Sponges from Funafuti. 353 



general surface. Lobes (mainly following the shape of the 

 meshes of the madrepore in the present specimens) usually 

 elongated and quadrangular, attaining a size of 3 5 x 1*5 

 millim., but varying in shape and size ; lobes connected with 

 each other by short tubular channels varying in diameter ; 

 the interior of the lobes occupied by a labyrinth of thin- 

 walled channels and spaces. Flagellated chambers oval, 

 eurypylous, 30 x 24 /*. 



Skeleton formed of scattered strongyles occasionally forming 

 an irregular reticulum, and mostly lying tangentially in the 

 walls of the lobes. 



Spicules. — Megascleres : strongyles varying from 126 to 

 246 fi in length and from 3 to 5*5 fi in thickness, curved at 

 the centre, smooth, rarely slightly roughened at the ends. 



Microscleres : microstrongyles (rather rare) 90 X 3 /x, 

 slightly curved, with truncate ends, and with spines arranged 

 in a regular closely-whorled spiral. 



Surface of the " galleries/' and especially of the connecting 

 channels, finely shagreened, but frequently smooth. 



The specimens consist of three fragments of the coral 

 Echinopora excavated by the sponge. Two of the pieces are 

 dried, the sponge being of a dark brown colour; the third 

 piece is in formol, the sponge here being orange-yellow when 

 the coral is freshly fractured, but gradually becoming brown 

 on exposure to light. The fragments are also excavated by 

 the purple boring-sponge Cliona Schmidti, Ridley. The 

 new species is nearly related to the Cliona purpurea of 

 Hancock (11, p. 343, pi. xii. fig. 6), the latter species now 

 coming under the new genus. 



Topsent (28, pp. 576-7) was of opinion that Hancock's 

 species was not a genuine boring-sponge, but possibly a 

 Desmacidine which had grown into excavations made by 

 some other organism. In answer to my request for informa- 

 tion about Hancock's collection, Mr. R. Howse, Curator of 

 the Newcastle Museum, very kindly forwarded to me the 

 type specimen of Cliona purpurea. Tn view of the close 

 affinity of this species to D. Davidi, a brief description of the 

 former is given here for comparison. 



The type specimen consists of several small fragments of 

 Tridacna gigas } with holes for the oscular and poral papillae 

 on the inner and outer surfaces. The broken edges show a 

 labyrinth of small oval cavities with finely shagreened surface 

 connected by tubes or foramina, and lined with a dull red 

 fluffy membrane. The megascleres are amphi-subtylote and 

 measure 246 x 5*25 /x, the ends being 6 ll in breadth, slightly 

 spined, and sometimes with a terminal tuft of spines. The 



