nucleus also occurs in the cells of the adult animal. The shape of the nuclei is somewhat oval, the longer diameter being perpendicular to the long axis of the protoplasmic process. On one side towards the center of the cell, the outline of the nucleus is more or less wavy. In some cases, the wavy outline is not very marked (Fig. 5) but in most cases, it is conspicuous and one is lead to compare it to the pseudopodia-like process of an amoeba (Figs. 1, 2, 4).

The nuclei showing the pseudopodia-like processes have been observed by several investigators; in the egg nuclei of the insects and Coelenterata by KORSCHELT ('89); in the spinning gland cells of a Swedish caterpillar and also in the spinal ganglion cells of the fishes, frogs, etc., by HOLMGREN ('95-'00); in the developing ovum of the Nassa by HOFFMANN ('02); in the nuclei of the ventral horn cells of various vertebrates by KOLSTER ('01); and the same thing is also shown in the illustrations accompanying a large number of papers in which, however, the authors do not describe this interesting phenomenon. Before going on to a further discussion of this appearance, I shall describe more in detail the histological characters of the wavy outline together with the structure of the adjoining part of the cell body which contains the centrosome.

The neuclear membrane which covers the pseudopodia is not completely continuous but is composed of separate portions when seen in thin sections; in other words, the surface of the nuclear membrane towards the cytoplasm is porous. A disappearance or dissolution of the membrane on this side of the nucleus has been observed by HOLMGREN, and PUGNAT ('98), but in the case of the white rat, it is always porous in character. This is clearly shown in Fig. 2. In many cases, however, the , local dissolution of the nuclear membrane is not as conspicuous as in Fig. 2, but the outline appears varicose in structure owing to an accumulation of basophile granules around the pores (Figs. 2, 5).

The nuclear membrane which lies towards the protoplasmic process is of uneven thickness. The thicker portions stain much more deeply than the rest of the membrane with the

basic dyes (Figs. 1, 5). This indicates that the thicker portions contain an accumulation of the nucleoproteid. This statement is supported by the fact that the preparations tested for iron show this area deeply stained (Fig. 2). The accumulation of the nucleoproteid along the part of the nuclear membrane which turns towards the process is a highly interesting phenomenon since it bears on the problem of the cell metabolism. This point will be discussed later on in detail. An accumulation of the nucleoproteid is frequently visible along the outer surface of the nuclear membrane as is shown in Fig. 3.

These pseudopodia-like processes of the nucleus are intimately related to the rays of the centrosome. The centrosome in nerve cells has been described by several investigators, and in the nerve cells of the white rat in both adult and young it has been described by the present writer ('01). Although some investigators deny the existence of the centrosome in the nerve cell, the structure is so definite and so clear that in properly prepared sections, its presence can not be disputed. The centrosome is especially clear in the case of the embryonic cells and every minute feature of the organ may be distinguished. As a rule, the centrosome lies very near the nucleus and in the concavity formed by it (Figs. 1, 2, 3, 4, 5). The centrosome is composed of two minute central corpuscles surrounded by still more minute granules (centrosphere). These granules arrange themselves in straight lines which run from the center towards the periphery radially (astral lines). The astrosphere is clearly distinguished from both surrounding cytoplasm and centrosphere, since it stains very lightly with iron-haematoxylin owing to a lack of the NISSL granules. By overstaining with acid dyes, however, astrosphere stains a more intense red than the surrounding substance. The minute structure of the centrosome in the nerve cells of the white rat has been reported already by the writer ('01) and therefore, with the exception of the astral structure, it need not be further described here. The astral rays which start from the centrospere run radially towards all parts of the cell body. Those rays which runs towards the nucleus extend not only as far as the nucleus but penetrate its

ter as rays.

membrane and become directly continuous with the linin network. This is shown in Figs. 1, 2, 4, 5. As is shown in the figures, the protoplasmic lines pass through the pores of the nuclear membrane and run into the nucleus where they fuse with the linin network. Some of the rays fuse together with the linin network as soon as they enter into the nucleus, but others run quite a distance without losing their original characUltimately, however, they fuse with the network and no rays can be seen near the periphery of the nucleus. The penetration of these rays was verified by careful examinations, many times repeated, all precautions being taken against possible mistake. Thus my observations on the spinal ganglion cells of the white rat show a direct continuation of the cytoplasm and nuclear network by means of the rays of the centrosome. In addition to these observations, HOLMGREN ('99) noticed in adult nerve cells of Lophius piscatorius that the NISSL granules were hung along the astral rays and that these granules could be traced with them into the nucleus. From this fact HOLMGREN Concluded that the NISSL granules are formed by the migration of the chromatin out of the nucleus, and that the granules thus formed are again passed back into the nucleus by means of the rays. He, therefore, regards the rays as a pathway by which NISSL granules re-enter the nucleus. The present writer was unable to find any formed NISSL granules along the line of the rays within the centrosphere. Therefore the returning of the formed NISSL granules into the nucleus was not found in the nerve cells of the white rat. I will return to this point in general discussion.

Distribution of the nuclear material.-At this stage of intrauterine life, the nucleus of the nerve cell contains a large number of the chromatic particles which are scattered through it. The particles vary in size from minute granules to comparatively large bodies. The larger granules do not exceed five in number and may be composed exclusively of basophile subtances. Subsequent to the stage in which basophile and oxyphile granules are distinct, the nucleolus appears. The minute granules as distinguished from large granules, mentioned above, stain

The iron test

with iron-haematoxylin a deep black to a grey. shows in the same granules also varying amounts of the iron. These granules are hung along the linin net as is shown in all the figures. By using BIONDI-EHRLICH'S tricolor stain, these granules stain from a deep blue to a brownish red; thus all the staining methods employed show that these granules are chemically heterogeneous. This gradation in the staining capacity probably indicates gradations in the chemical constitution of the granules which range from a substance rich in nucleic acid to a substance poor in nucleic acid. These granules are found most abundantly along the nuclear membrane, especially at the two poles of the nucleus on the side toward the cytoplasmic process (Figs. 1, 4. 5). It is a striking fact connected with the distribution of the granules that the appearance of the NISSL granules in the cytoplasm is clearly correlated with the accumulation of the granules at the two poles of the nucleus. granules first appear in the neighboring cytoplasm. This fact suggests that the NISSL granules are derived from the nucleus.' From facts similar to those just given, as well as from other evidence, SCOTT ('98-'99) arrived at the conclusion that the NISSL granules are of chromatic origin and are produced by the migration of the basophile granules from the nucleus. The nuclear origin of the nuclein compounds which are seen in the cytoplasm has been maintained by several investigators from observations on glandular tissues as well as on the muscles. My own observations therefore indicate the nuclear origin of the NISSL granules and thus corroborate the observations of these previous investigators, HOLMGREN ('99) on the nerve cells of Lophius piscatorious and ROHDE ('03) on the nerve cells of various vertebrates.

The NISSL

The formation of the NISSL granules in the nerve cells is comparable with the formation of the zymogen granules, muscle fibrils, and yolk granules. This is a highly important and fundamental problem in cellular biology and the subject is fully discussed and presented historically in WILSON'S "Cell in Development and Inheritance," 2nd edition, 1900.

Migration of the nucleoli.—A migration of the nucleolus as well as other nuclear material from the nucleus to the cell body has been reported by several investigators. In the case of the nerve cells, ROHDE ('96-'03) observed the migration of the nucleolus in the case of both lower vertebrates and mammals. In a recent paper on the nerve cells, ROHDE ('03) maintains the migration of the basophile substance from the nucleus, where it exists in solution, into the cytoplasm. LEVI ('96) believes the neurosomes in the cytoplasm to be derived from the nucleus. LEVI showed this relation by the electrical stimu

In the present work, I have noticed also nuclear substance or accessory nucleoli the cytoplasm. This is so often seen in stage of development that it cannot be

laton of nerve cells. the migration of the from the nucleus to the nerve cells at this regarded as an artifact. The migration is always towards the cytoplasmic process, no matter whether this process turns distally or toward the spinal cord. If the dislocation of the large masses or accessory nucleoli were due to mechanical forces, the knife or gravitation, or to other mechanical factors, one would expect to see them moved towards the less resistant side, that is towards uncovered side of the nucleus, as in the case of pathologically altered cells in which the nucleolus or intranuclear masses escape towards the side on which the cytoplasm is least abundant. Again, if such a migration were produced by the knife or gravitation, as HERRICK ('95) showed to be the case in the ovarian cells of the lobster, one would expect to find the displacement always in the same direction in all the cells of a given section, but such is not the case. Moreover, the general appearance within the cells does not suggest a mechanical bursting of the nuclear membrane. For these reasons, I believe that the observations of ROHDE and others are correct and my own observations strongly confirm their statements. The fate of the extruded granules has still to be considered. In order to

1 History of the observations on this subject in the tissue cells is given in detail by MONTGOMERY, T. H.-Comparative Cytological Studies with especial Regard to the Morphology of the Nucleolus, Journ. Morphology, VOL. 15, No. 2, 1898.