72 
MR. J. P. JOULE ON THE AIR-ENGINE. 
the moment that the pump-valve opens admitting' a fresh supply of air into the re- 
ceiver. The cylinder should of course be provided with proper expansion gear to 
cut off the air at the required part of the stroke, which must be a constant quantity 
for each engine. The valves of the pump would of course be self-acting. 
In an engine similar to that described, it will be obvious that if the temperature of 
the receiver be kept constant, the pressure of air in it will also remain constant. For 
whilst the same quantity of air is always introduced into the receiver by each stroke 
of the pump, the quantity expelled out of it would increase with an augmentation 
and decrease with a diminution of pressure. 
In conclusion, I would recommend the examples No. 3 and No. 5 of Table II. to 
the attention of those who may be willing to construct an air-engine. In both of 
these cases the capacity of the pump is two-thirds of that of the cylinder. In the 
cylinder of No. 3 the air is to be cut olf at one- third of the stroke; and in that of 
No. 5 at one-sixth of the stroke. The temperature of the air in the receiver (sup- 
posing that of the atmosphere to be 32° Fahr.) is 625°* 145 Fahr. in No. 3, and 
924°'66 Fahr. in No. 5. The consumption of fuel in No. 3 need not exceed one-half, 
nor that in No. 5 one-third of that in the most perfect steam-engines at present con- 
structed. 
Acton Square, Salford, Manchester, 
Maij 6, 1851. 
% 
Note to the foregoing Paper, with a New Experimental Determination of the Specific 
Heat of Atmospheric Air. 
Received March 23, 1852. 
Since the above was written. Professor W. H. Miller has directed my attention 
to the probable incorrectness of the value of h, as deduced from the experiments of 
Delaroche and Berard on the specific heat of air, and my own determination of the 
mechanical equivalent of heat ; in comparison with the value deduced from the 
numerous and excellent experiments on the velocity of sound. Mr. Rankine con- 
siders that the discrepancy between the two values arises from the incorrectness of 
Delaroche and Berard’s result, an opinion which would seem to be justified by the 
entire want of accordance between the determination of these philosophers, and those 
of SuERMANN, and Clement and Desormes. I have therefore been induced to make 
the following careful experiments in order to obtain a fresh and, if possible, more 
correct value of the specific heat of air at constant pressure. 
The apparatus I employed is represented by fig. 4, in which a and h are two vessels, 
each of which contains a coil of leaden piping, eight yards long and one quarter of 
an inch in internal diameter. The coil of the upper vessel passes three-eighths of an 
inch through the bottom, to which it is soldered at c, and is thence connected with 
the coil of the lower vessel by a piece of vulcanized india rubber tubing. This part 
