THE CONTINENTAL ELORA OF SOUTH SWEDEN 



333 



tation is naturally more favourable than 

 the western one, dismembered as the 

 latter is by the sea. 



If the said isolated occurrences of steppe 

 species are relics of a more continuous 

 distribution, they ought to be relics of 

 a more power fully developed western 

 branch of the Scandinavian distribution 

 area. 



As the North German plain should 

 have been able, with a more continental 

 climate, to offer favourable conditions to 

 a steppe flora, its continuation towards 

 the north-west, which the Danish Islands 

 and the peninsula of Jutland maybe said 

 to form, should at the same time have 

 been much more suitable for steppe 

 species than at present. Apart from 

 the climatic conditions, this should have 

 been the case during an earlier post- 

 glacial period, for the soil must have 

 been better in these parts, before it was 

 leached by the precipitation. 



Hence it seems to me to be fairly 

 probable that these isolated occurrences of steppe species should form remains 

 of a more continuous western branch of distribution that the species once had 

 in Western Scandinavia. 



In connection with this there might be brought forward the possibility of cer- 

 tain peculiar occurrences of species in Vastergotland having arisen as part of a 

 similar westerly branch of distribution: Prunella grandiflora (Plate 6), (which has 

 a locality in north of Jutland and one, but somewhat uncertain, on Zealand), Allium 

 montanum (Plate 6), Stipa pennata, Dracocephalum Ruyschiana (Plate 4), and Po- 

 tentilla rupestris (p. 326). 



Fig. 19. The distribution of Pulsatilla pratensis 



(L.) Mill, in the Scandinavian North. 



As to Denmark the map refers to P. » pratensis 



(L.) Mill.)) in Hayek's sence -f P. » nigricans 



St6rck». Cf. p. 402 in Appendi.\ I and Ostenfeld 



191 1, pp. 252 fif. 



O: uncertain occurrences. 



The two branches in the South=Scandinavian distribution. 



The division mentioned of the Scandinavian distribution-area into two branches 

 is also of interest from another point of view. It has been pointed out that the 

 reason of the division is that the South Swedish highland isolates from each other two 

 plains with localities suitable to the species in question. In some cases, however, 



23 Gcogra/iska Annaler ig:2. 



