CHROMOSOME NUMBERS AND THEIR ALTERATION 215 



plants is one that can hardly be escaped. To what extent the poly- 

 ploidy in nature has arisen by doubling in relatively pure strains (auto- 

 polyploidy), or b}^ the doubling in hybrids (allopolyploidy) to be discussed 

 in the next chapter, is a difficult problem to solve. At present the relative 

 importance of the two processes is a subject of debate, and it is well 

 realized that much observational and experimental work must be done 

 before the tangled situation in nature can be very thoroughly understood. 

 In the meantime it is to be borne in mind that the two forms of natural 

 polyploidy differ in degree rather than in kind: to form a hybrid at all, 

 two species must have a considerable degree of similarity in their genomes. 



The significance of autopolyploidy in the origin of polyploid species 

 in nature is strongly suggested by the fact that experimentally produced 

 polyploids and those in the field show many resemblances. They tend to 

 differ morphologically from related diploids in the same way; they often 

 show similar phj^siological peculiarities and ecological adaptability; both 

 show the same type of meiotic irregularity, including multivalent associa- 

 tions and sterility among the spores or gametophytes. The natural 

 polyploids are irregular in a less degree, presumably because of the past 

 action of natural selection. 



Whether induced autopolyploids gradually improve in fertility and 

 regularity of meiotic behavior after many generations is a significant 

 question now being studied. Should they do so and become distinct new 

 types with regular bivalent formation, it is possible that they might have 

 greater genetic stability than their ancestral diploid types, for the reason 

 that a new recessive mutation would remain hidden longer and affect the 

 phenotype only in those rare individuals having all the increased number 

 of controlling factors in the recessive state. 



The role of allopolyploidy is suggested by the absence of multivalents 

 in many natural polyploids, for when the high chromosome number 

 results from a combination of genomes each of which is present onl}^ 

 twice, bivalents only are expected at meiosis. Furthermore, as will be 

 pointed out in the next chapter, chromosomal doubling in diploid hybrids 

 is often, though not always, accompanied by increased fertility, and this, 

 together with the increased vigor and new character combinations due to 

 the hybridity, should contribute much to the success of such newly formed 

 types in nature. 



Answ^ers to many questions like those suggested above will be required 

 before the evolutionary role of heteroploidy can be described \\dth any 

 degree of precision. A role it surely has, but just how it should ho ranked 

 among other factors of speciation in different families of organisms we do 

 not know. 



