0. C. Farrington — Chemical Composition of Iolite. 15 



-on account of the strong tendency of the Haddam mineral to 

 alteration. The ratios, however, as will be seen, are almost 

 exactly the same as those given by the Guilford mineral. The 

 formula of iolite is therefore H 2 . 4(MgFe)0, 4Al a O„ 10SiO a , 

 the ratio of MgO : FeO being in these two analyses very nearly 

 7 : 2. The theoretical percentages according to this formula are 

 given below, and for comparison, the mean of each of the two 

 analyses calculated to 100 per cent, the small quantities of 

 Fe 2 3 and MnO being reckoned as A1 2 3 and MgO respectively. 



Theory. 



10SiO„ 49-40 



4A1 3 3 ._ 33-60 



|(4FeO) 5-27 



|(4MgO) 10-25 



HO 1-48 



, Oak 



!. to 100. 



Guilford. 



Haddam, 



49-41 



49-21 



33-17 



33-30 



5-40 



5-24 



10-40 



10-41 



1-62 



1-84 



100- 100- 100- 



These results show satisfactory agreement, and the per- 

 centages, it may be said, do not differ materially from those 

 of the hitherto published analyses except in the state of oxida- 

 tion of the iron. The fact that the iron is present as FeO, in 

 spite of the lack of green color, which caused Scbeerer's con- 

 clusion to the contrary, shows how little reliance is to be placed 

 on color. Indeed, in a recent description of colorless iolite 

 from Brazil,* Dr. Groth expresses the belief that the usual 

 violet color of the mineral must be merely due to a pigment 

 and not to any essential constituent. For the purpose of deter- 

 mining the nature of the water, about a gram of the Guilford 

 mineral was subjected to increasing temperatures until constant 

 weights were obtained at each. The results were as follows : 









Paint 



Full 







100° c. 



300° C. 



redness. 



redness. 



Total. 



Loss in weight. 



none 



0-63 



0-87 



o-io 



1-60 



Up to full redness the mineral remained light in color but 

 on further heating, over the blast lamp, it turned black, baked 

 together and showed a slight increase in weight, owing doubt- 

 less to oxidation of the iron. 



It will be seen that the percentage lost by heating to full 

 redness is the same as that of water found by actual determina- 

 tion. Hence loss by ignition at" this degree of temperature 

 can safely be taken as representing the amount of water. In 

 the Haddam iolite it was therefore determined in this way. 

 It is possible that too intense heating may account for the 



* Zeitschr. Kryst., vol. vii, p. 594. 



