144 BULLETIN OF THE BUREAU OF FISHERIES. 
Lendenfeld’s treatment tends to obscure the nice distinctions to which Sollas’s classifica- 
tion gives expression, and which should certainly not be lost sight of. 
Topsent in 1904 continues to use the three genera, Chrotella, Tetilla, Craniella. 
Dendy in 1905 (p. 89) uses but redefines Tetilla so as to include forms in which the 
ectosome is in part fibrous. His definition runs: “Cortex absent or feebly developed; 
no special cortical skeleton.’ This is one way out of the difficulty presented by the 
occurrence of intermediate forms, in that Tetilla is here made to include species that 
shade off toward Tethya (Craniella), and which certainly are intermediate. To be sure, 
another classifier using the same genera might include such or slightly different inter- 
mediate forms under Tethya, extending Tethya downward, so to speak, rather than 
Tetilla upward. Dendy also uses Craniella in 1905, and again uses Tetilla in 1916, in 
the sense in which he employed it in 1905. Row in 1911 uses Tetilla and Chrotella. 
Hentschel in r911 uses Tetilla, but in 1912 follows Lendenfeld and merges Tetilla (+ Chro- 
tella) in Tethya (Craniella auct.) 
As exploration goes on the number of sponge genera known to run into one another 
increases. Everywhere intermediate forms are found. We meet, then, very frequently 
the practical difficulty of finding the record of a known species or of deciding where to 
record a new species. If the genera exhibited a linear arrangement, we might have 
sharply defined genera alternating with less homogenous intermediate ones. But it 
frequently happens that the species of a sponge family fall into groups which shade off 
in all directions toward one another. In such a case, and it looks as if discovery would 
show that this is all but universal in sponges, the questions arise: Shall we give up any 
formal grouping of the species (it is of course not a mere question of names, genera or 
subgenera) ? Or shall we define all the species groups (genera or subgenera) in a compre- 
hensive, and therefore rather loose, way, which results in overlapping? Or shall we 
meet the difficulty by accepting some sharply defined and other lcosely defined genera? 
It is the latter method which is commonly employed, although not always explicitly, 
and no better treatment has as yet been found. 
Tetilla, the simplest, and therefore presumably the ancestral genus of the family, 
has been gradually enlarged in the practice of recent writers (Dendy, Topsent, Hent- 
schel) by the incorporation in it of species that depart in one direction or another from 
the central group of typical forms to which Sollas’s definition is applicable. Topsent, 
for instance, includes (1904, p. 97) T. longipilis, in which there is the beginning of a 
cortex, viz, an ectosome which is in part fibrous; the species having differentiated 
in the direction of Tethya. Dendy (1905, p. 89) includes T. hirsuta, in which there are 
a more or less fibrous cortex and surface depressions, the smooth floor of which is per- 
forated by pores or by oscula; a species with Tethya and Cinachyra-like features and 
which Lendenfeld (1903), in fact, lists asa Cinachyra. In the same paper (p. 91) Dendy 
includes T. anomala, in which the ectosome is pretty sharply differentiated from the 
choanosome, is fairly thick and to some extent fibrous, and ‘‘almost amounts to a 
cortex;” evidently a species approaching Tethya. Another species approaching Tethya 
has been more recently recorded by Dendy (1916, p. 105). This is T. barodensis, in 
which there is a well-developed dense cortex which is “‘perhaps to some extent fibrous.” 
T. cinachyroides Hentsch. (1911, p. 283) and T. limicola Dendy (1905, p. 93) also deserve 
mention as intermediate forms; in the anatomy of the peripheral canal system, Cina- 
chyra-like, although they lack the cortex of Cinachyra. Tetilla in this paper is accepted 
in the extended sense. 
