CAUSES OF STERILITY IN THE MULE. 37 



number in the ass is about sixty-five, thus making a difference of 

 twenty-eight chromosomes between the parents of the hybrid. 



4. The seminiferous tubules of the mule contain a much 

 smaller amount of germ cells than do the tubules of the horse. 

 Some of the tubules of the mule are entirely devoid of sex cells. 



5. The sex cells of the mule are larger than those of the horse 

 in the corresponding stages. 



6. All of the fifty-one chromosomes in the spermatogonial 

 cells of the mule enter the spindle for division. The mitotic 

 figures are normal and there are no straggling chromosomes. 



7. The accessory chromosome of the hybrid, which is undoubt- 

 edly maternal in origin, resembles entirely the accessory of the 

 horse, which fact shows that this sex-determining chromosome 

 retains its individuality. 



8. The period of synizesis which is so obvious in the primary 

 spermatocytes of the horse, is lacking in the mule. 



9. The spireme of the horse is also lacking in the mule, but is 

 replaced by a continuous network of chomatin threads, parts 

 of which sometimes resemble the spireme to a certain extent. 



10. There is no definite time for the pairing of threads or 

 chromosomes in the hybrid; the synaptic period begins at the 

 time the chromatin threads make their appearance and continues 

 through the prophase. 



11. The pairing of chromosomes, or pseudo-reduction, is 

 always incomplete and very inconstant. The number of chro- 

 mosomes in the late prophase of the primary spermatocytes 

 varies from thirty-four to forty-nine. The greatest majority of 

 counts of chromosomes lie between forty and forty-five. The 

 expected number, if reduction was complete, would be twenty- 

 five besides the unpaired accessory. 



12. In the late prophase of the primary spermatocytes the 

 bivalent chromosomes can as a rule be readily distinguished 

 from the univalent ones. The number of chromosomes which 

 the various cells lack in order to make the original total of fifty- 

 one, in terms of univalence, can usually be accounted for by the 

 proportional increase in the presence of bivalents in such cells. 



13. Up to the early prophase of the primary spermatocytes 

 there seems to be no necessity for the paternal and maternal 



