52 DE. J. F. GEMMILL ON THE DEVELOPMENT OF 



The ambulacral and adambulacral plates appear in a typical manner ; but the aboral 

 skeleton shows numerous small uniformly distributed plates in the disc and two 

 terminals in each arm-rudiment. Small nodules for the spines appear everywhere over 

 the aboral surface. 



The ambulacral and adambulacral plates increase in size and number; while the 

 plates of the aboral skeleton increase in size, and at a much later stage in number also. 

 For a time the latter fit closely with one another ; afterwards interspaces, the future 

 meshes of the skeletal reticulum, appear between them and are bridged over by new 

 plates. The spines increase in size and form small groups, which become associated 

 with the first-formed plates. 



(0) General. 



The development of Solaster agrees in very many points with that of Cribrella (as 

 described by Masterman), including the remarkable mode of blastula and gastrula 

 formation and the even more remarkable origin of the hypogastric coelom. Like that 

 of Cribrella, the swimming larva of Solaster with its three glandular arms and sucker 

 on the preoral lobe can readily be referred to the brachiolarian type. This is all the 

 plainer since in Solaster, as I could satisfy myself by observation of living specimens, 

 the tips of the arms lose their cilia and become capable of exercising slight and 

 temporary adhesive functions. Their primary object, I should judge, in Solaster is 

 to enable the muscular sucker to be brought into play at the proper time for 

 initiating the stage of definite fixation. 



If one were looking at Cribrella and Solaster alone among Echinoderms, nothing 

 could have greater apparent reason in point of form' than Masterman's comparison of 

 the hydrocoele with the epigastric coelom, and it would be alike attractive and 

 legitimate to accept the homologies he has sought to institute between the various 

 larval cavities in Cribrella and Balanoglossus. In the case of Solaster it might further 

 be argued that the posterior annulus referred to on p. 16 gave additional evidence 

 of enteropneust affinity ; while the manner in which much of the preoral lobe, including 

 the (rudimentary) apical nervous system, became incorporated at metamorphosis with 

 the oral surface of the Starfish indicated greater ontogenetic continuity than the more 

 abrupt changes occurring in the metamorphosis of Asterina and apparently of the 

 bipinnaria. But attractive as this hypothesis is, it seems to me that it would impose 

 diflSculties of interpretation on the development of other Starfish and of Echinoderms 

 in general, greater than those from which it could relieve Cribrella or Solaster. The 

 position has been stated by Macbride (15, 16, 16 a, 16 J), and I must put myself 

 down as a supporter of his view that in Cribrella (and Solaster) we are dealing with 

 a form of development modified from that which should be looked upon as typical and 

 primitive in the Order. 



Of course, the question, being one of evidence, may yet be reopened, e. g., by the 

 discovery of a feeding bipinnaria in which the hypogastric coelom originates wholly 



