moniliform condition is not an expression of the plasticity of the cell as certain authors have claimed.

In a later contribution GEIER ('02) presents important data upon the development of the protoplasmic processes as well as upon their form in the adult. He has studied by the GOLGI method the cells of the spinal cord of rabbits of different ages; one day, two weeks and one month; and of kittens at birth and at three days, seven days, one month and two months old. The two series of experiments lead to the same conclusions: the protoplasmic processes of the anterior horn cells are less regular in outline in the new-born than in the adult. As the animal grows the outline of the processes become more regular and the process straightens. In the new-born the dendrites have a more regular contour in the region of the cell body than in the distal region, and the process of straightening progresses from the cell body outward. As for the dendrites of the posterior horn, GEIER finds it difficult to determine whether there is any such change accompanying growth as he describes for the anterior horn cell. This difficulty arises from the fact which he demonstrated in his earlier work and which his present investigation confirms, that the dendrites of the posterior horn of the adult are normally very irregular in outline as compared with the dendrites of the anterior horn cells. He finds, however, that there are certain cells in the anterior horn which have very irregular dendrites. He considers them commissural cells. In his youngest specimens he found the varicose condition of the dendrites very conspicuous, but less marked in the older specimens. The condition tends to disappear as the animal grows.

Golgi's Endocellular Net.

SOUKHANOFF ('02) has made a specific study of GOLGI's endocellular net in the cells of the cerebral cortex of mammals by modifications of the GOLGI method. He finds this net only in a zone around the nucleus, although it does not lie directly upon the nucleus. Surrounding the net is a zone of protoplasm which is noticeably clearer than the rest. Some of the filaments of the net are fine, others are coarse, still others are ribbon-like. Often they are of irregular contour. In some small cells the net consists of only a few curls of the filaments, and it is a much simpler structure in the cortical cells than it is in the cells of the cord and spinal ganglia. In general form the net conforms rather closely to the shape of the cell. Sometimes one branch of the net, sometimes two or three, pass out into the process.

In the intepretation of the endocellular net SOUKHANOFF speaks

positively on three points: it is a strictly endocellular structure; it has nothing to do with the neurofibrils of APÁTHY and BETHE; it is not identical with the canaliculus HOLMGREN. Regarding its relation to "l'état spiremateau" of NELIS, he could not at the time of writing offer a positive opinion.

As already stated in the discussion of the structure of the ground substance, HATAI ('03) resolves the net in question into a modified protoplasmic reticulum. His position upon this point is strongly supported by almost every feature of the net which SOUKHANOFF emphasizes, the lack of uniformity in size and the irregularity in contour of the fibrils, the restricted perinuclear position of the net, its conformity in shape to the shape of the cell and its relation to the processes of the cell. In all of these features the net has a striking resemblance to the filaments which HATAI describes as formed by the matting together of numerous fibrillae of the neurosomic net. But it will be remembered

that HATAI considers that the neurofibrils of APÁTHY are identical with this modified reticulum. The interpretation would place him in direct opposition to SOUKHANOFF when the latter says that the GOLGI endocellular net has nothing to do with the neurofibrils of APÁTHY and ВЕТНЕ.

By HATAI's work, therefore, another of the manifold structura elements of the nerve cell is explained upon the basis of the funda mental structure of the protoplasm.

The Gemmules.

The structures which BERKELEY called "gemmules" have received various names by other investigators: "épines" by RAMÓN-Y-CAJAL; "appendices piriformes" by MLLE. STEFANOWSKA; and "appendices collatereaux" by others. As treated by many authors they may vary in form from short club-shaped to spindle-shaped or even filamentous structures. In his study of the cellular changes in the cerebral cortex under experimental anemia, SOUKHANOFF ('98) finds that as the symptoms become more acute the gemmules become modified and ultimately disappear. In another work of the same year, he discovered similar changes in the gemmules in animals which have been subjected to the vapors of ether, chloroform and alcohol. In other animals treated with injections of trional the pronounced modifications in the gemmules, as well as other changes in the central system, were attributed to the derangement of general nutrition and not to the specific action of the drug.

In further pathological studies with the GOLGI method, SOUKHAN

OFF ('98) observes that the gemmules are lost in varying degrees in the nerve cells of animals subjected to poisoning by arsenic, rabies, tuberculin and also of animals suffering from thyroidectomy. These studies were checked by preparations of nerve cells from a normal guinea pig killed by decapitation. The nervous tissue was quickly placed in the fixing reagents and received like treatments with the abnormal tissues. The preparations showed gemmules present almost universally upon the dendrites.

In collaboration with CZARNIECKI, SOUKHANOFF (02) has studied the dendrites of the ventral horn cells of the rabbit with the GOLGI method. By killing the animals quickly with chloroform the authors considered that the tissue was found in the normal condition.

Upon the dendrites of certain cells very few gemmules were found; in others they were numerous, beginning to appear on the dendrites nearer the cell body and becoming more numerous as the distance from the perikaryon increased till in the distal region they almost covered the process. In another type of cell, which lay with the protoplasmic processes partly within the white substance, that part of the dendrite which lay within the white substance bore no gemmules, while the part lying within the grey substance was abundantly supplied with them. And even when the dendrite lay in the border line between the grey and the white substance the side of the dendrite which faced the grey bore more gemmules than did the side facing the white. The authors conclude that the gemmules are much less numerous upon the ventral horn cells than upon the cells of the cerebral cortex, but that they are much more variable in size and shape. They may be pyriform, filamentous, club-shaped, tubercular, or finely branched. Certain of these forms which the authors call "rejitons," which are very irregularly fibrillar and club-shaped, are not found in the cells of the cerebral cortex.

GEIER ('01) has studied with the GOLGI method the cerebral cortex of mammals and birds which have been subjected to the vapors of ether and chloroform for from five to ten minutes. In some of the experiments the gemmules showed a tendency to disappear, while in others the cells appeared perfectly normal. The disappearance of the gemmules was found to be concomitant with the moniliform condition of the dendrites. As GEIER considers the moniliform condition of the dendrites as morbid, due probably to exhaustion or want of nutrition, he would also interpret the disappearance of the gemmules as indicative of morbid processes in the cell.

In a later work GEIER ('02) described the development of the pro

toplasmic processes of the cells of the spinal cord of the rabbit and cat. Besides the ordinary club-shaped gemmules there are others, in the animal just born, in the form of fine filaments or spines which often arise from a small conical projection of the dendrite. In the middle of this filament may sometimes occur a small thickening. The processes of the posterior horn cells have the same type of gemmules but have them in much larger numbers.

In a rabbit two weeks old the processes have become more regular in outline. This change appears only in the basal part of the dendrites and advances with age towards the periphery. The filamentous gemmules do not appear at this stage all along the dendrites as they did in the rabbit one day old, but occur only in the more distal, irregular part of the process. In a rabbit one month old they are found only on the extreme terminal branches of the dendrite. In cats from birth to one month old the same mode of development is followed. But in kittens two months old GEIER found no typical filamentous gemmules upon the anterior horn cells. In place of them were similar processes of much larger size. This, together with the fact that in an animal three days old the filamentous gemmules are two or three times larger than they are at birth, indicates that some at least of these gemmules grow into dendritic branches. GEIER believes that all such gemmules of the young cells are dendrites in the process of development.

This peculiar type of gemmule is not found upon the posterior horn cells, but the more complex forms are very abundant on the dendrites and occur also on the cell body. GEIER concludes that the gemmules are relatively rare upon the anterior horn cells and of a constant form while they are exceedingly numerous and variable in form upon the cells of the posterior horn. Their absence in any case is associated with the moniliform condition of the dendrites.

PATON (oo) interprets the gemmules as artifacts which mark the points at which the fibrillae of the pericellular net enter the cell or come in contact with it. He asserts that the appearance of the gemmules upon the cortical cells of the embryo is synchronous with the appearance of neurofibrils.

That the gemmules mark the point of continuity between the intracellular and the pericellular fibrils is denied by HATAI. He demonstrates the neurosomic net within the gemmule and the neurosomic nature of the axone terminals which compose this pericellular net. But, while the two fibril systems come into intimate touch with each other, especially at the gemmules, they differ sufficiently in structure

and reaction to enable one to determine a boundary line between them.

As to the physiological significance of the gemmules, their finer structure and their reaction to anesthesia as well as the features of their distribution as noted above, would seem to justify us in according to them an important part in the maintenance of the conduction paths within the central system.

Golgi's Pericellular Net.

The

The network of fine fibers which GOLGI first described as surrounding the nerve cells of the central and peripheral system has been treated by subsequent writers as both nervous and non-nervous. more recent work, however, by HELD, BETHE and others demonstrates that this net is continuous with the terminals of medullated axones. HATAI ('03) confirms this position and shows further that the fibrillae of the net are of the same nature as those of the axone. That is to say, they are a modified neurosomic reticulum. We have treated the relation of the net to the nerve cell in the foregoing section upon the gemmules.

HATAI'S observations would seem to be in harmony with HOLMGREN'S ('99) views regarding the pericellular net of the spinal ganglion cell of Lophius. This net lies between the capsule and the cell and between the lamellae of the capsule. It is continuous with fibers which come from other regions of the spinal ganglion and with fibrils which penetrate deeply into the cell. But these fibrils do not become continuous with the protoplasm of the cell into which they penetrate. A light area always separates them from the surrounding protoplasm.

Intracellular Canaliculi.

In 1886 FRITSCH first discovered an intracellular system of vessels in the nerve cell. His observations were made upon certain large cells in the medulla of Lophius. He interpreted the structures as genuinely vascular. HOLMGREN ('99) discovers similar structures in the spinal ganglion cell of Lophius and claims that he is first to confirm the observations of FRITSCH. But during the same year of HOLMGREN'S publication NELIS ('99) introduces "un nouveau detail" in the protoplasmic structure of the nerve cell, which has many features in common with the intracellular vessels of FRITSCH and HOLMGREN. These structures, the "Gefässe" of FRITSCH. the "Kanälchen" of HOLMGREN and "l'état spiremateux" of NELIS, we shall for convenience designate as intracellular canaliculi.