O52 REPORTS ON THE STATE OF SCIENCE. 
made arrangements to continue the work on these lines, and hopes to be 
able to carry his explosion experiments to about 3000° C. by a modified 
method. " 
Clerk determined the leakage of the piston by two methods and found 
that it did not exceed 0°3 per cent. per stroke, so that error by leakage 
is negligible. 
Hopkinson’s suggestion that heat may be absorbed by a body of air 
whose mean temperature is higher than that of the walls enclosing it has 
been supposed by some to be impossible. If, however, the case be put in 
the following way it will be readily admitted that his explanation is quite 
possible. 
Referring to Fig. 2, which is a diagrammatic section of a gas-engine 
cylinder having a very elongated compression space, if it be assumed 
that air be compressed within the cylinder until the piston is full in, as 
shown in the dotted position, then the mean temperature may rise to, 
say, 240° C. as a mean throughout the whole space, but the air at the 
extreme end of that space may be cooled down nearly to the wall tem- 
perature, assumed to be 16° C. 
500°C 
FIG.2. 
The temperature will thus range from nearly 16° C. at the small end 
of the cone to perhaps 300° C. in the open part of the compression 
chamber. Such a disposition may be made to produce a mean tempera- 
ture throughout the whole compression chamber of 240° C. If the air 
be now expanded by moving the piston from the dotted position to the 
full-line position, it is obvious that the temperature of the air, which 
has been reduced to 16° C. by contact with the walls, will fall below that 
temperature, and so heat may be added to the air at the extreme end 
while heat is still being lost to the cylinder by the air near the piston. 
Many gas-engine constructions have narrow spaces, and the engine ‘ R’ 
above its inlet valve has such a space, so that the temperature throughout, 
even during compression and expansion, may be very unequal. 
Although the unequal division of heat-flow between compression and 
expansion lines must be accepted as a fact, the Committee are not yet 
satisfied as to the explanation. The view put forward by Clerk, when 
discussing the possibility of an error from this source in his original 
paper,? was that the difference, if any, between the heat-flow during expan- 
sion and during compregsion was to be ascribed to a rise in the temperature 
of the surface of the metal or of a film adhering thereto. He expressed 
the opinion that any difference so caused could nof be great on account 
of the small possible variation in temperature of the metal surface. Hop- 
kinson’s suggestion, on the other hand, was based on the possibility of 
1 Proc. Roy. Soc., Series A, vol. 1xxvii. 1906. 
