478 COLLECTIONS FEOM MELANESIA. 



fascicles of spicule-shafts, a ray proceeding from the head of each of 

 the latter, extending along the membrane and supporting it (see 

 fig. I") ; in some parts stout acerate spicules (No. 2) take part in 

 the formation of the dermal skeleton. Sarcode transparent, of 

 very pale brown colour ; rendered subopaque, when seen in the 

 m£i8S, by immense numbers of small elongate stellate spicules. 



Spicules : — (1) Trixadiate of axis and dermal skeleton, consisting 

 of a straight shaft and two arms, one tapering to a sharp point and 

 boldly recurvate, the other ending abortively in a rounded extremity 

 shortly after its origin ; the arms are set at right angles to the 

 shaft and at angles of about 160° to each other, but lie in different 

 planes. Length of shaft and long arm probably variable, and depend- 

 ing on the position of the spicule ; the former attains a length of 5-5 

 millim., the latter of 2 millim. ; diameter about '05 millim. It is 

 the shaft of this spicfule which forms the longitudinal skeleton-bands. 

 (2) Large acerate of dermal skeleton, slightly curved, tapering from 

 centre to sharp points ; size about 1"8 by 'OS-'OTS miUim. (3) 

 Minute elongate stellate flesh-spicule, consisting of a straight or 

 occasionally curved or sinuous cylindrical shaft, beset with numerous 

 irregular blunt processes, about 20 to the spicule, varying- in length 

 from -001 to -002 millim., thickness about -001 millim. ; length of 

 spicule about -0095 millim., thickness of shaft alone -001 millim. 

 Crowded over aU parts of the soft tissues. 



Hah. Port Darwin, 7-12 fms., bottom sand and mud; Torres 

 Straits, 10 fms., bottom sand. 



Of the two specimens from Port Darwin the larger is 74 miUim. 

 (3 inches) long in its present state, viz. without its original base 

 and with the apex somewhat abraded: it probably did not much 

 exceed this length when perfect; its longest diameter (it is sub- 

 oblong in transverse section) is 8 miUim., its shortest 5 millim., at 

 the present base. The smaller specimen has the base attached, but 

 has lost the apex ; it is almost cylindrical, and has a diameter of 

 about 3'5 mUhm. throughout. In the dermis of the larger speci- 

 men no acerates have been found, but in the smaller one they 

 appear to replace the triradiates in this place ; it is in this specimen 

 that an axial canal traverses the sponge. The specimen from Torres 

 Straits is a fragment, forming the base of a specimen almost 

 certainly belonging to the same species, but very imperfect. Its 

 acerate differs from that of the typical form by having a diameter of 

 •075 instead of "05 millim. ; the flesh-spicule shows no divergence. 



The species differs very markedly from Stewart's — (1) outwardly, 

 in having the surface level instead of bearing sharp points ; (2) 

 inwardly, in the presence of an aasial cavity, in having tri- instead 

 of quadriradiate body-spicules, and in having a dermal acerate; 

 the steUates of T. columm'fera, further, are normal globostellates 

 and not elongate, as here ; in the general arrangement of the 

 skeleton this species differs by possessing a number of longitudinal 

 lines, instead of the condensed central mass of that species. 



The species is obviously nearly related to a form named Tri- 

 brachion Schmidtli, well described and illustrated as the type of 



