36 DE. J. p. GEMMILL ON THE DEVELOPMENT OF 



rendered less unaccountable when one considers the secondary communications which 

 in all three species make their appearance normally or otherwise between the anterior 

 ccelom and the tips of the dorsal or ventral horns of the posterior coelom. 



The subsequent history of these diverticula is similar to that which Macbride has 

 described in the case of Asterina (15). They become constricted off from the posterior 

 coelom as small cavities lying one in each interradius between the oral (left larval) 

 body-wall and the hydroccele ring. They next become U-shaped, the basal part of 

 each U being turned towards the centre of the oral surface, while the limbs stretch 

 out along the rays. Each ray thus has a pair of radial perihsemal canals which are 

 derived from adjacent diverticula. In each pair the radial perihsemal canals lie 

 close together on the deep side of the radial nerve-tract. Deeper still, and separated 

 by a layer of intervening tissue, is the radial canal of the hydroccele. 



The basal parts of the periheemal diverticula lengthen out, and, abutting on one 

 another, form, as was shown by Macbride, the series of segments of a circle which 

 constitutes the external oral circular sinus. They are at first closed off from one 

 another, and remain thus for the first three months or so. In the adult their radial 

 extensions communicate with each other between successive pairs of sucker-feet, and 

 also proximally to the first pair. 



Some time after the fourth month, a short vertical branch from each of the peri- 

 haemal pouches appears, passing aborally in the tissue of what will afterwards be the 

 interbrachial septum, and apparently ending in the tissue-spaces there. In the adult 

 S. endeca I have described the occurrence of such a perihsemal branch in each of the 

 interradii. It is unaccompanied by haemal tissue (8 «). 



Genital MacMs and Ahoral Perihcemal Ring. — A small pocket, the bottom of which 

 is lined by primitive germ-cells, forms during the middle stages of metamorphosis 

 from the tip of the dorsal horn of the posterior coelom and pushes its way between 

 the madreporic vesicle and the axial sinus in the Starfish horizontal plane. The 

 germ-cells grow also downwards a short distance in the core of the axial organ, but 

 the cavity of the pocket does not accompany this extension. The pocket seems to 

 become the aboral perihfemal sinus, as described by Macbride in Asterina ; but I was 

 unable to trace it, even in my oldest specimens, across the madreporic interradius, and, 

 indeed, in these specimens it was only just beginning to extend, together with the 

 contained genital rachis, on to the body-wall beyond the axial complex towards the 

 sinistral side as viewed from the aboral aspect, i. e. towards ray II. In the adult 

 ;S'. endeca there occurs an interruption of the aboral perihsemal sinus in the madreporic 

 interradius (8 a). 



(3) The Enteron. 

 It is the middle chamber of the young Solaster larva which gives rise to the 

 euteron. The mode of separation of this chamber presents certain peculiarities, 



