ART. 12 HEXACTINELLID SPONGES — OKADA 39 



parts are sufficiently preserved to make it possible to study the iso- 

 lated spicules. It is 11 mm in height and 13 mm in breadth. 



The chief difference from typical sollasii lies in the total ab- 

 sense of large onychasters and oxyhexasters, as well as in the presence 

 of the much smaller circular umbellate clavulae. The discohexaster 

 of the present subspecies is much smaller, measuring 40/a to 45/a in 

 diameter, and has a somewhat stronger and much shorter terminal. 

 Of the circular umbellate clavulae, some deviations occur in different 

 regions of the same sponge. This variation is chiefly associated with 

 the swelling on the shaft, just below the umbel. In many cases the 

 swelling is conspicuous ; occasionally it is totally lacking or is incon- 

 spicuous. The roughness of the surface is much pronounced on the 

 swelling or on parts just below the terminal umbel. The anchorate 

 hooked clavulae are quit« delicate. They have 8 to 10 weakly devel- 

 oped, slender, hooklike spines, widely diverged externally, and 55/x 

 to 67/A in breadth at the lower extension. The shaft is slender, smooth 

 on the surface, gradually attenuated to the pointed end, and 450jli 

 long and 8/a broad just below the umbel. The four tangential rays 

 of the dermal pentactins are much more distinctly tuberculous and 

 somewhat broader than those of the typical form. 



FARREA BERINGIANA, new species 



Figure 6 ; Plate 3, Figure 5 



Several large and small fragments (U.S.N.M. No. 22037) which 

 may represent parts of the lateral wall of a sponge body, were col- 

 lected from a depth of 64 fathoms, off Bering Island, Bering Sea 

 (Station 4790). 



Though both the shape of the sponge and its spiculation might 

 warrant establishing a new genus for this material, and indeed I find 

 that the entire spiculation, particularly the dictyonal framework and 

 total absence of umbel clavulae, is completely different from what 

 we find in other species of Farrea^ I venture to retain it in the pres- 

 ent genus. Instead of the network and tubular sponge body, which 

 occur in many species of Farrea^ there is here the moderately soft, 

 fairly thick, compact wall of a large cup. The sponge body was 

 broken into several fragments 30 mm to 34 mm broad, and accord- 

 ingly the complete outer configuration could not be studied. Yet I 

 believe that the body does not form the framework of a slender tube, 

 as in many species of Farrea; it rather seems to form a somewhat 

 larger cup or tube. 



The gastral membrane, which is quite clearly visible in alcoholic 

 specimens, extends over the whole inner surface in the form of a 

 delicate skin. A quadrate, latticelike network, formed of opposed 

 dermalia or gastralia, is entirely absent on all sides of both mem- 



