STUDIES IN THE RETINA. 443 Studies in the Retina: Rods and Cones in the Frog and in some other Amphibia. By H. M. Bernard, M.A.Cantab. (From the Biological Laboratories, Royal College of Science, London.) With Plates 30 and 31. PART II.1 THE first part of this paper was devoted to showing that the structures called "cones" in the amphibian retina were the earlier stages in the development of the new rods required by growth, and that they force their way in whei'ever there is room for them between already existing rods. The forms of these elements with the positions of their nuclei2 were alone dealt with. In this paper it is proposed to give some account of the intimate structure of the amphibian rod. The minute details to be described will necessitate some discussion of the physiological processes which, so far as I have been able to interpret them, underlie their forms. Little success has so far attended the attempts of naturalists to unravel the finer structure of the rods. Indeed, the sub- ject seems to have been temporarily closed by the classical researches of Max Schultze in the sixties, for since that time little or no advance has been made. The earlier 1 The final revision of this MS. was kindly undertaken by my friend Mr. Martin Woodward, during my temporary absence from England. s On page 44 I inadvertently attributed to Borysiekiewitz an observation of. my own. This will be fully dealt with in Part III.
444 H. M. BERNARD. literature is, however, full of scattered observations, and it is possible that some of them may have been overlooked by me or far too briefly noticed. I do not pretend to have mastered the whole of the literature on the retina. I do not wish, therefore, to make any claims to priority, but simply to describe my observations, referring briefly to those of former students, so far as I know of any covering the same ground. And here I should add that, while confining myself in this paper solely to the Amphibia,1 these researches have extended over other Vertebrates, and that the general conclusions arrived at are not drawn solely from the facts here described. Enveloping Membranes.—While the existence of the membrane investing the inner limb of the rod requires no demonstration, it has been much disputed whether the outer limbs possess any such envelope or not. Apart from the fact that such a covering is difficult to demonstrate, ifc is possible that the conception df the rod as a cuticular structure may have strengthened the doubt. It has long been known that the outer limbs of the rods can be made to divide up trans- versely into discs, and that on such a dissolution no investing membrane can be seen (cf. Max Schultze's figures, cArch. mikr. Anat.,' Bd. iii, pi. xiii, figs, lie, etc.). Merkel2 found membranes whei'ever he looked for them except in the Amphibia, while Landolts figured very thick homogeneous membranes covering the rods (frog and newt) and the cones (newt). I am not aware that these have ever been confirmed, and I doubt their existence. He figures them even passing in between outer limb and ellipsoid. Something like what he figures may be seen in my fig. 13, c—i, the significance of which will be discussed later on. In the meantime I may state that I do not regard the thick rind there shown as an outer covering. 1 The following forms have been examined:—Rana temporaria, Bufo vulgaris, Molge cristata and M. vulgaris, Salamandra maculosa, and Siredou pisciformis. 1 'Arch. Anat. u. Phys.,' 1870, pi. xiv. » • Arch, mikr. Auat.,' Bd. vii, 1871, p. 81, pi. ix,
STUDIES IN THE EETINA. 445 That enveloping membranes occur in the outer limbs of Amphibian rods is certain, both on theoretical grounds and because they can be demonstrated. As we saw in Part I of this paper (this Journal, p. 29), the rods are primarily protoplasmic vesicles protruded from the retina. The walls of the vesicles are of extraordinary delicacy and transparency, and it will be a triumph of microscopic technique when retinas can be fixed so as to show them iutact. They are best seen in their very earliest stages of protrusion, before any rods are formed and the pigment is ouly just being forced away from the retina by tlieir increase in number and size. From this early stage we traced them through tlieir principal form-phases till they became normal rods, and all these phases were not only consistent with their being long, membranous sacs, but even confirmatory of this conception of their essential structure. Lastly, the persist- ence of the membrane covering the inner limbs has, as we have seen, long been an established fact. But granted that in all the earlier stages of the rods we have a wall to the vesicle—a wall which persists in the inner limb,—we have still to ask whether that is the case with the outer limb when the rod is complete. May not the proto- plasmic wall merge in the substance which fills up the interior of the outer limb and lose its individuality, so that it would be impossible to speak of any investing membrane ? This is, of course, quite possible ; and, moreover, it is certain that even if it preserved its individuality one would rarely expect to demonstrate the existence of such a delicate film of transparent protoplasm round the outer limb of the amphibian rod, with its usually refractive contents. Actual observa- tions, however, show clearly that the protoplasmic wall does retain its individuality, and that to the last the rod is a thin protoplasmic vesicle filled up with matter, the origin and nature of which will be discussed in the following pages. As demonstration of this persistence of the protoplasmic wall of the vesicle, I will call attention to PI. 30, fig. 2, which is taken from the retina of a toad. All the rods in this
446 H. M. BERNABD. retina are obviously bags which have, under pressures and strains, lost their normal cylindrical shapes, and are now pulled out or crushed together into every variety of form, from short, rounded sacs to long, thin clubs with round knobs at the tips.1 Endless, too, are the instances in which the inner and outer limbs have been pulled somewhat apart, and the stretched or torn membrane becomes visible under good microscopic powers. Important, also, in this connection is PI. 30, fig. 1, from a newt, in which in one spot all the rods were broken away, but their basal portions persisted emptied of all contents except the remains of the ellipsoids. This last is a fortunate observation, because it shows that, in essence, the inner and outer limbs are simply two sacs separated by a thin wall, and that the great differences seen between them must be referred entirely to their contents. To this we shall return later. Lastly, I have sections of a newt's retina in which the thin coverings of the rods have taken stain, and are quite demon- strable in optical sections. External Markings.—The longitudinal striation of the outer limbs of the rods has long been seen, but its nature has never been satisfactorily settled. Max Scbultze regarded it as a furrowing of the surface, and figured the cross-sections of the rod as having an outline like that shown in PI. 30, fig. 8. With regard to the inner limbs of the rods in Amphibia, exact records of striation are few. The well-known " Faser- korb " of Max Schultze was found by him, " essentially the same," in all the classes of Vertebrates, including the Am- phibia. He records finding it in the axolotl, and he figures it in the newt.3 Hoffmann also figures the upper ends of these or similar threads round the bases of the inner limbs of amphibian rods, and forming a ring of needle-like points similar to those figured by Max Schultze as projecting from 1 Max Schultze gives a somewhat similar figure, viz. the sac-like rod of a pike, produced artificially ('Arch. mikr. Anat.,' Bd. iii, pi. xiii, fig. 18«). » ' Arch. mikr. Anat.,' Bd. v, p. 379, pi. xxii, fig, 2a.
STUDIES IN THE RETINA. 447 the membrana litn. externa when the rods and cones are broken away (cf. Hoffmann's figures [Bronn's ' Thierreich ; Amphibien,' pi. xxiii, 11—18, and pi. xxiv, 2—8] with Max Schultze's [' Arch. mikr. Anat.,' Bd. v, pi. xxii]). The latter author traced the threads of his " Faserkorb " proxi- mally into the connective tissue of the outer nuclear layer, but inasmuch as distally they ran on to the outer limbs of the rods, he clearly wished to see in them the ends of the nerves (cf. Strieker's ' Handbuch '). Hoffmann, who figures the basal threads as running only a short way down the inner limbs of the Amphibia, and then only loosely applied, appa- rently regarded them as nothing more than hair-like prolon- gations of the membrana lim. externa. Further, as stated, Max Schultze described a continuation of his " Faserkorb " a short way down the outer limbs. Hoffmann (loc. cit.) also figures somewhat similar threads running on to the outer limbs of amphibian rods; these he could not explain because his basal threads were not supposed to run the whole length of the inner limbs. There seems, then, to have been a distinct tendency to attribute to the inner limbs in the Amphibia a system of longitudinal fibres, though apparently not so pronounced or complete as the " Faserkorb" of the inner limbs of the human rods and cones. We may say, then, that the rods are thought to be longi- tudinally striated, but while the inner limbs are externally striated with fibrils the outer limbs are marked by furrows. My own observations entirely confirm the existence of longi- tudinal striae; but those on the inner limb and those on the outer limb are not distinct in kind from one another, but are parts of one system. Long before I had succeeded in discovering the true rela- tions of these striations to oue another, I had noticed that the markings on the outer limbs consisted far more of longi- tudinal rows of dots than of furrows. The rows, though mostly continuous, are not always strictly parallel; and the dots only occasionally fall into circular series running nearly
448 H; M. BERNARD. evenly round the rod. I find in my notes that at times two series of dots at right angles to one another are recorded as marking the exterior of the rods. The dots were usually slightly drawn out longitudinally. my earlier drawings. It was noticed that the dots appeared almost as if they raised the surface of the rod, and that, hence, between the rows there were slight furrows, but on this point I have never satisfied myself; if any furrowing exists, it must be very slight. While these longitudinal rows of dots on the outer limbs were clear with any well-preserved retinas stained in Bhrlich's hsematoxylin, it was not till I employed the iron-alum haamatoxylin method of staining that I saw any striation of the inner limbs, and then, while that on the outer limbs was very strong and regular, that on the inner limbs was hardly ever regular, often indeed not recognisable as a system of strife at all. Further, I then found, as stated, that the two are not distinct phenomena, but that the fine staining threads which run down in the walls of the inner limbs are continued on to the outer limbs, as Max Schultze observed: but they do not stop short, as he supposed ; on the contrary, they run down the whole way, swelling into small clumps of staining matter at short dis- tances from one another, these clumps being the rows of dots I had seen all along. Fig. 3, a, b,1 are from Fig. 11 shows diagrammatically the arrangements of this system of threads, while figs. 13, a—d} and 29, a, b, e—-J, a,ve from actual preparations of retinas from different Amphibia. Beginning usually faint near the nucleus, and seldom as a dis- tinct system, the arrangement gets more pronounced distally. It may be very pronounced indeed near the ellipsoid (e. g. in the toad, fig. 13, a, b). Here it passes on to the outer limbs, and, where inner and outer limbs are stretched a little apart, may be seen as a nearly regular ring of smooth, thin threads, 1 Cf. Max Schultze, ; Arch. mikr. Anat../ Bd. iii, pi. xiii, fig. 11, where he shows a rod covered with " pigment granules ; " another figure occurs in Bd. v, pi. xxii, fig. 17 a. The dots above referred to are quite distinct from pigment granules, one of which I have drawn in fig. 3, b.
STUDIES IN THE RETINA. 449 not free, like Max Sohultze's needle-like prolongations of his "Faserkorb," but rather as thickenings of the stretched membrane. On these threads clumps of staining matter soon appear (fig. 13, c). In the diagram, fig. 11, the system is drawn very symmetrically from the nucleus outward, but this is not by any means usually the case. The nearest approach to it has been found in the axolotl, preparations of which inspired this diagram. Fig. 29, b, represents more truly the ordinary conditions. We have a gradual formation of the symmetrical system of striae towards the distal ends of the inner limbs (though usually quite irregularly), and when formed it passes on to the outer limbs. There is some evi- dence that this is also what takes place in the human rods and cones, for the "fibrillation" is said to be limited to the outer portions of the inner limbs (cf. ' Quaiu's Anatomy,' 1894, vol. iii, part 3, p. 49, fig. 52, after Schwalbe). Some variation seems to occur in the numbers of the longi- tudinal threads on the outer limbs; they are sometimes very numerous (e. g. newt, fig. 30), at others very sparse ; and this is not only the case in different Amphibia, but in different specimens of the same. Figs. 3 and 6 are from different frogs; in one case the threads are crowded, and in the other quite far apart: the rods in this latter case have been greatly stretched, but one does not see why that should lessen the number of striae. The significance of some of the irregularities of this system of strise1 will be better understood when we have described the connection between these threads and the contents of the vesicles in whose walls they occur. The rods, then, are delicate protoplasmic vesicles, in the thin walls of which staining threads occur. In the walls of the outer limbs these threads are usually more or less beaded with clumps of staining matter. The claim made by Max Schultze and Hoffmann (see the figures and plates re- 1 The spiral twist of the striae oa the outer limbs has been rightly attributed, to torsion. I have only seen it, and then very marked, on rods broken off like that shown in fig. 13, c.
450 H. M. BERNARD. ferred to above) that the outer limbs of the cones are also striated will be discussed later. The Contents of the Rods. — According to Max Schultze the outer limbs of the rods are built up of discs joined together by some cementing substance. This descrip- tion, propounded by so great an observer, seetns to have had the effect of turning away attention from Hensen's figures of cross-sections of rods of the frog,1 which clearly showed some definite internal structure. It must, however, be admitted that Hensen's cross-sections differed among themselves; there were two kinds (see PI. 30, figs. 7, a, b, which reproduce them), and they were not easy to reconcile with one another. Nevertheless I think it cannot be doubted that the discs of Max Schultze, which are, I believe, artificial phenomena, helped to consign them to temporary oblivion. As a matter of fact, Hensen's figures, which were optical sections and hence hazy, come near the truth, and are, as we shall see presently, reconcilable with my own observations. It seems fairly clear, for instance, that his two sections may compare with my own figs. 9, b, 13, g, and 12, 13, h, respectively. Hensen, however, was too anxious to discover nerve-endings, and was therefore prepared to see fibrils in any clear space or small refractive portion of the section. In the case of fig. 7, a, he thought the meshes of the reticulum round the periphery of the sections were fibrils of doubtful significance, but in fig. 7, b, those in the centre were regarded as nerves,— three, he thought, in the centre of each rod. With regard to the contents of the inner limb of the rod, its- most conspicuous element, the. " ellipsoid," has long been known; it has been regarded as the organ in which the nerves end (cf. fig. 23, on the right), and deserves a separate section. This is readily accorded, inasmuch as it admits of being described separately, and what follows will be clearer if we temporarily ignore it. At the same time we shall find it necessary to discuss the contents of both outer and inner limbs together, passing by for the present this particular body. 1 Virchow's ' Arch. path. Anat.,' Bd. xxxix, 1867, pi. xii, figs. 7 and 8.
STUDIES IN THE RETINA. 451 For a clear understanding of the description and figures relating to the contents of the rods to be here given, it is worth while turning once more to their development, and noting that, in essence, they are protoplasmic vesicles ex- truded from the retina. As seen in the first part of this paper, the early stages of these vesicles are seldom found intact, but when they are they usually appear clear, and apparently with only fluid contents. Faint traces of delicate proto- plasmic networks may occasionally be seen (see Part I, PI. 3, fig. 16). Networks are, again, found in well-preserved and propei'ly stained preparations in the large basal vacuoles of the cones (see PI. 31, figs. 23, 27, 28). Later we find dis- tinct networks in the inner limbs of cones and rods, with usually a certain number of very pronounced threads running down in their delicate walls (see above and figs. 29, a, h, i, j) ; so also in the outer limbs—which, as we saw in Part I of this paper, began as fluid vesicles at the tips of the cones— a protoplasmic reticulum ultimately appears. The staining reticulum in the outer limbs is not often found as a simple meshwork, but this is sometimes the case, and we may assume that it first appears as sucb. Two instances are shown in the figures (4, b, and 6). We gather from these cross-sections that the clumps on the longitudinal threads running down the rods are the points of attachment of this internal reti- culum to the walls of the vesicle. As a rule this reticulum is not evenly distributed; we find a tendency for it to be compressed into the axis of the rod, always, however, remaining attached by its threads to the wall fibrils. As this compression increases the threads of the internal axial portion get very thick, coarse, and matted together. The compression may go so far that the reticulum merely consists of an axial strand with a few meshes in it, while the attaching threads are lengthened so as, in cross- section, to look like the spokes of a wheel (see figs. 12 and 13, k, and also cf. Hensen's optical section reproduced in my fig. 7, b). So far, then, the rods are protoplasmic vesicles, each
452 H. M. BERNARD. divided into two compartments by a cross-membrane;1 and as they assume their definitive shapes they become gradually filled with a staining reticulum, which, omitting the ellipsoid, develops especially strongly in the outer and, in the adult Amphibian rod, more important of the compartments. This account seems to justify the description of the rods as prolongations of the " visual cells." It is obvious that each may be regarded as a prolongation of the cytoplasm belong- ing to each rod nucleus, a prolongation at first filled with fluid, but sooner or later containing also the usual reticulum which ramifies through the cytoplasm of ordinary cells. My only objection to this description is to the term " visual cells." My researches long ago compelled me to abandon the usual conception of the retina as composed of cells, and I now regard it as a syncytium, in which the nuclei are arranged in layers, not as fixed morphological units, but solely as centres of physiological acti- vities which may at times require them to migrate outwards, ultimately, if life lasts long enough, to become rod nuclei. The evidence for this is, to my mind, so convincing that I have no hesitation in making the state- ment, even though it stands in such startling contrast to the conclusions of nearly all the most recent workers on the retina, such as Ramon y Cajal, Dogiel, and others, and though a criticism of the method and results of these authors is here out of the question. In the first part of this paper, p. 43, I referred to the migration of nuclei from the middle nuclear layer to the outer nuclear layer, and showed that, even if we could not see evidence of it in our sections, it would be necessary to assume it; and I here add figures of nuclei passing through the outer reticular layer in differeut Am- phibia (figs. 21—23, 25, 26); while, again, in fig. 24 one or perhaps two nuclei have moved outwards together, leaving a space vacant in the middle nuclear layer, and apparently 1 I have not yet been able to ascertain for certain the time of appearance of this membrane. As we-shall see below, it probably appears before the ellipsoid.
STUDIES IN THE RETINA. 453 dragging the cytoplasmic reticuluin after them. figures might be multiplied indefinitely, and, moreover, taken from nearly every retina that is closely enough exa- mined. I reserve fall discussion of this somewhat revolu- tionary conception of the retina as a syncytium for another communication. But in the meantime I feel compelled to state my conviction that the rods are not the prolon- gations of "visual cells," but protrusions of the cytoplasm of the retinal syncytium, each, at least in the Amphibia, dominated by a nucleus. Passing on from this digression, and regarding it for the moment as indifferent how we describe the rods in their relations to the nuclei, the evidence is abundant, as I shall now endeavour to show, that these nuclei are the centres of the physiological activity which gives rise to the rods. In the first place, a great part, if not all of the fluid or hyaline matter, here always spoken of as fluid, which first causes the vesicle to protrude, comes from the associated nuclei. Fig. 17 can hardly admit of any other interpretation than that fluid is extruded by the nuclei into the inner limbs of rods. If it is objected that these figures might as easily be inter- preted as representing phenomena due to the stimulation of fixing agents, this argument will not apply to fig. 23, where we see a " double cone,"1 in which one nucleus is still large and vesicular, while the other is collapsed, becauseits fluid contents have been discharged into the base of the cone belonging to it. Indeed, a study of cones with their basal vacuoles makes it very evident that the fluid of these vacuoles has been derived from their nuclei. Lai'ge vesicular nuclei in the position of cone-nuclei, i. e. well within the membrana limitans externa, are vei'y common and in striking contrast to the more condensed rod-nuclei (figs. 16, a, b, and 18). The same contrasts may also be found in the other nuclear layers, but here, again, it is impossible to give in this paper Such 1 For the correct interpretation of " double cones " in the Amphibia see Part I, p. 33.
454 H. M. BERNARD. an extended account of the observations made relating to this subject. Selecting one more instance, I would refer to fig. 20, in which a large fluid vesicle has been discharged from its associated nucleus, and apparently has not found a way down as a young cone between the adjoining rod-nuclei, or, if part of it has succeeded in doing so, that part did not come into the optical field. Lastly, fig. 19 shows a rod thrust outwards by an increase in size of its basal vacuole. In the second place, the staining reticulum of each rod is also certainly derived from its associated nucleus. Not only can the reticulum of the inner limbs be seen in direct con- nection with the linin network of the nucleus (see figs. 29, a, i,j), but a thick stream can be seen descending from the nucleus on to the ellipsoid (figs. 10,23, 27), a phenomenon to which we shall refer more fully later on. Indeed, if the form of the cone or young rod (figs. 13, d, 15, b, 23, 29, a1) with its nucleus surmounting its narrow basal neck be kept in mind, it is difficult to conceive of any other origin than the nucleus for the large amount of staining material which finds its way outwards into what was certainly originally a fluid vesicle, with, at the most, a few delicate reticular strands. The longitudinal fibrils running down the outer limbs are, in their shape and arrangement, evidence for this outward movement, while the clumps of staining matter along the whole length of their threads, and the density of the reticulum in the interiors of the rods, are witnesses of the immense quantity of this staining matter required. Actual demonstration of the derivation of this reticulum of the outer limb from that of the inner limb, and both from the nuclear reticulum, can be seen in the figures. For in- stance, there occur, in different parts of the inner limb, often in the wall low down and partly apparently embedded in the ellipsoid, deeply staining refractive bodies, usually globular, and, what is more important to note, always sur- rounded by clear zones as if they were the centres of small fluid vacuoles (figs. 15, a, and 29, c—g). These are certainly 1 Many more are figured ia Part I, PI, 3.
STUDIES IN THE- RETINA. 455 chromatin globules, and are usually found in young rapidly growing retinas.1 In well-stained preparations it is common to find that, from these bodies, fine threads run down the walls of the outer limb. In one figure of a developed rod this thread was the only one which took the stain (fig. 29, c). In another figure, two staining and rather straggling threads came from one of these bodies, which had apparently been flattened out against the membranous partition between inner and outer limb (fig. 15, a). To this phenomenon, i.e. this membrane acting as a barrier between inner and outer limb, we shall return. Even where there are no such bright globules of chroma- tin, the derivation of the reticulum of the outer limb from that of the inner can be at once seen if we study the figures of the developing cones shown in fig. 29, e—-j. These figures are merely a selection, and might be multiplied indefinitely. They show quite clearly that the staining material within the outer limb appears where the thin threads from the inner limb come down on its wall. This fact shows that the striation of the outer limbs of the cones figured by Max Schultze and Hoffmann may exist, not as a complete system as they repre- sented it, but as the first beginnings of the subsequent stria- tion of the rods. Unfortunately none of these figures (29) seem to show the true tips of the cones; still, enough is here seen to demon- strate the point we have immediately in hand. Lastly I would refer to fig. 26, which is by no means an uncommon phenomenon. A nucleus is seen passing through the outer reticular layer and about to join the outer nuclear layer (that of the rods and cones). It is preceded by a fluid space, while from it a very delicate reticulum streams out- wards. This I interpret as representing a very early stage in the formation of a rod, being still entirely within the 1 The only other figure I know of which shows such a body is one by Hensen (I. c, fig. 7, c), who, as we have seen, came so near discovering the structure of the rods, having failed apparently for the want of better micro, scopic technique.
456 H. M. BERNARD. retina. The fluid vesicle in the ordinary course of things would, on approaching the mem. lim. externa, form the usual conical protrusion, and into it the staining reticulum would follow. On the other hand, it is only fair to note that streams of very delicate staining reticulum occur elsewhere; one other, for instance, is shown running up from the left- hand rod-nucleus in fig. 27. The explanation of this must be deferred until I can at the same time give the evidence in full on which it rests, and this I hope to be able to do in the near future. Further Contents of the Rod. —So far, then, we have described the origin and structure of the rod as a protoplas- mic protrusion from the retina, containing the usual staining network very strongly developed in the outer limb, and with some clear fluid in the meshes or interstices. This network and this fluid are not, however, the sole contents of the normal rod, aud the striking difference between inner and outer limbs, apart from the difference in sh;ipe and density of the reticulum, is found in the fact that while the former remain protoplasmic vesicles, with apparently soft, flexible walls filled with these elementary constituents which we have described (passing over for the moment the ellipsoid), the outer limbs become filled with some highly refractive substance, which renders them turgid. The change from the loose, long terminal bag found at the tip of the advanced cone (c4) to the outer limb of the rod (rj) (see Part I, PI. 3, fig. 4) is seen to consist not only in the squeezing outwards of the staining matter to the distal end of the inner limb, but also in the filling up of the outer limb. Now while we have traced to its source some of the matter which helps to fill the outer limb, viz. the staining reticulum, this will not account for the refractive contents which now seem to make them turgid and cylindrical. Further, we saw that the outer limbs of the rods lengthened (from ?*! tor3), and hence apparently continued to take in more of this refractive constituent of their contents ; and not only lengthened, but as a rule became also much thicker, I have
STUDIES IN THR RETINA. 457 noticed also that the outer limbs of Schwalbe's rods (i\ and rs) were in most cases rather more deeply stained than the longer, thicker definitive rods, although I lay no great stress on this. The accidents which can never be eliminated from our technical methods are too numerous to allow conclusions to be based upon mere variations in diffuse staining. I mention the point, however, just because it is possible that the proportion of the refractive matter to the staining reticulum might be expected to be less in an outer limb, just beginning to fill up, than in a large swollen rod. It is this refractive matter which gives the rods their characteristic appearances, and which has led to their being classed among1 cutictilar structures. The source of this refractive matter is to be seen in the pigment epithelium into which the tips of the rods are pluuged, and it is largely composed of pigment granules, probably with some portion of the protoplasm of the epithelial cells. At least the absorption of cytoplasm as well as pigment by the rods can actually be shown to take place under special cir- cumstances, as we shall presently see. In the first place, dealing for the moment with general considerations, I again refer to the development of the rod; a fluid vesicle is thrust into the pigment layer, and slowly becomes filled with refractive ma.tter. Both the vesicle and the epithelial cells are, so far as we can see, naked proto- plasm in the very closest contact with one another,—indeed, tightly intei-locked, the pigment cells constantly forcing a passage up between the packed rods.1 Between these some interaction is almost certain to take place. This interaction is, I contend, in part at least an absorption of pigment by the rods. The pigment of the epithelial cells is constantly recruited by an outward streaming1 of granules from the choroidal layer adjacent to it, a streaming which can be seen in every successful preparation. So that we may conclude that pigment is being used up and as constantly replaced. The only other alternatives to this view are either that the refractive matter in the outer limbs of the rods comes from 1 J?or the evidence that the rod layer is normally compact see Part I. TOL. 44, PART 3.—NEW SERIES, Q 0
458 H. M. BERNARD, the retina, or that it is manufactured in situ within the rods. That it does not come from the retina, from which we can easily trace the fluid and the staining network, we gather from the total absence of any refractive matter in the inner limb except in the ellipsoid; and, as we shall presently see, the position of this body forms additional evidence that the sotirce of the refractive matter is from without inwards to- wards the retina, and not from the retina outwards. That the matter is not manufactured in situ we gather from the microscopic appearances, which show very clearly that it is forced in through the walls. The evidence for this is to be seen iu the changes already described, which take place in the character of the reticulmn within the outer limb of the rod. Figs. 4, b, aud 6, b, show this reticulura simply diffused equally across the section; figs. 13, c—h, and 12 show different stages iu its compression towards the axis of the rod. Now it is difficult to explain this compression except on the assumption of some matter passing in through the walls and crushing it inwards, stretching, or perhaps merely lengthening the threads which attach it to the walls. Fig. 13,,/, shows the process as being irregular, while fig. 14 shows that it may take place locally, i. e. along one side of a rod and not on the opposite side. This observation is im- portant, because it is in keeping with the fact that the tongues of the pigment cells run up lengthwise between the rods. Fig. 13, i, shows that at times the reticulum, though compressed towards the axis, may retain some of its concentric threads, the refractive matter passing them by. The refrac- tive layer was here 1'5 J U thick, the whole rod being 9 ft.1 Again, in eyes in which, after exposure to light, the pig- ment has been forced up to the membrana limitans externa, individual granules can be seen remaining behind after the general retreat of the pigment, and sticking to the clear protoplasmic walls of the inner limbs. Many of them can 1 Zenker ('Arch. mikr. Anat.,' iii, 1867, p- 259) discovered that the outer layer of the rod is more highly refractive than the axial portion.
STUDIES TN THE RETINA. 459 then be seen obviously fadiug away, the shape being re- tained, but the bright colour and sharpness of contour have disappeared, and the whole appearance suggests their being slowly absorbed. Although, as above stated, with the excep- tionof the ellipsoid (and the oil globule in thecones of the frog), I have never found refractive matter in the inner limbs in Amphibia, cases occur elsewhere in the animal kingdom in which large inner limbs become filled with it, but in a manner entirely confirmatory of my argument that its source is the pigment epithelium. The clinging of pigment granules to the protoplasmic walls of cones was noted in Part I. Again, in a series of sections of retinas of animals which had been exposed for three hours to the light of an arc lamp,1 the heat rays being screened off as far as possible, one interesting result is conspicuous. The pigment epithelium is here and there disorganised, and isolated pigment cells have forced their way up to various heights among the rods. These can be found in all stages of losing their pigment; some appear as nuclei still thickly enveloped in pigment, others with only a trace of pigment, while here and there nuclei alone persist from which all the pigment and the protoplasm have disappeared. section, selected because of the cross-sections of the rods, such a nucleus, bereft of all its pigment, embedded among rods, and in these latter the reticulum has been compressed into the axis, which, as above suggested, indicates the absorp- tion of extraneous matter through the walls. Other effects of this exposure to such a fierce light have still to be studied. For instance, the contents of the rods have a singularly blotchy appearance, but I cannot satisfy myself whether this lies in the object or in the accidents of staining. While these arguments are, I think, sufficient for the Fig. 12 shows in a tangential 1 I am indebted to my friend Mr. George Newtli, of the Royal College of Science, not only for the use of the necessary apparatus, but also for indis- pensable advice and assistance in making a series of experiments with pure monochromatic light, the results of which are still being worked out.
460 H. M. BKBNAED. present demonstration that the refractive matter within the outer limbs is absorbed by the rods from the pigment, I should like to mention two points on which I am in great uncertainty. It has appeared to me more than once as if the pigment granules could pass bodily into the rods, and, at least for a time, maintain their individuality. I do not see why this should not occasionally happen; indeed, I cannot explain some of the phenomena on any other hypothesis. Still, the evidence shows conclusively that this is not the normal method, but that the pigment granules are absorbed as a colourless or neai'ly colourless refractive and amorphous matter. The occasional finding of retinas in which the colour of this refractive matter within the rod is the same as that of the pigment granules without (I have seen this in sections of the retinas of the pigeon and of frog tadpoles, etc.) may be mentioned, in passing, as additional evidence of the origin of the former from the latter. One appearance suggestive of pigment granules within the rod seen in osmic acid preparations must be familiar to all students of the retina. It is the "disc" formation on which Max Schultze laid so much stress. this to a transverse flaking of the internal reticulum, perhaps a kind of coagulation of the same, as Max Schultze himself suggested. The transverse flakes are usually deeply coloured by osmic acid, and often appear exactly like layers of intruded pigment granules. In preparations not treated with osmic acid the appearance is not to be found. The second point is the relation of the phenomena here detailed to the visual purple. This is said to be produced in the dark through the interaction of the rods and the pigment epithelium, i. e. when the epithelium is only in contact with the tips of the rods, and, further, it is said to be bleached by the light, i. e. when the rods should, according to my own observations, be absorbing clear refractive matter from the epithelial cells, which are then in intimate association with the rods, inasmuch as tongues of the cells then travel up between the rods. I am of course awar that it is frequently I now, however, refer
STUDIES IN THE RETiNA. 46J maintained that fine protoplasmic processes of the pigment cells are permanently advanced as far forward as the meni- brana limitans externa, and are thus always in contact with the rods. Not in any single one of the retinas of some twenty-five vertebrates I have yet examined, and theirnumber must, I think, now amount to fully one hundred, fixed and stained by all the latest methods, and examined with the best available microscopic lenses, have I been able to find a trace of these processes of the epithelial cells permanently interlocking with the rods. On the contrary, when the pig- ment is retracted the contour of the pigment cells is per- fectly straight or rounded as the case may be. Had such processes existed, I am convinced that at least some evidence of their presence would have forced itself on my attention long ago. I have, therefore, so far no point of connection to offer between the physiological details here described and the visual purple, which appears when, according to my own observations, the rods should be getting rid of the matter absorbed when last the light forced the pigment cells into close contact with them, and is bleached when they ought to be absorbing, and at the same time clarifying, the warm colouring matter of the pigment. A reconciliation of these observations will doubtless some day be forthcoming, and there the matter must be left for the present. The Ellipsoid.—This somewhat inappropriate name is usually applied to the body found in the inner limbs of the Amphibia where these limbs abut against the outer limbs. Max Schultze regarded it as a plano-convex lens; the name here adopted was suggested by Krause ("Opticus Ellip- soid "). It is here preferred robbed of its prefix "opticus," so as not necessarily to suggest special functions.1 So far as the terms describe form alone, " plano-convex" is prefer- able to ellipsoid for the Amphibia, for that is the most usual definitive form assumed in the adult rod, i. e. when the rod is not very large and thick, as it is in the axolotl, in 1 Krause thought it was the uerve-end organ (' Anat. Untersuch.,' 1860)
462 H. M. BERNARD. which case the body is usually an irregular flattened disc (fig. 23). As a matter of fact, the body is of very various shapes. Fig. 15 shows a series of cones and rods (salamander) in which only in a young cone is the body egg-shaped, in others it takes the shape of the tip of the swollen inner limb of the cone: if the latter is large, the body is large ; if narrow, the body is narrow, while in the definitive rod it is uniformly plano-convex. It thus seems quite plastic in its earlier (cone) stage, and only assumes a definite form in the full-grown rod. Dealing, then, with this body as we have with the other contents of the rod, we must regard it as an aggregation of these contents which, for some reason or other, rests perma- nently against the transverse membrane separating the inner and outer limbs. It varies greatly in its staining. It is sometimes intensely stained, at others it is comparatively clear and refractive. In this latter case a dense stream of staining matter is very frequently seen descending upon it from the nucleus (see figs. 10, 23, 27). We cannot be far wrong, then, if we refer the variation in the intensity with which the body takes stain to the relative proportions of staining matter and refractive matter which compose it. For out of these two substances, which, as we have seen, together constitute the visible contents of the rods, it must surely consist. Regarding it for the moment in its definitive plano-convex form, it seems to me that we have, both in its shape and in its position, striking confirmation of our conclusion as to the origins of the contents of the rod. On the one hand, we have an outwardly streaming reticulum of staining matter which, so far as we can see, only manages to get further, i. e. into the outer limb by way of the outer walls. There cer- tainly seems to be some condensation of the reticulum against the blind end of the inner limb (see fig. 27, left-hand figure). On the other hand, coming into the rods from the opposite direction, viz. from the pigment epithelium, we have the re-
STUDIES IN THE RETINA. 463 tractive matter. This, as we have seen, is absorbed by the walls of the rods filling them up till they are turgid. This matter would thus find its way inevitably up against the transverse membrane Separating inner from outer limb, and, seeing that it passed through the outer wall into the rod, there is no apparent reason why it should not pass through this transverse membrane from the outer limb into the inner limb. This, then, I believe, is what takes place, the very form of the ellipsoid being suggestive of its having been forced through to form a kind of drop on the proximal side of the transverse membrane. Confirmatory evidence will later be adduced from other retinas, but sufficient to establish the point will be found in what follows. When we come to the ellipsoid in the cones (see figs. 15, c—e) it would seem that the explanation we have given of it in the rod could hardly apply. There appears to be a transverse membrane (fig. 29, /, j), but there is no swollen outer limb filling up with refractive matter. the explanation of the ellipsoid is practically the same, as we can gather from the conditions seen in the frog. In the cones of the frog there is invariably a round refractive globule at the junction of the basal and the conical portion. In well-stained specimens a mass of staining matter is generally seen abutting against this globule, as if they mutually blocked the way for one another. We thus get practically the same condition as in the rod, though in this case we do not know exactly where the transverse membrane is, i. e. whether the refractive globule is on its inner or outer side. This parallel assumes (1) that the refractive globule of the cones of the frog is of the same substance and has the same source as the refractive matter in the rod, and (2) that this refractive globule and the adjacent staining matter will later fuse together to form the definitive ellipsoid. The former of these assumptions is, I think, fully justi- fiable. We have seen how readily pigment granules cling to the thin protoplasmic walls of the cones, and can be seen fading away on the fine membranous walls of the inner limbs Nevertheless
464 H. M. BERNARD. of the rods, as if in the act of being absorbed. Hence it is but natural to assume that some of the refractive matter which later fills these vesicles to overflowing should early find its way into the tips of the cones and be squeezed out by the lateral pressure described in Part I as existing in the rod layer, so as to appear as refractive globules just above the line where the pressure of the rods ceases, i. e. on a line between the junctions of the inner and outer limbs. The secondary thrusting back again of these globules in cones (c8), described in Part I of this paper, needs no comment. Then, again, I mentioned in Part I that in young tadpoles it was possible at times to see these globules actually dis- appearing in the ellipsoids of young rods (see PI. 3, fig. 15), showing clearly that, in this refractive globule of the cone with its adjoining staining matter, we really have the ele- ments of the future ellipsoid, though not blended together. Further, in one of my slides of a young frog tadpole the refractive matter absorbed by the rod is not always dis- coloured; globules of bright reddish-brown matter exactly resembling the pigment in colour occur high up in the rod, near the transverse membrane, while as a complete confirma- tion of the argument, globules of exactly the same colour can here and there be found in the ellipsoids of the same rods. The condition found in the cones of the frog thus helps us to understand the ellipsoid in the cones of the other Am- phibia here dealt with. It has long been known that the refractive globule was absent from the cones of the toad, an absence which was disconcerting to the earlier investigators, who would attribute to it an important dioptric function. It is also absent from the cones of the salamander and the axolotl. In these cases, from our point of view, it is not so much that the refractive matter is absent, but that it never really forms as a distinct globule ; it is mixed with the staining matter to become the ellipsoid as fast as it collects. In the case of the newt, all students will remember that
STUDIES IN THE BETINA. 465 Max Schultze, and others after him,1 described and figured a combination of two "lenses," a biconvex and a plano-convex, as a higher specialisation than the simple plano-convex "lens" (the ellipsoid) of the frog, toad, salamander, etc. Schulfcze even claimed that this lens could be isolated. The body which he figured can be seen frequently enough, but not by any means always in the shape of a biconvex lens. It is nothing but a fluid vacuole, more sharply defined than usual. Fig. 30 shows two rods of a newt side by side; in one there is a well-defined vacuole resting on the ellipsoid, aud in the other a quite uudefined vacuole like that usually found in other Amphibia. The former is interesting because its origin from the nucleus can be seen, a second one appear- ing ready to escape. Most of the nuclei in this preparation have vacuoles about the same size as shown in fig. 30. Further, in very many of the outer limbs of the rods rows of fluid globules of different sizes can be seen. Compare the views as to the origin of the fluid on p. 453. Let us sum up the conclusions so far arrived at, forbearing to enter more fully into the physiological results obtained till the corroborative evidence yielded by the eyes of verte- brates other than Amphibia can be prepared for publi- cation. The rods in the Amphibia are specialised protrusions of the retina, consisting of extremely delicate protoplasmic vesi- cles, each divided by a transverse membrane into an inner and an outer compartment. The staining reticulum which traverses these vesicles is especially developed in the outer- most, into which it finds its way in threads down the walls. These threads, at short distances, give off other threads from small nodes into the interiors of these outer vesicles. These latter further become filled with refractive matter absorbed from the pigment epithelium, and certainly largely obtained from the pigment granules. This matter absorbed through the walls condenses the mass of the reticulum into the axes 1 Cf. 'Arch. mikr. Anat.,' Bd. v, 1869, pi. xxii, fig. 2a. See also'Broim's Thierreich' (Amphibia). Max
466 H. M. BEHNA.RD. of the rods. through the transverse membrane, where it mixes with the staining matter of the inner limb, and forms the body infe- licitously termed the ellipsoid. A portion of this refractive matter exudes (To be continued.) EXPLANATION OF PLATES 30 and 31, Illustrating Part II of Mr. H. M. Bernard's " Studies in the Retina : Rods and Cones in the Frog and in some other Amphibia." PLATE 30. PIG. 1.—The basal portions of two rods (newt) with remains of ellipsoids, and showing the relations of inner and outer limb as two vesicles separated by a thin membrane. PIG. 2.—Part of an adult rod (toad) distorted, and showing its sac-like character, the contents having ruptured down the middle. Cf. Max Schultze's Kg., 'A. M. A.,' iii, pi. xiii, fig. 18rf (Pike). PIG. 3.—a, b. Two surface views of rods (frog, Plemming), showing the dotted appearance of the longitudinal striation. b. With a pigment granule to show relative size of dots. PIG. 4.—a. Rod of young salamander (boiling corrosive sublimate), showing the longitudinal striation straggling irregularly, b. Cross-section of a rod of same eye, showing a simple reticulum attached to the dots in the walls. PIGS. 5 and 6.—Parts of rods and cross-sections from retina of frog (exposed to osmic vapour); in Pig. 6, a, the rod was abnormally stretched. Pic. 7.—a, b. Two cross-sections of rods (frog) according to Hensen. a. " Optical section of a fresh rod." b. " Optical section of a rod fixed with osmic acid." FIG. 8.—A cross-section of the same according to Max Schultze. PIG. S).—Part of a rod (a), with cross-section (b), and a small piece cut off tangentially (c) from retina of axolotl; note the density of reticulum near the tip of the rod. Cf. b with lieusen's section, Pig. 7, a- PIG. 10.—Upper portion of rod of same (Perenyi), partly in optical section, showing part of nucleus and inner limb ; the reticulum of the latter can be seen (1) coming from the nucleus, (2) condensing on what is called the ellip-
STUDIES IN THE RETINA. 467 soid (cf. Figs. 23 and 27), and (3) passing in fine threads on to the outer limb; the striae which continue them are obscured by the internal reticulum. FIG. 11.—A diagram of the same showing the surface arrangement of the threads ; the symmetry is, however, seldom so complete on the inner limb. The threads thicken towards the tip of the rod. FIG. 12.—Tangential section through a group of rods of a salamander after three hours' exposure to intense light. The cross-sections compare with Hensen's section, Fig. 7, b. Among the rods is the nucleus of a dislocated pigment cell, from which all the cytoplasm and pigment have been absorbed. FIG. 13.—From a toad fixed in boiling corrosive sublimate, a, b, show a thick basket-like reticulum ("Faserkorb") of inner limb near the ellipsoid ; c, part of a rod broken away, and showing the longitudinal threads beaded with dots of staining matter; d, a Schwalbe's rod with threads apparently running over the deeply stained ellipsoid and on to the outer limb ; e,f, g, sections of rods cut at different angles; from c toy what appears to be a homogeneous rind is seen; h, a rod 9/* thick, with a "rind" 1*5^i thick; part of the same in t, as interpreted when examined under high magnifica- tion (1500 times; apoch. 2 mm.; N.A. 1'4, comp. oc. 12) ; _;, a part of a cross-section, showing the material of the rind fitting into irregular spaces of the reticulum; k, cross-section of a narrow rod 5 p thick. FIG. 14.—Optical section of a rod from the other eye of the same animal; the "rind" seen only on one side. FIG. 15.—Salamander (boiling corrosive sublimate): a, the upper parts of a rod, showing dark body in or on ellipsoid, flattened like the latter against the transverse membrane (see Fig. 1), and from this body two staining threads descend on the outer wall of the rod; b, ellipsoid, egg-shaped in young cone; c, the same with its shape adapted to its position, and with mass of staining material in contact with it; d, e, Lhe same ; /, g, ellipsoids ill rods and in their definitive shapes, with dense staining matter condensing on them (of. Fig. 10). FIG. 16.—Toad (strong alcohol): a, two condensed rod-nuclei protruding through the mem. lim. externa, and a vesicular cone-nucleus in contact with the outer reticular layer; b, the same, with similar vesicular nuclei, one in the middle, and one in the outer, nuclear layer. FIG. 17.—The same, condensed rod-nuclei with fluid vacuoles in the inner limbs, as if exuded by these nuclei. FIG. 18.—Axolotl (Perenyi); a condensed rod-nucleus with a much Urger vesicular nucleus above it (cf. also Fig. 23). FIG. 19.—The same; a rod forced farther than its neighbours into the pig. ment, apparently by the intrusion of fluid into the inner limb, thereby dis- placing the ellipsoid.
468 H. M. BERNARD. FIG. 20.—Frog (Flemming); a nucleus iu position of cone-nuclei, but con- densed, with a vacuole near it and apparently discharged from it. FIG. 21.—Frog (Flemming); two nuclei passing, with amoeboid changes of shape, through the outer reticular layer. FIG. 22.—Axolotl, showing the similarity of the cone-nuclei to those of middle nuclear layer. PLATE 31. FIG. 23.—The same, showing a nucleus passing through the outer reticular layer; a double cone, one with condensed and the other with vesicular nucleus (cf. Fig. 18, right-hand nuclei). FIG. 24.—The same, showing a spot where one nucleus, or perhaps two nuclei, have passed outwards, and apparently dragged the tissue outwards with them and left a large gap in the middle nuclear layer. FIG. 25.—Salamander (Perenyi), showing three nuclei, two working out- wards through the outer reticular layer. FIG. 26.— Frog (Flemming); a nucleus passing through the outer reticular layer, preceded by a fluid space and an exquisitely fine staining network. FIG. 27.—Axolotl (Perenyi); a large nucleus with a similar fine staining network above it, aud the reticulum condensing at the distal end of its inner limb; on the right is a small condensed rod-nucleus, with the reticulum of the inner limb condensing on the ellipsoid (cf. also Figs. 10 and 23). FIG. 28.—Frog (Flemming), showing upper end of a cone with a reticulum traversing the basal vacuole. FIG. 29.—a—j, a series of figures from the retina of a young salamander (seven weeks old, Perenyi), showing in developing cones and in rods the origin of the staining reticulum of outer limbs from that of the inner limbs, and sometimes from bright round masses of chromatin. The connection of the reticulum with that of the nucleus is shown in a, i,j (cf. c, d, e,f, g, with Fig. 15,«;. FIG. 30.—Two rods from the retina of the newt. The very delicate longi- tudinal striation of the outer limbs is shown, and in connection with the outer- most rim of the coarsely granular ellipsoid. Above the ellipsoid is a fluid vesicle very variously developed, but not infrequently slightly flattened (Max Schultze's " biconvex lens "). Fluid vesicles or vacuoles are also very com- monly seen in the nuclei and in the axes of the rods.
i i i i i i I! I! I! iU'i iU'i Fift.3. Fift.3. iU'i Fift.3. Fia.lO. Fia.lO. Fia.lO. Fur U. Fur U. C d. e • f Fur U. C d. e • f C d. e • f uf. ZZ-. uf. ZZ-. uf. ZZ-. Fiy. 20. Fiy. 20. Fiy. 20.
tart. 3?tmi.^:S'a. '%'.M. MSM.31. Fiq. Z3. Fip.24. Fur. SO. A.Sflnft,LitliTL(mdo' H.M.Beraaii del.