— 377 — 



some time to complete these investigations, but they show 

 interesting facts. They confirm the results attained by the 

 two other methods, that the variations of the characters 

 are induced by genotypical differences. 



III. Are all Combinations of the Characters possible? 



In the paper from 1921 I reported about the combinations 

 of the 6 first mentioned characters. In the meantime, I have 

 examined a great number of individuals from different habitats 

 in order to decide if all combinations of the remaining characters 

 are likely to occur, and to get information as to which combi- 

 nations are the most frequent. The characters were examined in 

 sets of 4, 5 or 6 at a time. Within the sets practically all combi- 

 nations can be found. Characters from different sets have been 

 examined together too and, as they also combine in different ways, 

 I suppose it justifiable to infer that, on the whole, all combina- 

 tions can be realized. Of course I have not seen all possible com- 

 binations of all these characters. It would take more than a lifetime 

 to examine only one individual of each combination, and I do 

 not believe that they are realized all simultaneously. 



In table I next page a summary is given of the combinations 

 of four characters that are easy to determine even on exsiccata. 

 The results from 23 habitats are given in the table. In each place 

 50 or 100 individuals were taken and determined. The numbers 

 in the table are egalized so, that they in all cases indicate the 

 number of individuals of the combination (isoreagent) in question 

 per 100 individuals from that habitat. The four characters are: 

 size of petals, colour of petals, stipules and end-lobe of stipules. 

 The order of the characters is not indifferent; they ought to be 

 taken in the afore-mentioned order, as the bulk of the individuals 

 by this method is placed in two groups, one at the top and one 

 at the bottom, corresponding to the two species in question: 

 V. tricolor (at the bottom) and V. arvensis (at the top). In the 

 middle the fairly rare irrelevant types are to be found. As to the 

 flower-colours I have only taken two alternatives here, as there 

 is a great difference between the arvensis-colour albida on the one 

 hand, and all the tricolor- colours violacea, lutea, rosea and alba 

 on the other. 



Each of the 23 habitats has its column in the table indicated 



