker green, and soon afterwards with black; the upwardly curved piece of the subdorsal then becomes detached and more completely surrounded by black. The white fragment of the subdorsal which has become separated, in the next place broadens, and a black (dark) pupil appears in its centre.

Now the first rudiments of the eye-spot certainly appear very insignificant in a caterpillar two centimeters long, but we must not forget that in the ancestors of the existing Chærocampa-larvæ this character appeared in the adult state. If we conceive the curvature of the white subdorsal with the underlying dark pigment to be correspondingly magnified, its importance as a means of alarm can scarcely be denied, particularly when we consider that this marking stands on the enlarged fourth segment, which alone invests the caterpillar with a singular, and, to smaller foes, an alarming appearance. We know that in the case of those Chærocampa-larvæ which possess no eye-spots, the distension of this segment is employed against hostile attacks. (See the illustration of Darapsa Chærilus, [Pl. IV]., Fig. 34.) Those markings which even only remotely resembled an eye must, in such a position, have increased the terrifying action. On these grounds I believe that it may be safely admitted, that this kind of marking possessed the same significance in its initial stages as it now does when fully perfected. No functional change has here taken place.

Among all the facts brought together in the first section I only know of one group of phenomena which at least permit of an attempt to refer them to a phyletic vital force. This is the occurrence of dark ground-colours in adult larvæ which are of light colours in their young condition. I have already attempted to show that in the Chærocampa-larvæ this change of colour depends on a double adaptation, the young caterpillars being adapted to the green colour of the plant and the adults to the soil and dead leaves. This interpretation appears the more correct when we find the same process, viz. the gradual replacement of the original green by brown colours, among species of widely different genera, which, with the dark colouring, possess the necessarily correlated habit of hiding themselves by day when in the adult condition. This is the case with Sphinx Convolvuli, Deilephila Vespertilio, and Acherontia Atropos.

Thus far all has been easily explicable by natural selection; but when we also see a “tendency” to acquire a dark colour in the course of development, in those species which neither conceal themselves nor are adaptively coloured, but are very conspicuously marked—and if, further, it can be shown that these species, such for instance as Deilephila Galii, actually possess immunity from the attacks of foes,—how can this tendency to the formation of a dark colour be otherwise explained than by the admission of a phyletic vital force urging the variations in this direction?

Nevertheless I believe that also on this point an appeal to unknown forces can be dispensed with. In the first place, dark ground-colours can be of use to a species otherwise than as means of adaptation. In D. Galii, as well as in D. Euphorbiæ, the light ring-spots appear rather at their brightest on the pitchy-black ground; and if this caterpillar must (sit venia verbo!) become conspicuous, this purpose would be best attained by acquiring a dark ground-colour, such as that of D. Euphorbiæ.

The tendency, apparently common to all these Sphingidæ, to acquire a dark colour with increasing age, depends therefore on two quite distinct adaptations—first, in species sought by enemies, on an adaptation to the colour of the soil; and secondly, in species rejected by foes, on the endeavour to produce the greatest possible contrast of colour.

Moreover, the supposition from which this last plea for a vital force set out is not universally correct, since there are species, such for instance as D. Nicæa, which never acquire a dark colour; and in D. Galii also, although all the individuals abandon the protective green of the young stages, they by no means all acquire a dark hue in exchange for this colour; many individuals in their light ochreous-yellow colouring rather strikingly resemble the snake-like caterpillar of D. Nicæa.

---

VII.
Phyletic Development of the Markings of the Sphingidæ: Summary and Conclusion.

If, from the form possessed by many of the caterpillars of the Sphingidæ on their emergence from the egg, we may venture to draw a conclusion concerning the oldest phyletic stage, these larvæ were originally completely destitute of marking. The characteristic caudal horn must be older than the existing markings, since it is present in the younger stages (except in cases where it is altogether wanting), and is generally even larger than at a later age.

There is, however, further evidence that there were once Sphinx-larvæ without any markings. Such a species now exists. I do not mean the boring caterpillars of the Sesiidæ,[160] which live in the dark, and are therefore colourless, but I refer to a large larva (over six centimeters long) preserved in spirit in the Berlin Museum,[161] which, from its form, belongs to the Smerinthus group. It possesses a caudal horn, and on the whole upper surface is covered with short and sparsely scattered bristles, such as occur in the Sesiidæ. The colour of this unknown insect appears to have been light green, although it now shows only a yellowish shade. Every trace of marking is absent, and it thus corresponds exactly with the youngest stages of the majority of the existing Sphinx-larvæ—even to the short bristles sparsely scattered over the whole upper surface of its body. We have therefore, so to speak, a living fossil before us, and it would be of great interest to ascertain its history.

All the data furnished by the developmental history go to show that of the three kinds of markings which occur in the Sphingidæ, viz., longitudinal and oblique stripes and spots, the first is the oldest. Among the species which are ornamented with oblique stripes or spots there are many which are longitudinally striped in their young stages, but the reverse case never occurs—young larvæ never show spots or oblique stripes when the adult is only striped longitudinally.

The first and oldest marking of the caterpillars of the Sphingidæ was therefore the longitudinal striping, or, more precisely speaking, the subdorsal, to which dorsal and spiracular lines may have been added. That this second stage of phyletic development has also been preserved in existing species has already been sufficiently shown; the greater portion of one group, the Macroglossinæ, has indeed remained at this stage of development.

From the biological value which must be attributed to this kind of marking, its origination by natural selection presents no difficulty. The first rudiments of striping must have been useful, since they must have broken up the large surface of the body of the caterpillar into several portions, and would thus have rendered it less conspicuous to its enemies.

Thus it is not difficult to perceive how a whole group of genera could have made shift with this low grade of marking up to the present time. Colour and marking are not the only means of offence and defence possessed by these insects; and it is just such simply-marked larvæ as those of the Macroglossinæ which have the protective habit of feeding only at night, and of concealing themselves by day. Moreover, under certain conditions of life the longitudinal stripes may be a better means of protection, even for a Sphinx-larva, than any other marking; and all those species in which this pattern is retained at the present time live either among grasses or on Coniferæ.

It cannot be properly said that the second form of marking—the oblique stripes—has been developed out of the first. If these had arisen by the transformation of the longitudinal stripes, the two forms could not exist side by side. This is the case, however, both in certain species in the adult state (Calymnia Panopus[162]), as well as in others during their young stages (most beautifully seen in Smerinthus Populi, [Fig. 56]). Various facts tend to show that the oblique stripes appeared in the phyletic development later than the longitudinal lines. In the first place they appear later than the latter in the ontogeny of certain species. This is the case with Chærocampa Elpenor and Porcellus, in which, however, they certainly do not reach a high state of development. Then again, the longitudinal lines disappear completely in the course of the ontogeny, whilst the oblique stripes alone maintain their ground. Thus, the subdorsal line vanishes at a very early stage, with the exception of a small residue,[163] in all native species of Smerinthus. I have already attempted to show that new characters are only acquired in the last stage, and that if still newer ones are then added, the former disappear from the last stage, and are transferred back to a younger one. Characters vanish therefore from a stage in the same order as they were acquired.

Finally, among the genera with longitudinal stripes (e.g. Macroglossa) we know certain species which, when at an advanced age, possess oblique stripes (M. Fuciformis), although these slant in a direction opposite to those of most of the other larvæ of the Sphingidæ. These are, however, always species which differ from their allies in their mode of life, not feeding on grasses or low plants, but on large-leaved shrubs. If we were able to ascertain the ontogeny of these species, we should find that the oblique stripes appeared late in life, as has already been shown in the case of Pterogon Œnotheræ.

If it be asked why the longitudinal lines were first formed, and then the oblique stripes, it may be replied that the physical constitution of these caterpillars would be more easily able to give rise to simple longitudinal lines than to complicated oblique stripes crossing their segments.[164] It may perhaps also be suggested that the oldest Sphingidæ lived entirely on low plants among grasses, and in the course of time gradually took to shrubs and trees. At the present time the majority of the Sphinx-larvæ still live on low plants, and but few on trees, such caterpillars generally belonging to certain special genera.

The character of oblique stripes becomes perfected by the addition of coloured edges, the latter, as is self-evident, having been added subsesequently.

The third chief constituent of the Sphinx-markings, i.e. the spots—whether perfect ocelli or only ring-spots—in two of the special genera here considered, arise on the subdorsal, where they are either deposited (Deilephila), or built up from a fragment of this line (Chærocampa). That these markings can, however, also originate independently of the subdorsal, is shown by the ocellus of Pterogon Œnotheræ, situated on the segment bearing the caudal horn. In this case, however, the ontogeny teaches us that the spot also succeeds the subdorsal, so that we can state generally that all these spot-markings are of later origin than the longitudinal striping.

The question as to the relative ages of the oblique stripes and the spot-marking does not admit of a general answer. In some cases (C. Elpenor and Porcellus) the oblique stripes disappear when the ocelli reach complete development, and we may therefore venture to conclude that in these cases the former appeared earlier in the phylogeny. But it is very probable that oblique stripes arose independently at different periods, just as longitudinal lines occur irregularly in quite distinct families. It would be a great error if we were to ascribe the possession of oblique stripes solely to descent from a common ancestor. The oblique markings found on certain species of Macroglossa (M. Corythus from India) have not been inherited from a remote period, but have been independently acquired by this or by some recent ancestral species. They have nothing to do genetically with the oblique stripes which occur in some species of Chærocampa (e.g. in C. Nessus, from India), or with those of the species of Smerinthus and Sphinx. They depend simply on analogous adaptation (Seidlitz[165]), i.e. on adaptation to an analogous environment.

The case is similar with the spot-markings. I have already shown that under certain conditions ring-spots may assume the exact appearance of eye-spots by the formation of a nucleus in the “mirror,” such as occurs occasionally in Deilephila Euphorbiæ ([Fig. 43]), more frequently in D. Galii, and as a rule in D. Vespertilio. Nevertheless, these markings arise in quite another manner to the eye-spots of the Chærocampinæ, with which they consequently have no genetic relation; the two genera became separated at a time when they neither possessed spot-markings. Further, in Pterogon Œnotheræ we find a third kind of spot-marking, which is most closely allied to the ocelli of the Chærocampa-larvæ, but is situated in quite another position, and must have originated in another manner, and consequently quite independently of these eye-spots.

It can also be readily understood why the first and second elements of the markings of the Sphingidæ should be mutually exclusive, and not the second and third or the first and third.

A light longitudinal line cutting the oblique stripes, considerably diminishes that resemblance to a leaf towards which the latter have a tendency, and it is therefore only found in cases where an adaptive marking can be of no effect on account of the small size of the caterpillar, i.e. in quite young stages. (See, for instance, [Fig. 56], the first stage of S. Populi.) At a later period of life the old marking must give way to the new, and we accordingly find that the subdorsal line vanishes from all the segments on which oblique stripes are situated, and is only retained on the anterior segments where the latter are wanting. In some few cases both elements of marking certainly occur together, such as in Calymnia Panopus and Macroglossa Corythus; but the oblique stripes are, under these circumstances, shorter, and do not extend above the subdorsal line, and in Darapsa Chœrilus even become fused into the latter.[166]

In certain cases there may also be a special leaf structure imitated by the longitudinal lines, but on the whole the latter diminish the effect of the oblique stripes; and we accordingly find that not only has the subdorsal disappeared from those segments with oblique stripes, but that most larvæ with this last character are also without the otherwise broad spiracular and dorsal lines. This is the case with all the species of Smerinthus[167] known to me, as well as with all the species of the genera Sphinx, Dolba, and Acherontia.

Oblique stripes and spot-markings are not, however, necessarily mutually exclusive in their action, and we also find these in certain cases united in the same larva, although certainly never in an equal state of perfection. Thus, Chærocampa Nessus[168] possesses strongly marked oblique stripes, but feebly developed ocelli; and, on the other hand, Chærocampa Elpenor shows strongly developed eye-spots, but the earlier oblique stripes are at most only present as faint traces. This is easily explained by the mode of life. These caterpillars—at least such of them as are perfectly known—do not live on plants with large, strongly-ribbed leaves, and are even in the majority of individuals adapted to the colour of the soil; the oblique stripes have therefore in these cases only the significance of rudimentary formations.

That the first and third forms of markings also are not always mutually prejudicial in their action is shown by the case of Chærocampa Tersa, in which the eye-spots certainly appear to possess some other significance than as a means of causing terror. In most of the Chærocampa-larvæ the subdorsal line disappears in the course of the phylogeny, and it can be understood that the illusive appearance of the eye-spots would be more perfect if they did not stand upon a white line.

If we consider the small number of facts with which I have here been able to deal, the result of these investigations will not be deemed unsatisfactory. It has been possible to show that each of the three chief elements of the markings of the Sphingidæ have a biological significance, and their origin by means of natural selection has thus been made to appear probable. It has further been possible to show that the first rudiments of these markings must also have been of use; and it thus appears to me that their origin by means of natural selection has been proved to demonstration. Moreover, it has not been difficult to understand the displacement of the primary elements of the markings by secondary characters added at a later period, as likewise an essential effect of natural selection. Finally, it has been possible to explain also the subordinate or accessory elements of the markings, partly by the action of natural selection, and partly as the result of markings formerly present acting by correlation.

From the origin and gradual evolution of the markings of the Sphingidæ we may accordingly sketch the following picture:—

The oldest Sphinx-larvæ were without markings; they were probably protected only by adaptive colouring, and a large caudal horn, and by being armed with short bristles.

Their successors, through natural selection, became longitudinally striped; they acquired a subdorsal line extending from the horn to the head, as well as a spiracular, and sometimes also a dorsal, line. The caterpillars thus marked must have been best hidden on those plants in which an arrangement of parallel linear parts predominated; and we may venture to suppose that at this period most of the larvæ of the Sphingidæ lived on or among such plants (grasses).

At a later period oblique stripes were added to the longitudinal lines, the former (almost always) slanting across the seven hindmost segments from the back towards the feet in the direction of the caudal horn. Whether these stripes all arose simultaneously, or, as is more probable, whether only one at first appeared, which was then transferred to the other segments by correlation assisted by natural selection, cannot, at least from the facts available, at present be determined.

On the whole, as the oblique stripes became lengthened towards the back, the longitudinal lines disappeared, since they injured the deceptive effect of the stripes. In many species also there were formed dark or variegated coloured edges to the oblique stripes, in imitation of the shadow lines cast by the leaf-ribs.

Whilst one group of Sphingidæ (Sphinx, Smerinthus) were thus striving to make their external appearance approximate more and more to that of a ribbed leaf, others of the longitudinally striped species became developed in another manner.

Some of the latter lived indeed on bush-like leaved plants, but no oblique stripes were developed, because these would have been useless among the dense, narrow, and feebly-ribbed leaves of the food-plants. These caterpillars, from the earlier markings, simply retained the longitudinal lines, which, combined with a very close resemblance to the colour of the leaves, afforded them a high degree of protection against discovery. This protection would also have been enhanced if other parts of the food-plant, such as the berries (Hippophaës), were imitated in colour and position in such a manner that the large body of the caterpillar contrasted still less with its environment. In this way the first ring-spot probably arose in some species on only one—the penultimate segment.

As soon as this first pair of ring-spots had become an established character of the species, they had a tendency to become repeated on the other segments, advancing from the hind segments towards the front ones. Under certain conditions this repetition of the ring-spots might have been of great disadvantage to the species, and would therefore have been as far as possible prevented by natural selection (Hippophaës); in other cases, however, no disadvantage would have resulted—the caterpillar, well adapted to the colour of its food-plant, would not have been made more conspicuous by the small ring-spots, which might thus have become repeated on all the segments (Zygophylli). In cases like the two latter, striking colours must have been eliminated when inherited from an immediate ancestor; but on this point nothing can as yet be said with certainty.

In other cases the repetition of the ring-spots with strongly contrasted colours was neither prejudicial nor indifferent, but could be turned to the further advantage of the species. If a caterpillar fed on plants containing acrid juices (Euphorbiaceæ) which, by permeating its alimentary system, rendered it repulsive to other animals, the ring-spots commencing to appear (by repetition) would furnish an easy means for natural selection to adorn the species with brilliant colours, which would protect it from attack by acting as signals of distastefulness.

But if the dark spots stood on a light ground (Nicæa), they would present the appearance of eyes, and cause their possessors to appear alarming to smaller foes.

From the developmental histories and biological data at present before us, it cannot with certainty be said which of these two functions of the ring-spots was first acquired in the phylogeny, but we may perhaps suppose that their significance as a means of causing alarm was arrived at finally.

It may also be easily conceived that as the ring-spots became more and more complicated, they would occasionally have played other parts, being fashioned once again in these stages into imitations of portions of plants, such as a row of berries or flower-buds. For this, however, there is as yet no positive evidence.

As the ring-spots became detached from the subdorsal line out of which they had arisen, the latter disappeared more and more completely from the last ontogenetic stage, and receded towards the younger stages of life of the caterpillar—it became historical. This disappearance of the subdorsal may also be explained by the fact that the original longitudinal stripe imitating the linear arrangement of leaves would become meaningless, even if it did not always diminish the effect of the ring-spots. But characters which have become worthless are known in the course of time to become rudimentary, and finally to disappear altogether. I do not believe that disuse alone causes such characters to vanish, although in the case of active organs it may have a large share in this suppression. With markings it cannot, however, be a question of use or disuse—nevertheless they gradually disappear as soon as they become meaningless. I consider this to be the effect of the arrest of the controlling action of natural selection upon these characters (suspension of the so-called “conservative adaptation,” Seidlitz). Any variations may become of value if the character concerned is met with in the necessary state of fluctuation. That this process of extinction does not proceed rapidly, but rather with extreme slowness, is seen in the ontogeny of several species of Deilephila, which retain the now meaningless subdorsal line through a whole series of stages of life.

In another group of Sphinx-larvæ with longitudinal stripes, an eye-spot became developed independently of the subdorsal line, in the position of the caudal horn, which has here vanished with the exception of a small knob-like swelling. This character—which we now see perfected in Pterogon Œnotheræ—undoubtedly serves as a means of causing terror; but whether the incipient stages possessed the same significance, cannot be decided in the isolated case offered by the one species of the genus Pterogon possessing this marking.

In a third group of longitudinally striped caterpillars, the younger genus Chærocampa, eye-spots were developed directly from portions of the subdorsal line, at first only on the fourth and fifth segments. It can be here positively asserted that this character served as a means of alarm from its very commencement. It is certainly for this reason that we see the subdorsal line in the immediate neighbourhood of the spots disappear at an early stage, whilst it is retained on the other segments for a longer period. A portion of the younger (tropical) species of this group then developed similar, or nearly similar, ocelli on the remaining segments by correlation; and it may now have occurred that in solitary cases the eye-spots acquired another significance (C. Tersa?), becoming of use as a disguise by resembling berries or flower-buds. It is also conceivable that the eye-spots may in other cases have been converted into a warning sign of distastefulness.

In all those larvæ which possessed purely terrifying markings, however, not only was the original protective colouring preserved, but in most of them this colour gradually became replaced by a better one (adaptation of the adult larva to the soil). The oblique stripes imitating the leaf-ribs also are by no means lost, but are almost always present, although but feebly developed, and often only temporarily.

The pattern formed by the oblique stripes may also be retained, even with perfect adaptation to the soil, and may be converted to a new use by losing its sharpness, and, instead of imitating definite parts of plants, may become transformed into an irregular and confused marking, and thus best serve to represent the complicated lights and shadows, stripes, spots, &c., cast on the ground under low-growing plants from between the stems and dead leaves.

Just as in the case of ocellated species where caterpillars without eye-spots may retain and newly utilize their older markings, so larvæ having oblique stripes with the most diversely coloured edges may show the same markings in allied (younger?) species, both in a rudimentary and in a transformed condition. These markings may thus contribute to the formation of a latticed or reticulated pattern. Even the oldest marking, the subdorsal line, may still play a part, since its remnants cause certain portions of the complicated pattern to appear more strongly marked (S. Convolvuli). Finally, when an adaptation to a changing environment intersected by lights and shadows is required, new markings may be here added as in other cases, viz., dark streaks extending over the light surface of the whole caterpillar.

In concluding this essay, I may remark that, with respect to the wide and generally important question which gave rise to these investigations, a clearer and simpler result has been obtained than could have been expected, considering the complexity of the characters requiring to be traced to their causes, as well as our still highly imperfect knowledge of ontogenetic and biological facts.

For a long time I believed that it was not possible to trace all the forms of marking and their combinations to those causes which are known to produce transformation; I expected that there would be an inexplicable residue.

But this is not the case. Although it cannot yet be stated at first sight with certainty in every single instance how far any particular element of marking may have a biological value in the species possessing it, nevertheless it has been established that each of the elements of marking occurring in the larvæ of the Sphingidæ originally possessed a decided biological significance, which was produced by natural selection.

In the case of the three chief elements of the markings of the Sphingidæ, it can be further shown that not only the initial stages but also their ultimate perfection—the highest stages of their development, are of decided advantage to their possessors, and have a distinct biological value, so that the gradual development and improvement of these characters can be traced to the action of natural selection.

But although natural selection is the factor which has called into existence and perfected the three chief forms and certain of the subsidiary markings, in the repetition of the local character on the other segments, as well as in the formation of new elements of marking at the points of intersection of older characters now rudimentary, we can recognize a second factor which must be entirely innate in the organism, and which governs the uniformity of the bodily structure in such a manner that no part can become changed without exerting a certain action on the other parts—an innate law of growth (Darwin’s “correlation”).

Only once during the whole course of the investigations was it for an instant doubtful whether a phyletic vital force did not make itself apparent, viz., in the red spots accompanying the oblique stripes in several Smerinthus-larvæ. Closer analysis, however, enabled us to perceive most distinctly the wide gulf that separates “analogous variation” from the mystic phyletic vital force. Nothing further remains therefore for the action of this force in respect to the marking and colouring of the Sphingidæ, since several even of the subordinate markings can be traced to their causes, only the “dorsal spots” of our two native species of Chærocampa having been referred to correlation without decided proof. From the temporary inability to explain satisfactorily such an insignificant detail, no one will, however, infer the existence of such a cumbrous power as a phyletic vital force.

The final result to which these investigations have led us is therefore the following:—The action of a phyletic vital force cannot be recognized in the marking and colouring of the Sphingidæ; the origination and perfection of these characters depend entirely on the known factors of natural selection and correlation.

---

II.
ON PHYLETIC PARALLELISM IN METAMORPHIC SPECIES.

INTRODUCTION.

In the previous essay I attempted to trace a whole group of apparently “purely morphological” characters to the action of known factors of transformation, to explain them completely by these factors, and in this manner I endeavoured to exclude the operation of an internal power inciting change (phyletic vital force).

In this second study I have attempted to solve the problem as to whether such an innate inciting power can be shown to exist by comparing the forms of the two chief stages of metamorphic species, or whether such a force can be dispensed with.

Nobody has as yet apparently entertained the idea of testing this question by those species which appear in the two forms of larva and imago (insects), or, expressed in more general terms, by those species the individuals of which successively possess quite different forms (metamorphosis), or in which the different forms that occur are distributed among different individuals alternating with and proceeding from one another (alternation of generation). Nevertheless, it is precisely here that quite distinct form-relationships would be expected according as the development of the organic world depended on a phyletic vital force, or was simply the response of the specific organism to the action of the environment.

Assuming the first to be the case, there must have occurred, and must still occur, what I designate “phyletic parallelism,” i.e. the two stages of metamorphic species must have undergone a precisely parallel development—every change in the butterfly must have been accompanied or followed by a change in the caterpillar, and the systematic groups of the butterflies must be also found in a precisely corresponding manner in a systematic grouping of the caterpillars. If species are able to fashion themselves into new forms by an innate power causing periodic change, this re-moulding cannot possibly affect only one single stage of development—such as the larva only—but would rather extend, either contemporaneously or successively, to all stages—larva, pupa, and imago: each stage would acquire a new form, and it might even be expected that each would change to the same extent. At least, it cannot be perceived why a purely internal force should influence the development of one stage more than that of another. The larvæ and imagines of two species must differ from one another to the same extent, and the same must hold good for the larvæ and imagines of two genera, families, and so forth. In brief, a larval system must completely coincide with the system based entirely on imaginal characters, or, what amounts to the same thing, the form-relationships of the larvæ must correspond exactly with the form-relationships of the imagines.

On the other hand, the condition of affairs must be quite different if an internal power causing phyletic remodelling does not exist, the transformation of species depending entirely on the action of the environment. In this case dissimilarities in the phyletic development of the different stages of life must be expected, since the temporary, and often widely deviating, conditions of life in the two stages can and must frequently influence the one stage whilst leaving the other unacted upon—the former can therefore undergo remodelling while the latter remains unchanged.[169]

By this means there would arise an unequal difference between the two stages of two species. Thus, the butterflies, supposing these to have become changed, would bear a more remote form-relationship to each other than the caterpillars, and the differences between the former (imagines) would always be greater than that between the larvæ if the butterflies were, at several successive periods, affected by changing influences whilst the larvæ continued under the same conditions and accordingly remained unaltered. The two stages would not coincide in their phyletic development—the latter could not be expressed by parallel lines, and we should accordingly expect to find that there was by no means a complete congruity between the systems founded on the larval and imaginal characters respectively, but rather that the caterpillars frequently formed different systematic groups to the butterflies.[170]

Accordingly, the problem to be investigated was whether in those species which develope by means of metamorphosis, and of which the individual stages exist under very different conditions of life, a complete phyletic parallelism was to be found or not. This cannot be decided directly since we cannot see the phyletic development unfolded under our observation, but it can be established indirectly by examining and comparing with each other the form-relationships of the two separate stages—by confronting the larval and imaginal systematic groups. If the phyletic development has been parallel and perfectly equal, so also must its end-results—the forms at present existing—stand at equal distances from one another; larval and imaginal systems must coincide and be congruent. If the course of the phyletic development has not been parallel, there must appear inequalities—incongruences between the two systems.

I am certain that systematists of the old school will read these lines with dismay. Do we not regard it as a considerable advance in taxonomy that we have generally ceased to classify species simply according to one or to some few characters, and that we now take into consideration not merely the last stage of the development (the imago), but likewise the widely divergent young stages (larva and pupa)? And now shall it not be investigated whether caterpillars and butterflies do not form quite distinct systems? In the case of new species of butterflies of doubtful systematic position was not always the first question:—what is the nature of the caterpillars? and did not this frequently throw light upon the relationships of the imago? Assuredly; and without any doubt we have been quite correct in taking the larval structure into consideration. But in so doing we should always keep in mind that there are two kinds of relationship—form- and blood-relationship—which might possibly not always coincide.

It has hitherto been tacitly assumed that the degree of relationship between the imagines is always the same as that between the larvæ, and if blood-relationship is spoken of this must naturally be the case, since the larva and the imago are the same individual. In all groups of animals we have not always the means of deciding strictly between form- and blood-relationship, and must accordingly frequently content ourselves by taking simply the form-relationship as the basis of our systems, although the latter may not always express the blood-relationship. But it is exactly in the case of metamorphic species that there is no necessity for, nor ought we to remain satisfied with, this mode of procedure, since we have here two kinds of form-relationship, that of the larvæ and that of the imagines, and, as I have just attempted to show, it is by no means self-evident that these always agree; there are indeed already a sufficient number of instances to show that such agreement does not generally exist.

This want of coincidence is strikingly shown in a group of animals widely remote from the Insecta, viz. the Hydromedusæ, the systematic arrangement of which is quite different according as this is based on the polypoid or on the medusoid generation. Thus, the medusoid family of the oceanic Hydrozoa springs from polypites belonging to quite different families, and in each of these polypoid families there are species which produce Medusæ of another family.

Similarly, the larvæ of the Ophiuroidea (Pluteus-form) among the Echinodermata are not the most closely related in form to those of the ordinary star-fishes, but rather to the larvæ of quite a distinct order, the sea-urchins.

I will not assert that in these two cases the dissimilarity in the form-relationship, or, as I may designate it, the incongruence of the morphological systems, must depend on an unequal rate of phyletic development in the two stages or generations, or that this incongruence can be completely explained by the admission of such an unequal rate of development: indeed it appears to me probable that, at least in the Ophiureæ, quite another factor is concerned—that the form-relationship to the larvæ of the sea-urchins does not depend upon blood-relationship, but on convergence (Oscar Schmidt), i.e. on adaptation to similar conditions of life. These two cases, however, show that unequal form-relationship of two stages may occur.

From such instances we certainly cannot infer off-hand that a phyletic force does not exist; it must first be investigated whether and to what extent such dissimilarities can be referred to unequal phyletic development and, should this be the case, whether deviations from a strict congruence of the morphological systems are not compatible with the admission of an internal transforming power. That a certain amount of influence is exerted by the environment on the course of the processes of development of the organic world, will however be acceded to by the defenders of the phyletic vital force. It must therefore be demonstrated that deviations from complete congruence occur, which, from their nature or magnitude, are incompatible with the admission of innate powers, and, on the other hand, it must likewise be attempted to show that the departures from this congruence as well as the congruence itself can be explained without admitting a phyletic vital force.

In the following pages I shall attempt to solve this question for the order Lepidoptera, with the occasional assistance of two other orders of insects. Neither the Echinodermata nor the Hydromedusæ are at present adapted to such a critical examination; the number of species in these groups of which the development has been established with certainty is still too small, and their biological conditions are still to a great extent unknown. In both these respects they are far surpassed by the Lepidoptera. In this group we know a large number of species in the two chief stages of their development and likewise more or less exactly the conditions under which they exist during each of these phases. We are thus able to judge, at least to a certain extent, what changes in the conditions of life produce changes of structure. Neither in the number of known species of larvæ, nor in the intimate knowledge of their mode of life, can any of the remaining orders of insects compete with the Lepidoptera. There is no Dipterous or Hymenopterous genus in which ten or more species are so intimately known in the larval stage that they can be employed for the purposes of morphological comparison. Who is able to define the distinctions between the life-conditions of the larvæ of twenty different species of Culex or of Tipula? The caterpillars of closely allied species of Lepidoptera, on the other hand, frequently live on different plants, from which circumstance alone a certain difference in the life-conditions is brought about.

The chief question which the research had to reply to was the following:—Does there exist a complete phyletic parallelism among Lepidoptera or not? or, more precisely speaking:—Can we infer, from the form-relationships which at present exist between larvæ on the one hand and imagines on the other, an exactly parallel course of phyletic development in both stages; or do incongruences of form-relationship exist which point to unequal development?

Before I proceed to the solution of this question it is indispensable that one point should be cleared up which has not been hitherto touched upon, but which must be settled before the problem can be formally stated in general terms. Before it can be asked whether larvæ and imagines have undergone a precisely parallel development, we must know whether unequal development is possible—whether there does not exist such an intimate structural relationship between the two stages that every change in one of these must bring about a change in the other. Were this the case, every change in the butterfly would cause a correlative change in the caterpillar, and vice versâ, so that an inequality of form-relationship between the larvæ on one hand and the imagines on the other would be inconceivable—systems based on the characters of the caterpillars would completely coincide with those based on the characters of the butterflies and we should arrive at a false conclusion if we attributed the phyletically parallel development of the two stages to the existence of an internal phyletic force, whilst it was only the known factor, correlation, which caused the equality of the course of development.

For these reasons it must first be established that the larva and imago are not respectively fixed in form, and the whole of the first section will therefore be devoted to proving that the two stages change independently of one another. Conclusions as to the causes of change will then be drawn, and these will corroborate from another side a subsequent inquiry as to the presence or absence of complete congruence in the two morphological systems. The two questions the answers to which will be successively attempted are by no means identical, although closely related, since it is quite conceivable that the first may be answered by there being no precise correlation of form, or only an extremely small correlation, between the caterpillar and the imago, whilst, at the same time, it would not be thereby decided whether the phyletic development of the two stages had kept pace uniformly or not. A perfect congruence of morphological relationships could only take place if transformations resulted from an internal power instead of external influences. The question:—Does there exist a fixed correlation of form between the two stages? must therefore be followed by another:—Do the form-relationships of the two stages coincide or not—has their phyletic development been uniform or not?

STUDIES IN THE THEORY
OF DESCENT

BY
DR. AUGUST WEISMANN
PROFESSOR IN THE UNIVERSITY OF FREIBURG

WITH NOTES AND ADDITIONS BY THE AUTHOR

TRANSLATED AND EDITED, WITH NOTES, BY
RAPHAEL MELDOLA, F.C.S.
LATE VICE-PRESIDENT OF THE ENTOMOLOGICAL SOCIETY OF LONDON

WITH A PREFATORY NOTICE BY
CHARLES DARWIN, LL.D., F.R.S.
Author of “The Origin of Species,” &c.

IN TWO VOLUMES
VOL. II.

WITH EIGHT COLOURED PLATES

London:
SAMPSON LOW, MARSTON, SEARLE, & RIVINGTON
CROWN BUILDINGS, 188, FLEET STREET
1882
[All rights reserved]

---

I.
Larva and Imago vary in Structure independently of each other.

It would be meaningless to assert that the two stages above mentioned were completely independent of one another. It is obvious that the amount of organic and living matter contained in the caterpillar determines the size of the butterfly, and that the quantity of organic matter in the egg must determine the size of the emergent larva. The assertion in the above heading refers only to the structure; but even for this it cannot be taken as signifying an absolute, but only a relative independence, which, however, certainly obtains in a very high degree. Although it is conceivable that every change of structure in the imago may entail a correlative change of structure in the larva, no such cases have as yet been proved; on the contrary, all facts indicate an almost complete independence of the two stages. It is quite different with cases of indirect dependence, such, for example, as are brought about by ‘nurse-breeding.’ This phenomenon is almost completely absent in Lepidoptera, but is found in Diptera, and especially in Hymenoptera in every degree. The larvæ of ichneumons which live in other insects, require (not always, but in most instances) that the female imago should possess a sharp ovipositor, so that in this case also the structure and mode of life of the larva influences the perfect insect. This does not depend, however, on inherent laws of growth (correlation), but on the action of external influences, to which the organism endeavours to adapt itself by natural selection.

I will now let the facts speak for themselves.

It is shown by those species in which only one stage is di- or polymorphic that not every change in the one stage entails a corresponding change in the other. Thus, in all seasonally dimorphic species we find that the caterpillars of butterflies which are often widely different in the colour and marking of their successive generations are absolutely identical. On the other hand, many species can be adduced of which the larvæ are dimorphic whilst the imagines occur only in one form (compare the first and second essays in this volume).

There are however facts which directly prove that any one stage can change independently of the others; I refer to the circumstance that any one stage may become independently variable—that the property of greater variability or of greater constancy by no means always occurs in an equal degree in all the three stages of larva, pupa, and imago, but that sometimes the caterpillar is very variable and the pupa and imago quite constant. On the other hand, all three stages may be equally variable or equally constant, although this seldom occurs.

If variability is to be understood as indicating the period of re-modelling of a living form, whether in its totality or only in single characters or groups of characters, from the simple fact of the heterochronic variability of the ontogenetic stages, it follows that the latter can be modified individually, and that the re-modelling of one stage by no means necessarily entails that of the others. It cannot however be doubted that variability, from whatever cause it may have arisen, is in all cases competent to produce a new form. From the continued crossing of variable individuals alone, an equalization of differences must at length take place, and with this a new, although not always a widely deviating, constant form must arise.

That the different stages of development of a species may actually be partly variable and partly constant, and that the variable or constant character of one stage has no influence on the other stages, is shown by the following cases, which are, at the same time, well adapted to throw light on the causes of variability, and are thus calculated to contribute towards the solution of the main problem with which this investigation is concerned.

When, in the following pages, I speak of variability, I do not refer to the occurrence of local varieties, or to variations which occur in the course of time, but I mean a high degree of individual variability—a considerable fluctuation of characters in the individuals of one and the same district or of the same brood. I consider a species to be constant, on the other hand, when the individuals from a small or large district differ from one another only to a very slight extent. Constant forms are likewise generally, but not invariably, such as are poor in local varieties, whilst variable forms are those which are rich in such variations. Since the terms “variable” and “constant” are but relative, I will confine myself to the most extreme cases, those in which the individual peculiarities fluctuate within very wide or very narrow limits.

As no observations upon the degree of variability shown by a species in the different stages of its development were available, I was obliged to fall back upon my own, at least so far as relates to the larval and pupal stages, whilst for the imaginal stage the wide experience of my esteemed friend Dr. Staudinger has been of essential service to me.

Let us in the first place confine our attention to the three chief forms which every Lepidopteron presents, viz. larva, pupa, and imago. With respect to the constancy or variability of these three forms, we actually find in nature all the combinations which are theoretically conceivable.

(1.) There are species which possess a high degree of constancy in all three stages, such, for example, as Limenitis Camilla, Pieris Brassicæ,[171] Sphinx Ligustri, and Euchelia Jacobææ.

(2.) There are species showing a high degree of variability in all three stages. This case must be of rare occurrence, as I am only able to adduce Araschnia Prorsa-Levana, a fact which arises from the circumstance that the pupal stage is, as a rule, but seldom variable.

(3.) There are species which are variable in two stages and constant in the third. To this class, for example, belongs Smerinthus Tiliæ, of which the larva and imago are very variable, whilst the pupa is quite constant. The same is the case with Lasiocampa Pini, the well-known fir moth. Many butterflies show this same phenomenon in other combinations, such, for instance, as Vanessa Urticæ and Polychloros, in which the larva and pupa are very variable, and the imago very constant. In a less degree the same is also the case with Vanessa Atalanta, whilst in Pieris Napi the pupa and imago are variable, and the caterpillar remarkably constant, this likewise being the case with the local form Bryoniæ, which, according to my theory, is to be regarded as the parent form of Napi (See [Part I.] of the present volume).

(4.) There are species which are constant in two stages, and variable only in the third. Thus, a few species can be found in which the larva and pupa are constant and the imago variable. This is the case with Saturnia Yamamai, the imago of which is well known to present numberless shades of colour, varying from light yellow to greyish black, whilst the green caterpillar shows only slight individual differences of marking, and scarcely any differences of colour. The pupa of this species is quite constant. Arctia Caja and Hebe, and Chelonia Plantaginis belong to this same category.

There are a very large number of species which possess very constant imagines and pupæ, but extremely variable larvæ. The following are the cases known to me:—Macroglossa Stellatarum, Fuciformis and Bombyliformis; Chærocampa Elpenor, Celerio, and Nerii; Deilephila Galii, Livornica, Hübn., Hippophaës, Vespertilio, and Zygophylli; Sphinx Convolvuli; Acherontia Atropos; Smerinthus Ocellatus and Tiliæ; Callimorpha Hera; Cucullia Verbasci and Scrophulariæ.

Cases in which the variability depends entirely upon the pupa, while the larva and imago are extremely constant, are of great rarity. Vanessa Io is a case in point, the pupa being light or dark brown, or bright golden green, whilst in the two other stages scarcely any light shades of colour or variations in the very complicated marking are to be met with.

The facts thus justify the above view that the individual stages of development change independently—that a change occurring in one stage is without influence on the preceding and succeeding stages. Were this not the case no one stage could possibly become variable without all the other stages becoming so. Did there exist a correlation between larvæ, pupæ, and imagines of such a nature that every change in the larva entailed a corresponding change in the imago, as soon as a large number of larval characters became fluctuating (i.e. as soon as this stage became variable), a large number of imaginal characters would necessarily also become fluctuating (i.e. this stage would also become correspondingly variable).

There is one other interpretation which might perhaps be attempted from the point of view of the old doctrine of species. It might be said that it is a special property of certain larval or imaginal markings to be variable whilst others are constant, and since the larval and imaginal markings of a species are generally quite distinct, it may easily happen that a butterfly possessing markings having the property of constancy may belong to a caterpillar having variable markings.

There is a soul of truth underlying this objection, since it is true that the various forms of markings which occur in Lepidoptera apparently reach different degrees of constancy. If we speak of the constancy or variability of a species, a different meaning is attached to these expressions according as we are dealing e.g. with a species of Sphinx or a species of Arctia. That which in the latter would be estimated as a high degree of constancy, in the former would be taken as a considerable amount of variability. It is of interest, in connection with the question as to the causes of constancy, to note that the power of any form of marking to attain to a high degree of constancy is by no means inversely proportional to the complication of the marking, as would have been expected à priori.

Thus, the species of Sphinx and of allied genera possess on their fore-wings, which are mostly coloured with a mixture of dull grey, white and black, an exceedingly complicated arrangement of lines which, in constant species, show a high degree of uniformity: on the other hand, the checquered fore-wings of our Arctiidæ, which are far more coarsely marked, always show, even in the most constant species, well-marked individual differences. The different types of marking must therefore be measured by different standards.

But in granting this, we decidedly refute the statement that constancy and variability are inherent properties of certain forms of marking.

This reasoning is based on the simple fact that a given type of marking comprises both species of great constancy and of (relatively) great variability.

Thus, the fore-wings of Sphinx Ligustri and S. Convolvuli are extremely constant, whilst the very similarly marked Anceryx (Hyloicus) Pinastri is exceedingly variable. Similarly Deilephila Euphorbiæ is known by its great variability of colouring and marking, whilst D. Galii, which resembles this species so closely as to be sometimes confounded with it, possesses a high degree of constancy, and further, the Corsican and Sardinian D. Dahlii is very variable. Among the family Arctiidæ, Callimorpha Hera and the Alpine Arctia Flavia are cases of constancy, whilst A. Caja, which is so similar to the last species, is so generally variable that two perfectly identical specimens can scarcely be found together.

The same can be shown to hold good for the markings of caterpillars. Thus, the larva of D. Dahlii shows very considerable variability, whilst that of D. Galii is very constant in marking (disregarding the ground-colour). So also the larva of Vanessa Urticæ is very variable and that of V. Antiopa very constant, &c.

The great differences with respect to constancy or variability which are displayed by the different stages of one and the same species, must therefore find their explanation elsewhere than in the type of the marking itself. The explanation must be found in the circumstance that each stage changes independently of the others, and at different periods can enter a new phase of variability.

We are here led in anticipation to the main question:—Are changes produced by internal or external causes? is it the physical nature of the organism which is compelled to become remoulded spontaneously after the lapse of a certain period of time? or does such modification only occur when produced directly or indirectly by the external conditions of life?

In the cases before us the facts undoubtedly indicate a complete dependence of the transformations upon external conditions of life.

The independent appearance of variability in the separate stages of the metamorphosis might, however, be regarded as only apparent. It might still be attempted to attribute the changes to a purely inherent cause, i.e., to a phyletic vital force, by assuming that the latter acts periodically in such a manner that at first one and then the following stage becomes variable, until finally the entire species is transformed.

There is but little to be said in reply to this if we once take refuge in entirely unknown forces, the operation of which can be arbitrarily conceived to be either constant or periodic.

But granting that such a transforming power exists and acts periodically, the variability must always pass over the different stages in a fixed direction, like a wave over the surface of water—imago, pupa, and larva, or larva, pupa, and imago, must successively become variable. Cases like that of Araschnia Prorsa, in which all three stages are variable, may certainly be thus explained, but those instances in which the larva and imago are extremely variable, and the pupa quite constant, are entirely inexplicable from this point of view.

The latter can, however, be very simply explained if we suppose the changes to be dependent upon external influences. From this standpoint we not only see how it is possible that an intermediate stage should remain uninfluenced by the changes which affect the two other stages, but we can also understand why it should just be the pupal stage that plays this part so frequently. If we ask why most pupæ are constant and are relatively but very slightly variable, the answer will be found in the facts that all pupæ which remain concealed in the earth or inside plants (Sesiidæ), or which are protected by stout cocoons, show complete constancy, whilst any considerable amount of variability occurs only in those pupæ which are suspended or openly exposed. This is closely connected with a fact to which I have called attention on a former occasion,[172] viz., that dimorphism occurs in certain pupæ, but only in those which are openly exposed and which are therefore visible to their foes. I am only acquainted with such cases among the pupæ of butterflies, and it is likewise only among these that I have found any considerable amount of variability.

Facts of this kind indicate that Nature does not uselessly sport with forms, but that at any rate changes of this sort result from external influences. The greater frequency of variability among larvæ and its comparative rarity in imagines is also undoubtedly in favour of this view.

It has already been shown that species with variable larvæ and constant imagines are extremely common, but that those with constant larvæ and variable imagines are very rare. This confirms the conclusions, already drawn above, first, that the variability of the imago cannot owe its existence to the variability of the larvæ, and secondly, that the causes which produce variability affect the larval condition more commonly than that of the imago.

Where can these causes be otherwise sought than in the external conditions of life, which are so widely different in the two stages, and which are much more variable for the larva than for the imago?

Let us take the species of one genus, e.g. those of Deilephila. The imagines of our European species—as far as we know—all live in precisely the same manner; they all fly at twilight,[173] showing a preference for the same flowers and very often frequenting the same spots, so that in the haunts of one species the others are almost always to be met with, supposing them to occur in the same locality. They conceal themselves by day in similar places, and are attacked by similar foes.

It is quite different with the caterpillars. These, even in the case of the most closely allied species, live under different conditions, as appears from the fact that they feed on different plants. The latter can, however, produce changes both directly and indirectly. The larvæ may acquire adaptive colours and markings, and these would vary in accordance with the colour and structure of the food-plant; or they may become brightly coloured as a sign of distastefulness in cases where they are inedible. Then again the colour of the soil on which the larvæ live would act upon their colours making these adaptive. Certain habits of the caterpillars may also be dependent upon the nature of their food-plants. Thus, e.g. Deilephila Hippophaës feeds only at night, and conceals itself by day under moss and among the leaves at the base of the food-plant; but D. Euphorbiæ could not acquire such a habit, because Euphorbia Cyparissias generally grows on arid soil which is poor in vegetation, and which therefore affords no concealment, and furthermore, because a caterpillar, as long as it continues to feed, cannot, and as a matter of fact does not, ever wander far from its food-plant. A habit of concealment by burying in the earth also, such for example as occurs in Acherontia Atropos, could not be acquired by D. Euphorbiæ, because its food-plant generally grows on hard, dry, and stony ground.

In addition to these considerations, the foes would be different according as the caterpillar lived on plants which formed dense thickets covering large extents of the shore (Hippophae) or grew isolated on dry hillocks and declivities where the herbage was scanty or altogether absent; or again, according as the insect, in conjunction with such local differences, fed by day or had acquired the habit of feeding only by night. It must in fact be admitted that new and improved adaptations, or, in more general terms, that inducements to change, when depending on the environment, must be more frequently dissimilar for larvæ than for the imagines. We must accordingly expect to find actual change, or that condition of variability which may be regarded as initiative to change, occurring more commonly in larvæ than in perfect insects.

Since facts are in complete accordance with the results of these à priori considerations we may also venture to conclude that the basis of the considerations is likewise correct, viz., the supposition that the changes of colour and marking in caterpillars, pupæ, and imagines result from external influences only.

This must not be taken as signifying that the single stages of the larval development are also only able to change through the action of external influences. The larval stages are correlated with each other, as has already been shown (see the previous essay): new characters arise in the adult caterpillar at the last stage and are then gradually transferred back to the younger stages quite independently of external influences, this recession being entirely brought about by the laws of correlation. Natural selection here only exerts a secondary action, since it can accelerate or retard this transference, according as the new characters are advantageous or disadvantageous to the younger stages.

Now as considerable individual differences appear in the first acquisition of a new character with respect to the rapidity and completeness with which the individuals acquire such a character, the same must obtain for the transference of an improvement acquired in the last stage to the next younger stage. The new character would be acquired by different individuals in different degrees and at different rates—it would have, to a certain extent, to struggle with the older characters of the stage; in brief, the younger stage would become variable.

Variability of this kind might well be designated as secondary, in contradistinction to primary variability; the latter (primary) depends upon an unequal reaction of the individual organisms to external influences, the former (secondary) results from the unequal strength and rate of the action of the innate laws of growth governing the organism. In both cases alike exceeding variability may occur, but the causes producing this variability are dissimilar.

The different stages of larval development would thus frequently display independent variability in a manner similar to the pupal or imaginal stages, since they can show individual variability while the other stages of development remain constant. This appearance of independent variability in the different stages of the larval development, however, is in truth deceptive—we have here in fact a kind of wave of variability, which passes downwards through the developmental stages, becoming gradually weaker, and finally dying out completely.

In accordance with this, we very frequently find that only the last or two last stages are variable, while the younger stages are constant. Thus in Macroglossa Stellatarum, the larvæ are constant in the first, second, and third stages, but become variable in the fourth, and in the fifth stage first show that high degree of variability which has already been described in detail (See. [Pl. III]., Figs. 3–12). The larvæ, of Vanessa Cardui also, according to my notes, are extremely constant in the first four stages in spite of their complicated marking, but become variable in the fifth stage, although to no very great extent.

In Smerinthus Tiliæ, Ocellatus and Populi also, the greatest larval variability is shown only in the last stage, the preceding stages being very constant. These cases by no means depend upon the marking of the young stages being simpler and therefore being less capable of varying. The reverse case also occurs. In a somewhat similar manner as the young of the tapir and wild hog are striped, while the adult animals are plainly coloured, the young caterpillars of Saturnia Yamamai possess longitudinal black lines on a yellow ground, while as early as in the second stage a simple green colour appears in the place of this complicated but perfectly constant marking. If the young stages are so frequently constant, this rather depends upon the fact that the transference of a new character to these stages not only takes place gradually, but also with continually diminishing energy, in a manner somewhat similar to physical motion, which continually diminishes in speed by the action of resistance till it is completely arrested. This constancy of the younger stages may further be due to the circumstance that the characters would only be transferred when they had become fixed in the last stage, and were consequently no longer variable. The transferred characters may thus have acquired a greater regularity, i.e. a less degree of variability, than they possessed at their first origination. Extensive investigations in this special direction must be made if the precise laws, in accordance with which the backward transference of new characters takes place, are to be discovered. By such researches only should we arrive with certainty at the causes which determine the lesser variability of the young larval stages.

It may also occur that the early stages are variable, whilst the later stages are constant, although this case appears to happen less frequently. Thus, the caterpillars of Gastropacha Quercifolia vary considerably in the second stage but are constant at a later period, and the same is the case with Spilosoma Urticæ, which in the second stage may be almost considered to be dimorphic, but which subsequently becomes constant.

Cases in which the first stage is variable appear to be of the least frequent occurrence. I know of only one such instance, viz., Anceryx Pinastri, of which the newly hatched larvæ ([Pl. VI]., Fig. 53) show considerable differences in the brownish-black crescentic spots. The second ([Fig. 54]), third, and fourth stages are then tolerably constant, while the fifth stage again is very variable.

An instance of this kind can be easily explained by two waves of variation, the first of which now affects only the first stage, while the second has just commenced to affect the fifth stage. Such a supposition is not opposed to any theoretical considerations, but rather has much probability in its favour, since we know that species are from time to time subject to be remodelled; and further, that the coalescence of several stages of phyletic development in the ontogeny of one and the same species (see p. [226], development of the genus Deilephila) shows that during the backward transference of one character, new characters may appear in the last stage of the ontogeny, and indeed very frequently at a time when the next youngest character has not been transferred back so far as to the first stage.

That this secondary variability is to a certain extent brought about by the conflict between the old and new characters, the latter striving to suppress the former, is shown by the caterpillar of Saturnia Carpini which I have observed for many years from this point of view, and than which I do not know a more beautiful illustration.

When these larvæ leave the egg they are black, but in the adult state are almost bright green—this at least being the case in a local form which, from the district in the vicinity of Genoa where it is found, I will designate as the var. Ligurica. Now whilst these two extreme stages of development are relatively constant, the intermediate stages show a variability which becomes greater the nearer the last stage is approached, this variation in the marking depending simply on the struggle between the green colour and the more anciently inherited black. In this manner there arises, especially in the fourth stage of the German local form, an incredible mixture of the most diverse markings, all of which can, however, be very easily explained from the foregoing point of view.

The simpler and, as I am inclined to believe, the older form of the transformation is presented to us in the local variety Ligurica. In the last stage, when 7.5 centimeters long, this form is of a beautiful bright green colour without any trace of black marking[174] ([Pl. VIII]., Fig. 77). The colour of the six orange warts which are situated on each segment is also similar in all specimens, so that this stage is perfectly constant.

Our German S. Carpini shows different characters in the fifth stage. It is true that individual specimens occur which are entirely green without any black, but these are rare; the majority possess a more or less broad black ring encircling the middle of each segment ([Pl. VIII]., Figs. 78 and 79). Those specimens in which the black ring has become broken up into large or small spots surrounding the base of the warts constitute intermediate forms ([Fig. 80]). The last stage of the German local form, unlike that of the Genoese local form, is therefore very variable.

The two forms, moreover, do not simply differ in being more or less advanced in phyletic development, but also in several other points. As it is of great theoretical interest to show that a species can develop local differences only in the stage of larva, I will here subjoin the plain facts.

The differences consist in that the Genoese local form goes through five moults whilst the German local form, like most caterpillars, has only four moults. Further, in the Genoese form the light green, which is also possessed by the German form in the fourth stage, when it once appears, is retained to the end of the larval development, whilst in the fifth stage of the German form this colour is replaced by a dull greyish-green (compare [Figs. 77 and 78]). There is further a very considerable difference in the earlier stages which shows that the phyletic transforming process has taken a quite independent course in the two forms. Since the struggle between the green and black—retaining this idea—appears to be quite finished in the last stage of the Genoese form, we should expect that the new colour, green, would now also have encroached further upon the younger stages than in the German form. Nevertheless, this is not the case, but quite the reverse happens, the black maintaining its ground longer in the Italian than in the German form.

In the Genoese form the two first stages are completely black, and in the third stage an orange-yellow lateral stripe first appears. In the German form this stripe appears in the second stage, and there is not subsequently added, at least on the middle segments, a yellow border surrounding some of the warts of the median series. In the third stage, however, the yellow (which is but the precursor of the later green colour) becomes further extended, so that the caterpillars often appear of an orange colour, some or all of the warts and certain spots and stripes only being black ([Figs. 66 and 68]). The warts are also often yellow while the ground remains in most part black—in brief, the bright colour is in full struggle with the black, and an endless series of variations is the result of this conflict, whilst in the corresponding stage of the Genoese form almost complete constancy prevails.

This constancy remains also in the following (fourth) stage, the caterpillar still being deep black, only the yellow (sulphur-coloured) lateral stripe, which has now become brighter, indicating the impending change ([Fig. 67]). This takes place in the fifth stage, in which the ground-colour suddenly becomes bright green, the black remaining at most only in traces on the anterior edges of the segments.

This is the same marking as is shown by the fourth stage of the German form, only in this case individuals quite destitute of black do not occur. In many specimens indeed black forms the ground-colour, the green only appearing in certain spots ([Figs. 71 to 75]); in others the green predominates, and these two extremes are connected by innumerable intermediate forms, so that this stage must be regarded as the most variable of all.

The sixth stage of the Genoese and the fifth of the German form have already been compared together. The results may be thus tabulated:—

A. German form. B. Genoese form.
Stage I. 9 days. Black; constant. 9 days. Black; constant.
Stage II. 8 days. Black, with orange-yellow lateral stripe; variable. Black, with yellow; very variable. 11 days. Black; constant.
Stage III. 5 days (in some cases as much as 16 days). 12 days. Black, with orange-yellow lateral stripes; constant.
Stage IV. 16 days (in some cases only 5 days). Bright green and black mixed; very variable. 6 days. Black, with bright yellowish lateral stripe; constant.
Stage V. 6 days (frequently longer). Dark green, with or without black bands; variable. 6 days. Bright green, small traces of black; variable.
Stage VI. Pupation. 18 days. Bright green, without any black; constant.
Stage VII. Pupation.

From this comparison we perceive that the process of transformation has at least become preliminarily concluded in the Genoese form. Why the backward transference of the newly-acquired character to the young stages has not yet occurred, or, at least, why it is not in progress, does not appear; neither can it be stated whether this will take place later, although we may venture to suppose that such will be the case. At first sight but a relatively short time appears necessary for the single stage V., which is still in a state of fluctuation (variable), to become constant by continued crossing, like all the other stages.

That the transformation is still in full progress in the German form, is shown by the fact that in this case all the stages are variable with the exception of the first—the second stage being only variable to a small extent, the third to a much greater extent, and the fourth to the highest degree conceivable, whilst the fifth and last stage is again less variable—so that the greatest struggle between the old and new characters takes place in the fourth stage.

Among the innumerable variations presented by this last stage a complete series of transitional forms can be arranged so as to show the gradual conquest of the black by the green, and thus indicating, step by step, the course which the latter colour has taken.

In the blackest specimens there is nothing green but the lateral (infra-spiracular) line which was yellow in the preceding stage, and a crescent-shaped streak at the base of the middle warts together with a still smaller crescent at the base of the upper warts ([Figs. 71 and 81]). These spots become extended in lighter specimens and approximate so as to leave only narrow black bridges, a third spot being added at the posterior edge of the warts ([Figs. 72 and 82]). The three spots then extend on all sides, still leaving for a long period narrow black lines at the boundaries where their growth has caused them to abut. In this manner there frequently arises on the green ground a true hieroglyphic-like marking ([Figs. 85 and 86]). Finally the black disappears from the anterior edge and diminishes on the middle line of the back where it still partly remains as a T-shaped figure ([Figs. 73 and 74]), although generally replaced elsewhere by the green with the exception of small residues.

One point remained for a long time inexplicable to me, viz., the change of the light green into dark grey-green which appeared in the last stage in connection with a total change of the black marking.

Supposing that new characters are actually acquired only in the last stage, and that from this they are transferred to the younger stages, we should expect to find completely developed in the last stage the same colouring and markings as are possessed more or less incompletely in the fourth stage. Now since the developmental tendency to the removal of black and to the predominance of green—if we may thus venture to express it—is obvious in the fourth stage, we may expect to find in the fifth stage a bright green ground-colour, either without any mixture of black or with such black spots and streaks as were retained in the fourth stage as residues of the original ground-colour. But instead of this the fifth stage shows a dark green colour, and a more or less developed black marking which cannot in any way be derived from that of the fourth stage.

The Genoese local form observed last year first gave me an explanation to the extent that in this form the last stage is actually only the potential penultimate stage, or, more correctly expressed, that the same characters which at present distinguish the last stage of this form, are already more or less completely transferred to the penultimate stage.

The apparently paradoxical behaviour of the German form can be explained by supposing that before the pure bright green had become completely transferred to the penultimate stage a further change appeared in the last stage, the green ground-colour becoming darker, and black transverse bands being formed. The marking of the last stage would then be regarded as the reverse of that of the preceding stage; the absence of black would be the older, simple black spots at the base of the warts the next in succession, and a connected black transverse band the most advanced state of the development.

Whether this explanation is correct, and if so, what causes have produced the second change, may perhaps be learnt at some future time by a comparison with the ontogeny of other Saturniidæ; in the meantime this explanation receives support from another side by the behaviour of the Genoese local form. If the last stage of the German form has actually commenced to be again re-modelled, then this variety is further advanced in phyletic development than the Genoese form; and this corresponds entirely with the theory that in the former the light colour (the orange considered as preliminary to the transformation into green) has already been carried down into the second stage, whilst in the Genoese variety even in the fourth stage only the first rudiments of the colour-transformation show themselves.

The Genoese form is to a certain extent intermediate between the German form of Saturnia Carpini and the nearly related S. Spini, a species inhabiting East Germany. In this latter the larvæ, even in the adult state, are completely black with yellow warts. This form of caterpillar must therefore be regarded as phyletically the oldest, and this very well agrees with the character of the moth, which differs essentially from S. Carpini only in not being sexually dimorphic. In Carpini the male possesses a far more brilliant colouring than the female, the latter agreeing so completely with the female of Spini that it can hardly be distinguished therefrom, especially in the case of the somewhat larger South European specimens of the last species. Now as the more simple colouring of the female must in any case be regarded as the original form, we must consider Spini, both sexes of which possess this colouring, to be phyletically the older form, and Carpini, the male of which has become differently coloured, must be considered as the younger type. This completely accords with the characters of the larvæ.

I must here mention that I have also asked myself the question whether the variations of the different larval stages are connected together as cause and effect—whether the lightest specimens of the fifth stage may perhaps not also have been the lightest individuals of the third and fourth stages.

Such relationship is only apparent between the third and fourth stages; the darkest larvæ of the third stage become the darker varieties of the fourth stage, although it is true that the lighter forms of the third sometimes also become dark varieties in the fourth stage. Between the fourth and fifth stages there is scarcely any connection of this kind to be recognized. Thus, the darkest varieties of the fourth stage sometimes become the lightest forms of the fifth stage, whilst in other cases from the lightest individuals of the fourth stage there arise all the possible modifications of the fifth stage. Further details may be omitted: the negative result cannot cause any surprise, as it is a necessary consequence of the continued crossing that must take place.

We thus see that the three chief stages of development (larva, pupa, and imago) actually change in colour independently of each other, the single stages of the larval development being however in greater dependence upon one another, and being connected indeed in such a manner that a new character cannot be added to the last stage without being transferred in the course of time to the preceding stage, and at a later period from this again even to the youngest stage, supposing it not to be previously delayed in the course of its transference by unknown opposing forces. On this last point, however, the facts at present available do not admit of any certain decision.

But why do the individual larval stages behave in this respect so very differently to the chief stages of the whole development? why are the former so exactly correlated whilst the latter are not? If new characters have a general tendency to become transferred to the younger ontogenetic stages, why are not new imaginal characters first transferred to the pupa, and finally to the larva?

The answer to these questions is not far to find. The wider two stages of a species differ in structure, the less does correlation become possible; the nearer the two stages are morphologically related, the more powerful does the action of correlation become. It is readily conceivable that the more widely two succeeding stages deviate in structure and mode of life, the less possible does it become for characters to be transferred from one to the other. How is it possible, for example, that a new character in the proboscis or on the wings of a butterfly can be transferred to the caterpillar? If such correlation existed it could only manifest itself by some other part of the caterpillar changing in correspondence with the change of the proboscis or wings of the butterfly. That this is not the case has, in my opinion, been conclusively shown by all the foregoing considerations respecting the independent variability of the chief stages of the metamorphosis.

There are, moreover, an endless number of facts which prove the independence of the individual stages of development—I refer to the multitudinous phenomena presented by metamorphosis itself. The existence of that form of development which we designate as metamorphosis is alone sufficient to prove incontestably that the single stages are able to change independently of one another to a most remarkable extent.

If we now ask the question: how has the so-called “complete” metamorphosis of insects arisen? the answer can only be: through the gradual adaptation of the different stages of development to conditions of life which have continually deviated more and more widely from each other.[175]

But if individual stages of the post-embryonic development can finally attain to such complete diversity of structure as that of the larva and imago through gradual adaptations to continually diverging conditions of life, this shows that the characters acquired by the single stages are always only transferred to the same stages of the following generation, whilst the other stages remain uninfluenced thereby. This depends upon that form of heredity designated by Darwin “inheritance at corresponding periods of life,” and by Haeckel “homochronic heredity.”

---

II.
Does the Form-relationship of the Larva coincide with that of the Imago?

Having thus established the independence in the variability of the individual stages of metamorphosis, I will now turn to the consideration of the question as to how far a parallelism is displayed in the phyletic development of these stages. Is there a complete congruence of form-relationship between larvæ on the one hand and imagines on the other? does the classification founded on the morphology of the imagines agree with that based on the morphology of the larvæ or not?

If, according to Claus,[176] we divide the order Lepidoptera into six great groups of families, it is at once seen that these groups, which were originally founded exclusively on imaginal characters, cannot by any means be so clearly and sharply defined by the larval characters.

This is certainly the case with the Geometræ, of which the larvæ possess only ten legs, and on this account progress with that peculiar “looping” movement which strikes even the uninitiated. This group, which is very small, is however the only one which can be founded on the morphology of the larvæ; it comprises only two nearly related families (Phytometridæ and Dendrometridæ), and it is not yet decided whether these should not be united into one group comprising the family characters of the whole of the “loopers.”

Neither the group of Micro-lepidoptera, nor those of the Noctuina, Bombycina, Sphingina, and Rhopalocera, can be based systematically on larval characters. Several of these groups are indeed but indistinctly defined, and even the imagines present no common characteristics by which the groups can be sharply distinguished.

This is well shown by the Rhopalocera or butterflies. These insects, in their large and generally brilliantly coloured wings, which are usually held erect when at rest, and in their clubbed antennæ, possess characters which are nowhere else found associated together, and which thus serve to constitute them a sharply defined group.[177] The caterpillars, however, show a quite different state of affairs. Although the larval structure is so characteristic in the individual families of butterflies, these “larval-families” cannot be united into a larger group by any common characters, and the “Rhopalocera” would never have been established if only the larvæ had been known. It is true that they all have sixteen legs, that they never possess a Sphinx-like horn, and that they are seldom hairy, as is the case with many Bombycidæ,[178] but these common negative characters occur also in quite distinct groups.

In the butterflies, therefore, a perfect congruence of form-relationship does not exist, inasmuch as the imagines constitute one large group of higher order whilst the larvæ can only be formed into families. If it be admitted that the common characters of butterflies depend on their derivation from a common ancestor, the imagines must have retained certain common characters which enable them to be recognized as allies, whilst the larvæ have preserved no such characters from the period at which the families diverged.

Without going at present into the causes of these phenomena I will pass on to the consideration of further facts, and will now proceed to investigate both the form-relationships within the families. Here there can be no doubt that in an overwhelmingly large majority of cases the phyletic development has proceeded with very close parallelism in both stages; larval and imaginal families agree almost completely.

Thus, under the group Rhopalocera there is a series of families which equally well permit of their being founded on the structure of the larva or on that of the imago, and in which the larvæ and imagines therefore deviate from one another to the same extent. This is the case, for instance, with the families of the Pieridæ, Papilionidæ, Danaidæ, and Lycænidæ.

But there are also families of which the limits would be very different if the larvæ were made the basis of the classification instead of the butterflies as heretofore. To this category belongs the sub-family Nymphalinæ. Here also a very characteristic form of caterpillar indeed prevails, but it does not occur in all the genera, being replaced in some by a quite different form of larva.

In the latest catalogue of Diurnal Lepidoptera, that of Kirby (1871), 112 genera are comprised under this family. Of these most of the larvæ possess one or several rows of spines on most or on all the segments, a character which, as thus disposed, is not met with in any other family.

This character is noticeable in genera 1 to 90, if, from those genera of which the larvæ are known, we may draw a conclusion with reference to their allies. I am acquainted with larvæ of genus 2, Agraulis, Boisd. (Dione, Hübn.); of genus 3, Cethosia, Fabr.; 10, Atella, Doubl.; 12, Argynnis, Fabr.; 13, Melitæa,[179] Fabr.; 19, Araschnia, Hübn.; 22, Vanessa, Fabr.; 23, Pyrameis, Hübn.; 24, Junonia, Hübn.; 31, Ergolis, Boisd.; 65, Hypolimnas, Hübn. (Diadema, Boisd.); 77, Limenitis, Fabr.; 81, Neptis, Fabr.; 82, Athyma, Westw.; and finally with those of genus 90, Euthalia, Hübn.—which, according to Horsfield’s figures, possess only two rows of spines, these being remarkably long and curved, and fringing both sides. It may be safely assumed that the intermediate genera would agree in possessing this important character of the Nymphalideous larvæ, viz., spines.

After the genus 90 there are 22 more genera, and these are spineless, at least in the case of the two chief genera, 93, Apatura, and 104, Nymphalis. Of the remainder I know neither figures nor descriptions.[180] In the two genera named the larvæ are provided with two or more spine-like tentacles on the head, and the last segment ends in a fork-like process directed backwards. The body is otherwise smooth, and differs also in form from that of the larvæ of the other Nymphalinæ, being thickest in the middle, and tapering anteriorly and posteriorly; neither is the form cylindrical, but somewhat flattened and slug-shaped. If therefore we were to arrange these butterflies by the larvæ instead of by the imagines, these two genera and their allies would form a distinct family, and could not remain associated with the 90 other Nymphalideous genera.

We have here a case of incongruence; the imagines of the genera 1–90 and 91–112 are more closely allied than their larvæ.

From still another side there arises a similar disagreement. The larvæ of the genera Apatura and Nymphalis agree very closely in their bodily form and in their forked caudal appendage with the caterpillars of another sub-family of butterflies, the Satyrinæ, whilst their imagines differ chiefly from those of the latter sub-family in the absence of an enlargement of certain veins of the fore-wings, an essential character of the Satyrinæ.

This double disagreement has also been noticed by those systematists who have taken the form of the caterpillar into consideration. Thus, Morris[181] attempted to incorporate the genera Apatura and Nymphalis into the family Libytheidæ, placing the latter as transitional from the Nymphalidæ to the Satyridæ. But although the imagines of the genera Apatura, Nymphalis, and Libythea may be most closely related—as I believe they actually are—the larvæ are widely different, being at least as different as are those of Apatura and Nymphalis from the remaining Nymphalinæ.

Now if we could safely raise Apatura and Nymphalis into a distinct family—an arrangement which in the estimation of Staudinger[182] is correct—and if this were interpolated between the Satyridæ and Nymphalidæ, such an arrangement could only be based on the larval structure, and that of the imagines would thus remain unconsidered, since no other common characters can be found for these two genera than those which they possess in common with the other Nymphalideous genera.

The emperor-butterflies (Apatura), by the ocelli of their fore-wings certainly put us somewhat in mind of the Satyrinæ, in which such spots are always present; but this character does not occur in the genus Nymphalis, and is likewise absent in most of the other genera of this group. The genus Apatura shows in addition a most striking similarity in the markings of the wings to the purely Nymphalideous genus Limenitis, and it is therefore placed, by those systematists who leave this genus in the same family, in the closest proximity to Limenitis. This resemblance cannot depend upon mimicry, since not only one or another but all the species of the two genera possess a similar marking; and further, because similarity of marking alone does not constitute mimicry, but a resemblance in colour must also be added. The genus Limenitis actually contains a case of imitation, but in quite another direction; this will be treated of subsequently.

It cannot therefore be well denied that in this case the larvæ show different relationships to the imagines.

If the “natural” system is the expression of the genetic relationship of living forms, the question arises in this and in similar cases as to whether the more credence is to be attached to the larvæ or to the imagines—or, in more scientific phraseology, which of the two inherited classes of characters have been the most distinctly and completely preserved, and which of these, through its form-relationship, admits of the most distinct recognition of the blood-relationship, or, inversely, which has diverged the most widely from the ancestral form? The decision in single instances cannot but be difficult, and appears indeed at first sight impossible; nevertheless this will be arrived at in most cases as soon as the ontogeny of the larvæ, and therewith a portion of the phylogeny of this stage, can be accurately ascertained.

As in the Rhopalocera most of the families show a complete congruence in the form-relationship of the caterpillars and perfect insects, so a similar congruence is also found in the majority of the families belonging to other groups. Thus, the two allied families of the group Sphingina can also be very well characterized by their larvæ;[183] both the Sphingidæ and the Sesiidæ possess throughout a characteristic form of larva.

Of the group Bombycina the family of the Saturniidæ possess thick cylindrical caterpillars, of which the segments are beset with a certain number of knob-like warts. It is true that two genera of this family (Endromis and Aglia) are without these characteristic warts, but the imagines of these genera also show extensive and common differences from those of the other genera. A distinct family has in fact already been based on these genera (Endromidæ, Boisd.). Thus the congruence is not thereby disturbed.

So also the families Liparidæ, Euprepiidæ, and Lithosiidæ appear sharply defined in both forms; and similar families occur likewise under the Noctuina, although in this group the erection of families presents great difficulties owing to the near relationship of the genera, and is always to some extent arbitrary. It is important, however, that it is precisely the transitional families which present intermediate forms both as larvæ and as imagines.

Such an instance is offered by the Acronyctidæ, a family belonging to the group Noctuina. The imagines here show in certain points an approximation to the group Bombycina; and their larvæ, which are thickly covered with hairs, likewise possess the characteristics of many of the caterpillars of this group.[184]

A second illustration is furnished by the family Ophiusidæ, which is still placed by all systematists under the Noctuina, its affinity to the Geometrina, however, being represented by its being located at the end of the Noctuina. The broad wings and narrow bodies of these moths remind us in fact of the appearance of the “geometers;” and the larvæ, like the imagines, show a striking resemblance to those of the Geometrina in the absence of the anterior abdominal legs. For this reason Hübner in his work on caterpillars has termed the species of this family “Semi-Geometræ.”

All these cases show a complete congruence in the two kinds of form-relationship; but exceptions are not wanting. Thus, the family Bombycidæ would certainly never have been formed if the larval structure only had been taken into consideration, since, whilst the genera Gastropacha, Clisiocampa, Lasiocampa, Odonestis, and their allies, are thickly covered with short silky hairs disposed in a very characteristic manner, the caterpillars of the genus Bombyx, to which the common silkworm, B. Mori, belongs, are quite naked and similar to many Sphinx-caterpillars (Chærocampa). Are the imagines of the genera united under this family, at any rate morphologically, as unequally related as their larvæ? Whether it is correct to combine them into one family is a question that does not belong here; we are now only concerned with the fact that the two stages are related in form in very different degrees.

An especially striking case of incongruence is offered by the family Notodontidæ, under which Boisduval, depending only on imaginal characters, united genera of which the larvæ differed to a very great extent. In O. Wilde’s work on caterpillars this family is on this account quite correctly characterized as follows:—“Larvæ of various forms, naked or with thin hairs, sixteen or fourteen legs.”[185] In fact in the whole order Lepidoptera there can scarcely be found associated together such diverse larvæ as are here placed in one imago-family; on one side the short cylindrical caterpillars of the genus Cnethocampa, Steph. (C. Processionea, Pithyocampa, &c.), which are covered with fine, brittle, hooked hairs, and are very similar to the larvæ of Gastropacha with which they were formerly united; and on the other side there are the naked, humped, and flat-headed larvæ of the genus Harpyia, Ochs., with their two long forked appendages replacing the hindmost pair of legs, and the grotesquely formed caterpillars of the genera Stauropus, Germ., Hybocampa, Linn., and Notodonta, Ochs.

The morphological congruence between larvæ and imagines declares itself most sharply in genera, where it is the rule almost without exception. In this case we can indeed be sure that a genus or sub-genus founded on the imagines only will, in accordance with correct principles, present a corresponding difference in the larvæ. Had the latter been known first we should have been led to construct the same genera as those which are now established on the structure of the imagines, and these, through other circumstances, would have stood in the same degree of morphological relationship as the genera founded on the imagines. There is therefore a congruence in a double sense; in the first place the differences between the larvæ and imagines of any two genera are equally great, and, in the next place, the common characters possessed by these two stages combined cause them to form precisely the same groups defined with equal sharpness; the genera coincide completely.

So also the butterflies of the sub-family Nymphalinæ can well be separated into genera by the characters of the larvæ, and these, as far as I am able to judge, would agree with the genera founded on the imagines.

The genus Melitæa, for example, can be characterized by the possession of 7–9 fleshy tubercles bearing hairy spines; the genus Argynnis may be distinguished by always having six hairy unbranched spines on each segment, and the genus Cethosia by two similar spines on each segment; the genus Vanessa shows sometimes as many as seven branched spines; and the genus Limenitis never more than two branched blunt spines on each segment, and so forth. If we go further into details it will be seen that the most closely related imagines, as might indeed have been expected, likewise possess the most nearly allied larvæ, whilst very small differences between the imagines are also generally represented by corresponding differences in the larvæ. Thus, for instance, the genus Vanessa of Fabricius has been divided into several genera by later authors. Of these sub-genera, Grapta, Doubl. (containing the European C.-album, the American Fabricii, Interrogationis, Faunus, Comma, &c.), is distinguished by the fact that the larvæ not only possess branched spines on all the segments with the exception of the prothorax, but these spines are also present on the head; in the genus Vanessa (sensû strictiori), Doubl., the head and prothorax are spineless (e.g. V. Urticæ); in the tropical genus Junonia, Hübn., which was also formerly (Godart, 1819[186]) united with Vanessa, the larvæ bear branched spines on all the segments, the head and prothorax included.

It is possible to go still further and to separate two species of Vanessa as two new genera, although they have hitherto been preserved from this fate even by the systematists most given to “splitting.” This decision is certainly justifiable, simply because these species at present stand quite alone, and the practical necessity of forming a distinct genus does not make itself felt, and this practical necessity moreover frequently comes into conflict with scientific claims: science erects a new genus based on the amount of morphological difference, it being quite immaterial whether one or many species make up this genus; such an excessive subdivision is, however, a hindrance to practical requirements, as the cumbrous array of names thereby becomes still further augmented.

The two species which I might separate from Vanessa on the ground of their greater divergence, are the very common and widely distributed V. Io and Antiopa, the Peacock Butterfly and the Camberwell Beauty. In the very remarkable pattern of their wings, both show most marked characteristics; Io possesses a large ocellus on each wing, and Antiopa has a broad light yellow border which is not found in any other species of Vanessa. There can be no doubt but that each of these would have been long ago raised into a genus if similarly marked species of Vanessa occurred in other parts of the world, as is the case with the other species of the genus. Thus, it is well known that there is a whole series of species resembling our V. Cardui, and another series resembling our V. C.-album, the two series possessing the same respective types of marking; indeed on these grounds the sub-genera Pyrameis and Grapta have been erected.[187]

I should not have considered it worth while to have made these remarks if it had not been for the fact that the caterpillars of V. Io and V. Antiopa differ in small particulars from one another and from the other species of the genus. These differences relate to the number and position of the spines, as can be seen from the following table:—

Species of the Genus Vanessa, Fabr.

	Number of Spines on the head and segments of the larva.
	Head.	Segm. I.	Segm. II.	Segm. III.	Segm. IV.	Segm. V.	Segm. VI.–XI.	Segm. XII.
V. Io	0	0	2	2	4	6	6	4
V. Antiopa	0	0	4	4	6	6	7	4
V. Urticæ	0	0	4	4	7	7	7	4
V. Polychloros	0	0	4	4	7	7	7	4
V. Ichnusa	0	0	4	4	7	7	7	4
V. Atalanta	0	0	4	4	7	7	7	4
V. C.-album	2	0	4	4	7	7	7	4
V. Interrogationis	2	0	4	4	7	7	7	4
V. Levana	2	0	4	4	7	7	7	4

This character of the number of spines will not be considered as too unimportant when we observe how perfectly constant it remains in the nearly allied species. This is the case in the three consecutive forms, Urticæ, Polychloros, and Ichnusa. Now when we see that two species which differ in their imaginal characters present correspondingly small differences in their larvæ, this exact systematic congruence indicates a completely parallel phyletic development.

Exceptions are, however, to be met with here. Thus, Hübner has united one group of the species of Vanessa into the genus Pyrameis just mentioned, on account of certain characteristic distinctions of the butterflies. I do not know, however, how this genus admits of being grounded on the structure of the larvæ; the latter, as appears from the above table, agree exactly in the number and position of the spines with the caterpillars of Vanessa (sensû strictiori), nor can any common form of marking be detected which would enable them to be separated from Vanessa.

Still more striking is the incongruence in the genus Araschnia, Hübn. (A. Prorsa-Levana), which, like the genus Pyrameis, is entirely based on imaginal characters. This is distinguished from all the other sub-genera of the old genus Vanessa by a small difference in the venation of the wings (the discoidal cell of the hind-wings is open instead of closed). Now it is well-known that in butterflies the wing-venation, as most correctly shown by Herrich-Schäffer, is the safest criterion of “relationship.” It thus happens that this genus, typified by the common Levana, is in Kirby’s Catalogue separated from Vanessa by two genera, and according to Herrich-Schäffer[188] by forty genera! Nevertheless, the larvæ agree so exactly in their spinal formula with Grapta that we should have no hesitation in regarding them as a species of this sub-genus. It appears to me very probable that in this case the form-relationship of the caterpillar gives more correct information as to the blood-relationship of the species than that of the imago—in any case the larvæ show a different form-relationship to the imagines.

Just as in the case of butterflies there are many genera of Sphingidæ which can be based on the structure of the larvæ, and which agree with those founded on the imagines.

Thus, the genus Macroglossa is characterized by a straight anal horn, a spherical head, and by a marking composed of longitudinal stripes, these characters not occurring elsewhere in this combination. The nearly allied genus Pterogon, on the other hand, cannot be based on the larvæ only, since not only is the marking of the adult larva very distinct in the different species, but the anal horn is present in two species, whilst in a third (P. Œnotheræ) it is replaced by a knob-like eye-spot. The genus Sphinx (sensû strictiori) is distinguished by the simple, curved caudal horn, the smooth, egg-shaped head and smooth skin, and by a marking mainly composed of seven oblique stripes. The genus Deilephila is distinguished from the preceding by a dorsal plate, situated on the prothorax and interrupting the marking, as well as by the pattern, which here consists of a subdorsal line with ring-spots more or less numerous and developed; the skin also is rough, “shagreened,” although it must be admitted that there are exceptions (Vespertilio). The genus Chærocampa admits also of being based on the form-relationship of its caterpillars, although this is certainly only possible by disregarding the marking and taking alone into consideration the peculiar pig-like form of the larvæ. The genus Acherontia, so nearly related to Sphinx, possesses in the doubly curved caudal horn a character common to the genus (three species known[189]). Finally may be mentioned the genus Smerinthus, of which the larvæ, by their anteriorly tapering form, their shagreened skin and almost triangular head with the apex upwards, their simply curved anal horn, and by their seven oblique stripes on each side, constitute a genus as sharply defined as that formed by the moths.

Although in all the systematic divisions hitherto treated of there are cases where the form-relationship of the larva does not completely coincide with that of the imago, such incongruences are of far more frequent occurrence in the smallest systematic group, viz. species.

The larvæ of two species have very frequently a much nearer form-relationship than their imagines. Thus, the caterpillars of Smerinthus Ocellatus and S. Populi are closely allied in structure, marking, and colouring, whilst the moths in these two last characters and in the form of the wings are widely separated.[190] Judging from the larvæ we should expect to obtain two very similar moths, but in fact both Populi and Ocellatus have many near allies, and these closely related species sometimes possess larvæ which differ more widely than those of more distantly related species of imagines.

Thus, in Amur-land and North America there occur species of Smerinthus which closely resemble our Ocellatus in colour, marking, and form of wing, and which possess the characteristic large blue ocellus on the hind-wings. S. Excæcatus is quite correctly regarded as the representative American form of our Ocellatus, but its caterpillar, instead of being leaf-green, is of a chrome-yellow, and possesses dark green instead of white oblique stripes, and has moreover a number of red spots, and a red band on the head—in brief, in the very characters (colour and certain of the markings) in which the imagines completely agree it is widely different from Ocellatus. It appears also to be covered with short bristles, judging from Abbot and Smith’s figure.[191]

Just in the same way that the species having the nearest conceivable form-relationship to Ocellatus possesses a relatively strongly diverging larva, so does the nearest form-relation of Populi (imago) offer a parallel case. This species, which is also North American, lives on Juglans Alba. The imago of Smerinthus Juglandis differs considerably from S. Populi in the form of the wings, but it resembles the European species so closely in marking and colouring that no doubt can exist as to the near relationship of the two forms. The caterpillar of S. Juglandis,[192] however, differs to a great extent from that of Populi in colour—it is not possible to confound these two larvæ; but those of Populi and Ocellatus are not only easily mistaken for one another, but are distinguished with difficulty even by experts.

In this same family of the Sphingidæ cases are not wanting in which, on the other hand, the moths are far more closely allied than the larvæ. This is especially striking in the genus Deilephila, eight species of which are allied in the imaginal state in a remarkable degree, whilst the larvæ differ greatly from one another in colour, and to as great an extent in marking. These eight species are D. Nicæa, Euphorbiæ, Dahlii, Galii, Livornica, Lineata, Zygophylli, and Hippophaës. Of these, Nicæa, Euphorbiæ, Dahlii, Zygophylli, and Hippophaës are so much alike in their whole structure, in the form of the wings, and in marking, that few entomologists can correctly identify them off-hand without comparison. The larvæ of these four species, however, are of very different appearances. Those of Euphorbiæ and Dahlii are most alike, both being distinguished by the possession of a double row of large ring-spots. Zygophylli (see Fig. 50, [Pl. VI].) possesses only faint indications of ring-spots on a white subdorsal line; and in Hippophaës there is only an orange-red spot on the eleventh segment, the entire marking consisting of a subdorsal line on which, in some individuals, there are situated more or less developed ring-spots (see Figs. 59 and 60, [Pl. VII].). If we only compare the larvæ and imagines of D. Euphorbiæ and Hippophaës, we cannot but be struck with astonishment at the great difference of form-relationship in the two stages of development.

In the case of D. Euphorbiæ and Nicæa this difference is almost greater. Whilst these larvæ show great differences in colour, marking, and in the roughness or smoothness of the skin (compare Fig. 51, [Pl. VI]. with Figs. 43 and 44, [Pl. V].), the moths cannot be distinguished with certainty. As has already been stated, the imago of the rare D. Nicæa is for this reason wanting in most collections; it cannot be detected whether a specimen is genuine, i.e. whether it may not perhaps be a somewhat large example of D. Euphorbiæ.

An especially striking instance of incongruence is offered by the two species of Chærocampa most common with us, viz., Elpenor and Porcellus, the large and small Elephant Hawk-moths. The larvæ are so similar, even in the smallest details of marking, that they could scarcely be identified with certainty were it not that one species (Elpenor) is considerably larger and possesses a less curved caudal horn than the other. The moths of these two species much resemble one another in their dull green and red colours, but differ in the arrangement of these colours, i.e. in marking, and also in the form of their wings, to such an extent that Porcellus has been referred to the genus Pergesa[193] of Walker. If systemy, as is admitted on many sides, has only to indicate the morphological relationship, this author is not to blame—but in this case a special larval classification must likewise be admitted, in a manner somewhat similar to that at present adopted provisionally in text-books of zoology for the Hydroid Polypes and inferior Medusæ. This case of Porcellus, however, shows that those are correct who maintain that systemy claims to express, although incompletely, the blood-relationship, and that systematists have always unconsciously formed their groups as though they intended to express the genetic connection of the forms. Only on this supposition can it appear incorrect to us to thus separate two species of which the larvæ agree so completely.

I cannot conclude this review of the various systematic groups without taking a glance at the groups comprised within species, viz. varieties. Whilst in species incongruence is of frequent occurrence, in varieties this is the rule, for which reason it admits in this case of being more sharply defined, since we are not concerned with a double difference but only with the question whether in the one stage a difference or an absolute similarity is observable. By far the majority of varieties are either simply imaginal or merely larval varieties—only the one stage diverges, the other is quite constant.

Thus, as has already been shown, in all the seasonally dimorphic butterflies known to me the caterpillars of the two generations of imagines, which are often so widely different, are exactly alike; and the same obtains for the majority of purely climatic varieties of butterflies. Unfortunately there are as yet no connected observations on this point. The only certain instance that I can here mention is that of the Alpine and Polar form of Pteris Napi. This variety, Bryoniæ, the female of which differs so greatly in marking and colouring, possesses larvæ which cannot be distinguished from those of the ordinary form of Napi.(See part I. appendix I. p. [124].)

That caterpillars can also vary locally without thereby affecting the imagines is shown by the frequently mentioned and closely investigated cases of di- and polymorphism in the larvæ of a number of Sphingidæ (M. Stellatarum, A. Atropos, S. Convolvuli, C. Elpenor, and Porcellus, &c.). The same thing is still more clearly shown by those instances in which there are not several but only one distinct larval form occurring in each of two different localities.

To this class belongs the above-mentioned case of Chærocampa Celerio (p. [197]), supposing our information concerning this species to be correct; likewise the recently-mentioned case of the Ligurian variety of the caterpillar of Saturnia Carpini; and finally the case of Eriogaster Lanestris, so well known to lepidopterists. This insect inhabits the plains of Germany, and in the Alps extends to an elevation of 7000 feet, where it possesses a larva differently marked and coloured (E. Arbusculæ) to those of the lowlands whilst the moths are smaller, but do not differ in other respects from those of the plains.

Among the Alpine species many other such cases may occur, but these could only be discovered by making investigations having special reference to this point. Of the Alpine butterflies, for example, not a single species can have been reared from the caterpillar; for this reason but few observations have on the whole been given by entomologists respecting the Alpine larvæ, which are not known sufficiently well to enable such a question to be decided.

The investigation of the form-relationships existing between larvæ on the one hand and imagines on the other has thus led to the following results:—

We learn on comparison that incongruences or inequalities of form-relationship occur in all systematic groups from varieties to families. These incongruences are of two kinds, in some cases being disclosed by the fact that the larvæ of two systematic groups, e.g. two species, are more closely related in form than their imagines (or inversely), whilst in other cases the larvæ form different systematic groups to those formed by the imagines.

The results of the investigation into the occurrence of incongruences among the various systematic groups may be thus briefly summarised:—

Incongruences appear to occur most frequently among varieties, since it very frequently happens that it is only the larva or only the imago which has diverged into a variety, the other stage remaining monomorphic. The systematic division of varieties is thus very often one-sided.

Among species also incongruences are of frequent occurrence. Sometimes the imagines are much more nearly related in form than the larvæ, and at others the reverse happens; whilst again the case appears also to occur in which only the one stage (larva) diverges to the extent of specific difference, the other stage remaining monomorphic (D. Euphorbiæ and Nicæa).

The agreement in form-relationship appears to be most complete in genera. In the greater number of cases the larval and imaginal genera coincide, not only in the sharpness of their limits, but also—as far as one can judge—in the weight of their distinctive characters, and therefore in the amount of their divergence. Of all the systematic groups, genera show the greatest congruence.

In families there is again an increase of irregularity. Although larval and imaginal families generally agree, there are so many exceptions that the groups would be smaller if they were based exclusively on the larval structure than if founded entirely on the imagines (Nymphalidæ, Bombycidæ).

If we turn to the groups of families we find a considerably increased incongruence; complete agreement is here again rather the exception, and it further happens in these cases that it is always the larvæ which, to a certain extent, remain at a lower grade, and which form well defined families; but these can seldom be associated into groups of a higher order having a common character, as in the case of the imagines (Rhopalocera).

After having thus collected (so far as I am able) the facts, we have now to attempt their interpretation, and from the observed congruence and incongruence of form-relationship of the two stages to endeavour to draw a conclusion as to the underlying causes of the transformations.

It is clear at starting that all cases of incongruence can only be the expression or the consequence of a phyletic development which has not been exactly parallel in the two stages of larva and imago—that one stage must have changed either more rapidly or more slowly than the other. An “unequal phyletic development” is thus the immediate cause of incongruence.

Thus, the occurrence of different larvæ in species of which the imagines have remained alike may be simply understood as cases in which the imago only has experienced a change—has taken a forward step in phyletic development, whilst the larvæ have remained behind. If we conceive this one-sided development to be repeated several times, there would arise two larval forms as widely different as those of Deilephila Nicæa, and Euphorbiæ, whilst the imagines, as is actually the case in these species, would remain the same.

The more commonly occurring case in which one stage has a greater form-divergence than the other, is explicable by the one stage having changed more frequently or more strongly than the other.

The explanation of the phenomena thus far lies on the surface, and it is scarcely possible to advance any other; but why should one stage become changed more frequently or to a greater extent than the other? why should one portion be induced to change more frequently or more strongly than another? whence come these inducements to change? These questions bring us to the main point of inquiry:—Are the causes which give rise to these changes internal or external? Are the latter the result of a phyletic vital force, or are they only due to the action of the external conditions of life?

Although an answer to this question will be found in the preceding essay, I will not support myself on the results there obtained, but will endeavour to give another solution of the problem on fresh grounds. The answer will indeed be the same as before:—A phyletic force must be discountenanced, since in the first place it does not explain the phenomena, and in the second place the phenomena can be well explained without its assumption.

The admission of a phyletic vital force does not explain the phenomena. The assumption that there is a transforming power innate in the organism indeed agrees quite well with the phenomenon of congruence, but not with that of incongruence. Since a large number of cases of the latter depend upon the fact that the larvæ are more frequently influenced by causes of change than their imagines, or vice versâ, how can this be reconciled with such an internal force? On this assumption would not each stage of a species be compelled to change, if not contemporaneously at least successively, with the same frequency and intensity, by the action of an innate force? and how by means of the latter can there ever result a greater form-divergence in the larvæ than in the imagines?

It is delusive to believe that these unequal deviations can be explained by assuming that the phyletic force acts periodically. Granting that it does so, and that the internal power successively compels the imago, pupa, and finally the larva to change, there would then pass a kind of wave of transformation over the different stages of the species, as was actually shown above to be the case in the single larval stages. The only possible way of explaining the unequal distances between larvæ and imagines would therefore be to assume that two allied groups, e.g. species, were not contemporaneously affected by the wave, so that at a certain period of time the imago alone of one species had become changed, whilst in the other species the wave of transformation had also reached the larva. In this case the imagines of the two species would thus appear to be more nearly related than their larvæ.

Now this strained explanation is eminently inapplicable to varieties, still less to species, and least of all to higher systematic groups, for the simple reason that every wave of transformation may be assumed to be at the most of such strength as to produce a deviation of form equal to that of a variety. Were the change resulting from a single disturbance greater, we should not only find one-sided varieties, i.e. those belonging to one stage, but we should also meet as frequently with one-sided species. If, however, a wave of transformation can only produce a variety even in the case of greatest form-divergence, the above hypothetical uncontemporaneous action of such a wave in two species could only give rise to such small differences in the two stages that we could but designate them as varieties. An accumulation of the results of the action of several successive waves passing over the same species could not happen, because the distance from a neighbouring species would always become the same in two stages as soon as one wave had ended its course. In this manner there could therefore only arise divergences of the value of varieties, and incongruences in systematic groups of a higher rank could not thus be explained.

All explanations of the second form of incongruence from the point of view of a phyletic force can also be shown to be absurd. How can the fact be explained that larval and imaginal families by no means always coincide; or that the larvæ can only be formed into families whilst the imagines partly form sharply defined groups of a higher order? How can an internal directive force within the same organism urge in two quite distinct directions? If the evolution of a definite system were designed, and the admission of such a continually acting power rendered necessary, why such an incomplete, uncertain, and confused performance?

I must leave others to answer these questions; to me a vital force appears to be inadmissible, not only because we cannot understand the phenomena by its aid, but above all because it is superfluous for their explanation. In accordance with general principles the assumption of an unknown force can, however, only be made when it is indispensable to the comprehension of the phenomena.

I believe that the phenomena can be quite well understood without any such assumption—both the phenomena of congruence and incongruence, in their two forms of unequal divergence and unequal group-formation.

Let us in the first place admit that there is no directive force in the organism inciting periodic change, but that every change is always the consequence of external conditions, being ultimately nothing but the reaction—the response of the organism to some of the influences proceeding from the environment; every living form would in this case remain constant so long as it was not compelled to change by inciting causes. Such transforming factors can act directly or indirectly, i.e. they can produce new changes immediately, or can bring about a remodelling by the combination, accumulation, or suppression of individual variations already present (adaptation by natural selection). Both forms of this action of external influences have long been shown to be in actual operation, so that no new assumption will be made, but only an attempt to explain the phenomena in question by the sole action of these known factors of species formation.

If, in the first instance, we fix our attention upon that form of incongruence which manifests itself through unequal divergence of form-relationship, it will appear prominently that this bears precise relations to the different systematic groups. This form of incongruence constitutes the rule in varieties of the order Lepidoptera, it is of very frequent occurrence in species, but disappears almost completely in genera, and entirely in the case of families and the higher groups. On the whole, therefore, as we turn to more and more comprehensive groups, the incongruence diminishes whilst the congruence increases, until finally the latter becomes the rule.

Now if congruence presupposes an equal number of transforming impulses, we perceive that the number of the impulses which have affected larvæ and imagines agree with one another the more closely the larger the systematic groups which are compared together. How can this be otherwise? The larger the systematic group the longer the period of time which must have been necessary for its formation, and the more numerous the transforming impulses which must have acted upon it before its formation was completed.

But if the supposition that the impulse to change always comes from the environment in no way favours the idea that such impulses always affect both stages contemporaneously, and are equal in number during the same period of time, there is not, on the other hand, the least ground for assuming that throughout long periods the larvæ or the imagines only would have been affected by such transforming influences. This could have been inferred from the fact that varieties frequently depend only upon one stage, whilst specific differences in larvæ only also occur occasionally, the imagines remaining alike; but no single genus is known of which all the species possess similar larvæ. Within the period of time during which genera can be formed the transforming impulses therefore never actually affect the one stage only, but always influence both.

But if this is the case—if within the period of time which is sufficient for the production of species, the one stage only is but seldom and quite exceptionally influenced by transforming impulses, whilst both stages are as a rule affected, although not with the same frequency, it must necessarily follow that on the whole, as the period of time increases, the difference in the number of these impulses which affect the larva and of those which affect the imago must continually decrease, and with this difference the magnitude of the morphological differences resulting from the transforming influences must at the same time also diminish. With the number of the successively increasing changes the difference in the magnitude of the change in the two stages would always relatively diminish until it had quite vanished from our perception; just in the same manner as we can distinguish a group of three grains of corn from one composed of six, but not a heap of 103 grains from one containing 106 grains.

That the small systematic groups must have required a short period and the large groups a long period of time for their formation requires no special proof, but results immediately from the theory of descent.

All the foregoing considerations would, however, only hold good if the transforming impulses were equal in strength, or, not to speak figuratively, if the changes only occurred in equivalent portions of the body, i.e. in such portions as those in which the changes are of the same physiological and morphological importance to the whole organism.

Now in the lower systematic groups this is always the case. Varieties, species, and genera are always distinguished by only relatively small differences; deep-seated distinctions do not here occur, as is implied in the conception of these categories. The true cause of this is, I believe, to be found in the circumstance that all changes take place only by the smallest steps, so that greater differences can only arise in the course of longer periods of time, within which a great number of types (species) can, however, come into existence, and these would be related by blood and in form in different degrees, and would therefore form a systematic group of a higher rank.

The short periods necessary for the production of inferior groups, such as genera, would not result in incongruences if only untypical parts of the larvæ, such as marking or spines, underwent change, whilst in the imagines typical parts—wings and legs—became transformed. The changes which could have occurred in the wings, &c., during this period of time would have been much too small to produce any considerable influence on the other parts of the body by correlation; and two species of which the larvæ and imagines, had changed with the same frequency, would show a similar amount of divergence between the larvæ and between the imagines, although on the one side only untypical parts—i.e. those of no importance to the whole organization—and on the other side typical parts, were affected. The number of the changes would here alone determine whether congruence or incongruence occurred between the two stages.

The case would be quite different if, throughout a long period of time, in the one stage only typical and in the other only untypical parts were subjected to change. In the first case a complete transformation of the whole structure would occur, since not only would the typical parts, such as the wings, undergo a much further and increasing transformation in the same direction, but these changes would also lead to secondary alterations.

In this manner, I believe, must be explained the fact that in the higher groups still greater form-divergences of the two stages occur; and if this explanation is correct, the cause of this striking phenomenon, viz., that incongruence diminishes from varieties to genera, in which latter it occurs but exceptionally, whilst in families and in the higher groups it again continually increases, is likewise revealed. Up to genera the incongruence depends entirely upon the one stage having become changed more frequently than the other; but in families and groups of families, and in the orders Diptera and Hymenoptera, as will be shown subsequently, in sub-orders and tribes, it depends upon the importance of the part of the body affected by the predominant change. In the latter case the number of changes is of no importance, because these are so numerous that the difference vanishes from our perception; but an equal number of changes, even when very great, may now produce a much greater or a much smaller transformation in the entire bodily structure according as they affect typical or untypical portions, or according as they keep in the same direction throughout a long period of time, or change their direction frequently.

Those unequal form-divergences which occur in the higher systematic groups a re always associated with a different formation of groups—the larvæ form different systematic groups to the imagines, so that one of these stages constitutes a higher or a lower group; or else the groups are of equal importance in the two stages, but are of unequal magnitude—they do not coincide, but the one overlaps the other.

Incongruences of this last kind appear in certain cases within families (Nymphalidæ), but I will not now subject these to closer analysis, because their causes will appear more clearly when subsequently considering the orders Hymenoptera and Diptera. Incongruences of the first kind, however, admit of a clear explanation in the case of butterflies. They appear most distinctly in the groups composed of families.

Nobody has as yet been able to establish the group Rhopalocera by means of any single character common to the larvæ; nevertheless, this group in the imagines is the sharpest and best defined of the whole order. If we inform the merest tyro that clubbed antennæ are the chief character of the butterflies, he will never hesitate in assigning one of these insects to its correct group. Such a typical character, common to all families, is, however, absent in the larvæ; and it might be correctly said that there were no Rhopalocerous larvæ, or rather that there were only larvæ of Equites, Nymphales, and Heliconii. The larvæ of the various families can be readily separated by means of characteristic distinctions, and it would not be difficult for an adept to distinguish to this extent in single cases a Rhopalocerous caterpillar as such; but these larvæ possess only family characters, and not those of a higher order.

This incongruence partly depends upon the circumstance that the form-divergence between a Rhopalocerous and a Heterocerous family is much greater on the side of the imagines than on that of the larvæ. Were there but a single family of butterflies in existence, such as the Equites, we should be obliged to elevate this to the rank of a sub-order on the side of the imagines, but not on that of the larvæ. Such cases actually occur, and an instance of this kind will be mentioned later in connection with the Diptera. But this alone does not explain why, on the side of the imagines, a whole series of families show the same amount of morphological divergence from the families of other groups. There are two things, therefore, which must here be explained:—First, why is the form-divergence between the imagines of the Rhopalocera and Heterocera greater than that between their larvæ? and, secondly, why can the imagines of the Rhopalocera be formed into one large group by means of common characters whilst the larvæ cannot?

The answers to both these questions can easily be given from our present standpoint. As far as the first question is concerned, this finds its solution in the fact that the form-divergence always corresponds exactly with the divergence of function, i.e. with the divergence in the mode of life.

If we compare a butterfly with a moth there can be no doubt that the difference in the conditions of life is far greater on the side of the imagines than on that of the larvæ. The differences in the mode of life of the larvæ are on the whole but very small. They are all vegetable feeders, requiring large quantities of food, and can only cease feeding during a short time, for which reason they never leave their food-plants for long, and it is of more importance for them to remain firmly attached than to be able to run rapidly. It is unnecessary for them to seek long for their food, as they generally find themselves amidst an abundance, and upon this depends the small development of their eyes and other organs of sense. On the whole caterpillars live under very uniform conditions, although these may vary in manifold details.

The greatest difference in the mode of life which occurs amongst Lepidopterous larvæ is shown by wood feeders. But even these, which by their constant exclusion from light, the hardness of their food, their confinement within narrow hard-walled galleries, and by the peculiar kind of movement necessitated by these galleries, are so differently situated in many particulars to those larvæ which live openly on plants, have not experienced any general change in the typical conformation of the body by adaptation to these conditions of life. These larvæ, which, as has already been mentioned, belong to the most diverse families, are more or less colourless and flattened, and have very strong jaws and small feet; but in none of them do we find a smaller number of segments, or any disappearance, or important transformation of the typical limbs; they all without exception possess sixteen legs, like the other larvæ excepting the Geometræ.

Now if even under the most widely diverging conditions of life adaptation of form is produced by relatively small, and to a certain extent superficial, changes, we should expect less typical transformations in the great majority of caterpillars which live on the exterior of plants or in their softer parts (most of the Micro-lepidoptera). The great diversity in the forms of caterpillars depends essentially upon a different formation of the skin and its underlying portions. The skin is sometimes naked, and can then acquire the most diverse colours, either protective or conspicuous, or it may develop offensive or defensive markings; in other cases it may be covered with hairs which sting, or with spines which prick; certain of its glands may develop to an enormous size, and acquire brilliant colours and the power of emitting stinking secretions (the tentacles of the Papilionidæ and Cuspidate larvæ); by the development of warts, angles, humps, &c., any species of caterpillar may be invested with the most grotesque shape, the significance of which with respect to the life of the insect is as yet in most cases by no means clear: typical portions are not, however, essentially influenced by these manifold variations. At most only the form of the individual segments of the body, and with these the shape of the whole insect, become changed (onisciform larvæ of Lycænidæ), but a segment is never suppressed, and even any considerable lengthening of the legs occurs but very seldom (Stauropus Fagi).[194]

We may therefore fairly assert that the structure of larvæ is on the whole remarkably uniform, in consequence of the uniformity in the conditions of life. Notwithstanding the great variety of external aspects, the general structure of caterpillars does not become changed—it is only their outward garb which varies, sometimes in one direction, and sometimes in another, and which, starting from inherited characters, becomes adapted to the various special conditions of life in the best possible manner.

All this is quite different in the case of the imagines, where we meet with very important differences in the conditions of life. The butterflies, which live under the influence of direct sunlight and a much higher temperature, and which are on the wing for a much longer period during the day, must evidently be differently equipped to the moths in their motor organs (wings), degree of hairiness, and in the development of their eyes and other organs of sense. It is true that we are not at present in a condition to furnish special proofs that the individual organs of butterflies are exactly adapted to a diurnal life, but we may safely draw this general conclusion from the circumstance that no butterfly is of nocturnal habits.[195] It cannot be stated in objection that there are many moths which fly by day. It certainly appears that no great structural change is necessary to confer upon a Lepidopteron organized for nocturnal life the power of also flying by day; but this proves nothing against the view that the structure of the butterflies depends upon adaptation to a diurnal life. Analogous cases are known to occur in many other groups of animals. Thus, the decapodous Crustacea are obviously organized for an aquatic life; but there are some crabs which take long journeys by land. Fish appear no less to be exclusively adapted to live in water; nevertheless the “climbing-perch” (Anabas) can live for hours on land.

It is not the circumstance that some of the moths fly by day which is extraordinary and demands a special explanation, but the reverse fact just mentioned, that no known butterfly flies by night. We may conclude from this that the organization of the latter is not adapted to a nocturnal life.

If we assume[196] that the Lepidopterous family adapted to a diurnal life gives rise in the course of time to a nocturnal family, there can be no doubt but that the transformation of structure would be far greater on the part of the imagines than on that of the larvæ. The latter would not remain quite unchanged—not because their imagines had taken to a nocturnal life which for the larva would be quite immaterial, but because this change could only occur very gradually in the course of a large number of generations, and during this long period the conditions of life would necessarily often change with respect to the larvæ. It has been shown above that within the period of time necessary for the formation of a new species impulses to change occur on both sides; how much more numerous therefore must these be in the case of a group of much higher rank, for the establishment of which a considerably longer period is required. In the case assumed, therefore, the larvæ would also change, but they would suffer much smaller transformations than the imagines. Whilst in the latter almost all the typical portions of the body would undergo deep changes in consequence of the entirely different conditions of life, the larvæ would perhaps only change in marking, hairs, bristles, or other external characters, the typical parts experiencing only unimportant modifications.

In this manner it can easily be understood why the larvæ of a family of Noctuæ do not differ to a greater extent from those of a family of butterflies than do the latter from some other Rhopalocerous family, or why the imagines of a Rhopalocerous and a Heterocerous family present much greater form-divergences than their larvæ. At the same time is therefore explained the unequal value that must be attributed to any single family of butterflies in its larvæ and in its imagines. The unequal form-divergences coincide exactly with the inequalities in the conditions of life.

When whole families of butterflies show the same structure in their typical parts (antennæ, wings, &c.), and, what is of more importance, can be separated as a systematic group of a higher order (i.e. as a section or sub-order) from the other Lepidoptera whilst their larval families do not appear to be connected by any common character, the cause of this incongruence lies simply in the circumstance that the imagines live under some peculiar conditions which are common to them all, but which do not recur in other Lepidopterous groups. Their larvæ live in precisely the same manner as those of all the other families of Lepidoptera—they do not differ in their mode of life from those of the Heterocerous families to a greater extent than they do from one another.

We therefore see here a community of form within the same compass as that in which there is community in the conditions of life. In all butterflies such community is found in their diurnal habits, and in accordance with this we find that these only, and not their larvæ, can be formed into a group having common characters.

In the larvæ also we only find agreement in the conditions of life within a much wider compass, viz. within the whole order. Between the limits of the order Lepidoptera the conditions of life in the caterpillars are, as has just been shown, on the whole very uniform, and the structure of the larvæ accordingly agrees almost exactly in all Lepidopterous families in every essential, i.e. typical, part.

In this way is explained the hitherto incomprehensible phenomenon that the sub-ordinal group Rhopalocera cannot be based on the larvæ, but that Lepidopterous caterpillars can as a whole be associated into a higher group (order); they constitute altogether families and an order, but not the intermediate group of a sub-order. By this means we at the same time reply to an objection that may be raised, viz. that larval forms cannot be formed into high systematic groups because of their “low and undeveloped” organization.

To this form of incongruence, viz. to the formation of systematic groups of unequal value and magnitude, I must attach the greatest weight with respect to theoretical considerations. I maintain that this, as I have already briefly indicated above, is wholly incompatible with the admission of a phyletic force. How is it conceivable that such a power could work in the same organism in two entirely different directions—that it should in the same species lead to the constitution of quite different systems for the larvæ and for the imagines, or that it should lead only to the formation of families in the larvæ and to sub-orders in the imagines? If an internal force existed which had a tendency to call into existence certain groups of animal forms of such a nature that these constituted one harmonious whole of which the components bore to one another fixed morphological relationships, it would certainly have been an easy matter for such a power to have given to the larvæ of butterflies some small character which would have distinguished them as such, and which would in some measure have impressed them with the stamp of “Rhopalocera.” Of such a character we find no trace however; on the contrary, everything goes to show that the transformations of the organic world result entirely from external influences.

---

III.
Incongruences in other Orders of Insects.

Although the order Lepidoptera is for many reasons especially favourable for an investigation such as that undertaken in the previous section, it will nevertheless be advantageous to inquire into the form-relationships of the two chief stages in some other orders of metamorphic insects, and to investigate whether in these cases the formation of systematic groups also coincides with common conditions of life.

Hymenoptera.

In this order there cannot be the least doubt as to the form-relationship of the imagines. The characteristic combination of the pro- and meso-thorax, the number and venation of the wings, and the mouth-organs formed for biting and licking, are found throughout the whole order, and leave no doubt that the Hymenoptera are well based on their imaginal characters.

But it is quite different with the larvæ. It may be boldly asserted that the order would never have been founded if the larvæ only had been known. Two distinct larval types here occur, the one—caterpillar-like—possessing a distinct horny head provided with the typical masticatory organs of insects, and a body having thirteen segments, to which, in addition to a variable number of abdominal legs, there are always attached three pairs of horny thoracic legs: the other type is maggot-shaped, without the horny head, and is entirely destitute of mouth-organs, or at least of the three pairs of typical insect jaws, and is also without abdominal and thoracic legs. The number of segments is extremely variable; the larvæ of the saw-flies have thirteen besides the head, the maggot-shaped larvæ of bees possess fourteen segments altogether, and the gall-flies and ichneumons only twelve or ten. We should be much mistaken also if we expected to find connecting characters in the internal organs. The intestine is quite different in the two types of larvæ, the posterior opening being absent in the maggot-like grubs; at most only the tracheal and nervous systems show a certain agreement, but this is not complete.

The order Hymenoptera, precisely speaking and conceived only morphologically, exists therefore but in the imagines; in the larvæ there exist only the caterpillar- and maggot-formed groups. The former shows a great resemblance to Lepidopterous larvæ, and in the absence of all knowledge of the further development it might be attempted to unite them with these into one group. The two certainly differ in certain details of structure in the mouth-organs and in the number of segments, abdominal legs, &c., to a sufficient extent to warrant their being considered as two sub-orders of one larval order; but they would in any case be regarded as much more nearly related in form than the caterpillar- and maggot-like types of the Hymenopterous larvæ.

Is it not conceivable, however, that the imagines of the Hymenoptera—that ichneumons and wasps may be only accidentally alike, and that they have in fact arisen from quite distinct ancestral forms, the one having proceeded with the Lepidopterous caterpillars from one root, and the other with the grub-like Dipterous larvæ from another root?

This is certainly not the case; the common characters are too deep-seated to allow the supposition that the resemblance is here only superficial. From the structure of the imagines alone the common origin of all the Hymenoptera may be inferred with great probability. This would be raised into a certainty if we could demonstrate the phyletic development of the maggot-formed out of the caterpillar-formed Hymenopterous larvæ by means of the ontogeny of the former. From the beautiful investigations of Bütschli on the embryonic development of bees[197] we know that the embryo of the grub possesses a complete head, consisting of four segments and provided with the three typical pairs of jaws. These head segments do not subsequently become formed into a true horny head, but shrivel up; whilst the jaws disappear with the exception of the first pair, which are retained in the form of soft processes with small horny points. We know also that from the three foremost segments of the embryo the three typical pairs of legs are developed in the form of round buds, just as they first appear in all insects.[198] These rudimentary limbs undergo complete degeneration before the birth of the larva, as also do those of the whole[199] of the remaining segments, which, even in this primitive condition, show a small difference to the three foremost rudimentary legs.

The grub-like larvæ of the Hymenoptera have therefore descended from forms which possessed a horny head with antennæ and three pairs of gnathites and a 13-segmented body, of which the three foremost segments were provided with legs differing somewhat from those of the other segments; that is to say, they have descended from larvæ which possessed a structure generally similar to that of the existing saw-fly larvæ. The common derivation of all the Hymenoptera from one source is thus established with certainty.[200]

But upon what does this great inequality in the form-relationship of the larvæ and imagines depend? The existing maggot-like grubs are without doubt much further removed from the active caterpillar-like larvæ than are the saw-flies from the Aculeate Hymenoptera. Whilst these two groups differ only through various modifications of the typical parts (limbs, &c.), their larvæ are separable by much deeper-seated distinctions; limbs of typical importance entirely vanish in the one group, but in the other attain to complete development.

In the Hymenoptera there exists therefore a very considerable incongruence in the systems based morphologically, i.e. on the pure form-relationships of the larvæ and of the imagines. The reason of this is not difficult to find: the conditions of life differ much less in the case of the imagines than in that of the larvæ. In the former the conditions of life are similar in their broad features. Hymenoptera live chiefly in the air and fly by day, and in their mode of obtaining food do not present any considerable differences. Their larvæ, on the other hand, live under almost diametrically opposite conditions. Those of the saw-flies live after the manner of caterpillars upon or in plants, in both cases their peculiar locomotion being adapted for the acquisition and their masticatory organs for the reduction of food. The larvæ of the other Hymenoptera, however, do not as a rule require any means of locomotion for reaching nor any organs of mastication for swallowing their food, since they are fed in cells, like the bees and wasps, or grow up in plant galls of which they suck the juice, or are parasitic on other insects by whose blood they are nourished. We can readily comprehend that in the whole of this last group the legs should disappear, that the jaws should likewise vanish or should become diminished to one pair retained in a much reduced condition, that the horny casing of the head, the surface of attachment of the muscles of the jaws, should consequently be lost, and that even the segments of the head itself should become more or less shrivelled up as the organs of sense therein located became suppressed.

The incongruence manifests itself however in yet another manner than by the relatively greater morphological divergence of the larvæ: a different grouping is possible for the larvæ and for the imagines. If we divide the Hymenoptera simply according to the form-relationships of the imagines, the old division into the two sub-orders Terebrantia or Ditrocha and Aculeata or Monotrocha will be the most correct. The distinguishing characters of a sting or ovipositor and a one- or two-jointed trochanter are still of the greatest value. But these two sub-orders do not by any means correspond with the two types of larvæ since, in the Terebrantia, there occur families with both caterpillar-formed and maggot-formed larvæ.

The cause is to be found in that a portion of these families possess larvæ which are parasitic in other insects or in galls, their bodily structure having by these means become transformed in a quite different direction. The mode of life of the imagines is, on the other hand, essentially the same.

We have here therefore another case like that which we met with among the Rhopalocerous Lepidoptera, in which the imagines appear to be capable of being formed into a higher group than the larvæ, because the former live under conditions of life which are on the whole similar whilst the latter live under very divergent conditions.

The old division of the Hymenoptera into two sub-orders has certainly been abandoned in the later zoological text-books; they are now divided into three:—saw-flies, parasitic, and aculeate Hymenoptera; but even this arrangement has been adopted with reference to the different structure of the larvæ. Whether this system is better than the older, i.e. whether it better expresses the genealogical relationship, I will not now stop to investigate.[201]

DIPTERA.

The imagines of the Diptera (genuina), with the exception of the Aphaniptera and Pupipara, agree in all their chief characters, such as the number and structure of the wings, the number and joints of the legs, the peculiar formation of the thorax (fusion of the three segments);[202] and even the structure of the mouth organs varies only within narrow limits. This is in accordance with their mode of life, which is very uniform in its main features: all the true Diptera live in the light, moving chiefly by means of flight, but having also the power of running; all those which take food in the imago condition feed upon fluids. Their larvæ, on the other hand, are formed on two essentially different types, the one—which I shall designate as the gnat-type—possessing a horny head with eyes, three pairs of jaws, and long or short antennæ, together with a 12- or 13-segmented body, which is never provided with the three typical pairs of thoracic legs, but frequently has the so-called abdominal legs on the first and last segments. The other Dipterous larvæ are maggot-shaped and without a horny head, or in fact without any head, since the first segment, the homologue of the head, can in no case be distinguished through its being larger than the others; it is on the contrary much smaller. The typical insect mouth-parts are entirely absent, being replaced by a variously formed and quite peculiar arrangement of hooks situated on the mouth and capable of protrusion. Never more than eleven segments are present besides the first, which is destitute of eyes; neither are abdominal legs ever developed.

The mode of life differs very considerably in the two groups of larvæ. Although the Dipterous maggots are not as a rule quite incapable of locomotion like the grubs of the Hymenoptera (bees, ichneumons), the majority are nevertheless possessed of but little power of movement in the food-substance on which they were deposited as eggs. They do not go in search of food, either because they are parasitic in other insects in the same manner as the ichneumons (Tachina), or else they live on decaying animal or vegetable substances or amidst large swarms of their prey, like the larvæ of the Syrphidæ amongst Aphides. They generally undergo pupation in the same place as that which they inhabit as larvæ and indeed in their larval skin which hardens into an oval pupa-case. Some few leave their feeding place and pupate after traversing a short distance (Eristalis).

As in the case of the Hymenoptera the structure of the larvæ can here also be explained by peculiarities in their mode of life. Creatures which live in a mass of food neither require special organs of locomotion nor specially developed organs of sense (eyes). They have no use for the three pairs of jaws since they only feed on liquid substances, and the hooks within the mouth do not serve for the reduction of food but only for fastening the whole body. With the jaws and their muscular system there likewise disappears the necessity for a hard surface of attachment, i.e. a corneous head.

The mode of life of the larvæ of the gnat-type is quite different in most points. The majority, and indeed the most typically formed of these, have to go in search of their food, whether they are predaceous, such as the Culicidæ and many of the other Nemocera (Corethra, Simulium), or whether they feed on plants, which they in some cases weave into a protective dwelling tube (certain species of Chironomus). Many live in water and move with great rapidity; others bury in the earth or in vegetable substances; and even those species which live on fungi sometimes wander great distances, as in the well-known case of the “army worm” where thousands of the larvæ of Sciara Thomæ thus migrate.

Now the two types of larvæ correspond generally with the two large groups into which, as it appears to me correctly, the Diptera (genuina) are as a rule divided. In this respect there is therefore an equality of form-relationship—the grouping is the same, and the incongruence depends only upon the form-divergence between the two kinds of larvæ being greater than between the two kinds of imagines.[203]

That the form-divergence is greater in the larvæ than in the imagines cannot be doubted; that this distant form-relationship cannot, however, be referred to a very remote common origin, i.e. to a very remote blood-relationship, not only appears from the existence of transition-forms between the two sub-orders, but can be demonstrated here, as in the case of the Hymenoptera, by the embryonic development of the maggot-like larvæ.

Seventeen years ago I showed[204] that the grub-formed larvæ of the Muscidæ in the embryonic state possessed a well-developed head with antennæ and three pairs of jaws, but that later in the course of the embryonic development a marked reduction and transformation of these parts takes place, so that finally the four head segments appear as a single small ring formed from the coalesced pairs of maxillæ, whilst the so-called “fore-head” (the first head segment), together with the mandibles, becomes transformed into a suctorial-head armed with hooks and lying within the body. At the time of writing I drew no conclusion from these facts with reference to the phyletic development of these larval forms; nor did Bütschli, six years later, in the precisely analogous case of the larvæ of the bees. The inference is, however, so obvious that it is astonishing that it should not have been drawn till the present time.[205]

There can be no doubt that the maggot-like larvæ of insects are not by any means ancient forms, but are, on the contrary, quite recent, as first pointed out by Fritz Müller,[206] and afterwards by Packard[207] and Brauer,[208] and as is maintained in the latest work by Paul Mayer[209] on the phylogeny of insects.

The Dipterous maggots have evidently descended from a larval form which possessed a horny head with antennæ and three pairs of jaws, but which had no appendages to the abdominal segments; they are therefore ordinary Dipterous larvæ of the gnat-type which have become modified in a quite peculiar manner and adapted to a new mode of life, just as the grubs of the Hymenoptera are larvæ of the saw-fly type, which have become similarly transformed, although by no means in the same manner. The resemblance between the two types of larva is to a great extent purely external, and depends upon the process designated “convergence” by Oscar Schmidt, i.e. upon the adaptation of heterogeneous animal forms to similar conditions of life. By adaptation to a life within a mass of fluid nutriment, the caterpillar-formed larvæ of the Hymenoptera and the Tipula-like larvæ of the Diptera have acquired a similar external appearance, and many similarities in internal structure, or, in brief, have attained to a considerable degree of form-relationship, which would certainly have tended to conceal the wide divergence in blood-relationship did not the embryological forms on the one side and the imagines on the other provide us with an explanation.

It is certainly of great interest that in another order of insects—the Coleoptera—grub-formed larvæ occur quite irregularly, and their origin can be here traced to precisely the same conditions of life as those which have produced the grubs of bees. I refer to the honey-devouring larvæ of the Meloïdæ (Meloë, Sitaris, Cantharis). The case is the more instructive, inasmuch that the six-legged larval form is not yet relegated to the development within the egg, but is retained in the first larval stage. In the second larval stage the maggot-form is first assumed, although this is certainly not so well pronounced as in the Diptera or Hymenoptera, as neither the head nor the thoracic legs are so completely suppressed as in these orders. Nevertheless, these parts have made a great advance in the process of transformation.

The grub-like larvæ of the Hymenoptera and Diptera appear to me especially instructive with reference to the main question of the causes of transformation. The reply to the questions: what gives the impetus to change? is this impetus internal or external? can scarcely be given with greater clearness than here. If these larvæ have abandoned their ancestral form and have acquired a widely divergent structure, arising not only from suppression but partly also from an essentially new differentiation (suctorial head of the Muscidæ), and if these structural changes show a close adaptation to the existing conditions of life, from these considerations alone it is difficult to conceive how such transformations can depend upon the action of a phyletic force. The latter must have foreseen that at precisely this or that fixed period of time the ancestors of these larvæ would have been placed under conditions of life which would make it desirable for them to be modified into the maggot-type. But if at the same time the imagines are removed in a less degree from those of the caterpillar-like larvæ, this divergence being in exact relation with the deviations in the conditions of life, I at least fail to see how we can escape the consequence that it is the external conditions of life which produce the transformations and induce the organism to change. It is to me incomprehensible how one and the same vital force can in the same individual induce one stage to become transformed feebly and the other stage strongly, these transformations corresponding in extent with the stronger or weaker deviations in the conditions of life to which the organism is exposed in the two stages; to say nothing of the fact that by such unequal divergences the idea of a perfect system (creative thought) is completely upset.

Nor can the objection be raised that we are here only concerned with insignificant changes—with nothing more than the arrested development of single organs and so forth, in brief, only with those changes which can be ascribed to the action of the environment.

We are here as little concerned with a mere suppression of organs through arrested development as in the case of the Cirripedia; the transformation and reconstruction of the whole body goes even much further than in these Crustacea, although not so conspicuous externally. Where do we elsewhere find insects having the head inside a cavity of the body (sectorial head of the Muscidæ), and of which the foremost segment—the physiological representative of the head—consists entirely of the coalesced antennæ and pairs of maxillæ?

The incongruences in the form-relationships are, however, exceedingly numerous in the case of the Diptera, and a special treatise would be necessary to discuss them thoroughly. I may here mention only one case, because the inequality shows itself in this instance in a quite opposite sense.

Gerstäcker, who is certainly a competent entomologist, divides the Diptera into three tribes, viz. the Diptera genuina, the Pupipara, and the Aphaniptera. The latter, the fleas, possess in their divided thoracic segments and in their jointed labial appendages characters so widely divergent from those of the true Diptera and of the Pupipara that Latreille and the English zoologists have separated them entirely from the Diptera and have raised them into a separate order.[210] Those who do not agree in this arrangement, but with Gerstäcker include the fleas under the Diptera, will nevertheless admit that the morphological divergence between the Aphaniptera and the two other tribes is far greater than that which exists between the latter. Now the larvæ of the fleas are completely similar in structure to those of the gnat-type, since they possess a corneous head with the typical mouth parts and antennæ and a 13-segmented body devoid of legs. Were we only acquainted with the larvæ of the fleas we should rank them with the true Diptera under the sub-order Nemocera. On first finding such a larva we should expect to see emerge from the pupa a small gnat.

While the imagines of the Nemocera and Aphaniptera thus show but a very remote form-relationship their larvæ are very closely allied. Can any one doubt that in this case it is not the larva but the imago which has diverged to the greatest extent? Have not the fleas moreover become adapted to conditions of life widely different from those of all other Diptera, whilst their larvæ do not differ in this respect from many other Dipterous larvæ?

We have here, therefore, another case of unequal phyletic development, which manifests itself in the entirely different form-relationship of the larvæ and the imagines. Thus in this case, as in that of the Lepidoptera, it is sometimes the larval and at other times the imaginal stage which has experienced the greatest transformation, and, as in the order mentioned, the objection that a phyletic vital force produces greater and more important differentiations in the higher imaginal stage than in the lower or less developed larval stage, is equally ineffectual.

If, however, it be asked whether the unequal phyletic development depends in this case upon an unequal number of transforming impulses which the two stages may have experienced during an equal period of time, this must be decidedly answered in the negative. The unequal development obviously depends in this case, as in the higher systematic groups of the Lepidoptera, upon the unequal value of the parts affected by the changes. These parts are on the one side of small importance, and on the other side of great importance, to the whole structure of the insect. This is shown in the last-mentioned case of the fleas, where, of the typical parts of the body, only the wings have become rudimentary, whilst the antennæ, mouth-parts, and legs, and even the form and mode of segmentation (free thoracic segments), must have suffered most important modifications; their larvæ, on the other hand, can have experienced only unimportant changes, since they still agree in all typical parts with those of the gnat-type.

Although therefore in this and in similar cases a greater number of transforming impulses may well have occurred on the one side than on the other—and it is indeed highly probable that this number has not been absolutely the same—nevertheless the chief cause of the striking incongruence is not to be found therein, but rather in the strength of the transforming impulses, if I may be permitted to employ this figure, or, more precisely expressed, in the importance of the parts which become changed and at the same time in the amount of change.

In this conclusion there is implied as it appears to me an important theoretical result which tells further against the efficacy of a phyletic force.

If the so-called “typical parts” of an animal disappear completely through the action of the environment only, and still further, if these parts can become so entirely modified as to give rise to quite new and again typical structures (suctorial head of the Muscidæ) without the typical parts of the other stage of the same individual being thereby modified and transformed into a new type of structure, how can we maintain a distinction between typical and non-typical parts with respect to their origin? But if a difference exists with respect only to the physiological importance of such parts, i.e. their importance for the equilibrium of the whole organization, while, with reference to transformation and suppression, exactly the same influences appear to be effective as those which bring about a change in or a disappearance of the so-called adventitious parts, where is there left any scope for the operation of the supposed phyletic force? What right have we to assume that the typical structures arise by the action of a vital force? Nevertheless this is the final refuge of those who are bound to admit that a great number of parts or characters of an animal can become changed, suppressed, or even produced by the action of the environment.

---

IV.
Summary and Conclusion.

The question heading the second section of this essay must at the conclusion of the investigation be answered in the negative. The form-relationship of the larvæ does not always coincide with that of the imagines, or, in other words, a system based entirely on the morphology of the larvæ does not always coincide with that founded entirely on the morphology of the imagines.

Two kinds of incongruence here present themselves. The first arises from the different amount of divergence between two systematic groups in the larvæ and in the imagines, these groups being of equal extent. The second form of incongruence consists essentially in that the two stages form systematic groups of different extents, either the one stage constituting a group of a higher order than the other and therefore forming a group of unequal value, or else the two stages form groups of equal systematic value, these groups, however, not coinciding in extent, but the one overlapping the other.

This second form of incongruence is very frequently connected with the first kind, and is mostly the direct consequence of the latter.

The cause of the incongruences is to be found in unequal phyletic development, either the one stage within the same period of time having been influenced by a greater number of transforming impulses than the other, or else these impulses have been different in strength, i.e. have affected parts of greater or less physiological value, or have influenced parts of equal value with unequal strength.

In all these cases in which there are deep-rooted form-differences, it can be shown that these correspond exactly with inequalities in the conditions of life, this correspondence being in two directions, viz. in strength and in extent: the former determines the degree of form-difference, the latter its extent throughout a larger or smaller group of species.

The different forms of incongruence are manifested in the following manner:—

(1.) Different amount of form-divergence between the larvæ on the one side and the imagines on the other. Among the Lepidoptera this is found most frequently in varieties and species, and there is evidence to show that in this case the one stage has been affected by transforming influences, either alone (varieties), or at any rate to a greater extent (species). In the last case it can be shown in many ways that one stage (the larva) has actually remained at an older phyletic grade (Deilephila species). Incongruences of this kind depending entirely upon the more frequent action of transforming impulses can only become observable in the smaller systematic groups, in the larger they elude comparative examination. In the higher groups unequal form-divergence may be produced by the transforming impulses affecting parts of unequal physiological and morphological value, or by their influencing parts of equal value in different degrees. All effects of this kind can, however, only become manifest after a long-continued accumulation of single changes, i.e. only in those systematic groups which require a long period of time for their formation. By this means we can completely explain why the incongruences of form-divergence continually diminish from varieties to genera, and then increase again from genera upwards through families, tribes, and sub-orders: the first diminishing incongruence depends upon an unequal number of transforming impulses, the latter increasing incongruence depends upon the unequal power of these impulses.

Cases of the second kind are found among the Lepidopterous families, and especially in the higher groups (Rhopalocera and Heterocera), and appear still more striking in the higher groups of the Hymenoptera and Diptera. Thus the caterpillar shaped and maggot-formed larvæ of the Hymenoptera differ from one another to a much greater extent than their imagines, since the latter have experienced a complete transformation of typical parts; whilst in the caterpillar-formed larvæ these parts vary only within moderate limits. Similarly in the case of the Diptera, of which the gnat-like larvæ diverge more widely from those of the grub type than do the gnats from the true flies. On the other hand the divergence between the imagines of the fleas and gnats is considerably greater than that between their larvæ—indeed the larvæ of the fleas would have to be ranked as a family of the sub-order of the gnat-like larvæ if we wished to carry out a larval classification. By this it is also made evident that these unequal divergences, when they occur in the higher systematic groups, always induce at the same time the second form of incongruence—that of the formation of unequal systematic groups.

In general whenever such unequal divergences occur in the higher groups they run parallel with a strong deviation in the conditions of life. If these differ more strongly on the side of the larvæ, we find that the structure of the latter likewise diverges the more widely, and that their form-relationship is in consequence made more remote (saw-flies and ichneumons, gnats and flies); if, on the other hand, the difference in the conditions of life is greater on the side of the imagines, we find among the latter the greater morphological divergence (butterflies and moths, gnats and fleas).

(2.) The second chief form of incongruence consists in the formation of different systematic groups by the larvæ and the imagines, if the latter are grouped simply according to their form-relationship without reference to their genetic affinities. This incongruence again shows itself in two forms—in the formation of groups of unequal value, and the formation of groups equal in value but unequal in extent, i.e. of overlapping instead of coinciding groups.

Of these two forms the first arises as the direct result of a different amount of divergence. Thus the larvæ of the fleas, on account of their small divergence from those of the gnats, could only lay claim to the rank of a family, whilst their imagines are separated from the gnats by such a wide form-divergence that they are correctly ranked as a distinct tribe or sub-order.

The inequalities in the lowest groups, varieties, can be regarded in a precisely similar manner. If the larva of a species has become split up into two local forms, but not the imago, each of the two larval forms possesses only the rank of a variety, whilst the imaginal form has the value of a species.

Less simple are the causes of the phenomenon that in the one stage the lower groups can be combined into one of higher rank, whilst the other stage does not attain to this high rank. Such a condition appears especially complicated when the two stages can again be formed into groups of a still higher rank.

This is the case in the tribe Rhopalocera, which is founded on the imagines alone, the larvæ forming only families of butterflies. Both stages can however be again combined into the highest systematic group of the Lepidoptera.

In this case also the difference in the value of the systematic groups formed by the two stages corresponds precisely with the difference in the conditions of life. This appears very distinctly when there are several sub-groups on each side, and not when, as in the fleas, only one family is present as a tribe on the one side and on the other as a family. Thus in the butterflies, on the one side there are numerous families combined into the higher rank of a sub-order (imagines), whilst on the other side (larvæ) a group of the same extent cannot be formed. In this instance it can be distinctly shown that the combination of the families into a group of a higher order, as is possible on the side of the imagines, corresponds exactly with the limits in which the conditions of life deviate from those of other Lepidopterous families. The group of butterflies corresponds with an equally large circle of uniform conditions of life, whilst a similar uniformity is wanting on the side of the larvæ.

The second kind of unequal group formation arises from the circumstance that groups of equal value can be formed from the two stages, but these groups do not possess the same limits—they overlap, and only coincide in part.

This is most clearly seen in the order Hymenoptera, in which both larvæ and imagines form two well-defined morphological sub-orders, but in such a manner that the one larval form not only prevails throughout the whole of the one sub-order of the imagines, but also extends beyond and spreads over a great portion of the other imaginal sub-order.

Here again the dependence of this phenomenon upon the influence of the environment is very distinct, since it can be demonstrated (by the embryology of bees) that the one form of larva—the maggot-type—although the structure now diverges so widely, has been developed from the other form, and that it must have arisen by adaptation to certain widely divergent conditions of life.

This form of incongruence is always connected with unequal divergence between the two stages of the one systematic group—in this case the Terebrantia. The larvæ of this imaginal group partly possess caterpillar-like (Phytospheces) and partly maggot-formed (Entomospheces) larvæ, and differ from one another to a considerably greater extent than the saw-flies from the ichneumons.[211] The final cause of the incongruence lies therefore in this case also in the fact that one stage has suffered stronger changes than the other, so that a deeper division of the group has occurred in the former than in the latter.

The analogous incongruences in single families of the Lepidoptera may have arisen in a similar manner, as has already been more clearly shown above; only in these cases we are as yet unable to prove in detail that the larval structure has become more strongly changed through special external conditions of life than that of the imagines.

In the smallest systematic group—varieties, it has been possible to furnish some proof of this. The one-sided change here depends in part upon the direct action of external influences (seasonal dimorphism, climatic variation), and it can be shown that these influences (temperature) acted only on the one stage, and accordingly induced change in this alone whilst the other stage remained unaltered.

It has now been shown—not indeed in every individual case, but for each of the different kinds of incongruence of form-relationship—that there is an exact parallelism corresponding throughout with the incongruence in the conditions of life. Wherever the forms diverge more widely in one stage than in the other we also find more widely divergent conditions of life; wherever the morphological systemy of one stage fails to coincide with that of the other—whether in the extent or in the value of the groups—the conditions of life in that stage also diverge, either more widely or at the same time within other limits; whenever a morphological group can be constructed from one stage but not from the other, we find that this stage alone is submitted to certain common conditions of life which fail in the other stage.

The law that the divergence in form always corresponds exactly with the divergence in the conditions of life[212] has accordingly received confirmation in all cases where we have been able to pronounce judgment. Unequal form-divergences correspond precisely with unequal divergence in the conditions of life, and community of form appears within exactly the same limits as community in the conditions of life.

These investigations may thus be concluded with the following law:—In types of similar origin, i.e. having the same blood-relationship, the degree of morphological relationship corresponds exactly with the degree of difference in the conditions of life in the two stages.

With respect to the question as to the final cause of transformation this result is certainly of the greatest importance.

The interdependence of structure and function has often been insisted upon, but so long as this has reference only to the agreement of each particular form with some special mode of life, this harmony could still be regarded as the result of a directive power; but when in metamorphic forms we not only see a double agreement between structure and function, but also that the transformation of the form occurs in the two chief developmental stages in successive steps at unequal rates and with unequal strength and rhythm, we must—at least so it appears to me—abandon the idea of an inherent transforming force; and this becomes the more necessary when, by means of the opposite and extremely simple assumption that transformations result entirely from the response of the organism to the actions of the environment, all the phenomena—so far as our knowledge of facts at present extends—can be satisfactorily explained. A power compelling transformation, i.e. a phyletic vital force, must be abandoned, on the double ground that it is incapable of explaining the phenomena (incongruence and unequal phyletic development), and further because it is superfluous.

Against the latter half of this argument there can at most be raised but the one objection that the phenomena of transformation are not completely represented by the cases here analysed. In so far as this signifies that the whole organic world, animal and vegetable, has not been comprised within the investigation this objection is quite valid. The question may be raised as to the limit to which we may venture to extend the results obtained from one small group of forms. I shall return to this question in the last essay.

But if by this objection it is meant that the restricted field of the investigation enables us to actually analyse only a portion of the occurring transformations,[213] and indeed only those cases, the dependence of which upon the external conditions of life would be generally admitted, I will not let pass the opportunity of once more pointing out at the conclusion of the present essay that the incongruences shown to exist by no means depend only upon those more superficial characters the remodelling of which in accordance with the external conditions of life may be most easily discerned and is most difficult to deny, but that in certain cases (maggot-like Dipterous larvæ) it is precisely the “typical” parts which become partly suppressed and partly converted into an entirely new structure. From the ancient typical appendages there have here arisen new structures, which again have every right to be considered as typical. This transformation is not to be compared with that experienced by the swimming appendages of the Nauplius-like ancestor of an Apus or Branchipus which have become mandibulate, nor with the transformation which the anterior limbs must have gone through in the reptilian ancestors of birds. The changes in question (Dipterous larvæ) go still further and are more profound. I lay great emphasis upon this because we have here one of the few cases which show that typical parts are quite as dependent upon the environment as untypical structures, and that the former are not only able to become adapted to external conditions by small modifications—as shown in a most striking manner by the transformations of the appendages in the Crustacea and Vertebrata—but that these parts can become modelled on an entirely new type which, when perfected, gives no means of divining its mode of origin. I may here repeat a former statement:—With reference to the causes of their origination we have no grounds for drawing a distinction between typical and untypical structures.

It may be mentioned in concluding that quite analogous although less sharply defined results are arrived at if, instead of fixing our attention upon the different stages of a systematic group in their phyletic development, we only compare the different functional parts (organs in the wide sense) of the organisms.

A complete parallel can be drawn between the two classes of developmental phenomena. From the very different systematic values attached by taxonomists to this or that organ in a group of animals, it may be concluded that the individual parts of an organism are to a certain extent independent, and that each can vary independently, when affected either entirely alone or in a preponderating degree by transforming impulses, without all the other parts of the organism likewise suffering transformation, or at least without their becoming modified in an equal degree. Did all the parts and organs in two groups of animals diverge from each other to the same extent, the systematic value of such parts would be perfectly equal; we should, for example, be able to distinguish and characterize two genera of the family of mice by their kidneys, their liver, their salivary glands, or by the histological structure of their hair or muscles, or even by differences in their myology, &c. equally as well as by their teeth, length of toes, &c. It is true that such a diagnosis has yet to be attempted; but it may safely be predicted that it would not succeed. Judging from all the facts at present before us, the individual parts—and especially those connected in their physiological action, i.e. the system of organs—do not keep pace with reference to the modifications which the species undergoes in the course of time; at one period one system and at another period some other system of organs advances while the others remain behind.

This corresponds exactly with the result already deduced from the unparallel development of the independent ontogenetic stages. If the inequality in the phyletic development is more sharply pronounced in this than in the last class of cases, this can be explained by the greater degree of correlation which exists between the individual systems of organs in any single organism as compared with that existing between the ontogenetic stages, which, although developed from one another, are nevertheless almost completely independent. We should have expected à priori that a strong correlation would have here existed, but as a matter of fact this is not the case, or is so only in a very small degree.

Just as in the stages of metamorphosis the inequality of phyletic development becomes the more obliterated the more distant and comprehensive, or, in other words, the greater the period of existence of the groups which we compare, so does the unequal divergence of the systems of organs become obliterated as we bring into comparison larger and larger systematic groups.

It is not inconceivable—although a clear proof of this is certainly as yet wanting—that a variety of the ancestral species would differ only in one single character, such as hairiness, colour, or marking, and such instances would thus agree precisely with the foregoing cases in which only the caterpillar or the butterfly formed a variety. All the more profound modifications however—such for instance as those which determine the difference between two species—are never limited to one character, but always affect several, this being explicable by correlation, which, as Darwin has shown in the case of dogs, may cause modifications in the skull of those breeds having hanging ears in consequence of this last character alone. It must be admitted however that one organ only would be originally affected by a modifying influence. Thus, I am acquainted with two species of a genus of Daphniacea which are so closely allied that they can only be distinguished from one another by a close comparison of individual details. But whilst most of the external and internal organs are almost identical in the two species the sperm-cells of the males differ in a most striking manner, in one species resembling an Australian boomerang in form and in the other being spherical! An analogous instance is furnished by Daphnia Pulex and D. Magna, two species which were for a long time confounded. Nearly all the parts of the body are here exactly alike, but the antennæ of the males differ to a remarkable extent, as was first correctly shown by Leydig.

Similarly in the case of genera there may be observed an incongruence of such a kind that individual parts of the body may deviate to a greater or to a less extent than the corresponding parts in an allied genus. If, for instance, we compare a species of the genus of Daphniacea, Sida, with a species of the nearly allied genus Daphnella, we find that all the external and internal organs are in some measure dissimilar—nevertheless certain of these parts deviate to an especially large extent, and have without question become far more transformed than the others. This is the case, for example, with the antennæ and the male sexual organs. The latter, in Daphnella, open out at the sides of the posterior part of the body as long, boot-shaped generative organs, and in Sida as small papillæ on the ventral side of this region of the body. If again we compare Daphnella with the nearly allied genus Latona, it will be found that no part in the one is exactly similar to the corresponding part in the other genus, whilst certain organs differ more widely than others. This is the case for instance with the oar-like appendages which in Latona are triramous, but in Daphnella, as in almost all the other Daphniacea, only biramous.

In families the estimation of the form-divergence of the systems of organs and parts of the body becomes difficult and uncertain: still it may safely be asserted that the two Cladocerous families Polyphemidæ and Daphniidæ differ much less from one another in the structure of their oar-like appendages than in that of their other parts, such as the head, shell, legs, or abdominal segments. In systematic groups of a still higher order, i.e. in orders, and still more in classes, we might be inclined to consider that all the organs had become modified to an equally great extent. Nevertheless it cannot be conclusively said that the kidneys of a bird differ from those of a mammal to the same extent as do the feathers from mammalian hair, since we cannot estimate the differences between quite heterogeneous things—it can only be stated that both differ greatly. Here also the facts are not such as would have been expected if transformation was the result of an internal developmental force; no uniform modification of all parts takes place, but first one part varies (variety) and then others (species), and, on the whole, as the systematic divergence increases all parts become more and more affected by the transformation and all tend continually to appear changed to an equal extent. This is precisely what would be expected if the transforming impulses came from the environment. An equalization of the differences caused by transformation must be produced in two ways; first by correlation, since nearly every primary transformation must entail one or more secondary changes, and secondly because, as the period of time increases, more numerous parts of the body must become influenced by primary transforming factors.

A tempting theme is here also offered by attempting to trace the inequality of phyletic development to dissimilar external influences, and by demonstrating that individual organs have as a rule become modified in proportion to the divergence in the conditions of life by which they have been influenced, this action, during a given period of time, having been more frequent in the case of one organ than in that of the others, or, in brief, by showing the connection between the causes and effects of transformation.

It would be quite premature, however, to undertake such a labour at present, since it will be long before physiology is able to account for the fine distinctions shown by morphology, and further because we have as yet no insight into those internal adjustments of the organism which would enable us à priori to deduce definite secondary changes from a given primary transformation. But so long as this is impossible we have no means of distinguishing correlative changes from the primary modifications producing them, unless they happen to arise under our observation.

---

APPENDIX I.[214]
Additional Notes on the Ontogeny, Phylogeny, &c., of Caterpillars.

Ontogeny of the Noctua larvæ.—References have already been given in a previous note ([67], p. [166]) to observations on the number of legs and geometer-like habits of certain Noctua-larvæ when newly hatched. This interesting fact in the development of these insects furnishes a most instructive application of the principle of ontogeny to the determination of the true affinities, i.e. the blood-relationship of certain groups of Lepidoptera. While the foregoing portions of this work have been in course of preparation for the press, some additional observations on this subject have been published, and I may take the present opportunity of pointing out their systematic bearing—not, indeed, with a view to settling definitively the positions of the groups in question, as our knowledge is still somewhat scanty—but with the object of stimulating further investigation.

Mr. H. T. Stainton has lately recorded the fact that the young larva of Triphæna Pronuba is a semi-looper (Ent. Mo. Mag. vol. xvii. p. 135); and in a recently published life-history of Euclidia Glyphica (Ibid. p. 210) Mr. G. T. Porritt states that this caterpillar is a true looper when young, but becomes a semi-looper when adult. To these facts Mr. R. F. Logan adds (Ibid. p. 237) that “nearly all the larvæ of the Trifidæ are semi-loopers when first hatched.” The Cymatophoræ appear to be an exception, but Mr. Logan points out that this genus is altogether aberrant, and seems to be allied to the Tortricidæ. Summing up the results of these and the observations previously referred to, it will be seen that this developmental character has now been established in the case of species belonging to the following families of the section Genuinæ:—Leucaniidæ, Apameidæ, Caradrinidæ, Noctuidæ, Orthosiidæ, Hadenidæ, and Xylinidæ, as well as the other Trifidæ (excepting Cymatophora).[215] The larvæ of the Minores and Quadrifidæ are as a rule semi-loopers when adult and may be true loopers when young, although further observations on this point are wanted. These facts point to the conclusion that the Noctuæ as a whole are phyletically younger than the Geometræ, whilst the Genuinæ and Bombyciformes have further advanced in phyletic development than the Minores and Quadrifidæ. The last two sections are therefore the most closely related to the Geometræ, as correctly shown by the arrangement given in Stainton’s “Manual;” whilst that adopted in Doubleday’s “Synonymic List,” where the Geometræ precede the Noctuæ, is most probably erroneous.

Additional descriptions of Sphinx-larvæ.—In the foregoing essay on “The Origin of the Markings of Caterpillars,” Dr. Weismann has paid special attention to the larvæ of the Sphingidæ and has utilized for this purpose, in addition to his own studies of the ontogeny of many European species, the figures in the chief works dealing with this family published down to the time of appearance of his essay (1876).[216] In order to amplify this part of the subject I have added references to more recent descriptions and figures of Sphinx-larvæ published by Burmeister and A. G. Butler, and I have endeavoured in these cases to refer the caterpillars as far as possible to their correct position in the respective groups founded on the ontogeny and phylogeny of their allies. It is, however, obvious that for the purposes of this work figures or descriptions of adult larvæ are of but little value, except for the comparative morphology of the markings; and even this branch of the subject only becomes of true biological importance when viewed in the light of ontogeny. As our knowledge of the latter still remains most incomplete in the case of exotic species, it would be at present premature to attempt to draw up any genealogy of the whole family, and I will here only extend the subject by adding some few descriptions of species which are interesting as having been made from the observations of field-naturalists, and which contain remarks on the natural history of the insects.

Mr. C. V. Riley in his “Second Annual Report on the Noxious, Beneficial, and other Insects of the State of Missouri, 1870,” gives figures and describes the early stages and adult forms of certain grape-vine feeding larvæ of the sub-family Chærocampinæ. The full-grown larva of Philampelus Achemon, Drury, “measures about 3½ inches when crawling, which operation is effected by a series of sudden jerks. The third segment is the largest, the second but half its size, and the first still smaller, and when at rest the two last-mentioned segments are partly withdrawn into the third.... The young larva is green, with a long slender reddish horn rising from the eleventh segment and curving over the back.” Mr. Riley then states that full grown specimens are sometimes found as green as the younger ones, but “they more generally assume a pale straw or reddish-brown colour, and the long recurved horn is invariably replaced by a highly polished lenticular tubercle.” The specimen figured was the pale straw variety, this colour deepening at the sides, and finally merging into a rich brown. The markings appear to consist of an interrupted brown dorsal line, a continuous subdorsal line of the same colour, and six oblique scalloped white bars along the side. Whether the colour and marking is adapted to the vine, as is the case with the two varieties of the dimorphic Chærocampa Capensis (q.v.), is not stated. The larva of Philampelus Satellitia, Linn., when newly hatched, and for some time afterwards is “green with a tinge of pink along the sides, and with an immensely long straight pink horn at the tail. This horn soon begins to shorten, and finally curls round like a dog’s tail.” The colour of the insect changes to a reddish-brown as it grows older, and the caudal horn is entirely lost at the third moult. The chief markings appear to be five oblique cream-yellow patches with a black annulation on segments 6–10, and a pale subdorsal line. The caterpillar crawls by a series of sudden jerks, and often flings its “head savagely from side to side when alarmed.” “When at rest, it draws back the fore part of the body and retracts the head and first two joints into the third.” Two points in connection with these species are of interest with respect to the present investigations. The green colour and the possession of a long caudal horn when young shows that these larvæ, like those of Chærocampa Elpenor (p. [178]), C. Porcellus (p. [184]), and Philampelus Labruscæ (p. [195], [note]), are descended from ancestors which possessed these characters in the adult state.[217] The next point of interest is the attitude of alarm assumed by these larvæ, and effected by withdrawing the head and two front segments into the third.[218] The importance of this in connection with the similar habit of ocellated species will be seen on reading the remarks on page 367 bearing upon the initial stages of eye-spots. The other species figured by Mr. Riley are Chærocampa Pampinatrix, Smith and Abbot, and Thyreus Abboti, Swains. The latter has already been referred to (p. [256]).

In a paper “On a Collection of Lepidoptera from Candahar” (Proc. Zoo. Soc., May 4th, 1880), Mr. A. G. Butler has described and figured, from materials furnished to him by Major Howland Roberts, the larvæ of three species of Sphingidæ. Chærocampa Cretica, Boisd., feeds on vine; out of 100 specimens examined, there was not one black variety, while in another closely allied species, found at Jutogh and Kashmir, the larva is stated to be as often black as green. The general colour of the caterpillar harmonizes with that of the underside of the vine leaves; it possesses a thread-like dorsal, and a pale yellow subdorsal line; also “a subdorsal row of eye-spots, each consisting of a green patch in a yellow oval, the first spot on the fifth segment being the largest and most distinct, those on each following segment becoming smaller, more flattened, and less distinct, till lost on the twelfth segment, sometimes becoming indistinct after the seventh or eighth segment; these spots are only distinct as eye-spots on the fifth and sixth segments, that on the sixth being flatter than that on the fifth, those on the remaining segments appearing like dashes while the larvæ is green, but more like eyes on its changing colour when full fed.” The change here alluded to is the dark-brown coloration so generally assumed by green Sphinx-larvæ previous to pupation, and which, as I have stated elsewhere (Proc. Zoo. Soc., 1873, p. 155), is probably an adaptation advantageous to such larvæ when crawling over the ground in search of a suitable place of concealment. Making the necessary correction for the different mode of counting the segments, it will be seen that the primary ocelli of this species are in the same position as those of the other species of this genus as described in a previous part of this essay, and that it belongs to the second phyletic group treated of at p. [193]. The interesting fact that this species does not display dimorphism, whilst the closely allied form from Kashmir is dimorphic, shows that in the present species the process of double adaptation has not taken place; and this will probably be found to be connected with the habits of life, i.e. the insect being well adapted to the colour of its food-plant may not conceal itself on the ground by day. The caterpillar of Deilephila Robertsi, Butl., is found at Candahar on a species of Euphorbia growing on the rocky hills, and is so abundant that at the end of May every plant with any leaves left on it had several larvæ feeding upon it. “The larvæ are very beautiful and conspicuous, and are very different in colouring according to their different stages of growth.” The general colour is black with white dots and spots; a subdorsal row of large roundish spots, one on each segment, either white, yellow, orange or red; dorsal stripe variable in colour, and sometimes only partially present or altogether absent. “At the end of May most of the larvæ found presented a different appearance; the black disappears more or less, and with it many of the small white spots. In some cases the black only remains as a ring round the larger white spots; the ground-colour therefore becomes yellowish-green or yellow, varying very considerably.” The larva does not change colour previous to pupation. This species, according to the outline figure given (loc. cit., Pl. XXXIX., Fig. 9), appears to belong to the first of Dr. Weismann’s groups, comprising D. Euphorbiæ, D. Dahlii and D. Nicæa (see p. [199]), and is therefore in the seventh phyletic stage of development (p. [224]). From the recorded habits it seems most probable that the colours and markings of this caterpillar are signals of distastefulness. It is much to be regretted that Major Roberts has not increased the value of his description of this species by adding some observations or experiments bearing on this point. Eusmerinthus Kindermanni, Lederer, feeds on willow. “General colour green, covered with minute white dots and seven long pale yellow oblique lateral bands. (The ground-colour is the same as the willow-leaves on which the larva feeds, the yellow stripes the same as the leaf-stalks, and the head and true legs like the younger branches).” As no subdorsal line is mentioned or figured, this species must be regarded as belonging to the third stage of phyletic development (see p. [242]).

I have recently had an opportunity of inspecting a large number of drawings of Sphinx-larvæ in the possession of Mr. F. Moore, and of those species not mentioned in the previous portions of this work the following may be noticed:—Chærocampa Theylia, Linn., like Ch. Lewisii (note [82], p. [194]), appears to be another form connecting the second and third phyletic groups of this genus. Ch. Clotho, Drury, belongs to the third group (figured by Semper; see note 3 to this Appendix). The larva of Ch. Lucasii, Walk., offers another instance of the retention of the subdorsal line by an ocellated species. The larva of Ch. Lycetus, Cram., of which Mr. Moore was so good as to show me descriptions made at the various stages of growth, presents many points of interest. It belongs to the third phyletic group, and all the ocelli appear at a very early stage. The dimorphism appears also in the young larvæ, some being green, and others black, a fact which may be explained by the law of “backward transference” (see p. [274]). A most suggestive feature is presented by the caudal horn, which in the young caterpillar is stated to be freely movable. It is possible that this horn, which was formerly possessed by the ancestors of the Sphingidæ, and which is now retained in many genera, is a remnant of a flagellate organ having a similar function to the head-tentacles of the Papilio-larvæ, or to the caudal appendages of Dicranura (see p. [289]).

Lophostethus Dumolinii, Angas.—The larva of this species differs so remarkably from those of all other Sphingidæ, that I have thought it of sufficient interest to publish the following description, kindly furnished by Mr. Roland Trimen, who in answer to my application sent the following notes:—“My knowledge of the very remarkable larva of this large and curious Smerinthine Hawk-moth is derived from a photograph by the late Dr. J. E. Seaman, and from drawings and notes recently furnished by Mr. W. D. Gooch. The colour is greenish-white, inclining to grey, and in the male there is a yellow, but in the female a bluish, tinge in this. All the segments but the second and the head bear strong black spines, having a lustre of steel blue, and springing from a pale yellow tubercular base. The longest of these spines are in two dorsal rows from the fourth to the eleventh segment, the pairs on the fourth and fifth segments being longer than the rest, very erect, and armed with short simple prickles for three-fourths of their upper extremity. The anal horn, which is shorter than the spines, is of the same character as the latter, being covered with prickles, and much inclined backwards. Two lateral rows of similar shorter spines extend from the fourth to the 12th segment, and on each of the segments 6–11 the space between the upper and lower spines is marked with a conspicuous pale yellow spot. Two rows of smaller similar spines extend on each side (below the two rows of larger ones) from the second to the thirteenth segment, one spine of the lowermost row being on the fleshy base of each pro-leg. All the pro-legs are white close to the base, and russet-brown beyond. Head smooth, unarmed in adult, greenish-white with two longitudinal russet-brown stripes on face.

“The young larvæ have proportionally much longer and more erect spines with distinct long prickles on them. There is a short pair besides, either on the back of the head or on the second segment. Moreover, the dorsal spines of the third and fourth segments, and the anal horn (which is quite erect, and the longest of all), are longer than the rest, and distinctly forked at their extremity.

“Mr. Gooch notes that these young larvæ might readily be mistaken for those of the Acrææ, and suggests that this may protect them. He also states that the yellow lateral spots are only noticed after the last moult before pupation, and that the general resemblance of the larva as regards colour is to the faded leaves of its food-plant, a species of Dombeia.”

The forked caudal horn in the young larva of this species is of interest in connection with the similar character of this appendage in the young caterpillar of Hyloicus Pinastri, p. [265].

Retention of the Subdorsal Line by Ocellated Larvæ.—It has already been shown with reference to the eye-spots of the Chærocampa-larvæ, that these markings have been developed from the subdorsal line, and that, in accordance with their function as a means of causing terror, this line has in most species been eliminated in the course of the phylogeny from those segments bearing the eye-spots in order to give full effect to the latter (see p. [379]). In accordance with the law that a character when it has become useless gradually disappears, the subdorsal is more or less absent in all those species in which the ocelli are most perfectly developed; and it can be readily imagined that in cases where adaptation to the foliage exists the suppression of this line would under certain conditions be accelerated by natural selection. On the other hand, it is conceivable that the subdorsal line may under other conditions be of use to a protectively coloured ocellated species by imitating some special part of the food-plant, under which circumstances its retention would be secured by natural selection.

Such an instance is offered by Chærocampa Capensis, Linn.; and as this case is particularly instructive as likewise throwing light upon the retention of the subdorsal by certain species having oblique stripes (see p. [377], and note [166], p. [378]), I will here give some details concerning this species which have been communicated to me by Mr. Roland Trimen, the well-known curator of the South African Museum, Cape Town. The caterpillar of C. Capensis, like so many other species of the genus, is dimorphic, one form being a bright (rather pale) green, and the other, which is much the rarer of the two, being dull pinkish-red. Both these forms are adapted in colour to the vine on which they feed, the red variety according to some extent with the faded leaves of the cultivated vines, but to a greater extent with the young shoots and underside of the leaves of the South African native vine (Cissus Capensis), on which it also feeds. There are two eye-spots in this species in the usual positions; they are described as being blue-grey in a white ring, and raised so as to project a little. The subdorsal is white, and is bordered beneath by a wide shade of bluish-green irrorated with white dots, and crossed by an indistinct white oblique ray on each segment. These last markings are probably remnants of an oblique striping formerly possessed by the progenitor of this and other species of the genus (see, for instance, Fig. 25, [Pl. IV]., one of the young stages of C. Porcellus). It is possible that these rudimentary oblique stripes are now of service in assisting the adaptation of the larva to its food-plant, but this cannot be decided without seeing the insect in situ.

The subdorsal line extends from immediately behind the second eye-spot to the base of the very short and much curved violet anal horn. With reference to the protective colouring Mr. Trimen writes:—“The difficulty of seeing these large and beautifully-coloured larvæ on the vines is quite surprising; six or more may be well within sight, and yet quite unnoticed. The subdorsal stripe greatly aids in their concealment, as it well represents in its artificial light and shade the leaf-stalks of the vine.” When this larva withdraws its front segments the eye-spots stand out very menacingly; but in spite of this it is greedily eaten by fowls and shrikes (Fiscus Collaris), and Mr. Trimen also found that a tame suricate (Rhyzæna Suricata) and a large monitor lizard (Regenia Albogularis) did not refuse them. The failure of the eye-spots in causing terror in these particular cases cannot be regarded as disproving their utility in all instances. It must always be borne in mind that no protective character can possibly be of service against all foes; natural selection only requires that such characters should be advantageous with respect to the majority of the enemies of any species, and further experiments with this caterpillar may show that in the case of smaller foes the eye-spots are effective as a means of causing alarm. The dimorphism of the larva of C. Capensis is of special interest, although we are not yet sufficiently acquainted with the habits of this species to offer a complete explanation. According to Dr. Weismann’s conclusions (p. [297]), the dimorphism of the Chærocampa-larvæ is due to a double adaptation, the insects first having acquired the habit of concealing themselves by day, and the dark form having then been produced by the action of natural selection, in order to adapt such varieties to the colour of the soil, whilst others retained the green colour which adapts them to the foliage of their food-plants. In accordance with this, C. Capensis may have a similar habit of concealment, or (should this be found not to be the case) it is possible that this insect at a former period possessed this habit and fed upon some other plant, when it would have become dimorphic in the manner explained, and the existing dimorphism may be a survival of the more ancient dimorphism, the red form (corresponding to the older dark form) having been subsequently modified so as to become also adapted to the new food-plant. Much light would be thrown upon this by studying the ontogeny of the species.

Phytophagic Variability.—A number of observations bearing on the phytophagic variability of the Sphinx-larvæ and other caterpillars have been recorded in a previous [note] (p. [305]), and reference has also been made to the food-plants of Acherontia Atropos in South Africa (note [121], p. [263]). I am now enabled to add some further observations on this species, from notes furnished to me by Mr. Roland Trimen, who states that for many years he has noticed that at the Cape this larva varies greatly in the depth and shade of the green ground-colour, the variability being in strict accordance with the colour of the leaves of the particular plant on which the individual feeds. The phenomenon was particularly noticeable in larvæ feeding on Buxia Grandiflora, a shrub in common cultivation in gardens, and of which the foliage is of a very dull pale greyish-green. Another striking instance was noticed in some very fine caterpillars feeding on a large shrubby Solanum, which, excepting the bright yellow bands bordering the dorsal violet bars, were generally dull ochreous-yellow, like the leaves and stalks of the Solanum. On plants with bright green or deep green leaves, the colour of the larvæ is almost in exact agreement. Mr. Trimen adds:—“These remarks apply principally to the underside and pro-legs and lower lateral regions, the dorsal colours of violet and yellow varying but little. The protection afforded is very considerable, as the larvæ almost always cling to the lower side of the twigs of their food-plants, so that their uniformly-coloured under-surface is upwards, and turned towards the light, and their variegated upper surface turned downwards.”

These observations are of the highest importance, not only as adding another instance to the recorded cases of phytophagic variation, but likewise as showing that with this variability a protective habit has been acquired. It is to be hoped that such a promising field for experimental investigation as is offered by this and analogous cases will not long remain unexplored. In attacking the problem two chief questions have in the first place to be settled: (1) Is the variability truly phytophagic, i.e. are the colour variations actually brought about by the chemico-physiological action of the food-plant? and (2) Are the larvæ at any period of growth susceptible to the action of phytophagic influences? The first question could be decided by feeding larvæ from the same batch of eggs on different food-plants from the period of their hatching. The second question could be settled by changing the food-plants of a series of selected specimens at various stages of growth, and observing whether any change of colour was produced. In accordance with the principles advocated in a previous [note] (p. [305]), it is conceivable à priori that phytophagic variability may occur by direct chemico-physiological action, quite irrespective of any of the changes of colour being of protective use. In the case of brightly-coloured distasteful species phytophagic variability might thus have full play, but in the case of protectively-coloured edible species, phytophagic variability would be under the control of natural selection. These considerations raise a question of the greatest theoretical interest in connection with this phenomenon. If phytophagic variability can have full play uncontrolled by natural selection in brightly-coloured caterpillars, ought not this phenomenon to be of more common occurrence in such species than in those protectively coloured? Although our knowledge of this subject is still very imperfect, as a matter of fact brightly coloured larvæ, so far as I have been able to ascertain, do not appear to be susceptible of phytophagic influences. But this apparent contradiction, instead of opposing actually confirms the foregoing views, as will appear on further consideration. The colours of protected species are as a whole much inferior in brilliancy to those of inedible species, so that any phytophagic effect would be more perceptible in the former than in the latter, in which the highest possible standard of brilliancy appears in most cases to have been attained. Now phytophagic variations of colour appear to be of but small amount, or, in other words, such variations fluctuate within comparatively restricted limits, and as the cases at present known are mostly adaptive it is legitimate to conclude that they have been produced and brought to their present standard by natural selection, i.e. that they have arisen from phytophagic influences as a cause of variability. The initial stages of phytophagic variations must therefore have been still less perceptible than the now perfected final results; and this leads to the conclusion that minute variations of this character were of sufficient importance to protectively-coloured species to be taken advantage of by natural selection. But minute variations in a dull-coloured larva would, as previously pointed out, produce a comparatively much greater effect than such variations in a brilliantly-coloured species; and as protection is required by the former, the initial phytophagic effects would be accumulated, and the power of adaptability conferred by the continued action of natural selection, whilst in vividly-coloured species where no power of adaptability is required this cause of variation would not only produce a result which, as compared with its effects upon dull species, may be regarded as a “vanishing quantity,” but this result would be too insignificant to be taken advantage of by natural selection, which is in these cases dealing only with large “quantities,” and striving to make the caterpillars as brilliant as possible. The fact that vividly-coloured distasteful larvæ do not show phytophagic variation is to my mind explained proximately by these considerations; the ultimate cause of phytophagic variability regarded as a chemico-physiological action requires further investigation.

Sexual Variation in Larvæ.—Since most of the markings of caterpillars can be explained by the two factors of adaptation and inheritance, or, in other words, by their present and past relations to the environment, and since sexual selection can have played no direct part in producing these colours and markings, I feel bound to record here some few observations on the sexual differences in larvæ in addition to the cases of Anapæa and Orgyia already recorded (note [i.], p. [308]) and of Lophostethus Dumolinii (p. [527]).

Mr. C. V. Riley states[219] with reference to the larva of Thyreus Abboti that the ground-colour appears to depend upon the sex, Dr. Morris having described the insect as “reddish-brown with numerous patches of light green,” and having expressly stated that “the female is of a uniform reddish-brown with an interrupted dark-brown dorsal line and transverse striæ.” Mr. W. D. Gooch, who has reared the South African butterflies Nymphalis Cithæron and N. Brutus from their larvæ, states[220] that these “differed sexually in both instances.” Of Brutus only a few were bred, but of Cithæron many. “The sexual difference of the latter was that the females had a large dorsal sub-cordate cream mark, which was wanting, or only shown by a dot, in the males, and the colour was more vivid in the edgings to the frontal horns.”

Although such cases appear to be at present inexplicable, they are of interest as examples of those “residual phenomena” which, as is well known, have in many branches of science so often served as important starting-points for new discoveries and generalizations.[221]

---