OF ARTS AND SCIENCES. 201 



applied it, that the precipitate separates from such a dilute solution as 

 to carry but a small proportion of soluble salts into the wholly un- 

 washed deposit. The following are the figures of such an analysis of 

 the precipitate pressed between folds of filter-paper without washing, 

 and air-dried. 



Weight of the reduction-tube -|- precipitate 



" " precipitate 



" " reduction-tube after ignition 

 Loss by ignition 

 "Weight of residue 



Weight of CaCl tube after ignition 

 a u « before " 



" " water 

 " " oxygen = .2514 — .2142 



Weight of crucible -j- CrjOa 

 « " « + filter ash 

 " " CrA 

 " " soluble salts = .3822 — .2882 = 



(Equiv. CrA) : (Equiv. O) = (Cr^Os found) : (1 equiv. 0). 



The proportion 76.48 : 8 = .2882 : x = .0302 gives the theoreti- 

 cal amount of oxygen necessary to bring the chromic oxide found up to 

 the composition expressed by the formula CrgOa CrOg at 0.0302 gram. ; 

 the oxygen actually found was 0.0372 gram., being seven milligrammes 

 in excess of the theoretical amount. When it is remembered that the 

 unavoidable loss in the water determination, however small, tends to 

 increase the amount of oxygen found, and that the loss in the chromic 

 oxide determination tends to decrease the theoretical amount of ox- 

 ygen, a discrepancy of only seven milligrammes will be considered 

 allowable. A second analysis of the same precipitate by the same 

 method gave even a nearer result, viz. : — 



Oxygen required to raise the CrgOs found to the 



composition Cr.^Og CrOj 0.0235 gram. 



Oxygen actually found 0.0264 " 



These two analyses, in connection with the preceding determinations 

 by the other methods, seem to us to prove conclusively that the precip- 

 itate produced by passing nitric oxide through a solution of bichromate 

 VOL. V. 26 



