1 92 1. No. 8. 



ox THE ECLOGITES OF NORWAY. 



shows the same relation in pure mixtures of pyrope and ahiiandite, ac- 

 cording to Ford. 



The proportion of Fe.,0.5 to Al.,0.j was not taken into account, firstl}' 

 because the quantity of Fc^O.^, in the series under consideration, is alwa3's 

 small and little \ariable, and secondly because, an exact determination of 

 FeO in the garnets being diflicult though not impossible, most analyses 

 of garnets are more or less unreliable in this respect. 



As could be expected, the lime-bearing pyrope-almandite mixtures have 

 higher refracti\e indices than those in which no MgO has been replaced 

 by CaO. The difference is largest at about 60 pt FeO, owing to the fact 

 that these garnets generally are richest in lime. The relation is not quite 

 regular, however. 



At first sight the réfringence appears as a linear function of the ratio 

 FeO : (MgO + CaO). Really it is not so, the apparent approximation being- 

 due to the fact that the amount of lime in mixtures varies with the varia- 

 tion o( the ratio MgO : FeO. The latter, in turn, depends upon the bulk 

 composition of the rocks and the attendant pl\vsical conditions, these inter- 

 relations being indeed the main subject of study in this paper. Now, 

 both these factors, the physical conditions and the chemical composition, 

 account for the actual characters of rocks. Therefore we will hnd, by 

 grouping rocks in any system developed on a natural basis, that a definite 

 group will contain garnets varying in composition between definite limits. 

 This will be shown bv the following table where the molecular percentages 

 of the lime compound in the garnets in different rocks are given. The 

 anal3'tical material is mainly the same as used further on in this work. 

 The garnets in the Norwegian labradorite-rocks ha\-e been treated as 

 eclogite-garnets and not as gabbro-garnets, on grounds that will be obvious 

 from the description. 



