HOLMGREN ('99, 'oo) asserts that the canaliculus is continuou s with the pericellular lymph space and that it is accompanied by trabeculae of the connective tissue capsule of the cell. Along with it occur

also the nuclei of its membranous walls and connective tissue nuclei which have migrated in from the capsule of the cell, while corpuscles of the lymph circulate through the canaliculus.

KOLSTER ('00) observes canaliculi also in the nerve cell of the spinal cord and spinal ganglia of Petromyzon. He has demonstrated them successfully in unstained sections of cells fixed in osmic acid. He traces them from the periphery of the cell into the deepest part where they seem to be continuous with a perinuclear space of the same nature. A slight invagination of the endothelial capsule may occur where the canaliculus enters the cell and free nuclei are found in the protoplasm which are interpreted as nuclei of the capsule; but KOLSTER, even with the technique employed by HOLMGREN, fails to find a nucleated membranous wall surrounding the lumen of the canaliculus. It is bordered simply by the granular protoplasm of the cell. a later work, HOLMGREN ('00) confirms his first observations by demonstrations of the canaliculus in the nerve cells of mammals and birds with all the features originally described for Lophius. He further shows that electrical stimulation of the nerve cell causes an expansion of the canaliculi.

Still, in

HOLMGREN'S interpretation of the canaliculi receives positive support from the observations of PUGNAT YO1) upon the embryological development of the canaliculus in the nerve cell of the chick. PUGNAT finds that the canaliculi appear first in the outer zone of the spinal ganglion cell on the eleventh day. By the fifteenth day they have reached the central zone. These canaliculi, according to PUGNAT also, have membranous walls and are continuous with extracellular vessels of the same nature.

Canaliculi have been recognized also by BOCHENEK in the largest nerve cells of Helix. As to the structural and topographical features of the system, BOCHENEK agrees in the main with HOLMGREN. He finds the connective tissue fibrillae and cells very conspicuous in the protoplasm of the nerve cell and even invading the basal portion of the axone. He explains the structure as a simple invagination of the capsule into the body of the nerve cell or as clefts in the cell. His figures and descriptions would not give one the idea of a clearly defined membranous wall about these clefts, yet he says "Si, dans cet exposé des faits, nous sommes en complet accord avec le travail de HOLM

GREN,

nous ne pouvons pourtout pas souscrire a ses déductions théoretíques."

The theoretical deduction of HOLMGREN to which BOCHENEK cannot subscribe is regarding the significance of the canaliculus in Helix. HOLMGREN holds that the nerve cells are poorly or richly supplied with canaliculi according to the functional condition. BOCHENEK has observed, however, that the canaliculi of Helix are equally developed in winter and in summer; that is, during the periods of activity and inactivity. He believes, therefore, that they are constant features and are to be explained as an adaptation of a large cell for increased absorbing surface.

The "nouveau détail" which NELIS ('99) calls "l'état spiremateux" is in the form of a spireme-shaped or much coiled, continuous band. In material hardened in GILSON's fluid or 5% formalin and stained in HEIDENHAIN'S iron haematoxylin and eosin or erythrosin these bands appear uncolored and amorphous, and marked off from the colored ground substance by regular, parallel lines. The position of the spireme varies in position and extent in different cells. It may lie near the periphery of the cell, near the center or in close relation to the nucleus. In some cases one of its borders is indistinguishable from the nuclear membrane; although the author considers that the two elements do not in reality coincide.

NELIS finds the spireme in the plexiform ganglion of the dog, and, less conspicuously, in the superior cervical ganglion. It is present in the pyramidal cells of the cerebral cortex and in the spinal ganglion cells also. He considers that this is a normal structure, but that under certain pathological conditions it may become much more extensively developed and more easily demonstrated. As to its significance, NELIS is undecided, but he is inclined to interpret it as a protoplasmic element of the cell.

DE BUCK and DE MOOR ('99), in a work upon the lesion of the nerve cells accompanying experimental tetanus in the guinea pig, find a system of vacuoles which they identify with the spireme of NELIS. These vacuoles may approach very near the border of the cell, but are always separated from it by a thin layer of protoplasm. VAN GEHUCHTEN and NELIS ('00) have observed the structure also under pathological conditions and conclude that it is accentuated by arsenic poisoning, tetanus, etc.

Among these various descriptions of intracellular vessels it is quite easy to identify the type described by KOLSTER with the canaliculus of HOLMGREN, by assuming that KOLSTER has failed to demonstrate the

It only re

membranous walls of the vessel. Furthermore, the general form and position of the spireme of NELIS harmonizes well with the canaliculus of HOLMGREN. Though NELIS does not find nucleated membranous walls, he does find parallel walls with a sharp contour. mains to demonstrate the continuity of the structure with extracellular lymph vessels in order to identify it with the canaliculus of HOLMGREN. But as to the interpretation of BOCHENEK'S clefts we find greater difficulty. He has compared his preparations with the original preparations of the spireme made by NELIS and has decided that the two structures are totally different. He expresses himself as surprised, also, that HOLMGREN should homologize the canaliculi of Helix with the spireme. This suggests that HOLMGREN and BOCHENEK may be dealing with altogether different structures. It may be that BOCHENEK has not seen the genuine canaliculus with membranous walls, but that he has demonstrated only deep clefts in the cell which are filled with the capsular tissue.

The Nucleus.

HOLMGREN (99) finds that the size of the nucleus in the spinal ganglia of Lophius varies with the size of the cell. The linin net and chromatic granules, which stain red with toluidin blue and erythrosin, are massed thickly around the nucleolus and radial branches stream out from this to the nuclear membrane which is acidophile also. On the side of the nucleus which lies nearest the center of the cell the contour may become flattened or even indented by a mass of tigroid substance which accumulates at this point. The nuclear membrane of this region thickens and changes its reaction to the stain, for it now stains dark blue in toluidin-erythrosin, deep black in iron haematoxylin and green with the triple BIONDI stain. Simultaneously with this change in the nuclear membrane there come changes in the tigroid mass, which is heaped upon the nucleus in this region. The large tigroid bodies, which, earlier in the process, were packed closely together here, have become resolved into small granular heaps or scattered granules suspended in a homogeneous or purely granular ground substance. The latter stains blue with toluidin-erythrosin, gray or perhaps black with iron haematoxylin and red with the triple acid stain. Following this condition the thickened part of the nuclear membrane disappears either throughout the entire extent or at intervals. Through this cleft the cytoplasm becomes continuous with the karyoplasm. The union takes place through protoplasmic bands (Züge) which radiate from the centrosome into the nuclear substance and become continuous

with the linin net. The part of the linin net which enters into the radiating bands contains granules which stain a darker red than the other acidophile elements and black with iron haematoxylin. In the radiating bands beyond the limits of the nucleus there occur granules which stain black also in iron haematoxylin but dark blue in toluidinerythrosin. They are sharply differentiated from the tigroid elements by their reaction to DELAFIELD's haematoxylin and the triacid stain. In the former they stain blue-black, while the tigroid elements stain faint blue; in the latter, they appear green while the tigroid elements appear red. This reaction places them in the category of basichromatin and they must be considered as such elements which have migrated out from the nucleus. They become basophile during their migration from the nucleus and represent the nuclear chromatin.

While these changes are going on the nucleus enlarges and migrates towards the periphery of the cell, till, in some cases, a mere film of protoplasm separates it from the cell membrane. In the meantime, also, the acidophile granules of the nucleus have increased greatly in number and during the process also the nucleolus sends off fragments into the cytoplasm.

KOLSTER ('00) also contributes interesting facts upon the general morphology of the nucleus. In a large series of preparations he finds that only a part of the nuclear border has a regular outline and limiting membrane. As studied in serial sections the nuclei, practically without exception, lose their membrane in a particular region and the karyoplasm pushes out in pseudopodia-like processes into the cytoplasm. This appears to the author to be a constant feature and not due to any peculiar physiological condition. He interprets it as concerned in the nutrition of the nucleus. But, compared with HOLMGREN'S work, KOLSTER'S results are noteworthy from the fact that the side of the nucleus which is marked by this irregular contour may lie opposite the centrosome while HOLMGREN finds it directed toward the centrosome.

Irregularities in the border of the nucleus have been observed also by BOCHENEK ('01) in certain large nerve cells of Helix.

The idea that the nuclear membrane breaks down and that there is an interchange of formed substance between the nucleus and the cytoplasm, especially as described by HOLMGREN, is opposed by SCOTT (99). He believes this appearance is due to the action of the knife in cutting. However, it hardly seems probable that both HOLMGREN and KOLSTER should be utterly deceived in this manner. Moreover, it does not seem that SCOTT's fundamental thesis or his results

exclude the possibility of such ruptures in the nuclear membrane as HOLMGREN and KOLSTER describe.

The Nucleolus.

The behavior of the nucleolus in the spinal ganglion cells of Lophius as described by HOLMGREN ('99) has already been mentioned in the section relating to the nucleus. Regarding the finer structure and chemical nature of the nucleolus the works of KOLSTER, HATAI and SCOTT are noteworthy.

In the spinal ganglion cell and the FREUD' schen cells of Petromyzon, according to KOLSTER, the nucleolus appears homogeneous only in preparations fixed in sublimate and stained in iron haematoxylin. In all other methods which the author used he found two rather sharply differentiated zones, a dark, relatively large central portion surrounded by a shell of substance which stains more faintly than the central body. In some preparations there appeared a granular, intermediate zone between the central core and outer shell. KOLSTER is inclined to interpret the nucleolus, not as a solid mass, but as a vacuole with peculiar liquid contents. The reaction of the contents to reagents accounts for the structural features which appear by different methods. PUGNAT, on the other hand, considers that the central, granular appearance of the nucleolus is due to the presence of formed bodies.

SCOTT (99) and HATAI ('03) both demonstrate an oxyphile center with basophile peripheral zone in the nucleolus. Only the basophile part represents the chromatin, while the oxyphile center is the true nucleolar substance..

nerve cell during mitosis While in the adult nerve

In

The behaviour of the nucleolus of the has been carefully worked out by HATAI. cell the larger part of the nucleolus stains blue with toluidin erythrosin, in the germinal cells the entire nucleolus stains a deep red. the later telophase stage of mitosis the nuclear substance, which has become closely massed around the chromosomes, dissolves and accumulates again in small spherical masses which collect near the center of the nucleus. Each granule of this group sends out a process from either pole. These processes from the various granules anastomose to form a net with granules at the angles of the meshes. The linin and the basophile nuclear substance now collect around this group of acidophile granules to form the nucleolus of the adult nerve cell. But if the cell is to divide again, the linin which has accumulated around the nucleolus breaks up in the early prophase and the nucleolar granules separate. As the spireme forms they collect upon