— 375 — 



As will be seen from the foregoing, the characters in several 

 cases are rather difficult to distinguish from each other, and the 

 list here given is not exhaustive. As for stipules for instance 

 there are more types than the four mentioned, but on the other 

 hand each V. tricolor or arvensis can as regards stipules be placed 

 under one of the 4 types. The limitation of all the quantative, 

 transgrediating characters against each other is also difficult, but 

 I suppose that a limit, even an arbitrary one, is always to prefer 

 to a description, in common terms, of the habitual appearance, 

 that never can be so concise and so easy to compare as the state- 

 ment of a numeric relation. 



II. Are the Variations induced by Genotypical Differences? 



This problem I have tried to solve in different ways. 



1. By observation in nature. In many habitats I found 

 the different characters realized in plants growing together 

 under the same conditions. Plants with violet petals and 

 yellow petals, with long spur and short spur, with straight 

 spur and incurved spur, with erect stems and prostrate stems 

 grow intermixed in the same habitat, and when they are 

 different, it must be due to the genotypical differences. The 

 tables I and II (p. 378 and 381) state the extent in which the 

 different characters found are realized in the same habitat. 



2. By cultivation of the types under the same con- 

 ditions. In several cases I found marked differences between 

 the composition of the populations from habitats the con- 

 ditions of which differed considerably. On the narrow isth- 

 mus Holmsland Klit between Ringkøbing Fjord and the 

 North Sea I found two such populations. A road runs length- 

 ways through the isthmus and separates the dunes on the 

 western seaside from the clayey and more fertile fields on 

 the eastern fjord-side. There is but a few steps between 

 these two localities. Both in the dunes and in the fields 

 there were populations of Viola tricolor but of two very 

 different types what can be seen in fig. 5. In the dunes all 

 the plants were parvifoliata (leaves of the thick, somewhat 

 fleshy type common in the dunes of West-Jutland), the stems 

 and the leaves were atro purp urea, most of them were pros/ rale 

 and caespitose perennials (fig. 5 b, c). Otherwise in the fields 



