38 DE. J. F. GEMMILL ON THE DEVELOPMENT OF 



present convex surfaces to each other. These surfaces now flatten, at the same time 

 increasing in area. Next they become slightly concave, so that, in each, anterior and 

 posterior lips can be distinguished. The two posterior lips now unite across 

 the isthmus, the fusion advancing from the dorsal towards the ventral side. The two 

 anterior lips also unite, but in this case the fusion begins ventrally. 



Dorsally the greater size of the right enteric fold causes its anterior lip slightly to 

 overarch the corresponding lip of the left fold. The opening between the middle and 

 the anterior chamber is thus gradually reduced to a slit situated dorsally and looking 

 to the left. On the other hand, the slit to which the opening between the middle 

 chamber and the posterior coelom gradually becomes reduced is situated ventrally, 

 but the early predominance in this region of the left enteric fold is not sufficiently 

 well maintained to cause this opening to be deflected towards the right, and accordingly 

 it remains practically in the middle line. 



The processes above described convert the enteric isthmus into a short flattened 

 canal leading ventrally and posteriorly, with a slight inclination to the right, from the 

 anterior to the posterior coelom (PL V. fig. 49). The canal now separates itself off 

 completely from the cavities in front and behind, the anterior aperture remaining open 

 slightly longer than the other. On the whole, the right gastric fold contributes more 

 than the left one to the total area of the enteron. One expression of the difference is 

 seen in the slight pouching to the right of the wall of the enteron, which is indicated 

 in PI. III. figs. 31 & 32. This pouching is much better marked at a corresponding 

 stage in the development of S. pajyposa, and for purposes of comparison a drawing 

 from such a stage is given in PI. III. fig. 38. 



After the separation of the enteron is complete, it is no longer possible to locate its 

 former junctions with the anterior and posterior coeloms, so as to be able to trace their 

 position during and after metamorphosis. The growth of the posterior coelom with 

 its two horns induces the enteron to assume the form of a crescent lying within that 

 of the coelom in question. Dorsal and ventral horns may accordingly be distinguished, 

 and the tips of these, like the corresponding parts of the posterior coelom in the course 

 of metamorphosis, come to project into the disc arm-rudiments which we have 

 numbered 1 and 2 (p. 22). The shape of the enteron as seen from the side is now 

 that of a disc with a deep notch opposite the region of the preoral lobe, i. e. between 

 arm-rudiments 1 and 2. In course of time this notch becomes filled out, owing to 

 continued expansion of the enteron and to reduction in size of the strtictures (axial 

 complex) occupying the notch, in much the same way as the similar notch in the 

 epigastric coelom (p. 27) gradually disappears. 



By this time the enteron shows six or seven short outgrowths into the arm-rudiments. 

 The two largest of these are constituted by the tips of the horns of the enteron, which 

 have been already referred to as entering arm-rudiments 1 and 2. The others are 

 in connection with succeeding arm-rudiments. The last two (8 and 9) are latest in 

 formation, like the rays to which they belong. All these outgrowths are lobes of 



