262 



rnt: American naturalist [Vol. lv 



mutant "Xew Specio" corresponds to (EnotJiera gigas 

 and is l}rouglit al)ont l)y a doubling of the chromosome 

 number. Tlie color ratios in our tetraploicl daturas indi- 

 cate that (Enoihera nanella is a Mendelian segregate and 

 suggest that other of the CEnothera mutants which give 

 monohybrid ratios in crosses may be of the same nature. 

 Our evidence in regard to 0. nmiella comes from the oc- 

 currence of this mutant in cultures of 0. gigas, DeVries 

 (14) reports that certain races of gigas when selfed regu- 

 larly produce from 1 to 2 per cent, nanella mutants, while 

 certain pedigrees give monohybrid ratios which, on ac- 

 count of the lesser vitality of the recessive nanellas, show 

 a higher proportion of the dominant gigas forms. From 

 the pedigrees approaching a 3 : 1 ratio he obtained plants 

 which bred true, except again for the 1 to 2 per cent, of 

 nanella mutants in their offspring. A glance at the chart 

 (Table III) will show that, if our theory of tetraploidy 

 be correct, the 1 to 2 per cent, of mutant nanellas which 

 deVries obtained by setting plants from 3 : 1 pedigrees 

 must have been the recessives in a 35 : 1 ratio since no 

 dominant plants in a 3 : 1 pedigree of a tetraploid race 

 could be expected to breed true. The dominant pheno- 

 types must either throw 3:1 ratios again or 35:1 ratios. 

 The deviations of the nanella mutants in this case from a 

 35 : 1 ratio is accounted for by a similar proportionate 

 deviation in the 3:1 ratio. The work of Muller (8) on 

 balanced lethals strongly suggests that such of the CB]no- 

 thera mutants as are not caused by chromosomal duplica- 

 tion are due to cross-overs from a balanced lethal con- 

 dition. 



What then is a mutation! I do not feel we need to be 

 bound by its a])plication to the evening primrose for 

 rca.-oii^ (>r pi iority, since Waagen (15) had previously 

 used ihc \vv\i\ ill paleontology in an entirely different 



