156 SCIENCE PROGRESS. 



numerous localities in the island, and are considered by 

 Backstrom 5 to be products of the differentiation of an 

 original basic magma. It is worthy of note that they are 

 soda-rhyolites, potash being always subordinate to soda. 

 The intermediate rocks, which occur sparingly, are not 

 typical andesites, but have a more peculiar composition. 

 The great predominance of basic lavas is observed not 

 only in Iceland, but in Northern Ireland, Western Scotland, 

 the Fseroer Isles, and probably a considerable part of the 

 Arctic regions, all of which have belonged in Tertiary times 

 to one petrographical province. The existence of such 

 provinces, each having its varied group of igneous rocks 

 agreeing in certain characteristics and differing from the 

 types developed in adjacent provinces, naturally leads to 

 the idea of a differentiation in space as well as in time, and 

 that on a regional scale. Judd long ago pointed out the 

 strong contrast between the Tertiary volcanic rocks of 

 Bohemia and of Hungary, divided by the Carpathian Chain, 

 and the idea can be extended to other parts of Europe. 

 Iddings, in the memoir already cited, remarks that America 

 is divided by the long chain of the Rocky Mountains and 

 the Andes into two vast regions, the igneous rocks of which 

 offer a like contrast. Rocks relatively rich in alkalies occur 

 along the eastern side of the Rocky Mountains, in the 

 Central and Eastern United States and in Canada, in Brazil 

 and along the eastern part of the Andes, in Argentina and 

 Paraguay ; while in the western region quite different types 

 are found, the alkali-rocks {e.g., those carrying leucite or 

 nepheline) being, with one exception, unknown there. The 

 Yellowstone Park district lies near the dividing line of these 

 two great regions, and the rocks of the three centres 

 described there by Iddings show considerable differences 

 according to their situations. 



It is not a little interesting to observe how regions 

 petrographically different are divided by mountain ranges ; 

 the problem of the differentiation of igneous rock-magmas 

 on a large scale seems thus to connect itself with that of 

 crust-movements, but this is an almost untouched field of 

 speculation. Perhaps materials are not yet available for 



