It will much simplify our conception of the nature of the air-tubes when we learn that they originate in the embryo as tubular ingrowths of the integument (ectoderm), these branching and finally reaching every part of the interior of the body. They are elastic tubes, and being filled with air are silvery in color, though at their origin near the spiracles they are reddish or violet bluish; or, in the larva of Æschna, reddish brown, this tint being due to a finely granular pigment situated in the peritoneal membrane.

In their essential structure the tracheæ consist of the chitinous intima, which is a continuation of the cuticle of the integument, and of a cellular membrane or outer layer of cells (a continuation of the hypodermis) called the peritoneal membrane, or ectotrachea (Figs. 392, 393).

Leydig discovered that the spiral filaments are not distinct and separate, but intimately connected with the inner membrane (intima), and he detected the outer or peritoneal membrane, which Chun afterwards found to be epithelial in its nature, Minot stating that it is a true pavement epithelium.

Figure 393 represents a longitudinal section of a large trachea of Hydrophilus, showing the peritoneal membrane (ectotrachea, ep) and the intima or endotrachea, divided into the cuticula (cu), with the darker colored inner layer, in which are embedded the dark-colored tænidia (f).

Distribution of the tracheæ.—The distribution of the air-tubes, as Lubbock and also Minot state, depends first upon the shape of the organs, and upon the size of those whose size is variable. Around the large, hollow organs (digestive canal, sexual organs) the tracheæ ramify in all directions, forking so that the branches diverge at a wide angle. In the organs which have muscular walls, like the oviduct, the tracheæ run straight when the walls are distended, but have a sinuous course when the walls are contracted. (Minot.)

“Around the organs of more elongated form the branches of the tracheæ run more longitudinally, as is shown by the air-tubes of the muscles, which also present some peculiarities worthy of especial notice.

“A short, thick trunk arrives at the muscular bundle, and dividing very rapidly, breaks up into a large number of delicate tubes, which penetrate between the muscular fibres, then terminating in tubes of exceeding fineness, which at first sight seem to form a network that might well be called a rete mirabile. A closer examination, however, reveals that it is not a real network, but rather an interlacing confusing to the eye. The longitudinal direction of the tracheæ of the muscles presents a striking contrast to the system of divarication represented in Figs. 13 and 14. The course of the tracheæ of the Malpighian tubes is also very curious. There is one large trachea which winds around the tube in a long spiral, giving off numerous small branches which run to the surface of the tube, upon which they form delicate ramifications. Each tube has but a single main trachea, and I think the trachea continues the whole length of the tube, but of this last point I am not quite sure.” (Minot.)

While in the nymphs of Orthoptera the tracheæ very closely resemble those of the adult, in larvæ of insects with a complete metamorphosis the tracheæ differ very much in distribution from those of the adult. The larval tracheæ are also more generalized and more like those of the original type than the tracheæ of perfect insects. (Lubbock.)

In general there are two main tracheæ, one passing along each side of the body, near the digestive canal, connected with its mate by a few transverse anastomosing branches, and sending off a branch to each spiracle, this arrangement being most simple and apparent in the maggots of Diptera. From these two main branches smaller twigs branch off into every part of the body with its appendages, passing among the different organs, often serving as cables to hold them loosely in place; they also penetrate into the component parts of the organ themselves, passing into the fat-bodies, and among the fibres of muscles, where they become finely attenuated and refined like the capillaries of the vascular system of vertebrates. (Figs. 395, 396.)

In the youngest larva of Corethra plumicornis Weismann ascertained the thickness of the longitudinal stem to be 0.0017 mm. That of the finest tracheal endings in the silk-glands of the silkworm was found by Von Wistinghausen to be 0.0016 mm. (Zeits. f. Wiss. Zool. xlix, 1890, p. 575.) Weismann states that in the larvæ of Corethra and Chironomus the tracheal system is only incompletely developed; the tracheæ are not united with each other, and in the youngest larvæ they do not contain air.

Each of the two main tracheæ, as Kolbe states, sends off into each segment of the body three branches.

1. An upper or dorsal branch, which supplies the muscles of the dorsal region.

2. A middle (visceral) branch, whose twigs pass to the digestive canal and back to the organs of reproduction.

3. A lower (ventral) branch, whose twigs are distributed to the ganglia and to the muscles of the ventral region.

In certain Thysanura, as a species of Machilis (Fig. 397), we probably have the primitive condition of the tracheal system, the longitudinal and transverse anastomoses being absent, but in other Thysanura (Japyx, Nicoletia, Lepisma, and a few species of Machilis) they are present.

As Kolbe remarks, whether the fine ends of the tracheæ are closed or open, whether after the analogy of the blood capillaries of vertebrates they anastomose with each other, whether the ends of the air-tubes pass between the cells or penetrate into them, these questions are not fully settled. According to Leydig’s^[61] latest views the tracheæ penetrate into the cells and unite with the hyaloplasma. Hence the process of respiration in the last instance takes place in the hyaloplasma. This assumption accords with the fact that in the tracheate Arthropods the terminations of the tracheæ carry the atmospheric air into the space bounded by the cellular network, also to the hyaloplasma filling the spaces. Leydig^[62] also thinks that the finest tracheal endings penetrate into the muscular tissue and unite with the primitive muscular fibres.

Kupffer is likewise of the opinion that the fine tracheæ penetrate into the cells, and Lidth de Jeude asserts that they enter the epithelial cells, “each cell containing several branches.” Kölliker, Emery, etc., maintain, however, that the tracheal endings lie between the cells. Wielowiejski,^[63] in describing the line tracheæ of the phosphorescent organs, thinks that the tracheal endings (tracheal capillaries) rarely end blindly, but anastomose with one another, forming an irregular network. The latest observer, Gilson (1893), asserts that tracheal twigs penetrate deeply into the epithelial cells of the silk glands of larval Trichoptera as well as of caterpillars, passing through their protoplasm.

A late investigator, C. von Wistinghausen, finds in the tracheæ of the spinning-glands of caterpillars a completely formed network between the terminal branches of two or several tracheal groups. The tracheal tubes of this series of terminal branches pass into this network, which he calls the tracheal capillary end-network (Figs. 398, 400). This last varies in thickness and spreads out under the membrana propria of the glandular mass over the entire surface of the large gland-cells and on a level with the tracheal capillaries. The tracheal endings do not penetrate into the cells, but are separated from the plasma of the cells by a thin membrane. The tracheal capillary end-network appears as a system of fine tubes like the tracheal capillaries, consisting of a peritoneal layer and a chitinous intima (Fig. 400). The walls of these tubes are homogeneous, not porous, though readily permeable by the parenchymatous fluid. The interchange of gases consequently may go on easier and more vigorously in a system of richly anastomosing tubules of the net-like mass of tracheal capillaries, than in tubes ending blindly.

While the diameter of the tracheal capillaries is 0.0016 mm. or 1 µ, that of the tubules composing the tracheal capillary end-network is scarcely measurable, but is less than 1 µ.

These tracheal capillaries also occur on the seminal and other sexual tubes, on the intestine, on the urinary tubes, on the fat-bodies, but are most easily detected on the silk-glands.

The latest researches are those of E. Holmgren, who has studied the branching of the tracheæ in the spinning-glands of caterpillars. He prefers to call the end-cells “transition cells,” as they lead from the tracheal tubes proper to the capillary network. This latter is formed by slender nucleated cells, often with an intracellular lumen, and, according to the author, probably constituting a respiratory epithelium. He finds that both large and small tracheæ may penetrate the gland-cells. (Anat. Anzeiger, xi, 1895, pp. 340–6, 3 figs.; Jour. Roy. Micr. Soc., 1896, p. 182.)

b. The spiracles or stigmata

The spiracles are segmentally arranged openings in the sides of the thorax and abdomen, through which the air passes into the air-tubes. In its essential structure a spiracle, or stigma, is a slit-like opening surrounded by a chitinous ring, the lips or edges of the opening being membranous and closed by a movable valve of the spiracle attached by its lower edge, which is closed by an occlusor muscle (Fig. 401). The aperture when open forms a narrow oval slit; and in most insects the slit is within guarded by a row of projecting spines or setæ, which form a lattice work or grate to keep out dust, dirt, fluids, etc.

Krancher^[64] has described five leading types of stigmata, not, however, taking into account those of the Synaptera.

I. Stigmata without lips (Primitive or generalized stigmata).

a. The simplest stigma is an aperture which is kept open by a chitinous ring (Acanthia). The opening may be round or elliptical. There are no lips nor any movement of the edges to be observed. Such air-holes occur in the abdomen of bugs (Hemiptera) and beetles (Coleoptera); within the opening of the stigmata in the same insects is a funnel-like contraction. Also in the Diptera the abdominal stigmata are of the same type.^[65] The stigmata of the Pulicidae (Siphonaptera) are more complicated, as the edges of the openings are provided with setæ (Fig. 402).

b. The stigma consists of a series of minute single stigmata, which are usually surmounted by a common chitinous ring, and whose tubular continuations unite within in a common trachea, so that the single tubes pass off from the stigma like the fingers on the hand. This form is found in the larvæ and puparia of Diptera.

II. Stigmata with lips (Secondary more specialized stigmata).

c. The lips are represented by a single chitinous ring, with sparse spines. One side of the stigma is a little higher, and partly overlaps the other posteriorly; this form is peculiar to the Orthoptera and Libellulidae.

d. The lips are roof-like, bent inwards and densely hairy, forming a peculiar kind of felting. The setæ of the lips are in most beetles and many Lepidoptera separate, and more or less branched. In caterpillars, the setæ are so finely branched as to form a loose felt, or sieve-like arrangement.

e. The stigmata are round, with a very broad border and a concentric middle portion, the structure being complicated. The concentric middle portion is pouch-like and bears the occlusor muscle. This form occurs in the larvæ of lamellicorn beetles, and can be seen with the naked eye, or with a lens, in Oryctes, Cetonia, and Melolontha (Fig. 403).

f. Over the outer opening of the spiracle is an incurved chitinous projection, on one side of which the trachea takes its origin. It is thus in the Hymenoptera.

The remarkable grate-like stigma of the lamellicorn larvæ has the appearance as if the outer closing plate or valve were impenetrable. The earlier observers considered these stigmata to be open, but Meinert regards them as closed; Schiödte, however, has observed by pressing a preserved specimen of a Melolontha larva the alcohol within passing out in drops, through the grate-like plate, and hence he considers this a proof that the stigma is permeable (Kolbe).

More recently (1893) Boas has examined the same structure in the same species of larva as examined by Schiödte, and he finds it to be open only during the process of moulting. He finds that on each side of the larva there are nine short and wide stigmatic branches, each of which is shut off from the exterior by a brown plate; this consists of a reniform sieve-plate, and of a curved bulla which fits into the cavity of the plate. The stigmatic branch, however, is provided with a large external opening, which is homologous with the stigma, but which is usually closed by the plate and bulla, and is only open during the moulting; at first it is circular, but later becomes a cleft. A transverse section shows that the bulla is a simple tegumentary fold, the outer chitinous layer of which has become especially firm. The plate forms a horizontal half-roof, which springs from one side of the tracheal orifice, and is supported by obliquely set bases, which spring from the adjoining part of the inner side of the tracheæ. The plate and bars are purely cuticular structures. (Zool. Anz., 1893; also Journ. Roy. Micr. Soc., p. 54.)

The tracheal system of libellulid nymphs is not closed; on the other hand, in the fully-grown nymphs the anterior stigmata occurring on the dorsal side are large, and the tracheæ arising from them are thick. These stigmata are permeable by the air. In half-grown and still younger stages of Æschna the two anterior thoracic stigmata are undeveloped. In order to breathe, the fully-grown nymph either rises up on the upper side and elevates the end of the body to the surface in order to take the air into the rectum, or it rests with the back of the thorax at the surface in order to breathe through the large stigmata. The young nymphs take in air only through the rectum. The young nymphs of Libellula and its allies, on the other hand, possess large thoracic stigmata, but they prefer to breathe through the rectum. The fully-grown nymphs of Agrion breathe through the thoracic stigmata. (Dewitz, in Kolbe.)

The position and number of pairs of stigmata.—The spiracles are usually situated in the soft membrane between the tergites and pleurites, but their exact position varies in different groups. In the Coleoptera they occupy on the thorax a more ventral position, and on the abdomen are placed near the edge of the dorsal side, under the elytra. In the dragon-flies, the first pair is situated much more dorsally than the second and third pairs; the following seven pairs are almost wholly ventral and lie concealed in the membranous fold near the external plate. In the Hemiptera, also, the abdominal stigmata, though entirely free and visible, are situated ventrally.

Primarily, in the embryo a pair of stigmata appear on each segment of the thorax and abdomen, except the 10th and 11th, and even possibly in the head, for a pair of stigmata are said to occur in the head of Podurids (Smynthurus) (Lubbock), though this statement needs confirmation. Scolopendrella, however, is known to possess a pair of cephalic spiracles.

From the foregoing statement it will be seen that while in existing winged insects no more than 10 (in Japyx 11) pairs of stigmata are to be found in any one species, yet that 12 segments of the body, in different groups taken collectively, bear them. The primitive number of pairs of spiracles, therefore, in winged insects, was 12, i.e. a pair in each thoracic segment, and a pair in each of the first nine abdominal segments. Insects were originally all holopneustic, and gradually as the type became differentiated into the different orders they became peripneustic or amphipneustic, and, in certain aquatic forms, apneustic. (See pp. 459, 461.)

In the still more primitive, probably wingless, ancestors of insects there was a larger number of stigmata. Hatschek, in 1877, discovered a pair of tracheal invaginations in each of the three posterior head-segments of the embryo of a moth, with stigmatal openings in the 1st and 2d maxillary segments.

Thus early in embryonic life every segment of the body, except those bearing the eyes and the last abdominal, bore a pair of stigmata, so that the primitive insect had at least 15, and perhaps more, pairs of stigmata.

The position of the stigmata is subject to much variation, the result of adaptation to this or that mode of life. Examples are those insects which live in dusty situations or usually more or less concealed in the earth, as in most beetles, and in the Hymenoptera. In such beetles, the stigmata are situated in the thin membrane between the segments; in the Hymenoptera, on the upper edge of the segments. In the Siphonaptera, Pediculina, bed-bug, and similar forms, which breathe an air freer from dust, the spiracles lie free on the outside of the body.

“When the stigmata are free and without any protection on the abdomen, there are other ways by which the entrance of foreign bodies into the tracheæ is prevented. In such cases the body is covered with dense hairs, as in most Diptera and Neuroptera, as well as many Lepidoptera; or there is situated in front of the stigma either a small fissure which is covered over by a number of hairs arising from the edge, as in many Orthoptera; or, as in most insects, a luxurious growth of hairs on the inside of the stigma forms a thick filter for the air. Thus we see that also in this respect each species of insect is completely adapted to its surroundings.” (Krancher.)

The closing apparatus of the stigma.—Whether the external opening of the stigma is permanently open or closed, communication with the tracheæ may be cut off at pleasure during respiration by an internal apparatus of elastic chitinous bands and rods and the occlusor muscle.

The parts concerned in this operation are: 1. The closing bow; 2. The closing lever or peg; 3. The closing band; 4. The occlusor muscle (Figs. 405, 406).

“The first three parts are chitinized; they form a ring around the stigmatic opening, and are united to each other by joints. The bow is usually crescentic and as a rule surrounds one-half of the trachea. On the other side is the closing band which, by different contrivances, representing the closing lever or peg, becomes closely pressed against the closing bow. This lever is usually of the shape of a slender chitinous rod, which causes the closure; but it can also bend rectangularly, become converted into a typical lever as in the Lepidoptera, or it may assume the form of two peg-like processes, which press with their base against the closing bow.” (Krancher.)

“The closure of the spiracular opening is effected by the contraction of the muscles, while the opening is due to the elasticity of the chitinous parts. When at rest the spiracle is naturally open, so that the air in the trachea can directly communicate with the external air. Usually one end of the muscle is attached to the closing peg, and the other end to the closing bow. Where, as in Melolontha, the closing apparatus is provided with two levers, then naturally the muscle binds these two together and brings about by powerful contractions a firm closure of the trachea”; but, remarks Krancher, “this is not the only kind; there are numerous modifications. Besides the form just described, the levers assume the form of valves (Sirex), or of a brush (Pulex); or of a ring (larvæ of Diptera) with a circular muscle attached to it; or of a ring which simply becomes compressed (thoracic stigmata of Diptera).”

c. Morphology and homologies of the tracheal system

As first shown by Bütschli, the tracheal system is a series of segmentally arranged tubular invaginations of the ectoderm; a pair of stigmata primitively occurring on every segment of the body except perhaps the most anterior, and the last two or last one, a reduction in their number having since taken place, until in the Podurans none have survived. In the supposed ancestor of myriopods and insects, Peripatus, there are tracheæ; but they are very fine, simple, not-branched chitinous tubes which are united into tufts at the base of a flask-shaped depression of the integument, the outer aperture of which depression is regarded as a stigma. In one species (P. edwardsii) these tufts and their openings are scattered irregularly over the body; but in another kind (P. capensis) some of the stigmata at least show traces of a serial arrangement, being disposed in longitudinal rows—two on each side, one dorsally and one ventrally, those of each row, however, being more numerous than the pairs of legs. (See p. 9 and Fig. 4, D.)

It should be observed that in Peripatus, which does not possess urinary tubes, the segmental organs or nephridia are well developed, hence the tracheal tubes coexisting with them cannot be their homologues. We are therefore compelled to regard the tracheal system as of independent origin, arising in the earliest terrestrial air-breathing arthropod, and not indebted for its origin to any structure found in worms, unless perhaps, as both Kennell and Lang suggest, to dermal glands, since, according to Kennell, certain Hirudinea and many Turbellarian worms possess long, mostly unicellular, glands which spread far through the parenchyma of the body. (Kennell.)

Thus Kennell supposes that the ancestors of the Tracheates had spiracles on every segment of the body where the internal organization allowed them to exist. “The reduction of the breathing holes to a smaller number, and their restriction of a pair only to a single segment, was brought about partly by adaptation to a peculiar mode of life,—as insect larvæ especially teach us,—partly also—I may say mechanically—as a result of the obstruction to their development made by the growth or excessive development of other organs.” Among these he reckons the thick, dense cuticula of the integument, the internal fusion of several segments to form body-regions, and the arrangement and great development of the muscles in the head and thorax, etc. (p. 29.)

Kennell has suggested the origin of the tracheæ of Peripatus from the unicellular dermal glands of annelidan ancestors, since he has found glands in certain land-leaches of tropical America, which are provided with enormously long tubular passages united into bundles and opening externally, these tubes appearing to be slightly chitinized. Fig. 407 will show the appearance of a bundle of fine tracheal tubes of Peripatus ending at the bottom of a follicle formed by a deep invagination of the integument, which may be regarded as a primitive spiracle. (See Kennell, Ueber einige Landblutegel des tropical America, Zool. Jahrb. ii, 1886; also Die Verwandtschaftsverhältnisse der Arthropoden, 1891, p. 25.) We may add that Carrière supposes from his study of the embryology of the wall-bee (Chalicodoma muraria), published in 1890, that not only the salivary glands, but also the tentorium, are homologues of the tracheæ, while other structures than tracheæ may have evolved from unicellular dermal glands, which are widely distributed. It may in this connection be observed that some authors derive the book-lungs or book-leaf tracheæ of Arachnida from the gills of Limulus; hence if those of Arachnida arose from quite different and more specialized organs than dermal glands, it is not impossible that the tracheæ of Peripatus, Myriopods, and insects arose de novo, and then we need not look for any primitive structures in worms from which they arose.

Although Bütschli in 1870 in his embryology of the honey-bee called attention to the “great similarity which the eleven pairs of invaginations in the eleven first trunk-segments in their first indication (anlage) have with the spinning-glands, and also with the segmental organs of Annelids,” he did not go further than this, and it is now known that in the 2d maxillary segment open not only spinning-glands, but in the embryo a pair of stigmata.

Paul Mayer, however, regarded the tracheæ and urinary tubes as homodynamous structures, and this view was advocated by Grassi (1885) for the reason that while in the embryo honey-bee there are ten pairs of stigmata, the first thoracic and two last abdominal segments wanting them, the germs of the urinary tubes arise in a corresponding situation on the two last abdominal segments. To this view Emery (Biol. Centralb., 1886, p. 692) objects that in Peripatus the nephridia and tracheæ “have nothing to do with the segmental organs,” as Peripatus besides nephridia possesses both coxal glands and tracheæ.

Both Kennell and Lang derive the coxal glands of Arthropoda from the setiparous or parapodial glands of annelid worms, and the recent endeavor of Bernard to show that the tracheæ arose from setiparous glands seems to be disproved by the fact that in insects as well as in other Arthropoda coxal glands with their outlets exist in the same segments as those bearing stigmata. Reasoning by exclusion, we are led to regard Kennell’s original view as the soundest.

Patten, however, regards the tracheæ as modified ends of nephridia, remarking: “Since in Acilius some of the abdominal tracheæ at first communicate with the cavities of the mesoblastic somites, it is probable that all the tracheæ represent the ectodermic portions of the nephridia.” (Origin of Vertebrates from Arachnids, p. 355.)

It is probable, therefore, that the tracheæ first arose as modifications of dermal glands, as in mites and Peripatus, and that at first they were not provided with tænidia (as in Chilopoda), while in later forms tænidia were developed. In the earliest tracheate forms the stigmata were not segmentally arranged, probably appearing irregularly anywhere in the body, but afterwards in the myriopods and insects became serially arranged.

d. The spiral threads or tænidia

It is generally supposed that the so-called “spiral thread” forms a continuous thread from one end of a tracheal branch to the other. This was first shown not to be the case by Platner in 1844. Minot has proved that “there is not a single spiral thread, but several, which run parallel to one another and end after making a few turns around the trachea.”

The tænidia we have found to be in some cases separate, independent, solid rings, though when there is more than one turn the thread necessarily becomes spiral. The tænidia of a main branch stop at the origin of the smaller branches, and a new set begins at the origin of each branch. The tænidia at the origin of the branch do not pass entirely around the inside of the peritoneal membrane; in the axils they are short, separate, spindle-shaped bands (Fig. 409).

At one point in the main trachea of the larva of Datana the tænidia were seen to end singly on one side (at a considerable distance from any branch or axil) at intervals, with a tænidium situated between them, making four or five turns; then there is only one band situated between two ends; this band or thread is succeeded by a set with five turns between the two ends, this set being succeeded by one complete ring situated between two ends; in all cases the ends vary in length, some threads being short and others long, so that they apparently end anywhere along the circumference of the trachea, and this arrangement is seen to apparently extend along the whole length of the trachea. Hence it is seen that as a rule the tænidia vary much in length, and never, as generally supposed, pass continuously from one end to another of a tracheal branch, for there are many spirals in a branch, each making only from one to five turns, most usually four turns. Fig. 408, part of a trachea of Dyticus marginatus, shows that at a slight bend in a trachea the tænidia is interrupted, and short, incomplete, wedge-shaped tænidia (e) are interpolated; at A, d is seen a split in one of the tænidia (compare also MacLeod, Pl. 1, Fig. 9). The threads are quite irregular in width. In the axils of the branches there is, as seen in Fig. 409, a basketwork of independent, short, often spindle-shaped tænidia; these are succeeded by longer ones, until we have threads passing entirely around near the base of each new branch; these being succeeded by others which make from two to five spiral turns.

The shape of the tænidia appears to vary to a great extent. In lepidopterous insects we have observed them to be in their general shape rather flat and slightly concavo-convex, the hollow looking towards the centre of the trachea. Minot’s section (Fig. 393) shows that in Hydrophilus they are cylindrical and solid, and Chun states that those of Stratiomys are round, while in Eristalis they are round, with a ridge projecting into the cavity of the trachea; in Æschna the thread is quadrangular. MacLeod states that sometimes it is cylindrical, in other cases flat, likewise prismatic; Macloskie believes that the spiral threads of the centipede are “fine tubules, externally opening by a fissure along their course.”

Stokes confirms Macloskie’s statements, stating that in the hemipterous Zaitha fluminea “the tænidia are fissured tubules formed within and from chitinized folds of the intima, the convexity of the folds looking towards the lumen of the tracheæ.” In Fig. 414, 1, are represented portions of several tænidia showing the fissure, which is sometimes interrupted; at 2 are seen “the formation of what may be called apertures in a chitinous bridge.” Stokes regards the tænidia as “inwardly directed folds of the membrane.” Near the spiracles the tracheal membrane is externally studded with minute papillæ, as shown at 3, where are represented three broad and incomplete tænidia, with the tapering end, or the beginning, of another. Stokes adds, “Here they are only broad grooves, with no appearance of the narrow fissure of the completed tænidium. At 4 is figured a portion of the internal surface of a large trachea near the external orifice, the tænidia being in an incipient stage, evidently forming more or less of a network, as is usually the case next to the stigma” (compare p. 451, and Fig. 414).

The tracheæ of chilopod myriopods appear to be like those of insects. A number of authors have failed to detect the spiral threads in the Julidæ. As to the Arachnida, several observers, including Menge and Bertkau, have denied the existence of the spiral thread in the spiders with the exception of the Attidæ; and MacLeod finds them “scarcely visible” in Argyroneta.

Besides the tracheæ, the salivary duct is kept permanently distended by tænidia, which, however, are not spiral. They usually form incomplete rings, as in Stomoxys, arranged as shown in Fig. 410.

The labella (proboscis) of flies are supported by incomplete chitinous tubes or “pseudo-tracheæ,” the ends of which form the scraping teeth, this being, according to Dimmock, their primary function. Dimmock describes them as cylindrical channels opening on the surface in zigzag slits. These channels are held open by incomplete rings, one end of which is forked. “These rings are apparently arranged so that one has its fork on one side of the opening of the channel, the next ring the fork on the opposite side of the channel, and so on, in alternation. Their true structure is revealed when flattened out.”

The use of the elastic tænidia is to render the tracheæ elastic, and to keep them permanently open, as is the case with the parallel rings of the trachea of the higher vertebrates. The tracheæ are thus rendered firm and solid, at the least expense of chitinous material. The spiral thread, as MacLeod remarks, “is the realization in nature of what engineers call a form of the greatest resistance.”

The tænidia are wanting in the fine endings of the tracheæ (tracheal capillaries); also in the cockroach, according to Miall and Denny, they are not developed in the large tracheæ close to the spiracles, and the intima or wall of the tube has a tessellated instead of a spiral marking (Fig. 411). The same structure is seen in the Perlidæ (Nemoura, Gerstaecker, Zeit. f. wissen. Zool. xxiv, Taf. xxiii, Figs. 5 and 7); also in Æschna (Hagen, Zool. Anz. 1880, p. 159). In certain fine tracheæ of the eyes of the fly no spiral threads are developed. (Hickson.) The air-sacs or dilated tracheæ are also without tænidia.

While in the living insect the main and smaller tracheæ are filled, with air, it is stated by Von Wistinghausen that the fine capillary ends contain a fluid.

e. Origin of the tracheæ and of the “spiral thread”

While we owe to Bütschli the discovery of the mode of origin and morphology of the tracheæ, which as he has shown^[66] arise by invaginations of the ectoblast; there being originally a single layer of epiblastic cells concerned in the formation of the tracheæ; we are indebted to Weismann^[67] for the discovery of the mode of origin of the “intima,” from the epiblastic layer of cells forming the primitive foundation of the tracheal structure.

Weismann did not observe the earliest steps in the process of formation of the stigma and main trunk of the tracheæ, which Bütschli afterwards clearly described and figured.

Weismann, however, thus describes the mode of development of the intima; after describing the cells destined to form the peritoneal membrane, he says: “The lumen is filled with a clear fluid and already shows a definite border in a slight thickening of the cell-wall next to it.

“Very soon this thickening forms a thin, structureless intima, which passes as a delicate double line along the cells, and shows its dependence on the cells by a sort of adherence to the rounded sides of the cells (Taf. vii, 97 A, a b c). Throughout the mass, as the intima thickens, the cells lose their independence, their walls pressing together and coalescing, and soon the considerably enlarged hollow cylinder of the intima is surrounded by a homogeneous layer of a tissue, whose origin from cells is recognized only by the regular position of the rounded nuclei (Taf. vii, Fig. 97, B).

“Then as soon as the wavy bands of the intima entirely disappear, and it forms a straight, cylindrical tube, a fine pale cross-striation becomes noticeable (vii, 97, B, int), which forms the well-known ‘spiral thread,’ a structure which, as Leydig has shown, possesses no independence, but arises merely from a partial thickening of the originally homogeneous intima.

“Meyer’s idea that the spiral threads are fissures in the intima produced by the entrance of air is disproved by the fact that the spiral threads are present long before the air enters. Hence the correctness of Leydig’s view, based on the histological structure of the tracheæ, is confirmed by the embryological development, and the old idea of three membranes, which both Meyer and Milne-Edwards maintain, must be given up.”

Weismann also contends that the elastic membrane bearing the “spiral thread” is in no sense a primary membrane, not corresponding histologically to a cellular membrane. On the contrary, the “peritoneal membrane comprises the primary element of the trachea; it is nowhere absent, but envelops the smallest branches, as well as the largest trunks, only varying in thickness, which in the embryo and the young larva of Musca stands in relation to the thickness of the lumen.”

The trachea, then, consists primarily of an epithelial layer, the “peritoneal membrane,” or the invaginated epiblast; from this layer an intima is secreted, just as the skin or cuticle is secreted by the hypodermis. We may call the peritoneal membrane the ectotrachea, the intima or inner layer derived from the ectotrachea the endotrachea. The so-called “spiral threads” are a thickening of the endotracheal membrane, sometimes arranged in a spiral manner. For these chitinous bands we have proposed the name tænidia (Greek, little bands).

As to the origin of the spiral thread our observations^[68] have been made on the caterpillar of a species of Datana, which was placed in alcohol, just before pupation, when the larva was in a semipupal condition, and the larval skin could be readily stripped off. At this time the ectotrachea of the larva had undergone histolysis, nothing remaining but the moulted endotrachea, represented by the tænidia, which lay loosely within the cavity of the trachea. The ectotrachea or peritoneal membrane of the pupa is meanwhile in process of formation; the nuclear origin of the tænidia is now very apparent.