38 
EXPERIMENTAL WORK IN FUEL TESTING. 
liminary heating and seven minutes over the full flame of the Bunsen 
burner gave an average of 35.6 per cent, with a variation of 0.75 per 
cent between results. A result obtained by four minutes of prelimi- 
nary heating and then seven minutes over the full flame gave 35.42 
per cent. The difference in results obtained by the three-, four-, and 
five-minute preliminary treatment is small, and in all subsequent 
experimental tests the time of the preliminary heating was four 
minutes. To determine the mechanical losses and difference in vola- 
tile compounds given off, a number of ash determinations were made 
after the driving off of the volatile matter by the official method and 
after driving off the volatile matter in connection with the prelimi- 
nary heating. The results of volatile matter and ash on three deter- 
minations by the official method and on two determinations by the 
modified method of four minutes of preliminary heating and then 
seven minutes over the full flame are as follows: 
Determinations of volatile matter by two methods. 
Constituent. 
Official method. 
Modified 
method. 
1. 
2. 
3. 
4. 
5. 
Volatile matter 
66. 72 
4.30 
67.47 
4.38 
54.82 
a 7. 25 
36.06 
11.16 
36.65 
Ash 
11.15 
a This result is possibly explained by the fact that this sample stood for two hours in the crucible 
after weighing out, and a considerable amount of the moisture content may have escaped before the 
sample was treated for the determination of the volatile matter. 
That mechanical losses occurred during the rapid evolution of the 
volatile matter by the official method was also indicated by the shower 
of solid carbon particles driven off as sparks during the first few min- 
utes, while with the preliminary heating these sparks were nearly or 
entirely absent. The average volatile matter on the first two deter- 
minations was 67.1 per cent, the average ash 4. 34 per cent. The aver- 
age volatile matter on the two results by the modified method was 
36.35 per cent, ash 11.15 per cent. The moisture in the sample deter- 
mined at this time was 19.78, giving fixed carbon 32.72 per cent. The 
difference in the ash results on the two pairs is 6.81 per cent, or the 
part of the ash driven off mechanically by the regular method is 6.81-^- 
11.15, or 61 per cent. Taking this portion of the fixed-carbon result 
by the modified method gives 20 per cent as the amount of fixed car- 
bon expelled mechanically in the first determinations. The results by 
the official method after making this correction, and also after taking 
the correct ash value, are shown in column 1 of the table below. 
After making this correction for mechanical losses the difference 
in the fixed carbon by the two processes is still 10.75 per cent, which 
difference must be due to the difference in the breaking down of the 
hydrocarbon compounds by the different heat treatment. The ash 
