184 EILIF DAHL 



3. Amphi-Atlantic species with vicariant taxa of lower rank 



in eastern America and western Europe 10 species 



4. Species of polymorphic groups with vicariants in the 



Pacific area 22 species 



5. Species of polymorphic groups with vicariant species else- 

 where in Europe, Asia and America but not in the 



Pacific area 25 species 



6. Other Amphi-Atlantic species 33 species 



Of the groups listed above, only Group 3 can be said to support the 

 hypothesis that the plants dispersed across the Bering Sea, whereas Groups 

 1 and 2 and, to some extent, also Groups 4 and 5 contradict the hypo- 

 thesis. It will be seen that the data clearly suggest a direct migration between 

 Europe and America, 



Within Group 3, supporting the hypothesis of dispersal across the Bering 

 Sea, are seven temperate plants and only three Arctic-Sub-Arctic ones. 

 As a comparison, in Group 1, supporting the hypothesis of a direct Atlantic 

 dispersal, all except one species (the sea-shore plant Ligusticum scoticum) 

 have an Arctic-Sub-Arctic type of distribution. Thus, the hypothesis of 

 migration across the Bering Sea seems more likely concerning temperate 

 than Arctic-Sub-Arctic plants. 



The outline given above can be exemplified by numerous individual 

 examples. Here, I shall consider only one example, where more detailed and 

 advanced information is available. 



In the Papaver radicatum complex studied by Knaben (1959a, b) one 

 species, P. lapponicum, is Amphi-Atlantic, as far as is known. In Europe it 

 grows in northern Scandinavia as far east as Kola, whereas in North America 

 it has a wide area including most of Greenland, northern Labrador and the 

 Canadian Arctic Archipelago as far west as Banks Island. Knaben {he. cit.) 

 recognizes one subspecies in northern Norway and three different subspecies 

 west of the Atlantic Ocean. The morphological differences between the 

 American subspecies seem to be about as great as the differences between the 

 Scandinavian and the American subspecies. 



More quantitative information on the degree of genetic differentiation is 

 available through crossing and subsequent cytological analysis. When 

 mutations take place the pairing of corresponding chromosomes in meiosis 

 of hybrids is often impaired and the number of bivalents, thus formed, in the 

 meiosis of such hybrids seems to be a fair measure of the degree of genetic 

 differentiation. Knaben has carried out a number of intra-racial and inter- 

 racial crosses, and Table 5 is based on information contained in her work. 

 The maximum number of bivalents is 28. The number of bivalents in intra- 

 racial crosses of ssp. occidentale is fairly low, suggesting that it is genetically 

 heterogeneous and may one day be split up into more subspecies. But the 



