72 PROCEEDINGS OF THE AMERICAN ACADEMY 



In another preparation, — 

 1.0324 gr. lost on ignition with WO.Naj 0.0922 gr. = 8.33% TsHg 



and up 

 2.0670 gr. gave 0.1255 gr. PA^Ig, = 3.88% Pp^ 

 2.0352 gr. " 0.1220 gr. " =3.84% " 



These analyses lead to the formula 



22 M0O3 . V,,0. . 3 (NHJP + 9 aq, 



3628 100.00 



If from the analyses of the two salts above described we calculate 

 the composition of the combination of molybdic and phosphoric oxides 

 supposed to be isolated, and compare this with the percentages calcu- 

 lated upon the two hypotheses of a ratio of 22 : 1 and a ratio of 24 : 1, 

 we have : — 



Calc'd. I. II. Calc'd. 



22 M0O3 3168 95.76 95.76 95.76 96.06 3456 24 MoOg 

 PA 142 4.24 4.24 4.24 3.94 142 PA 



2^fi 



3310 100.00 100.00 100.00 100.00 3598 



In both cases the phosphoric oxide was precipitated twice, but the 

 ammonia-magnesian phosphate was not treated with ammonic sulphide. 

 According to the results of Dr. Gooch already cited, the probable 

 error of this method does not exceed 1 % in excess of the quantity of 

 phosphoric oxide present. It appears, therefore, that the correction 

 to be applied to the phosjjhoric oxide in the above analyses does not, 

 at most, exceed 0.04%. The mean of Dr. Gooch's analj'ses would 

 require a deduction of 0.02% only. The yellow ammonium salt 

 analyzed by Rammelsberg corresponds to the formula 



22 M0O3 • P2O5 ■ 3 (NHJ^O + 12 aq, 



which requires (Rammelsberg) : — 



3682 100.00 99.37 



