538 Palache and Merwin — Connellite and Chalcojphyllite. 



interference color was red. From either of these colors the 

 difference of retardation of the two refracted rays could be 

 closely estimated. But when green was the color assumed, the 

 interference color might have been either orange, yellow or 

 green. In this case the one-quarter undulation mica plate was 

 inserted so as to raise or lower the color a sufficient amount to 

 produce either blue or purple. Then by making proper allow- 

 ance for the mica plate the interference color was calculated. 

 By measuring the thickness of the prism the birefringence of 

 the mineral was calculated from the interference color and 

 found to be -021. This figure, the mean of seven observations, 

 is as likely to be correct as the one derived from the refraction 

 indices, '022. 



Chemical composition — (Merwin). Material sufficient for 

 chemical analysis (0*73 gram) of ideal purity was easily 

 obtained by hand-picking. The analysis is compared below 

 with the only other one we have, that of Fenfield made on 

 •074 gram, and leads to a somewhat different formula. The 

 water was given off in three distinct fractions at about 235°, 

 260° and incipient redness. A little of the second fraction of 

 water was seen to come off before the heat was removed in 

 estimating the first fraction. 



Connellite, Molecular Eatios Connellite, 



Merwin , * , Penfield , «- x 



S0 3 3-48 -043 -98 1 4-9- -061 1' 



CI 6-37 '18 4-09 4 7'4 '209 3-4 



CuO 75-96 -959 21-84 22 72-3 -913 15 



H,0 below 220° -25 -4 



H 2 0, 220° -260° -- 12-06 '61 15*24) ) 



H 2 0, 260°-300° .... 2-10 -117 2-65 [ 20 f-.16"8 '933 15-3 



H 2 above 300° 1-66 '092 2'11 ) ) 



101-83 101-8 



LessO = Cl 1-42 1-67 



Total 100-41 100-13 



As shown by the molecular ratios, the composition of this 

 connellite may be expressed by the empirical formula 

 Cu 22 Cl 4 S0 23 + 20H 2 O. The corresponding formula for Fen- 

 field's analysis is Ou 1B (C1.0H) 4 S0 16 +l5H s O. 



The distribution of the water might be accounted for by 

 supposing some such molecular grouping as this : 



[CuS0 4 .3Cu(OH),.H s O].2[CuCl 3 .Cu(OH)J.14[Cu(OH)J. 



The first part of this formula is identical with that of the 

 mineral langite, which loses one equivalent of water at a 

 moderate temperature. The fact that when first heated con- 

 nellite yields water alone indicates the presence of a copper 

 hydrate molecule easily decomposed by heat; thus the fifteen 



