6 
EXPERIMENTAL WORK IN FUEL TESTING 
While many of the investigations can only be regarded as pre- 
liminary, it is believed that the mass of the results will be of value in 
shedding light on some matters of general interest. 
ACCURACY OF METHOD OF TAKING CAR SAMPLES. 
The regular method of taking car samples has been fully described 
under " Sampling " in Professional Paper No. 48, page 175. To check 
the accuracy of this method, check samples on two cars of coal rather 
high in moisture, ash, and sulphur were taken at the same time that 
the regular sample was taken, by opening a gate in the conveyor run- 
way every four or five minutes, and thereby allowing the contents of 
one or two buckets passing at that time to be emptied into a small bin. 
In this way a sample of 1,200 to 1,500 pounds was taken during the 
unloading of the car. This sample was then put through a small 
crusher, crushed to i-inch size, and resampled at the conveyor buckets, 
and a sample was sent to the chemical laboratory for analysis. In the 
following table the Indiana coal was a run-of-mine coal, and the 
amount sampled was approximately 20 tons; and the Ohio coal was 
a run-of-mine coal, and the amount sampled was approximately 26 
tons. 
Moisture, ash, and sulphur determinations fr 
es of two cars of run-of-mine coal. 
Indiana coal. 
Ohio coal. 
Constituent. 
Regular 
sample. 
Check 
sample. 
Regular 
sample. 
Check 
sample. 
Moisture 
10.80 
12.62 
4.39 
10.47 
12.92 
4.43 
9.01 
11.33 
4.02 
9.43 
Ash 
11.59 
Sulphur 
4.02 
For coals containing such high percentages of ash and sulphur the 
agreement between these independently taken samples is satisfactory. 
It indicates, so far as the sampling is concerned, that the analytical 
work may be taken as closely representing the average of the carload. 
MOISTURE LOSS IN COARSE SAMPLES FROM STANDING IN COVERED 
METAL PAILS. 
The large samples from the fuel-testing plant were sent to the labora- 
tory in closed metal pails, and some of them were allowed to stand in 
the pail several hours, or occasionally over night, before sampling. 
To see if any considerable amount of moisture might be lost in this 
way, a portion of the wet sample from the washery (Pennsylvania No. 
9 coal) was taken as the most favorable case, and allowed to stand for 
five days in the closed pail in the laboratory before sampling. The 
average temperature for the five days was 18° C, and the atmospheric 
humidity 60 per cent. The percentages of moisture determined by 
