1904.] NATURAL SCIENCES OF PHILADELPHIA. 711 



tropic, as it should be, following the law of alternating cleavage (fig. 29). 

 While this division is in progress, the lateral cells of the second quartette 

 are again in division (figs. 26 and 27). This division is leiotropic; its 

 result is that a cell similar to but smaller than 2a^ and 2c^ is placed to 

 the left of these cells. Lastly, 2b^ divides similarly, the products being 

 shown in fig. 41. For the six cells thus formed from the second quar- 

 tette, 2a^, 2a^-^, etc., I have adopted Mead's name of "secondary trocho- 

 blasts/' though they arc not precisely identical in origin with those 

 to which Mead gave this name. At a stage of about 81 cells the second- 

 ary trochoblasts 2a^ and 2c^ divide leiotropically and equally (figs. 34 

 and 35). No doubt 2b^ also follows suit, though I have not seen this 

 division. 



The origin of the "accessory trochoblasts" la^-^-^, etc., by the dexio- 

 tropic division of the intermediate girdle cells has already been de- 

 scribed. There is now an irregular row, formed of 22 small cells, 

 which encircles the embryo at a level just between the first and second 

 quartettes. These cells are arranged as follows, passing from left to 

 right: ld2-2-kP-i-2a2-»-2ai-2-2ai-i-la- ■^ etc. It is very probable that 

 the accessory and primary trochoblasts soon divide again, since in 

 the embryo from which fig. 50 was drawn Ic^-^ and Ic^-^-^ were both 

 undergoing an equal and meridional division. The row" of cells thus 

 formed is clearly distinguishable up to the closure of the blastopore, 

 and it was by means of this row as well as by the large size of 2a---' and 

 2c^-^-' that I have been able to trace the boundary of the first quartette, 

 as shown in text fig. IV, A. The trochoblasts are very transparent, 

 and become elongated in the direction of the cell row^ 



At the stage shown in fig. 55 the prototroch is nearly transverse to 

 the long axis of the embryo, as is also shown by the dotted line in text 

 fig. IV, E, but in front of the blastopore it bends sharply forward in a 

 semicircle. In these figures the trochoblasts are represented by the 

 most anterior row of that group of small cells lying just anterior to 

 the blastopore. Beyond the stage represented in fig. 55 I have not 

 been able to trace this group, since all the cells of the ectoderm, through 

 rapid mitosis, soon become nearly uniformly small in size. In fig. 44, 

 where the trochoblasts were last identified, they were on the ventral 

 side, but a short distance anterior to the blastopore, and j^assed up on 

 each side in a row nearly transverse to the long axis of the embryo (cf. 

 text fig. IV, E). Since the stomodseum appears at the point where 

 the blastopore closed and is then pushed forward through growth of 

 the ventral ectoderm (ventral plate), it follows that the trochoblasts 

 must still remain as a more or less transverse row passing in front of 

 the stomodseum. At the close of the eml^rvonic life, when the trans- 



