448 BULLETIN 100, UNITED STATES NATIONAL MUSEUM. 



ules grading over in some species to the dichotriaene form; with 

 microrhabds that are usually strongylate but sometimes oxeate. 



Thiele, 1900, discusses Sollas' definition of the genus, and would 

 not limit it to cup or vase shaped forms, nor to those with the pores 

 in sieves, but would leave only the skeletal characters as diagnostic. 

 Lendenfeld 1903 adopts this change. 



Lendenfeld, 1900 (pp. 343-44), makes a change in the customary 

 distinction between Theonella and Discodermia, owing to his dis- 

 covery of a form (7\ discifera, 1906, p. 351), in which characters 

 hitherto regarded as differentials — namely, phyllo- or discotriaenes — 

 intergrade. This is a good case of the familiar phenomenon that 

 genera which seem to be natural are found, as the number of known 

 species increases, to intergrade with respect to one or more of the 

 differential characters (Wilson, 1919, 19195). The possession of 

 microxeas, larger than the microscleres (strongylate) proper, would 

 still continue, in Lendenfeld's scheme, to set off Discodermia from 

 Theonella. But these two classes of spicules intergrade in some spe- 

 cies, as for instance, Discodermia ramifera Topsent (1904, p. 58), 

 where they form but a single class, the range in length being 20-45 

 [x. The same holds for Discodermia emarginata, var. lamellaris of 

 this report. In most of the recorded species of Discodermia the 

 difference in size between the microrhabds and the microxeas is not 

 very great, the former ranging about 10-20 \i, the latter up to 50 [x, 

 in length. In a few species, however, the categories are quite dis- 

 tinct, as in D. panoplia Sollas, where the microxeas have a length 

 of 70-100 [a, the microrhabds a length of only 10-15 [jt,. 



THEONELLA SWINHOEI Gray. 



Theonella swinhoei Gray, 1868&, p. 565. — Sollas, 1888, p. 2S4. — Lendenfeld, 

 1903, p. 126. 



Four specimens, two from station D5218, one from D5593, and 

 one from D5252, cannot be separated from one another nor from 

 Gray's species. They show however that the species idea must be 

 somewhat modified. Thus the ridges and projecting lobes which 

 are conspicuous on the surface of some specimens (Sollas, 1888, 

 Lendenfeld, 1903) may be absent in others. The vasiform habit with 

 single cylindrical cloaca extending nearly through the entire sponge, 

 again, is not constant. For the cloaca may extend only through the 

 upper half of the sponge, or instead of being vasiform the sponge 

 may be massive, the upper surface bearing several oscula, each lead- 

 ing into a cloacal canal which extends through one-quarter to one- 

 third the total height of the sponge. The pores are not strictly in 

 groups (sieves, Sollas), but the whole dermal membrane is uniformly 

 porous. The system of larger subdermal cavities which may be 



