L. V. PIRSSON 



Since the analyses were made on very small quantities the 

 ordinary anal3'tical errors become considerable and it is, there- 

 fore, probable that the average of the two would be more correct 

 than either alone. The water in No. II is evidently too high 

 and may be excluded. The average is shown in the third 

 column and its molecular ratios in the fourth. The lime, iron 

 and magnesia are of course due to admixed microlites of pyr- 

 oxene, and deducting sufficient silica to turn them into the gen- 

 eral formula RSiOg the remaining ratios have the following rela- 

 tions : 



SiO, - - - - - -799 ^4-1=4 



AI2O3 ----- .222 = 1.1 = 1 



NaoO + KaO - - - .193 = 1.0=1 



H2O ----- .416 = 2.1 = 2 



which gives the analcite formula Na Al (8103)31120 with a fair 

 degree of exactness, some of the soda being replaced by a little 

 potash as in the Brazilian rocks. 



In his article Lindgren' speaks of the difficulty of distinguish- 

 ing the isotropic analcites from glass and in a review of the 

 paper Iddings^ emphasizes this point and suggests that isotropic 

 minerals may have been determined as glass in some cases. It is 

 »now becoming evident how often this has, in all probability, 

 been done. 



It is now clear from what has been stated above that the 



'Op. cit., p. 55. 



^Journal of Geology, Vol. I, p. 638. 



