No. 3-] THE GASTRULATION OF AMPHIOXUS. 577 



There are, in fact, two ways in which we may think of the 

 changes shown in Figs. VI-X and XI-XV as taking place. 

 The wall that grows backward to close the blastopore may 

 either simply elongate as a result of cell growth, and the 

 bending at the anterior end remain fixed, or the growth back- 

 wards might be due to a gradual shifting of the part of 

 greatest bending along the opposite side ; in consequence of 

 this the dorsal wall would increase in length at the expense 

 of the ventral in one case, and the ventral at the expense of 

 the dorsal in the other. It is also possible that both a change 

 in the bending and cell growth might take place at the same 

 time. 



In a fourth series we have tried to represent our own idea of 

 the method of closure of the blastopore. In the first figure of 

 this series, XVI, the gastrula is represented as being somewhat 

 unsymmetrical. The dorsal lip is at c, the ventral at a. The 

 point of greatest bending of the gastrula is near the center, but 

 a little toward the ventral side. At a later stage, XVII, both 

 dorsal and ventral sides of the gastrula have come nearer 

 together, and the embryo has, in consequence, become longer. 

 The dorsal and ventral sides are of about equal length, and the 

 point of greatest bending is opposite the blastopore. In a later 

 stage, XVIII, the walls grow longer and the dorsal side flattens 

 somewhat. The opening of the blastopore is still opposite the 

 point of greatest bending. In the last stage, XIX, the ventral 

 lip grows faster and the blastopore opens more on the dorsal 

 side. Fig. XX shows the method of closure as seen from 

 the posterior side. The series of concentric circles represent 

 the successive stages. The final stage is excentric, owing 

 to the more rapid growth of the ventral lip. 



If instead of a dorso-ventral series we had made use of a 

 lateral one, that is, one from right to left, to illustrate our 

 idea of the method of closure of the blastopore, we would have 

 given a series exactly like that drawn in I-V. Even in the 

 dorso-ventral series our idea of the method of closure corre- 

 sponds more nearly to that of the first series than to any other ; 

 the points of difference being : first, the unsymmetrical gas- 

 trula at the earliest stage (XVI) ; second, the flattening of the 



