﻿500 DE. WHITMAN CROSS ON THE NATURAL [Aug. I9IO, 



I do not mean to claim that our ideas in this matter were as 

 clear as they are to-day. We may have builded better than we 

 knew. But there was every reason why we should not go into 

 theoretical discussion of the molecular constitution of the magma 

 iu connexion with the norm, for we laid stress on the belief that 

 systematic classification should not be based oh theory. The 

 constitution of the magma is still imperfectly known ; it was much 

 less so in 1902. Some degree of electrolytic dissociation in magmas 

 is assumed by Iddings, Togt, and Harkcr. 



Whatever our views on the magmatic solution may have been, 

 the important point, so far as the character of the norm is con- 

 cerned, is that the normative molecules were chosen after a very 

 careful consideration of the laws of chemical affinity, which have 

 been evident in synthetic work and are expressed in crystallization, 

 and would control the combinations of oxides and other primary 

 constituents of rock-magmas, /whenever combination was 

 possible. These considerations are stated very concisely, but 

 definitely, in explaining the norm and the order of procedure in 

 its calculation. For all the purposes of this discussion it is imma- 

 terial whether these laws become effective at crystallization or 

 before it. The significant fact is that the laws of 

 chemical affinity, recognized in formulating the norm, 

 appear to have been in the main correctly understood; 

 th at they have operated in the magmatic stage, and still 

 hold good to a large degree under the average conditions 

 of crystallization. 



The remark of Mr. Harker that the ' rangs and subrangs are 

 made to depend directly on chemical characters, viz., the molecular 

 ratios of certain oxides,' while other divisions rest on proportions of 

 mineral molecules, shoAvs that he did not understand why this was 

 done and deserves explanation, destroying the critical force of the 

 comment. The molecular ratios of K.,0, Xa.,0, and CaO, on which 

 rangs and subrangs depend, in Classes I, II, and III, are not the 

 ratios of the total amount of these oxides in the rock as shown by 

 analysis, but of that part of these oxides assigned to molecules of 

 orthoclase, leucite, albite, nepheline, sodalite, noselite, and anorthite. 

 In other words, the ratios are the ratios of these mineral molecules, 

 in three groups. It is a concise and convenient way of indicating 

 the character of the silicate molecules the proportions of which, 

 in the groups called felspar and lenad (felspathoid), enter largely 

 into the division of Orders. The same explanation applies to the 

 ratios of oxides used in making rangs and subrangs in Classes IV 

 and V. They stand for certain mineral molecules. 



The view that magmatic molecules correspond to simple mineral 

 molecules has been expressed by many petrographers. While 

 testing the ' Kernhypothese ' of Eosenbusch in its application to 

 laurdalite and its descendant rocks, Erogger came to the conclusion 

 that magmas consist of simple compounds ordinarily found in rock- 

 minerals, and that their number exceeded not very greatly the 



