No. 3-] ' THE CELL-LINEAGE OF NEREIS. 431 



are in a curved and not a plane layer, and differ more or less in 

 size. Still the hexagonal form may be taken as the type with- 

 out serious error. The arrangement of regular hexagons (as 

 for instance in the honeycomb) may be roughly character- 

 ized as alternate, since, as shown in the accompanying diagram, 

 the cells of each row alternate with those of the adjoining rows. 

 Sooner or later this arrangement 

 is assumed by the blastomeres of 

 all segmenting ova, and the dif- 

 ference between the radial and 

 the spiral types is caused simply 

 by the earlier appearance of the 

 alternation in the latter, as a re- 

 sult, undoubtedly, of different me- 

 chanical conditions. In the true 

 radial type, beautifully shown in 

 the cleavage of EcJiiniis and Synapta (Selenka, No. 23), it does 

 not appear until a late stage. In the spiral type it appears 

 more or less distinctly from the beginning of development. 

 This is obvious as far back as the eight-celled stage (Diagram 

 VIII, E), but it is easy to show that the peculiarities of the 

 four-celled stage are due to the same cause. The cross-furrows 

 of this stage do not exist in the true radial type (Diagram VIII, 

 A). In the spiral type they are owing to an actual or virtual 

 displacement of the cells, so that they assume the cross-form 

 shown in Diagram VIII, B, C, D, the diagonally opposing cells, 

 A and C, lying above, and the others, B and D, below. This 

 displacement may be characterized as a rotation, which is pri- 

 marily in the vertical transverse plane, as may be seen from a 

 comparison of Diagram VIII, A, B, C. In typical cases, of which 

 Sagitta (Hertwig, No. 13) and Asterina (Ludwig, No. 16) are 

 examples, the rotation is complete and symmetrical, the upper 

 and lower cross-furrows being equal and at right angles to each 

 other. Many gradations exist between this and the true radial 

 type, one of the most interesting being the embryo of Hydroides, 

 in which the two cross-furrows are equal and at right angles to 

 one another, but so short as to escape any but the closest exami- 

 nation. It appears to be invariably the case that in telotethical 

 ova, with unequal cleavage, the cross-furrows are unequal, the 

 upper being more or less reduced, as in the embryos of many 



