726 The Trials of Oil Engines at Cambridge. 
quently to be remarkably free from backlash errors. The indi- 
cators were well and frequently lubricated with Russian tallow, 
which was found to work veiy satisfactorily. They were all 
afterwards tested under steam on a mercury column in the 
engineering laboratory at King’s College, London, and correc- 
tion curves plotted. The errors were not beyond average indi- 
cator errors, but were in no case to be disregarded, and the true 
spring scales have in all cases been employed. 
In the half-power and light trials, half-minute cards in most 
cases showed such constant and wide changes in mean pressure 
from explosion to explosion that no reliable determination of 
the indicated power could be made without an integrating 
indicator. The only half-power diagrams which could be worked 
out with accuracy were those of Messrs. Crossley, which were 
very constant. 
On the full-power trial, the diagrams obtained from the 
Britannia Co.’s engine showed considerable fluctuations, due to 
the method of governing. But a comparison of half-minute 
diagrams with a number of double cards taken at successive 
intervals shows mean results within 1 per cent, of one another. 
The indicated power given in the table may therefore be relied 
upon as correct within those limits. 
The water was measured as it was discharged from the 
jackets into standard thirty-gallon tanks, which had been 
previously calibrated. The mean rate of filling these was 
ascertained and the discharge per minute calculated. Readings 
for temperature were taken every quarter of an hour. 
In the tables on pp. 728-729 will be found the mean 
quantities for all three trials of the six engines which completed 
them. 
Oil consumption. — Five engines used less than a pound of 
oil per indicated horse-power per hour, and two, Messrs. 
Crossley’s and Messrs. Hornsby’s engines, developed a brake 
horse-power on 082 lb. and 0-977 lb. per hour respectively. 
In the short two hours’ trial Messrs. Hornsby’s engine 
necessarily appears less economical than over a longer run, be- 
cause the starting lamp consumes an appreciable fraction of the 
whole supply. Where much stopping and starting again are 
required this larger consumption would, however, show itself. 
The record of consumption of Messrs. Crossley’s engine is 
remarkable : 0-82 lb. of oil represents a cost for fuel of 0 37 d. 
per brake horse-power per hour, or the equivalent of 3 lb. of 
Welsh coal. Few condensing steam-engines of equal size work 
on a less amount. 
Reckoned on the three days’ run, the cost per brake horse- 
