Manganese as the Pyrophosphate. 



241 









Table V. 















A. 



In Platinum. 









Mn 2 P 2 



7 equiva- 















lent to Mn0 2 . 



Error in 



Error in Saturated 







Man- 



, A ^ 



terms of 



terms of solution of 





Total ganese 



Taken. 



Found. 



Mn 2 P„0 7 . ! 



tfangaaese. HN T aXH 4 P0 4 . 



NH 4 Cl. 



vol. 



in the 



grm. 



grm. 



grm. 



grm. 



cm 3 . 



grm. 



cm 3 . : 



filtrate. 



0-1885 



0-1903 



0-0018 + 



0-0007 -•- 



5 



20 



200 



none 



0-1885 



0-1910 



0-0025 + 



0-0010 + 



5 



20. 



200 



a 



0-1885 



01913 



0-0028 + 



0-0011 + 



5 



20 



200 



u 



0-1885 



0-1911 



0-0026 + 



0-0010 + 



5 



20 



200 



u 



0-3770 



0-3776 



0-0006 + 



0-0002 + 



5 



20 



200 



a 



0-3770 



0-3773 



0-0003 + 



00001 + 



5 



20 



200 



(< 



0-3770 



0-3778 



0-0008 + 



0-0003 + 



5 



20 



200 



u 



0-3770 



3783 



0-0013 + 



0-0005 + 



5 



20 



200 



£( 







B. In Glass. 











0-1885 



0-1904 



0-0019 + 



0-0007 + 



5 



20 



200 



a 



0-1885 



0-1898 



0-0013 + 



0-0005 + 



5 



20 



200 



u 



0-3770 



03767 



0-0003 — 



o-oooi- 



5 



20 



200 



a 



0-3770 



0-3784 



0-0014 + 



0-0005 + 



5 



20 



200 



« 



In this series of experiments the mean indication is, for the 

 first time, in excess of the theory. Previously the error has 

 been one of deficiency, and that in proportion to the amount of 

 manganese handled, no doubt because the amount of uncon- 

 verted trimanganese phosphate is proportioned to the entire 

 amount of the phosphate. The positive error which is devel- 

 oped in this last series of determinations is probably due to 

 the appearance of the natural error of all precipitation pro- 

 cesses — viz., the tendency on the part of the precipitate to inclnde 

 matter in solution. In the previous experiments this effect 

 was doubtless obscured by the incompleteness of the conver- 

 sion of the trimanganous phosphate to the ammonium man- 

 ganese phosphate. Indeed it is quite possible that even in the 

 last determinations the conversion is not absolute, and that 

 this is so suggested by the fact that the errors of excess are 

 larger in the case of the smaller amounts of manganese for 

 which the conversion throughout the entire work has appeared 

 to be more complete. From the consideration of the results 

 tabulated and described it would seem to be obvious that not 

 only is the presence of ammonium chloride not objectionable 

 in this analytical process, which depends upon obtaining the 

 ammonium manganese phosphate from the trimanganese phos- 

 phate precipitated from a pure solution of manganese, but that 

 its presence in not too small amount, or that of a substitute, is 

 absolutely essential to make this conversion complete. For a 

 given amount of manganese and a given volume of solution it 



