— 383 



quarry at Frederikshavn (see table III F in this paper). The 

 seeds were derived from uncontrolled fertiHzation in nature. 

 Seeds of two of these piants, both pari\ n.lab, viol, mac (the size 

 of flower was that of arvensis, the colour was tricolor's) were sown 

 as V. 246 and V. 247. The result can be seen from table III. 



When these four characters are taken into consideration, the 

 two piants have given, besides the mother type, 9 different com- 

 binations and among these one that I did not find in the original 

 population (parv, lut). Some of these must be due to segregation 

 and some of them to cross-fertilization, but in both cases it sup- 

 ports my theory. Moreover, from populations with arvensis or 

 tricolor alone I have not obtained such segregations. 



Some of the segregated types were selfed in 1921 and they 

 gave in 1922 the following results: 



V. 414: Mother plant: grand, lab, lilacina, eradiata, n.mac. 

 petalis latis (in all a very peculiar type, that I have not seen 

 in nature) gave only 9 piants, but they were all just as the 

 mother plant. 



V. 415: Mother plant: parv, n.lab, lut, n.mac, petalis angustis gave 

 8 piants all of the same type as the mother plant, but the 

 size of the petals varied somewhat, as some of the piants had 

 petals of the size of the sepals; other piants had smaller petals. 



V. 416: Mother plant: parv, n.lab, albida {with a hue of violacea)^ 

 mac segregated the following types (table IV): 



Table IV. 



All 36 individuals were }nac. 

 Some of the piants had bro ad 

 petals and some of them nar- 

 row. 



lilac = lilacina (a hue of 

 violacea). 



V. 417: Mother plant: grand, lab, viol, mac, petalis latis gave 11 

 piants all as the mother plant (tricolor-iype). 



V. 418: Mother plant: parv, n.lab, albid, mac segregated the 

 following types (table V): 



