1104 
MANGANESE DEPOSIT? OP INDIA : DESCRIPTIVE. [ PaRT IV ! 
^ . . . , (usually with apatite), have been intruded, 
Origin of the ores. " ^ 
and that as usual the ores were subsequently 
formed at the expense of the manganese-garnet. Since the latter mineral 
is often seen in quite -fresh grains scattered through psilomelane other- 
wise of good quality, and since the kodurite is seen, when it undergoes 
decomposition, frequently to have its spandite changed to ochre, it 
seems that when the kodurite undergoes alteration the spandite of one 
portion is decomposed and the manganese removed, leaving the iron 
behind as ochreous spots, and that the manganiferous solutions so 
produced percolate to another part of the rock (often close by, some- 
times in the same hand-specimen — see page 1 109) and there replace 
felspar without affecting the spandite, so as finally to produce psilome- 
lane studded with the unaltered spandite. 
The intruded mass of rock, including all the felspathic rocks except 
the quartzite, must have been at least 60 feet wide, to judge from the 
abundant kaolin, ochre and wad now visible in the Sonpuram pits (see 
fig. 91, page 1107) ; probably if properly exposed by cross-cuts traversing 
the whole width of these rocks, the total width would be found to be at 
least 100 feet including partings of the original quart zites. As regards 
the age of these quartzites, they may be regarded as probably part of 
the khondalite series. 
In the Sonpuram workings I saw but little spandite, so that the ore 
visible there was probably largely formed by replacement of rock that 
originally contained no spandite. In the Mamidipilli pits spandite is 
very abundant and the remarks in the preceding paragraphs apply 
especiallv to these. I did not examine the Bankuruvalsa pits carefully 
enough to state that spandite was absent there, but it was very notice- 
able that the sample (No. 26) taken from the Bankuruvalsa ore-stacks 
showed no spandite, whilst that (No. 25) from the Mtlraidipilli stacks 
exhibited abundance of it. 
The ores are psilomelane and pyrolusite, the latter being most abun- 
dant at Bankumvalsa ; but in all three villages 
thfores'' ''"'^ psilomelane is the predominant mineral. The 
character of the ores of each village can be 
best judged from the following samples : — 
Sample A. 24. (Sonpuram) taken from some 350 tons of stacked 
ore. Composed mainly of psilomelane, both compact and cavernous 
