136 Mr. Dyke on the Determination of the Mean Spherical 



It should also be noted that Rayleiglr's method is equally 

 applicable to all the simpler types of vibration. The pro- 

 cedure is the same in all cases, and like accuracy is obtained. 

 In the case even of lateral vibrations, when the rod is 

 " supported," or is clamped at both ends, Mr. Garrett's 

 method would seem to require modification, and supposing 

 such modification possible, no assurance could be felt in the 

 absence o£ some guiding general principle as to the degree 

 of accuracy of the results, or as to the direction in which 

 they would differ from the true frequencies. In the two 

 cases actually considered, Mr. Garrett's results for the 

 frequency and those obtained by Lord Rayleiglr's method 

 from the same assumed type of vibration differ from the 

 truth in opposite directions. If Garrett's method admits of 

 generalization, and can be proved to supply a minimum as 

 Rayleiglr's method supplies a maximum value for the 

 frequency, it may prove of real service in acoustics. 



December 3, 1904. CHARLES Chree. 



XII. On the Practical Determination of the Mean Spherical 

 Candle- Power of Incandescent and Arc Lamps. By G. 

 B. Dyke, P.Sc, Salomons Scholar of the Institution of 

 Electrical Engineers* . 



[Plate III.] 



ri^HE following experiments were undertaken at the sug- 

 JL gestion of Dr. J. A. Fleming, F.R.S., as a piece of 

 work during the tenure of a Salomons Scholarship, and were 

 carried out at the Pender Laboratory, University College, 

 London, under his direction. 



The principal object of the research was to investigate the 

 relation between the mean spherical candle-power of various 

 types of incandescent and arc lamps, and the candle-power 

 measured in a horizontal direction. 



It is obvious that some improved method of expressing the 

 efficiency of electric lamps is needed, having regard to the 

 fact that a nominal 16-candle power incandescent lamp may 

 give from 12 to 20-candle power in different directions in a 

 horizontal plane, corresponding to a variation in so-called 

 efficiency of from 5 to 3 watts per candle. 



Such a method has been pointed out by Dr. Fleming f, 



* Communicated bv the Physical Society : read November 11. 1904. 

 t " The Photometry of Electric Lamps/' by Dr. J. A. Fleming, F.R.S. 

 Journal Instit. Electric Engineers, vol. xxxii. p. 165. 



