10 DR. J. F. GEMMILL ON THE 



ccelom, as well as the left and right anterior and middle coelomic 

 regions (see 4, p. 245). The left posterior ccelom gives rise to 

 the main part of the adult perivisceral cavity, and in the late 

 larva the circulation of the fluid is from the ventral round to the 

 dorsal horn of the ccelom in question, then through the left and 

 right middle and anterior coelomic regions back to the ventral 

 horn. At metamorphosis, the dorsal and ventral horns become 

 united and are closed off from the other coelomic regions named, 

 which for the most part become obliterated, the left middle one, 

 however, giving rise to the hydroccele. The circuit thus becomes 

 limited to the perivisceral ccelom and the flow is naturally dextral 

 as viewed from the aboral aspect. These points are illustrated 

 in the appended text-figure. 



The ciliation of the perihsemal sinuses in the adult is in agree- 

 ment with the fact of their direct enteroccelic origin. 



As regards endoderm, the greater part of the oesophagus of 

 the larva is retained at metamorphosis, and gives rise to the 

 pharyngeal portion of the adult gastric cavity. The larval 

 oesophagus is ciliated towards the stomach, and we note that the 

 pharyngeal portion of the adult gastric cavity is similarly ciliated. 

 Within the stomach of the early larva, food-particles are churned 

 round dextrally as viewed from its anterior extremity (4, p. 240). 

 Something similar occurs within the gastric cavity of the adult 

 (p. 7). In late larvae, when the stomach has readied full size, 

 oralward currents appear over a large part of its fundus. These 

 are probably represented in the adult by the oralward and centri- 

 fugal currents, wdiich one finds on the interiadial portions of the 

 stomachal and pyloric sac-regions of the gastric cavity. 



II. Ciliation and Feeding. 



In Asterids a large part of the actinal epiderm is derived 

 phylogenetically and also in many cases ontogenetically (Solaster 2, 

 Asterias 4) from the preoral lobe and from the region which 

 in the feeding types of larvae (Asterias) surrounds the larval 

 mouth and includes the circumoral or food-gathering area (4, 

 p. 240). No doubt the bilateral ancestor of Echinoderms obtained 

 its nutriment through ciliary action, as is still done by all feeding 

 Echinoderm larvae and by the Crinoids. The oral or centripetal 

 currents which one finds in the ambulacral grooves of starfishes 

 are probably of ancestral origin. Recently, certain observations 

 raised the question in my mind whether particular starfishes do 

 not still obtain a portion of their nutriment through ciliary 

 action. 



PORANJA PULVILLUS. (PI. I. fig. 1.) 



The observations in question were concerned in the first place 

 with the pin-cushion starfish, Porania pulvillus (0. F. M.), their 

 starting-point being the fact that in investigating the actinal 

 ciliation in this starfish with the help of suspended carmine 



