branches supplying the palate. Furthermore the endings of the medullated fibers in the sensory spots of the epithelium were not demonstrated, endings which, under normal conditions, are almost always stained.

It is clear from such preparations that isolation from the central nervous system was completely accomplished by the operation. For the medullated fibers of the palatine nerves were not only wholly degenerated, but also stimulation of the iso: lated region failed to call forth any response. According to BETHE the myelin substance does not begin to undergo degeneration until some time after the axis cylinders have lost their staining properties, and it follows that the latter have been degenerate for a considerable period. But is this the case with the non-medullated fibers of the peripheral networks which BETHE and others have asserted to be directly connected with true nerve cells?

In every degeneration preparation of the palate methylene blue failed to stain the fine network of the sensory fibrillae which I have described and shown in figure 2. This result was to be expected, as the network consists only of non-medullated fibrillae, all branches of medullated fibers, and goes to show that the degeneration of the latter is complete. But in both cases in which the palatine nerves had been isolated for five weeks the subepithelial and perivascular networks were beautifully stained. Neither fibers nor cells showed the least trace of degenerative changes. In regions where the stain was well differentiated the neurofibrillae could be easily made out, proving that these elements were in a normal condition, and that the staining of the fibers was not alone due to the impregnation of the perifibrillar substance.

A portion of the subepithelium network from one of the prepations of series 3 is shown in figure 11. It will be noted that a number of cells are grouped together at the points a and a'. The spot where such a group of cells occurs is in every case immediately beneath a sensory organ, and the fibers radiate out from these regions. Certain fibrils pass from the cells and apparently end in the sensory organs. I was not able to settle

this point definitely but there are other fibrils which end in the epithelium between the sensory spots. From this it is evident that a part at least of the fibrillae of this network are sensory in function. And it is not clear to me why the cells should be grouped together beneath the sensory organs unless they have some connection with them.

Fig. 11. A portion of a subepithelial network from a degeneration preparation (period of degeneration five weeks); the cells and fibers apparently show normal structure, and were normally stained. a, a', groups of cells beneath sensory organs; b, b', degenerated medullated fibers of which is connected with the network; c, a medullated fiber continuous with the network and showing normal structure for a short distance at its distal end. X 225.

It has already been shown that certain of the fibers of the subepithelium network are directly continuous with medullated nerves. In the figures two such cases are seen; at times such fibers could be traced a considerable distance along the degenerate nerves of the medullated plexus. These fibers showed all the normal characteristics of medullated nerves, but were never observed in the large nerve trunks. Other medullated fibers connected with the subepithelial network showed degenerative changes up to the point where they lost their myelin sheaths. As far as observed the only normal medullated fibers to be found in these degeneration preparations, were those connected with the subepithelial network. This network was found stained in all parts of the palate. The perivascular networks were not so completely demonstrated, but this was probably due to incomplete staining rather than to the degeneration of the networks about the vessels. For these networks are never completely stained in normal preparations. When stained in the degeneration preparations, neither fibers nor cells of the perivascular network show histological differences when compared with normal preparations (Fig. 12). As far as structure and staining qualities go, the one cannot be distinguished from the other.

Fig. 12. A portion of a nerve network about the walls of a small vessel; from the same preparation as figure 11. X 360.

These series of degeneration preparations prove therefore that the peripheral networks of fibers and cells will retain their normal structure after five weeks of isolation from the central ner

vous system. The objection may be raised that the period which elapsed was not long enough, that the degeneration was simply incomplete, that the distal endings of the fibers form the networks and that these were still in their normal condition. But

we have already seen that the distal endings of most of the medullated nerves, the terminal fibrillae in the sensory organs, have completely degenerated. Why do the fibers of the networks alone fail to show similar changes?

It has been shown in the first part of this paper that the fibers of these networks are nerve fibers. It cannot be maintained therefore that they do not degenerate because they are non-nervous structures. The only explanation of their immunity from the degenerative changes which affect the isolated medullated nerves is that the cells of the network exert upon them a distinct trophic influence. The cells are then something more

than sheath cells or connective tissue cells as BARDEEN and others have asserted. The medullated fibers of the palate are well supplied with sheath cells but this does not prevent their degeneration when separated from their ganglion cells. We can only conclude that the cells of these networks are true nerve cells in that the integrity of the fibers is dependent upon them. This is in strict agreement with the conditions which we find in the nervous system of the lower animals, and substantiates the conclusions of BETHE and LEONTOWITSCH

As to whether these networks of nerve cells and fibers will retain their integrity indefinitely when severed from all connection with the central nervous system, we do not know at present. Experiments are in progress to determine whether they also will degenerate in the course of a few months, or whether they possess the power of regenerating new fibers. BETHE (:03) maintains that the sheath nuclei are modified nerve cells and still retain their primitive function of neurogenesis. It might be expected that these peripheral nerve cells possess a similar function. Our present data are not sufficient however to warrant the assumption of LEONTOWITSCH that there is a constant process of physiological regeneration going on in the skin, and that the subepithelial network is transformed into the

peripheral portion of a neurite from a sensory ganglion cell. It seems to me rather that in these networks of nerve cells and fibers we have to do with primitive nervous structures more or less independent of the central nervous system, structures which, as BETHE points out, correspond to the diffuse nervous system of many invertebrates, and which are connected, on the one hand with the integument, and on the other hand with the nonstriated musculature.

SUMMARY.

I. The palatine branch of the seventh cranial nerve forms a plexus of medullated fibers in the palate of the frog; from this plexus fibers pass, to end by branching in the sensory organs of the epithelium.

2.

The innervation of the sensory organs of the palate is not as diagrammatic as has been asserted; a diffuse network of neurofibrillae connects different sensory neurones, and puts the sensory organs into communication.

3. A network of cells and non-medullated fibers extends throughout the deeper layers of the palate and forms a close meshwork about the walls of the vessels.

4. Immediately beneath the epithelium is found another network of cells and fibers; sensory fibrils from it end in the epithelium, and it is also connected with the perivascular network.

5. The fibers of the networks are nervous structures for (a) they are not demonstrated by specific stains for elastic and connective tissue; (b) they are composed of neurofibrillae; (c) they are often directly continuous with medullated nerves.

6. Neurofibrillae are present in the cells of the networks, but most of them pass through without forming a basketwork about the nucleus.

7. When the nerves of the palate are isolated from their ganglion cells the medullated fibers which end in the epithelium degenerate at the expiration of 25 to 35 days; the myelin sheaths disintegrate, and the axis cylinders fail to stain.

8. Under the same conditions both the cells and fibers