272 Ilillebrand and Dana — Notes on Tyrolite from Utah* 



mineral. That these hopes were fallacious a comparison of the 

 subjoined analysis with the earlier ones at once shows : 



CuO 45-08 



CaO 6-78 



As 2 Oo 28-52 



fl 2 . 17-21 



S0 3 ._ 2-23 



Ee 2 3 ._. 008 



Insoluble.. 016 



100-06 



The percentage of S0 3 is practically the same as formerly 

 found (2*27 and 2*45 per cent). Under the microscope there 

 appeared no foreign impurity approximating in any degree to 

 6*8 per cent of the whole mass, which would be the amount of 

 gypsum represented by the S0 3 found. It therefore appears" 

 necessary to consider the latter as a proper constituent of the 

 mineral, but the construction of a probable formula is not 

 rendered thereby less difficult, and will not now be attempted. 



A few words may properly be added in reference to the 

 behavior of the mineral on drying over sulphuric acid and on 

 heating. That the large amount of water given off to dry air 

 is not altogether or probably even in large measure hygro- 

 scopic water, appears from the following table, which shows 

 the loss sustained by 1 gram of the powdered mineral in a 

 desiccator over sulphuric acid. 



Hours exposed. Loss. 



18. -0231 gram. 



26 -0083 ■' 



23 ._ -0029 " 



24 0012 " 



23 . '0008 " 



24 0001 " 



25 -0003 " 



24 -0002 " 



24 '0003 ' 



48 -0006 " 



24.. . -0002 " 



283 -0380 " 



Purely hygroscopic water would probably have been entirely 

 removed from a mineral like this after a very few hours ex- 

 posure. The table shows also that the element time may often 

 play an important part in determinations of this kind. The 

 experiment might reasonably have been considered ended after 

 the 158th hour when a loss of but T 'o mg. was shown during 

 24 hours, but nevertheless a nearly steady loss of T 3 „ mg. per 

 day took place for six days more, and might have been longer 

 observed but for the interruption of the experiment. This 

 may account for a marked discrepancy between the amount 

 lost at 280° C. in a former experiment (1. c. p. 302) — 10*34 per 



