172 BULLETIN OF THE BUREAU OF FISHERIES. 
The size of the flagellated chambers has been adduced as a differential feature 
marking off Hircinia and Stelospongia (Lendenfeld, loc. cit., p. 484). The chambers 
are said to be larger in Stelospongia, the diameter being given as 41 to 48 w. But in 
Hircinia variabilis F. E. Schulze, the diameter may be as much as 4o p. In such a 
case this generic differential can not be applied, although it may hold for the majority 
of the species of the two genera. 
The fundamental character of the fascicular fibers appears to be constant in Stelo- 
spongia. They are compound fibers (Lendenfeld, loc. cit., p. 478; Pl. XXXI, fig. 7; 
Pl. XXXII, figs. 7, 8, 9, 10). Lendenfeld sometimes calls them “plexus bands’’ (PI. 
XXXI, figs. 4, 10, 12, 14; Pl. XXXII, figs. 9, 10). The constituent parallel fibers may 
be well apart, or may come together so closely as to fuse (Pl. XXXI, fig. 14). The 
width of the compound fiber thus varies greatly from less than 200 yp to several milli- 
meters. It is unimportant whether the several main fibers of a ‘‘fascicle’” arise as 
branches of a common fiber or not. Farther in the interior doubtless they often unite, 
although separate peripherally. Compound fibers of this kind, as we have seen, un- 
doubtedly occur here and there in Hircinia ectofibrosa, but in this species certainly, and 
probably in the genus at large, when the main fibers become fascicular the character- 
istic formative method practiced is that of incorporating the roots of the connectives. 
The characteristic fascicular fibers of the two genera are thus probably quite different 
structures. 
Since in Lendenfeld’s system both under Hircinia and Stelospongia there are forms 
with single main fibers, H. variabilis and S. (Cacospongia) vesiculifera, for example, we 
are driven back in the separation of these forms, and hence in the separation of the 
genera, to the presence or absence of filaments and of ‘‘root plates” formed by the 
divergent roots of the connectives. 
