168 
THE VOYAGE OF H.M.S. CHALLENGER. 
with long bent cladi arise. But it is a rule amongst sponges that bending is a precursor 
to budding or bifurcating ; a bud thus arises at the point of flexure, and a bifurcate cladus 
is the result (Fig. 1e). Thus the various forms of trisenes which characterise this sponge 
may be explained as special modifications produced by the special mechanical conditions 
involved in the structure of the cloacal tube. The disenes with a single bifurcate cladus 
might be regarded as reduced forms derived from dichotrisenes like those of Thenea. Thus 
let the figure (Fig. If) represent the cladome of a dichotrisene seen in plan ; if all the cladi 
except those shown by heavy lines be suppressed we shall obtain the disene of Tethyopsis. 
I by no means deny the possibility of such an origin, but regard it at the same time as 
less probable, for not only do the conjecturally least modified trisenes of the cortex present 
no trace of bifurcation, but in Tribrachium, the nearest ally of this sponge, they are 
similarly absent. (Since writing this I find however that dichotrisenes are sometimes 
present in Tribrachium in the cortex near the base of the cloacal tube, and so far the 
probabilities in favour of a dichotrisene origin of the disenes are increased.) Moreover, 
they are confined in the sponge under consideration to special regions, and are exactly 
adapted to the circumstances under which they there occur ; the point of bifurcation of 
the cladus is invariably situated at the point of rapid flexure of the circumference 
of the excurrent canal, and the relative lengths of the proto- and deuterocladi alter 
according to the distance of this point from the cladal origin. Finally, a whole series of 
gradations can be traced between the usual monsene and the disene with one dichocladus. 
The importance of this contention is obvious, for if the conclusions here reached be held 
correct, we have additional evidence in favour of the origin of similar forms of spicules, 
and by equal reason of other structures, from different ancestors, but under similar 
mechanical conditions. 
Besides cladoxeas, oxeas are present in cloacal tubes of the stage of four, but not 
in that of sixteen (fifteen) ; they lie in the walls of the excurrent tubes longitudinally, 
and where these tubes form the outer wall of the cloacal tube, they lie both 
longitudinally and transversely, parallel to the surface, like those of the incurrent 
conditus. 
The lumen or canal of the excurrent tubes is much narrower than the spicular wall, 
within which it lies more or less concentrically (PL XVIII. figs. 18, 19). The face of 
this canal is lined with epithelium and its sanidasters, and between it and the collenchy- 
matous layer of the spicular wall run connecting strands and fenestrated lamellae of 
collenchyma coated with epithelium, beneath which sanidasters occur, but less densely than 
usual ; on the other hand additional spicular elements, the orthodragmas, make their 
appearance within the collenchyma. The collenchymatous strands and lamella are so 
disposed, radiating between the inner and outer walls of the canals, curving, branching 
and anastomosing, as to subdivide the intervening space into widely communicating more 
or less vesicular lacunae. The collenchyma of the spicular wall also is fenestrated, so 
