FORMATION OF THE FIRST SPERMATOCYTE CHROMOSOMES 579 



These facts are somewhat contrary to the view expressed 

 by Baumgartner ('04) for Gryllus. In several species of this 

 genus he describes each chromosome as having a definite shape 

 and definite method of condensation into the spermatocyte 

 chromosome. In Forficula, as shown above, some of the chro- 

 mosomes may or may not pass through the ring stage. The 

 rule however, seems to be the ring formation. Baumgartner 

 brings his facts to the support of the hypothesis of the genetic 

 continuity of the chromosomes, and I think rightly so. I do 

 not consider the fact that a chromosome may or may not pass 

 through the ring stage as evidence against such an hypothesis. 

 The end result is the same in either case and I do not see how an 

 occasional variation by the omission of a bend in the chromosome 

 can be used as evidence against the hypothesis. No such rigid 

 demands are made of any other hypothesis and certainly we have 

 no laws of development or behavior without their occasional 

 variations. If variations are found in what may be considered 

 laws of development why is such rigidity demanded of a workirg 

 hypothesis? 



DISCUSSION 



We have seen that the chromosomal variations in Forficula 

 are different from any which have been previously described, 

 and they can be explained without the assumption that Forficula 

 is a composite species made up of several smaller ones. For the 

 discussion, let us leave out of account the spermatogonial vari- 

 ations described in specimen 34-2, as they may be pathological. 

 Restricting ourselves then to the spermatocyte divisions, we find 

 not only variations in number, but also that some of the chro- 

 mosomes behave irregularly in these divisions. The variations 

 in number, as I have shown, are caused by the failure of some of 

 the spermatogonial chromosomes to pair at synapsis, and hence 

 such chromosomes appear in the spermatocyte divisions as single 

 or univalent instead of bivalent chromosomes. From what we 

 know of the behavior of chromosomes, we should expect these 

 univalent ones to divide in one of the two maturation divisions 

 and not in the other. In Forficula the behavior of these chromo- 



