We have seen that the embryo rapidly passes through extraordinary changes of form, and now, after hatching, especially in the insects with a complete metamorphosis, the animal continues to undergo striking changes in form, in adaptation to different modes of life.

The life of a winged insect, such as a butterfly, fly, or bee, may be divided into four stages: the embryo, or egg state, the larva, pupa, and imago,—the term metamorphosis being applied to the changes after birth, or post-embryonic stages of life. The transformations of the more specialized orders of insects involve wonderful changes of form, which are only paralleled in other types of animals by the metamorphoses of the echinoderms, of certain worms, and of the Crustacea, as well as by those of the frog. An insect, such as a butterfly or bee, during its post-embryonic life lives, so to speak, three different lives, having distinct bodily structures and existing under quite dissimilar surroundings and habits; so that a caterpillar is practically a different animal from the pupa, and the latter from the imago, with different organs, the appendages and other structures being so modified as to be, so far as regards their functions, radically different. These changes of functions or of habits have also been plainly enough the exciting cause of the divergency in structure of what fundamentally is one and the same organ, the change having been brought about by adaptation of the same organs to quite different uses.

The changes are not only observable in the body and its appendages, but also in the internal organs, and consequently are both structural and physiological. The term larva, as applied to the first stage of animals, is a very variable and indefinite one, that of insects in general being a much more highly organized animal than the larva of a worm, starfish, or crustacean.

a. The nymph as distinguished from the larval stage

As there is no marked difference between the different stages of the young in the insects with an incomplete metamorphosis (Heterometabola), the chief difference being the possession of the rudiments of wings and the absence of a resting stage, the terms larva and pupa are in reality scarcely applicable to them, and we much prefer the term nymph, first proposed by Lamarck for the active “pupa” of Orthoptera, Hemiptera, the Odonata and Ephemeridæ, and adopted in part by many. Indeed, in the more generalized and older orders, the larval and pupal stages are not differentiated, though the term larval, in its general sense, will probably always be used; just as we speak of the larval stages of worms, echinoderms, or Crustacea.

Eaton in his elaborate work on the Ephemeridæ employs the term nymph to designate all the aquatic or early stages in the development of the young after hatching, and he urges that the old-fashioned usage of larva and pupa seem scarcely worth retention. “Nymphs are young which live an active life, quitting the egg at a tolerably advanced stage of morphological development and having the mouth-parts formed after the same main type of construction as those of the adult insect.” The word nymph is used in the same sense by McLachlan, and by Cabot. Calvert also applies the term nymph “to the stage of odonate existence between the egg and the transformation into the imago.” On the other hand, Brauer applies the term nymph to the pupa of holometabolous insects. For larval Hyatt proposes the term nepionic.

b. Stages or stadia of metamorphosis

The intervals or periods between the moults or ecdyses of caterpillars and other eruciform larvæ are called stages or stadia; thus, as most caterpillars moult four times, we have five stages or stadia, or stage (stadium) I to V. As observed by Sharp, there is, unfortunately, no term in general use to express the form of the insect at the various stadia; “entomologists say, ‘the form assumed at the first moult,’ and so on.” Hence he adopts a term suggested by Fischer,^[86] and calls the insect as it appears after leaving the egg the first instar, and what it is after the first moult the second instar, and so on; hence the pupa, or chrysalis, which assumed that condition after moulting five times would be the sixth instar, and the butterfly itself would be the seventh instar.

c. Ametabolous and metabolous stages

In the Synaptera development is direct, the young differing neither in form, structure, or habits from the adult. Hence they are said to be ametabolous. Since there is an absence of even a tendency to a partial metamorphosis, it is evident that the insects have not inherited a tendency to undergo a transformation, but that it is an adaptation induced in the hexapod type after the first winged insects appeared, and which became more marked in the more specialized insects and at a period comparatively late in geological history, i.e. perhaps at or soon after the beginning of the Carboniferous period.^[87]

The transformations of the pterygote insects vary greatly in degree, and it is difficult to draw the line between the grades. Those in which the adults differ from the freshly hatched young only or mainly in having wings are generally said to have an incomplete or gradual metamorphosis. There is no inactive, resting, or pupal stage, and the wings are acquired only after successive moults. Insects with an incomplete metamorphosis are the Orthoptera, Dermaptera, Platyptera (Mallophaga, Plecoptera, Corrodentia, Embidæ), Ephemeridæ, Odonata, Thysanoptera, and Hemiptera, with the exception of the male Coccidæ, in which there is a resting or sub-nymph stage. As regards the number of moults in the Synaptera, Grassi states that in Campodea there is a single fragmentary ecdysis, while Sommers tells us that Macrotoma plumbea sheds its skin throughout life, even after attaining its full size.

As an example of the partial metamorphosis of the hemimetabolous insects we may select that of the locust, in which there are five moults and six stages (instars), as seen in Fig. 558, five of which are nymphal. In the first two stages there are no rudiments of wings, these appearing after the second moult. Besides the acquisition of wings there are slight differences after each moult, both in structure and color, besides size, so that we may always recognize the comparative age and the particular stage of growth of any individual.^[88]

We have watched the development of Melanoplus spretus from the egg to the imago, and examined thousands of specimens which show the six stages. On the other hand, European authors differ as to whether there are three, four, or five moults in the migratory locust.^[89] It is not improbable that, as is the case with many other insects, the number of moults may vary according to the temperature and food, variation in these agencies causing either retardation or rapidity in development.

Those with a complete metamorphosis are said to be metabolous or holometabolous. (Lang.)

Leach^[90] in 1815 gave the name of Ametabola to insects without, and Metabola to insects with a metamorphosis.

Latreille (1831) called insects with an incomplete metamorphosis homotenous (which means similar to the end of life), and those with a complete metamorphosis, polymorphous. For the different degrees of metamorphosis of insects he employed two terms: for the incomplete degree, metamorphosis dimidia, and for the total or pupal, metamorphosis perfecta.

Westwood in his Introduction to the Modern Classification of Insects (1839), taking into account the relation of the larva with the imago, divided insects into two divisions: the Heteromorpha, or those in which there is no resemblance between the parent and its offspring, and Homomorpha, in which the larva resembles the imago, except in the absence of wings.

From the point of view of the degree of metamorphosis, insects have been divided into Heterometabola and Metabola.

I. Heterometabola.—This group may be divided as follows:

1. Manometabola,^[91] embracing those forms with a slight or gradual metamorphosis, but which are active in all the stages, without any resting stage. The orders passing through this degree of metamorphosis are the following: Orthoptera, Dermaptera, Platyptera, Thysanoptera, and Hemiptera (Coccidæ excepted).

In all these groups, the only external differences of importance between the freshly hatched nymph and the adult is the presence of wings. The chief difference internally is the complete development of the sexual glands.

It should be observed, however, that in the last nymph stage of the Thysanoptera the articulations of the limbs are enveloped by a membrane and the wings enclosed in short fixed sheaths; the antennæ are turned back on the head, and the insect, though it moves about, is much more sluggish than in the other state. (Haliday.) Hence here we have a close approach to the following degree.

2. Heremetabola,^[92] including those forms with a gradual though slight or incomplete metamorphosis, but with a quiescent or resting stage at the close of the nymph life. Lang has emphasized this stage, calling attention to the fact that the fore legs of the nymph of the 17–year Cicada, which lives underground on the roots of trees, are thick and adapted for digging. The transition from the nymph to the winged adult is signalized by the decided change in form of the fore legs, as well as by the acquisition of the wings. “The last larval stage is, then, what is called quiescent, i.e. the organization of the imago develops within the chrysalis at the expense of the accumulated reserve material.” (Lang.) There seems to be a resting stage, when the insect does not perhaps suck the sap from the roots, and awaits in its chamber its approaching change to the imago; but we should scarcely apply the term pupa to this stage, though the antennæ of the freshly hatched larva are larger and longer than in the fully grown nymph and are distinctly 8–jointed.

3. Hemimetabola.—In this division, so named by Brauer, the changes are more marked, though there is no truly inactive pupa-like stage. The orders are Perlaria (Plecoptera), Odonata, and Plectoptera (Ephemeridæ). The freshly hatched nymphs of these three groups are much alike in shape, that of Perlidæ, and indeed most of the Platyptera, being more generalized, unless we except that of Chloëon; all closely recall Campodea, and are therefore in the Campodea-stage. These nymphs are indeed more generalized than the freshly hatched nymph of Blattidæ, or any other of the orders mentioned except the Platyptera, to which perlids belong. They all have feet, and the body is more or less flattened. (Fig. 560.)

II. Holometabola.—In this division we have for the first time a true larva, and a pupa stage as distinguished from the imago. Moreover, the insect at each stage is distinguished by radical differences in form, surroundings, and in the nature of the food, while the pupa is inactive, usually immovable, and incapable of taking any food, and is often protected by a cocoon spun by the larva. The holometabolous orders are the Neuroptera, Coleoptera, Mecoptera, Trichoptera, Lepidoptera, Siphonaptera, Diptera, and Hymenoptera.

As we have among worms, echinoderms, and Crustacea certain exceptional species in a metamorphic group whose metamorphosis is suppressed, their development being direct, so there is in pterygote insects, though in a very much less degree, cases of direct development. In the wingless cockroaches such as Pseudoglomeris, etc., of the tribe of Periphæriides, in some of which, however, the males are winged, and in the Hemiptera, occur wingless forms such as the lice and bed-bug. The Mallophaga are all wingless, while certain Dermaptera (Chelidura, Anisolabis) are also apterous. The absence of wings in such cases is due to disuse from parasitism, or to a life under stones or in cracks and fissures, where the insects are driven to avoid their enemies, and hence do not need wings. The growth of wings and consequently the development of a metamorphosis is suppressed, so that, as Lang says, “in contrast to the original ametabola of the Apterygota, we have here an acquired ametabola.”

It is rare that, after the rudiments of wings have once appeared in the very young, they should disappear in the late nymph stage; this is, however, said by Walsh to be the case with the Ephemerid Bætisca (Fig. 440). This is a case of retardation in an acquired ametabolesis.

THE LARVA

The term larva is peculiarly applicable to the young of the holometabolous orders. The name (Latin, larva, a mask) was first given to the caterpillar because it was thought by the ancients to mask the form of the perfect insect. Swammerdam supposed that the larva contained within itself “the germ of the future butterfly, enclosed in what will be the case of the pupa, which is itself included in three or more skins, one over the other, that will successively cover the larva.” What led to his conception of the nature of these changes was probably his observations on the semitransparent larva of the gnat, in which the body and limbs of the pupa can be partially seen; for Weismann has shown that the great Dutch observer’s belief that the pupal and imaginal skins were in reality already concealed under that of the larva is partially founded in fact. Swammerdam states: “I can point out in the larva all the limbs of the future nymph, or Culex, concealed beneath the skin,” and he also observed beneath the skin of the larvæ of bees, just before pupating, the antennæ, mouth-parts, wings, and limbs of the adult. But, as we shall see farther on, the discovery by Weismann in the larva of the germs of the imago has completely changed our notions as to the nature of metamorphosis, and revolutionized our knowledge of the fundamental processes concerned in the change from larva to pupa, and from pupa to imago.

Not only are the larvæ of each order of insects characteristic in form, so that the grub or larva of beetles is readily distinguished from those of other orders, or the maggot of flies from the apodous larva of wasps and bees, but within the limits of the larger orders there is great diversity of larval forms, showing that they are the result of adaptation to their surroundings. This is especially the case with the larvæ of the Coleoptera, Lepidoptera, Diptera, and Hymenoptera.

In general, the larvæ of insects may be divided into two types,—the Campodea-form, or campodeoid, sometimes called thysanuriform, and the eruciform.

a. The Campodea-form type of larva

This is the most primitive and generalized type of larva (Fig. 560). A Campodeoid larva is one nearest in general shape to Campodea, the form which we have seen to be the nearest allied to the probable ancestor of the insects, and it also resembles the nymphs of the heterometabolous insects, before the appearance of their rudimentary wings.

Brauer, in 1869,^[93] first suggested that the larvæ of a great number of insects may be traced back to Campodea and Iapyx. The Campodea-form larva is active, with a more or less flattened body, well developed mandibulate mouth-parts, and usually long legs. The nearest approach to the form of Campodea is the freshly hatched nymph of cockroaches (Blattidæ), Forficula, Perlidæ, Termitidæ, Psocidæ, Embidæ, Ephemeridæ, Odonata, especially the more generalized Agrionidæ, the nymphs of Hemiptera, the larvæ of certain Neuroptera, the active pedate larvæ of the more generalized Coleoptera, such as those of Carabidæ, Cicindelidæ, Dyticidæ, etc., and the first larva (instar) of Stylopidæ and Meloidæ (Fig. 560, d).

While the Campodea-shape is retained throughout nymphal life, of the orders above mentioned the Neuroptera and Coleoptera alone have a true resting pupal stage.

It should also be observed that great changes in the form of the nymph occur within the limits of the Orthoptera; the nymph of all the families except that of the Blattidæ, evidently the most generalized and primitive, being more or less specialized, while the nymphs of the other orders all vary in degree of specialization and modification. The process of adaptation once begun went on very rapidly, as it has in many other orders of insects, as well as in animals of other phyla.

Brauer also sagaciously pointed out that “a larger part of the most highly developed insects assume another larva form, which appears not only as a later acquisition, through adaptation to certain definite conditions, but also arises as such before our eyes. The larvæ of Lepidoptera, of saw-flies, and Panorpidæ show the form most distinctly, and I call this the caterpillar form (Raupenform). That this is not the primitive form, but one later acquired, we see illustrated in certain beetles. The larvæ of Meloë and of Sitaris, in their fully grown conditions, possess the caterpillar form, but the new-born larvæ of these genera show the Campodea-form. The last form is lost as soon as the larva begins its parasitic mode of life.... The larger part of the beetles, the Neuroptera (in part), the bees and flies (the last with the most degraded maggot form), possess larvæ of this second form.” In 1871 we adopted these views, giving the name eruciform to this type of larvæ, and afterwards Lubbock adopted Brauer’s views. Brauer considered that the eruciform larva was the result of living a stationary semi-parasitic life on plants, in carrion, or burrowing in the trunks and branches or leaves and buds of trees, where they do not have to move about in search of their food. The change from the Campodea-form to the eruciform larva is a process of degeneration and often of atrophy of the limbs, and, in the footless forms of dipterous and hymenopterous insects, of the gnathites, accompanied by a tendency of the body to become more or less cylindrical.

The first steps in the origination of the eruciform larva were apparently taken in the order Neuroptera, as restricted by Brauer and by myself, where, though the larvæ are campodeoid, there is a true resting pupal stage. The most generalized larval form is perhaps that of the Sialidæ (Fig. 560, l), in which the body tends to be slightly cylindrical, though the legs are long, and the gnathites well developed for seizing and biting their living prey. The terrestrial larvæ of the Hemerobiidæ, though modifications of the sialid larval form, are considerably specialized in adaptation to their active carnivorous habits. But the life-history of Mantispa, where there are two larval stages, gives us plainly enough the key to the mode in which the complete metamorphosis was brought about. The larva, born a true Campodea-like form, with large, long, 4–jointed legs, has a structure which would enable it to move about freely after its prey, beginning at once to live a sedentary life in the egg-sac of a spider; before the first moult it loses the use of its legs, while the antennæ are partly aborted. The result is that, owing to this change of habits and surroundings from those of its active ancestors, it changes its form, and the fully grown larva becomes cylindrical, with small slender legs, and, owing to the partial disuse of its jaws, acquires a small, round head.

Its antennæ, mouth-parts, and legs not only retarded in growth, but retrograding and becoming vestigial, the body meanwhile becoming fat and cylindrical, an apparent acceleration of growth goes on within, with probably an enlargement of the intestine and fat-body, and thus the pupal form is perfected while the larva is full-fed and quiescent. It is not improbable that in the primitive neuropteron, as the result of a mode of life like that of Mantispa, the quiescent life of the later stages graduated into a quiescent, inactive pupal life, allowing the changes going on in the internal organs to result in a complete metamorphosis, which was transmitted to the later Neuroptera, thus making the complete metamorphosis a fixed, normal condition. It thus appears that a change of habits and of food, and more especially the fact that the nymph became so surrounded with an abundance of food close at hand that it did not have to run actively about and seize it in a haphazard manner, were the factors bringing about a change from the Campodea-form nymph to the eruciform larva, thus inducing a hypermetamorphosis.

The larvæ of the Mecoptera (Panorpidæ, Fig. 562, b) are still more caterpillar-like, and besides their cylindrical body, rounded head, small short gnathites, small thoracic legs, they have what appear to be 2–jointed legs to each of the nine abdominal segments, and the close resemblance to caterpillars is farther carried out by the presence of a pair of prothoracic spiracles, none existing on the other two thoracic segments.

In the Meloidæ (Fig. 560, d) and Stylopidæ the first larval stage is Campodea-form; the changes will be described in the subsequent section on Hypermetamorphosis, and while these cases of change from a campodeoid to an inactive eruciform larva are very salient, if we compare the graduated series of larval forms throughout the order of Coleoptera, as represented by the illustrations in Fig. 561, we shall see that in nearly, if not each, case the form of the boring or mining, or bark or bud or seed-inhabiting grub is the result of a change of habit and commissariat from active predaceous larvæ, like those of the Carabidæ and other adephagous families, together with those of the Staphylinidæ, with their flat body, big mandibles, and well-developed maxillæ, to the cylindrical bodies of such larvæ as those of Dermestes and Anthrenus, which live a more sedentary life, to the root-feeding wire-worm or elaterid larvæ, and scarabæid grubs, onward to the phytophagous Chrysomelidæ, with the mining and boring buprestids and cerambycids,—in all these forms we see a gradual atrophy of the legs, which is fully carried out in the vermiform or maggot-like larva of the weevils. These changes throughout the members of the entire order are epitomized in the life-history of the Meloidæ, in which there are three typical forms of larva: the Campodea-form (triungulin stage), eruciform (second or carabidoid stage), and vermiform (coarctate) larva.

In the Lepidoptera the eruciform, pedate type is adhered to throughout the order, with the rare exception of the nearly apodous mining larva of Prodoxus (Fig. 563, a), Phyllocnistis, and Nepticula, which have no thoracic legs, and the limacodid larvæ, whose abdominal legs are totally aborted, while the thoracic ones are much reduced (Fig. 564).

In the Hymenoptera the phytophagous forms are eruciform, while by the agency of the same factors as already mentioned, i.e. a sedentary or parasitic life and abundance of food within constant reach, the larvæ lose their legs and become vermiform.

In the Diptera, which are the most highly specialized of insects, the maggot or vermiform shape, and absence of any legs, prevails throughout the order, though the eucephalous larvæ show their origin from a primitive eruciform type of larva. The highly specialized larvæ of the Culicidæ and Simuliidae are undoubtedly related to the earliest and most generalized types, while the maggots of the parasitic flies (Tachinidæ) and other muscids are later degradational forms, and the result of adaptation induced, as in the previous cases, by a sedentary or parasitic mode of life, living as they do immersed in an abundance of rich nitrogenous food, with the result that the mouth-parts have become atrophied by disuse, while the limbs have become entirely aborted, though the thoracic imaginal discs develop normally in the embryonic or pre-larval stages.

It appears, therefore, highly probable that the metamorphoses of insects are the result of the action of change of conditions, just as the polymorphism of Termites is with little doubt the result of differences of food and other conditions. These matters will be farther discussed under the head of Causes of Metamorphosis.

LITERATURE ON ANCESTRY OF INSECTS, ETC.

Müller, Fritz. Für Darwin, 1869, pp. 144, 67 Figs.

Brauer, Friedrich. Betrachtung ueber die Verwandlung der Insekten im Sinne der Descendenz-Theorie. (Verhandlung d. k.k. zool. bot. Gesell. Wien., 1869, 1 Taf., pp. 1–20.)

Packard, A. S. Amer. Naturalist, iii, p. 45, March, 1869.

—— Proc. Boston Soc. Nat. Hist., xiv, 1870, p. 61.

—— Amer. Nat., iv, Feb. 1871, p. 756; v, 1871, pp. 52, 567.

—— Embryological Studies. (Memoirs Peabody Acad. Sc. Salem, 1871–72.)

—— Our common insects, 1873, Chapter on Ancestry of Insects, pp. 175–178.

—— Third Report U. S. Ent. Commission, 1883, pp. 295–304.

Lubbock, John. On the origin of insects. (Journ. Linn. Soc., London, xl, 1873.)

—— Origin and metamorphoses of insects. (Nature, 1873 [in book form, 1874], pp. 108, 66 Figs.)

Mayer, Paul. Ueber Ontogenie and Phylogenie der Insekten. (Jena. Zeitschr. Wissens., x, 1876, pp. 125–221, 4 Taf.)

Hyatt, A., and Arms, J. M. Insecta. (Bost. Soc. Nat. Hist. Guides for science-teaching, viii.) Boston, 1890, pp. 300, 13 Pls., 223 Figs.

The rapidity of growth and enormous increase in size in early life is especially noticeable in caterpillars and other phytophagous larvæ. The latest observations are those of Trouvelot on Telea polyphemus. When this silkworm hatches, it weighs 1
20 of a grain.

When

“When,” he says “a worm is 30 days old, it will have consumed about 90 grains of food; but when 56 days old, it is fully grown and has consumed not less than 120 oak leaves, weighing ¾ of a pound; besides this it has drank not less than ½ an ounce of water. So the food taken by a single silkworm in 56 days equals in weight 86,000 times the primitive weight of the worm. Of this about ¼ of a pound becomes excrementitious matter, 207 grains are assimilated, and over 5 ounces have evaporated.”^[94]

Dandolo stated that the Asiatic silkworm (Bombyx mori) weighs on hatching not over 1
100 of a grain, but when fully grown about 95 grains. During this period, therefore, it has increased 9500 times its original weight, and has eaten 60,000 times its weight of food. Newport thought this estimate of the amount of food was a little too great. But comparing it with Trouvelot’s estimate for the American silkworm, which weighs when hatched five times as much, it would not appear to be so. Newport found that the larva of Sphinx ligustri at the moment of leaving the egg weighs about 1
80 of a grain, and when fully fed 125 grains, so that in the course of 32 days it increases about 9976 times its original weight. This proportion of increase is exceeded by the larva of Cossus ligniperda, which, boring in the trunks of trees, remains about three years in the larva state, and increases, according to Lyonet, to the amount of 72,000 times its first weight.

Newport adds that those larvæ in which the proportion of increase is the greatest, are usually those which remain longest in the pupa state, as in the silkworm. “Thus Redi observed in the maggots of the common flesh-flies a rate of increase amounting to about 200 times the original weight in 24 hours, but the proportion of increase in these larvæ does not at all approach that of the Sphinx and Cossus.” From his observations on the larva of one of the wild bees (Anthophora retusa) Newport believes that this is also the case with the Hymenoptera. The weight of the egg of this insect is about 1
150 of a grain, and the average of a full-grown larva 68
10 grains, so that its increase is about 1020 times its original weight; “which compared with that of the Sphinx of medium size, is but as 1 to 9¾, and to a Sphinx of maximum size only as 1 to a little more than 11.”

The growth is most rapid after the last moult. “Thus a larva of Sphinx ligustri, which at its last change weighed only about 19 to 20 grains, at the expiration of eight days, when it was fully grown, weighed nearly 120 grains.” (Newport.)

d. The process of moulting (ecdysis)

Insects periodically shed the exoskeleton, together with the chitinous lining of their internal organs of ectodermal origin, which thus sloughed off are called the exuvia. The process in the locust has been described by Riley.^[95] It occupies from half to three-quarters of an hour (Fig. 565). This process has naturally, from the ease with which it can be observed, been most frequently examined in the Lepidoptera, though careful and detailed observations of the inner and outer changes are still greatly needed, especially in other orders. In the caterpillar of most moths, especially one of the more generalized bombycine moths, on slipping out of its egg-shell the head is of enormous size as compared with the body, but the latter soon fills out after the creature has eaten a few hours; the head, of course, does not during this time increase in size, and the larvæ of different instars may be exactly distinguished, as Dyar has shown, by the measurements of the head.

Before the caterpillar moults, it stops feeding, and the head is now small compared with the body; the head of the second instar is now large, situated partly under the much-swollen prothoracic segment, and pushes the head of the first instar forward.

Newport has well described the mode of shedding the skin in Sphinx ligustri, and his detailed description will apply to most lepidopterous larvæ.

The whole body is wrinkled and contracted in length, and there are occasionally powerful contractions and twitchings of its entire body; the skin becomes dry and shrivelled, and is gradually separated from the new and very delicate one of the next instar beneath. After several powerful efforts of the larva the old skin cracks along the middle of the dorsal surface of the mesothoracic segment, and by repeated efforts the fissure is extended into the 1st and 3d segment, while the covering of the head divides along the vertex and on each side of the clypeus. “The larva then gradually presses itself through the opening, withdrawing first its head and thoracic legs, and subsequently the remainder of its body, slipping off the skin from behind like the finger of a glove. This process, after the skin has once been ruptured, seldom lasts more than a few minutes. When first changed the larva is exceedingly delicate, and its head, which does not increase in size until it again changes its skin, is very large in proportion to the rest of the body.” (Art. Insecta, etc.)

Trouvelot’s account is more detailed and an advance on that of Newport’s view. He explicitly states, and we know that he was a very close observer, that the old skin is detached by “a fluid which circulates between it and the worm.” His account is as follows: The polyphemus worm, like all other silkworms, changes its skin five times during its larval life. The moulting takes place at regular periods, which comes around about every 10 days for the first four moultings, while about 20 days elapse between the fourth and fifth moulting. The worm ceases to eat for a day before moulting, and spins some silk on the vein of the under surface of a leaf; it then secures the hooks of its hind legs in the texture it has thus spun, and there remains motionless; soon after, through the transparency of the skin of the neck, can be seen a second head larger than the first, belonging to the larva within. The moulting generally takes place after four o’clock in the afternoon; a little before this time the worm holds its body erect, grasping the leaf with the two pairs of hind legs only; the skin is wrinkled and detached from the body by a fluid which circulates between it and the worm; two longitudinal bands are seen on each side, produced by a portion of the lining of the spiracles, which at this moment have been partly detached; meanwhile the contractions of the worm are very energetic, and by them the skin is pulled off and pushed towards the posterior part; the skin thus becomes so extended that it soon tears just under the neck, and then from the head. When this is accomplished the most difficult operation is over, and now the process of moulting goes on very rapidly. By repeated contractions the skin is folded towards the tail, like a glove when taken off, and the lining of the spiracles comes out in long white filaments. When about one-half of the body appears, the shell still remains like a cap, enclosing the jaws; then the worm, as if reminded of this loose skull-cap, removes it by rubbing it on a leaf; this done, the worm finally crawls out of its skin, which is attached to the fastening made for the purpose. Once out of its old skin, the worm makes a careful review of the operation, with its head feeling the aperture of every spiracle, as well as the tail, probably for the purpose of removing any broken fragment of skin which might have remained in these delicate organs. Not only is the outer skin cast off, but also the lining of the air-tubes and intestines, together with all the chewing organs and other appendages of the head. After the moulting, the size of the larva is considerably increased, the head is large compared with the body, but 8 or 10 days later it will look small, as the body will have increased very much in size. This is a certain indication that the worm is about to moult. Every 10 days the same operation is repeated. From the fourth moulting to the time of beginning the cocoon the period is about 16 days. (Amer. Naturalist, i, pp. 37, 38.)

Little has been recorded as to the exact mode of casting the larval skin in Coleoptera. Slingerland states that Euphoria inda when pupating sheds the larval skin off the anal end in the same way as in caterpillars, while in Pelidnota punctata the larval skin splits down the whole length of the back, retains the larval shape, and forms a covering for the pupa which remains inside. (Can. Entomologist, xxix, p. 52.) The old larval skin in the Coccinellidæ and certain Chrysomelidæ is retained crumpled up at the end of the body, while in Dermestes, Anthrenus, etc., it cloaks the pupa.

Not only is the integument, with its hairs, setæ, and other armatures, as well as the cornea or facets of the eyes, shed, but also all the lining or intima of those internal organs which have been originally derived by an ingrowth or invagination of the ectoderm are likewise cast off, with the probable exception, of course, of the mid-intestine, which is endodermal in its origin. Even so early an observer as Swammerdam noticed that the internal lining of the alimentary canal comes away with the skin. He states that the larva of Oryctes nasicornis sheds both the lining of the colon, and of the smaller as well as the larger branches of the tracheæ.

Careful observations are still needed on the internal changes at ecdysis of most insects. Newport seems to have observed more closely than any one else, notwithstanding the great number who have reared caterpillars but have not carefully observed these points, the extent of the process internally. He informs us: “The lining of the mouth and pharynx, with that of the mandibles, is detached with the covering of the head, and that of the large intestines with the skin of the posterior part of the body, and besides these also the lining of the tracheal tubes. The lining of the stomach itself, or the portion of the alimentary canal which extends from the termination of the œsophagus to the insertion of the so-called biliary vessels, is also detached, and becomes completely disintegrated, and appears to constitute part of the meconium voided by the insect on assuming its imago state.” (Art. Insecta, p. 876.) Newport states on another occasion that he had “noticed the remarkable circumstance [now explained by the fact that the mid-intestine is of endodermal origin] that the mucous lining of the true ventriculus was not cast off with the rest, but was discharged with the fecula.”^[96] Burmeister also observed that the smaller tracheæ as well as the internal tunic of the colon of Libellulæ are shed.

In the apodous larvæ of Hymenoptera which live in cells, as we have observed in those of Bombus, during the process of moulting, the delicate skin breaks away in shreds, probably owing to the tension due to the unequal growth of the different parts of the body. “Thus after the skin beneath has fully formed, shreds of the former skin remain about the mouth-parts, the spiracles, and anus. Upon pulling upon these, the lining of the alimentary tube and tracheæ can be drawn out, sometimes, in the former case, to the length of several lines.”^[97] We then added, “As all these internal systems of vessels are destined to change their form in the pupa, it may be laid down as a rule in the moulting of insects and Crustacea, that the lining of the internal organs, which is simply a continuation of the outer tegument, or arthroderm, is, in the process of moulting, sloughed off with that outer integument.” We have satisfied ourself that in the larvæ of the Lepidoptera (e.g. Datana) the tracheæ at the time of ecdysis undergo a complete histolysis, and arise de novo from hypodermal cells, the so-called spiral threads originating from elongated peritracheal nuclei. (See p. 449, Fig. 412.) This is undoubtedly also the case with the salivary ducts, which are strengthened and rendered elastic by tænidia like those of tracheæ. As the urinary tubes are diverticula of the proctodæum, itself an ectodermal invagination, they may also, though not lined with a chitinous intima, be renewed. With little doubt the intima of the ducts of poison, spinning, and most, if not all the other glands, though certainly the dermal glands, is exuviated. We have found that the lobster in moulting sheds, besides the skin with the most delicate setæ, the lining of the proventriculus, and the apodemes of the head and thorax, hence it is most probable that the tentorium of the head of insects as well as the apodemes and phragmas of the thorax are exuviated.

The formation of the inner skin, or that of any succeeding stage (instar), is due to the secretion of the structureless chitinous layer by the cells of the hypodermis, during the process of histogenesis. These cells at this time are very active, and the formation of the new layer of chitin arrests the supply of nourishment to the old skin, so that it dries, hardens, and with the aid of the fluid thrown out at this time separates from the new chitinous layer secreted by the hypodermis.

Mention of this fluid, which Newport was the first to observe, and which he says causes the separation of the old from the underlying fresh integument of the caterpillar, recalls a passage in Hatchett-Jackson’s Studies in the morphology of the Lepidoptera, which we quote on a succeeding page, where he calls attention to the formation of such a liquid, which in the reptiles facilitates the process of moulting, adding, “Whether such is the case with the moult of the caterpillar, I do not know.” Is it not also possible that the growth of the setæ or tubercles on the cuticle of the caterpillar may likewise serve to loosen and detach the overlying skin about to be cast off? After writing the foregoing, we find that Miall and Denny have suggested that the setæ of the cockroach probably serve the same purpose as the casting-hairs of the crayfish and reptiles.

It is well known that in the crayfish and in lizards the skin is first loosened by the growth of temporary hairs or setæ, which locally grow inward from the old cuticle and push the skin away when it is shuffled off by the movements of the body, jaws, and limbs, as well as the body in general.^[98]

Such spines arise in the pupa of many insects, for Verhoeff finds that the spines and teeth of pupal fossorial and other Hymenoptera, as well as Coleoptera, function as moulting-processes for loosening and pushing off the last larval skin, rather than for locomotion. He also claims that the spines of the pupa of the dipterous Anthrax are both for locomotion and for boring, especially the spines on the head and tail. He therefore divides these pupal spines into helcodermatous (boring or tearing) and locomotor spines.

Gonin has fully confirmed Newport’s discovery of the exuvial fluid. He states that during pupation the outside of the pupa, especially the parts of the head and thorax “is coated with a viscous liquid secreted by special glands.” The parts only harden subsequent to pupation after exposure to the air (p. 41). His observations were made under the direction of Professor Bugnion, who kindly writes us:—

“M. Gonin has proved the formation of a liquid which passes under the cuticle at the time of the last moult and facilitates exuviation. We think that this liquid is secreted by large cells (unicellular glands) which we see especially on the surface of segments 1–3. These cells form part of the hypodermis, and their pores open under the cuticle.”

In a subsequent letter enclosing a sketch kindly made for me by M. Gonin (Fig. 566), Professor Bugnion writes me Aug. 24, 1897, regarding the functions of the large hypodermal cells (l. hy), as follows: “It seems to me, in fact, after having again examined the sections, that the function of these cells is not sufficiently elucidated. Indeed these cells occur only in the section passing through the 1st segment, between the head and 1st thoracic segment. It would seem, if these cells supply the liquid which lubricates the surface at the time of ecdysis, that they should be spread over the entire surface of the body. Moreover, these cells have no distinct orifice, and although there is seen at times to issue streams of a substance (coagulated by the reagents), they cannot be compared with true unicellular glands like those of the epidermis of fishes, amphibians, etc.

“On the other hand, if it is the blood which oozes out on the surface (according to your hypothesis), it would seem that the loss of blood would cause the death of the larva. I believe then it is due to the secretion of the hypodermis which spreads over the whole surface when the cells are still soft (not yet hardened from contact with the air). At all events, there is a liquid spread over the surface; it is this liquid which glues the wings and the legs to the body at the moment the caterpillar issues from the rent in its skin. If at this instant we plunge the pupa in the water the liquid is dissolved, and the feet, wings, etc., are not glued to the body.”

Dr. T. A. Chapman also writes us: “There is no question about the existence of a fluid between the two skins at moulting. In hairy larvae the hairs are always wet at first, or if the skin be renewed rather more quickly than the larva does it naturally, the wetness of both surfaces is obvious. I do not know the nature of the fluid, but it is related to that which hardens into the dense pupal case, and also hardens in a less degree the skin of the larva. I suppose it must contain some chitin in a soluble form. If a newly cast larva skin be taken, there is no difficulty in extending the shrivelled mass to its full length and dimensions, but if a short time elapses, this chitin hardens, and the skin cannot be extended after soaking in water, alcohol, ammonia, or any other solvent I have tried.”

It has been stated that there is a subimaginal pellicle in Lepidoptera, but as Dr. Chapman writes me, “what has been observed has been some of the inner pupal dissepiments, such as the pupal cases of the under wings,” etc. They may be observed in the head of the tineid pupæ, and other small moths. We have thought that the delicate, purplish, powdery layer left in the cast shells of the pupæ of saturnians, Catocalæ, and other moths, might possibly be such a pellicle, but this view has been dispelled by the following statement of Professor Bugnion in a letter answering an inquiry whether he had noticed such a pellicle.

“A liquid which is secreted in a few minutes at the time of the last moult, forms in drying a yellowish layer spotted with black (in Pieris brassicæ). This layer extends around the entire pupa, and serves both to protect it and to glue together the wings, legs, etc., in their new position. The dried liquid on the surface of the pupa, and by means of which the appendages are glued to the surface, very likely corresponds to the pellicle of which you speak.” The newly exposed integument is at first pale and colorless, but soon assumes the hues peculiar to the species, and the insect, at first exhausted, after a short rest becomes active.