PROCEEDINGS OF THE ACADEMY AND AFFILIATED 



SOCIETIES 



THE PHILOSOPHICAL SOCIETY OF WASHINGTON 



The 760th meeting was held on October 16, 1915, at the Cosmos 

 Club, President Eichelberger in the chair; 35 persons present. 



Mr. E. C. Crittenden presented an illustrated paper giving the re- 

 sults of an investigation in collaboration with Messrs. E. B. Rosa and 

 A. H. Taylor on Effect of atmospheric pressure on the candlepower of 

 various flames. In order to calibrate flame standards of candlepower 

 by comparison with electric standards it is necessary to know the 

 effects of various atmospheric conditions on the flames. Humidity and 

 atmospheric pressure are the conditions which cause most variation. 

 The former varies considerably from season to season, and its effect 

 has been determined by observations extending over several years. 

 The natural variations in pressure from time to time at any one place are 

 not great enough to determine their effect with precision, and at any 

 rate extrapolation to other pressures would be unreliable. By using 

 a set of tanks in which air could be supplied at high or at low pressure, 

 the variation of the candlepower of pentane and Hefner lamps over a 

 wide range of barometric pressure has now been determined. The varia- 

 tion is not linear. In general the intensity of a flame increases with in- 

 creasing pressure, but at a decreasing rate, until the flame becomes 

 smoky, when a further increase in pressure may cause a slight decrease 

 in intensity. Conversely, decrease of pressure in general causes a de- 

 crease of candlepower, which is more and more rapid as the flame gets 

 farther from the smoking condition. The pressure presumably affects 

 the rate of diffusion, of oxygen through the fuel. When the pressure 

 is increased the diffusion is slower, the process of combustion is re- 

 tarded, and the time during which the carbon particles exist in the 

 glowing state is increased. Consequently at any one time there are 

 more particles giving out light. On the other hand the average tempera- 

 ture of the particles is lowered, as is shown by the increasing redness of 

 the flame, so that the light emitted by each particle is greatly reduced. 

 Eventually such a condition is reached that the decrease in temperature 

 counterbalances any further increase in number of glowing particles. 

 Besides the data for the standard lamps, curves were shown to indicate 

 the effect of variation in air pressure on the candlepower of gas burned 

 in several types of burners. The effect of humidity on a gas flame, an 

 acetylene flame, and a kerosene standard lamp was also shown. As an 

 application of the data presented, a detailed discussion of the signifi- 

 cance of gas candlepower tests, as made at present, was given. 



51 



