399 



a very wide spread opinion, that tricolor has pinnate stipules 

 (Becker 1905, pag. 48, about tricolor: "Stipulae pinnatifidae"), 

 but I presiime it is wrong. Palmate stipules are more frequent 

 than pinnate. Many of the tricolor (grandiflora) individuals descend 

 from crosses between tricolor and arvensis and I suppose that all 

 the tricolor individuals with pinnate stipules are stabilized segre- 

 gation products from these crosses. In other words: that the 

 original typical tricolors had palmate and the original typical 

 arvensis had pinnate stipules. If \ve had a territory with tricolor 

 alone, the statistic proportion from this territory might decide 

 the question. Such a territory we do not have, as arvensis occupies 

 a territory being co-extensive with and extending beyond trico- 

 lor's. When the facts are so, we must try to compare a territory 

 where tricolor is far more common than arvensis with one where 

 arvensis is just as frequent as tricolor. In Jutland we have such 

 two territories. V. tricolor (grand, viol) is one of the most predom- 

 inant piants in West- Jutland. In early summer the fields are 

 blue with these flowers. In many of these fields I never saw an 

 arvensis. In East- Jutland fields with arvensis alone alternate 

 wdth fields with tricolor alone and with fields with both arvensis 

 and tricolor. The following table gives the distribution of palmate 

 and pinnate individuals among the tricolors in 7 habitats from 

 West- Jutland and 7 habitats from East- Jutland: 



The table shows that in West- Jutland a far greater proportion 

 of tricolors have palmate stipules than in East-Jutland and there- 

 fore I suppose palmate stipules to be typical of the pure tricolor. 



As to the dark spot in front of the style it is rare in parvi- 

 flora individuals and very frequent in grandiflora individuals and 

 I suppose it to be a ^rico/or-character. 



