in the Presence of Bismuth, Antimony, etc. 139 



All the determinations were allowed to stand fifteen hours 

 or more before filtering to insure completeness of precipita- 

 tion. The filtering was performed upon asbestos in a perfor- 

 ated crucible, the precipitate thoroughly washed with cold 

 water, dried at 105° to a constant weight, and weighed as 

 cuprous sulphocyanide. 



Table I. 



Final volume 500 cm3 . 



>7Qcm3 f approximately decinormal NH^SON used for each experiment. 







As 











Calcu- 









taken as 



Tartaric 



HNH 4 S0 3 



Cu a S 2 (CN) 2 



lated 





Cu taken 



H 3 3 As 



H 2 S0 4 



acid. 



sat. sol. 



found. 



as Cu. 



Error. 





grm. 



grm. 



cm 3 . 



grm. 



cm 3 . 



grm. 



grm. 



grm. 



1. 



•3144 



0-2 



5 





5 



•5889 



•3078 



— •0066 



2. 



a 



0'2 



c: 





10 



•5978 



•3124 



— •0020 



3. 



Ct 



0-05 



a 





10 



•5919 



-3,091 



— •0053 



4. 



cc 



0'2 



u 

 HC1 





15 



•6020 



•3146 



+ •0002 







sp.gr. about 











4 









1*17 



















cm 3 . 













5. 



ct 



0'4 



15 





10 



•5969 



•3120 



— •0024 



6. 



u 



0-05 



cc 





a 



•5995 



•3133 



—•0011 



7. 



a 



0-6 



10 



1 



a 



•6035 



•3154 



+ •0010 



8. 



cc 



O'o 



u 



1 



a 



•6032 



•3153 



+ •0009 



9. 



a 



0-4 



a 





a 



•6015 



•3144 



•0000 



10. 



a 



0-2 



a 





a 



•6024 



•3148 



+ •0004 



11. 



a 



o-i 



a 





it 



•6017 



•3145 



+ •0001 



12. 



« 



o-i 



9 



1 



a 



•6029 



•3151 



+ •0007 



13. 



u 



o-oi 



10 





a 



•6022 



•3147 



+ •0003 



The negative errors of experiments 1 to 3 of Table I are 

 plainly due to the large amount of acid present, for they dis- 

 appear when the bisulphite is sufficiently increased (experiment 

 4), and the deficiencies in nos. 5 and 6 may be ascribed to the 

 same cause, for they are not found in the subsequent experi- 

 ments with arsenic where less acid was used. Tartaric acid 

 was added in three of the above determinations, but its use in 

 separations from arsenic is obviously unnecessary. 



From these results it is evident that the estimation of copper 

 in the presence of arsenic by this method is entirely practica- 

 ble in either sulphuric or hydrochloric acid solution. 



When bismuth is present a much closer adjustment of con- 

 ditions is necessary. Experiments 14 to 19 show the effect of 

 attempting to keep the bismuth in solution with hydrochloric 



