318 RIKARD STERNER 



have been able to hold on to one or other locality in these parts even without 

 human intervention. In any explanation of the rich occurrence of steppe-species 

 in the districts mentioned, both these circumstances should be taken into con- 

 sideration.) 



The Distribution Boundaries of the species in South Sweden. 



In most cases the steppe species of the South Swedish flora reach definitive 

 distribution limits within the borders of South Sweden. In the sequel I shall 

 try to explain the latter with the help of what has been said above about the 

 mode of occurrence of species and the geographical conditions of South Sweden. 



For all but one of the species the nortJiern limit of their European distribu- 

 tion passes through South Sweden or in some cases through the southernmost 

 section of Norrland. 



The species that has a considerable distribution in Norrland too is Viola rii- 

 pestris (Fig. 13). 



This species has a number of occurrences in northernmost Scandinavia, in 

 Sweden at Torne Trask, which are, in fact, the westernmost occurrences of the 

 great distribution of this species in northern Finland and Russia. In Medelpad 

 and Jamtland runs the limit of the South Scandinavian distribution. Viola ru- 

 pestris occurs in southern Norrland, above all on sandy slopes in the river 

 valleys, but it also belongs to the flora of southerly hill slopes (Swed. y>sydhctory> 

 Andersson och Birger 191 2, p. 93). 



Some other steppe species stretch into southern Norrland with a small number 

 of occurrences. Foremost amongst them should be mentioned Raimnc7ilus 

 polyanthemos (Fig. 14) and Crepis praemorsa (p. 315). 



The wide distribution towards the north of these two species may be con- 

 nected with their rich mountainous distribution in Central Europe and their less 

 decided character of steppe species (cf. above pp. 281 and 315). 



The species have numerous occurrences in southern Dalarne. They reach 

 their northern limit with a cluster of occurrences in the Silurian district of 

 Lake Siljan. 



From the coast district of southern Norrland there are only very uncertain 

 locality-reports concerning Crepis praemorsa. Ranunculus polyanthemos, on the 

 other hand, has numerous quite positive occurrences there. Its northernmost 

 occurrence is in the parish of Nordingra in Angermanland (H. W. Arnell). (Its 

 locality in Jamtland, reported by Olsson 1894, must be judged very uncertain.) 



The distribution of Ranunculus polyanthemos in the districts named might be 

 put in connection with local occurrences of calcareous soil, Archaean lime (cf. 

 Grevillius 1894), shell-gravel banks (cf. Halden 1917 and 1920) or marl occur- 



