CHROMOSOMES AND EVOLUTION 91 



animals one finds that in the case of the Angiosperms 

 there is an excess of even over odd haploid numbers 

 amounting to over 40 per cent, while in the case of 

 animals the numbers of even and odd haploid num- 

 bers are not significantly unequal (Tables VIII and 

 IX). This figure of 40 per cent gives a lower limit 

 to the number of polyploid species of Angiosperms. 

 At least 40 per cent of all Angiosperm species are 

 tetraploid, hexaploid, &c. There is no other reason 

 why more even than odd haploid numbers should 

 exist and a detailed investigation of the frequency of 

 particular numbers ^^^ bears out the general conclu- 

 sion. One cannot set an upper limit to the extent 

 to which polyploidy has occurred in plants, since 

 aneuploids and other ' derived polyploids ' are not 

 included in the above minimum figure. Neither is it 

 possible to estimate the extent to which all Angio- 

 spt^rms have a more or less remote polyploid ancestry. 

 It is also difficult to judge the relative importance of 

 auto- and allo-polyploidy. The distinction between 

 them is, however, only a relative one, since, on the 

 one hand, allo-polyploids exist which are the result 

 of hybridization between very closely related parent 

 species or varieties and, on the other hand, inde- 

 pendent mutation in the two diploid sets of an auto- 

 tetraploid will give rise to an organism which will, 

 like an allo-tetraploid, have a number of genetic 

 differences between its two diploid chromosome sets. 

 Thus in two ways forms will arise which are essen- 

 tially intermediate between auto- and allo-polyploids 

 in their genetic constitution. 



Whatever their origin, all even-numbered poly- 

 ploids (tetraploids, hexaploids, &c.) will, as a result of 

 independent mutation in their multiple chromosome 

 sets, tend to evolve towards a new condition of 

 genetical diploidy, in which no gene will be repre- 

 sented more than twice in the chromosome set. There 

 can be no doubt that many plant species represent 

 stages in this process. Mutations are less likely to 



