258 THORODDSEN 



mixed with a large quantity of tuff-dust (mohella). The surface is 

 divided into squares or more or less regularly formed polygons, by 

 bands of small stones or gravel, while the clay of the interior of the 

 squares or polygons is destitute of stones. The surface resembles a 

 net the meshes of which are formed by irregular bands of gravel. 

 Usually it is only the gravel which supports plant-growth; there 

 the plants can find shelter between the stones, while the middle of 

 the clay-cakes is too wet for plant-growth. But when the "rude- 

 mark" becomes drier, vegetation may gradually extend over the 



wr, ?3' 



Fig. 18. Portion of a '"rudemark" in the neighbourhood of Reykjavik, 

 showing the position of the stone and gravel bands. (Drawn by Th. Thoroddsen.) 



cakes of clay, first forming a scattered growth upon them and ulti- 

 mately soil and a plant-carpet, especially when tuff-dust and drifted 

 soil have settled on the surface. The polygonal cakes vary greatly 

 in form and size, but generally the} 7 have a diameter of only 1 /2- 

 1 metre. The knolls (Jmfa, pi. fmfur; see Fig. 17), which play such 

 an important part with reference to Icelandic vegetation and agri- 

 culture, stand in close genetic relation to the "rudemarks" and we 

 will therefore try to give an account of the way in which they both 

 have probably originated, but questions bearing on this point require 

 to be elucidated by the experimental investigations of persons living 

 on the spot. My investigations in Iceland have confirmed me in 

 my opinion that the Icelandic pufur as already mentioned 

 stand in a close genetic relation to the "rudemarks" and that sub- 

 surface ice is an essential condition for the formation of both; 



