40 CLASSIFICATION BASED ON MINEEAL COMPOSITION. 



Even supposing the species were well established, have we any methods 

 whereby these species or divisions can be positively discriminated ? 



In 1876, Descloizeaux gave a method, whereby he thought the diffei'- 

 ent plagioclastic feldspars could be distinguished from one another. This 

 method required a thin transparent section, either cleaved or ground par- 

 allel to the plane of easiest cleavage (0, OP, 001, p) — the trielinic feld- 

 spars being as a rule twinned, so as to show color bands parallel with the 

 plane of the next easier (or less perfect) cleavage. The sections are 

 placed on the stage of the polarizing microscope, with the color bauds 

 jjarallel to any diagonal of the crossed nicols (plane of vibration or 

 plane of polarization) ; the section is then revolved until one set of color 

 bands becomes dark, or the light is extinguished in them. The angle 

 between this point and the former position of the section is taken. The 

 section is then revolved in the opposite direction until extinction of light 

 takes place in the alternate set of color bands. The angle between 

 this point and the first or original position of the section is taken. If 

 the section is properly cleaved, or ground, the two angles observed are 

 equal. By experiment on feldspars of known composition, the angles 

 between the nicol diagonal and the extinguished color bands, or the angle 

 (double the others) between the two positions in which the alternate bands 

 are rendered dark, have been determined ; and it is assumed that all feld- 

 spars having the same angle of extinction as any one of these previously 

 determined angles belongs to the same species of feldspar. Descloizeaux 

 held, that this supposed fixity of optical characters was opposed to Tscher- 

 mak's theory.* 



Prof R. Pumpelly endeavored to make Descloizeaux's method practi- 

 cally applicable to thin sections of rocks, which he did in the following 

 manner : If instead of cutting sections parallel to the principal cleavage 

 they should be cut at any angle with that cleavage but in the zone : 

 el(p:h'; 001: 100; OP: go P co) we should have every variation of 

 angle, from 0' up to the maximum for that feldspar. In a thick rock sec- 

 tion in which the feldspars are cut at random, it is necessary first to ascer- 

 tain whether any feldspar section has been cut in the zone 0: it. This is 

 done by ascertaining on trial if the extinction in the alternate color bands 

 takes place at equal angles on oi)posite sides of the nicol diagonal. If it 

 does, the section of feldspar was cut as required. A numl^er of such sections 



* Comptes Rcmlus, 1S76, Ixxxii. 8S5-891 ; Ann. Cliiiiiie Pli^siquc, 1S76 (5), Ix. 4.33-499. 



