575 
DE. H. DEBUS OH THE CHEMICAL THEO&Y OE GOTPOWDEB. 
(VIII.), hence the calculated heat generated by the reactions between saltpetre, sulphur, 
and pure carbon of 1 grm. of English service powder is : 
704 x93-75 
100 
= 660 cal. 
This theoretical quantity is 59‘9 units, or 8‘4 per cent, less than the amount found by 
experiment, a difference which would be much smaller if the amount of heat produced 
by the action of the sulphur upon the iron of the apparatus were known and could be 
subtracted from the experimental number. It is worthy of notice that the differences 
in the amounts of heat found in several experiments made with the same description 
of powder are nearly as great as the differences between the calculated and observed 
results. 
A sample of powder manufactured by Curtis and Harvey, and marked No. 6, gave 
in four experiments 
I. II. III. IV. 
732-9 744-9 755-7 784 
units of heat, hence, between the first and last experiment a difference of 51T units. 
Another heat determination may be here introduced. 16KN0 3 + 16'4C + 5'5S of 
the Spanish pebble powder burnt according to the reactions on which equation 
(VIII.) is based, 16’4 for y, and 5*5 for 2 put in equation (X.), make W= 1501250 cal. 
As 
16KN0 3 +16-4C + 5-5S 
are equal to 1988*8 parts by weight, 1 grm. of the powder generates 754 cal. 
If 5 per cent, are deducted for hydrogen, oxygen, and ash, we obtain : 
754x95 
~ 100 
= 716-3 cal. 
for the calculated heat of 1 grm. of powder. Experiment gave 762*3, or 46 units 
more. 
It follows as a general result from these considerations that the mean quantities of 
heat generated by the combustion of the English service and Spanish pebble powders 
are about 60 units greater than the theoretical values. If, however, the amount of 
heat generated by the action of the sulphur upon the iron of the explosion apparatus 
were known, and could be subtracted from the observed quantities of heat, the 
difference would become much smaller. 
On the other hand, the theoretical numbers stand to each other nearly in the same 
ratios as do the corresponding experimental values. 
The equations (VIII.), (IX.), and (X.) will now be used to determine the composition 
of an ideal powder, that is to say, of a powder composed of saltpetre, pure carbon, and 
MDCCCLXXXII. 4 E 
