8. L. Penfield — Connellite from Cornwall. 85 



OH so that when added to the CI it will bring the ratio up to 

 the first whole number, the analysis will be as follows : 



Ratio. 



so 3 





4-9 



•061 



1-00 



CI 

 OH 





7.4 

 •6 



■208 > 

 •035 f "^ 



4-00 



CuO 





72-3 



•911 



15-00 



H 2 





16-5 



•913 



15-04 



Loss at 100° 



C. 



•4 

 202-1 







equivalent to CI and OH 



i 



2-0 







100-1 



With this interpretation the ratio is very exact, and from 

 the repeated instances in which CI and OH mutually replace 

 one another it seems best to make this assumption to explain 

 the analysis. The formula can then be written Cu 15 (CI . OH) 4 

 S0 16 , 15H 2 0, although it is of course probable that all of the 

 Cu which is not in combination with CI and S0 4 is combined 

 with hydroxyl. The compound is very similar to spangolite 

 in composition, an analysis of which is given above for com- 

 parison, both minerals being very basic sulphato-chlorides. In 

 spangolite, moreover, the closest relation was found between 

 S0 3 , A1 2 3 and CuO, while CI was slightly deficient and H 2 

 high, and it was also stated in discussing that analysis that the 

 ratios could be made almost exact by assuming, as in this case, 

 that a little OH is isomorphous with the CI. The method of 

 analysis was on the whole like that employed in the analysis of 

 spangolite. The air-dry powder lost very little by drying in a 

 desiccator over H 2 S0 4 or in an air bath at 100° C. For deter- 

 mining the water the mineral was weighed in a platinum boat, 

 covered with dried Na 2 C0 3 and ignited in a combustion tube, 

 the water being collected in a weighed chloride of calcium 

 tube : it was found to be neutral. Before making the experi- 

 ment on the mineral a blank trial was made in which the chlo- 

 ride of calcium tube gained only -0002 gr., showing that dried 

 !N~a 2 C0 3 can be handled quickly in the air without taking on 

 any appreciable quantity of moisture. The greatest pains was 

 taken with the S0 3 determination : before filtering off the pre- 

 cipitated BaS0 4 the solution containing the precipitate was re- 

 peatedly evaporated with HC1 to remove as far as possible all 

 H]ST0 3 . . After filtering off the BaS0 4 the filtrate was again 

 evaporated to dryness, taken up in very dilute acid and water 

 and the least trace of BaS0 4 filtered off. The BaS0 4 precipi- 

 tate originally weighed O0110 gr. after purifying in the ordi- 

 nary way by fusion with Na 2 C0 3 and reprecipitating it weighed 

 0-0103 gr., showing that the first precipitate was nearly pure, 

 and as there is some chance of loss during the manipulation of 



