96 
Some EXxpeERIMENTS ON LocoMoTIVE COMBUSTION. 
By J. W. SHEPHERD. 
Through the courtesy of the T. H. & I. Railroad officials, a study of 
the combustion in a locomotive while in operation was undertaken, the 
study being made from the analyses of the stack gases. The analyses 
were made with a modified Orsat apparatus. 
The experiments were conducted on the large Schenectady passenger 
engines and on fast runs between Terre Haute and St. Louis, and Terre 
Haute and Indianapolis. 
The apparatus for sampling the gases consisted of a half-inch gas pipe 
extending eight or ten inches into the center of the stack and bent uni- 
formly following the outside of the stack to near its base, where another 
bend led it back to the cab on the fireman’s side. Through this pipe, with 
the proper connections inside the cab, the gases were drawn into bottles 
by means of a steam jet. The bottles were fitted with ground-glass 
stoppers. The end of the gas pipe within the stack was fitted with a 
thimble, the lower end of which was solid steel and the sides perforated. 
This particular fitting was found to be essential to the successful operation 
of the apparatus. 
Samples were taken in three different ways: (1) For periods of one to 
two minutes by displacement of water; (2) for continuous samples, from 
Terre Haute to terminals of road if desired, by displacement of water, 
and (3) for any period of time (brief as desired and whenever desired) by 
air displacement. In methods Nos. 1 and 2 the water displaced was 
acidulated with sulphuric acid. In method No. 3 the gas was passed five 
times before the bottle was disconnected from sampler. 
Method No. 1 was not satisfactory, because the fireman did not fire 
normally during the sampling. 
Method No. 2 showed that the value of a fire does not always vary 
directly as the increase of carbon dioxide and decrease of free oxygen in 
the stack gases. Samples Nos. 1 and 10 in the following table will serve 
as an illustration. No. 1 was taken continuously from Terre Haute to 
Effingham, Ill. (sixty-eight miles), and No. 10 from Effingham to East St. 
Louis (100 miles), both on same train, but different crews. No. 1 required 
less coal per car mile than No. 10. It is to be observed that No. 10 shows 
a percentage of carbon monoxide, which means that the rapid evolution 
