200 REPORTS ON THE STATE OF SCIENCE.—1912. 
diameter but of different lengths, the density of the mixture fired being 
varied in inverse proportion to the length (so as to keep the total 
quantity constant), would be of great interest. Such a comparison 
would throw direct light on the heat-flow in an actual gas-engine if 
among the cylinders tried were some in which the length was a 
fraction—say, one-sixth to one-third—of the diameter. In many gas- 
engines the shape of the combustion space is approximately of this 
character. 
5. Turbulence.—During the suction stroke of a gas-engine, or 
during the period of injection in an engine charged from a separate 
compressor, the mixture of gas and air which is subsequently to be 
exploded enters the engine through the valves or ports at a high 
velocity, so that the gas within the cylinder is in a state of eddying 
or turbulent motion. This motion gradually dies away after the valves 
are closed, but will persist for some time during the compression 
stroke, so that at the moment of explosion there may still be a good 
deal of turbulence. In consequence of this motion of the gas the con- 
vection of heat will go on more rapidly, and what may be called the 
‘ effective conductivity ’ of the gas will be increased. 
Perhaps the most obvious direction in which to look for the effects 
of turbulence in gas-engines is the relation between speed and fuel 
economy, and this aspect of the matter has been discussed by Callendar, 
Lanchester, and others. On the one hand it may be expected that 
the longer the time taken over the operations of compression and 
expansion the more heat, other things being the same, will pass into 
the walls during that period. As against this must be set the con- 
sideration that, with a given valve opening, slow speed means less 
turbulence, first, because the velocity of entry of the gas is less, and 
second, because the time available for the resulting turbulence to die 
out under the influence of viscosity is longer. Reduction of speed 
therefore means less effective conductivity, and it is even conceivable 
that on this account the heat-flow per cycle may be less and the fuel 
economy greater at the lower speed. ‘The effect of heat-flow upon 
economy is not very marked, and it is therefore not surprising that 
no decisive verdict has yet been pronounced on the relation between 
economy and speed. There is no doubt that, given satisfactory 
ignition, economy is somewhat improved by increasing the speed, but 
the relation betwéen these two things has not been so precisely deter- 
mined as to permit a conclusion to be drawn about the part played 
by turbulence, nor in view of the complication of the question does 
it seem likely that much information can be derived from this source. 
A more promising line of inquiry would be a direct measurement of 
jacket-losses at different speeds. The Committee are not aware that 
any very accurate measurements of jacket-loss at different speeds, 
other conditions being kept rigorously the same, have ever been under- 
taken. From some rather rough measurements of this character made 
by various members of the Committee, it appears that the heat-loss per 
cycle does undoubtedly diminish with increase of speed, but not in 
proportion thereto. 
The complete elucidation of the part played by turbulence in the 
