Chap. IV. ] hollandite. 95 
It should be noticed that in calculating the mineral composition of 
composite specimens of ore from their analysis, a difference is usually 
found between the amount of oxygen determined and that required for 
the manganates left over after the braunite has been extracted from the 
analysis. This difference is, however, never sufficient to point to the 
presence of some other compound rather than manganates corre- 
sponding to the theoretical acid H4Mn05 ; in fact the errors are always 
sufficiently small to be attributable to experimental errors in their 
determination. The differences between the amounts of oxygen required 
for the manganates in the above analyses and the amounts actually 
found are as foUows : — 
per cent. 
Kajlidongri —0-37 
Balaghat +0-61 
Gowari Warhona +0'85 
Junawani — 0'52 
These differences may, as already stated, be due to experimental errors 
in the determination of the oxygen, but the discrepancies can be explained 
in another way. In the cases where there is a surplus of oxygen it may be 
that a portion of the manganese forming the basic radicle of the mangan- 
ates is not in the protoxide form corresponding to the formula Mn2Mn05, 
but is in the sesquioxide form corresponding with the formula 
Mn4(Mn05)3. The conversion of the requisite amount of Mn2Mn05 into 
Mn4(Mn05)3 would account for this surplus oxygen. In the same way 
when there is a deficit in the amoimt of oxygen according to the above 
way of representing the mineralogical composition of the ores, this deficit 
can be accounted for by the conversion of an equivalent amount of ferric 
manganate, Fe4(Mn0.5)3, mto the ferrous manganate, Fe2Mn05. This 
point must for the present remain open to doubt ; but it is to ))e 
noticed that hi the case of some of the analyses of psilomelane the 
discrepancy is perhaps too great to be attributable to experimental 
errors. 
Froni the foregoing paragraphs it follows that the composition of 
these crystalline manganates can be represented by the general formulae 
given below, in which the constituents are placed in order of import- 
ance in their respective brackets : — 
TO (]\fci,Ba,K2, H2,Ca,Mg,Fe,Co)2]VIn05 + M(Fe,Al,Mn)i(Mn05)3. 
This is, of course, the most general formula, but it can be simplified, 
by omitting the least important constituents, as follows : — 
TO (iMn.Ba,K2.H2)2Mii05 ^ n {Fe,Al)4(MnOr,)..,. 
