In the simple eye the non-faceted cornea and the retinula are readily made out, but the crystalline cones are not developed as such. The morphological key to both structures is found in the integument, of which the whole eye, simple or compound, is a modification. A defined tract of the chitinous cuticle becomes transparent, and either swells into a lens (fig. 53), or becomes regularly divided into facets (fig. 55), which are merely the elaboration of imperfectly separated polygonal areas, easily recognised in the young cuticle of all parts of the body. Next, the chitinogenous layer is folded inwards, so as to form a cup, and this, by the narrowing of the mouth, is transformed into a flask, and ultimately into a solid two-layered cellular mass (fig. 53). The deep layer undergoes conversion into a retina, its chitinogenous cells developing the nerve-rods as interstitial structures, while the superficial layer, which loses its functional importance in the simple eye, gives rise by a similar process of interstitial growth to the crystalline cones of the compound eye (fig. 55). The basement-membrane, underlying the chitinogenous cells, is transformed into the fenestrated membrane. The nerve-rods stand upon it, like organ pipes upon the sound-board, while fibrils of the optic nerve and fine tracheæ pass through its perforations. The mother-cells of the crystalline cones and nerve-rods are largely replaced by the interstitial substances they produce, to which they form a sheath; they are often loaded with pigment, and the nuclei of the primitive-cells can only be distinguished after the colouring-matter has been discharged by acids or alkalis.

Dr. Hickson112 has lately investigated the minute anatomy of the optic tract in various Insects. He finds, in the adult of the higher Insects, three distinct ganglionic swellings, consisting of a network of fine fibrils, surrounded by a sheath of crowded nerve-cells. Between the ganglia the fibres usually decussate. In the Cockroach, and some other of the lower Insects, the outermost ganglion is undeveloped. The fibres connecting the second ganglion with the eye take a straight course in the young Cockroach, but partially decussate in the adult.

The investigation, like so many other trains of biological inquiry, begins with Leeuwenhoeck (Ep. ad Soc. Reg. Angl. iii.), who ascertained that the cornea of a shardborne Beetle, placed in the field of a microscope, gives images of surrounding objects, and that these images are inverted. When the cornea is flattened out for microscopic examination, the images (e.g., of a window or candle-flame) are similar, and it has been too hastily assumed that a multitude of identical images are perceived by the Insect. The cornea of the living animal is, however, convex, and the images formed by different facets cannot be precisely identical. No combined or collective image is formed by the cornea. When the structure of the compound eye had been very inadequately studied, as was the case even in Cuvier’s time (Leçons d’Anat. Comp., xii., 14), it was natural to suppose that all the fibres internal to the cornea were sensory, that they formed a kind of retina upon which the images produced by the facets were received, and that these images were transmitted to the brain, to be united, either by optical or mental combination, into a single picture. Müller,113 in 1826, pointed out that so simple an explanation was inadmissible. He granted that the simple eye, with its lens and concave retina, produces a single inverted image, which is able to affect the nerve-endings in the same manner as in Vertebrates. But the compound eye is not optically constructed so as to render possible the formation of continuous images. The refractive and elongate crystalline cones, with their pointed apices and densely pigmented sides, must destroy any images formed by the lenses of the cornea. Even if the dioptric arrangement permitted the formation of images, there is no screen to receive them.114 Lastly, if this difficulty were removed, Müller thought it impossible for the nervous centres to combine a great number of inverted partial images. How then can Insects and Crustaceans see with their compound eyes? Müller answered that each facet transmits a small pencil of rays travelling in the direction of its axis, but intercepts all others. The refractive lens collects the rays, and the pigmented as well as refractive crystalline cone further concentrates the pencil, while it stops out all rays which diverge appreciably from the axis. Each element of the compound eye transmits a single impression of greater or less brightness, and the brain combines these impressions into some kind of picture, a picture like that which could be produced by stippling. It may be added that the movements of the insect’s head or body would render the distance and form of every object in view much readier of appreciation. No accommodation for distance would be necessary, and the absence of all means of accommodation ceases to be perplexing. Such is Müller’s theory of what he termed “mosaic vision.” Many important researches, some contradictory, some confirmatory of Müller’s doctrine,115 have since been placed on record, with the general result that some modification of Müller’s theory tends to prevail. The most important of the new facts and considerations which demand attention are these:—

Reasons have been given for supposing that images are formed by the cornea and crystalline cones together. This was first pointed out by Gottsche (1852), who used the compound eyes of Flies for demonstration. Grenacher has since ascertained that the crystalline cones of Flies are so fluid that they can hardly be removed, and he believes that Gottsche’s images were formed by the corneal facets alone. He finds, however, that the experiment may be successfully performed with eyes not liable to this objection, e.g., the eyes of nocturnal Lepidoptera. A bit of a Moth’s eye is cut out, treated with nitric acid to remove the pigment, and placed on a glass slip in the field of the microscope. The crystalline cones, still attached to the cornea, are turned towards the observer, and one is selected whose axis coincides with that of the microscope. No image is visible when the tip of the cone is in focus, but as the cornea approaches the focus, a bristle, moved about between the mirror and the stage, becomes visible. This experiment is far from decisive. No image is formed where sensory elements are present to receive and transmit it. Moreover, the image is that of an object very near to the cornea, whereas all observations of living Insects show that the compound eye is used for far sight, and the simple eye for near sight. Lastly, the treatment with acid, though unavoidable, may conceivably affect the result. It is not certain that the cones really assist in the production of the image, which may be due to the corneal facets alone, though modified by the decolorised cones.

Grenacher has pointed out that the composition of the nerve-rod furnishes a test of the mosaic theory. According as the percipient rod is simple or complex, we may infer that its physiological action will be simple or complex too. The adequate perception of a continuous picture, though of small extent, will require many retinal rods; on the other hand, a single rod will suffice for the discrimination of a bright point. What then are the facts of structure? Grenacher has ascertained that the retinal rods in each element of the compound eye rarely exceed seven, and often fall as low as four—further, that the rods in each group are often more or less completely fused so as to resemble simple structures, and that this is especially the case with Insects of keen sight.116

Certain facts described by Schultze tell on the other side. Coming to the Arthropod eye, fresh from his investigation of the vertebrate retina, Schultze found in the retinal rods of Insects the same lamellar structure which he had discovered in Vertebrata. He found also that in certain Moths, Beetles, and Crustacea, a bundle of extremely fine fibrils formed the outer extremity of each retinal or nerve-rod. This led him to reject the mosaic theory of vision, and to conclude that a partial image was formed behind every crystalline cone, and projected upon a multitude of fine nerve-endings. Such a retinula of delicate fibrils has received no physiological explanation, but it is now known to be of comparatively rare occurrence; it has no pigment to localise the stimulus of light; and there is no reason to suppose that an image can be formed within its limits.

The optical possibility of such an eye as that interpreted to us by Müller has been conceded by physicists and physiologists so eminent as Helmholz and Du Bois Reymond. Nevertheless, the competence of any sort of mosaic vision to explain the precise and accurate perception of Insects comes again and again into question whenever we watch the movements of a House-fly as it avoids the hand, of a Bee flying from flower to flower, or of a Dragon-fly in pursuit of its prey. The sight of such Insects as these must range over several feet at least, and within this field they must be supposed to distinguish small objects with rapidity and certainty. How can we suppose that an eye without retinal screen, or accommodation for distance, is compatible with sight so keen and discriminating? The answer is neither ready nor complete, but our own eyesight shows how much may be accomplished by means of instruments far from optically perfect. According to Aubert, objects, to be perceived as distinct by the human eye, must have an angular distance of from 50″ to 70″, corresponding to several retinal rods. Our vision is therefore mosaic too, and the retinal rods which can be simultaneously affected comprise only a fraction of those contained within the not very extensive area of the effective retina. Still we are not conscious of any break in the continuity of the field of vision. The incessant and involuntary movements of the eyeball, and the appreciable duration of the light-stimulus partly explain the continuity of the image received upon a discontinuous organ. Even more important is the action of the judgment and imagination, which complete the blanks in the sensorial picture, and translate the shorthand of the retina into a full-length description. That much of what we see is seen by the mind only is attested by the inadequate impression made upon us by a sudden glimpse of unfamiliar objects. We need time and reflection to interpret the hints flashed upon our eyes, and without time and reflection we see nothing in its true relations. The Insect-eye may be far from optical perfection, and yet, as it ranges over known objects, the Insect-mind, trained to interpret colour, and varying brightness, and parallax, may gain minute and accurate information. Grant that the compound eye is imperfect, and even rude, if regarded as a camera; this is not its true character. It is intended to receive and interpret flashing signals; it is an optical telegraph.

Plateau117 has recently submitted the seeing powers of a number of different Insects to actual experiment. The two windows of a room five metres square were darkened. An aperture fitted with ground glass was then arranged in each window. At a distance of four metres from the centre of the space between the windows captive Insects were from time to time liberated. One of the windows was fenced with fine trellis, so as to prevent the passage of the Insect, or otherwise altered in form, but the size of the aperture could be increased at pleasure, so as exactly to make up for any loss of light caused thereby, the brightness of the two openings being compared by a photometer.

It was found that day-flying Insects require a tolerably good light; in semi-obscurity they cannot find their way, and often refuse to fly at all. By varnishing one or other set in Insects possessing both simple and compound eyes, it was found that day-flying Insects provided with compound eyes do not use their simple eyes to direct their course. When the light from one window was sensibly greater than that from the other, the Insect commonly chose the brightest, but the existence of bars, close enough to prevent or to check its passage, had no perceptible effect upon the choice of its direction. Alterations in the shape of one of the panes seemed to be immaterial, provided that the quantity of light passing through remained the same, or nearly the same. Plateau concludes that Insects do not distinguish the forms of objects, or distinguish them very imperfectly.

It is plain, and Plateau makes this remark himself, that such experiments upon the power of unaided vision in Insects, give a very inadequate notion of the facility with which an Insect flying at large can find its way. There the animal is guided by colour, smell, and the actual or apparent movements of all visible objects. Exner has pointed out how important are the indications given by movement. Even in man, the central part of the retina is alone capable of precise perception of form, but a moving object is observed by the peripheral tract. Plateau (from whom this quotation is made) adds that most animals are very slightly impressed by the mere form of their enemies, or of their prey, but the slightest movement attracts their notice. The sportsman, the fisherman, and the entomologist cannot fail to learn this fact by repeated and cogent proofs.

Sense of Smell in Insects.

The existence of a sense of smell in Insects has probably never been disputed. Many facts of common observation prove that carrion-feeders, for example, are powerfully attracted towards putrid animal substances placed out of sight. The situation of the olfactory organs has only been ascertained by varied experiments and repeated discussion. Rosenthal, in 1811, and Lefebvre, in 1838, indicated the antennæ as the organs of smell, basing their conclusions upon physiological observations made upon living insects. Many entomologists of that time were inclined to regard the antennæ as auditory organs.118 Observations on the minute structure of the antennæ were made by many workers, but for want of good histological methods and accurate information concerning the organs of smell in other animals, these proved for a long time indecisive. It was by observation of living insects that the point was actually determined.

Hauser’s experiments, though by no means the first, are the most instructive which we possess. He found that captive insects, though not alarmed by a clean glass rod cautiously brought near, became agitated if the same rod had been first dipped in carbolic acid, turpentine, or acetic acid. The antennæ performed active movements while the rod was still distant, and after it was withdrawn the insect was observed to wipe its antennæ by drawing them through its mouth. After the antennæ had been extirpated or coated with paraffin, the same insects became indifferent to strong-smelling substances, though brought quite near. Extirpation of the antennæ prevented flies from discovering putrid flesh, and hindered or prevented copulation in insects known to breed in captivity.

Following up these experiments by histological investigation of many insects belonging to different orders, Hauser clearly established the following points, which had been partially made known before:—

The sensory elements of the antennæ are lodged in grooves or pits, which may be filled with fluid. The nerve-endings are associated with peculiar rods, representing modified chitinogenous cells. The number of grooves or pits may be enormous. In the male of the Cockchafer, Hauser estimates that there are 39,000 in each antenna. He remarks that in all cases where the female Insect is sluggish and prone to concealment, the male has the antennæ more largely developed than the female.

Sense of Taste in Insects.

F. Will119 gives an account of many authors who have investigated with more or less success the sense organs of various Insects. He relates also the results of his own experiments, and gives anatomical details of the sensory organs of the mouth in various Hymenoptera.

Wasps, flying at liberty, were allowed to visit and taste a packet of powdered sugar. This was left undisturbed for some hours, and then replaced by alum of the same appearance. The Wasps attacked the alum, but soon indicated by droll movements that they perceived the difference. They put their tongues in and out and cleansed them from the ill-tasted powder. Two persisted at the alum till they rolled on the table in agony, but they soon recovered and flew away. In a few hours the packet was quite deserted. After a day’s interval, during which the sugar lay in its usual place, powdered, and of course perfectly tasteless, dolomite was substituted. The wasps licked it diligently and could not be persuaded for a long time that it could do nothing for them. Similar experiments were made with other substances, and Insects whose antennæ and palps had been removed were subjected to trial. The result clearly proved that a sense of taste existed, and that its seat is in the mouth.120 Peculiar nerve-endings, such as Meinert and Forel had previously found in Ants, were found in abundance on the labium, the paraglossæ, and the inner side of the maxillæ of the Wasp. Some lay in pits, through the bases of which single nerves emerged, and swelled into bulbs, or passed into peculiar conical sheaths. Interspersed among the gustatory nerve-endings were setæ of various kinds, some protective, some tactile, and others intended to act as guiding-hairs for the saliva.

Will observes that the organs described satisfy the essential conditions of a sense of taste. The nerve-endings pass free to the surface, and are thus directly accessible to chemical stimulus. Further, they are so placed that they and the particles of food which get access to them are readily bathed by the saliva. Moistened or dissolved in this fluid, the sapid properties of food are most fully developed.

The sensory pits and bulbs appropriated to taste are believed to be unusually abundant in the social Hymenoptera.

Sense of Hearing in Insects.

The auditory organs of Insects and other Arthropoda are remarkable for the various parts of the body in which they occur. Thus they have been found in the first abdominal segment of Locusts, and in the tibia of the fore-leg of Crickets and Grasshoppers, and more questionable structures with peculiar nerve-endings have been described as occurring in the hinder part of the abdomen of various larvæ (Ptychoptera, Tabanus, &c). The auditory organ of Decapod Crustacea is lodged in the base of the antennule, that of Stomapods in the tail, while an auditory organ has been lately discovered on the underside of the head of the Myriopod Scutigera.

Auditory organs are best developed in such Insects as produce sounds as a call to each other. The Cockroach is dumb, and it is, therefore, not a matter of surprise that no structure which can be considered auditory should have ever been detected in this Insect.121

The sensory hairs of the skin have been already noticed (p. 31).

CHAPTER VII.

The alimentary canal of the Cockroach measures about 2 3/4 inches in length, and is therefore about 2 3/4 times the length of the body. In herbivorous Insects the relative length of the alimentary canal may be much greater than this; it is five times the length of the body in Hydrophilus. Parts of the canal are specialised for different digestive offices, and their order and relative size are given in the following table:—

Considered with respect to its mode of formation, the alimentary canal of all but the very simplest animals falls into three sections—viz., (1) the mesenteron, or primitive digestive cavity, lined by hypoblast; (2) the stomodæum, or mouth-section, lined by epiblast, continuous with that of the external surface; and (3) the proctodæum, or anal section, lined by epiblast folded inwards from the anus, just as the epiblast of the stomodæum is folded in from the mouth. The mesenteron of the Cockroach is very short, as in other Arthropoda, and includes only the chylific stomach with its diverticula. The mouth, œsophagus, and crop form the stomodæum, while the proctodæum begins with the Malpighian tubules, and extends thence to the anus. Both stomodæum and proctodæum have a chitinous lining, which is wanting in the mesenteron. At the time of moult, or a little after, this lining is broken up and passed out of the body.

The mouth of the Cockroach is enclosed between the labrum in front, and the labium behind, while it is bounded laterally by the mandibles and first pair of maxillæ. The chitinous lining is thrown into many folds, some of which can be obliterated by distension, while others are permanent and filled with solid tissues. The lingua is such a permanent fold, lying like a tongue upon the posterior wall of the cavity and reaching as far as the external opening. The thin chitinous surface of the lingua is hairy, like other parts of the mouth, and stiffened by special chitinous rods or bands. The salivary ducts open by a common orifice on its hinder surface. Above, the mouth leads into a narrow gullet or œsophagus, with longitudinally folded walls, which traverses the nervous ring, and then passes through the occipital foramen to the neck and thorax. Here it gradually dilates into the long and capacious crop, whose large rounded end occupies the fore-part of the abdomen. When empty, or half-empty, the wall of the crop contracts, and is thrown into longitudinal folds, which disappear on distension. Numerous tracheal tubes ramify upon its outer surface, and appear as fine white threads upon a greenish-grey ground.

Three layers can be distinguished in the wall of the crop—viz., (1) the muscular, (2) the epithelial, and (3) the chitinous layer.122 The muscular layer consists of annular and longitudinal fibres, crossing at right angles. (See fig. 58.) In most animals the muscles of organic life, subservient to nutrition and reproduction, are very largely composed of plain or unstriped fibres. In Arthropoda (with the exception of the anomalous Peripatus) this is not generally the case, and the muscular fibres of the alimentary canal belong to the striped variety. The epithelium rests upon a thin structureless basement-membrane, which is firmly united in the œsophagus and crop to the muscular layer and the epithelium. The epithelium consists of scattered nucleated cells, rounded or oval. These epithelial cells, homologues of the chitinogenous cells of the integument, secrete the transparent and structureless chitinous lining. Hairs (setæ) of elongate, conical form, and often articulated at the base, like the large setæ of the outer skin, are abundant. In the œsophagus they are very long, and grouped in bundles along sinuous transverse lines. In the crop the hairs become shorter, and the sinuous lines run into a polygonal network. The points of the hairs are directed backwards, and they no doubt serve to guide the flow of saliva towards the crop.

The gizzard has externally the form of a blunt cone, attached by its base to the hinder end of the crop, and produced at the other end into a narrow tube ( 1/4 to  1/3 in. long), which projects into the chylific stomach. Its muscular wall is thick, and consists of many layers of annular fibres, while the internal cavity is nearly closed by radiating folds of the chitinous lining. Six of the principal folds, the so-called “teeth,” are much stronger than the rest, and project so far inwards that they nearly meet. They vary in form, but are generally triangular in cross section and irregularly quadrilateral in side view. Between each pair are three much less prominent folds, and between these again are slight risings of the chitinous lining. A ridge runs along each side of the base of each principal tooth, and the minor folds, as well as part of the principal teeth, are covered with fine hairs. The central one of each set of secondary folds is produced behind into a spoon-shaped process, which extends considerably beyond the rest, and gradually subsides till it hardly projects from the internal surface of the gizzard. Behind each large tooth (i.e., towards the chylific stomach) is a rounded cushion set closely with hairs, and between and beyond these are hairy ridges. (See fig. 61.) The whole forms an elaborate machine for squeezing and straining the food, and recalls the gastric mill and pyloric strainer of the Crayfish. The powerful annular muscles approximate the teeth and folds, closing the passage, while small longitudinal muscles, which can be traced from the chitinous teeth to the cushions, appear to retract these last, and open a passage for the food.123

The gizzard ends below, as we have already mentioned, in a narrow cylindrical tube which is protruded into the chylific stomach for about one-third of an inch. Folds project from the wall of this tube, and reduce its central cavity to an irregular star-like figure. Below it ends in free processes slightly different from each other in size and shape. The chitinous lining and the chitinogenous layer beneath pass to the end of the tube and are then reflected upon its outer wall, ascending till they meet the lining epithelium of the cæcal tubes. Between the wall of the gizzard-tube and its external reflected layer, tracheal tubes, fat-cells, and longitudinal muscles are enclosed.

The chylific stomach is a simple cylindrical tube, provided at its anterior end with eight (sometimes fewer) cæcal tubes, and opening behind into the intestine. Its muscular coat consists of a loose layer of longitudinal fibres, enclosing annular fibres. Internal to these is a basement membrane, which supports an epithelium consisting of elongate cells which are often clustered into regular eminences, and separated by deep cavities. The epithelium forms no chitinous lining in the chylific stomach or cæcal tubes; and this peculiarity, no doubt, promotes absorption of soluble food in this part of the alimentary canal. Short processes are given off from the free ends of the epithelial cells, as in the intestines of many Mammalia and other animals.

Between the cells a reticulum is often to be seen, especially where the cells have burst; it extends between and among all the elements of the mucous lining, and probably serves, like the very similar structure met with in Mammalian intestines,124 to absorb and conduct some of the products of digestion. Different epithelial cells may be found in all the stages noticed by Watney—viz., (1) with divided nuclei; (2) small, newly produced cells at the base of the epithelium; (3) short and broad cells, overtopped by the older cells around; (4) dome-shaped masses of young cells, forming “epithelial buds”;125 (5) full-grown cells, ranging with those on either side, so as to form an unbroken and uniform series. The regeneration of the tissue is thus provided for. The cells come to maturity and burst, when new cells, the product of the epithelial buds, take their place.

The epithelium of the chylific stomach is continued into the eight cæcal tubes, where it undergoes a slight modification of form.

At the hinder end of the chylific stomach is a very short tube about half the diameter of the stomach, the small intestine. At its junction with the chylific stomach are attached, in six bundles, 60 or 70 long and fine tubules, the Malpighian tubules.126 The small intestine has the same general structure as the œsophagus and crop; its chitinous lining is hairy, and thrown into longitudinal folds which become much more prominent in the lower part of the tube. The junction of the small intestine with the colon is abrupt, and a strong annular fold assumes the character of a circular valve (fig. 68).

The rectum is about  1/4 inch long, and is dilated in the middle when distended. Six conspicuous longitudinal folds project into the lumen of the tube. These folds are characterised by an unusual development of the epithelium, which is altogether wanting in the intermediate spaces, where the chitinous lining blends with the basement-membrane, both being thrown into sharp longitudinal corrugations. Between the six epithelial bands and the muscular layer are as many triangular spaces, in which ramify tracheal tubes and fine nerves for the supply of the epithelium. The chitinous layer is finely setose. The muscular layer consists of annular fibres strengthened externally by longitudinal fibres along the interspaces between the six primary folds.127

The corrugated and non-epitheliated interspaces may be supposed to favour distension of the rectal chamber, while the great size of the cells of the bands of epithelium is perhaps due to their limited extent. Leydig128 attributed to these rectal bands a respiratory function, and compared them to the epithelial folds of the rectum of Libellulid larvæ, which, as is well known, respire by admitting fresh supplies of water into this cavity. It is an obvious objection that Cockroaches and other Insects in which the rectal bands are well developed do not take water into the intestine at all. Gegenbaur has therefore modified Leydig’s hypothesis. He suggests (Grundzüge d. Vergl. Anat.) that the functional rectal folds of Dragon-flies and the non-functional folds of terrestrial Insects are both survivals of tracheal gills, which were the only primitive organs of respiration of Insects. The late appearance of the rectal folds and the much earlier appearance of spiracles is a serious difficulty in the way of this view, as Chun has pointed out. It seems more probable that the respiratory appendages of the rectum of the Dragon-fly larvæ are special adaptations to aquatic conditions of a structure which originated in terrestrial Insects, and had primarily nothing to do with respiration.

The number of the rectal bands (six) is worthy of remark. We find six sets of folds in the gizzard and small intestine of the Cockroach, six bundles of Malpighian tubules, with six intermediate epitheliated bands. There are also six longitudinal bands in the intestine of the Lobster and Crayfish. The tendency to produce a six-banded stomodæum and proctodæum may possibly be related to the six theoretical elements (two tergal, two pleural, two sternal,) traceable in the Arthropod exoskeleton, of which the proctodæum and stomodæum are reflected folds.

The anus of the Cockroach opens beneath the tenth tergum, and between two “podical” plates. Anal glands, such as occur in some Beetles, have not been discovered in Cockroaches.

Appendages. The Salivary Glands.

The three principal appendages of the alimentary canal of the Cockroach are outgrowths of the three primary divisions of the digestive tube; the salivary glands are diverticula of the stomodæum, the cæcal tubes of the mesenteron, and the Malpighian tubules of the proctodæum.