FREQUENCY OF POLYPLOIDS 



231 



factors, viz., 8, 12, 16, 24, 27, 32 and 36. This means that these 

 numbers are favoured by the conditions of origin or survival of new 

 forms ; such would be the case if they owed their origin largely to 

 a process of reduplication which gave in the first instance doubling, 

 in the second instance trebling, quadrupling, and so on ; in a word, 

 to polyploidy. 



This conclusion is supported by the observation of polyploid 

 behaviour of various kinds in nearly all species with these higher 



Crocus 



Hap /old /iumber of CAromosorrjes. 



Fig. 78. — Graph showing the distribution of species with different 

 numbers of chromosomes in Crepis and Crocus where polyploidy 

 has played little part in the change in numbers. Crocus can be 

 regarded as having the same type of differentiation as Crepis, 

 but at a more advanced stage of development. (From Mather, 

 1932, cf. Fig. 16.) 



numbers, but it is not supported by observations on species with 

 8 chromosomes or, for the most part, on species with 12 (cf, Nicotiana, 

 Ch. VIII). The discrepancy is important, for it indicates that many 

 of these were polyploid in origin but have since lost all trace of 

 their origin, except in their number, through later differentiation of 

 the kind already discussed. 



Fernandes' enumeration is summarised in the list (Table 38 cf. 

 Wanscher, 1934). It is chiefly derived from the lists compiled by 

 Tischler (1936) and Gaiser (1933). It is a fairly random sample 



