55G 
0)1 t/tc Food of Plant. 'i. 
mum, which is placed upon a little weighed filter, well washed with the 
alcoholic mixture, dried at 212\ and weighed. Every hundred parts of 
this salt contain 6*346 parts of nitrogen. 
One great advantage of this process is that the error is divided by 16, 
ill consequence of the nitrogen being reckoned from the heavy platinum 
salt. It appears peculiarly applicable to the analysis of bodies containing 
but little nitrogen, a problem which has hitherto presented much practical 
difficulty. 
To judge how far the new nitrogen process could be applied to the 
analysis of such a body as meal to determine the proportion of gluten 
(and a more severe trial of the process itself could hardly have been de- 
vised), it was thought right to repeat the analysis of two of the samples 
of wheat-flour described in the essay, which were made by the new plan, 
using for the purpose the old " absolute method," and adopting every 
precaution that could be thought of to diminish the error of excess to 
which this method is subject. A large quantity of flour was burned, 
and the exhaustion of the tube made as complete as possible by the aid 
of a very powerful air-pump, instead of the little hand-syringe generally 
used. The following are the results :■ — 
1 . Flour from Battersea wheat. 
Quantity taken . . 19* 19 grains. 
This produced 1 • 25 cubic inches of nitrogen at 60^, which, corrected 
for moisture and pressure, was reduced to 1*24 cubic inches, weighing 
O'.SISO grains, or 1 '95 per cent. 
The other process, the error of which is in defect, gave 1 "64 and 1 •61 
per cent. 
2. Flour from nitred wheat, Berkshire. 
Quantity taken . . 19*89 grains. 
This produced 1 "2 cubic inches of nitrogen at 60°, which, corrected for 
moisture and pressure, fell to 1 ' 187 cubic inches, weighing 0*358 grains, 
or 1 • 8 per cent. 
The new method gave . . l*48andl'56. 
6, Coventry Street, Huymarkel, 
October 2, 1843. 
