Alfred Denny

The Outer Skeleton.

---

SPECIAL REFERENCES.

Krukenberg. Vergleichend-Physiologische VortrÄge. IV.—Vergl. Physiologie der Thierischen GerÜstsubstanzen. (1885.) [Chemical Relations of Chitin.]

Graber. Ueber eine Art fibrilloiden Bindegewebes der Insectenhaut. Arch. f. mikr. Anat. Bd. X. (1874.) [Minute Structure of Integument.] Also,

Viallanes. Recherches sur l’Histologie des Insectes. Ann. Sci. Nat., Zool. VI^e SÉrie, Tom. XIV. (1882).

Audouin. Recherches anatomiques sur le thorax des Insectes, &c. Ann. Sci. Nat. Tom. I. (1824.) [Theoretical Composition of Insect Segments.] Also,

Milne-Edwards. LeÇons sur la Physiologie et l’Anatomie ComparÉe. Tom. X. (1874.)

Savigny. MÉmoires sur les animaux sans vertÈbres. Partie I^e. ThÉorie des organes de la bouche des CrustacÉes et des Insectes. (1816.) [Comparative Anatomy of the Mouth-parts.]

Muhr. Ueber die Mundtheile der Orthopteren. Prag. 1877. [Mouth-parts of Orthoptera.]

---

Chitin.

When the skin of an Insect is boiled successively in acids, alkalies, alcohol, and ether, an insoluble residue known as Chitin (C₁₅H₂₆N₂O₁₀) is obtained. It may be recognised and sufficiently separated by its resistance to boiling liquor potassÆ. Chitin forms less than one-half by weight of the integument, but it is so coherent and uniformly distributed that when isolated by chemical reagents, and even when cautiously calcined, it retains its original organised form. The colour which it frequently exhibits is not due to any essential ingredient; it may be diminished or even destroyed by various bleaching processes. The colouring-matter of the chitin of the Cockroach, which is amber-yellow in thin sheets and blackish-brown in dense masses, is particularly stable and difficult of removal. Its composition does not appear to have been ascertained; it is white when first secreted, but darkens on exposure to air. Fresh-moulted Cockroaches are white, but gradually darken in three or four hours. Lowne30 observes that in the Blow-fly the pigment is “first to be met with in the fat-bodies of the larvÆ. These are perfectly white, but when cut from the larva, and exposed to the air, they rapidly assume an inky blackness.... When the perfect insect emerges from the pupa, and respiration again commences, the integument is nearly white, or a faint ashy colour prevails. This soon gives place to the characteristic blue or violet tint, first immediately around those portions most largely supplied with air vessels.” Professor Moseley31 tells us that, thinking it just within the limits of possibility that the brown coloration of the Cockroach might be due to the presence of silver, he analysed one pound weight of Blatta. He found no silver, but plenty of iron, and a remarkable quantity of manganese. That light has some action upon the colouring matter seems to be indicated by the fact that in a newly-moulted Cockroach the dorsal surface darkens first.

Chitin is not peculiar to Insects, nor even to Arthropoda. The pen of cuttle-fishes and the shell of Lingula contain the same substance,32 which is also proved, or suspected, to occur in many other animals.

The chemical stability of chitin is so remarkable that we might well expect it to accumulate like the inorganic constituents of animal skeletons, and form permanent deposits. Schlossberger33 has, however, shown that it changes slowly under the action of water. Chitin kept for a year under water partially dissolved, turned into a slimy mass, and gave off a peculiar smell. This looks as if it were liable to putrefaction. The minute proportion of nitrogen in its composition may explain the complete disappearance of chitin in nature.

The Chitinous Cuticle.

Fig. 8.—Diagram of Insect integument, in section. bm, basement membrane; hyp, hypodermis, or chitinogenous layer; ct, ct', chitinous cuticle; s, a seta.

The chitinous exoskeleton is rather an exudation than a true tissue. It is not made up of cells, but of many superposed laminÆ, secreted by an underlying epithelium, or “chitinogenous layer.” This consists of a single layer of flattened cells, resting upon a basement membrane. A cross-section of the chitinous layer, or “cuticle,” examined with a high power shows extremely close and fine lines perpendicular to the laminÆ. The cells commonly form a mosaic pattern, as if altered in shape by mutual pressure. The free surface of the integument of the Cockroach is divided into polygonal, raised spaces. Here and there an unusually long, flask-shaped, epithelial cell projects through the cuticle, and forms for itself an elongate chitinous sheath, commonly articulated at the base; such hollow sheaths form the hairs or setÆ of Insects—structures quite different histologically from the hairs of Vertebrates.

The polygonal areas of the cuticle correspond each to a chitinogenous cell. Larger areas, around which the surrounding ones are radiately grouped, are discerned at intervals, and these carry hairs, or give attachment to muscular fibres.

Viallanes (loc. cit.) has added some interesting details to what was previously known of Insect-hairs. There are, he points out, two kinds of hairs, distinguished by their size and structure. The smaller spring from the boundary between contiguous polygonal areas, and have no sensory character. The larger ones occupy unusually large areas, surmount chitinogenous cells of corresponding size, and receive a special nervous supply. The nerve34 expands at the base of the hair into a spindle-shaped, nucleated mass (bipolar ganglion-cell), from which issues a filament which traverses the axis of the hair, piercing the chitinogenous cell, whose protoplasm surrounds it with a sheath which is continued to the tip of the hair. Such sensory hairs are abundant in parts which are endowed with special sensibility.

Fig. 9.—Nerve-end­ing in skin of Stratio­mys larva. h, hairs; b, their chi­tin­ous base; c, nuc­leus of gen­er­at­ing cell; g, gan­glion cell. ×250. Copied from Viallanes.

Fig. 10.—Diagram of sen­sory hair of Insect. Cc, chi­ti­nous cu­ti­cle; h, hair; c, its gen­er­at­ing cell; g, gan­glion cell; bm, base­ment-mem­brane.

The chitinous cuticle is often folded in so as to form a deep pit, which, looked at from the inside of the body, resembles a lever, or a hook. Such inward-directed processes serve chiefly for the attachment of muscles, and are termed apodemes (apodemata). A simple example is afforded by the two glove-tips which lie in the middle line of the under-surface of the thorax (p.58, and fig.27). In other cases the pit is closed from the first, and the apodeme is formed in the midst of a group of chitinogenous cells distant from the superficial layer, though continuous therewith. Many tendons of insertion are formed in this way. The two forked processes in the floor of the thorax (p.58, and fig.27) are unusually large and complex structures of the same kind. In the tentorium of the head (p.39, and fig.17) a pair of apodemes are supposed to unite and form an extensive platform which supports the brain and gullet.

Fig. 11.—Nymph (in last larval stage) escaping from old skin. ×2 1/2.

Like other Arthropoda, Insects shed their chitinous cuticle from time to time. A new cuticle, at first soft and colourless, is previously secreted, and from it the old one gradually becomes detached. The setÆ probably serve the same purpose as the “casting-hairs” described by Braun in the crayfish, and by Cartier in certain reptiles,35 that is, they mechanically loosen the old skin by pushing beneath it. In many soft-bodied nymphs the new skin can be gathered up into a multitude of fine wrinkles, which facilitate separation, but we have not found such wrinkles in the Cockroach, except in the wings. The integument about to be shed splits along the back of the Cockroach, from the head to the end of the thorax,36 and the animal draws its limbs out of their discarded sheaths with much effort. It is remarkable that the long, tapering, and many-jointed antennÆ are drawn out from an entire sheath. At the same time the chitinous lining of the tracheal tubes is cast, while that of the alimentary canal is broken up and passed through the body.

Fig. 12.—Cast skin of older nymph (“pupa”). ×2 1/2.

Prolonged boiling in caustic potash, though it dissolves the viscera, does not disintegrate the exoskeleton. This shows that the segments of the integument are not separate chitinous rings, but thickenings of a continuous chitinous investment. Nevertheless, their constancy in position and their conformity in structure often enable us to trace homologies between different segments and different species as certainly as between corresponding elements of the osseous vertebrate skeleton.

Parts of a Somite.

Audouin’s laborious researches into the exoskeleton of Insects37 resulted in a nomenclature which has been generally adopted. He divides each somite (segment) into eight pieces, grouped in pairs—viz., terga (dorsal plates), sterna (ventral plates), epimera (adjacent to the terga), and episterna (adjacent to the sterna).

While admitting the usefulness of these terms, we must warn the reader not to suppose that this subdivision is either normal or primitive. The eight-parted segment exists in no single larval or adult Arthropod. Lower forms and younger stages take us further from such a type, instead of nearer to it; and Audouin’s theoretical conception is most fully realised in the thorax of an adult Insect with well-developed legs and wings.

The morphologist would derive all the varieties of Arthropod segments from the very simple and uniform chitinous cuticle found in Annelids and many Insect-larvÆ. This becomes differentiated by unequal thickening and folding in, and a series of rings connected by flexible membranes is produced. Locomotive and respiratory activity commonly lead to the definition of terga and sterna, which are similarly attached to each other by flexible membranes. A pair of limbs may next be inserted between the terga and sterna, and the simple segment thus composed occurs so extensively in the less modified regions and in early stages that it may well be considered the typical Arthropod somite.

Special needs may lead to the division of the sterna into lateral halves, but this is purely an adaptive change. The third thoracic sternum of the male Cockroach, and the second and third of the female are thus divided, as is also the hinder part of the seventh abdominal sternum of the female.

In an early stage every somite has its tergal region divided into lateral halves, owing to the late completion of the body on this side. Traces of this division may survive even in the imago. There is often a conspicuous dorsal groove in the thoracic terga, and at the time of moult the terga split along an accurately median line (see fig.12).

Additional pieces may be developed between the terga and sterna, and these have long been termed pleural.38 There may be, for example, single stigmatic plates, as in the abdomen of the Cockroach, pieces to support the thoracic legs, and pieces to support the wings; but the number and position of these plates depends upon their immediate use, and their homologies become very uncertain when Insects of different orders are compared. In general, the pleural elements of the segment are late in development, variable, and highly adaptive.

Somites of the Cockroach.

The exoskeleton of the Cockroach is divisible into about seventeen segments, which are grouped into three regions, as follows:—

Head		Procephalic lobes	3
		Post-oral segments
Thorax			3	39
Abdomen			11
			—
			17
			—

It is a strong argument in favour of this estimate that many Insects, at the time when segmentation first appears, possess seventeen segments.40 The procephalic lobes, from which a great part of the head, including the antennÆ, is developed, are often counted as an additional segment.41

The limbs, which in less specialised Arthropoda are carried with great regularity on every segment of the body, are greatly reduced in Insects. Those borne by the head are converted into sensory and masticatory organs; those on the abdomen are either totally suppressed, or extremely modified, and only the thoracic limbs remain capable of aiding in locomotion.

The primitive structure of the Arthropod limb is adapted to locomotion in water, and persists, with little modification, in most Crustacea. Here we find in most of the appendages42 a basal stalk (protopodite), often two-jointed, an inner terminal branch (endopodite), and an outer terminal branch (exopodite), each of the latter commonly consisting of several joints. It does not appear that the appendages of Insects conform to the biramous Crustacean type, though the ends of the maxillÆ are often divided into an outer and an inner portion.

We shall now proceed to describe, in some detail, the regions of the body of the adult Cockroach.

Head; Central Parts.

Fig. 13.—Front of Head. ×10.

The head of the Cockroach, as seen from the front, is pear-shaped, having a semi-circular outline above, and narrowing downwards. A side-view shows that the front and back are flattish, while the top and sides are regularly rounded. In the living animal the face is usually inclined downwards, but it can be tilted till the lower end projects considerably forward. The mouth, surrounded by gnathites or jaws, opens below. On the hinder surface is the occipital foramen, by which the head communicates with the thorax. A rather long neck allows the head to be retracted beneath the pronotum (first dorsal shield of the thorax) or protruded beyond it.

On the front of the head we observe the clypeus, which occupies a large central tract, extending almost completely across the widest part of the face. It is divided above by a sharply bent suture from the two epicranial plates, which form the top of the head as well as a great part of its back and sides. The labrum hangs like a flap from its lower edge. A little above the articulation of the labrum the width of the clypeus is suddenly reduced, as if a squarish piece had been cut out of each lower corner. In the re-entrant angle so formed, the ginglymus, or anterior articulation of the mandible, is situated.

The labrum is narrower than the clypeus, and of squarish shape, the lower angles being rounded. It hangs downwards, with a slight inclination backwards towards the mouth, whose front wall it forms. On each side, about halfway between the lateral margin and the middle line, the posterior surface of the labrum is strengthened by a vertical chitinous slip set with large setÆ. Each of these plates passes above into a ring, from the upper and outer part of which a short lever passes upwards, and gives attachment to a muscle (levator menti).

Fig. 14.—Top of Head. ep, epi­cra­nial plate; oc, eye; ge, gena. ×10.

The top and back of the head are defended by the two epicranial plates, which meet along the middle line, but diverge widely as they descend upon the posterior surface, thus enclosing a large opening, the occipital foramen. Beyond the foramen, they pass still further downwards, their inner edges receding in a sharp curve from the vertical line, and end below in cavities for the articulation of the mandibular condyles.43

The sides of the head are completed by the eyes and the genÆ. The large compound eye is bounded above by the epicranium; in front by a narrow band which connects the epicranium with the clypeus; behind, by the gena. The gena passes downwards between the eye and the epicranial plate, then curves forwards beneath the eye, and just appears upon the front of the face, being loosely connected at this point with the clypeus. Its lower edge overlaps the base of the mandible, and encloses the extensor mandibulÆ.

Fig. 15.—Side of Head. oc, eye; ge, gena; mn, man­dible. ×10.

Fig. 16.—Back of Head. ca, cardo; st, stipes; ga, galea; la, lacinia; pa, palp; sm, sub­mentum; m, men­tum; pg, para­glossa. ×10.

The occipital foramen has the form of an heraldic shield. Its lateral margin is strengthened by a rim continuous with the tentorium, or internal skeleton of the head. Below, the foramen is completed by the upper edge of the tentorial plate, which nearly coincides with the upper edge of the submentum (basal piece of the second pair of maxillÆ); a cleft, however, divides the two, through which nerve-commissures pass from the sub-oesophageal to the first thoracic ganglion. Through the occipital foramen pass the oesophagus, the salivary ducts, the aorta, and the tracheal tubes for the supply of air to the head.

Fig. 17.—Fore-half of Head, with ten­tor­ium, seen from be­hind. ×12.

The internal skeleton of the head consists of a nearly transparent chitinous septum, named tentorium by Burmeister, which extends downwards and forwards from the lower border of the occipital foramen. In front it gives off two long crura, or props, which pass to the ginglymus, and are reflected thence upon the inner surface of the clypeus, ascending as high as the antennary socket, round which they form a kind of rim. Each crus is twisted, so that the front surface becomes first internal and then posterior as it passes towards the clypeus. The form of the tentorium is in other respects readily understood from the figure (fig.17). Its lower surface is strengthened by a median keel which gives attachment to muscles. The oesophagus passes upwards between its anterior crura, the long flexor of the mandible lies on each side of the central plate; the supra-oesophageal ganglion rests on the plate above, and the sub-oesophageal ganglion lies below it, the nerve-cords which unite the two passing through the circular aperture. A similar internal chitinous skeleton occurs in the heads of other Orthoptera, as well as in Neuroptera and Lepidoptera. PalmÉn44 thinks that it represents a pair of stigmata or spiracles, which have thus become modified for muscular attachment, their respiratory function being wholly lost. In Ephemera he finds that the tentorium breaks across the middle when the skin is changed, and each half is drawn out from the head like the chitinous lining of a tracheal tube.

AntennÆ; Eyes.

Fig. 18.—Base of Antenna of Male (to left) and Fe­male (to right). ×24.

A pair of antennÆ spring from the front of the head. In the male of the common Cockroach they are a little longer than the body; in the female rather shorter. From seventy-five to ninety joints are usually found, and the three basal joints are larger than the rest. Up to about the thirtieth, the joints are about twice as wide as long; from this point they become more elongate. The joints are connected by flexible membranes, and provided with stiff, forward-directed bristles. The ordinary position of the antennÆ is forwards and outwards.

Each antenna is attached to a relatively large socket (fig.15), which lies between the epicranium and clypeus, to the front and inner side of the compound eyes. A flexible membrane unites the antenna to the margin of the socket, from the lower part of which a chitinous pin projects upwards and supports the basal joint.

It is well known that in many Crustacea two pairs of antennÆ are developed, the foremost pair (antennules) bearing two complete filaments. Some writers have suggested that both pairs may be present in Insects, though not simultaneously, the Crustacean antennule being found in the larva, and the Crustacean antenna in the adult. This view was supported by the familiar fact that in many larvÆ the antennÆ are placed further forward than in the adult. The three large joints at the base of Orthopterous antennÆ have been taken to correspond with those of Crustacean antennules, and it has been inferred that in Insects with incomplete metamorphosis, only antennules or larval antennÆ are developed.45 This reasoning was never very cogent, and it has been impaired by further inquiry. Weismann has shown that in Corethra plumicornis, the adult antenna, though inserted much further back than that of the larva, is developed within it,46 and Graber has described a still more striking case of the same thing in a White Butterfly.47 There is, therefore, no reason to suppose that Insects possess more than one pair of antennÆ, which is probably preoral, not corresponding with either of the Crustacean pairs.

We have already noticed (p.26) the superficial points in which the antenna of the male Cockroach differs from that of the female.

The eyes of some Crustacea are carried upon jointed appendages, but this is never the case in Insects, though the eye-bearing surface may project from the head, as in Diopsis and Stylops. Professor Huxley48 supposes that the head of an Insect may contain six somites, the eyes representing one pair of appendages. The various positions in which the eyes of Arthropoda may be developed weakens the argument drawn from the stalk-eyed Crustacea. Claus and Fritz MÜller go so far on the other side as to deny the existence of an eye-segment even in Crustacea.

Mouth-parts of the Cockroach.

Before entering upon a full description of the mouth-parts of the Cockroach, which present some technical difficulties, the beginner in Insect anatomy will find it useful to get a few points of nomenclature fixed in his memory. Unfortunately, the terms employed by entomologists are at times neither convenient nor philosophical.

There are three pairs of jaws, disposed behind the labrum, as in the diagram:—

Labrum.

1st pair of Jaws

(Mandibles).

2nd "

(MaxillÆ).

3rd "

(Labium, or 2nd pair of MaxillÆ).

Fig. 19.—Diagram of Cock­roach Jaws, in hori­zon­tal sec­tion.

The mandible is undivided in all, or nearly all, Insects. Each maxilla may consist of

A palp on the outer side,
A galea (hood),
A lacinia (blade), on the inner side.

The galea (hood) of the 3rd pair of jaws is sometimes called the paraglossa.

A tongue-like process may be developed from the front wall of the mouth (epipharynx), or from the back wall (hypopharynx or lingua).49 Both epipharynx and hypopharynx project into the mouth, and, in some Diptera, far beyond it.

The tip of the labium is sometimes produced into a long tongue, called the ligula (strap).

The mouths of Insects may be classed as:—

Biting.—Orthoptera, Neuroptera, Coleoptera (in some Coleoptera a licking tongue is developed), most Hymenoptera.

Licking and Sucking.—Some Hymenoptera—e.g., Honey Bee.

Sucking.—(a) With lancets—Diptera, Hemiptera. (b) Without lancets—Lepidoptera.

The reference of these to a common plan, and the determination of the constituent parts, is mainly the work of Savigny. Mouth-parts were made the basis of the classification of Insects by Fabricius (1745–1808).

The mandibles of the Cockroach are powerful, single-jointed50 jaws, each of which is articulated by a convex “condyle” to the lower end of the epicranial plate, and again by a concave “ginglymus” to the clypeus. The opposable inner edges are armed with strong tooth-like processes of dense chitin, which interlock when the mandibles close; those towards the tip of the mandible are sharp, while others are blunt, as if for crushing. Each mandible can be moved through an angle of about 30°. A flexible chitinous flap extends from its inner border to the labrum. The powerful flexor of the mandible arises within the epicranial vault; its fibres converge to a chitinous tendon, which passes outside the central plate of the tentorium, and at a lower level through a fold on the lower border of the clypeus, being finally inserted near the ginglymus. A short flexor arises from the crus of the tentorium. The extensor muscle arises from the side of the head, passes through the fold formed by the lower end of the gena, and is inserted close to the outer side of the condyle of the mandible.

Fig. 20.—The Jaws, separated. Mn, mandible, seen from behind (to left) and front (to right); Mx' maxil­la (first pair); Mx" labi­um, or sec­ond pair of maxil­lÆ. The other let­ters as be­fore. ×20.

The anterior maxillÆ lie behind the mandibles, and like them are unconnected with each other. They retain much more of the primitive structure of a gnathite than the mandibles, in which parts quite distinct in the maxillÆ are condensed or suppressed. The constituent pieces are seen in fig.20. There is a two-jointed basal piece, consisting of the cardo (ca) and the stipes (st). The cardo is a transverse plate bent upon itself, and enclosing muscles; it is attached to the outward-directed pedicel of the occipital frame, and carries the vertical stipes. To the side and lower end of the stipes is attached the five-jointed palp (pa), a five-jointed limb used in feeding and in exploration, while the lacinia (la) and galea (ga) are articulated to its extremity. The lacinia is internal and posterior to the galea; it is broad above, but narrows below to a bifid tooth of dense chitin; its inner surface is beset with a cluster of strong setÆ. The galea is more flexible, and forms an irregular three-cornered prism with an obliquely truncated end, upon which are many fine hairs. A flexible and nearly transparent flap connects the inner edges of the stipes and cardo, and joins both to the labium. The muscles which move the bases of the maxillÆ spring from the crura, central plate, and keel of the tentorium.

On the posterior surface of the head, below the occipital foramen, we find a long vertical flap, the labium, which extends downwards to the opening of the mouth. It represents a second pair of maxillÆ, fused together in their basal half, but retaining elsewhere sufficient resemblance to the less modified anterior pair to permit of the identification of their component parts. The upper edge is applied to the occipital frame, but is neither continuous with that structure nor articulated thereto. By stripping off the labium upwards it may be seen that it is really continuous with the chitinous integument of the neck. The broad shield-like base is incompletely divided by a transverse hinge into an upper and larger piece, the submentum, and a distal piece, the mentum. To the mentum are appended representatives of the galeÆ (here named paraglossÆ) and laciniÆ, while a three-jointed palp with an additional basal joint (distinguished as the palpiger) completes the resemblance to the maxillÆ of the first pair.51 In front of the labium, and lying in the cavity of the mouth is a chitinous fold of the oral integument, the lingua, which lies like a tongue in the floor of the mouth. The common duct of the salivary glands enters the lingua, and opens on its hinder surface. The lingua is supported by the chitinous skeleton represented in the figures of the salivary glands. (Chap. vii., infra.)

The epipharynx, which is a prominent part in Coleoptera and Diptera, is not recognisable in Orthoptera.

Functions of the AntennÆ and Mouth-parts.

We must now shortly consider the functions of the parts just described. The antennÆ have long been regarded as sense-organs, and even the casual observer can hardly fail to remark that they are habitually used by the Insect to gain information concerning its immediate surroundings. Long antennÆ, such as those of the Cockroach, are certainly organs of touch, but it has been much disputed whether they may not also be the seat of some special sense, and if so, what that sense may be. Several authors have found reason to suppose that in the Insect-antenna resides the sense of hearing, but no evidence worth the name is forthcoming in favour of this view. Much better support can be found for the belief that the antenna is an olfactory organ,52 and some experiments which seem conclusive on this point will be cited in a later chapter.

In the Cockroach the mandibles and maxillÆ are the only important instruments of mastication. The labium is indirectly concerned as completing the mouth behind and supporting the lingua, which is possibly of importance in the ordinary operations of feeding. Plateau53 has carefully described the mode of mastication as observed in a Carabus, and his account seems to hold good of biting Insects in general. The mandibles and maxillÆ act, as he tells us, alternately, one set closing as the others part. The maxillÆ actually push the morsel into the buccal cavity. When the mandibles separate, the head is slightly advanced, so that the whole action has some superficial resemblance to that of a grazing quadruped.

The palps of the maxillÆ and labium have been variously regarded as sensory and masticatory instruments. Not a few authors believe that they are useful in both ways. The question has lately been investigated experimentally by Plateau,54 who finds that removal of both maxillary and labial palps does not interfere either with mastication or the choice of food. He observes that in the various Coleoptera and Orthoptera submitted to experiment the palps are passive while food is being passed into the mouth.

Plateau’s experiments are conclusive as to the subordinate value of the palps in feeding. The observation of live Cockroaches has satisfied us that the palps are constantly used when the Insect is active, whether feeding or not, to explore the surface upon which it moves. We have seen no ground for attributing to the palps special powers of perceiving odours or flavours, nor have we observed that they aid directly in filling the mouth with food.

It is worthy of note that Leydig has described and figured in the larva of Hydroporus (?), and Hauser in Dytiscus, Carabus, &c., a peculiar organ, apparently sensory, which is lodged in the maxillary and labial palps. It consists of whitish spots, sometimes visible to the naked eye, characterised by unusual thinness of the chitinous cuticle and by the aggregation beneath it of a crowd of extremely minute sensory rods. Of this organ no satisfactory explanation has yet been given.55

Comparison of Mouth-parts in different Insects.

The jaws of the Cockroach form an excellent standard of comparison for those of other Insects, and we shall attempt to illustrate the chief variations by referring them to this type.56 Mouth-parts are so extensively used in the classification of Insects that every entomologist ought to have a rational as well as a technical knowledge of their comparative structure. No part of Insect anatomy affords more striking examples of adaptive modification. In form, size, and mode of application the jaws vary extremely. It would be hard to find feeding-organs more unlike, at first sight, than the stylets of a Gnat and the proboscis of a Moth, yet the study of a few well-selected types will satisfy the observer that both are capable of derivation from a common plan. Nor is this common plan at all vague. It is accurately pictured in the jaws of the Cockroach and other Orthoptera. These correspond so entirely with the primitive arrangement, inferred by a process of abstraction from the most dissimilar Insects, as to furnish a strong argument for the descent of all higher Insects from forms not unlike Orthoptera in the structure of their mouth-parts.

Fig. 21.—Embryo of Aphis. Copied from Mecz­ni­kow, Zeits. f. wiss. Zool., Bd. XVI., taf. xxx., fig.30. Ref­er­ences in text. ×220.

Though the jaws of the Cockroach are eminently primitive with respect to those of most other Insects, they are themselves derived from a far simpler arrangement, which is demonstrable in all embryonic Insects. Fig.21 shows an Aphis within the egg. The rudiments of the antennÆ (At), mandibles (Mn), and maxillÆ (Mx¹, Mx²) form simple blunt projections, similar to each other and to the future thoracic legs (L¹, L², L³). We see, therefore, that all the appendages of an Insect are similar in an early stage of growth; and we may add that a Centipede, a Scorpion, or a Spider would present very nearly the same appearance in the same stage. A Crustacean in the egg would not resemble an Insect or its own parent so closely.57 Aquatic life favours metamorphosis, and most Crustacea do not begin life with their full quota of legs, but acquire them as they are wanted.

Paired appendages of perfectly simple form are therefore the first stage through which all Insect-jaws must pass. Our second stage is a little more complex, and not nearly so universal as the first. A caterpillar (fig.22) has its own special wants, and these are met by the unequal development of its jaws. The mandibles are already as complete as those of the Cockroach, which they closely resemble, but the maxillÆ are stunted cylinders formed mainly of simple rings, and very like the antennÆ. They show, however, the beginnings of three processes (palp, galea, and lacinia), which are usually conspicuous in well-developed maxillÆ. The second pair of maxillÆ (Lm) are coalesced, as usual, and form the spinneret. The mouth-parts of the Caterpillar do not therefore in all respects represent a universal stage of development, but show important adaptive modifications. The mandibles are rapidly pushed forward, and attain their full development in the larva; the first pair of maxillÆ are temporarily arrested in their growth, and persist for a long time in a condition which Orthopterous embryos quickly pass through; the maxillÆ of the second pair are not only arrested in their growth, but converted to a special use, which seems to stop all further progress. The labial palps, indeed, which are not at all developed in the caterpillar, survive, and become important parts in the moth; but the greater part of the labium disappears when the time for spinning the cocoon is over.

We come next to the Orthopterous mouth, which is well illustrated by the Cockroach. This is retained with little modification in all the biting Insects (Coleoptera and Neuroptera). The mandibles may become long and pointed, as in Staphylinus and other predatory forms; in some larvÆ of strong carnivorous propensities (Ant-lion, Dytiscus,58 Chrysopa) they are perforate at the tip, and through them the juices of the prey are sucked into the mouth, which has no other opening. The labium undergoes marked adaptive change, without great deviation from the common plan, in the “mask” of the larva of the Dragon-fly. This well-known implement has a rough likeness, in the arrangement and use of its parts, to a man’s fore-limb. The submentum forms the arm, the mentum the fore-arm. Both these are simple, straight pieces, connected by an elbow-joint. The hand is wider, and carries a pair of opposable claws, the paraglossÆ. In some Coleoptera the labium is reduced to a stiff spine, while in the Stag-beetle it is flexible and hairy, and foreshadows the licking tongue of the Bee. The maxillÆ become long and hairy in flower-haunting Beetles, and even the mandibles are flexible and hairy in the ScarabÆus-beetles. Fritz MÜller has found a singular resemblance to the proboscis of a Moth in a species of Nemognatha, where the maxillÆ are transformed into two sharp grooved bristles 12mm. long, which, when opposed, form a tube, but are incapable of rolling up.59

In the Honey Bee (fig.23) nearly all the mouth-parts of the Cockroach are to be made out, though some are small and others extremely produced in length. The mandibles (Mn) are not much altered, and are still used for biting, as well as for kneading wax and other domestic work. The mandibular teeth have proved inconvenient, and are gone. The lacinia of the maxilla (Mx') forms a broad and flexible blade, used for piercing succulent tissues, but the galea has disappeared, and there is only a vestige of the maxillary palp (Mxp). In the second pair of maxillÆ the palp (Lp) is prominent; its base forms a blade, while the tip is still useful as an organ of touch. The paraglossÆ (Pa) can be made out, but the laciniÆ are fused to form the long, hairy tongue. This ends in a spoon-shaped lobe (not unlike the “finger” of an elephant’s trunk), which is used both for licking and for sucking honey.

The proboscis of the Bee is therefore more like a case of instruments than a single organ. The mandibles form a strong pair of blunt scissors. The maxillÆ are used for piercing, for stiffening and protecting the base of the tongue, and when closed they form an imperfect tube outside the tongue, which, according to Hermann MÜller, is probably suctorial. The labial palps are protective and sensory. Lastly, the central part, or tongue, is a split tube used for suction; it is very long, so as to penetrate deep flower-cups, and hairy, so that pollen may stick to it. When the proboscis is not in use it can be slid into the mentum (M), while it and the mentum together can be drawn out of the way downwards and backwards.60

In the singular suctorial mouth of Moths and Butterflies we observe, first of all, the great development of the maxillÆ. Each forms a half-tube, which can be accurately applied to its fellow, so as to form an efficient siphon. In many species the two halves can be held together by a multitude of minute hooks.61 At the base of each maxilla is a rudimentary palp (Mxp). The mandibles (Mn) are also rudimentary and perfectly useless. The labium, which was so important to the larva as a spinneret, has disappeared almost completely, but the labial palps (Lp) are large and evidently important.

In Diptera both piercing and sucking parts are usually present. The Gad-fly (fig.25) is typical. Here we recognise the labrum (Lbr), mandible (Mn), and maxilla (Mx') of the Cockroach transformed into stylets. The maxillary palp (Mxp) is still sensory. A pointed process, stiffened by chitinous ribs, is developed from the back of the labrum. This is the epipharynx (Ep), a process undeveloped in the Cockroach, though conspicuous in some Coleoptera. All these parts are overtopped by the suctorial labium (Lm), which has a two-lobed expansion at the end. In the more specialised Diptera this becomes a kind of cupping-glass. The Gad-fly is intermediate between the Gnat, in which all the mouth-parts are converted into piercing organs of extraordinary length and sharpness, and such flies as the House-fly and Blow-fly, where the sucking labium forms an organ of the most elaborate kind, the piercing organs undergoing a marked reduction. Except where the labium is short, it is doubly or trebly hinged, so that it can be readily tucked away under the chin.

In Hemiptera the long four-jointed labium (Lm) forms a sheath for the stylets. When not in use the whole apparatus is drawn up beneath the head and prothorax. The mandibles (Mn) are sharp at the tip, and close like a pair of forceps, enclosing the maxillÆ (Mx). These are of unequal length, only one reaching the end of the mandibular case. Both have saw teeth on the free edge. Palps are entirely wanting.

Comparing the four kinds of suctorial mouths, of which the Bee, the Moth, the Fly, and the Bug furnish examples, we observe that the sucking-tube is formed in the Moth out of the two maxillÆ, in the other three out of the labium. Of these last the Bee has the edges of the labium turned down, so that the siphon becomes ventral; in the Bug and Fly the edges are turned up, and the siphon becomes dorsal. The more specialised flies have the simple arrangement of the Bug complicated by a system of branching tubes, which are probably a special modification of the salivary duct. Similar as the mouth-parts of the four types may be in regard to their mode of working, they cannot be reduced to any common plan which differs materially from that presented by the jaws of the Cockroach.

Composition of Head.

In all Insects fusion of the primitive elements of the head begins so early and is carried so far, that it is extremely difficult to discover the precise way in which they are fitted together. The following facts have been ascertained respecting the development of the parts in question. At a very early stage of embryonic life the body of the Insect becomes divided into a series of segments, which are at fewest fourteen (in some Diptera), while they are not known to exceed seventeen.62 Each segment is normally provided with a pair of appendages. The foremost segment soon enlarges beyond the rest, and becomes divided by a median groove into two “procephalic lobes.”63 Of the appendages the first eight pairs are usually more prominent than the rest, and of different form; those of the eighth segment, which may be altogether inconspicuous, never attain any functional importance. The first four pairs of appendages are budded off from the future head, while the next three pairs form the walking legs, and are carried upon the thoracic segments. All the existing appendages of the fore part of the body are thus accounted for, but the exact mode of formation of the head has not yet been made out. The chief part of its walls, including the clypeus, the compound eyes, and the epicranial plates, arise from the procephalic lobes, and represent the much altered segment of which the antennÆ are the appendages. The labrum is a secondary outgrowth from this segment, and, in some cases at least, it originates as a pair of processes which resemble true appendages, though it is unlikely that such is their real character. No means at present exist for identifying the terga and sterna of the head, nor have the gena, the occipital frame, and the cervical sclerites (described below) been assigned to their segments.64 It is worthy of notice that in the stalk-eyed Crustacea, the head, or what corresponds to the head of Insecta, consists of either five or six somites, taking into account a diversity of opinion with respect to the eyestalks, while only four pairs of appendages can be certainly traced in the head of the Insect. The mandibles and maxillÆ exist to the same number in both groups, and are homologous organs, so far as is known; the numerical difference relates therefore to the antennÆ, of which the Crustacean possesses two pairs, the Insect only one. Whether the pair deficient in the Insect is altogether undeveloped, or represented by the pair of prominences which give rise to the labrum,65 is a question of much theoretical interest and of not a little difficulty.

The following table shows the appendages of the head and thorax in the two classes. The homologies indicated are, however, by no means established.66

Neck.

The neck is a narrow cylindrical tube, with a flexible wall strengthened by eight plates, the cervical sclerites, two of which are dorsal, two ventral, and four lateral. The dorsal sclerites lie immediately behind the head (fig.14); they are triangular, and closely approximated to the middle line. The inferior plates (fig.27) resemble segments of chitinous hoops set transversely, one behind the other, rather behind the dorsal sclerites, and close behind the submentum. There are two lateral sclerites on each side of the neck (fig.27), a lower squarish one, which is set diagonally, nearly meeting its fellow across the ventral surface, and an oblong piece, closely adherent to the other, which extends forwards and upwards towards the dorsal side.

Thorax.

The elements of the thoracic exoskeleton are simpler in the Cockroach than in Insects of powerful flight, where adaptive changes greatly obscure the primitive arrangement. There are three segments, each defended by a dorsal plate (tergum) and a ventral plate (sternum). The sterna are often divided into lateral halves. Of the three terga the first (pronotum) is the largest; it has a wide free edge on each side, projects forwards over the neck, and when the head is retracted, covers this also, its semi-circular fore-edge then forming the apparent head-end of the animal. The two succeeding terga are of nearly equal size, and each is much shorter than the pronotum, contrary to the rule in winged Insects.67

All the terga are dense and opaque in the female; in the male the middle one (mesonotum) and the hindmost (metanotum) are thin and semi-transparent, being ordinarily overlaid by the wing-covers. While the thoracic terga diminish backwards, the sterna increase in extent and firmness, proportionally to the size of the attached legs. The prosternum is small and coffin-shaped; the mesosternum partly divided into lateral halves in the male, and completely so in the female. The metasternum is completely divided in both sexes, while a median piece, carrying the post-furca, intervenes between its lateral halves in the male. Behind the sterna, especially in the case of the second and third, the flexible under-surface of the thorax is inclined, so as to form a nearly vertical step. In the two hinder of these steps a chitinous prop is fixed; each is Y-shaped, with long, curved arms for muscular attachment, and a central notch, which supports the nerve-cord. The hindmost of these, known as the post-furca, lies immediately behind the metasternum, and its short basal piece is attached between the lateral halves of that plate. Behind the mesosternum is a somewhat slighter prop, the medi-furca. A third piece of similar nature (the ante-furca), which is well developed in some Insects—e.g., in Ants—is apparently wanting in the Cockroach, though there is a transverse oval plate behind the prosternum, which may be a rudimentary furca.

Fig.27 shows two conical processes which lie in the middle line of the ventral surface of the thorax, one in front of the metasternum, the other in front of the mesosternum. These are the thoracic pits, tubular apodemata, serving for the insertion of muscles. The occurrence of stink-glands in the thorax of Hemiptera,68 and of so-called poison-glands in the thorax of Solpuga, led us to look for glands in connection with these processes, but we have found none.

Thoracic Appendages. Legs; Wings.

Three pairs of legs are attached to the thoracic segments; they regularly increase in size from the first to the third, but hardly differ except in size; the peculiar modifications which affect the fore pair in predatory and burrowing Orthoptera (Mantis, Gryllotalpa), and the third pair in leaping Orthoptera (Grasshoppers, &c.), being absent in the cursorial Blattina. Each leg is divided into the five segments usual in Insects (see fig.28). The coxa is broad and flattened. The trochanter is a small piece obliquely and almost immovably attached to the proximal end of the femur, on its inner side. The femur is nearly straight and narrowed at both ends; along its inner border, in the position occupied by the stridulating apparatus of the hind leg of the Grasshoppers, is a shallow longitudinal groove, fringed by stiff bristles. The tibia is shorter than the femur in the fore leg, of nearly the same length in the middle leg, and longer in the hind leg; it is armed with numerous stiff spines directed towards the free end of the limb. There are usually reckoned five joints in the tarsus, which regularly diminish in length, except that the last joint is as long as the second. All the joints bear numerous fine but stiff hairs upon the walking surface. The extremity of the fifth joint is segmented off, and carries a pair of equal and strongly curved claws.69

At the base of each leg are several chitinous plates (fig.28), upon which no small labour has been bestowed by different anatomists. They are arranged so as to form two joints intermediate between the coxa and the sternum, and these two joints admit of a hinge-like movement upon each other, while their other ends are firmly attached to the coxa and sternum respectively. (Compare III and IIIA, fig.28.) These parts in the Cockroach may be taken for two basal leg-joints which have become adherent to the thorax. In other cases, however, they plainly belong to the thorax, and not to the leg. In the Mole-cricket, for instance, similar plates occur; but here they are firmly united, and form the lateral wall of the thorax. In the Locust they become vertical, and lie one in front of the other. Most authors have looked upon them as regular elements of a typical somite. They regard such a segment as including two pleural elements—viz., a dorsal plate (epimeron), and a ventral plate (episternum). We have already (p.34) given reasons for doubting the constancy of the pieces so named. It is not inconvenient, however, to denote by the term episternum the joint which abuts upon the sternum; for the joint which is applied to the coxa no convenient term exists, and its occurrence in Insects is so partial, that the want need not be supplied at present.70 Both joints are incompletely subdivided. In the first thoracic segment of the Cockroach they are less firmly connected than in the other two.

Cockroaches of both sexes are provided with wings, which, however, are only functional in the male. The wing-covers (or anterior pair of wings) of the male are carried by the second thoracic segment. As in most Orthoptera genuina, they are denser than the hind wings, and protect them when at rest. They reach to the fifth segment of the abdomen, and one wing-cover overlaps the other. Branching veins or nervures form a characteristic pattern upon the surface (figs.4, 29), and it is mainly by means of this pattern that many of the fossil species are identified and distinguished. The true or posterior wings are attached to the metathorax. They are membranous and flexible, but the fore-edge is stiffened, like that of the wing-covers, by additional chitinous deposit. When extended, each wing forms an irregular quadrant of a circle; when at rest, the radiating furrows of the hinder part close up fan-wise, and the inner half is folded beneath the outer.71 The wing reaches back as far as the hinder end of the fourth abdominal segment. The wing-covers of the female are small, and though movable, seem never to be voluntarily extended; each covers about one-third of the width of the mesonotum, and extends backwards to the middle of the metanotum. A reticulated pattern on the outer fourth of the metanotum plainly represents the hind wing; it is clearly rather a degeneration or survival than an anticipation of an organ tending towards useful completeness.

The rudimentary wing of the female Cockroach illustrates the homology of the wings of Insects with the free edges of thoracic terga, and this correspondence is enforced by the study of the development of the more complete wings and wing-covers of the male. The hinder edges of the terga become produced at the later moults preceding the completely winged stage, and may even assume something of the shape and pattern of true wings; it is not, however, true, though more than once stated, that winged nymphs are common. Adults with imperfectly developed wings have been mistaken for such.

Origin of Insect Wings.

The structure of the wing testifies to its origin as a fold of the chitinous integument. It is a double lamina, which often encloses a visible space at its base. The nervures, with their vessels and tracheal tubes, lie between the two layers, which, except at the base, are in close contact. Oken termed the wings of an Insect “aerial gills,” and this rather fanciful designation is in some degree justified by their resemblance to the tracheal gills of such aquatic larvÆ as those of EphemeridÆ, PerlidÆ, PhryganidÆ, &c. In the larva of Chloeon (Ephemera) dipterum (fig.30), for example, the second thoracic segment carries a pair of large expansions, which ultimately are replaced by organs of aerial flight. The abdominal segments carry similarly placed respiratory leaflets, the tracheal gills, which by their vigorous flapping movements bring a rush of water against their membranous and tracheated surfaces.

Gegenbaur72 has argued from the resemblance of these appendages to wings, that the wing and the tracheal leaflet are homologous parts, and this view has been accepted as probable by so competent an observer as Sir John Lubbock.73

The leaflets placed most advantageously for propulsion seem to have become exclusively adapted to that end, while the abdominal gills have retained their respiratory character. At the time of change from aquatic to terrestrial life, which takes place in many common Insects when the adult condition is assumed, and which, according to Gegenbaur, was a normal event among primitive Insects, the tracheal gill is supposed to disappear, and in its place, at the next moult, an opening, the stigma, is formed by the rupture of an air-tube. Gegenbaur supposes that the primitive Insects were aquatic, and their tracheal system closed. The tracheal gill he takes to be the common structure which has yielded organs so unlike as the wing and the stigma.

The zoological rank of the Insects (EphemeridÆ, PerlidÆ, and LibellulidÆ), in which tracheal gills are most usual, is not unfavourable to such an explanation. Lubbock has given reasons for regarding Campodea and the Collembola (of the order Thysanura) as surviving and not very much altered representatives of the most primitive Insects, and he has shown that no great amount of modification would be required to convert the terrestrial Campodea into the aquatic Chloeon-nymph.74 We must not forget, however, that tracheal gills are by no means restricted to these families of low grade. Trichoptera, a few Diptera, two Lepidoptera (Nymphula and Acentropus), and two Coleoptera (Gyrinus and Elmis),75 have tracheal gills, and a closed tracheal system in the larval condition. We cannot suppose that these larvÆ of higher orders represent an unbroken succession of aquatic forms, but if we refuse to adopt this alternative, we must admit that the closed tracheal system with tracheal gills may be an adaptive modification of the open system with stigmata.

It is well known76 that in certain EphemeridÆ (e.g., Tricorythus and Coenis) a pair of anterior tracheal gills may become transformed into large plates, which partly protect the gills behind (fig.31). A similar modification of the second and third thoracic gills in Prosopistoma and BÆtisca brings all the functional respiratory organs under cover, and these enlarged plates resemble stiff and simple wings very closely.

PalmÉn77 has subjected Gegenbaur’s hypothesis to a very searching examination. He observes that:—

1. In Campodea, and presumably in other primitive Insects, the tracheal system is not closed and adapted for aquatic respiration, but open. Tracheal gills are not by any means confined to the lowest Insects. (See above, p.65.)

2. Tracheal gills are not always homodynamous or morphologically equivalent. In EphemeridÆ, some are dorsal in position, some ventral (first abdominal pair in Oligoneuria and Rhithrogena); they may be cephalic, springing from the base of the maxilla, as in Oligoneuria and Jolia; Jolia has a branchial tuft at the insertion of each of the fore legs.78 In PerlidÆ the tracheal gills may have a tergal, pleural, sternal, or anal insertion. In some LibellulidÆ also, anal leaflets occur.79

3. Tracheal gills never perfectly agree in position and number with the stigmata throughout the body. Sometimes they occur on different rings, sometimes on different parts of the same ring. Gegenbaur’s statements on this point are incorrect.

4. Tracheal gills may co-exist with stigmata. In PerlidÆ the tracheal gills persist in the imago, and may be found, dry and functionless, beneath the stigmata. In Trichoptera they gradually abort at successive moults, and in some cases remain after the stigmata have opened.

5. Stigmata do not form by the breaking off of tracheal appendages, but by the enlargement of rudimentary tracheal branches, which open into the main longitudinal trunks. In larvÆ with aquatic respiration these branches exist, though they are not functional.

PalmÉn’s objections must be satisfactorily disposed of before Gegenbaur’s explanation, interesting as it is, can be fully accepted. PalmÉn has proved, what is on other grounds clear enough, that stigmata are more ancient than tracheal gills, aerial tracheate respiration than aquatic. But there is nothing as yet to contradict the view that the first Insect-wings were adapted for propulsion in water, and that they were respiratory organs before they became motor. It is Gegenbaur’s explanation of the origin of stigmata, and not his explanation of the origin of wings, which is refuted by PalmÉn.

Abdomen.

In the abdomen of the female Cockroach eight terga (1–7; 10) are externally visible. Two more (8, 9) are readily displayed by extending the abdomen; they are ordinarily concealed beneath the seventh tergum. The tenth tergum is notched in the middle of its posterior margin. A pair of triangular “podical plates,” which lie on either side of the anus, and towards the dorsal surface, have been provisionally regarded by Prof. Huxley as the terga of an eleventh segment. Seven abdominal sterna (1–7) are externally visible. The first is quite rudimentary, and consists of a transversely oval plate; the second is irregular and imperfectly chitinised in front; the seventh is large, and its hinder part, which is boat-shaped, is divided into lateral halves, for facilitating the discharge of the large egg-capsule.

In the male Cockroach ten abdominal terga are visible without dissection (fig.33, p.70), though the eighth and ninth are greatly overlapped by the seventh. The tenth tergum is hardly notched. Nine abdominal sterna are readily made out, the first being rudimentary, as in the female. The eighth is narrower than the seventh, the ninth still narrower, and largely concealed by the eighth; its covered anterior part is thin and transparent, the exposed part denser. This forms the extreme end of the body, except that the small sub-anal styles project beyond it. The podical plates resemble those of the female.

Pleural elements are developed in the form of narrow stigmatic plates, with the free edge directed backwards. These lie between the terga and sterna, and defend the spiracle.80

The modifications of the hindmost abdominal segments will be more fully considered in connection with the reproductive organs.

The high number of abdominal segments found in the Cockroach (ten or eleven) is characteristic of the lower orders of Insecta. It is never exceeded; though in the more specialised orders, such as Lepidoptera and Diptera, it may be reduced to nine, eight, or even seven. The sessile abdomen of the Cockroach is primitive with respect to the pedunculate abdomen found in such insects as Hymenoptera, where the constricted and flexible waist stands in obvious relation to the operations of stinging and boring, or to peculiar modes of oviposition. The first abdominal segment, which is especially liable to dislocation and alteration in Insects, occupies its theoretical position in the Cockroach, though both tergum and sternum are reduced in size. The sternum is often altogether wanting, while the tergum may unite with the metathorax.

The externally visible appendages of the abdomen are the cerci and the styles of the male Cockroach. The cerci are found in both sexes; they are composed of sixteen rings each, and project beneath the edge of the tenth tergum. They are capable of erection by special muscles, and are supplied by large nerves.81 The sub-anal styles are peculiar in their insertion, being carried upon the sternum of their segment (the ninth).

The abdominal segments are never furnished with functional legs in adult Insects, but representatives of the lost appendages are often met with in larvÆ. According to BÜtschli,82 all the abdominal segments are provided with appendages in the embryo of the Bee, though they disappear completely before hatching. Some Hymenopterous larvÆ have as many as eight pairs of abdominal appendages, Lepidopterous larvÆ at most five (3–6; 10).83