356 BULLETIN 100, UNITED STATES NATIONAL MUSEUM. 



(George and Wilson, 1919). it seems best to retain both genera. 

 The family name (Tetillidae) remains unchanged. 



" Tetilla, the simplest and therefore presumably the ancestral 

 genus of the family, has been gradually enlarged in the practice of 

 recent writers by the incorporation in it of atypical species that 

 depart in one direction or another from the central group of typical 

 forms to which Sollas' definition (1888) is applicable" (George 

 and Wilson, 1919). Thus species that have differentiated in the di- 

 rection of Craniella, in that they have the beginnings of a fibrous 

 cortex, are here included (T. longipilis Topsent, 1904, p. 97; T. 

 anomala Dendy, 1905, p. 91; T. horodensis Dendy, 19166, p. 105). 

 Other species are classed here that have differentiated toward Cina- 

 chyra in that there are superficial poriferous depressions, the floor 

 of which exhibits pore-like apertures, the inhalent or exhalent na- 

 ture of which is not quite certain (see Kirkpatrick, 1905, Dendy, 

 1905), but these species lack the fibrous cortex and cortical skeleton 

 of Cinachyra. I propose to group them under the subgenus 

 Cinachyrella. 



Tetilla and its relatives offer excellent illustrations of the fact 

 that sponge genera become more and more difficult to distinguish as 

 the number of known species increases. While this is perfectly well 

 known, a few words on the matter may not be amiss. (See Wilson, 

 1919, 19196.) 



Each genus is of course only a group that has been gradually built 

 up round a type embodying a certain combination of well-marked 

 features or " characters.'' In the cases of Tetilla, Craniella, and 

 Cinachyra, the main '" characters," grouped in pairs (Mendel-wise), 

 the members of which contrast, are as follows: 



1. A nonfibrous cortex (a) or a fibrous cortex (A). 



2. Radiating cortical oxeas not present (b) or present (B). 



3. Pores scattered (c) or grouped in poriferous depressions (C). 

 Of these characters Tetilla typically embodies a, b, and c; Craniella^ 



A, B, and c; Cinachyra, A. B, and C. 



Comparison of many species shows that the characters vary, more 

 properly have varied during the evolution of the present races, in- 

 dependently of one another and hence the number of combinations 

 actually found increases with the number of species known. In the 

 case of the above genera there are more combinations than there are 

 genera recognized, and this is of course often the case in systematics. 

 Thus in the group of species congregated under Tetilla we find not 

 only the typical combination but others as well: A, b, and c; a, b, 

 and C. 



In such a state of affairs we can either combine genera until we 

 get groups so heterogeneous as to be useless to biology, or we can 

 go on splitting up genera on the plan that each genus shall represent 



