28 -DE. J. p. GEMMILL ON THE DEVELOPMENT OP 



next obliterated altogether during the process whereby aboral arm-rudiment 1 travels 

 rio-ht across the region where the neck existed in order to reach its final position over 

 the first radial pouch of the hydroccele. The internal cavity of arm-rudiment 1 is 

 thus derived from the tip of the (larval) ventral horn of the posterior coelom (p. 32). 



Left Lateral Diverticulum of Anterior Coelom. — The bottom and sides of the left 

 lateral diverticulum give rise to the hydroccele and to part of the internal oral circular 

 sinus, while the central portion of the diverticulum becoming obliterated assists in 

 providing the area on which the mouth afterwards opens. The development of the 

 hydroccele will now be described, but that of the internal oral circular sinus is better 

 taken under the heading of the axial coelom (p. 29). 



While at an early stage in growth the cells lining the greater part of the anterior 

 coelom and its derivatives become flattened, those which form the bottom of the left 

 lateral diverticulum remain cylindrical and crowded together. This part of the diver- 

 ticulum may be called the hydroccele groove. It is in the form of a crescent with 

 dorsal and ventral horns, the convexity being directed backwards. The concave side 

 looks forward into the main cavity of the left lateral diverticulum. The dorsal horn of 

 the crescent ends with a definite tip raised a little from the coelomic lining and bending 

 slightly to the right of the plane in which the rest of the crescent lies. 



The wall of the hydroccele crescent and that of the posterior coelom (p. 32) along 

 with the intervening mesenchyme form a somewhat thick mesentery attacliing the gut 

 to the left body-wall. The radial pouches appear as hollow outgrowths on the convex 

 side of the hydroccele groove, and push their way into the tissue of this mesentery. 

 Increasing in size they project into the posterior ccelom covered by a lining from it, 

 and keeping as close as they can to the left body-wall. Five radial pouches appear on 

 the hydroccele crescent at practically the same time. One end of this series is at the 

 tip of the dorsal horn of the crescent, and here, as was stated previously, our numbering 

 of the series begins. The other end falls short of the tip of the ventral horn, and it is 

 to this region that the later-formed pouches are added. 



To begin with, the third and fourth rather outgrow the others. Next come the second 

 and fifth, while the first lags a little behind, probably because it has less freedom to 

 expand by reason of its position with reference to the preoral lobe and the body (p. 29). 



After the formation of the first five pouches there is a short pause before the sixth 

 makes its appearance. It does so just beyond the fifth on the ventral horn of the 

 crescent, and is succeeded after a similar delay by pouch VII. By this time the free- 

 swimming stage is at an end, so that pouches VIII and IX remain to be added during 

 the course of metamorphosis. 



Meantime the hydroccele groove in the region of pouches III and IV has become 

 nipped ofi" from the rest of the left lateral diverticulum, forming a short canal open at 

 both ends and gradually extending dorsally and ventrally along the crescent. At the 

 time of fixation, conversion of the groove into a canal is not quite complete for the 

 whole of the dorsal horn, and for the ventral horn is accomplished only as far as the 



