NUCLEAR OSMOSIS AS A FACTOR IN MITOSIS. 149 
drawing out from the cytoplasmic reticulum a series of fine, delicate threads. If we 
now examine closely the kinoplasm which has been thus differentiated, we will find that 
the changed configuration of the cytoplasm is much more sharply defined near the 
nuclear membrane, and as we follow the threads outward to the periphery, they 
eradually fade out and become lost in the reticulum. From this it would seem that 
the tension expressed in these fibrils decreases in proportion with the distance from the 
nuclear membrane. This is obviously beyond actual proof, but if it is true, as appear- 
ances seem to indicate, it may account for the conical shape of the groups or sheaves 
of fibrils, for these clearly show a like tendency to attenuate as they approach the 
periphery. 
A comparison of the serial stages shown in figs. 14 to 21 is sufficiently convincing 
that while there is a marked but gradual decrease in the volume of karyolymph there is 
no evidence whatever that the nuclear membrane breaks down, collapses, or disappears. 
On the contrary, the membrane remains intact throughout all of these stages. As 
indicated in figs. 19, 20, 21, and 22, the nuclear vacuole may not always retain its 
spherical form, and the shape of the membrane may vary accordingly, but even in such 
extreme cases as shown in fig. 22 the contour of the membrane in section may easily 
be followed. It has simply receded with the gradual diffusion of the nuclear sap. 
Now, the nuclear membrane not only remains intact throughout the prophase, but 
| it continues to form the base of the lines of tension expressed in the drawn-out 
threads, for it is in reality a continuation of the same cytoplasmic substance as the 
kinoplasm. It would therefore naturally follow that, as the membrane receded with 
the diminution of the nuclear vacuole, the lines of tension would shift accordingly. 
| Such a shifting does not mean the changing of the threads bodily from one position to 
another. It means the relaxing of the tension along certain threads, which would 
consequently fall back into the form of the original reticulum, and the setting up of 
| new lines of tension, with the drawing out of new threads from the hitherto undiffer- 
| entiated reticulum. In this fashion not only individual threads, but entire sheaves or 
cones may appear to assume different positions. I believe the conditions shown in 
| figs, 20 and 21 represent transition stages in the shifting of the lines of tension, and 
| consequently the apparent shifting in the position of certain cones or sheaves in this 
| way. In both these figures there are portions of the cytoplasm which can be in- 
| terpreted in no other way than transitions between a reticulum and kinoplasmic threads. 
I regard such demonstrations as very important, because I believe they suggest a 
rational explanation not only for the apparent changes in the position of the fibril 
| sheaves, but also for the ultimate resolution of the multipolar into a bipolar arrange- 
| ment. In figs. 17, 18, 23, and 24, it will be seen that the apices of the cones may 
| be a considerable distance apart. It is generally believed that the apices approach 
one another and the cones ultimately coalesce in two groups. Such a movement of 
| the cones bodily seems to me not only improbable but impossible, without some dis- 
| turbance of the cytoplasmic reticulum. In the series of stages shown in figs. 16 to 
