286 KARL li. KRISTOFFERSON 



ween these species (3 X 10) a plant was found to be phenotypically 

 identical with my line 2. In a cross between two arvensis forms 

 (2 X 4) some Fo-plants with the same appearance as my pure lines 

 12 and 35 were found. However, I could only state the phe- 

 uotypical identity between these products of segregations and the ty- 

 pes living and collected in nature; their true-breeding was not ascer- 

 tained. On account of the above studies in Viola I dare say that pro- 

 bably the majority of Witti\ock's forms would be found to be present 

 in the F:.-generation of a cross between certain extremes of his forms, 

 for instance V. tricolor genuina versicolor X V. orvensis patens. 



I have never observed any mutations in my cultures, which is 

 quite explainable on account of the relatively small number of plants. 

 However, in spite of the occurrence in F^ of forms resembling forms 

 already existing in nature the assumption of the origin of the forms 

 of V. tricolor by mutations is not therefore to be rejected as quite 

 absurd. Even negative mutations might give rise to such forms; 

 positive mutations would give rifie to any form. For example, a violet 

 and small-flowering type with small labellum would result in a form 

 similar to V. arvensis ciirtisepala if the colour factors and the inhibi- 

 ting factor for the length of petals and labellum were lost etc. How- 

 ever, the »true» negative mutations hitherto found and investigated 

 make this assumption less probable. They are all less vital than their 

 original form, and low vitality has not been found in any form of 

 V. tricolor. 



On the other hand, the fact that types similar to types already 

 existing in nature are found in F. of crosses between different forms is, 

 according to my opinion, not a proof of the theory of evolution by 

 means of hybridization. As to the origin of the mass of phenotypes 

 known as V. tricolor nothing can l)e said on the basis of my ex- 

 periments. They only .indicate that the variation in this species is 

 due to re-combination of Mendelian factors. 



SUMMARY. 



1. In one cross between an autogamous form of Viola arvensis 

 (line 2) and a herkogamous of \'. tricolor (line 10) Fi became herkoga- 

 mous. /% showed segregation in autogamous and herkogamous plants 

 in the ratio 1 : 2,35. This ratio lies within the statistically allowed 

 ranges for a monohybrid segregation. However, the segregation may 



