proceedings: philosophical society 477 



In conclusion, graphs were thrown on the screen exhibiting the de- 

 flection of the rays of Hght as shown especially by the observations 

 made by Dr. A. C. D. Crommelin of the British party at Sobral. The 

 detailed data as well as a photograph showing the deflected star images 

 had been courteously supplied the author by the Astronomer Royal, 

 of England, Sir Frank W. Dyson, in time to be presented at the meeting. 



Discussion: Mr. Bauer's paper was discussed by Messrs. Hull 

 and Humphreys. 



82 8th meeting 



The 828th meeting was held at the Cosmos Club, January 17, 1920, 

 President Sosman presiding and 35 persons present. 



The program was devoted to the general subject of Physical labora- 

 tory methods applied to aircraft power plants. 



The first paper was by Mr. D. Mackenzie on The velocity of flame 

 propagation in gas-engine cylinders. 



For many years automotive engineers have discussed the probable 

 velocity of the flame in the compressed and burning gas in the cylinder 

 of an internal combustion engine. Many have maintained that this 

 velocity was comparatively low and that in a high speed engine, in 

 order to obtain maximum efficiency and power, at least two spark plugs 

 must be used. It has been maintained by these engineers that by 

 igniting the mixture at two points in the cylinder, the necessary time 

 for complete combustion would be greatly reduced as compared with 

 that needed when but one spark plug is used. However, until very 

 recently, no measurements have been made in an actual engine and the 

 determination of this velocity has always been considered to present 

 extreme difficulties. 



For several months, the Bureau of Standards has been investigating 

 this subject. A single cylinder gasoline engine, the combustion space, 

 valves, piston, etc., of which are identical with those used on the Liberty 

 aircraft engine, has been employed in this work. Three spark plugs 

 are placed in the cylinder; the first is used to ignite the charge and is 

 connected to the regular ignition system of the engine, the other two 

 plugs are connected at approximately the proper time to a source of 

 direct current, the voltage of which is insufficient to break down the 

 gap between the sparking points of the plugs while under compression, 

 but which is sufficient to cause a spark to pass as soon as ionization of 

 the gap occurs, due to combustion of the surrounding mixture. These 

 spark plugs are so connected with an oscillograph that the time when the 

 flame reaches them is recorded on a strip of photographic film. The 

 distance between the plugs is accurately known and the speed of the 

 oscillograph film is also easily determined. It is obvious, therefore, 

 that by measuring the distance between the points on the film indicating 

 discharge of current across the gaps, the average speed of the flame 

 between these points may be measured. 



