fibrillae which most authors describe, VAN GEHUCHTEN and NELIS find a delicate, granular striation which they fail to connect certainly with the cytoplasmic net. The granules of the striae are so arranged as to give in some cases a cross-striated effect.

PATON ('00), in his studies of the neurofibrils of the cerebral cortex of the pig, finds that the fibrillae of the axone run independently of each other into the cell body, where they enter into a most intricate net-work. He holds that the fibrillae of this net are continuous with those of a pericellular reticulum, which he interpets as identical with GOLGI's pericellular net.

PATON believes that the fibrils are very quickly affected by postmortem granular disintegration which begins at the center of the cell. He believes this fact may account for the view of certain authors that fibrils exist only in in the apical process of the cell.

PRENTISS has recently published in this Journal his latest results upon the neurofibrillar structures in Hirudo and Asticus. As to the general arrangement of the fibrils in the cell body he confirms APÁTHY'S position. He concludes, however, that "Neither in vertebrates nor in crustacea do the neurofibrillae of the nerve cells show any marked correlation in size and function." He believes that the differences in size which APÁTHY observed are due to incomplete impregnation of the fibrillae and perhaps to the cleaving together of smaller fibrillae in the cell process. As to the relation of the fibril to the cell and its processes PRENTISS supports BETHE's view that fibrils may enter one process and leave by another or by a collateral without coming into relation with the perikaryon itself. In Hirudo he finds a very limited fibrillar network in the neuropil. Such nets are more extensive in Astacus but they are not diffuse in their relation. They put relatively few fibrillae into communication with each other. PRENTISS Considers that his preparations tend to confirm BETHE's theory that the cells are not the centers of nervous activity, and that the fibrillae are continuous from cell to cell.

PUGNAT, in his recent review on the finer structure of the nerve cells, comes to the conclusion that the formed substance of the cytoplasm is a net, in some cases of fibrillae, in others of trabeculae, and that this net is continuous with the fibrillae of the processes of the cell. The relation which this net holds to the life aud function of the cell, he thinks, can best be explained upon BARD's theory of "derived substance." Although PUGNAT does not accept the sharp chemical and physical distinction between the protoplasm and the derived substance as BARD proposed, he believes that fibrillae are a product of the

protoplasm, that they are the conducting element and are therefore. the seat of katabolic processes, while the protoplasm and nucleus are the seat of the anabolic process. How the fibrillae are repaired by the protoplasm, PUGNAT says, we are absolutely ignorant. Through this theory of derived substance PUGNAT attempts to bring the neurone theory into harmony with the results of APÁTHY and BETHE and their followers. He thinks that whether the fibrillae as derived products. of the cell are continuous or not from cell to cell, the nerve cell itself may be considered as an anatomically distinct unit. He would, in other words, place the neurofibrillae in the same category with the fibrillae of the muscle or connective tissue cell. However, before such a compromise of the neurone theory is conceded HATAI's methods, by which he has received such brilliant pictures of the finer structures of the nerve cells and of the axone terminals, must be given a thorough test. The results which he has recently published are remarkably convincing.

In the afferent neurones of the electric lobe of Torpedo HATAI ('01) demonstrates fine fibrillae in enormous numbers, crowding the cell processes and the perikaryon. By serial sections through one of

these large cells he shows the behavior of the fibrillae within the cell body. Upon emerging from the process into the cell body they diverge in clusters. Some sweep around the nucleus to form here a dense net, others pass to various processes of the cell in such a manner that there is direct fibrillar connection established between each dendrite and every other dendrite and between the axone and all the dendrites. By this coursing of the fibrillae from the dendrites into the axone there is a beautiful spiral configuration given to the ground substance of the cone of origin. Superficially HATAI's figures in this case have a striking resemblance to the familiar drawing of the fibrillar elements in the nerve cell by Max Shultze.

HATAI, however, makes an important step in advance by demonstrating that these same fibrillae in the electric neurones of Torpedo can be resolved into rows of neurosomes. Furthermore, he asserts that these neurosomic fibrils in reality are a modified reticulum. It is only in thicker sections and under lower magnification that the structure gives the fibrillar appearance.

HATAI has demonstrated the reticular structure more exhaustively in the nerve cells of the white rat. He has studied these cells ('03) by the methods of BETHE and DOGIEL, but finds no such fibrillar structures as they describe. On the other hand, he demonstrates by other methods a neurosomic reticulum which is modified into a pseudo

fibrillar structure. The fibrillae are made up of rows of neurosomes connected by slender protoplasmic filaments. Generally, around the periphery of the cell body the meshes are larger than in other regions. Around the nucleus and in the cone of origin they become more narrow and very much elongated. A more pronounced modification of this character is found in the axone so that the reticular condition is difficult to see. In the axone, however, the neurosomes stain brighter than elsewhere. Especially is this true of the neurosomes of the axone terminals, where they are also larger than in other parts of the cell. The dendrites contain a relatively small amount of the ground substance and the neurosomes stain more faintly than in the axone. This fact enables HATAI to differentiate the finest dendritic branches from the contiguous terminals of the axone. Even in the gemmules he demonstrates the neurosomic reticulum as continuous with that of the rest of the cell, but he finds that its neurosomes differ both in size and staining reaction from those of the axone terminals. He concludes that there is no continuity of the so-called fibrillar structures between the nerve cells.

Now HATAI finds that a number of these neurosomic filaments may mat together into thicker strands and that several of these strands in some cases form a complicated network around the nucleus. Such a network does not appear in the cone of origin or in the periphery of the cell body. He homologizes this network with the intracellular anastomosing fibrils of APÁTHY and also with the endocellular network of GOLGI. Thus HATAI resolves the neurofibrils of APÁTHY into the protoplasmic elements of the cell and denies that they pass continuously from cell to cell. The diffuse nets in the neuropil and pericellular nets also may be resolved into the reticulum of axone terminals which would be strictly protoplasmic and not extra-cellular, derived substance.

Moniliform Condition of the Dendrites.

Since DOGIEL'S discovery of the dendritic varicosities in the cells of the retina this feature of the nerve cell has held an important place in neuro-cytology. Among the most comprehensive contributions upon the subject are those of SOUKHANOFF. This author made four experments by ligature of the abdominal aorta, three upon guinea pigs which died from the experiment in from one-half to twenty-four hours, and one upon a rabbit which was killed after twenty-four hours. In these experiments he found that the diminished blood supply had produced a very rapid modification in the nerve cell, varying directly in

intensity with the duration of the anemia. The modifications of least intensity consisted in the appearance of swellings along the dendrite which made its contour irregular. As this condition became more intensified the swollen regions of the dendrites became fusiform and then spherical. During this process the connecting regions between the enlargements of the dendrites became more attenuated till a final beaded or moniliform condition resulted. This modification he considered pathological.

In another article (98) upon the modification of the dendrites under the influence of narcotics SOUKHANOFF reports in considerable detail the results of nine experiments upon mice, rabbits and guinea pigs. These animals had been subjected to the vapor of ether, chloroform or alcohol for various periods of time. Two of the experiments were made by injection of trional. As a result of these experiments and of a critical review of the work of other investigators, SoUKHANOFF draws the following conclusions: (1) The moniliform condition may occur in certain dendrites under normal conditions, a conclusion reached by practically all authors; (2) under the influence of ether, chloroform or alcohol there is not a very appreciable increase in the moniliform condition; (3) injection of trional causes a moniliform condition of nearly all the dendrites of the cerebral cortex; (4) this change is accompanied by a more or less complete loss of the gemmules; (5) a loss in weight in guinea pigs subjected to trional injections can be attributed only to profound nutritive derangement. This suggests that the moniliform condition of dendrites may be "une degenerescence sui generis" manifested whenever nutrition is severely affected.

A third contribution by SOUKHANOFF ('98) deals with the varicose atrophy of the dendrites in the cerebral cortex under pathological conditions. His pathological studies were checked by examination of a normal specimen. From an exhaustive study of the latter case he concluded that the moniliform condition of the dendrites is occasionally found under normal conditions, but that it is very slight and not to be compared with that found in cases of poisoning. From nine experiments upon guinea pigs which had been subjected to arsenic poison in varying degrees for periods of from six to thirty-three days, he concludes that in acute and subacute poisoning by arsenic there occurs a moniliform degeneration of the dendrites of the cerebral cortex. This is very slight in some eases, but very marked in others. This difference cannot be accounted for by difference in the duration of the poisoning. Its cause is more likely to be found in individual differences in the specimens as regards resistive power and general health.

SOUKHANOFF'S experiments include also poisoning by rabies and tuberculin and a case of thyroidectomy. In a rabbit infected with rabies during eighteen days and fatally, the varicose degeneration of the dendrites was slight. The experiment upon the influence of tuberculin was performed upon a dog that had some time previously been inoculated with tuberculin for immunization. It was then inoculated with an emulsion of the Bacillus of KOCH. From his study of this case SOUKHANOFF concludes that the varicose condition occurs in many dendrites of the cerebral cortex but that this is due to derangement in nutrition. In the case of thyroidectomy, upon a dog which lived a week after extirpation of the thyroid gland, profound changes were observed in both large and small dendrites of the cerebral cortex. This also SOUKHANOFF attributes to nutritive derangements which are known to follow thyroidectomy.

In general SOUKHANOFF concludes that although some dendrites in the normal nervous system are found in the varicose condition, the large numbers found changed under certain conditions are indicative of a morbid process; and that all poisonings which cause profound derangement in general nutrition cause a pronounced varicose degeneration of the dendrites.

The condition of the dendrites in the spinal cord of the rabbit has been studied by SOUKHANOFF and CZARNIECKI ('02). The spinal cord of two specimens, which were killed quickly with chloroform were treated by the GOLGI method. To bring about quicker penetraaion of the fluid incisions were made along the cord. The authors find that the cells of the cord show very pronounced differences in the form of the dendrites. The cells of the anterior horns differ strikingly in this respect from all other cells in the cord. Some of the dendrites of the anterior horn cells have comparatively regular contour, others are in a distinct varicose condition while the majority are in a condition intermediate between the two extremes. On the other hand, the dendrites of many of the small cells of the cord were found in a very marked varicose condition.

GEIER ('01) reports a series of experiments which are especially noteworthy for their thoroughness. They were made upon mammals and birds killed with chloroform or ether, and involved an examination of at least an entire cerebral hemisphere in each case. He describes each case in detail and draws the general conclusion that anesthesia by chloroform or ether does not of itself cause a moniliform condition of the dendrites, but that such a condition must be considered as the expression of a morbid or fatigued state of the cell. He holds also that the