100 CEYLON PEARL OYSTER REPORT. 



effected. There are no transitional forms intermediate between these desmas and the 

 discotrisenes of the surface, such as Mr. CARTER describes (4), and the desmas are 

 evidently not derived from discotriaenes. Each originates from a tetract crepis 

 (calthrops) by branching of the actines and secondary deposition of siliceous material. 

 The initial stages in the development of the desmas are to be found in abundance 

 just beneath the superficial layer of discotrisenes, where the skeleton is evidently 

 growing. The young crepis (Plate IV., fig. 4, a) is a regular tetract calthrops with 

 minutely roughened surface. In one of the youngest I have seen, the actines measure 

 about 0'04 millim. by 0'004 millim. One actine may be longer than the other three ; 

 they are sometimes conical and suddenly constricted at the base. Having attained a 

 length of about 0*08 millim., or it may be less," the actines swell out at their ends 

 and begin to branch irregularly (Plate VII. , fig. 4, c). Presently the original minute 

 roughening of the surface is covered over by smooth secondary deposits of silica, and 

 the adult, much larger tuberculation appears, though the more central portions of the 

 spicule may often be nearly smooth when fully developed. The diameter of the 

 actines of the fully formed spicule (" epactines " of SOLLAS), between the centre and 

 the point where they commence to branch, is about 0'036 millim. This account of 

 the development of the desmas, while differing widely from that given by CARTER, 

 agrees with the views expressed by SOLLAS in his " Challenger" report. 



(2.) Discotrisenes (Plate IV., fig. 4,/ g) ; disk flat, about 0'23 millim. in diameter ; 

 outer surface beset with numerous minute, conical papillaa ; margin approximately 

 circular, but minutely crenulated and usually strongly emarginate at one, two or three 

 places, so as to form semi-circular bays or indentations ; shaft well developed, stout, 

 conical, fairly sharp-pointed, up to about 0'147 millim. long. These spicules form a 

 single layer on the surface, with overlapping disks. The youngest stage I have 

 recognised in their development is shown in Plate IV., fig. 4,f. 



(3.) Monaxonid spicules (rhabdi) ; probably oxeote ; long and slender, and commonly 

 tapering very gradually to a fine point. These spicules are so long and slender and 

 so interwoven with the desmas that I have never succeeded in seeing both ends of 

 the game spicule in situ, and even in boiled-out preparations they are always broken. 

 They are particularly numerous in the neighbourhood of the vent, but do not, at any 

 rate usually, project beyond the surface. I have measured one up to a length of 

 about 1'2 millims., with a maximum diameter of about O'OOG millim. 



(4.) Microrhabds (Plate IV., fig. 4, h) ; usually fusiform and strougylote, with very 

 minutely roughened surface; sometimes slightly curved; size about O'OIG millim. 

 by 0'0027 millim. These spicules are very abundant beneath the layer of disco- 

 triaenes ; and especially in the membrane which immediately surrounds the vent, 

 where they form a continuous layer, the discotrisenes being absent. 



It is with some reluctance that I have been constrained to add another species of 

 Discodermia to the six already described by Mr. CARTER (4, 5) from the Gulf of 

 Manaar (D. papillata, D. aspcra, D. spinispindifera, D. latviiliscus, D. sinuosa, and 



