The duration of a caterpillar's life varies. It may extend only to a fortnight, or to three years in Cossus. The number of moults is generally three or four. In this process the old cuticle as a rule splits on the back of the thoracic somites, and the split extends to a variable degree both forwards and backwards. The three scales of the head may separate also. The cuticle of the oesophagus and intestine appears to be cast at least in the final moult. Previous to pupation the caterpillar may suspend itself either by the anal prolegs, e.g. Vanessa, or secure itself by the anal prolegs and a band of silk round the thorax, e. g. Pieris; or may fashion a silken cocoon with or without an admixture of foreign bodies; or else it buries itself as do the majority of Sphingidae. A few members of this family, however, e.g. Chaerocampa Nerii, form a cocoon on or near the surface of the ground.

Note the following points of internal anatomy. The nervous system comprises a supra- and an infra-oesophageal ganglion and a series of ten ventral ganglia, of which three are thoracic and seven abdominal. The tenth is often double, e. g. in Acherontia. The commissures between the three thoracic ganglia are generally divaricated. The respiratory sympathetic system consists of a nerve running on the dorsal aspect of the nerve-chain from one ganglion to the ganglion next following. In the abdominal region this nerve branches to right and left just in front of the ganglion, with which two short filaments connect the branches at their origin. The branches are the nervi transversi or respiratorii, and supply the tracheae and stigmata. In the thoracic region the nerve breaks up into three primary branches. The median divides right and left as in the abdominal region, and supplies the longitudinal trunk which unites the prothoracic to the first abdominal stigma, as well as the tracheae it gives off. Each side branch runs backwards and unites with the ganglion, first of all giving off a lateral branch, the nervus lateralis transversus of Cattie, which joins the first nerve given off by the ganglion.

The stomatogastric system consists, according to Newport, of a ganglion frontale and nervus recurrens: and of two ganglia in connection (like the paired system of the Cockroach) with the posterior aspect of the supra-oesophageal ganglion. The latter supply nerves to the tracheae of the head. The ocelli, as usual, have each a single lens. Beneath the lens lies a number of black pigment cells imbedding a crystalline body composed of several parts (Carriere).

The stigmata possess a complete apparatus by which they may be closed. The tracheal stems arising from them are all connected on each side by a longitudinal trunk which in the meso- and meta-thoracic somites gives off an external branch furnished with the usual closing apparatus of a stigma. No external indications of stigmata, however, are visible in these somites in Sphinx, but Prof. Westwood possesses a dried specimen of Cossus in which they are clearly visible. W. Muller-Blumenau has found an aquatic Lepidopterous larva (Cataclysta pyropalis) living in Brazil, in which all the stigmata are closed and respiration is carried on by filamentous branchiae, but there are two closed thoracic stigmata, which he places between the pro- and meso-thorax and between the meso- and meta-thorax. He states that he has also found them in many terrestrial larvae. The wings arise in close connection with the meso- and meta-thoracic branches. There are three pairs of cephalic stigmata in the embryo (Hatschek).

The digestive tract consists of an oesophagus, and an intestine and rectum lined by a cuticle, and a chylific stomach or mesenteron. There are six Malpighian tubes with a beaded exterior which open as usual into the beginning of the intestine. Two short salivary glands open into the mouth, and a pair of serictaria or silk glands by a common duct on the spinneret1. The rudimentary genital organs lie one on either side the heart under the fifth abdominal tergum. A delicate filament may be traced from each organ round the intestine. Herold figures these filaments as extending in Pieris Brassicae, in the male to the anterior border of the ninth somite; in the female to the anterior border of the eighth somite, and into connection with two oval bodies at its posterior border. An abundant fat body fills the body-cavity, or coelome. The blood is acid as in all caterpillars with one exception hitherto examined, and contains amoeboid blood-corpuscles. It is green, and the colour is due to metachlorophyl (Poulton).

There are two types of larval (or young) Insecta: one known as Campodei-form, from a more or less close resemblance to the genus Campodea among Thy-sanura; the other as eruciform, of which a caterpillar may be taken as a good example. The Campodeiform larva has the typical regions of the body clearly defined, mouth parts adapted for biting, well developed ambulatory thoracic limbs, and frequently terminal abdominal jointed appendages. The outlines of the body are even, and the series of somites follow one another without any abrupt change of shape. Such a larva is seen, more or less adapted of course, in most Ametabola and Hemimetabola. The condition of the adult insect is acquired in a graduated series of moults, whilst organs such as wings make their appearance; and others e.g. the genitalia, are brought to maturity. It is also seen in some Metabola, as in certain families of Coleoptera, e.g. Cicindilidae, Dytiscidae, etc.; in some Neuroptera, e.g. Sialidae; some Megaloptera. In these instances it is occasionally highly specialised e.g. in the Ant-lion. It appears in others as a first larval form, subsequently modified in accordance with a change of habit in the direction of the eruciform type, constituting what is known as Hypermetamorphosis. Such a change occurs in Mantispa among Panorpaiae, in the Strepsiptera, and in various Meloidae among Coleoptera, e.g. Meloe, Sitaris, Hornia, Epicauta, Cantharis. In Mantispa the second larval form is only sub-eruciform: in the Meloidae there is a number of forms one after the other more and more degraded. (See Packard, American Naturalist, vii. 1883, pp. 938-944; Riley, op. id. xii. 1878, pp. 213, 282.)

1 Poletajew (Z. A. viii. 1885) states that the silken thread is single in the Tenthredinidae, but double and twisted in Lepidoptera. Hence he disputes the statement that the ducts of the serictaria have a common outlet.

The eruciform type of larva is very generally found among Metabola. It may succeed, as just mentioned, a Campodeiform larva, and in its simplest shape, e. g. in Trichoptera, which may be termed sub-eruciform, it is little more than a persistent embryonic form such as is observable in the earliest stages of most Insecta. The caterpillar of the Lepidoptera, of the Panorpatae, of Tenthredinidae among Hymenoptera, and the grubs of some Coleoptera are typical examples of the type. The head is well-defined, but the somites of the body are simple and cylindrical (homono-mous), and the animal has a vermiform aspect. Nevertheless a thoracic region with articulated limbs is distinguishable, and what is more the abdomen (except in Coleopterous forms) possesses functional abdominal limbs. Such limbs, but not functional, exist in Thysanura, and as rudiments in the embryoes of most Insecta, e.g. in Hydrophilus (Dytiscidae), which has a modified Campodeiform larva. The limbless grub of Aculeate Hymenoptera, and still more the maggot of Diptera, must be regarded as degenerate examples of this type: and where limbs are present in a Dipterous larva they are probably secondary and special developments.

The larvae of all existing Insecta, and even the primitive Ametabolous order Thysanura have been modified to a greater or less extent by Natural Selection. The embryo however has a type of structure which is readily modified in the direction of one of the two larval types; and it is not surprising to find larval forms, such as those of Trichoptera, which may be regarded as actually transitional between the two.

Larvae of British Lepidoptera and their Food plants, Wilson, London, 1880. Figures in Horsefield and Moore, Catalogue of Lepidopterous Insects in East India House Museum, 2 vols., 1858-59, and in Dewitz, Jugendstadien exot. Lepidoptera, Nova Acta, 44, 1883. Aquatic Lepidopterous larvae. Muller-Blumenau, A. N. 50, 1884; Maurice, Bull. Scientifiques du Departement du Nord, iv.

Anatomy.Cossus, Lyonet, Traite anatomique de la Chenille qui ronge le bois du Saule, Hague, 1762. Sphinx, Newport (and also as to pupa and imago) 'Insecta'; Encyclopaedia of Anat. and Phys. ii., London, 1836-39.

Skin glands.Klemensiewics, Verhandl. K. K., zool. bot. Gesellschaft, Wien, xxxii. 1883. Glandular hairs. Dimmock, Psyche, iv. 1885.

Muscles.Lubbock, on Pygaera bucephala. Tr. L. S. xxii. 1859, Stigmata.Krancher, Z. W. Z. xxxv. 1881.

Serictaria.Helm, Z. W. Z. xxvi. 1876; Lidth van Jeude, Z. A. i. 1878; Joseph, Z. A. iii. 1880.

Nervous system.Newport, Ph. Tr. cxxii. 1832; cxxiv. 1834 (includes Vanessa Urticae); Cattie, Z. W. Z. xxxv. 1880.

Blood.E. B. Poulton, P. R. S. xxxviii. 1885.

Protective coloration and markings. Poulton, op. cit.; ibid. xl. 1886, and Trans. Entomol. Soc. 1884 and 1885; Weismann, Studies in Theory of Descent (translated by Meldola), London, 1880-82; Cameron, on Smerinthus, Trans. Entomol. Soc, 1880. On colour, see also Hagen, Proc. American Acad. (2) ix. 1882; F. Muller, Kosmos, xii.; Wallace, Tropical Nature, London, 1878, pp. 158, 249.