— 407 — 



In other cases, as in the population F from Holmsland, no ex- 

 ternal signs of a preceding cross and recombination have been found, 

 but there can be no objection to that supposition that a factor 

 for optimum at a high pu.-value might cross over to the factor- 

 complex for grand, viol without being accompanied by factors for 

 morphological characters, so that the explanation, also in this 

 case, can be a preceding cross between tricolor and arvensis 

 witli the following recombination of factors. 



X. Adaptation as a Result of Mutation, Crosses and Selection. 



In a short paper Turesson (1922) has shown that some 

 so-called '"adaptative" types inhabit certain localities, so that 

 the re-appearance of a distinct locality coincides with the re- 

 appearance of the variety typical to that locality. This faet is 

 most strikingly shown in Hieraciiim umbellatiim. In the south- 

 eastern corner of Skåne (Sweden) woods, dunes and sand-fields 

 alternate within short distances. The woods are inhabited by 

 an erect broad-leaved type, the dunes by a lesser erect, more 

 narrow-leaved and strongly shoot-regenerating type; in the sand- 

 fields grows a prostrate, hirsute type. Where woodland is replaced 

 by dunes, the woodland varieties are replaced by the dune vari- 

 eties and these in their turn are replaced by the prostrate types 

 when the sand-fields begin. At the boundary hne between two of these 

 zones a narrow zone inhabited by all kinds of intermediate types 

 can be found. — The same applies to three Atriplex-ij^Q&. The east 

 coast of South-Sweden is inhabited by one type, the west coast 

 along the sound by another and north of the sound on the ex- 

 posed west coast lives a more deviating, small-leaved, prostrate 

 type. On the whole Turesson found a parallelism between the 

 extremeness of type and the extremeness of locality. He sup- 

 poses the varieties differentiated from previously existing types 

 and their combinations when crossed. 



The parallelism between these Hieracium- and Atriplex-iy^QS 

 and the extreme West-Jutland types of Viola is striking. Certain 

 populations are composed by isoreagents which are ecologically 

 "adapted" to the conditions of the habitat. In the vagabondizing 

 Viola populations on cultivated ground such an adaptation is 

 of course not so pronounced. These populations do not live so 

 long at the same place that they can become balanced. But 



