92 THE CHROMOSOMES 



prove lethal in a tetraploid than in a diploid, since 

 two of the four genes (and similarly four of the six 

 in a hexaploid) are more or less superfluous and can 

 consequently mutate without vitally affecting some 

 important process as usually happens when the genes 

 of a diploid mutate. 



It seems fairly clear that a number of new Angio- 

 sperm genera have arisen as a result of hybridization 

 between fairly widely separated species followed by 

 allo-polyploidy, just as Raphanobrassica (allo-tetra- 

 ploid Raddish- Cabbage hybrid) and Aegilotriticum 

 (allo-polyploid Wheat -Rye hybrid) have arisen under 

 experiment. It is at any rate certain that well- 

 marked species like Aesculus camea, Galeopsis Tet- 

 rahit, Spartina Townshendii, and the American form 

 of Phleum pratense, have arisen in the first place as 

 allo-polyploids.^^' i^^' '2* ^^ On the other hand where 

 several chromosome numbers which are multiples of 

 the lowest one exist within the same species of plant 

 (as in Biscutella laevigata (see Chap. Ill) and Nastur- 

 tium officinale, ^^^ they are clearly due to auto -poly- 

 ploidy. Such auto -polyploid varieties are usually 

 taxonomically distinguishable from the diploid form 

 and from one another by slight morphological dif- 

 ferences and may form the starting-point for new 

 species if, as is usually the case, their hybrids with 

 the diploid form or with one another are more or 

 less sterile. It is probably in this way that the 

 species of the genus Rumex (section Lapathum) have 

 originated (Table V). 



Although polyploidy has undoubtedly taken place 

 in nearly all the larger plant genera, and been re- 

 sponsible for the origin of a large number of species, 

 there are some genera such as Carex where the known 

 somatic chromosome numbers (18, 30, 32, 38, 52, 54, 

 56, 62, 64, 68, 70, 72, 74, 76, 80, 82) do not suggest 

 that it has played any important part in the forma- 

 tion of new species. The reason for this is not clear ; 

 perhaps the actual formation of pol3rploid cells is 



