THE DEVELOPMENT OF THE EARTH-WORM. 327 



become moved more on to the back and towards the oral 

 extremity, in such a way that, in some sections through the 

 posterior portion of the embryo, the transverse sections of 

 the two streaks are found at the lower part and at the upper 

 part the two large cells, together with the last part of the 

 streaks (PI. IX, fig. 12). Leaving this point, on the dorsal 

 surface, the germinal streaks descend abruptly downwards, 

 embracing a somewhat triangular space at the top of the 

 posterior extremity, and, having reached the ventral surface, 

 approach, with their convexities, both each other and the 

 median line, without, however, adhering or coming into 

 mutual contact. Figs. 11, 12, 13, and 14 of PI. IX make 

 this process of displacement quite clear. Thus approximated 

 to one another the germinal streaks stretch along the ventral 

 surface, but at the anterior part of the embryo they again 

 separate, and arching over the lateral surfaces, ascend on the 

 back to join in the cephalic commissure. 



Besides this displacement, the development of the germinal 

 streaks must cause gradually an alteration in the general 

 form of the embryo, and the more so since the streaks grow, 

 not only in length, but also in width and depth. Hence the 

 transverse section of the body loses its lens shape and be- 

 comes circular, then the ventral surface becomes more and 

 more convex, and the anterior and posterior ends, curving 

 towards the dorsal surface, this becomes depressed and con- 

 cave, so that the embryo assumes a kidney or bean shape. 



Turning to the development of the cephalic germinal 

 streak, we find the mesoderm separated completely from the 

 ectoderm, consisting of a mass of small roundish cells, which 

 fills completely the space between the ascending external 

 lamina and the descending inflexion of the fold of ectoderm. 

 Some time later two narrow fissures appear in the lateral 

 region of this mass of mesoderm, then enlarge towards the 

 median dorsal line, where they then unite with one another, 

 thus splitting the mesoderm into two concentric layers, one 

 external and one internal. But as the split begins nearer 

 the external surface than the surface bounding the oesophagus, 

 the layers are, from the beginning, of unequal thickness ; the 

 external consists nearly everywhere of a single layer of cells, 

 while the internal has two or three layers. The first adapts 

 itself to the external wall of the cephalic ring, the second 

 joins itself to the oral epithelium. 



The splitting of the mesoderm in the cephalic germinal 

 streak is followed by an analogous process in the ventral 

 germinal streaks, beginning from the front and progressing 

 gradually towards the posterior end. About this most 



