SPONGES. 105 



desmas. Maximum length from apex to apex of branches about 074 millim., with 

 main axis about - 06 millim. thick. 



(2.) Stout and slender monaxonid spicules (Plate IV., fig. 5, b-c) ; more or less 

 curved and variously ended, ranging from oxeote to strongylote and stylote ; size 

 about 0-98 millim. by 0*0369 millim., 1*2 millims. by 0'0246 millim., 1-28 millims. by 

 0"0328 millim., &c. ; very variable. 



The most interesting feature of this sponge is the strong development of the 

 monaxonid spicules (rhabdi), which, in three out of the four specimens (the fourth 

 being dead and infested by another sponge\ play an equal part with the desmas in 

 the composition of the skeleton. This condition of the skeleton suggests that 

 certain Monaxonellid sponges usually associated with the Axinellidse may be derived 

 from Lithistid ancestors by the suppression of the desmas, or that certain Lithistids 

 with monocrepid desmas may possibly be derived from Monaxonellid ancestors. 



Petromica massalis is evidently very closely related to Topsent's P. grimaldii, the 

 type of the genus, from the Azores. The principal difference appears to concern 

 the ornamentation of the desmas, which, in P. grimaldii, are ornamented at the 

 extremities with little conical tubercles, not pointed. These are absent or very feebly 

 developed in the Ceylon species, but subsequent researches may make it desirable to 

 unite the two. 



R.N. 198 ; 216 (dead, and infested by another sponge ; deep water outside pearl 

 banks, Gulf of Manaar) ; 257 (type) ; 269 (young ; deep water oft' Galle and onwards 

 up West Coast of Ceylon). 



Grade : MONAXONELLIDA. 



Tetraxonida in which the primitive tetraxonid and tetractinellid coDdition of the 

 megascleres has been entirely lost and none but monaxonellid megascleres 

 remain. No desmas are developed. 



With the recognition of the true nature of the monaxonellid sponges as reduced 

 Tetraxonida, it becomes once more desirable to modify the name of the group. The 

 old name, " Monactinellidre," which we owe to Zittel, was objectionable because 

 implying that the spicules are necessarily monactinal (one-rayed), while, as a matter 

 of fact, they may be also diactinal (two-rayed). Sollas therefore altered the name 

 to " Monaxonidse," which was afterwards altered to " Monaxonida," under which 

 name the group appears in the ' " Challenger" Reports' and other recent works. 



The name " Monaxonida," however, becomes distinctly objectionable when used for 

 a mere sub-division of the order Tetraxonida, contrasted with the Tetractinellida ; 

 and I therefore propose to again modify the name of the group, choosing this time 

 the term " Monaxonellida," which is, I believe, both correct in meaning and in 

 harmony with the scheme of classification adopted. 



The Monaxonellida form, however, an unnatural group of polyphyletic origin, 

 which we can only retain as a matter of convenience until we know more about their 



p 



