The Descriptive Anatomy of the Brain and Cranial Nerves of Bdellostoma Dombeyi

The investigation of the brain and cranial nerves of Bdellostoma was suggested to me by Dr. Howard Ayers, who kindly furnished me with a great deal of his own material ; and I am glad to have this opportunity of expressing my thanks for the interest with which he has assisted and directed every stage of the work. I also wish to express my thanks to Mr. John G. Koch for his kindness in drawing for me figs. 1, 2, and 3, and in assisting me with fig. 14.

From the standpoint of the comparative anatomist there is interest attaching to a study of the brain of Bdellostoma. The Myxinoid fishes are the most primitive craniates known, and, to judge from the constitution of their central nervous system, they are several degrees lower in the scale than their nearest cousins, the Petromyzonts. For this reason we have in them a simpler pattern, a more primitive or ancestral arrangement of parts than in the higher brains previously studied. Moreover, the embryological work of Price, 1896, and Dean, 1899, shows that this simplicity is primitive, and not degenerative, as has sometimes been stated. The older naturalists, deceived by the curious jaw formation of the Myxinoids, ranked them as a side branch, decidedly separated from other vertebrate forms, but Dr. Ayers and Mr. C. M. Jackson, in their paper published in 1900, solved this puzzle, proving the homology of the Myxinoid jaws with those of other gnathostoines, and so placing these fishes in the main line of development of the gnathostomes. For all these reasons a thorough working out of the Myxinoid brain promised to throw much light upon that most interesting of problems, the origin and development of the vertebrate brain.

The present paper is a general survey of the ground covered by my studies. The following detailed account of the gross anatomy of the brain and the distribution of the cranial nerves prepares the way for a detailed description of the finer anatomy of the several parts of the brain of Bdellostoma. The work includes the dissection of Bdello-stoma heads preserved in formalin, and the microscopic study of sections both of the brain and of the entire head stained in haematoxylin or in alum carmine. A discussion of the methods used is reserved for a later paper.

The fishes, from which the brains studied were taken, measured from 62 to 64 cm. from the tip of the nose to the end of the tail.

The brain and spinal cord of Bdellostoma lie in a tough but flexible capsule of dense fibrous connective tissue. The fibres are very thick and tough, some of them being 6·7 μ in diameter, and may be either straight or wavy. The capsule varies very much in thickness in different parts of its walls, these walls being much thinner at the cephalic than at the caudal end.

‘The general shape of the cranium is that of a flattened cylindrical flask, with its closed base at the anterior end, and with its open mouth in the form of a tube passing into the sheath of the spinal cord behind. At the cephalic face, in the centre of the area perforated by the olfactory fibres, the wall is ·192 mm. thick. In the median dorsal line it is ·168 mm. thick over the fore part of the olfactory lobes, and increases gradually in thickness as it proceeds caudad, measuring ·296 mm. over the habenular ganglia, and ·5 mm. over the cerebellum and fore part of the medulla. It then decreases gradually until, passing into the capsule of the spinal cord, it is only ·364 mm. thick. On the ventral surface the thickness is distributed a little differently. The cranium is ·167 mm. thick under the anterior end of the olfactory lobes, and ·224 mm. under the cephalic border of the cerebrum. As it passes under the ’tween brain, however, it becomes thinner, being only T68 mm. thick immediately anterior to the hypophysis. Caudad to the infundibular process it becomes thicker again, measuring here ·28 mm., and under the central part of the medulla it measured ·296 mm. The capsule becomes very much thicker as it approaches the notochord, and finally receives that structure within itself, surrounding it with a thin sheath, cf. Ayers and Jackson, 1900. The lateral walls of the cranium measure ·168 mm. in thickness by the side of the olfactory lobes, and then thin out to only ·085 mm. beside the fore brain. At the side of the mid brain, just anterior to the foramen of exit of the trigeminus nerve, it is ‘207 mm. thick, and just cephalad of the ear it measures ·267 mm. At the side of the spinal cord the thickness is ·28 mm.

The density of the cranial structure varies inversely with the thickness. Where the wall is thinnest the fibres are straight and very closely packed. As it increases in thickness they separate more and more, some of them becoming wavy ; and where it is thickest, as in the dorsal surface over the cerebellum and medulla, and just in front of the notochord, the fibres are comparatively loosely packed, and almost all of them are more or less wavy. At the hind end of the medulla, ventral to it, and just anterior to the notochord, lying between the layers of the fibrous capsule, is a thick plate of cartilage, connecting the cartilaginous ear capsules. This is the only cartilage found in the cranium.

The cranial cavity is 11 mm. long and 3 mm. deep. The width varies, measuring 6 mm. across in the region of the olfactory lobes, 5 mm. across at the foramen of the trigeminus, and mm. at the foramen of the vagus. The lateral and anterior walls of the cranium are pierced with numerous openings for the exit and entrance of the cranial nerves and the blood vessels that supply the brain. Owing probably to the fibrous composition of the cranium, the nerve foramina are not definite openings in the wall, but rather places where the fibres have been pushed apart, and the meshes between them consequently enlarged. Through these meshes pass the bundles of fibres composing the nerve trunks as threads through a sieve (fig. 10). The inner surface of the capsule is lined throughout with a layer of endothelial cells.

The Cranial Membranes.—The brain of Bdellostoma has but one enveloping membrane—the dura mater. This is composed, as usual, of areolar tissue, and consists of two layers—an outer one lying close to the inner surface of the cranial capsule, but separated from its lining membrane by the extra-dural lymph space, and an inner one following the contour of the brain, leaving, however, a small subdural space between it and the brain (fig. 7). Both ventrad and dorsad of the ’tween-brain, and ventrad of the anterior end of the olfactory lobes, one or both of these layers delaminate into several finer ones. The space between the layers is filled with coagulated lymph, which is infiltrated with connective-tissue cells, and shot through with rather fine wavy fibres, some running singly, others in small bundles, connecting the two layers. These fibres are comparatively few in number around the fore part of the brain, but increase greatly dorsad of the cerebellum, and dorsad and ventrad of the medulla.

There are also large blood sinuses in this intra dural space. The fibres of both inner and outer layers, throughout the greater part of the dura, are rather larger as a whole than those connecting the layers, and hold in their meshes the usual connective-tissue cells. Toward the caudal end of the medulla, at the level of the anterior end of the plate of cartilage described above, the entire composition of the dura changes. The finer fibres and most of the small cells disappear, and are replaced by heavy, coarse, stiffly-curled fibres, packed closely in the two layers, and fairly closely in the intra-dural space, particularly on the ventral side. In the dorsal intradural space, they are farther apart, and enclose in their meshes numbers of large ovoid cells, not found elsewhere. These cells have a fine protoplasmic network, staining lightly in haematoxylin and alum carmine, and a small, deeper-staining nucleus containing a very distinct nucleolus, and several granules.

General Description.—The brain of Bdellostoma is about 10 mm. long. It is from 5 to 7 mm. wide through the fore end, and from 4 to 4·5 mm. wide through the hind end. The following table gives the external measurements of four brains. Of these, brain A is that of B. Dombey i, and B, C, and D belong to B. Forsteri.

The brain is bilaterally symmetrical in every respect, and when viewed dorsally appears to be divided into main parts —a fore part, consisting of four sets of paired lobes, with a smaller median lobe, situated between the second and third pairs ; and a hind part, an unpaired, shield-shaped division, with a V-shaped cleft cut into its cephalo-dorsal surface, in which cleft the posterior pair of lobes belonging to the fore part rest (fig. 1). The four sets of paired lobes proceeding cephalo-caudad are the olfactory lobes, fore brain, mid brain, and cerebellum, and the small median lobe lying between them consists of the habenular ganglia. The large, unpaired hind portion is the medulla.

In a view of the ventral surface these same divisions appear, but are of somewhat different form (fig. 2). The olfactory lobes alone appear in full size here. Slightly caudad of them the ‘tween brain is to be seen, forming the most ventral prolongation of the brain. This important lobe extends caudad as far as the hind end of the mid brain, and laterad about half way from the median line to the lateral face of the brain, thus hiding from view the mesial portions of fore brain and mid brain. Close behind the caudal border of the ‘tween brain, lies the cephalic border of the medulla. The medulla is ventral to almost the whole of the cerebellum, only the extreme cephalo-lateral parts of that section showing above it (fig. 2).

The Olfactory Lobes.—The two olfactory lobes are, next to the medulla, the largest divisions of the brain, and thus give a line at once as to the relative importance of the several sections of the brain in the physiology of that organ. The lobes are of equal size and symmetrical, and almost entirely separated from each other, being united only at their ventro-mesial angles. Their mesial faces are flat and are apposed to each other, with a sufficient space between for a double fold of the inner layer of the dura to penetrate (fig. 7). Dorsally, the two lobes, taken together, are somewhat convex, the lateral faces rounding off from the dorsal to the ventral surface. Ventrally they are flat across the caudal end, and slightly concave across the cephalic end, their greatest thickness, dorso-ventrally, being not near the mesial face, but about half-way between that and the lateral edge. Examined superficially each lobe is seen to have three divisions—an inner, middle, and outer or lateral one (figs. 1, 2). On the dorsal surface the inner and middle divisions are nearly equal in width, and cover the crown of the lobe, the outer border of the middle division being about where the arch begins to fall away, so that the lateral division does not extend as far dorsally as the other two. On the ventral surface they are more plainly marked than on the dorsal, and it is the inner lobe that does not extend as far ventral as the other two, thus making the ventral surface concave. These three divisions have definite relations to the ending of the olfactory nerves.

The olfactory lobes, viewed in section, are highly vascular, and are seen to be very loose in structure. The cephalic half of each lobe is composed mainly of the olfactory glomeruli. On the ventral and lateral surfaces this glomerular layer extends almost as far caudad as the end of the lobes; on the dorsal surface it does not spread as far. It is the divisions of this layer that are indicated on the surface of the lobes, and the absence of the layer from the inner ventral division that causes the concavity of the ventral surface. The glomeruli themselves are quite large, many of them measuring about 80 μ × 120 μ. They are packed close together, and lie with their axes arranged more or less longitudinally. Aside from the glomerular layer, the olfactory lobes consist of a loose mass of cells and fibres. The cells are, roughly speaking, of two kinds :—comparatively large multipolar cells, staining lightly in hæmatoxylin, and having large, round or ovoid nuclei, and smaller cells, spindle or multipolar, that stain an intense blue black. They are not arranged in any such definite way as the cells in other parts of the brain. The fibres are the secondary olfactory fibres. They come from all parts of the lobe, gathering into tracts as they leave it. The two chief tracts are the tract us olfacto-habenularis, and the tractus olfacto-cerebellaris. The tractus olfacto-habenularis leaves the olfactory lobe at the median dorsal angle and goes directly to the habenular ganglion, without any way station in the fore brain as is usual in the higher forms. The tractus olfacto-cerebellaris leaves each lobe at about the middle of the caudo-ventral border, and courses back, through the floor of the ‘tween brain, ascending finally to the roof of the cerebellum.

The fibres of the olfactory nerves enter each olfactory lobe in the dorsal and lateral parts of its cephalic face. Owing to the fact that the nasal organ lies immediately adjacent to the cranium, and covers its anterior end wall, its nerves have neither the need nor the opportunity of uniting into one great nerve-trunk, as in other forms, but pass through the cranium and enter the olfactory lobes in the same bundles in which they leave the nasal organ.

The Fore Brain.—The fore brain, as mentioned above, consists of two lobes, a right and a left, equal and symmetrical. They are slightly smaller than the olfactory lobes, narrower across the caudal than across the cephalic face, and are separated cephalo-dorsad by a deep cleft, the contination of the cleft between the olfactory lobes. In the median caudal part they fall away from each other, admitting between them the dorsal part of the ‘tween brain, the habenular ganglia. The dorsal cleft between them extends ventrad as far as the base of the habenular ganglia. Ventral to this they are united. There are no superficial divisions of the lobes of the fore brain as there are of the olfactory lobes.

On examining sections of the fore brain with the microscope the lobes are seen to consist of two parts, an outer shell, which is, however, from one to two thirds the thickness of the lobe, and extends over all but the ventro-caudal part, and a small inner core (6g. 8). The cells are of the same two kinds occurring in the olfactory lobes, but they are very differently arranged in the two parts. Those of the core are scattered through it without definite order. The shell, on the contrary, shows a very marked arrangement of parts (fig. 8). First of all comes an outer layer, about ‘22 mm. thick, composed, probably, mainly of glia fibres and the dendrites of cells, and having cells scattered sparsely through it. Next comes the main cell layer, about the same thickness, for the most part, as the outer layer, and consisting of cells, crowded close together, with their long axes radially arranged. Inside of this is a third layer, similar to the first, and about the same thickness, only probably without so many glia fibres, and within this a thinner secondary cell layer, similar to the primary one. The space between the secondary cell layer and the core is largely filled with fibre tracts. Of these, the most conspicuous one leaves the dorsal part of the primary cell layer, and runs caudo-ventrad, laterad of the core. As it nears the ventral surface it turns mesad, crossing to the other side in the post-optic decussation, and distributes itself in the floor of the ‘tween brain.

In the adult the lobes of the fore brain are solid, the ventricles having been obliterated by the encroachment of their walls on the ventricular cavities. In an embryo some little time before hatching the ventricles are very large, lying within the core, which is here much larger in proportion to the rest of the fore brain than in the adult. In this stage the shell lies cephalad and laterad of the core, but not dorsal to it, the dorsal wall over the ventricle being very thin. It will be necessary for the later stages in the development of the brain of Bdellostoma to be more thoroughly worked out before it will be possible to homologise the parts of the adult fore brain with those found in higher forms.

The ‘Tween Brain.—This important section of the brain is wedge-shaped, its broad flat base covering almost the whole under surface of the fore part of the brain, and its sides sloping inward as they rise, until they meet in a ridge in the habenulai’ ganglia. The wedge is bilaterally symmetrical, and also leans somewhat cephalad, so that a sagittal median section is somewhat trapezoidal in shape, ▱. The habenular ganglia are comparatively large, appearing externally as a separate division of the brain (see above). They lie, as before mentioned, close to the median line,—about midway between the lobes of the mid brain and those of the fore brain (fig. 1), but do not project as high as either of these; consequently they cannot be seen in a side view of the brain. They form the most anterior portion of the ‘tween brain, projecting decidedly cephalad of the rest of this segment of the brain. In the specimen measured the right ganglion was 1·2 mm. long, ·5 mm. wide, and ·93 mm. deep; the left ganglion was ·78 mm. long, ·39 mm. wide, and ·74 mm. deep, and was so placed that it extended ·27 mm. cephalad of the right ganglion. Bdellostoma possesses no trace of an epiphysis.

The base of the ‘tween brain extends nearly 1 mm. further ventrad than the surrounding parts of the brain, is shield-shaped in appearance, and slightly convex (fig. 2). It is marked by five small rounded hillocks. Through the first two of these, lying one on either side of the median line and about one third of the way back from the anterior edge of the ‘tween brain, the optic nerves enter the brain. About half way between these and the hind end of the ‘tween brain, situated in the median line, is the stalk of the infundibular process, and immediately behind it, one on each side, are two smaller projections, whose significance I have not been able to discover. The infundibular process is a small round body, embedded in the brain capsule, and very apt to be torn away in dissecting out the brain. It will be described later.

The habenular ganglia.—These ganglia, though asymmetrical in form, are alike in structure. Their cells are smaller than those of any other part of the brain, but are similar to the two kinds already described. They are pressed close together throughout the ganglia, without definite grouping, and with their long axes running in nearly every direction. There are a good many more of the light staining cells in the ventral parts of the ganglia than in the dorsal. In addition to the cells there are great strands of fibres, the beginnings and ends respectively of the great fibre tracts that centre here. Most of these fibres belong to the tractus olfacto-habenularis. The fibres of this tract come from all over the dorsal half of the olfactory lobes, and gather into thick strands—one for each lobe,—at the caudo-mesial angle of the lobes. Running in this way through the cephalomesial parts of the fore brain lobes, they enter the habenular ganglia and cross in their ventral part, just caudad of the fore end of the right ganglion, in angles of from 45° to 90°, thus forming the habenular commissure. Across the caudal part of the ganglia run the fine fibres of the posterior commissure.

The tracts of the habenular ganglia, in addition to the tractus olfacto-habenularis, are the tractus habenulo-tectalis, and the bundles of Meynert. The fibres of the tractus habenulo-tectalis arise in the dorsal part of the ganglia. Some of them cross immediately to the opposite ganglion, others remain with the tract leaving the ganglion of their own side. The tracts are unsymmetrically placed in the ganglia, but are symmetrically placed in regard to the rest of the brain, the left tract leaving the left ganglion just before its caudal end, and the right tract being directly opposite to it. The tracts run ventrally in a slight caudo-laterad direction for about mm., then turn sharply laterad, though still keeping a slight caudal direction, and penetrate among the cells of the mid brain tectum, at about the centre of its lateral surface. The bundles of Meynert connect the habenular ganglia with the medulla. The bundles are symmetrically placed, but unequally developed, the right bundle measuring about ‘28 mm. by 485 mm., the left 496 mm. by ‘12 mm. Their fibres arise mainly in the dorsal part of the ganglia, and gather in one common median strand at the cephalic end of the right ganglion. They run ventrad until they reach the ventral level of the ganglia, then, turning sharply, they run caudad, still as one common strand, in the median line, directly ventral to the ganglia. While on this course the strand is increased by fibres coming from the more ventral part of the right ganglion. At the posterior end of the right ganglion the bundles separate and turn caudo-ventrad, running to the base of the mid brain. Here they decussate for the first time, and pass on into the medulla, decussating again at the level of the entrance into the medulla of the fibres of ramus ophthalmicus V, the decussation stopping about 42 mm. cephalad of the exit of the motor trigeminus fibres from the medulla.

The Thalamus.—The thalamus of Bdellostoma possesses no features of marked interest from the standpoint of gross anatomy.

The Hypothalamus.—The base of the ‘tween brain is rather a passage way for tracts than a place of origin and ending. The principal tract ending there is the one described above coming from the fore brain, and the crossing of whose fibres forms the great post-optic decussation. This tract and the tractus olfacto-cerebellaris occupy most of the floor of the ‘tween brain.

Anterior to the optic nerves there are, in the base of the ‘tween brain, two small groups of cells, one on each side. These groups consist both of the dark and of the light cells, similar to those of the habenular ganglia, only larger. The cells in the caudal end of the ‘tween brain floor, also of the same two kinds already described, are arranged somewhat in longitudinal strands, their long axes running athwart the brain. Bundles of fibres of the fore brain tracts run between these strands of cells.

The optic nerve enters the ‘tween brain, as has already been mentioned, through a protuberance on its ventral surface. It is a small nerve, with fine, straight, deeply staining fibres. Upon entering the ‘tween brain its fibres take a medio-dorsal direction, at an angle of about 45° to the floor of the brain. They are easily followed as far as the chiasma, which lies at the ventro-cephalic angle of the postoptic decussation. So closely applied is the chiasma to the decussation, that it is only by the different direction of the fibres,—those of the decussation running parallel to the base of the brain, and at right angles to its long axis,—that it can be distinguished from it. After the crossing the fibres no longer remain in the compact bundle in which they entered the brain, but separate into a loose strand, and are directed caudo-dorso-laterad through the post-optic decussation. Beyond this they separate still more from each other, running singly or in small groups, and consequently are quickly lost among the numerous other fibres of this part of the brain.

The Infundibular Process, Hypophysis, and Saccus Vasculosus. The infundibular process is, as has already been mentioned, attached to a stalk in the median line, situated about two thirds of the way from the cephalic to the caudal end of the base of the ‘tween brain. The process itself is about 1·35 mm. long, 1·12 mm. wide, where the stalk enters it, and ·22 mm. thick. Its cavity is continuous with that of the infundibulum, and is lined with ependyma cells. These cells are of the cylindrical variety, and have long, stout processes, generally continuous with the long axis of the cell, but which may be deflected slightly from it. The cavity is lined with a single layer of the cells, set very close together, their long axes at right angles to the plane of the floor, and their processes coming from the end farthest from the cavity and penetrating through the wall to its other side. Under a magnification of forty diameters a section of the floor resembles very closely a section of columnar epithelium. The roof of the cavity is from two to three times as thick as the floor, and is composed of two layers, (1) an inner lining layer, similar to the floor, except that the cells are not placed so close together, and, in consequence, some lie over on their sides, and the columnar effect is not so marked; and (2) an outer (dorsal) layer, containing a great many ependyma cells lying separate from each other, with their long axes in all planes. This dorsal layer also contains occasional large round cells, with round nuclei. Near the caudal end of the roof there is a small evagination, directed forward (fig. 13) ; the dorsal layer just referred to appears to be a prolongation of this evagination. In places there is a space between these two layers, in other places they are held firmly together by transfixing processes of ependyma cells. Immediately caudal to the stalk the two layers are fused together and merged completely into one. The canal of the stalk, instead of being placed in the centre, lies just behind the anterior surface, so that the cephalic wall of the stalk is the continuation of the thin floor of the process, and the caudal wall the continuation of the thick roof. The infundibular process lies embedded among the thick fibres of the membranous cranium, and directly beneath it, though separated from it by a layer of fibrous tissue, is the hypophysis. This consists of a much coiled tube, probably glandular (Retzius, 1895), but I have not been able to find any outlet for it, nor to determine the cell structure. The cells themselves have not stained at all, or only with the faintest tinge, in my preparations, so that I cannot tell their shape or size. The nuclei, however, which are fairly large, have stained intensely black or deep blue, so that the hypophysis in section looks like a collection of more or less rounded bodies filled with black granules (fig. 13).

A saccus vasculosas does not exist in Bdellostoma, but the elements out of which it develops in higher forms are there, and Bdellostoma consequently marks an interesting stage in its evolution. The foundation of the nervous element lies in the thickening of the dorsal wall of the infundibular process and of the caudal side of its stalk, already described (fig. 13). From such a thickening the saccus arises in higher forms (Bdinger and Hall, 1899), and the slight evagination photographed in fig. 13 is probably its rudiment. The vascular element is also found. The vertebralis impar artery that runs cephalad ventral to the notochord divides at the level at which the fibres of the vagus leave the brain into the right and left intra-cranial arteries. These arteries penetrate into the membranous cranium, and run cephalad in its ventral wall (figs. 10 and 11). At the level of exit of the motor fibres of the trigeminus each of these arteries gives off a large branch that divides immediately into three. The first of these branches, entering the dura, is directed backward, ventral to the medulla, and supplies the caudal end of that section of the brain. The second runs dorsad, entering the medulla at once, and running on into the cerebellum. The third runs forward, first in the membranous capsule, then in the dura, until the anterior end of the medulla is reached. It then turns sharply mediad, and forms, with its fellow of the opposite side, a vascular plexus that lies in the cleft between the anterior end of the medulla and the posterior end of the ’tween brain (fig. 13). This plexus sends arteries into the medulla, mid brain, and ‘tween brain. Anterior to the foramen of the trigeminus the intra-cranial arteries give off a large branch that leaves the cranium, but they themselves continue forward in the cranial wall, supplying the fore part of the brain. Referring again to fig. 13, and noting again the evagination and thickening of the dorsal wall of the infundibular process, it is very easy to imagine an increase of this evagination that would eventually find its way into the cranial cavity through the foramen of the stalk. If it continued to grow it would be forced candad after entering the cranium, the stalk blocking its passage cephalad, and would come into connection with, and finally envelop, the vascular plexus. This would bring the saccus to the position it actually does occupy in higher forms.

The Ventricles of the ‘Tween Brain.—These are of the usual type in Bdellostoma, but differ somewhat from those of higher forms in some of the details. The iter, running directly ventral to the habenular ganglia, is always small, its diameter varying in different specimens. In some the lumen, though contracted, can be traced unbroken into the third ventricle, in others the lumen is entirely closed shortly after the entrance of the iter into the ‘tween brain, if not, indeed, before, but its course may still be traced by an irregular column of ependyma cells. In other brains still even the ependyma cells are scattered or absent. The third ventricle, though small, is always open; it is situated a little caudad of the anterior end of the right habenular ganglion, and is bounded as usual by the lamina terminalis. It is very much smaller in the adult than it is in the embryo. The infundibulum arises from what appears in the adult to be the cephalic end of the iter, but what the embryo shows to be in reality the posterior part of the third ventricle. In the adult the dorsal part of the infundibulum is always completely closed, and its presence indicated only by ependyma cells. In some specimens these cells are arranged in separated groups, and are hard to find; in others they are grouped compactly together, and their line may be followed with comparative ease. A little more than half-way between the third ventricle and the ventral surface of the ‘tween brain, the infundibulum takes definite form. It is a long, narrow slit, lying in the median plane, showing longer in horizontal sections through its ventral end, than in those through its dorsal end, and directed caudo-ventrad (figs. 5, 13). In sagittal sections it is seen to be somewhat boot-shaped. The lumen becomes wider as the ventral surface of the ‘tween brain is approached. At the anterior end of the most ventral level of this slit-shaped lumen is given off a small, unpaired, median recess directed forward. The hind end of the infundibular cavity is inserted into the middle point of a semi circular canal (fig. 5). This canal lies in the horizontal plane, its ends directed cephalad. A longitudinal section of the brain running through this canal and the infundibulum makes the two together look like the head of an anchor. There is a bulb-like dilatation of each end of the semi-circular canal, these dilatations lying in the two hillocks seen on the ventral surface of the ’tween brain just candad of the infundibular stalk. This same relation of parts is found in Myxine (Retzius, 1895), but Johnston (1902) makes no mention of it in describing the infundibulum of Petromyzon.

The Mid Brain.—The mid brain consists, dorsally, of two equal and symmetrical lobes, divided by a median longitudinal cleft that becomes shallower as it proceeds candad. Anteriorly the cleft widens out, the mid brain lobes separate from each other on the dorsal surface, and admit between them the hind part of the habenular ganglia (fig. 1). The lobes are about the same in length as those of the fore brain, but are narrower behind (fig. 1). Ventrally they taper to a comparatively small point that can be seen in section (fig. 6), but is obscured superficially by the medulla and ‘tween brain.

Structurally the mid brain is somewhat similar to the fore brain, and merges into it gradually, so that in examining a series of cross-sections it is difficult to tell definitely just where the mid brain begins and the fore brain ends. Like the fore brain it may be divided into two parts—an outer shell, the tectum opticum, and a central mass. The relative size of the two parts, however, is more nearly equal (fig. 9), nor are they as sharply marked out one from the other. The tectum has but two layers, as against the four in the outer division of the fore brain, an outer glia layer, and an inner layer of cells. These layers are in every respect similar to the corresponding layers of the fore brain. The central mass consists of cells densely grouped together, except where separated by the fibre-tracts passing through this part of the brain.

The mid brain, though not of the same importance in Bdellostoma as in higher forms possessing better developed eyes, and consequently larger optic nerves, is still an important section of the brain, and its fibre-tracts are more conspicuous than those of any other section except the medulla. The most striking of these is the dorsal decussation, whose fibres connect not only the roofs of the two lobes, but their sides and floors as well. Unlike Petromyzon (Johnston, 1902) the decussation does not extend through the whole mid brain roof, but is confined to its caudal part. Numerous small bundles of fine fibres leave the cell layer of the tectum and proceed towards the central mass, but they are so very fine that it is impossible to trace them in ordinary haematoxylin sections, or to separate from among them the fibres of the optic nerve. Conspicuous in the base of the mid brain, filling entirely its basal cone (fig. 6), are the bundles of Meynert, already described, on their way to the medulla; and, dorsal to their decussation, the ansulate commissure, comparatively small in Bdellostoma. This commissure is caudad of the bundles themselves, and lies so close to them that a small bundle of fibres of the right bundle is deflected by it and pierces down through the commissure instead of passing in front of it, rejoining its fellows later on. Leaving the central mass on each side is the tractus tecto-bulbaris et spinalis, running to the medulla and spinal cord. These, with the tractus habenulo-tectalis, already described, are the principal tracts found in the mid brain of Bdellostoma.

The iter enters the dorsal part of the mid brain as a straight open tube, whose lumen is about ^-09 mm. in diameter. But before the habenular ganglia are reached the tendency to close asserts itself, the ependyma cells crowd in, the tube narrows, and the lumen becomes very small. In some specimens the lumen, in the fore part of the mid brain, is, in places, entirely closed.

The Cerebellum.—The cerebellum is the smallest division of the brain of Bdellostoma, though very much larger and more conspicuous than the same section in the brain of Petromyzon. It consists, in Bdellostoma, of two small lobes, equal and symmetrical (fig. 1). Its roof is continuous with that of the mid brain, although superficially a slight depression marks the boundary between the two sections. The two lobes are divided in the dorso-median line by a continuation of the dorsal cleft that divides the mid brain lobes. The cerebellum rests solidly upon the dorsal surface of the forward end of the medulla (figs. 1, 6), covering the anterior end of the fourth ventricle, and, in consequence, has no peduncles. In structure it consists of an outer glia layer, similar to and a continuation of the outer glia layer of the mid brain, and an inner mass of cells. In the base of the cerebellum, on each side of the median line, lying partly in the cerebellum, and partly in the mid brain, is a group of from ten to fifteen multipolar giant-cells. Most of these cells send their axones back into the medulla, though whether all do or not I am not yet prepared to say. These cells are apparently like the Mauthner cells of the medulla. Lying close under the dorsal surface of the cerebellum is a decussation of fibres connecting its two lobes, a continuation backward of the dorsal decussation of the mid brain.

The cerebellum has also two other very prominent sets of fibres. Coming up from the floor of the ‘tween brain, running dorso-caudad in the lateral walls of the cerebellum to end among the cells of its roof, are the fibres of the great tractus offacto-cerebellaris, described above. Mesial to these, lying between them and the giant cells, are the fibres of another tract, equally important, the tractus cerebello-spinalis. These fibres, arising in the roof of the cerebellum, run ventro-caudad, and gather into two strong bundles, one on each side, at the caudal end of the cerebellum, just lateral to the groups of giant cells. On leaving the cerebellum, these bundles turn sharply latero-caudad. When the brain is dissected out, and its fore part is deflected downward, away from the medulla, as in fig. 4, these tracts are seen for a little distance in relief upon the dorsal surface of the medulla (fig. 4, tr. c. s.). Not for long, however, for they quickly plunge beneath the surface, penetrating between the fibres of the sensory trigeminus, still running caudo-laterad, until they near the sides of the medulla. Here they turn again, and run caudad on their way to the spinal cord. The tractus tecto-bulbaris et spinalis, after leaving the mid brain, runs through the floor of the cerebellum, where it is fused with the roof of the medulla. These tracts, after leaving the cerebellum, are also raised above the surrounding dorsal surface of the medulla, and, when the fore, part of the brain is deflected, may be seen in relief as two longitudinal strands, one on either side of the median line (fig. 4, tr. t. b. et s.). With the cerebellum in place, both these tracts and the tractus cerebello-spinalis are hidden.

The cerebellar ventricle leaves the fourth ventricle in the median line at its anterior dorsal angle, just before the beginning of the iter. The ventricle runs dorsad and dilates into an elliptical cavity that lies immediately ventrad of the glia layer of the cerebellum, and has its distal end projecting caudad (fig. 5).

The Medulla.—The medulla is the most intricate and complex of all the divisions of the brain of Bdellostoma. Although conforming in the main to the usual vertebrate type, it is, nevertheless, owing to causes as yet undiscovered, in some ways decidedly different, and is well worth careful study. It is the largest single division of the Bdellostoma brain, and is, in fact, very nearly as large as the cerebellum, mid brain, ’tween brain, and fore brain put together (figs. 1—3). It is bilaterally symmetrical, a slight depression marking its median line on both dorsal and ventral surfaces. It is about as broad as the two lobes of the mid brain, and about as long as mid brain and fore brain combined—for its actual measurements see the table on pages 142-3. On its ventral surface it is shield shaped, the top of the shield resting against the ‘tween brain, the point at the base cut off by the spinal cord. At the upper angles of the shield the trigeminus leaves the medulla, at the lower angles the vagus. The facialis, acusticus, except acusticus a, and glosso-pharyngeus leave along the sides. Directly caudad of the apex there is a semi-circular band, whose cut ends point forward, sculptured in slight relief (fig. 2). This ring consists of the two bundles of Meynert and their second decussation. A longitudinal depression, not as clean cut as the median one, divides each side of the ventral surface of the medulla into approximately equal parts, so that the ventral surface consists of four longitudinal divisions—two lateral and two median. In the caudal end of the median divisions are continued the ventral fibre-tracts of the spinal cord; in the lateral divisions are lodged the lateral motor columns of the medulla..

Turning now to the dorsal surface, we see that the raised dorsal part of the spinal cord expands into a large Y, whose thick arms compose almost the entire dorsal surface of the medulla, not separating until more than half-way to their anterior end, where they admit between them the lobes of the cerebellum (fig. 1). Three distinct divisions are sculptured on the arms of the Y :—(1) The great strands that run from the lateral part of the raised portion of the cord and form the caudo-lateral part of the arms. These strands are the ascending fibres of the sensory trigeminus; (2) two small bands running from the median part of the dorsal swelling of the cord along the median faces of the arms. These are the fasciculus communis of each side, and are visible on the dorsal surface from the caudal end of the medulla to the fork of the Y (fig. 1,f. c.). They are much thicker at the caudal end than at the cephalic ; (3) the third division lies between the communis and the sensory trigeminus, and consists of the tuberculum acusticum, the “dorso-lateral strands “of Goronowitsch and some other writers. After the arms of the Y have separated, the acustic bundles run along their inner border to the cephalic end of the medulla (fig. 1, ac.). Entering the acusticum, T2 mm. caudad of the hind end of the cerebellum, is the acusticolateral nerve acusticus b—in young forms, where the cerebellum extends farther caudad, relatively, over the medulla, this nerve enters the medulla in front of the hind end of the cerebellum; this condition is shown in fig. 14. Cephalad of this strands are seen running from the acusticum athwart the arms of the Y to the sides of the medulla. These are composed of fibres of the acustico-lateral nerve acusticas a, and of the ear nerves—acusticas c and acusticus d.

The Fourth Ventricle.—Leaving the external anatomy of the medulla of Bdellostoma, and turning to a study of sections through it, its most striking feature is found to be the smallness of the fourth ventricle, and the entire absence, except for about 11 μ. behind the cerebellum, of the exceedingly thin roof found over the fourth ventricle of other vertebrates. The fourth ventricle of Bdellostoma is composed of three distinct parts, an anterior and a posterior dilatation and a connecting canal. The anterior dilatation is 1·07 mm. long, ·036 mm. wide, and ·59 mm. deep. It lies in the cephalic end of the medulla, and is almost entirely covered by the cerebellum. The thin part of the roof lies over the hind end of this dilatation. The iter leaves the fourth ventricle at its cepbalo-dorsal angle, and immediately ventral to the iter is a small finger-like diverticulum (figs. 5, 6). The hind end of the dilatation gives off several small diverticula dorsal to the connecting canal. The number of these varies in different specimens, one brain having two, and two others three. Sometimes one of these diverticula may divide into two. The connecting canal leaves the anterior dilatation at its caudo-dorsal angle, and runs straight caudad to the canal of the spinal cord. It is ·065 mm. in diameter. The posterior dilatation is the most curious part of the fourth ventricle. It begins ·57 mm. cephalad of the commissura infima Halleri, and ends ·21 mm. caudad of it, having a total length of ·78 mm. Before it is reached the connecting canal has become narrower, so that it is about half as wide as it is deep. The posterior dilatation is semicircular in cross section, ·168 mm. wide and ·073 mm. deep, and is placed dorsal to the canal, which opens into it. Canal and dilatation together look, in cross section, like a longitudinal section of a toadstool. As the spinal cord is reached the whole ventricle contracts, preserving, however, its peculiar form, which is continued in the central canal of the spinal cord (fig. 12), the only difference being that in the cord the dorsal part is much smaller compared to the ventral than in the medulla.

Regions of the Medulla.—In analysing the medulla of Bdellostoma into its component regions it is easily seen that the great ascending tract of the sensory trigeminus is the dominant structure. This tract, judging from Johnston, 1902, is relatively much larger in Bdellostoma than it is in Petromyzon. It is the largest and most important tract of the hind brain, just as the olfactory tracts are the largest and most important tracts of the fore part of the brain, and it dominates all the anatomical arrangements in the medulla, displacing other tracts. It springs from the nucleus funiculi, a diffuse group of cells lying near the dorsal surface of the hind end of the medulla. Whether this nucleus is in Bdellostoma a development of the dorsal horn of the spinal cord, as stated by Johnston for Petromyzon, I am not prepared to say, but its position makes it likely. The nucleus extends cephalad for quite a distance, and the fibre tract, constantly growing larger and more pronounced, takes its way in the dorso-lateral part of the medulla to emerge at its cepbalo-lateral angle as the greater part of the trigeminus nerve (figs. 1, 10, 11).

The fasciculus communis, the continuation in part of the dorsal columns of the spinal cord, lies in the dorsal part of the medulla, next to the median line (fig. 1). A little cephalad of the commissura infima Halleri, a crescentshaped line of cells, the nucleus of the fasciculus communis, appears on the ventral surface of the fibre tract, and continues as a semicylindrical layer as far forward as the exit of the facialis from the brain. A little behind this the fibre tract begins to spread farther over the surface of the medulla, gradually shifting its position to one more removed from the median line (fig. 11). While shifting its position it also sends fibres laterad over the sensory trigeminus to the sides of the medulla, and, shortly after the level of the posterior end of the cerebellum is reached, the fasciculus communis is entirely displaced from the dorsal to the lateral surface. As soon as its fibres reach the lateral surface they turn caudad, running backwards to the facialis, glosso-pharyngeus, and vagus respectively.

The acusticus nucleus makes its appearance as a small clump of cells a little caudad of the level of the facialis, and lies not far from the median line between the fasciculus communis and the sensory trigeminus. It increases in diameter as it proceeds cephalad, and stains more deeply in hæmatoxylin than the surrounding tissue. As the fasciculus communis passes over to the side the acusticus comes to lie on the extreme doiso-mesial surface of the medulla (fig. 10). The fibres of the two sets of nerves arising from it, those of the lateral line and those of the ear, leave it in entirely separate groups. The most cephalic set of fibres is that of Acusticus a. These leave the dorsal surface of the acusticus in a well-defined group · 57 mm. from its anterior end, and pass laterad to the cranium above the dorsal surface of the medulla just cephalad of the acusticus ganglion. The fibres of the ear nerves are given off from just behind those of Acusticus a to within about ·5 mm. from the caudal end of the nucleus. Most of them pass laterad over the sensory trigeminus, leaving the medulla along its dorso-lateral edge, but some are deflected by the sensory trigeminus, pass ventrad of it, and leave the medulla in separate fibre groups along its mid-lateral surface, entering the ganglion near its ventral border. The fibres of Acusticus b arise in the ventral portion of the acusticus about three fifths of its length behind its cephalic end (fig. 11), and pass latero-dorsad to emerge upon the dorsal surface of the medulla. In the specimen from which the photograph of fig. 11 was taken ganglion cells lie among the fibres immediately outside the brain. If fig. 11 be examined with a magnifying glass a few can be seen in the nerve of the left side.

The motor cells of the medulla are gathered in two main groups, a lateral and a ventral motor column. The latter is the continuation in the medulla of the ventral horn of the spinal cord. It is comparatively small, and lies near the median line, ventral to the commissural fibres of the median raphe, and continues cephalad nearly to the anterior end of the medulla. It gives fibres to the vagus, but so far I have not been able to find any fibres going from it to the other nerves. Near its anterior end, about the level of the hind end of the cerebellum, it contains a number of cells of Mauthner, very large and irregular in shape. The lateral motor column begins at the hind end of the medulla and runs along its ventro-lateral angle as far forward as the motor trigeminus. It consists of three divisions, an anterior, closely packed sphere of cells, a middle division, in which the cells lie farther from each other and without definite arrangement, and a posterior division, where the cells are arranged in several longitudinal rows. This last division supplies the vagus and glosso-pharyngeus. The facialis draws its motor fibres mainly from the posterior’ half of the middle division, though taking a few from the anterior half. Most of the cells of the anterior half of the middle division, and all of those of the anterior division, send their fibres into the trigeminus.

The two sides of the medulla are connected along its entire length by a system of commissural fibres, which seem to come from all the regions of the medulla, and which cross in the median raphe immediately ventral to the fourth ventricle.

Bdellostoma has only seven of the ten to twelve cranial nerves found in most vertebrates, viz. the olfactorius, opticus, trigeminus, facialis, acusticus, glossopharyngeus, and vagus. Of these the olfactory and the trigeminus are the largest and most strongly developed, the optic is the smallest, and the acustic may be divided into four parts, two lateral line nerves, acusticus a and acusticus b, and two ear nerves proper, acusticus c and acusticus d. There is no trace in the adult of any of the eye muscle nerves.

N. Olfactorius.—Before describing the olfactory nerve it is necessary to give some idea of the organ innervated by it. The nasal organ of Bdellostoma is about 8 mm. long, 7 mm. wide at cephalic end, and 9 mm. wide at its caudal end, and 7 mm. deep. The measurements belong to specimen B of the previous table. The organ is bilaterally symmetrical, and consists of seven trapezoidal folds, hanging by their broader bases to the dorsal wall of the nasal capsule, and tapering to a point at the cephalo-dorsal surface. The median fold expands into a bulb at the cephalic end. Besides these seven free-hanging folds there are two smaller half folds, or flaps, fastened throughout almost their whole extent to the lateral walls of the capsule. Each half of the median fold belongs physiologically to its own side of the organ, and is innervated exclusively by the nerve of that side. These folds are related in a very definite way to the divisions marked on the surface of the olfactory lobes. To make the description simpler I shall call the half of the median fold A, the other folds in order B, C, D, and the lateral flap E. The nerve-fibres of each of these folds gather into bundles as they leave them, one bundle for each half of a fold, and in this form pierce the cranial wall. Once within the cranium the bundles break up, the different sets of fibres separating to seek their respective endings in the olfactory lobes. The nerve-fibres of A and B run exclusively to glomeruli in the dorsal part of the olfactory lobes, those of A to the inner division only, those of B to the inner and middle divisions. The C fibres run to both the dorsal and ventral parts of the lobe, the dorsal fibres being distributed to the middle division, the ventral fibres ending in all three, but principally in the inner and lateral ones. The fibres of D and E run exclusively to the ventral half of the lobes, those of D going to the ventral surfaces of the middle and lateral divisions, those of E to the lateral division only (figs. 1 and 2).

Interesting questions arise here : (1) What physiological significance has this difference in distribution, and (2) what, if any, are the differences of connection of these different divisions of the glomerular layer with other parts of the brain ? I hope, in a later paper, to throw some light upon at least the last of these questions.

N. Opticus.—The optic nerve of Bdellostoma is very small and delicate, owing to the undeveloped condition of the eye, the latter consisting merely of a round, two-layered cup. On leaving the eye the nerve runs raedio-caudad, passing ventrad of r. ophthalmicus V and r. maxillaris posterior V, and dorsad of r. maxillaris anterior V and r. mandibularis V (fig. 14). It enters the cranium in the ventral part of its lateral wall, running caudad in the cranial wall for ·36 mm., and, emerging on the floor of the cranial cavity, runs dorso-mediad to enter the ‘tween brain on its ventral surface as described above. There is no external chiasma.

N. Trigeminus.—This is the greatest of all the cranial nerves except the olfactory. It has two roots, a dorsal sensory, and a ventral motor. The sensory root is decidedly the larger and stronger of the two, and lies cephalad as well as dorsad oE the motor. Most of the sensory fibres are fine, but a few coarser ones are found among them, notably in the ophthalmic branch ; the motor fibres are coarse and stiff. All the fibres are medullated outside the brain. From the moment the nerve leaves the brain the ophthalmic fibres can be distinguished from those of the maxillo-mandibular trunk. They emerge cephalo-mediad of the others, and lie mesial to them in the nerve trunk until the ganglion is reached.

Leaving the medulla at its cephalo-lateral angle (figs. 1 and 2) the Trigeminus runs cephalo-laterad as a thick fibre bundle to the cranial wall, where it breaks up into several smaller strands in order to pass through the sieve-like foramen. Once outside, the sensory fibres enter the Gasserian ganglion, while the motor fibres run, a compact bundle, along the ganglion’s ventral surface. The ganglion itself is cylindrical in shape, its long axis directed cephalo-caudad ; it lies just outside the cranium, and extends from the anterior end of the cerebellum to about the middle of the olfactory lobes. The ophthalmic and maxillo-mandibular parts of the ganglion overlap each other, the former extending farther cephalad, and not so far caudad as the latter. The ophthalmic cells of the ganglion, like the ophthalmic fibres of the nerve, lie dorso-mediad of the maxillo-mandibular portion.

B. Ophthalmicus.—This first branch of the Trigeminus is purely sensory. P. Fürbringer declares that it has a few motor fibres, but I have found absolutely no trace of them, either in the origin or distribution of the branch, whether working with sections or by dissection. After leaving the ganglion the ophthalmicus runs forward, passing dorsad of the optic nerve, between the eye and the skull. Immediately in front of the eye it gives off a stout supraorbital branch, r. ophthalmicus cutaneus (r. cutaneus superficialis posterior, P. Fürbringer; r. dorsalis, Bowers, 1900). This branch runs dorsad between the muscles, appearing on the surface of the head muscles just cephalad of the eye. Running cephalo laterad across the subdermal lymph space, it reaches the skin, runs cephalad immediately under it for some millimetres, and divides and subdivides into numerous fine branches, innervating the skin on the dorsal and lateral surfaces between the levels of the third tentacle and the anterior end of the nasal capsule (fig. 14). About half way from the Gasserian ganglion to the head, the main trunk of the ophthalmicus divides into a superficial and a deep branch. The superficial branch, which is the smallor of the two, runs forward near the nasal tube, and along the dorsal surface of m. palato-ethmoideo-superficialis (the names of the various muscles are those given by Fürbringer). On reaching the anterior third of the nasal tube it divides, sending one, or possibly two small twigs to innervate the lateral face of the tube, while the rest goes on to the skin, innervating it in the extreme anterior dorsal head region, near the roots of the first and second tentacles. R. ophthalmicus profundus also runs forward, lyingbetween mm. palato-ethmoideo-superficialis and palato-ethmoideo-profundus. On nearing the anterior end of the head it turns dorsad, piercing the former muscle, and dividing into small ramuli, supplying the skin around the nasal opening, and the second tentacle. The innervation of the first three tentacles is practically the same. The nerve enters in one or two large branches and immediately subdivides into several fine ones that fill the entire space between the cartilage core and the skin. These branches lie so close together that in longitudinal sections there seems to be a solid nerve mass surrounding the core ; and it is only in cross sections of the tentacle that the nerve mass is seen to consist of fine ramuli, running in converging lines toward the tip.

Truncus Maxillo-Mandibularis.—This great division, containing both sensory and motor fibres, on leaving its own part of the Gasserian ganglion, divides quickly into four branches, two of them belonging to the r. maxillaris, and two to the r. mandibularis. R. maxillaris anterior (Fürbringer’s r. anterior maxillaris externus) is a broad, stout trunk, running forward beneath the sub-ocular arch, and through m. palato-coronaris, where it divides into a superficial and a deep branch. The superficial branch runs forward between m. palato-coronaris and m. copulo-quadratus-profundus, dividing at the anterior end of the latter muscle into four branches. Of these the outer and smaller is purely motor, supplying m. copulo-palatinus, the others are sensory, supplying the first, third, and fourth tentacles, and the skin immediately around them, The nerve of the first tentacle is particularly rich in small skin rami. The fourth tentacle is different in shape from the other three. They have narrow bases and taper to a sharp point, while the fourth, which is really a fold of the lip, has a broad base, and is short, blunt, and flabby. It has, in consequence, much more space inside of it than the other three, the nerve does not pack it closely as it does them, and its fine fibres are lost in the connective tissue. In addition to supplying the tentacle and the skin immediately around it, the nerve to the fourth tentacle has a large cutaneous branch that curves caudad, and supplies the lip at the lower angle of the mouth.

The deep branch is purely motor. It runs cephalad beneath m. palato-ethmoideo-profundus, into which it sends a large branch. This branch, after supplying the muscle, forks, sending a dorsal branch to supply m. nasalis and m. palato-ethmoideo-superficialis, and a ventral branch to m. tentacularis posterior. The main branch turns ventrad and forks, sending one branch running along the lateral face of m. copulo-ethmoidalis, into which it sends twigs, and then into m. copulo-teutaculo-coronaris ; the other branch runs straight forward to end in in. ethmoideo-nasalis.

R. maxillaris posterior is sensory, and helps to supply the mucous membranes of nose and mouth. It has a very short trunk, dividing, before it reaches the sub-ocular arch, into r. nasalis and r. palatinus. R. nasalis passes over the sub-ocular arch and under r. ophthalmicus, lying so close to this nerve that Fürbringer thought it arose from it. It moves toward the nasal tube just in front of the olfactory capsule, and runs cephalad along its side, giving off several branches, thus supplying the posterior two-thirds of the tube. R. palatinus passes under the sub-ocular arch, and runs cephalo-ventrad inside the palatine bar, giving off several twigs to the pharynx by the way. At the anterior end of m. velo-quadratus it divides into four small branches and innervates the roof of the mouth. This nerve is Fürbringer’s palatine branch of r. medius maxillaris externus.

R. Mandibularis (Fürbringer’sr.posterior maxillaris externus V + r. maxillaris internus V).—This nerve is mainly concerned with the innervation of the dental plate and its accompanying muscles, thus furnishing another proof of the statement made by Ayers and Jackson, 1900, that the tooth-plate represents the lower jaw, and not the tongue, as assumed by Müller and Fürbringer. R. mandibularis, like r. maxillaris, consists of two divisions—an anterior and a posterior,—both of which contain both sensory and motor fibres. The anterior division consists of. four branches, the first of which, r. dentalis, is a large, purely sensory nerve, running ventrad m. palato-coronaris and the wall of the pharynx to the tooth-plate, in which it subdivides profusely. The second branch is motor, and is small and slender. It divides almost immediately into a short posterior branch, running into m. copnlo-quadratus profundus, and anterior branch running cephalad outside of r. maxillaris anterior, and on the inner face of m. copulo-quadratus profundus, to which it appears to give small twigs. It then runs on to m. copulo-ethmoidalis, and apparently ends there. The third branch is also motor. It sends a tiny twig into m. velo-quadratus, and then runs ventrad between m. palato-coronaris and m. copulo-quadratus-profundus, supplying this latter muscle and m. hypo-copuloglossus. Last of all is a small branch—motor and sensory combined,—the motor twig going to m. velo-quadratus, the sensory to the roof of the pharynx beneath.

R. mandibularis posterior consists of two divisions. The anterior division, though mainly sensory, carries a couple of motor twigs to m. velo-quadratus. The sensory trunk runs between this muscle and the roof of the pharynx, sending twigs into the latter, and ending in the pharynx behind the first branchial arch. The posterior branch of r. mandibularis posterior is the great nerve supplying the muscles of the lower jaw, the “club muscle” and its accessories. It runs ventrad between m. copulo-quadratus-prof undus and the pharynx, supplying the caudal part of this muscle and m. hyo-copulo-glossus (Fiirbringer’s r. pro. hyo-copulo-glossus), and then divides into two, an anterior branch, supplying m. copulo-glossus-superficialis, and tn. copulo-glossus-profundus, and a posterior branch to the “club muscle,” the retractor of the lower jaw. This last branch also divides into two, an outer and an inner. The outer branch runs straight caudad in the fascia of the outer circular muscle, and on reaching the muscle itself divides into dorsal and ventral branches. The ventral branch runs on the upper surface of the circular muscle just beneath the long central muscle, sending off tiny threads all along the way. At the end of the circular muscle it passes diagonally upward, running across m. perpendicularis. The dorsal branch runs in the depths of the dorsal part of the circular muscle just over the central muscle, and it also gives off tiny threads along the way. About half way back it anastomoses with its fellow of the opposite side, and from this anastomosis three or four strands are given off, all running straight backward. Among them are two principal ones, a dorsal strand that passes very obliquely upward and ends at the broad dorsal tendon of the circular muscle, and a ventral one that remains on the deeper level and, running through m. perpendicularis, ends in the long central muscle. The inner branch goes back a little, then turns suddenly mediad and enters the tendon of the long central muscle, running backward in the tendon.

Caudad of r. mandibularis the main trunk of the trigeminus gives off a small motor branch to m. veloquadratus.

N. Facialis.—The facial nerve is exceedingly small and fragile, very difficult to dissect out, and hard to follow to its termination even in sections. It leaves the medulla caudad of acusticus a and acusticus b, passes through the acusticus ganglion, and breaks through the cranial wall just cephalad of the ear capsule. Haematoxylin sections give no hint as to its containing more than motor and communis fibres. On leaving the cranium the facialis runs caudoventrad under the hyoid arch. Its ganglion lies close to the ear capsule, and is always present, though variable in shape, consisting sometimes of a compact group of cells, forming a cluster about twice the diameter of the nerve, and sometimes having its cells lying singly, embedded in the nerve cord. After leaving the ganglionic region the facial nerve continues caudo-ventrad until it reaches the anterior border of the crani-hyo muscle, into which it sends a small twig. Then turning, the nerve runs caudo-cephalad, curving over the hyoid arch, and courses cephalad on the mesial surfaces of mm. copulo-quadratus-profundus, and hyo-copulopalatinus. It innervates the second of these muscles, but the innervation is hard to find, for instead of sending in the usual nerve twigs to subdivide in the heart of the muscle, it sends the fibres in as clusters of separate threads, and they are lost almost immediately. I have not found, either in sections or in dissecting, any twig answering to Furbringer’s r. cutaneus. While running on the inner surface of m. hyo-copulo-palatinus, the facialis divides into two or three branches, the dorsal branch carrying the motor fibres. There is nothing fixed as to the level of this division or the number of branches, the nerves of opposite sides differing sometimes in the same head. The branches turn ventro-laterad, passing between m. hyo-copulo-palatinus and m. copulo-palatinus, and forward over the latter muscle to the skin. They are so very small that I have not been able to trace them definitely to their endings in the skin, but I have every reason to believe that they terminate in the region of the fourth tentacle. I have found no communicating branch between the facialis and the trigeminus, or between the facialis and the glossopharyngeus.

The Nerves of the Lateral Line.—These two nerves, acusticus a and acusticus b, it is more convenient for purposes of comparison to consider together, for they represent in Bdellostoma what Johnston, 1902, calls the “lateral line VII” of Petromyzon, and Strong, 1895, the “dorsal VII “or “VII b “ of Amphibia. The position of these nerves in Bdellostoma is exceedingly interesting, for here only, so far as I know, do they arise and remain free from complication with the facialis or trigeminus. They arise from the tuberculum acusticum, and run, free from all entangling alliances, from there to the skin ; hence it seems best to cast aside the old names and call them acusticolateralis a and b, or, more simply, acusticus a and acusticus b.

Acusticus a (Johnston’s “‘ventral root of lateral line VII,” Strong’s “dorsal half of dorsal VII “), leaves the side of the medulla a little caudad of the trigeminus and immediately cephalad of the acusticus ganglion. Once past the cranium, its fibres run cephalo-laterad into a small ganglion lying caudad of, and in close contact with, the Gasserian ganglion. On leaving the ganglion the nerve runs cephalo-laterad toward the skin, passing caudad of the eyeball. It continues forward, directly beneath the skin, until near the front end of the nasal capsule, where it divides, sending one branch ventro-caudad, one ventrocephalad, and one cephalo-dorsad, supplying the skin at the side of the head, between the mouth opening and the anterior end of the brain.

Acusticus b [Johnston’s “dorsal root of lateral line VII,” Strong’s “ventral half of dorsal VII,” M. Fiirbringer’s “first spino-occipital nerve” (I)], leaves the medulla on its dorsal surface near the level of the hind end of the cerebellum, and runs dorso-laterad to the dura, then curving sharply, it runs caudo-ventrad in the dura, entering the cranium posterior to acusticus c, the first ear nerve (fig. 1). It generally has a ganglion lying within the cranial wall. Sometimes this ganglion is a very decided one, sometimes it consists only of a few cells. In addition to this it may have a second ganglion lying outside the cranial wall, or a few ganglion cells among its fibres immediately after leaving the brain. After leaving the cranium the nerve runs cephalolaterad without incident, entering the subcutaneous lymph space immediately behind acusticus a. It runs boldly out to the skin, turns sharply dorsad, and then caudad, supplying the skin immediately over the fore part of the brain.

N. Acustici c and d.— The acusticus ganglion lies inside the cranium and the dura, close to the cephalic half of the lateral surface of the medulla. Most of the fibres from it enter the medulla along the dorso-lateral angle, but small clusters enter the lateral surface farther ventrad. The ganglion is pear-shaped, with its broader end cephalad, and is placed immediately caudad of the fibres of acusticus a. The acustic nerve consists of two branches, r. utricularis, acusticus c, and r. saccularis, acusticus d, entirely separate and distinct, and leaving the ganglion at different places. R. utricularis is the more cephalad of the two; it leaves the ganglion a little in front of its hind end, curves ventro-laterad, and enters the ear at its cephalo-ventral angle. R. saccularis forms the caudal continuation of the ganglion, running caudad between the glosso-pharyngeus and the cranial wall, its fibres passing mediad of the endolymphatic duct. At the hind end of the brain capsule, just cephalad of the foramen of glosso-pharyngeus and vagus, r. saccularis curves ventro-laterad,passing through the membrane into the ear capsule at its caudo-ventral angle.

N. Glosso-pharyngeus.—(Mülle’s and M. Fürbringer’s r. glosso-pharyngeus X.) The glosso-pharyngeus leaves the medulla along the caudal half of its lateral face. It has from four to seven roots of origin lying one behind the other between the acusticus ganglion and the exit of the vagus, and contains both sensory and motor fibres. Its fibres run caudad between acusticus d and the vagus, lying so close to the latter that it is difficult to distinguish one from the other. The joint foramen of the glosso-pharyngeus and vagus lies at the caudal end of the cranium, in the angle formed by the junction of the ear capsule and the cranium, and the glosso-pharyngeal fibres pass out dorsal, anterior, and lateral to those of the vagus. The two nerves pass so close together, though, that I cannot say whether or not a few fibres of each stray into the trunk of the other. Once outside the cranium the glosso-pharyngeus and vagus run together in the same sheath, and a few millimetres beyond the capsule each gives a few fibres to the trunk of the other. From this point on, however, the two nerves may easily be dissected apart. The glosso-pharyngeal ganglion is small, causing no enlargement of the nerve trunk, and lies a little caudad of the brain capsule. The glosso-pharyngeus runs in the same sheath with the vagus as far as the second branchial arch ; then separating, it runs ventro-caudad, passes through m. constrictor-pharyngis, into which it sends a couple of branches, and so to the pharynx wall, dividing into several branches, and innervating the side of the pharynx near the end of the backward position of the second branchial arch.

N. Vagus.—The vagus is the most posterior of the cranial nerves of Bdellostoma, arising from the caudal end of the medulla by four or five roots. These roots run caudolaterad as separate, though close lying, strands, passing between the dorsal and ventral roots of the first and second spinal nerves, to the wall of the brain capsule. After the give and take of fibres with the glosso-pharyngeus the vagus runs straight caudad, near to the notochord, until it nears the first gill sac. It has no ganglion near the cranium, although a few scattered ganglion cells may be found near the brain, and also outside the cranial wall. In the region of the branchial rami, however, these cells occur in great numbers. They are rather smaller than the cells of the Gasserian ganglion, are of a long, oval shape, as though compressed by the nerve, and are arranged in longitudinal strands. These strands are found from cephalad of the first to caudad of the last branchial ramus, and lie all through the thickness of the nerve, but are more plentiful at the periphery than at the centre. All of the gill sacs are innervated in the same way. A ramus branchialis is given off that divides into an interior and exterior branch. The interior branch runs to the duct connecting the gill sac with the oesophagus, and divides again into dorsal and ventral rami.

The dorsal ramus passes above the duct, innervating its dorsal surface ;, the ventral ramus divides once more, giving off a small pharyngeal branch, that also sends tiny ramuli to the proximal end of the duct, and another branch that penetrates the duct at its distal end near the gill. In one specimen dissected I found on the first gill of the right side another fine branch that entered the anterior edge of the gill itself on the outer side. After the last gill sac the vagus enters m. constrictor cardiæ, where.it forms a dense plexus ; emerging from this it joins its fellow of the opposite side to form n. intestinalis impar, running caudad on the dorsal surface of the intestine, and sending fibres into it all along its course.

These are complete nerves, such as are found in the higher forms, with a sensory ganglion lying just outside the sheath of the spinal cord, and sensory and motor roots. There is a difference between the first two nerves emerging immediately caudad of the cranial nerves and those that come after, and Fiirbringer’s division into spino-occipital nerves and spinal nerves is a good one, though I cannot place my division where he places his. Judgingfrom his paper (M.Fiirbringer, 1897), he worked with mature adults only, and imperfect material. This explains in part his confusion in regard to these nerves, for natural growth causes somewhat of a shifting of relative position outside the brain between the young animal aud the mature adult. in fig. 14, for example, which was plotted from a specimen the diameter of whose head was about half the diameter of the adult head, acusticus b and the first spino-occipital nerve are seen to be separated by a considerable space, but in the mature adult, such as fig. 1 was drawn from, the foramen of acusticus b is farther caudad with reference to the medulla—the medulla has apparently grown backward, carrying this nerve with it—and, as a result, when outside the cranium, acusticus b lies for a little while immediately in contact with the first spino-occipital nerve. This is apparently one reason for Fürbringer’s calling acusticus b the first spino-occipital nerve, in spite of the fact that it arises from the acusticus centre of the medulla. In regard to his possible ventral root to this nerve, which he admits not having been able to trace as far as the skull, I find nothing whatever answering to it, unless, indeed, he has in mind one of the separated sets of acusticus fibres that enter the acusticus ganglion from the ventral part of the lateral surface of the medulla.

The First Spino-occipital Nerve.—The roots of this nerve are seen emerging from and entering the spinal cord immediately caudad of the last root of the vagus, in that transition region where the medulla ends and the cord begins. It has one sensory root and two motor ones, the sensory root a trifle the most anterior of the three. The sensory and dorsal roots curve away from each other to admit of the passage between them of theglosso-pharyngeusand vagus, and then approach each other as they reach the cranial wall. Here a curious thing happens : the sensory root divides, passing the wall in two sets of fibres, through two distinct foramina. The anterior set has a small ganglion, apparently apart from the main ganglion, lying in or inside the wall. After passing the wall both strands unite to enter the main ganglion of the nerve. The two motor roots divide also, passing the wall through three foramina and uniting again on the other side, and motor and sensory roots unite to form one nerve-trunk laterad of the ganglion. It is possible that this division of the roots in passing the wall indicates the fusion in the first spino-occipital of two distinct nerves, and that study of different stages of embryos will throw some light on the question. The main trunk of the nerve runs cephalo-laterad, then, bending ventrad, it divides into two branches, that run ventro-cephalad on the inner face of the lateral longitudinal body muscle, supplying finally the lateral body muscles and the skin.

The Second Spino-occipital Nerve.—This second spino-occipital nerve is similar, in some ways, to the first, in other ways it is more like the purely spinal nerves that follow it. It also arises from the transition region between brain and cord. It has one sensory and four motor roots, which belong, however, to two main stems, of two roots each, corresponding to the two motor roots of the spinal nerves. In this nerve, as in the following ones, the motor roots are cephalad of the sensory ones. The sensory root passés out through its single foramen to the ganglion, the motor roots are variable on the right side of one specimen examined passing through two foramina, and on the left side uniting inside the wall and passing as a single trunk. It must be noted that in all the spinal nerves, however, the motor roots emerge from the wall as a single trunk, the separate foramina having run together and united before the lateral face of the wall is reached. The motor root, as in the case of the first spino-occipital nerve, runs ventrad of the ganglion, joining the sensory root as it emerges therefrom. The nerve then runs cephalo-caudad for a space, like the first spino-occipital, then turns ventrad, running on the inner face of the lateral body muscles. Unlike the first spino-occipital it does not divide into two branches until near the ventral surface.

The First Spinal Nerve.—This nerve is the first to arise from the region of the spinal cord proper beyond the area of transition, and there are other features about it that distinguish it sharply from the two nerves preceding. It has one sensory and two motor roots. As in all the spinal nerves, the motor roots pass the wall by a Y-shaped foramen, and emerge as one trunk. It is outside the wall that a striking difference is seen between this nerve and those that have gone before, for the spinal nerves supply the dorsal as well as the latero-ventral skin and muscles. The ganglion, instead of being a spherical or ovoidal cluster of cells, is elongate, with two arms, a dorsal and a lateral. The dorsal arm, running upwards outside of the sheath of the spinal cord, gives off the dorso-sensory branch of the first spinal nerve. This branch runs dorso-cephalad until it reaches the top of: the sheath, then mediad to the median line, and straight dorsad to the skin. On reaching the skin it runs cephalad, running near the median line, and supplying the skin until its twigs finally disappear near the end of acusticus b. The sensory nerve coming from the lateral branch is the one joined by the motor trunk, and the united nerve runs straight laterad beneath the dorsal body muscles. Several strong twigs, representing together Fürbringer’s dorso-motor branch, penetrate this muscle and ramify inside it, while the ventral branch, motor and sensory combined, corresponding to the two spino-occipital nerves that have preceded the first-spinal nerve, takes its way down the body wall on the inner face of the lateral longitudinal muscle.

Illustrating Miss Julia Worthington’s paper on “The Descriptive Anatomy of the Brain and Cranial Nerves of Bdellostoma Dombeyi.”

ac. Acusticus region of the medulla, ac. g.,acus. g. Acusticus ganglion. a. v. Semicircular canal at base of infundibulum, b. M. Bundles of Meynert. c. a. Ansulate commissure, cb. Cerebellum, cb. v. Cerebellar ventricle. c. c. s. Central canal of the spinal cord. c. M. Cells of Mauthner. d”.b.M. Second decussation of the bundles of Meynert, d. c. s. Dorsal commissure of the spinal cord. d. e. Dorsal evagination of the roof of the infundibular process, d. h. Dorsal horn of the spinal cord. d. r. Dorsal root of spinal nerve, f. b. Fore brain, f. b. 1. Outer shell of the fore brain, f. b. 2. Inner core of the fore brain, f. c. Fasciculus communis. Gass.g. Gasserian ganglion, g.olf. Olfactory glomeruli. h.Hypophysis, h.g. Habenular ganglia. h.Iter. inf. Infundibulum, inf.p. Infundibular process, ini.a. Intra-cranial artery, ip. g. Interpeduncular ganglion, l.m.c. Lateral motor column. m.b. Mid-brain, md. Medulla, m. V. Motor Trigeminus. N.Acus.a., N. Acus. b. Acustico-lateral nerves. NN. Gl-ph. &Vag. Glosso-pharyn-. geus and Vagus. N. Olf. Olfactory nerve. N. Opt. Optic nerve. 0. ch. Optic chiasma. olf. g. Olfactory glomeruli, olf. I. Olfactory lobe. olf. o. Olfactory organ. R.Nas. Ramus nasalis. R.Hand,ant. 2, R. Hand,ant. 3, R. Hand. ant. 4. Second, third, and fourth branches of ramus mandibularis anterior. R. Hand. dent. Ramus mandibularis dentalis. R. Hand. post. 1, R. Hand. post. 2. First and second branches of ramus mandibularis posterior. R, Hand. post. 2a. Anterior branch of ramus mandibularis posterior 2. R. Hand. post. Hb. Posterior branch of ramus mandibularis posterior 2, running to the “club muscle.” R. Max. ant. Ramus maxillaris anterior. R. Oph. cut. F. Ramus ophthalmicus cutaneus Trigemini. R. Oph. sup. V. Ramus ophthalmicus superficialis Trigemini. R. Oph. pro. V. Ramus ophthalmicus profundus Trigemini. R. Pal. Ramus palatinus. R. pro. v. q. Ramus profundus maxillaris velo-quadratus. Sp. c. Spinal cord. Sp. 1. First spinal nerve. Sp. o. 1. First spino-occipital nerve. Sp. o. 2. Second spino-occipital nerve, s. V. Sensory Trigeminus, t. b. ‘Tween brain. Ten. 1, Ten. 2, Ten. 3, Ten. 4. First, second, third, and fourth tentacles, tr.c.s. Tractus cerebello-spinalis. tr.f.t. Tract running from the fore brain to the ‘tween brain, tr.h.t. Tractus habenulotectalis. tr.o.c. Tractus olfacto-cerebellaris. tr. t. b.el s. Tractus tecto-bulbaris et spinalis, v.p. Vascular plexus, v. h. Ventral horn of the spinal cord. III v, IV v. Third and fourth ventricles, I, II, V, VII, VIII a, VIII b, VIII c, VIII d_t IX, X. The cranial nerves respectively.

Fig. 1.—The brain, dorsal view, × 8⁠.

Fig. 2.—The brain, ventral view. × 7.

Fig. 3.—The brain, lateral view. × 7.

Fig. 4.—The medulla. The fore part of the brain is bent down till it forms with the medulla an angle of 120°. × 7.

Fig. 5.—Diagram of the ventricles of the brain. The dotted parts are always open, × 7.

Fig. 5 a.—The semicircular canal at the base of the infundibulum. × 7.

Fig. 6.—Photograph of a sagittal section of the brain, a little removed from the median plane. × 15.

Fig. 7.—Photograph of a horizontal section of the brain, about three fifths of the way from the dorsal to the ventral surface. × 15.

Fig. 8.—Photograph of a cross section of the brain through the entrance of the optic nerve, × 15.

Fig. 9.—Photograph of a cross section of the brain through the hypophysis. × 15.

Fig. 10.—Photograph of a cross section of the medulla, through the exit of the Trigeminus. × 15.

Fig. 11.—Photograph of a cross section of the medulla through the entrance of Acusticus a. × 15.

Fig. 12.—Photograph of a cross section of the spinal cord, × 42.

Fig. 13.—Photograph of a sagittal section through the infundibular process, in the median plane, × 40.

Fig. 14.—Diagram of the origin and distribution of the cranial, spino-occipital, and first spinal nerves. The outlines of the brain, eye, and ear are shown in dotted lines, ×about 9.