3 



INTRODUCTION. 



It is generally conceded that the segments of each chromosome are 

 sisters, and it is conceivable that, no matter in what manner or when 

 the daughter spirem may segment during division, the spindle fibers, 

 or those parts of the cell which have to do with the arrangement of 

 the chromosomes in the nuclear plate, are able to bring the sister seg- 

 ments together in pairs. 



Strasburger, Guignard, and others regard each long loop or turn of 

 the daughter spirem as representing a V or U of the preceding mitosis, 

 and that, consequently, the spirem segments exactly as it was con- 

 structed, i. e., the chromosomes simply separate at the points marking 

 the free ends of the V's and U's. The spirem accordingly breaks up 

 into pieces equal to the length of two segments or two granddaughter 

 chromosomes. It is claimed by Strasburger (1900, pp. 23, 24) that 

 these V's or U's are fastened to the spindle in the same manner as in 

 the first division, namely, at the angles or at the place of bending. 



Theoretically, there may be little objection to this view. The vast 

 majority of facts, however, show that there is no such regularity in 

 the shape of the chromosomes, or in their manner of attachment to the 

 spindle. We have seen that, in the daughter nucleus, the identity of 

 the individual chromosomes cannot be recognized, and we do not 

 know whether the spirem segments in the same manner in which it 

 was constructed. 



But if the spirem should segment by transverse division at the points 

 marking the angles of the V-shaped chromosomes instead of at the free 

 ends, then it is clear that the two segments of each chromosome would 

 not be sisters. The result might be that two or more sister chromo- 

 somes would go to the same daughter nucleus, a condition that might 

 furnish a basis for greater variation. We cannot prove either propo- 

 sition, and the author is not disposed to enter into any speculation here 

 upon the subject. The observed facts are these : The identity of the 

 individual chromosomes is lost in the daughter nucleus, and we do not 

 know whether the segments of the respective chromosomes appearing 

 in the nuclear plate of the second mitosis are sisters or not. There is 

 also no basis in fact for the conclusion that one chromosome is heredi- 

 tarily different from another. 



The first two nuclear divisions in the embryo-sac mother-cell, so far 

 as is known, are quite similar and homologous to those in the pollen 

 mother-cell. In Lilium mariagon, the species more carefully investi- 

 gated by the author, there is no important difference in the behavior of 

 the chromosomes. It may be mentioned, moreover, that the daughter 

 nuclei resulting from the first mitosis approach more closely the resting 

 condition than in the pollen mother-cell. 



