94 GEODIA ATAXASTRA. 



oblique, strongly inclined inwards (Plate 44, figs. 43-45). The umbilicus is 

 about 10 /( deep. 



The eight specimens of the two varieties of this species were collected on 

 October 26, 1904, on the shore of Pcrico Island, Panama. 



Although these specimens belong without doubt to the same species, they 

 are not quite identical. Seven of them are quite or nearly wliite and have the 

 same structure, one is gray and differs from the rest by its ataxasters being 

 larger, bj' its megascleres and sterrasters being smaller, by the clade-angles of 

 its large anatriaenes being considerably wider (averages 34° and 47° respec- 

 tively), by its mesoproclade-cladomes being different in shape, by the rays 

 of none of its large oxyasters being branched, by possessing a few oxysphaer- 

 asters with thick, smooth rays and irregular sterraster-derivates, and by being 

 Restitute of minute dermal anaclades. Some of these differences, particularly 

 those of the asters and the ana- and mesoproclades, seem to me to be germinal 

 in character. 



The spiculation and the cribriporal nature of the afTerents and efferents 

 show that these sponges belong to Geodia. For the reasons given in the descrip- 

 tion of Geodia agassizii I have compared them not only with the described 

 species of Geodia but also with those of Sidonops. The species most nearly 

 alUed to them are Geodia tuber (tuberosa) 0. Schmidt, G. distinda Lindgren, G. 

 hilgendorfi Tliiele, G. midleri (mullert) Fleming, that described by Dendy as 

 G. ramodigitata Carter and here described as G. mesotriaenella. Geodia tuber, G. 

 distincta, G. midlei'i, G. ramodigitata (Dendy), and G. mesotriaenella are dis- 

 tinguished from G. ataxastra by being destitute of ataxasters and by having 

 much larger dermal strongylosphaerasters. Geodia hilgendorfi, differs from G. 

 ataxastra by possessing small oxysphaerasters instead of the strongylo- 

 sphaerasters. 



