136 U.S. NATIONAL MUSEUM BULLETIN 27 2 



Megascleres straight to feebly curved, broadly fusiform, sharply 

 pointed amphioxea, covered with stout and sharp spines in varying 

 patterns; length range 170-235 /z, width range 11-15 m- 



Microscleres absent. 



Gemmoscleres minute birotulates with a stout, smooth, and short 

 shaft, and terminally with rotules of more or less recurved circular 

 margins and of equal size and shape; length of shaft about 11 M; its 

 thickness about 4-5 n; diameter of rotules ranging 9-12 /x. 



Gemmules rather abundant, confined to lower parts of sponge, 

 spherical and moderately small; usually encased in capsules of normal 

 megascleres; diameter ranging 475-540 n; pneumatic layer well 

 developed, but never thicker than the length of gemmoscleres, con- 

 sisting of rather large air spaces; gemmoscleres embedded in this coat 

 in a single layer; foramen produced into a conical and short porus 

 tube. 



Distribution. — Apparently discontinuous but wide dispersal in 

 all continents of the Northern Hemisphere, with preference to cold- 

 temperate regions. 



Color in life. — Ranging from light yellow to dark brown. 



Discussion. — Although not very abundant, T. horrida is a well- 

 described species, and the present studies did not reveal any additional 

 criteria for discussion. As can be expected in sponges with a widely 

 scattered distribution, this species displays some slight variations in 

 distant populations which are not significant enough to permit a clear 

 discrimination of races. T. horrida ningpoensis Gee (1926a) differs from 

 T. horrida only in the smaller size of its spicular components and must 

 therefore be considered an unimportant ecomorphic variation. T. sol 

 Annandale, relegated by Gee (193 le) to varietal rank, merely displays 

 pronounced radial reinforcements on both rotules of its gemmoscleres, 

 a condition often also found in a number of shdes of the typical T. 

 horrida examined. 



Our reexamination of the type slides of T. polysclera supports 

 Gee's (1931e) view that this species is fully synonymous with T. 

 horrida. These slides clearly contain spicular components of two 

 spongiUid species, one of them, containing the pecuHar microscleres 

 described by Rezvoj (1926a) as well as smooth megascleres, probably 

 belonging to the Spongilla lacustris group. The rapid growth of all 

 Trochospongilla species, studied in detail in Austrahan waters (Racek, 

 MS.) , often results in their fastening to already developed sponges of 

 other species, so that frequently a spicular mixture can be observed. 

 This phenomenon has also been recorded by previous authors and 

 accounts for the confusion of a number of previous species of this and 

 other genera (e.g., Spongilla tanganyikae Evans, Metania lissostrongyla 

 Burton, and others). 



