1915] on Gaseous Explosions 291 



combustion of 10 per cent and of 15 per cent mixtures is nearly the 

 same. The weaker mixture seems to lose heat a little less rapidly, 

 but this may possibly be due to a difference in the relative amounts of 

 heat lost from the copper to the backing in these two cases. No correc- 

 tion has been made in the fignres for this loss. It will be noticed 

 that at the higher gas temperatures corresponding to 100 lb. per square 

 inch and over, the heat loss with 15 per cent mixtures is less affected 

 by turbulence than at lower temperatures. This Prof. Hopkinson 

 ascribes to the influence of radiation, which he states is probably 

 unaffected by the motion of the gas, and is a very important cause 

 of loss when the gas is hot. Experiments were made to determine the 

 relation between the velocity of the air and the speed of the fan. 

 The velocity of the air was estimated by the cooling effect on a 

 platinum wire heated by an electric current, and they show that at a 

 fan speed of 2500 revolutions per minute the velocity of air near the 

 walls is probably greater than that obtaining in a gas engine at the 

 end of the suction period. From this it is inferred that the turbulence 

 in the engine is insufficient to produce any great effect on heat loss. It 

 will be noted that in Fig. 18 the time of explosion with the fan run- 

 ning at 3600 revolutions per minute is about one-fourth that obtain- 

 ing with the fan at rest. The time of explosion with the fan in 

 motion is about 0*0125 second, and with the fan at rest 0*05 second. 

 In Fig. 18 at 2300 revolutions the time is about 0*03 second ; while 

 at rest it is ' 14 second. 



That residual turbulence exists at the end of the ordinary com- 

 pression stroke in a gas-engine cylinder has also been proved by 

 Hopkinson by similar experiments on an engine driven by belt, in 

 which experiments the rate of heat flow was determined at the end of 

 the first compression and also at the end of subsequent compressions. 

 For this purpose a platinum wire was mounted in the combustion 

 space of an engine of 7-inch diameter cylinder and 15-inch stroke. 

 The wire was heated by an electric current. Within moderate limits 

 of temperature the heat loss from such a wire is proportional to the 

 temperature difference between it and the surrounding gas. The 

 ratio between heat loss and temperature difference is a measure of the 

 effective conductivity of the gas, and depends on the temperature 

 density and state of motion. If the first two factors are the same, 

 then the effective conductivity depends only upon the state of motion, 

 and may be taken as the measure of its amount. The engine was 

 motored round so as to compress and expand charges of air, the gas 

 supply being cut off, and comparative measurements of effective 

 conductivity at the top of compression were made, first, with the 

 engine valves working in the ordinary way, and, second, with the 

 valves closed, so that the same charge was continually compressed and 

 expanded, and there was therefore no turbulence resulting from 

 suction. It was found that at 240 revolutions per minute conduc- 

 tivity was more than 60 per cent greater in the first case than in 



