ON THE SOFT PARTS IN EUPLECTELLA ASPERGILLUM. 667 
The membranous wall, which is usually very delicate and thin, is furnished 
with smooth-edged roundish pores of different sizes, irregularly arranged, and 
varying very much in number. These form an open communication between 
_ the cavities of the chambers and the duct-like spaces surrounding them, which 
as cleft-like diverticula of the inhalent lacunar system, penetrate everywhere 
between the ciliated chambers, and extend even to the oral edges of the 
chambers, where they end in a somewhat tough and solid membrane, which 
hounds and connects laterally the chamber walls (Pl. XVII. fig. 4). Besides this 
membrane there extends beyond the oral edges of the neighbouring chambers, 
which here and there touch one another and partly adhere, numerous flat 
or linear cords of tissue proceeding from the chamber walls, traversing the 
surrounding spaces, and serving to keep the chambers expanded and in 
position. 
The general arrangement of the chambers is best understood by the 
disposition of the exhalent canal system, as they almost all run into it. 
These exhalent passages begin with digitate ceca of 100 to 200 pw. in 
transverse sections, which open terminally or laterally into wide canals. 
These latter open either directly by a round aperture through the inner 
wall of the sponge-tube, or are again united into still larger canals, which 
open with a circular aperture about 3 mm. wide into the large cavity of the 
sponge. 
It is to be observed that in the whole inner surface of this exhalent canal 
system there stretches a network consisting of flat or thread-shaped bands of 
tissue, in which, besides the long five- or six-rayed spicules, numerous free 
starlets appear, which BowerBank called “trifurcate hexradiate stellate 
spicules,” and which, although in lesser number, are also found in the cords of 
tissue which traverse the inhalent lacunar system. I may here incidentally 
remark that the number of pointed secondary rays proceeding from each of the 
six principal rays of these stars need not necessarily be three, but may be four, 
or even five, but the number is always the same at all the six points of one and 
the same star. 
The inner surface of the large excretory canals coming out from the ridges 
is formed of an extended, almost’ membranous, network, which stretches 
between the terminal crosses of the long-stalked five-rayed spicules, which are 
numerous here, and surround the quadrate meshes. 
The structure of each portion of the tube-wall, which projects externally 
between the ridges in the form of flattened bosses, corresponds essentially to 
this structure of the ridges. As a difference it may, however, be noted that 
the exhalent canals, which are here naturally much shorter, are not united into 
one or two larger principal excretory openings, but open directly into a large 
number of pores close to one another in a quadrate area in the inner side of 
