36 ROBERT T. HANCE. 



by additions, as the total length of the chromosomes of all com- 

 plexes, regardless of the number of elements involved, falls with- 

 in the total length limits established for the spermatogonial 

 chromosomes. Furthermore, it appears to be the chromosomes 

 of the long end of the series that are breaking up. In the normal 

 complexes the chromosomes exist in pairs (not obvious until 

 arranged according to length) and there is a relatively constant 

 length difference between pairs. The normal complexes of the 

 soma and of the spermatogonia agree in all particulars. 



The fragments are not lost but continue as a part of the com- 

 plex. They are never aligned with other chromosomes, as in 

 (Enothera scintillans, in such a way as to suggest that they may 

 be a part of those chromosomes separated from them by a clear 

 space. The fact that in complexes made up of more than forty 

 chromosomes fewer V-shaped elements are found led to a tenta- 

 tive suggestion in a recent paper (Hance, '17) that breaks oc- 

 curred at the point of fibre attachment, i. e., at the apex of the V. 

 This gains some confirmation from the fact that the fragments 

 (those chromosomes lying below chromosome 40 in the length 

 series) are approximately the length of the short arm of the V. 



GENERAL CONSIDERATIONS. 



It is not possible, at present at least, to state the cause of frag- 

 mentation. Since the condition was found in a plant known to 

 be unstable and in the pig which may very likely be genetically 

 impure led to the suggestion (Hance, '17) that the phenomena 

 might be in some way associated with hybridity. Such close 

 breeding forms as the grasshoppers and the mosquitos are known 

 to have a constant number of chromosomes in their body cells. 

 There seems to be other evidence sufficiently strong and con- 

 vincing on the other side to make the above idea a doubtful 

 possibility rather than a reality. The chromosomes which break 

 up may be multiples such as McClung and others have described 

 but I think that there is considerable evidence against this. 

 There is not enough known of somatic cytology, at present, to 

 permit any sweeping generalizations. 



Certain facts of chromosome behavior which are of general 

 interest were disclosed through the employment of metrical 



