Ixvi 
THE VOYAGE OF H.M.S. CHALLENGEE. 
1, Microxea. — A minute oxea which may be smooth (Fig. XIL, 77) or spined (Fig. XIL, 
y), centrotylote (Fig. XII., Q oi" iiol- 
2. Microstrongyle. — A minute strongyle, smooth, or spined, centrotylote or not 
(Fig. XII., y). This spicule is the “ bacillus ” of Carter. The centrotyle in both forms of 
microrabd represents the centrum of the aster from which it has been derived. These 
spicules are of polyphyletic origin, an aster in one genus of sponge being represented 
by a microrabd derived from it in another closely related genus. 
PhYLOGENY op the ScLERES op the DEMOSPONGIiE. 
The megascleres in their ontogenetic development commence as very minute spicules, 
which if their growth were arrested would be indistinguishable from microscleres ; 
the leading forms of the megascleres are repeated by the microscleres ; and in the 
simplest Tetractinellid sponges the skeleton consists wholly of microscleres, without any 
admixture of megascleres. On these and d priori grounds we may fairly infer that the 
megascleres have a microscleral origin. The interest attaching to the microscleres 
is in inverse ratio to their size ; they not only represent the primitive material 
from which the megascleres have been derived by increased growth and mechanical 
selection, but they present among themselves most perfect transitional series which help 
us to understand the history and causation of spicular forms ; while last but not least 
they reveal in an unexpected manner the filiation between the different families of 
sponges constituting the order Demospongise. 
The simplest form of microsclere is probably the globule, which occurs associated with 
the sigmaspires of the genus Tetilla, and next to this the sigmaspires themselves. 
Within the family Tetillidse the sigmaspire presents but slight modifications, but those 
which do occur are of great interest. 
1. The transition from the sigmaspire to the toxaspire can be traced in Chrotella 
macellata (vide p. 20). 
2. From the sigmaspire to a curved microstrongyle in Tetilla stipitata, Carter {vide 
p. 49). 
3. In Tetilla geniculata, Marenzeller (p. 46), the sigmaspire is centrally tylote. This 
seems to point to the globule as the original form of the sigmaspire, the central tylus 
representing the persistent globule. 
4. In Tetilla japonica, Lampe (p. 46), and in Craniella atropurpurea, Carter 
(p. 50), the sigmaspire becomes spined, and in the latter of the two sponges it acquires 
an unusually large size, though the spines remain small in proportion to the spire. The 
largest spines, however, occur at the termination of the spire and thus suggest a tendency 
towards the Desmacidine chela. 
In Demospongise which are not Tetractinellida we meet with sigmas, which may be 
