VOLATILE MATTER TN FUELS. 37 
Laboratory analyses on samples 2734 and 2764 of Texas No. 3 lignite. 
Samples. 
Constituent. 
2734. 
2764. 
1. 
2. 
3. 
Moisture 
9.88 
36.17 
43. 65 
10.30 
1.30 
20.24 
58.48 
10.85 
10.43 
1.03 
20.24 
35.42 
33.91 
10.43 
1 03 
Volatile matter 
Fixed carbon 
Ash 
Sulphur 
This great difference in the fixed-carbon results shown in columns 1 
and 2 could not be accidental, as all of the determinations on both 
samples were duplicated. A series of determinations was begun, to 
learn, if possible, the cause of this great variation. The two follow- 
ing causes were suspected, both of which were found to be partly 
responsible for the difference: (1) Mechanical loss due to the throwing 
out of solid particles by the too rapid expulsion of the volatile matter. 
The possibility of loss from this source is mentioned in the report of 
the committee of the American Chemical Society, but the results of 
their experiments are negative. (2) A different breaking down of the 
hydrocarbon compounds when expelled under different conditions and 
in the presence of variable amounts of moisture. 
The results of Mr. N. M. Austin's preliminary treatment of tne 
sample with a low heat and then with the application of the full flame 
of the Bunsen burner gives higher results in fixed carbon than where 
the full flame of the Bunsen burner is applied from the beginning. 
The proximate analysis of sample 2764, giving the unusual results, 
which was finally reported by the laboratory, is shown in column 3. 
A series of seven results by the official method gave for volatile 
matter on this sample an average of 62.5 per cent, with a variation 
between high and low results of over 12 per cent. Three results of 
volatile matter on this sample made after previous expulsion of the 
moisture at 105° C. gave average volatile matter 39.6 per cent, with a 
variation between high and low results of 5.9 per cent. Preliminary 
treatment by driving off the moisture and most of the volatile matter 
at a low heat was then tried, the flame of the Bunsen burner being 
turned down to 10 centimeters and the crucible gradually heated. The 
application of the heat was regulated by holding the burner in the 
hand and heating in such a way as to expel the moisture slowly and 
gradually smoke off most of the volatile matter, the volatile matter 
escaping freely enough during the last minute of this preliminary 
heating to burn with a small flame around the edge of the crucible 
cover. Two results with five minutes of preliminary heating and then 
seven minutes over the full flame of the Bunsen burner gave an 
average in volatile matter of 35.08 per cent, the variation between the 
two results being 0.23 per cent. Two results with three mi mites' pre- 
