82 
THE VOYAGE OF H.M.S. CHALLENGER. 
enclosing a spherical, darkly stained translucent nucleolus 0’004 mm. in diameter, which 
is connected with the walls of the nucleus by radiating processes. These cells have 
somewhat the appearance of the large elements of nervous ganglia, so much so that a 
distinguished zoologist on seeing them for the first time exclaimed, “ Cells from the 
spinal cord ! ” It is scarcely probable, however, that they are nervous in function ; they 
are “too good to be true,” i.e., we should expect such large and striking cells to exist 
in evident connection or association with nervous tracts, and to occur more usually in 
other and similar sponges. They appear to me to resemble the large amoeboid cells 
which I previously described in Thenea ivallichii (loc. cit., p. 448, pi. xvii. fig. 48), 
but in a different state of extension. Of probable hypotheses as to their nature there are 
three open to us : they may be parasites, or immature ova, or amoeboid wandering cells. 
If parasites, they can only be Protozoa, and we should expect to find other stages 
of their life-history represented, but we do not ; the same is true if we regard them as 
ova, they are so large that we should expect to find some examples in that stage in 
which they present a regular oval outline, but we do not ; if they are wandering cells, 
they are much larger than those cells are usually ; and they do not appear to wander, 
since they lie in distinct cavities, and their pseudopodial extensions are thread-like 
towards their extremities and not lobose, as is usually the case in wandering cells. The 
possibility of their being nervous elements must by no means be disregarded. 
The flagellated chambers (PI. IX. fig. 29) are usually spherical in form, and about 
0'04 mm. in diameter ; they occur in areas frequently bounded on all sides, in thin slices 
of the sponge, by collenchyme, but the ultimate branches of the canals in direct com- 
munication with them are without any special collenchymatous investment, and the 
chambers are truly eurypylous. 
Skeleton . — As in all species of Pcecillastra, the spicules are not united together in 
fibrous tracts, though they sometimes lie parallel to one another in groups. The large 
oxeas tend to run more or less transversely and longitudinally through the sponge, the 
longitudinal spicules diverging obliquely outwards as they proceed upwards. Some lie 
tangentially beneath the skin, others are directed at right angles to it, their distal ends 
either lying below the epithelium, which often rises tent-like over the points, as much as 
O' 15 to 0'25 mm. beyond the general surface, or they perforate it and protrude as 
hispidating spicules. 
The calthrops lie within the choanosome, irregularly disposed, though there is a 
tendency for their actines to lie tangentially to the surface of the adjacent canals. The 
orthotriaenes are disposed with their cladi tangential to the surface, the rhabdomes 
radiating from it at right angles. 
The spirasters are most abundant immediately beneath the external epithelium, but 
are not confined to it. The metasters are generally distributed through the sponge, but 
are rare beneath the outer epithelium. 
