98 PROP. p. M. DTJNCAN ON THE ANATOMY OF THE 



never become demi plates, but remain primaries*. The cause 

 of the persistence of the low primaries in Diadema relates to the 

 method and the rapidity of the growth of the tubercle-bearing 

 second plate, as well as to the amount of downward pressure 

 developed during the growth of tbe new plates at the radial end 

 of the ambulacra. The plates are all primaries there. (Plate Y. 

 fig. 4, 4 a.) 



In the communication by Mr. Percy Sladen and myself, already 

 alluded to f, we noticed especially the peculiar growth of some 

 plates not far from the radial plate in Gcelopleurus Sindensis, 

 nobis. It was noticed that although, in specimens of that spe- 

 cies, the triplets of the tubercle-bearing plates at the ambitus 

 were as is usual in the genus, yet more abactinally, and where 

 the downward growth-pressure was not assisted by th.e expansive 

 growth of tbe second or tubercle-bearing plate, the adoral and 

 aboral plates of the triplet were not forced to become demi 

 plates : tbey still remained as long and low primaries. This 

 state of things is exceptional in the Arbaciadse, but it is normal 

 in the genus Diadema. 



III. The Stetjcttjee of the Edges of the Plates 



AT the StJTIJEES. 



The Vertical Sutural Edges between the Ambulacra and the In- 

 terradia. — The line of suturing is a series of convex triplet-plate 

 ends received into corresponding concavities in the interradial 

 plates. On the exposed edges of botb the ambulacral and in- 

 terradial plates there is a thin, and rather solid, layer at the 

 inner part of the section, and a corresponding structure at the 

 outer or superficial part ; but all the intermediate thickness is 

 occupied by a number of excessively thin, straight or wavy, dis- 

 tinct, and long laminae. They are separated by similarly shaped 

 spaces, which are rarely crossed by any ofi'shoots of the lamellar 

 structure. The lamellae conform to the windings of the sutural 

 surfaces, and those of one compound plate, or of an interradial 

 plate, are continuous. 



The laminse are not continued across the sutures, and it is 

 evident that the interspaces were once occupied by connective 



* Quite at the peristomial edge a plate may become a demi. 



t Journ. Linn. Soc, Zool. toL xix. p. 36, pi. i. fig. 9. (The figures 6 and 9 

 have been misplaced on this plate ; but although on p. 36 and p. 57 the 

 reference is to fig. 6, the drawing is marked 9.) 



