extreme variability of appearance and size of mesh (fig. 363) even in the course of the same nerve-fibre, lends much probability to the view that it is produced artificially by the precipitation by the reagent employed, of some material entering into the constitution of the myelin of the medullary sheath. Gerlach has in fact shown that the character of the reticulum varies with the strength of the alcohol employed to act upon the nerve.

It will be seen from the above that there is still much diversity of opinion with regard to the minute structure of the medullary sheath of the nerves. As to its chemical composition. the white substance or myelin consists chiefly of lecithin and neurin together with cholesterin and one or two other substances in less amount. When escaped from the nerve-fibres it forms drops either rounded or irregular in shape, which always show the double contour which is so characteristic of the medullated nerve-fibres, the appearance being due to the peculiar manner in which myelin refracts the light. In contact with water it combines with that fluid, and as a result of the imbibition the myelin-drops undergo a considerable increase in bulk, accompanied by remarkable changes of form, often growing out into tube-like filaments for a considerable distance into the fluid. In this behaviour myelin is not peculiar but resembles certain other substance of a fatty and resinous nature.

It is generally believed that medullated nerve-fibres occur only in vertebrates, but Retzius has described the nerve-fibres in Palæmon (a crustacean) as also possessing a myelin-sheath, with nodes and internodal nuclei.

Sheath of Schwann, primitive sheath, or neurolemma.-The sheath of Schwann forms the outermost covering of the white nerve-fibres. It has the appearance of a delicate homogeneous membrane with nuclei disposed at intervals along its inner surface. As already mentioned, these nuclei bear a definite relation to the segments of the nerve-fibre, for they lie about midway between the nodes, only one nucleus being found in each internode (fig. 359). The nuclei are oval and sometimes flattened, they usually lie in a depression of the medullary sheath, and at each end of the nucleus, especially in young nerves, there is a small amount of granular protoplasm which may spread for a short distance between the primitive and the medullary sheath. The primitive sheath usually dips in at the nodes and is there only separated from the axis-cylinder by the annular band of Ranvier before mentioned.

So long as the primitive sheath is accurately filled by the contained medullary substance its outline can seldom be distinguished, but sometimes, when the white substance separates at various points from the inside of the tube, the contour of the fibre becomes indented and irregular, and then the membrane in question may, in favourable circumstances, be discerned as an extremely faint line, running outside the deeply shaded border formed by the white substance.

In the white fibres of the brain and spinal cord the nucleated sheath is absent, and these are only invested by a medullary sheath. In consequence of the absence of the comparatively tough primitive sheath the fibres from these situations cannot be isolated for any distance without rupture, and it is found that for the same reason the medullary sheath readily breaks away from the axis-cylinder, so that this is thus left bare. It is usually stated that the nodes of Ranvier do not exist on these white fibres of the brain and spinal cord; but various observers have described appearances which seem to indicate that the nodes may be present in these fibres also.

Pale or non-medullated fibres; fibres of Remak.-These occur principally in branches of the sympathetic nerve, but they are found also in greater or less amount in the cerebro-spinal nerves. They are transparent, faintly striated fibres of varying size, which exhibit nuclei at frequent intervals. The nuclei are applied to the surface of the fibre, and, according to the generally received account, belong to a delicate homogeneous sheath, similar to the primitive sheath of the medullated fibre. It must be admitted, however, with Ranvier that it is difficult or impossible to exhibit the sheath, and if this is the case, the nuclei must be regarded as embedded in the peripheral layer of the fibre itself, and as belonging to this. Many of these

fibres are branched and united with neighbouring fibres, so as to form a network along their course; a condition which is never found in the course of the medullated

Fig. 364.-PORTION OF THE NETWORK OF FIBRES OF REMAK FROM THE PNEUMOGASTRIC

OF THE DOG (Ranvier).

n, nucleus; p, protoplasm surrounding it; b, striation caused by fibrils.

Fig. 365.-A SMALL BUNDLE OF NERVE FIBRES FROM THE SYMPATHETIC NERVE (Key and Retzius). The bundle is composed of pale nerve-fibres, with the exception of the fibre m, m, which is enclosed here and there by a thin medullary sheath; n, n, nuclei of pale fibres.

nerves.

The branches of the olfactory nerve consist wholly of pale fibres, but these are different from the ordinary pale fibres in being provided with a distinct nucleated sheath.1

NERVE-CELLS.

These are found in the grey matter of the brain and spinal cord and in the ganglia; they exist also in some of the nerves of special sense near their terminations, and occur here and there in the course of certain other nerves. They are often named ganglion-cells.

In shape, nerve-cells vary greatly. Thus they may be spheroidal or ovoidal with a general even outline, or they may be of an angular or irregular figure, and it is found that the nerve-cells have a characteristic shape in certain parts of the 1 On the structure of the non-medullated fibres of the sympathetic system, see a paper by I. L. Tuckett, in the Journal of Physiology, vol. 19, 1896, p. 267.

nervous system. For example, the cells from the spinal ganglia of man and most vertebrates are of a rounded shape (see fig. 366), those from the sympathetic ganglia are more angular (fig. 367), those from the grey matter of the spinal cord of an irregular form, and provided with numerous branching processes (fig. 368); those from the cerebral convolutions conical, those from a particular layer of the

Fig. 366.-Two NERVE-CELLS FROM A SPINAL GANGLION (HUMAN) (Retzius). sh, nucleated sheath; n, n, nuclei of the primitive sheath of the nerve. From each cell a fibre can be seen to arise, and after a convoluted course on the surface of the nerve-cell, to bifurcate (opposite d); from which point the divisions pass either in the opposite direction to one another, as in A, or at first in the same direction as in B. The nuclei of the sheath of the nerve-cell are all represented in B, but only those seen in profile have been represented in A.

Fig. 367.-A GANGLION-CELL WITHIN ITS SHEATH; FROM THE HUMAN SYMPATHETIC.

HIGHLY MAGNIFIED. (Key and Retzius.)

grey matter of the cerebellum, flask- or retort-shaped (fig. 374, c), and so on. Other cells, situated in the course of certain nerve-fibres, are somewhat spindle-shaped, the two poles of the spindle being prolonged into the nerve-fibre.

Nerve-cells vary much in size as well as in shape. Many of the nerve-cells in the body are very large, but there are others which are quite small. The latter are especially abundant in the deeper part of the grey matter of the cerebellum and in the retina, where they form what are known as the granule-layers (fig. 374, d).

Structure. The nucleus of a nerve-cell is usually a large clear round vesicle containing a very distinct highly refracting nucleolus, and, in some cases, an intranuclear net-work. The cell-substance is finely granular or punctated, sometimes indistinctly striated, or reticular in appearance. Cells are often to be seen which contain very distinct brown or yellow coloured patches caused by an accumulation

of pigment granules (fig. 372, b). The colour is deeper in adult age than in infancy.

Every nerve-cell has one or more processes, and the cells are often named according to the number of processes they possess, uni-, bi-, and multi-polar; terms not perhaps well chosen, but rendered current by use. A fibrillation similar to that in the axis-cylinder of a nerve-fibre is visible in the cell-processes, and it may also be traced passing from them into the body of the cell and even through this from

A

B

Fig. 368.-Two NERVE-CELLS FROM THE SPINAL CORD OF THE OX, ISOLATED AFTER MACERATION IN VERY DILUTE CHROMIC ACID. MAGNIFIED 175 DIAMETERS. (E. A. S.)

Each cell has a well-defined, clear, round nucleus, and a large nucleolus. The cell-processes are seen to be finely fibrillated, the fibrils passing from one process into another through the body of the cell. a, axis-cylinder process broken a short distance from the cell.

one process into another (fig. 368). Some affirm that they have been able to trace a connection of these fibrils with the nucleus of the cell, but this is at present doubtful.

Nerve-cells which are entirely destitute of processes have sometimes been described. It is possible that such may exist, but there is always a probability that these may

be cells, the processes of which have become broken away during the manipulation required for isolation.

a

B

R

Fig. 369. A NERVE-CELL FROM THE CEREBRAL CORTEX (Golgi).

a, axis-cylinder process.

Fig. 370.-A BIPOLAR NERVE-CELL, WITH ITS POLES PROLONGED INTO MEDULLATED

NERVE-FIBRES (Key and Retzius).

The whole is invested by the primitive sheath. R, R, nodes of Ranvier.

It has been shown that many cells have at least one of their processes prolonged as a nerve-fibre (axis-cylinder process) and the statement is probably true of all