216 REPORTS ON THE STATE OF SCIENCE. 



from 1 inch to 4 inches diameter. As the air supply was reduced for 

 the same rate of gas consumption, the size of the flame increased and 

 also the heat radiated. A maximum of 15 to 20 per cent, was reached 

 for these burners when a brilliant and well-defined inner cone was 

 formed. If the amount of air supplied was in excess of that required 

 for complete combustion, the radiation fell off considerably in conse- 

 quence of the reduction in size and fall in temperature of the flame. 

 When the air supply was reduced until the inner cone disappeared, 

 with burners of this type, the flame became unsteady and was reduced 

 in temperature, the radiation falling to about 12 to 16 per cent. With 

 steady luminous flames, of the Argand or bat's-wing type, there was a 

 considerable increase of radiation on excluding air from the flame. 

 With small flames of low temperature the proportion of heat radiated 

 might be as low as 2 or 3 per cent. 



These results appeared to indicate that the radiation depended largely 

 on the size of the flame as well as on the temperature, and on the 

 presence of CO or solid C when the air was insufficient for complete 

 combustion. The mixtures employed corresponded fairly with the 

 range available in a petrol motor, but the temperature of the flame 

 in a motor, with ignition at constant volume, would certainly be much 

 higher. A considerable percentage of the loss of thermal efficiency in 

 such cases might evidently be ascribed to radiation. The exact propor- 

 tion could not be directly estimated, but it occurred to me in preparing 

 this note that the probable effect of radiation on the variation of 

 efficiency with size could be deduced by a more complete study of one 

 particular type of flame, and by measuring the radiation and absorption 

 for different thicknesses. With the assistance of Mr. G. Nelson, I have 

 accordingly repeated and extended some of these observations. 



Experiments with a Meker Burner. 



The type of burner selected for these experiments was the Meker 

 burner, with a nickel grid of 3 cm. diameter, consuming gas at the 

 rate of 0185 cubic feet a minute. The heat radiated was measured 

 in calories per square cm. per minute by an Angstrom pyrheliometer 

 at a distance of 52 cm., and the result multiplied by 4522 to deduce 

 the total radiation in calories per minute, assuming the flame to radiate 

 equally in all directions. The lower calorific value of the gas was 

 measured wet under the temperature and pressure of the experiment, 

 and was found to vary from 470 to 500 B.T.U. per cubic foot. With 

 full air-supply, the gas and air being nearly in the proportions required 

 for complete combustion, the burner gives a solid homogeneous conical 

 pointed flame, with no indications of an inner cone. As the air supply 

 is reduced minute cones make their appearance over the grid and 

 finally coalesce into a single steady brilliant inner cone, which increases 

 in size. The percentage of heat radiated rises steadily with increase of 

 size of the flame, from 10 - 5 with full air-supply to 16 per cent, as a 

 maximum with a large and bright inner cone. Beyond this point the 

 inner cone becomes ill-defined, the flame flickers, and the radiation 

 falls off to 14 per cent., rising again to over 16 as the flame becomes 



