( 1144 ) 
is in most cases narrower at the top than lower down, and in many 
cases it is clearly seen, that the wide central canal towards the 
top of the process branches into a number of much narrower 
ones. Finally I found that the wall of the canal is not smooth, but 
shows several transverse rugae~ Such a canal, according to the 
theory of PEKELHARING and myself, would be very little adapted to 
act as a cloaca. In view of all this I thought it necessary to recon- 
struct the canal system by means of series of sections. The result 
has been, that I found the positively excurrent canals not to com- 
municate with the central cavity, but exactly the contrary: the 
incurrent canals communicate with it. Consequently, the central 
canal is not a cloaca but an incurrent reservoir. Numerous minute 
apertures (stomata) lead into a system of canals, which open in the 
central cavity, which also may communicate directly with the sur- 
rounding medium by means of a comparatively wide opening at 
the top of a process. From this reservoir canals start and enter, 
ramifying, the parenchyma; they ultimately communicate with the 
flagellated chambers or mastichorions as I have called them. As the 
‘anal system is eurypulous, the mastichorions open with a wide 
apopyle in the excurrent lacunae or canals, which finally likewise 
open at the sponge surface with small apertures, procts. 
There is still another interesting feature in the canal system of 
Spirastrella purpurea. It is generally accepted that the principium 
movens for the watercurrent is to be songht for exclusively in the 
flagella of the choanocytes. In our sponge a second factor appears: 
‘in the wall of the larger canals, more especially of the central 
canal, numerous undoubtful muscle cells occur; they are situated in 
concentrical and in radial bundles. In the rugae we find the former 
in maximal dilatation. If on the other hand the concentric muscle 
cells contract the rugae are stretched out to a kind of membraneous 
diaphragms with the result that the lumen of the canal becomes 
considerably narrower. In some specimens this is in fact the case. 
Suppose that the terminal aperture has first closed, water must 
forcibly be pressed into the canals which lead to the inner parts 
and in this way the current-producing power of the flagella helped. 
In this connection it is worth while to remark that the total quantity 
of mastichorions is comparatively small. 
The parenchyma is almost entirely composed of a remarkable 
tissue to which I more than once have drawn attention and which 
can best be compared with lymphoid connective tissue. It consists 
of flat cells with delicate membraneous processes, forming together 
a sort of syncytium. At least I could not distinguish cell limits. In 
