148 PHYSICAL BASIS OF HEREDITY 



and form two different sets.^ Theoretically, the number 

 of different genes in a species might in this way be in- 

 creased. If changes in the same gene in the same direction 

 sometimes occur, as the evidence indicates that they do, 

 then identical new mutant genes, derived from the same 

 kind of original ones, might later arise in different pairs. 



There is, however, another way in which the number 

 of chromosomes may be doubled without doubling the 

 number of genes. If the chromosomes break in two, 

 double the number will be produced. It is not easy to 

 explain how this could occur in all of the chromosomes at 

 the same time if the process is supposed to be accidental. 

 If it be supposed that the break first occurred accidentally 

 in one member of the pair, it is not clear why such a 

 broken chromosome could establish itself on the theory 

 of chance, for the intermediate condition of one broken 

 and one intact chromosome would seem of no apparent 

 advantage. The same reasoning applies to the converse 

 process, viz., the coming together of chromosomes end 

 to end which would reduce the number by half. Such a 

 process would not increase the number of genes in 

 the total complex. Until we knoAv more about the 

 physical or chemical forces that hold the genes in chains, 

 and more about the way new genes arise, it is not worth 

 while to speculate about the causes or probabilities 

 of such occurrences. 



What has just been said in regard to doubling and 

 halving of the whole set of chromosomes applies also to 

 doubling in one pair of chromosomes. If doubling 

 occurred in one pair of a ten-chromosome type, a twelve- 

 chromosome type would result ; if in two pairs, a f ourteen- 

 chromosome type, etc. Unless tetraploidy is the simpler 

 procedure we should a priori suppose that increasing (or 

 decreasing) in pairs would, on the theory of chance alone, 



^ The question as to whether the four chromosomes involved would or 

 would not mate at random introduces a difficulty (as shoA\Ti in the 

 primula case). 



