OP ARTS AND SCIENCES. 91 



greater number of cases. In its place we may often, but not always, 

 employ the following, which usually gives good results in the cases of 

 the alkaline salts. Potassic cyanide, as free as possible from carbo- 

 nate and cyanate, is to be fused in a crucible of Berlin porcelain, and 

 then allowed to cool. A weighed portion of the antimonoso-tungstate 

 is then to be placed upon the fused mass, and the whole cautiously 

 heated to low redness, and kept for a short time at this temperature. 

 The antimony is completely reduced to metal, and after the solution 

 of the alkaline salts in cold water, and washing at first with water 

 and afterward with alcohol, may be weighed upon an asbestos filter. 

 It will be seen that this process is the same as that devised many years 

 ago by myself for the separation of tin from tungsten, worked out in 

 my laboratory by Mr. F. H. Talbutt,* and known as Talbutt's method. 

 The process unfortunately does nut apply to cases in which insoluble 

 compounds are formed together with the reduced metallic antimony. 

 As a rule, when the process is successful, the antimony is linally in the 

 form of large bright metallic globules, but in some cases it forms on 

 reduction wholly or partially a black powder containing tungsten, and 

 the method then fails. In these cases an accurate result may be ob- 

 tained by mixing the salt or the potassic cyanide with a quantity of 

 potassic or sodic carbonate sufficient to completely saturate botli the 

 tuugstic and antimonie oxides present. This variation in the applica- 

 tion of the method is due to my assistant, Mr. Herman Schmidt. In 

 determining antimouious or antimonie oxides by this process, there is, 

 of course, the disadvantage that a higher is always estimated from a 

 lower molecular weight. The following analyses will serve to show 

 tlie degree of accuracy which may be obtained under favorable circum- 

 stances, the method applying to antimonio-tungstates as well as to the 

 antimonoso series. In an antimonio-tungstate of potassium not of 

 definite constitution, 



1.0924 gr. gave 0.4407 gr, metallic antimony = 53.79 ^ SboOs 

 1.0316 gr. " 0.4155 gr. '• " = 53.71 % " 



In another compound, 



0.9059 gr. gave 0.2251 gr. metallic antimony = 33.13 <^ Shfi^ 

 1.0312 gr. " 0.2558 gr. " " =33.08'}^ " 



Antimouious oxide may also be determined in these compounds by 

 titrition with iodine, after adding to the solution an alkaline tartrate 



* Am. Journal of Science and Arts, 1. 7. 



