578 M. W. MACBBIDE. 



make tlieir appearance. Tliey are supported by braiiclies 

 from the rods which support the antero-latei-al arms. At 

 the same time the posterior ends of both the right and the left 

 anterior coslomic sacs become thickened and the thickenings 

 are hollowed out so as to form vesicles (fig. 37 a). These 

 vesicles are the rudiments of the right and left hydrocceles 

 respectively. On the left side the vesicle becomes constricted 

 from the left anterior coelom (figs. 10 and 11), so that it is 

 connected therewith by a naiTow neck only, which is the 

 rudiment of the stone-canal. On the right side no such con- 

 striction takes place, but the cavity of the right hydrocccle 

 becomes entirely shut off from that of the right anterior 

 ccelom (figs. 38 a and /;). Bury supposed (5) that the 

 left hydrocoele was formed by an outgrowth from the left 

 posterior coelom. This supposition is an exceedingly 

 natural one, for it corresponds with the appearances presented 

 by slightly older larVcD than those which we are at present 

 discussing. The left posterior ccfilom grows forward so as to 

 become closely apposed to the rudiment of the future hydro- 

 code, and an observer who had not an unbroken series of 

 stages before him might easily imagine that the hydrocoele 

 was being nipped off from the posterior coelom. With 

 regard to the right hydrocoele, this is larger relatively to tlie 

 left rudiment than it is in the case of either Aster in a 

 gibbosa or Echinus esculent us. In both the last-named 

 forms it occupies from the beginning a position near the mid- 

 dorsal line, for which reason Masterman (20) calls it the central 

 coelom and identifies it with the pericardial cavity of 

 Balanoglossus. Goto (11) also on the same ground denies 

 that it is the fellow of the left hydrocoele. But in Ophiothrix 

 fragilis there can be no doubt on the subject; the structure 

 in question is on the right side from the beginning and 

 remains on the right side, and occasionally, as Miiller (25) has 

 .figured and as is shown in fig. 53, takes on a five-lobed form. 

 Shortly after this period when all the larval arms have 

 attained their full length, the left hydrocoole begins to take 

 on a five-lobed form (figs. 12 and 39). The five lobes soon 



