GEODIA BREVIANA. 



155 



Carter the oxyasters are much larger and destitute of spines. In Dendy's Geodia 

 areolata (Carter) the orthotriaene-rhabdomes are much more slender, the ana- 

 triaene-clades shorter, the mesoprotriaenes absent, and the surface marked with 

 a reticulate tracing. In Geodia ataxastra the mesoproclades and anatriaenes 

 have differently shaped cladomes, the dermal strongylosphaerasters are much 

 smaller, and ataxasters are met with; in G. hreviana minute, dermal anacladcs 

 are present, the clades of the large anatriaenes are much stouter and the strongy- 

 losphaerasters larger; in G. agassizii some of the spicule-dimensions are con- 

 siderably greater, the anatriaene-clades stouter and more angularly bent, the 

 orthotriaene-clades often provided with one or more spine-like branches, some of 

 the mesotriaene-epirhabds much longer than the clades of the same cladome, 

 and smaller oxyasters abundant; and in G. ynesotriaena most of the spicule- 

 dimensions are from two to three times as great, the anatriaene-clades stouter 

 and curved in a different manner, the mesotriaene-epirhabds usually longer than 

 the clades of the same cladome, the rays of the strongylosphaerasters more 

 slender, and the oxysphaerasters much more spiny and provided with relatively 

 smaller central thickenings. 



Geodia agassizii and G. mesotriaena appear to be more closely allied to G. 

 mesotriaenella than the other four. It is true that many of the spicule-dimensions 

 of these sponges, particularly of G. mesotriaena, are much greater; in view of the 

 fact, however, that the specimen of G. mesotriaenella is very much smaller, and 

 therefore probably much younger, than the specimens of G. agassizii and G. 

 mesotriaena examined, this might be of no systematic importance. The differ- 

 ences between these species and G. mesotriaenella in regard to the shape of the 

 spicules may, of course, also be due to differences in the degree of development 

 (age), in which case G. mesotriaenella would have to be considered as a young 

 form of one or the other of them. Since, however, this seems very doubtful, 

 I deem it better, in order to avoid confusion, to describe this sponge as a dis- 

 tinct species. 



1893. Cydonium miilleri L. M. Lambe (non Fleming), Trans. Roy. soc. Canada, 1893, 10, p. 72, pi. 4, 

 fig. 1; pi. 6, fig. 1, la-i. 



I establish this species for a specimen dredged at Station 2894, off southern 

 California. The large anatriaenes have unusually stout and relatively short 

 clades, and to this the specific name refers. The study of the sponge deter- 

 mined in 1893 by Lambe (loc. cit.) as Cydonium miilleri Fleming which is in the 



Plate 35, figs. 1-27; Plate 36, figs. 1-12. 



