REVISION OF FRESHWATER SPONGES OF SPONGILLIDAE 141 



rarely smooth, more often with a number of uTsgular tubular pro- 

 jections; oscula comparatively few but conspicuous, often present on 

 conical eminences. Skeleton consisting of distinct vertical spicule 

 fibers, tightly bound together by spongin, and an irregular arrange- 

 ment of transverse fibers. Consistency of live sponge very rigid but 

 often brittle. 



Alegascleres almost straight or only feebly curved cylindrical am- 

 phioxea, as a rule entirely smooth; length range 220-310 fi, wddth 

 range 11-16 /x. 



Microscleres absent. 



Gemmoscleres minute birotulates with a slender shaft, and termi- 

 nally Avith circular rotules of unequal diameter; upper rotule often 

 considerably recurved to form a bowUike structure; length of shaft 

 13-17 n, its thickness 3-4 ju, diameter of lower rotule 16-19 n, of upper 

 13-16 n. 



Gemmules rather abundant, scattered through and loosely held in 

 position by skeletal mesh work, not encased in capsules of megascleres ; 

 diameter ranging 175-235 ix] pneumatic layer relatively thin and 

 granular; gemmoscleres embedded in this coat in one layer, foramen 

 produced into a conical and short porus tube. 



Distribution. — Ranging from India through SE. Asia north to 

 China, and south to eastern Australia (Racek, MS.). 



Color in life. — Varying shades of bro\vn, usually dark tan. 



Discussion. — The justification for Annandale's (1919b) separation 

 of T. latouchiana sinensis as a distinct and geographically isolated 

 subspecies has been doubted by Gee (1932b), who found the gemmo- 

 scleres in that subspecies, as well as in the Indian species, to vary to 

 such a degree that they did not seem to represent a reliable dis- 

 tinguishing criterion. Observations on the extensive material available 

 for the present study fully support Gee's views, and many intergrading 

 series of gemmoscleres were found in the same individual of this com- 

 plex. The retention of T. I. sinensis as a geographic race of T. latou- 

 chiana is therefore impossible. 



Gee's (1932b) "variety" T. I. pasigensis from the Philippines is 

 obviously a freak. Its spicules display a large number of striking 

 irregularities most common in sponges living in an adverse environ- 

 ment: Its megascleres are thinner and shorter and bear bulblike 

 central enlargements; its gemmoscleres have thinner and often spiny 

 shafts, usually protruding through both rotides; both rotules are 

 frequently at various angles to the shaft, which itself is not always 

 straight. While T. paulula, as mentioned above, is clearly separable 

 from the T. latouchiana group, no purpose can be served by dividing 

 the latter into subspecies or "varieties," the characters of which are 

 highly variable and cannot be demonstrated. 



