552 A. C. WALTON 



division the octads are separated into two daughter tetrads by 

 a spHtting along one of the previous longitudinal constrictions. 

 In this case the Querkerbe corresponds with the plane of telo- 

 syndesis. In a Sycon sponge, hkewise, a similar longitudinal 

 division of the double tetrads is found (Jorgensen, '10). In A. 

 megalocephala the first division is generally regarded as being 

 longitudinal, and perhaps as being foreshadowed by the longi- 

 tudinal split noted by Bonnevie ('08). Kiihtz ('13) reports a 

 longitudinal division of the chromosomes at the formation of the 

 first polar cell in several species of the Sclerostomum of the 

 horse. 



c. Anaphase. The daughter chromosomes (fig. 38), as they 

 begin to separate, have their long axes in a plane parallel to the 

 equatorial plate. As division progresses (fig. 39) each chromo- 

 some changes its orientation by rotation around an axis which 

 is perpendicular to the axis of the spindle, until its long axis be- 

 comes parallel to the spindle fibers. This fact seems to argue 

 against the active influence of the spindle fibers in the division, 

 a,s they were originally attached to the sides, not the ends, of the 

 ^chromosomes. The fibers are somewhat more accurately to be 

 regarded as being mere passive formations of the alveolar walls 

 of the karyoplasm acting in response to forces such as might 

 operate between two electrical centers, that is, the centrosomes. 

 That the eighteen daughter chromosomes of each plate are 

 tetrads, is shown by a comparison of edgewise (fig. 40) and face 

 (fig. 41) views of nuclear plates of the same stage, in which they 

 appear respectively as tetrads and as dyads. As a result of this 

 division into daughter chromosomes the Querkerbe, without any 

 shifting of its position in relation to the chromatic matter, now 

 lies in the plane of the short axis rather that in that of the long 

 axis of the tetrad; but by the ninety degree rotation of the tetrad 

 its plane is now perpendicular to the axis of the spindle, whereas 

 in the beginning it was parallel to it. That constriction in 

 each of the tetrads which is parallel to the axis of the spindle is 

 very plainly marked (fig. 40) and represents the original plane 

 of parasj^ndesis. Inasmuch as this first division is longitudinal, 

 and each daughter tetrad is, as far as can be determined, exactly 



