MULTIPLE CHROMOSOMES 563 



4 and photomicrographs P., Y., pi. 8). Such a structure would 

 result if the V were unfolded on the spindle with fibers attach- 

 ing at each end, but in this case the parts on either side of the 

 equatorial constriction would be unequal and division at this 

 place would mean separation of whole tetrads. Because of the 

 relatively slight difference in size of the joined tetrads, I have not 

 yet been able to satisfy myself of the exact relation of parts in 

 this elongated chromosome, but I suspect that separation of 

 chromatids has occurred in a plane at right angles to that which 

 obtains in the V. A full study of the prophase history should 

 clear up this matter. 



Emphasis needs to be placed upon the statement already 

 made that the mere form of the chromosome is not conclusive 

 evidence, either of its valence or its method of division. Valence 

 can be determined only by a study of the element through the 

 various generations of cells and in numerous individuals of the 

 species. Planes of division are certainly identified only by some 

 morphological character which marks one of the synapsed 

 elements. Instances of the exactness in configuration of chromo- 

 somes of tetrad, hexad and octad valence have aheady been given, 

 but it would seem desirable to consider the ring chromosome a 

 little more at length. In former papers I have called attention 

 to the importance of such elements as indicators of the planes of 

 division, and in one ('14) I pointed out the distinction between the 

 Hippiscus and Stenobothrus types and indicated how failure to 

 recognize such distinctions has led to much confusion. At the 

 same time it was recognized that, in the presence of parasynap- 

 sis, the opening in the ring might represent either the space be- 

 tween homologous chromosomes or their longitudinal division. 

 Since there was, so far, nothing to suggest two types of ring 

 formation, it was considered that only one exists and the avail- 

 able evidence would indicate that the annular space lies between 

 homologous chromosomes. In the presence of one type of ring 

 there must necessarily be two forms of first spermatocyte divi- 

 sion, for the Hippiscus ring divides along the length of the ring 

 while the Stenobothrus type separates transversely. Owing to 

 the fact that, up to that time, all rings of the Stenobothrus type 



