Gland-Cells of Internal Secretion in Spinal Cord of Skates. 17 
scopic size, this behavior (phagocytosis) is shared with the polymorphonuclear 
ele nents of the blood. But towards ultramicroscopic particles the macro- 
phages specifically drink them in.” 
By making use of this specific reaction, Evans, Bowman, and Winter- 
nitz have made an experimental study of the histogenesis of the miliary 
tubercle in vitally stained rabbits. They found that the giant tubercle 
cell is produced by a fusion of many endothelial cells, the stimulus of 
the tubercle bacilli being enough to cause endothelial cells to separate 
from the vessel-wall, form a syncytium, and begin their role of phago- 
cytosis of the bacilli. 
Might not the large, irregular cells of the skate have arisen in response 
to some stimulus, perhaps the stimulus caused by the presence of the 
granules, just as in response to the presence of the tubercle bacilli the 
giant tubercle cells arise? Morphologically the cells of the skate may 
be compared to the giant tubercle cells in several particulars. Both 
have a large amount of nuclear material, and the nucleus of the skate 
cells is often of such a distributed character that it may have arisen 
from more than one original cell, at least in some cases. As has been 
pointed out, the cytoplasm of the skate cells is occasionally distorted 
with vacuoles, precipitate, and granules. These vacuoles and precipi- 
tate might well be interpreted as granules in the process of ingestion and 
liquefaction. The fact that there is no definite cell-membrane would 
also make it easy for ingestion to occur. The conditions of surface ten- 
sion seem to be like those of the macrophages. There is, moreover, a 
large blood-supply to these cells, a condition which would fit in well 
with a function of phagocytosis. 
There is, furthermore, in the skate some evidence in favor of cell 
movement. The cell is quite variable in its position in the anterior 
horn. Sometimes the bulk of the cell is close to the central canal, with 
a long process reaching almost to the ventral artery. In other cases 
the cell is located much nearer the lateral edge of the anterior horn, 
with processes that extend laterally almost to the membranes of the 
spinal cord. The long processes might be regarded as pseudopodia 
capable of a very limited amount of motion. Figure 24, taken from a 
skate stimulated electrically, shows how the long medial ventral pro- 
cess has crowded and pushed aside the central canal cells so that Reiss- 
ner’s fiber, which normally is located in the middle of the central canal, 
has become entirely surrounded by the cell cytoplasm at one point. 
Another remarkable condition is occasionally met with in a few cases. 
Sometimes a part of a process, together with some of the nuclear 
material, is found inside the central canal (fig. 42). It is difficult to 
explain how this state of affairs is brought about. It seems certain 
that the cell processes must be capable of at least a limited amount of 
movement. In this same figure a study of the appearance of the 
nucleus and even of the chromatic granules indicates that a probable 
