1879.] on Coining Silver into Rupees. 57 
Having the weight and fineness of each sample of silver in a pot, we 
are in a position to compute its fineness on the supposition that no change 
takes place in melting: this I call the “ Theoretical Fineness.” When the 
contents are melted and well mixed, a small spoonful of the fused alloy is 
granulated and from this a muster is delivered to the Assay Master: the 
fineness of this I have called the “‘ Fineness of Pot,” it is generally greater 
than the Theoretical fineness. In the later processes and especially in that 
of “ pickling,” preparatory to coining, the fineness is further increased, and 
the final result is determined from an assay of the coins by taking a propor- 
tion of coins for assay singly, and also some for assay after melting them up, 
This last determination is the least satisfactory ; however uniform the melted 
mixture may be, the alloy is not equally distributed in the resultant 
ingots and every after process tends to increase this irregularity ; so that at 
last, not only are the various coins different in their fineness, but portions 
taken from different parts of the same coin are so. I have used as a measure 
of the fineness of the coins of one day, the mean result derived from 20 
esingle coins—the sample piece being always cut out from the centre of 
the coin, and I have cailed the result “‘ Fineness of Coins.” 
During these experiments 10 pots were daily alligated to the same 
Theoretical Fineness: I have thus had a measure of the accuracy of the 
Assay Reports, and I have used this for calculating the probable errors of 
the theoretical finenesses, in a way which (though somewhat arbitrary) seems 
to me sufficiently accurate for the purpose. When the probable error of an 
Assay Report is known, it is easy to calculate that of one heap, made of 
several samples of one quality, on the supposition that the whole is fairly 
mixed. As, however, the mixture must at best be very imperfect, I have 
preferred assigning to each quality of silver the same probable error of 
fineness as though all had depended on a single report. 
As any erroneous hypothesis as to the quality of scissel used would 
‘clearly have vitiated the results, I had a quantity melted down, assayed, and 
laminated, each pot being kept separate, and thus [had metal which was 
of known fineness—save the small change from lamination which would 
equally be shared by all scissel—but which I conceived would be subject 
in melting to the same changes as scissel itself. 
I had intended to keep the work from each pot separate all through, 
but after a certain point this was found impracticable, and the coins from a 
single day’s melting have been mixed. After I had completed the greater 
part of the calculation for this paper, I found that, by a careless blunder, 
there had been a mixing of the coins of the second and third days’ meltings : 
and though I could only prove that it had been slight, and it probably would 
not have seriously affected the result, I had the work of those days repeated 
