42 DE. J. P. GEMMILL ON THE DEVELOPMENT OP 



the various aboral arm -rudiments. So far accordingly they correspond with the single 

 terminal ossicle in the typical Starfish ray. The first pairs to appear belong to 

 rays II to V. That of ray I, which lies in the (larval) ventral lip of the nuchal notch, 

 follows almost immediately. The remainder arise in the same series as the arm- 

 rudiments to which they belong. 



In a five-rayed Starfish such as Asterina the terminals on the one hand, and the central 

 and primary interradials on the other, develop over cavities corresponding respectively 

 to the posterior and the right lateral (epigastric) coeloms of Solaster. In Solaster 

 the central and the primary interradials are represented by the plates of the middle 

 area above mentioned, but I was unable to discover any such orderly arrangement of 

 these plates as would make it possible to recognise among them the homologues of the 

 typical series. They numbered 25 or 30 in all at the time when there were five pairs 

 of terminals. 



Similar plates now begin to appear all over the marginal area, and a plateless 

 interval can no longer be made out. Just prior to the formation of the mouth the 

 total number of plates in the whole aboral body-wall may be put down as over a 

 hundred. They are uniformly distributed and of practically uniform size, but still 

 sufficiently small to leave much uncalcified space between them. Meantime the 

 aboral surface of the disc has been enlarging with greater rapidity than the oral one, 

 so that the former now curls round at the margin, particularly in the interradii and 

 along the sides of the rays. This brings the marginally situated plates into contiguity 

 with the adambulacrals, when these develop in their turn. 



All the plates now increase in size at a rate out of proportion to the general growth 

 of the disc, so that the adjacent edges tend to meet or overlap, the interspaces left 

 being few and small. It is in one of these interspaces that the anus opens. 



The plates are now disc-like as seen from above, and their margins, instead of 

 showing numerous sharp projecting points as during the earlier period of active 

 growth, are for the most part smooth and adapted for movement against or over one 

 another. 



Still later (fifth month) larger interspaces begin again to appear on the disc and 

 arms, but before these have reached any considerable size a secondary plate develops in 

 each which, by gradually increasing in size, particularly in one direction, comes to 

 bridge across the space and finally to subdivide it, thus giving rise to the first meshes 

 of the skeletal reticulum. No doubt succeeding meshes are added in a similar fashion, 

 as the area of the aboral body-wall continues to increase. 



In the adult the form of the plates has changed. Many of them (doubtless 

 including all the original, as well as the earlier secondary plates) have developed four or 

 five angular processes for articulation with their neighbours, and form principal nodes 

 in the skeletal network. The rest, much more numerous and no doubt including the 

 more recently formed elements, remain somewhat rod-like in shape, since they have only 



