20 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. QI 



in exceptional cases as much as 170 /x in diameter, and number about 

 10 to the square millimeter. The oscules are not evident, and we 

 may assume that the exhalant openings resemble the inhalant. The 

 ectosome contains numerous tangent spicules and may be removed 

 with moderate ease, so that there may be said to be a special 

 dermal skeleton. The genotype and no. 22305 possess in the center 

 of the upper surface a peculiar depression, which in general tapers 

 to the bottom like an inverted cone, and on the walls of this depres- 

 sion at intervals between the top and the bottom occur sharp ridges 

 running quite around it. Were this a single ridge that descended 

 to the bottom in a spiral, one would suspect that the sponge had 

 grown around a snail shell, but the indications are positively to the 

 contrary. There are relatively large canals that ascend through the 

 sponge parallel to the curved sides, perpendicular to the flat base. 

 These are about 2 or 3 mm in diameter. They do not communicate 

 with the exterior by any conspicuous opening through the dermis at 

 all but are roofed over by the above-mentioned special tangent dermal 

 skeleton. The internal structure in general is cavernous. Some for- 

 eign material is present, and in a few cases there are vague tracts 

 about 150 ju, in diameter crowded with spicules and containing some 

 spongin. The principal spicules are strongyles of astonishingly regu- 

 lar size and shape, the thickness varying only between about 12 fx and 

 13 fi, and the length only between about 330 /a and 380 /a. In addition 

 to these there are microrhabds or oxeas 2 /x to 3 /i, in diameter and 

 100 ft to 300 ju long. In the type specimen I find what I take to be 

 embryos; these are about 650 /x in diameter, subspherical, and nearly 

 black. 



Strongylophora in general is an East Indian or Indian Ocean genus, 

 and none of the other hitherto described species has microscleres 

 nearly so long as the microxeas of ram.pa, and furthermore all the 

 other species have at least some of the megascleres much larger than 

 any of those in rampa. That two specimens out of four have the same 

 sort of peculiar concavity is significant. This species is so common 

 that one would strongly svispect its occurrence in some earlier collec- 

 tions, but since there seems no way of telling which of many earlier 

 names was given to it, we are forced to describe it as a new species. 



Family COELOSPHAERIDAE Hentschel 



COELOSPHAERELLA, n. gen. 



This designation is proposed for a small group of species resembling 

 Coelosphaera but having palmate isochelas instead of arcuate, and 



