282 KARL 15. KRISTOFFERSON 



is dominant, but the colour factors developed the colour in the three 

 lower petals to a limited degree only. The growth form in a cross 

 between forms with erect main axis and branches and a prostrate 

 type was also in this case about intermediate etc. 



The F 2-gcncration of the cross 45 y(. 25. 



I have only grown the /«^.-generation of one of these crosses, 

 45 X 25. As I had only about 20 F^-plants, I have not made any 

 calculations of the ratios. 



All of the /'".-plants were of a different appearance, and they 

 showed segregation in all characters, so far as I could see. Probably 

 a segregation took also place as to the duration of life. In any case 

 about Vi — V:! of the Fa-population died away during the winter. Thus, 

 the character of perennialty in Viola is a dominant one, it shows 

 probably monohybrid segregation in F., as is the case in other plants 

 investigated w^th regard to this character. 



The reason of mentioning this Fo-generation, in spite of its small 

 number of individuals, lies in the fact that it exemplifies a species- 

 hybrid showing segregation in Fo. 



V. Miinbyana seems to be nearly related to the /r/co/or-group. The 

 fertility of the pollen of the primary hybrid is very good. By germi- 

 nation in glucose-agar I found a germination-power of 95 — 100 % . 

 The ovules seemed also to be quite fertile. In any case, the fructifica- 

 tion was very good. Becker (1910) has put V. Miinbyana in the cal- 

 carata group. It is possible, however, that other 'species of this group 

 will show the same genetical behaviour to V. tricolor and V. arvensis 

 as does V. Miinbyana. 



DISCUSSION OF THE RESULTS. 



The systematics of the great mass of phenotypes known as V. 

 tricolor L. and V. arvensis Murr, is rather difficult. Linn.i:us grouped 

 them together in one single species, V. tricolor; the small and yellow 

 flowering type was later characterized as a variety of this species (de 

 Candolle 1824). Still later it was broken out and put up as a separate 

 species, V. arvensis Murr. As the variation in these species was very 

 large an effort was made to divide them in subspecies and lower syste- 

 matical units. WiTTROCK (1897) puts up about 40 well distinguished 

 forms. When cultivated they were all found to breed true. They 

 fill thus the systematical demands on subspecies. Becker (1910) deals 

 Avith about 20 subspecies. 



