DEVELOPMENT OF CUCUMARIA. 637 



forms a spacious oral atrium into which the five tentacles project, 

 and the ring-canal now surrounds the anterior end of the gut, its 

 plane being inclined to the larval axis (as in G. lilanci) in such a 

 way that its dorsal part is nearer the anterior end of the animal 

 than its ventral. It is possible now to identify the radii and to 

 determine the point of closure of the ring. In six of my pre- 

 parations of this stage the ring is still open in the left dorsal 

 interraclius. 



Radial canals and tentacles are given oflf alternately from the 

 ring, but there is already discernible, in some individuals at any 

 rate, the beginning of that curious grouping of tentacles which 

 is found in the later pentacula. Of the radial canals the mid- 

 ventral is much the largest. It pi^ojects directly backwards from 

 the ring- canal, and has at its posterior end a rhombic dilatation 

 the two laterally directed angles of which represent the internal 

 rudiments of the primary tube-feet, which are thus, from their 

 first appearance, not terminal (PI. II. fig, 10). In C. planci, 

 C. hirchshergii, Holothitria tremula, Psolus fahricii, and Phyllo- 

 phorun urna the first two podia have been described, by various 

 authors, as arising simultaneously from the posterior end of the 

 mid-ventral radial canal, Holothuria Jioridana being exceptional, 

 in that it forms at first a single terminal tube-foot on the mid- 

 ventral canal (Edwards, 4). 



The right and left dorsal radial canals project outwards in the 

 plane of the ring-canal, and their ends have already begun to 

 turn backwards ; the lateral ventral canals are short, blunt, and 

 unbent. There is no difference in the degree of development of 

 the five tentacles. In those larvae in which the ring-canal is 

 closed a small blunt outgrowth — the rudiment of the Polian 

 vesicle — has appeared on the posterior wall of the ring at the 

 point of closure ; but whether this belongs to the dorsal or 

 to the ventral limb of the hydrocoel I find it impossible to 

 determine. 



The relations of the stone-canal are what they were in the 

 last stage described. At about the middle of its length, however, 

 there is now a slight enlargement of the lumen, caused by an up- 

 pushing of its antero-dorsal wall. This marks the point at which 

 the secondary madreporite (Madreporenblase of Ludwig) will 

 later be formed. 



On the fourth day (84 hours) the water-vascular system presents 

 an interesting transition stage, in which the three more dorsal 

 tentacles are connected with the radial canals from which they 

 spring in the adult, while the two more ventral ones still retain 

 their interradial communication with the ring-canal. As in 

 C. planci, the left dorsal radial canal has appropriated two 

 tentacles and the right dorsal canal only one — that which was 

 developed in the right dorsal interradius- The two lateral 

 ventral radial canals have no tentacles associated with them, 

 those developed in the ventral interradii being appropriated later 

 by the mid-ventral canal, to which, indeed, their bases already 



