430 WILSON. [Vol. VI. 



mesoblast axis is rotated through 90°. The same is true of the 

 neural axis. 



The truth of this statement becomes at once apparent on 

 considering the development of LopadorhyncJius, where the ro- 

 tation of the axis is effected very late in the development. In 

 the young lava (Diagram V, D) the mesoblastic and neural 

 bands are horizontal and parallel to the prototroch {cf. Kleinen- 

 berg, No. 14). In the course of the metamorphosis both are 

 gradually rotated by the downward growth of the dorsal region 

 {l.d., m.d.) until they are at right angles to the prototroch, 

 always, hoiuever, remaining parallel to each other ; i.e. the shifting 

 of the neural and of the mesoblastic axes takes place at the 

 same rate. I call special attention to the fact that the rotation 

 of the axis is not ideal, but actual ; it is caused by the more 

 rapid growth of the dorsal region, which as it were swings the 

 whole ventral plate into place. In E2ipomatns the process is 

 somewhat obscured owing to the early differentiation of the 

 primary mesoblasts and to the fact that the mesoblast-bands are 

 not formed until after the axial shifting has taken place. More- 

 over, the precise history of the neural plate has not been made 

 out. When these facts are taken into account, however, a study 

 of Hatschek's figures shows that the axial relations agree essen- 

 tially with Lopadorhynchns. 



Viewed in the light of these comparisons the axial relations 

 of Nereis are easily intelligible. The primary mesoblasts, as 

 usual, lie at the posterior lip of the blastopore, and a com- 

 parison of the various figures {e.g. 34, 51, 56) will show that 

 during the closure of the blastopore they are shifted somewhat 

 downwards and forwards, as is proved by their position with 

 reference to the cross-furrow. The actual change of position 

 is, however, very slight, since, as the figures show, the mesoblast 

 {second proteloblast) is from tJie first prevetited from occupy iizg its 

 usual position by the huge mass of the first somatoblast. It is 

 accordingly segmented off from the macromere in nearly its 

 definitive position. ^ 



The shifting of the mesoblast-axis is in fact brought to pass 

 primarily by the rotation of the cleavage-spindle {i.e. the plane 

 of division), by which the somatoblast is separated from the 



1 The justice of this statement will, I think, be realized after an examination of 

 Figs. 47 to 50. XVII. 



