194 REPORTS ON THE STATE OF SCIENCE.—1912. 
effective external circulation of water is possible, cooling does not 
present great difficulty ; but at places which are not near to the cooling 
water, so that the heat has to travel a long way, the temperature must 
be high to give the necessary gradient. Thus the central portion of the 
head of an ordinary flat- faced. piston if not water-cooled gets very hot, 
reaching a temperature of perhaps 600° C. in a four-cycle engine of 
24 inches bore. The piston expands considerably in consequence, the 
expansion being greater at the centre than at the edge which is accord- 
ingly put into tension. In larger cylinders the stresses in the piston 
set up by unequal heating, and the danger of pre-ignition arising from 
the hot metal, necessitate the cooling of this part by the circulation 
of oil or water. Even then the great thickness of metal in certain 
portions of the combustion chamber, and the difficulty of keeping 
the water flowing properly in every corner, may cause high local 
temperatures. 
‘The heat carried away by the cooling water and by radiation is the 
total given to every part of the walls, and its measurement gives no 
information on the important question of the manner in which the flow 
is distributed over the walls. It is certain, however, that the greater 
part of the heat-flow in a cycle occurs in a comparatively short time 
just after the moment of ignition, and passes therefore into the surface 
of the combustion chamber and valves and into the face of the piston. 
But little goes into the barrel of the cylinder, which is not uncovered 
until the density and temperature of the gases have fallen. That this 
must be so is obvious, but the magnitude of the effect is perhaps not 
generally recognised. Dugald Clerk found in his experiments on the 
compression ‘and expansion of flame* that the average heat- flow per 
square foot per second in the first three-tenths of the stroke is’ three 
times that of the average over the whole stroke for equal temperature 
differences, and he calculates that the actual rate of heat-flow in the 
first three-tenths is six times that of the whole stroke in ordinary gas- 
engines working at full load. This estimate, however, does not include 
loss due to radiation before maximum temperature. In the actual firing 
and expansion stroke of a gas-engine the difference must be even more 
when radiation and other losses incurred before maximum temperature 
are included, ‘and it is probable that in discussing the problem of cool- 
ing the metal it is a sufficiently good approximation to neglect the heat- 
flow into the outer half of the barrel altogether, ‘Professor Hopkinson 
informs the Committeé that he-has worked a gas-engine cylinder of over 
30 inches diameter in which there was no water circulation round the 
barrel.at all. The whole of the heat passing into the barrel was in this 
case remoyed either by radiation or by conduction into the piston, nor 
was the cooling which was applied to the piston much more than that 
found necessary on other parts of the walls of the combustion chamber, 
In small engines with uncooled pistons the water-jacket round the 
barrel is necessary to keep the piston cool. 
In the scientific analysis of gas-engine phenomena the facts stated 
in the last paragraph are important because they show that the heat- 
Proc. Roy. Soc. fae: i vol, 77 (1906), p. 500. 
