388 BULLETIN 100, UNITED STATES NATIONAL MUSEUM. 



rearrangement of the species of the two genera, to avoid overlapping 

 is probably already desirable. 



As Lundbeck, 1902 (p. 78), says, since sigmas and toxas occur to- 

 gether in the new species which he describes, Dendy's earlier sug- 

 gestion, that it may prove necessary to merge Toxochalina (with 

 toxas), might now be followed. But, as he intimates, it would be 

 better to keep the two genera separate by stressing the regular, 

 chalinid, arrangement of the skeletal fibers in Toxochalina in con- 

 trast to the more irregular arrangement which is on the whole char- 

 acteristic of Gellzodes. And yet in some species (type species, for 

 example, Ridley, 1884) of the latter, radial fibers and connectives 

 are distinguishable. We here meet again the common phenomenon 

 of the intergradation of a group of species with several other such 

 groups, the explanation of which seems to be that the characters of 

 the ancestral group have varied during the evolution of the existing 

 species independently of one another, and thus numerous combina- 

 tions of characters have resulted. 



GELLIUS VARIUS (Bowerbank), var. FIBROSA, new variety. 



Plate 40, fig. 3 ; plate 49, fig. 1. 



Halichondria varia Bowerbank, 1875, p. 292. 

 Gellius varius Ridley, 1884, p 424. 



Bowerbank's name of Halichondria varia (1875, p. 292) was given 

 to several specimens from the Straits of Malacca, in which the body 

 was in some cases a rough irregular mass and in others consisted of 

 short cylindrical branches; with renieroid skeletal network, uni- 

 or bi-spiculous; with unispiculous dermal network. Ridley (1884, 

 p. 424) states that in this type the oxeas measure 220 by 16 p, the 

 sigmas 25 to 32 jjl long. 



With this species Ridley (1884) combined Isodictya virgata Bow- 

 erbank, also from Straits of Malacca (Bowerbank, 1875, p. 294). 

 The sponge so named consisted of two branches, not exceeding 2 

 inches in height. The renieroid skeletal network shows well-pro- 

 nounced multispiculous primary (namely, radial) lines. Ridley 

 gives the oxeas as measuring 220 by 14 ja, the sigmas 25 jx long. 



Ridley (1884) had two fragments from Port Darwin; of an erect, 

 cylindrical habitus; dimensions not given. In these the oxeas meas- 

 ured 250 by 15 ji, the sigmas 19 p. long. Ridley and Dendy (1887, p. 

 38) had several fragments from the Philippines, all more or less 

 cylindrical, the largest 80 mm. long by about 6 mm. in diameter. 

 The pieces branch and the branches show a strong tendency to 

 anastomose. The size and distribution of the oscula are as in the 

 Albatross specimen. The oxeas measure 220 by 13-14.5 ji, sigmas 

 about 40 \i long. 



