342 BULLETIN 100, UNITED STATES NATIONAL MUSEUM. 



the ectosomal styles are much shorter and stouter than in the Philip- 

 pine species, forming moreover fairly distinct diverging brushes. T. 

 aaptos has been recorded from the Mediterranean and Gulf of Mexico 

 (see Topsent, 1900), from Porto Rico (Wilson, 1902, p. 388), south- 

 west Australia (Hentschel, 1909, p. 377), Aru Islands (Hentschel, 

 1912, p. 319), Okhamandal in Kattiawar (Dendy, 19166, p. 101). 



Family SPIRASTRELLIDAE. 



Spirastrellidae Ridley arid Dendy, 1886, 1887. 



Megascleres usually styles or tylostyles, sometimes diactinal. 

 Asters of various forms (euasters, spirasters, sanidasters, discasters) 

 occur, often forming a dermal crust. 



The genera are Timea Gray, 1867 (=zllymedesmia Bowerbank. 

 Authors), Spirastrella Schmidt, 1868, Latrunculia du Bocage, 1869, 

 Sceptrintus Topsent, 1898, Placospongia Gray, 1867, Negombo Dendy, 

 1905, Anthosigmella Topsent, 1918, and Diplastrella Topsent, 1918. 

 Xenospongia Gray, placed in this family by Topsent, is removed to 

 the Donatiidae. Dendy, 19216 (p. 131), has recently added to the 

 family Barbosia, Didiscus, and Sigmosceptrella, new genera, and also 

 Sceptrella O. Schmidt, Podospongia du Bocage, Trachycladus Carter 

 (see Epipolasidae), and Axos Gray (see Vosmaer, 1887, p. 243). 



The spiraster of the Spirastrellidae has generally been regarded as 

 a modification of the "true aster" (euaster). But Vosmaer (1909) 

 concluded that this spicule is not an aster but a spiral monaxon with 

 spines, since the latter contain no axial canals as do the actines of a 

 true aster. Dendy (19166, p. 96; 19216, p. 129yf) in pursuance of 

 Vosmaer's argument and as a result of the discovery of some funda- 

 mentally interesting forms, transfers the Spirastrellidae and along 

 with them the Clionidae, Suberitidae, and Polymastidae to the neigh- 

 borhood of the Desmacidonidae which he would regard as the parent 

 family. This is a change of great importance, one which demands 

 time for a consideration of all the facts involved. Topsent in a paper 

 published in 1922 goes so far in this direction as to transfer Latrun- 

 culia and the very similar forms Podospongia and Sigmosceptrella 

 to the Poeciloscleridae (nearly equivalent to Desmacidonidae), 

 grouping them in a new subfamily, Latrunculiinae. He regards the 

 position of Dendy's other new genera, Didiscus and Barbozia, as 

 doubtful. Whatever resemblance there is between the special micro- 

 scleres of the latter two and the former three genera may, he thinks, 

 be attributed to convergent evolution. The remaining Spirastrellidae 

 he would apparently retain as a family in the usual position (as 

 here). 



There is not perfect agreement among writers as to the nomen- 

 clature of the genera. Thiele, 1903 (p. 955) has ruled that 



