8 l'KNTTI KSKOI.A. M.-N. Kl. 



vvcTc remoxcd l)y hoiliiij^ at 30 iinrhr nrlucrd [jrcssurc, the apparatus 

 and the nu.tliod Ixing those proposed by V. M. Goldschmidt'. 



yXniong- the optical properties of the minerals, the indices of retraction 

 have the greatest importance. I-'or tiieir determination I useid a set of 

 li(|iiid.s prepared l)y myself after the flirections of Mr. Olaf Andersen, ac- 

 coixling to the methods used in the Geophysical Laboratory of the Carnegie 

 Institution of Washington. 



In the case of the garnet the prism method was practised. Using 

 grains of [)ut 2 or 3 mm in diameter 1 could prepare four or five prisms 

 at the same time, the whole procedure taking but half an hour. 



J. IJhIig- proposed to designate the composition of the garnets with 

 the moleculai- percentage figures of each of the simple compounds, as 

 grossulai-ite, andradite, almandite etc., each of which is expressed by the 

 s^-mbols of its most characteristic element. 



Against this mode of designation it ma\' be objected that, in the 

 garnets, the bivalent elements are variable independently of the trivalent 

 elements. It is not stricdy correct to designate a garnet, in which enter 

 MnO, CaO, Al^Og and Fe.^O.:,, as a mixture of spessartite, grossularite and 

 andradite, as in such a «mixture the Fe^O-j can not be said to be combined, 

 with CaO more than with MnO. Nor is it correct to calculate the percentages 

 of the bivalent and the trivalent elements together to the sum of 100. 

 1 have therefore calculated the atomic percentages of the elements separately 

 in each group of the isomorphic substituents. The designations thus re- 

 sulting, e.g. (P'e.j^Mn^Mggj),, (Aly-Fe.jlo, are clear without any explanation, as 

 they are simply molecular formulas, the group (SiO^lg^o omitted, being 

 common to all. Where no stress is laid on the trivalent elements which 

 all except AloO.^, in the pyrope-almandite series, are of litde importance, 

 the garnets will simply be characterized by the biwilent elements, and a 

 designation like FcgoMn^MggCag for a mixture of 82 mol. almandite, 7 mol. 

 spessartite, 8 mol. pyrope and 3 mol. grossularite is as clear as it is simple. 

 Thus we have a system of designations adaptable to anv degree of accuracy 

 and completeness needed. 



hi the determination of pyroxenes and olivines by means of refractive 

 indices and other optical properties, F. Becke's compilation in C. Doelter's 

 „Handbuch der Mineral chemie" ^, were adopted while for the monoclinic 

 amphiboles Ford's diagrams were used'*. 



' Videnskapssclskapets Skrifter. I. Mat.-naturv. Klasse, ig 16. No. 2, p. 4. 



2 J. Uhlig. Beitrag zur Kenntnis der Granaten in vulkanischen Gesteinen und Auswürf- 

 lingen des Niederrheins. Verb. d. Nat. Ver. d. preuss. Rheinlande u. Westfalens. Bonn 

 1910, 67, 307^ 403. 



3 Band II, I, p. I. 



4 W. E. Ford, A Contribution to the Optical Study of the Amphiboles. Am. Journ. Sei. 

 XXXVII, 1914, p. 185. 



