THE ORIGIN OF SPECIES 



valent. Ordinarily, a tetravalent divides so that two members go to each 

 pole, yielding normal gametes. But tetravalents also give a 3 and 1 distri- 

 bution. Now if only one or a few chromosomes are missing or are in excess 

 in an otherwise tetraploid zygote, it may be successful. Such trisomic and 

 monosomic strains are well known to plant breeders. But if more than a 

 small portion of the chromosome pairs are so unbalanced, lethality results. 

 This is believed to be the cause of the reduced fertility of the autotetra- 

 ploids. Reversion to the diploid condition is probably based upon develop- 

 ment of unfertilized ovules. 



Now the study of meiosis in naturally occurring autotetraploids reveals 

 abnormalities comparable in kind and in degree with those of the experi- 

 mental autotetraploids. In the face of such a disadvantage, one may well 

 ask how it is that these have ever become established in nature, let alone 

 become more widespread than the diploid parent, as is so commonly the 

 case. However, the tetraploids are commonly more vigorous, and adapted 

 to more severe environments, and it seems probable that their selective 

 value more than compensates for their reproductive liability. Yet most of 

 the naturally occurring polyploids which have been analyzed have been 

 of the allopolyploid type, and it may well be that it is this reproductive 

 liability which has restricted the role of autopolyploidy. 



ALLOTETRAPLOIDY IN EXPERIMENT AND IN NATURE 



Allotetraploidy has also been produced experimentally. Many methods are 

 available, of which much the simplest is the treatment of the Fi hybrids 

 between two species with colchicine. Other methods, however, give more 

 insight into the means by which naturally occurring allotetraploids may 

 be formed. One method is to cross two diflFerent autotetraploids. For ex- 

 ample, if A and B each represent difiFerent haploid chromosome sets, then 

 AAAA and BBBB would he the corresponding autotetraploids. With nor- 

 mal reduction, these will form gametes with the formulae AA and BB. 

 Upon cross fertilization, the allotetraploid, AABB, will be formed. Because 

 this actually comprises two different diploid groups existing side by side 

 in the same nucleus, the terms amphidiploid and double diploid are often 

 used as synonyms for allotetraploid. But, because allotetraploids are more 

 common in nature than are autotetraploids, and because it is generally 

 more difficult to cross two tetraploids than to cross the corresponding dip- 

 loids, it seems unlikely that this method has had general importance in 

 nature. 



Two methods are based upon the occasional failure of reduction divi- 

 sions, which is especially frequent in plants with chromosome comple- 

 ments which do not synapse readily. Thus, in the cross AA X BB (using 

 the terminology introduced above), the Fi should be AB. But if there is 

 insufficient homology between the chromosomes of the A and B genomes 

 to permit synapsis, the probability of nonreduction becomes considerable. 

 Thus a significant percentage of AB gametes may be produced. In a self- 

 fertilized plant, fertilization of some AB ovules by AB pollen would be 

 quite probable, thus producing the allotetraploid, AABB, at once. Each 



298 



