SECONDARY POLYPLOIDY ' 243 



They therefore show how secondary polyploidy may be supposed 

 to have arisen in Dahlia and in the Pomoideae, viz., in the derivatives 

 of crosses between original polyploid forms with different numbers 

 of chromosomes. Thus, selfing a pentaploid, derived from crossing 

 a tetraploid and hexaploid, would give, in the Rosa type with 

 seven chromosomes, trisomic and tetrasomic seedlings with numbers 

 between 28 and 42. The latter, if they bred true, would be secondary 

 polyploids. 



Further investigation will probably show that secondary 

 polyploidy, and other variations due to change in balance (such as 

 reduplication of segments), is an important source of species- 

 formation in plants. Secondary pairing indicates such an explana- 

 tion of the origin of the chromosome complements in many genera 

 with basic numbers of 13, 17 and 19 (e.g., Empetrum, Gossypium, 

 Salix ; cf. Lawrence, 1931 ; Skovsted, 1933). The observation of 

 secondary pairing consistent with a particular assumption of 

 primary or secondary polyploidy is not alone a sufficient basis of 

 inference, for a secondary association can also presumably arise in 

 a diploid from small reduplicated translocations which, as we shall 

 see later, are very general in diploids. On the other hand, the 

 evidence of secondary pairing has shown in certain instances that 

 secondary polyploidy involving change of balance is an effective 

 means of variation. This being established as an evolutionary 

 principle the precise mechanism in particular and doubtful cases 

 is not of immediate importance. 



