CAUSES OF STERILITY IN THE MULE. 21 



chromosomes at this stage is the fact that the bivalent ones can 

 as a rule be readily distinguished from the univalents, and the 

 loss of univalents can be accounted for by a proportional increase 

 of the bivalents. If, for example, forty-one univalent chromo- 

 somes can be detected, it means a loss of ten others. This 

 loss of ten univalent chromosomes can usually be accounted for 

 by the presence of five large or bivalent chromosomes, making a 

 total of forty-six. If only thirty-one chromosomes can be 

 counted, which means the disappearance of twenty univalents 

 as such, one can usually find ten large bivalent chromosomes, 

 making a total of forty-one. This part of the problem was so 

 fascinating that the chromosomes of hundreds of cells, in which 

 they were well scattered, were carefully counted and an attempt 

 was made to account for the loss or gain, depending on the type 

 of chromosomes that were considered first. In some cases those 

 which appeared to be univalent were counted first and then the 

 bivalents. In other cases those which appeared to be bivalent 

 were counted first and then the univalents, and it is surprising in 

 what a large percentage of cases the necessary number of fifty- 

 one, in terms of univalents, was actually obtained. 



Occasionally a discrepancy of one or two, and in rare cases,, 

 three bivalent chromosomes, was noted, or in other words, it was 

 sometimes impossible at first sight to distinguish all of the 

 bivalent chromosomes. Even in cases when the individual chro- 

 mosomes were carefully examined in regard to their single or 

 double nature, one could not always be positive whether he was 

 dealing with a single or a compound body. This fact, however, 

 should not be so very surprising because two of the smallest 

 type of single chromosomes may fuse to form a bivalent one of 

 practically the same size as some of the larger univalent chromo- 

 somes. The matter suggests rather strongly that it is not 

 always the same chromosomes that fuse, even in cases where the 

 same counts are obtained. On the other hand there is a possi- 

 bility of some of the large chromosomes being tri- instead of 

 bivalent, since the fusion of three threads, as well as three chro- 

 mosomes, was observed in several cases (Fig. 12). In cells in 

 which decay is noticeably under way it frequently occurs that 

 many chromosomes fuse together and form a large body, but this 

 cannot be regarded as fusion in the same sense as synapsis. 



