No. 3-] THE CELL-LINEAGE OF NEREIS. 411 



2. The Second Somatoblast. Origin of tJie Mesoblast-bands. 



We may now return to the end of the spiral period of cleav- 

 age (Fig. 29) and follow the history of the second somatoblast 

 {M). Its mode of formation is shown in Fig. 47 (PI. XVII). The 

 spindle, which lies in the granular portion of the macromere D, 

 is inclined at an angle of 45° to the second cleavage (median 

 line), and as the somatoblast forms, it is pushed over towards 

 the right so as to lie nearly in the middle line in front of the 

 first somatoblast. As a rule the division is completed before 

 the formation of x^, but I have seen some specimens in which 

 the two cells were simultaneously formed, and one (Fig. 33) in 

 which the somatoblast was completely separated before the 

 appearance of the A'^-amphiaster. ^\\q first division of the soma- 

 toblast cuts it into two equal parts. These are the primary 

 mesoblasts {M. M., Figs. 49, 50). They are invariably dis- 

 placed towards the left — probably, as I conjecture, because on 

 the right side (Fig. 48) the second somatoblast abuts against 

 the first and there is less resistance towards the other side. 



At the second division each of the primary mesoblasts buds 

 forth a small cell at the surface, near its anterior margin (Figs. 

 53, 72. The lessened diameter of the cells in the former figure 

 is owing to their rounding out ; in their quiescent state they 

 are considerably flattened against the entomeres). Examina- 

 tion of many specimens shows that there is a good deal of 

 variation in the direction of the spindles, and I found two 

 specimens in which one of the spindles was longitudinal, the 

 other exactly transverse. The budding of the mesoblasts is 

 continued in the same way for a considerable period, until a 

 group of small cells is formed, which lie superficially and form 

 the posterior lip of the blastopore (Figs. 54-57, 73). They 

 become continuous with the anterior cells of the ventral plate, 

 but, as has been already mentioned, they can usually be dis- 

 tinguished from the x-cells, even before the appearance of the 

 pigment, by the granular appearance of their nuclei. 



Through this process and by the forward growth of the ;tr-cells, 

 the primary mesoblasts are soon covered in, and lie quite be- 

 neath the surface. At the same time the blastopore closes in 

 completely (by a process to be described later). A longitudinal 

 (actual) section through this stage (passing slightly to one side 



