TEIE ORIGIN OF SPECIES 



bers might go to each pole, or any intermediate result might occur. The 

 other division is usually equational. Most of the gametes so produced are 

 inviable, but there is a possibility that the genomes from the parent spe- 

 cies might be accurately separated by random division, with fertile 

 gametes resulting. There are authenticated records of several mule mares 

 vi'hich have produced offspring. And in plants, the inclusion of all of the 

 chromosomes of the hybrid within a single gamete may also result in 

 viable offspring, which, however, will be polyploid ( see Chapter 19 ) . 



Yet the failure of the chromosomes to synapse cannot be the whole 

 explanation of hybrid sterility, for the reproductive systems of species 

 hybrids are often grossly abnormal. In hybrids between Drosophila mela- 

 nogaster and D. simulans, the gonads are rudimentary, and meiotic divi- 

 sions never begin. In other cases, the chromosomes actually do synapse, 

 but the hybrid is sterile anyway. Dobzhansky has described a particularly 

 instructive case in hybrids between Drosophila pseiidoobscura and D. 

 persimilis. If the proper strains of these two species are selected for hy- 

 bridization, synapsis in the hybrid will be complete; but other strains give 

 partial synapsis, or none at all. If failure of synapsis were the essential 

 factor in hybrid sterility, then the first group of hybrids should be fertile, 

 the second partially fertile, and the last highly sterile. But in point of fact 

 the outcome is the same in every case: the first maturation division pro- 

 ceeds normally up to the metaphase, but the anaphase is grossly abnormal 

 and results in a single, binucleate cell. The second division does not occur 

 at all, and the giant spermatids degenerate. It is evident that hybrid steril- 

 ity may result from a derangement at any point in the long and compli- 

 cated series of processes which extend from the zygote to the mature 

 gametes for the production of the next generation. 



Whenever synapsis of the chromosomes fails completely, no insight into 

 its causes can be obtained. But paitial synapsis frequently occurs, and 

 this may permit a study of the factors which prevent its completion. It 

 frequently turns out that chromosomal rearrangements are present, and 

 that these place purely mechanical obstacles in the way of completion of 

 synapsis. Horton's study of synapsis in hybrids between Drosophila mela- 

 nogaster and D. simulans is most illustrative. The meiotic chromosomes 

 of this hybrid are not available for study because the gonads are rudi- 

 mentary, as was mentioned above. However, there is every reason to 

 believe that the salivary gland chromosomes, which are well developed in 

 the hybrid, give an accurate picture of the synaptic behavior of tlie chro- 

 mosomes of the gonads. Any rearrangements which the chromosomes of 

 the two species show with respect to one another may be assumed to have 

 arisen since their separation but there is no way of judging which species 

 has the more primitive arrangement. 



In general, the salivary gland chromosomes of this hybrid are well 

 synapsed, but there are a considerable number of regions in which synap- 

 sis is irregular or lacking (Figure 103). In these regions, ten rearrange- 

 ments have been identified with certainty. Six of these are inversions, five 

 of them being quite small ( two to twelve bands in length ) and one rather 

 large (involving nine sections of the third chromosome, and including 



280 



