THE VOYAGE OF H.M.S. CHALLENGER. 
cviii 
There can be no question as to the validity of the family Tetillidse, its members are 
united by a nexus of characters, one or another of which may disappear without any 
danger of the family character being lost. 
The most persistent characters are the following : — the aggregation of the megascleres 
into radially directed fibres ; the form of the protrisene, anatrisene, and sigmaspire. 
Oxeas are always present, but these spicules are too generally distributed to be of 
much service in classification ; they are so commonly anisoactinate in the Tetillidse that 
in this particular they might be regarded as distinctive. 
The cortex, once supposed to be of ordinal value, for Schmidt’s families are equivalent 
to the orders of more modern spongologists, subsequently of subordinal value (Wyville 
Thomson), later of family value (Sollas), is now reduced to the distinguishing of genera, 
though in some cases it may define a family. In the Tetillidse it is of generic value only. 
The chamber-system may in other groups furnish characters for the definition of 
families, but not in the Tetillidse, since the change from the eurypylous type to the 
aphodal takes place too gradually, the genus Craniella with its aphodal system being 
united to Tetilla with a eurypylous system through the intermediate genera, Cinachyra 
and Chrotella ; lest it may be objected that this remark is in opposition to those already 
made with reference to the possibility of intermediate forms existing between families, I 
would add that as a matter of convenience the subdivision of the Tetillidse into two or 
more families has nothing to recommend it, for Cinachyra and Chrotella, while inter- 
mediate genera as regards the chamber system, are not intermediate but rather aberrant 
in other respects. Till our knowledge of the group is increased we must leave the 
aphodal and eurypylous Tetillidse to form together one family. 
The protrisene is highly characteristic and constant, but while it serves well to define 
the family, it is of no further use since it is not exactly repeated in any other group of 
sponges, or only very rarely. 
The aggregation of the megascleres into radial fibres, chosen by Wyville Thomson as a 
character for the definition of an order — the Eadiantia, — on which considerable stress has 
lately been laid by Vosmaer in the classification of the Monaxonida,^ is of very doubtful 
importance ; in primitive forms, such as Placina and its immediate derivatives, it is 
wholly absent, the only approach to it being found in the orientation of the trilophous 
candelabras of Placina trilopha, which are directed near the surface in the manner of 
trisenes ; throughout the whole of the Asterophora and the Sigmatophora the radial 
arrangement is prevalent, the most striking exception occurring in the family Pacha- 
strellidse, and even here in some genera the rhabdous spicules lie in bundles which tend 
to a radial direction. On the other hand, within the limits of a single family the radial 
arrangement may break down ; thus in the Theneidse, which appears to be a very natural 
family, a perfectly radial arrangement characterises the genus Thenea, but none of the 
^ Vosmaer, Bronn’s Klass. u. Ord. d. Thierreiclis, Porifera, pp. 326, 328, 
