136 
R. A. HARPER 
or cylinder of spore-plasm tends to be rounded off so that the clefts 
are wider at the surface than they are deeper down in the sporange. 
The furrows which have originated from the capillitial openings 
curve and branch in the peripheral region in such a fashion that they 
come to lie in a plane more or less tangential to the surface of the 
sporange and in this way begin to cut off blocks from the ends of the 
cylinders of spore-plasm formed in the initial cleavage stages. These 
tangential furrows begin to be formed before the radially placed 
cylinders and prisms have been entirely separated from each other. 
The cleavage is in reality a progressive process working from the 
surface of the spore sack inward, both by radially and by more or 
less tangentially placed furrows. An early appearance of these 
tangential furrows is shown in figure 9, and a slightly later stage in 
which the first peripheral blocks are almost or entirely cut off is 
shown in figure 1 1 at a and h. 
These same stages in the cleavage process can be observed also in 
the inner layer of the spore-plasm next to the columella cavity (fig. 11). 
But in this inner layer the cleavage is always a little less advanced 
than in the outer layer. The whole process parallels apparently a 
continuation of the extrusion of cell sap which began at an 
earlier stage in the formation of the sap cavities around the capillitial 
threads (figs. 3 and 4), and the extrusion of sap is apparently most 
rapid where evaporation can go on most rapidly. We cannot, of 
course, regard the process* of cutting up the spore-plasm as a mere 
matter of the drying out of the protoplasm due to evaporation. The 
clefts and furrows may be more or less filled with liquid at all times, 
and there is thus only a minor difference possible between the super- 
ficial and the central portions of the spore-plasm. Still, loss of water 
by way of the columella and stipe, which is filled with coarse con- 
cretions (see fig. 18), is doubtless slower than from the outer surface 
of the sporange and we find that cleavage is also slower next the 
columella than on the peripheral surface of the spore-plasm. There 
seems to be an obvious parallel between facility for water loss and 
rate of cleavage, but this in no wise excludes the possibility that 
chemical changes in the plasma membrane also favor the active 
extrusion of moisture through all these cleavage stages. 
At a stage when the shrinkage about the capillitial threads has 
led to the appearance of the radial clefts shown in figures 6-10, the 
nuclei show quite commonly a tendency to be arranged in series 
