Cœlenterata. The actual evolution of the eye is best shewn in the Hydrozoa. The simplest types are those found in Oceania and Lizzia[182]. In Lizzia the eye is placed at the base of a tentacle and consists of (fig. 276) a lens (l) and a percipient bulb (oc). The lens is a simple thickening of the cuticle, while the percipient part of the eye is formed of three kinds of elements:—(1) pigment cells; (2) sense cells, forming the true retinal elements, and consisting of a central swelling with the nucleus, a peripheral process representing a hardly differentiated rod, and a central process continuous with (3) ganglion cells at the base of the eye. In this eye there is present a commencing differentiation of a ganglion as well as of a retina.

The eye of Oceania is simpler than that of Lizzia in the absence of a lens. Claus has shewn that in Charybdea amongst the Acraspeda a more highly differentiated eye is present, with a lens formed of cells like the vertebrate eye.

It has (fig. 277 A) the form of a vesicle, with a small opening in the outer wall, placing the cavity of the vesicle in free communication with the exterior. The cells lining the posterior face of the vesicle form a retina (R); and are continuous with the fibres of the optic nerve (N.op). We have no knowledge of the development of this eye.

In the Gasteropods the eye (fig. 277 B) has the form of a closed vesicle: the cells lining the inner side form the retina, while the outer wall of the vesicle constitutes the cornea. A cuticular lens is placed in the cavity, on the side adjoining the cornea. This eye originates from the ectoderm, within the velar area, and close to the supraœsophageal ganglia, usually at the base of the tentacles. According to Rabl (Vol. II. No. 268) it is formed as an invagination, the opening of which soon closes; while according to Bobretzky (Vol. II. No. 242) and Fol it arises as a thickening of the epiblast, which becoming detached takes the form of a vesicle. It is quite possible that both types of development may occur, the second being no doubt abbreviated. The vesicle, however formed, soon acquires a covering of pigment, except for a small area of its outer wall, where the lens becomes formed as a small body projecting into the lumen of the vesicle. The lens seems to commence as a cuticular deposit, and to grow by the addition of concentric layers. The inner wall of the vesicle gives rise to the retina.

A brief description of its adult structure[183] will perhaps render more clear my account of the development. The most important features of the eye are shewn in fig. 277 C. The outermost layer of the optic bulb forms a kind of capsule, which may be called the sclerotic. Posteriorly the sclerotic abuts on the cartilaginous orbit, which encloses the optic ganglion (G. op); and in front it becomes transparent and forms the cornea Co, which may be either completely closed, or (as represented in the diagram) perforated by a larger or smaller opening. Behind the cornea is a chamber known as the anterior optic chamber. This chamber is continued back on each side round a great part of the circumference of the eye, and separates the sclerotic from a layer internal to it.

In the anterior optic chamber there are placed (1) the anterior part of the lens (l¹) and (2) the folds of the iris (Ir). The whole chamber, except the part formed by the lens, is lined by the epidermis (Int¹ and Int²). Bounding the inner side of the anterior optic chamber is a layer which is called the choroid (Int¹) which is continued anteriorly into the fold of the iris (Ir). The most superficial layer of the choroid is the epithelium already mentioned, next comes a layer of obliquely placed plates known as the argentea externa, then a layer of muscles, and finally the argentea interna. The argentea interna abuts on a cartilaginous capsule, which completely invests the inner part of the eye.

The lens is a nearly spherical body composed of concentric lamellæ of a structureless material. It is formed of a small outer (l¹) and large inner (l) segment, the two being separated by a thin membrane. It is supported by a peculiar projection of the wall of the optic cup, known as the ciliary body (Co.ep), inserted at the base of the iris, and mainly formed of a continuation of the retina. This body is however muscular, and presents a series of folds on its outer and inner surfaces, which are especially developed on the latter.

The first satisfactory account of the development of the eye is due to Lankester (No. 365). The more important features in it were also independently worked out by Grenacher (No. 363), and are beautifully illustrated in Bobretzky’s paper (No. 362). The eye first appears as an oval pit of the epiblast, the edge of which is formed by a projecting rim (fig. 278 A). The epiblast layer lining the floor of the pit soon becomes considerably thickened. By the growth inwards of the rim the mouth of the pit is gradually narrowed (fig. 278 B), resembling at this stage the eye of Nautilus, and finally closed. There is thus formed a flattened sack, lined by epiblast, which may be called the primary optic vesicle. Its cavity eventually forms the inner optic chamber. The anterior wall of the sack is lined by a much less columnar layer than the posterior, the former giving rise to the epithelium on the inner side of the ciliary processes, the latter to the retina.

The cavity of the sack rapidly enlarges, and assumes a spherical form. At the same time a layer of mesoblast grows in between the walls of the sack and the external epiblast. Two new structures soon arise nearly simultaneously (fig. 279),—which become in the adult eye the iris (cc) and the posterior segment of the lens. The iris is formed as a circular fold of the skin in front of the optic vesicle. It consists both of epiblast and mesoblast, and gives rise to a pit lined by epiblast. The posterior segment of the lens arises as a structureless rod-like body, which is shewn in fig. 279 depending from the inner side of the anterior wall of the optic vesicle. Its exact mode of origin is somewhat obscure. The following is Lankester’s account of it[184]: “It is formed entirely within the primitive optic chamber, and at first depends as a short cylindrical rod from the middle point of the anterior wall of that chamber, that is to say, from the point at which the chamber finally closed up. It grows subsequently by the deposition of concentric layers of a horny material round this cone. No cells appear to be immediately concerned in effecting the deposition, and it must be looked upon as an organic concretion, formed from the liquid contained in the primitive optic chamber.”

The lens would thus appear to be a cuticular structure. It gradually assumes a nearly spherical form; and is then composed of concentrically arranged layers (fig. 280, hl).

While the lens is being formed, the ciliary epithelium of the optic vesicle becomes divided into two layers, an outer layer of large cells and an inner of small cells. Both layers are at first continuous across the anterior wall of the optic chamber in front of the lens, but soon become confined to the sides (fig. 280 A, cc and gz). The inner layer is stated by Lankester to give rise to the muscles present in the adult. The mesoblast cells also disappear from the region in front of the lens, and the outer epithelium is converted into a kind of cuticular membrane. By these changes the original layers of cells in front of the lens become reduced to mere membranes,—a change which appears to be preparatory to the appearance of the anterior segment of the lens. The formation of the latter has not been fully followed out by any investigator except Bobretzky. His figures would seem to indicate that it is formed as a cuticular deposit in front of the membrane already spoken of (fig. 280 B, vl). The two segments of the lens appear at any rate to be separated by a membrane continuous with the ciliary region of the optic vesicle.

Grenacher believes that the front part of the lens is formed in a pocket-like depression of the epiblastic layer covering the outer side of the optic cup; and Lankester thinks that the lens “pushes its way through the median anterior area of the primitive optic chamber, and projects into the second or anterior optic chamber where the iridian folds lie closely upon it.”

While the lens is attaining its complete development there appears a fresh fold round the circumference of the eye, which gradually grows inwards so as to form a chamber outside the parts already present. This chamber is the anterior optic chamber of the adult. In most Cephalopods (fig. 277 C) the edges of the fold do not quite meet, but leave a larger or smaller aperture leading into the chamber containing the iris, outer segment of the lens, etc. In some forms however they meet and coalesce, and so shut off this chamber from communication with the exterior. The edge of the fold constitutes the cornea while the remainder of it gives rise to the sclerotic.

The retina is at first a thick layer of numerous rows of oval cells (fig. 279). When the inner segment of the lens is far advanced towards its complete formation pigment becomes deposited in the anterior part of the retina, and a layer of rods grows out from the surface turned towards the cavity of the optic vesicle (fig. 280 A, st). At a slightly later stage the retina becomes divided into two layers (Bobretzky), a thicker anterior layer, and a thinner posterior layer (fig. 280, rt and rt´´). The former is composed of two strata, (1) the rods and (2) a stratum with numerous rows of nuclei which becomes in the adult the granular layer with its pigment. The posterior layer gives rise to the cellular part of the posterior division of the retina, while layers of connective tissue around it give rise to the connective tissue of this portion of the retina (layer 6 in the scheme on p. 474). The nervous layer is derived from the optic ganglion which attaches itself to the inner side of the connective tissue layer.

The greater part of the choroid is formed from the mesoblast adjoining the retina, but the epithelium covering its outer wall is of epiblastic origin.

It is difficult to decide from development whether the Molluscan eyes, so far dealt with, originated in the first instance pari passu with the supraœsophageal ganglia or independently at a later period. On purely à priori ground I should be inclined to adopt the former alternative.

In addition to the above eyes there occur amongst Mollusca highly complicated eyes, of a very different kind, in two widely separated groups, viz. certain species of a genus of slug (Onchidium), and certain Lamellibranchiata. These eyes, though they have no doubt been evolved independently of each other, present certain remarkable points of agreement. In both of them the rods of the retina are turned away from the surface, and the nerve-fibres are placed, as in the Vertebrate eye, on the side of the retina which faces outwards.

The peculiar eyes of Onchidium, investigated by Semper[185], are scattered on the dorsal surface, there being normal eyes in the usual situation on the head. The eyes on the dorsal surface are formed of a cornea, a lens composed of 1-7 cells, and a retina surrounded by pigment; which is perforated in the centre by an optic nerve, the retinal elements being in the inverted position above mentioned.

The development of these eyes has been somewhat imperfectly studied in the adult, in which they continue to be formed anew. They arise by a differentiation of the epidermis at the end of a papilla. At first a few glandular cells appear in the epidermis in the situation where an eye is about to be formed. Then, by a further process of growth, an irregular mass of epidermic cells becomes developed, which pushes the glandular cells to one side, and constitutes the rudiment of the eye. This mass, becoming surrounded by pigment, unites with the optic nerve, and its cells then differentiate themselves, in situ, into the various elements of the eye. No explanation is offered by Semper of the inverted position of the rods, nor is any suggested by his account of the development. As pointed out by Semper these eyes are no doubt modifications of the sensory epithelium of the papillæ.

The eyes of Pecten and Spondylus[186] are placed on short stalks at the edge of the mantle, and are probably modifications of the tentacular processes of the mantle edge. They are provided with a cornea, a cellular lens, a vitreous chamber, and a retina. The retinal elements are inverted, and the optic nerve passes in at the side, but occupies, in reference to its ramifications, the same relative situation as the optic nerve in the Vertebrate eye. The development has unfortunately not yet been studied.

Our knowledge of the structure or still more of the development of the organ of vision of the Platyelminthes, Rotifera, and Echinodermata is too scanty to be of any general interest.

Chætopoda. Amongst the Chætopoda the cephalic eyes of Alciope (fig. 281) have been adequately investigated as to their anatomy by Greeff. These are provided with a large cuticular lens (l), separated from the retina by a wide cavity containing the vitreous humour. The retina is formed of a single row of cells, with rods at their free extremities, continuous at their opposite ends with nerve-fibres. The development of this eye has not been worked out. Eyes not situated on the head are found in Polyophthalmus, and have probably been evolved from the more indifferent type of sense-organ found by Eisig in the allied Capitellidæ.

Chætognatha[187]. The paired cephalic eyes of Sagitta are spherical bodies imbedded in the epidermis. They are formed of a central mass of pigment with three lenses partially imbedded in it. The outer covering of the eye is the retina, which is mainly composed of rod-bearing cells; the rods being placed in contact with the outer surface of each of the lenses. In the presence of three lenses the eye of Sagitta approaches in some respects the eye of the Arthropoda.

Arthropodan eye. A satisfactory elucidation of the phylogeny of Arthropodan eyes has not yet been given.

All the types of eyes found in the group (with exception of that of Peripatus)[188] present marked features of similarity, but I am inclined to view this similarity as due rather to the character of the exoskeleton modifying in a more or less similar way all the forms of visual organs, than to the descent of all these eyes from a common prototype. In none of these eyes is there present a chamber filled with fluid between the lens and the retina, but the space in question is filled with cells. This character sharply distinguishes them from such eyes as those of Alciope (fig. 281). The types of eyes which are found in the Arthropoda are briefly the following:

(1) Simple eyes. In all simple eyes the corneal lens is formed by a thickening of the cuticle. Such eyes are confined to the Tracheata.

There are three types of simple eyes. (a) A type in which the retinal cells are placed immediately behind the lens, found (Lowne) in the larvæ of some Diptera (Eristalis), and also in some Chilognatha.

(b) A type of simple eye found in some Chilopoda, and in some Insect larvæ (Dytiscus, etc.) (fig. 282), the parts of which are entirely derived from the epidermis. There is present a lens (l) formed as a thickening of the cuticle, a so-called vitreous humour (gl) formed of modified hypodermis cells, and a retina (r) derived from the same source. The outer ends of the retinal cells terminate in rods, and their inner ends are continuous with nerve-fibres.

The typical compound eye is formed (fig. 283) of a series of corneal lenses (c) developed from the cuticle; below which are placed bodies known as the crystalline cones, one to each corneal lens; and below the crystalline cones are placed bodies known as the retinulæ (r) constituting the percipient elements of the eye, each of them being formed of an axial rod, the rhabdom, and a number of cells surrounding it.

The crystalline cones are formed from the coalescence of cuticular deposits in several cells, the nuclei of which usually remain as Semper’s nuclei. These cells are probably simple hypodermis cells, but in some forms, e.g. Phronima, there may be a continuous layer of hypodermis cells between them and the cuticle. In various Insect eyes the cells which usually give rise to a crystalline cone may remain distinct, and such eyes have been called by Grenacher aconous eyes, while eyes with incompletely formed crystalline cones are called by him pseudoconous eyes.

The development of the compound eye has so far only been satisfactorily studied in some Crustacea by Bobretzky (No. 367); by whom it has been worked out in Palæmon and Astacus, but more fully in the latter, to which the following account refers:

After the separation of the supraœsophageal ganglia from the superficial epiblast, the cells of the epidermis in the region of the future eye become columnar, and so form the above-mentioned epidermic layer of the eye. This layer soon becomes two or three cells deep. At the same time the most superficial part of the adjoining supraœsophageal ganglion becomes partially constricted off from the remainder as the neural layer of the eye, but is separated by a small space from the thickened patch of epidermis. Into this space some mesoblast cells penetrate at a slightly later period. Both the epidermic and neural layers next become divided into two strata. The outer stratum of the epidermic layer gives rise to the crystalline cones and Semper’s nuclei; each crystalline cone being formed from four coalesced rods, developed as cuticular differentiations of four cells, the nuclei of which may be seen in the embryo on its outer side. The lower ends of the cones pass through the inner stratum of the epidermic disc, the cells of which become pigmented, and constitute the pigment cells surrounding the lower part of the crystalline cones in the adult. The outer end of each of the crystalline cones is surrounded by four cells, believed by Bobretzky to be identical with Semper’s nuclei[189]. These cells give rise in a later stage (not worked out in Astacus) to the cuticular corneal lenses.

Of the two strata of the neural layer the outer is several cells deep, while the inner is formed of elongated rod-like cells. Unfortunately however the fate of the two neural layers has not been worked out, though there can be but little doubt that the retinulæ originate from the outer layer.

The mesoblast which grows in between the neural and epidermic layers becomes a pigment layer, and probably also forms the perforated membrane between the crystalline cones and the retinulæ.

The above observations of Bobretzky would appear to indicate that the paired compound eyes of Crustacea belong to the type of cerebral eyes. How far this is also the case with the compound eyes of Insects is uncertain, in that it is quite possible that the latter eyes may have had an independent origin.

The relation between the paired and median eye of the Crustacea is also uncertain.

In the genus Euphausia amongst the Schizopods there is present a series of eyes placed on the sides of some of the thoracic legs and on the sides of the abdomen. The structure of these eyes, though not as yet satisfactorily made out, would appear to be very different from that of other Arthropodan visual organs.

The Eye of the Vertebrata. In view of the various structures which unite to form it, the eye is undoubtedly the most complicated organ of the Vertebrata; and though its mode of development is fairly constant throughout the group, it will be convenient shortly to describe what may be regarded as its typical development, and then to proceed to a comparative view of the origin of its various parts, and to enter into greater detail with reference to some of them. At the end of the section there is an account of the accessory structures connected with the eye.

The formation of the eye commences with the appearance of a pair of hollow outgrowths from the anterior cerebral vesicle or thalamencephalon, which arise in many instances, even before the closure of the medullary canal. These outgrowths, known as the optic vesicles, at first open freely into the cavity of the anterior cerebral vesicle. From this they soon however become partially constricted, and form vesicles (fig. 284, a), united to the base of the brain by comparatively narrow hollow stalks, the rudiments of the optic nerves. The constriction to which the stalk or optic nerve is due takes place obliquely downwards and backwards, so that the optic nerves open into the base of the front part of the thalamencephalon (fig. 284, b).

The external or superficial epiblast which covers, and is in most forms in immediate contact with, the most projecting portion of the optic vesicle, becomes thickened. This thickened portion is then driven inwards in the form of a shallow open pit with thick walls (fig. 285 A, o), carrying before it the front wall (r) of the optic vesicle. To such an extent does this involution of the superficial epiblast take place, that the front wall of the optic vesicle is pushed close up to the hind wall, and the cavity of the vesicle becomes almost obliterated (fig. 285 B).

The bulb of the optic vesicle is thus converted into a cup with double walls, containing in its cavity the portion of involuted epiblast. This cup, in order to distinguish its cavity from that of the original optic vesicle, is generally called the secondary optic vesicle. We may, for the sake of brevity, speak of it as the optic cup; in reality it never is a vesicle, since it always remains widely open in front. Of its double walls the inner or anterior (fig. 285 B, r) is formed from the front portion, the outer or posterior (fig. 285 B, u) from the hind portion of the wall of the primary optic vesicle. The inner or anterior (r), which very speedily becomes thicker than the other, is converted into the retina: in the outer or posterior (u), which remains thin, pigment is eventually deposited, and it ultimately becomes the tesselated pigment-layer of the choroid.

By the closure of its mouth the pit of the involuted epiblast becomes a completely closed sac with thick walls and a small central cavity (fig. 285 B, l). At the same time it breaks away from the external epiblast, which forms a continuous layer in front of it, all traces of the original opening being lost. There is thus left lying in the cup of the secondary optic vesicle, an isolated elliptical mass of epiblast. This is the rudiment of the lens. The small cavity within it speedily becomes still less by the thickening of the walls, especially of the hinder one.

At its first appearance the lens is in immediate contact with the anterior wall of the secondary optic vesicle (fig. 285 B). In a short time however, the lens is seen to lie in the mouth of the cup (fig. 288 D), a space (vh) (which is occupied by the vitreous humour) making its appearance between the lens and anterior wall of the vesicle.

In order to understand how this space is developed, the position of the optic vesicle and the relations of its stalk must be borne in mind.

The vesicle lies at the side of the head, and its stalk is directed downwards, inwards and backwards. The stalk in fact slants away from the vesicle. Hence, when the involution of the lens takes place, the direction in which the front wall of the vesicle is pushed in is not in a line with the axis of the stalk, as for simplicity’s sake has been represented in the diagram (fig. 285), but forms an obtuse angle with that axis, after the manner of fig. 286, where s´ represents the cavity of the stalk leading away from the almost obliterated cavity of the primary vesicle.

Fig. 286 represents the early stage at which the lens fills the whole cup of the secondary vesicle. The subsequent condition is brought about through the rapid growth of the walls of the cup. This growth however does not take place equally in all parts of the cup. The walls of the cup rise up all round except that point of the circumference of the cup which adjoins the stalk. While elsewhere the walls increase rapidly in height, carrying so to speak the lens with them, at this spot, which in the natural position of the eye is on its under surface, there is no growth: the wall is here imperfect, and a gap is left. Through this gap, which afterwards receives the name of the choroidal fissure, a way is open from the mesoblastic tissue surrounding the optic vesicle and stalk into the interior of the cavity of the cup.

From the manner of its formation the gap or fissure is evidently in a line with the axis of the optic stalk, and in order to be seen must be looked for on the under surface of the optic vesicle. In this position it is readily recognised in the embryo seen as a transparent object (fig. 118, chs).

Bearing in mind these relations of the gap to the optic stalk, the reader will understand how sections of the optic vesicle at this stage present very different appearances according to the plane in which the sections are taken.

When the head is viewed from underneath as a transparent object the eye presents very much the appearance represented in the diagram (fig. 287).

A section of such an eye taken along the line y, perpendicular to the plane of the paper, would give a figure corresponding to that of fig. 288 D. The lens, the cavity and double walls of the secondary vesicle, the remains of the primary cavity, would all be represented (the superficial epiblast of the head would also be shewn); but there would be nothing seen of either the stalk or the fissure. If on the other hand the section were taken in a plane parallel to the plane of the paper, at some distance above the level of the stalk, some such figure would be obtained as that shewn in fig. 288 E. Here the fissure f is obvious, and the communication of the cavity vh of the secondary vesicle with the outside of the eye evident; the section of course would not go through the superficial epiblast. Lastly, a section, taken perpendicular to the plane of the paper along the line z, i.e. through the fissure itself, would present the appearances of fig. 288 F, where the wall of the vesicle is entirely wanting in the region of the fissure marked by the position of the letter f. The external epiblast has been omitted in this figure.

With reference to the above description, taken with very slight alterations from the Elements of Embryology, Pt. 1., two points require to be noticed. Firstly it is extremely doubtful whether the invagination of the secondary optic vesicle is to be viewed as an actual mechanical result of the ingrowth of the lens. Secondly it seems probable that the choroid fissure is not simply due to an inequality in the growth of the walls of the secondary optic cup, but is partly due to a doubling up of the primary vesicle from the side along the line of the fissure, at the same time that the lens is being thrust in in front. In Mammalia, the doubling up involves the optic stalk, which becomes flattened (whereby its original cavity is obliterated) and then folded in on itself, so as to embrace a new central cavity continuous with the cavity of the vitreous humour. And in other forms a partial phenomenon of the same kind is usually observable, as is more particularly described in the sequel.

The inner or anterior wall is, from the first, thicker than the outer or posterior; and over the greater part of the cup this contrast increases with the growth of the eye, the anterior wall becoming markedly thicker and undergoing changes of which we shall have to speak directly (fig. 289).

In the front portion however, along, so to speak, the lip of the cup, anterior to a line which afterwards becomes the ora serrata, both layers cease to take part in the increased thickening, accompanied by peculiar histological changes, which the rest of the cup is undergoing. Thus a hind portion or true retina is marked off from a front portion.

The front portion, accompanied by the mesoblast which immediately overlies it, is behind the lens thrown into folds, the ciliary ridges; while further forward it bends in between the lens and the cornea to form the iris. The original wide opening of the optic cup is thus narrowed to a smaller orifice, the pupil; and the lens, which before lay in the open mouth of the cup, is now inclosed in its cavity. While in the hind portion of the cup or retina proper no deposit of black pigment takes place in the layer formed out of the inner or anterior wall of the vesicle; in the front portion forming the region of the iris, pigment is largely deposited throughout both layers, though first of all in the outer one, so that eventually this portion seems to become nothing more than a forward prolongation of the pigment epithelium of the choroid.