W. L. Stevens — Sensitive Flame as a means of Research. 259 



tions when the attempt is made to apply it to purposes of 

 measurement. 



I have attempted by means of the whistle and flame to 

 verify acoustically the experiment in light first performed by 

 Grimaldi and analyzed by Dr. Thomas Young, that of produc- 

 ing diffraction bands by transmitting waves in the same phase 

 through two small openings, and exploring the air with the 

 sensitive flame for the hyperbolic lines of maximum and mini- 

 mum motion. The whistle, giving forth waves 1*05 inch in 

 length, was placed 34 inches from the screen of cardboard, 

 whose width was two feet. Near the middle of this were cut 

 two vertical slits, 3 inches apart, and each J inch wide. The 

 position required by theory for the hyperbolic bands was deter- 

 mined, the screen being at right angles to the direction of the 

 whistle from its middle point. The middle line of maximum 

 motion behind the screen was detected without difficulty. It 

 was discontinuous, as might be expected when the wave length 

 is so considerable in comparison with the distance between the 

 apertures. The nearest hyperbolas on the two sides of this 

 were found in their right position, and traced back rather 

 more than a foot from the screen, but they were not so well 

 defined as the middle line. The next pair of hyperbolas was 

 also found, but with poor definition. By using slits a half -inch 

 in width results were perhaps a little better, though in neither 

 case could any measurements approximate to exactness. 



Fresnel's celebrated experiment of producing interference 

 bands by reflection of light from two mirrors inclined at an 

 angle of nearly 180° was tried by Professor A. M. Mayer and 

 myself conjointly, using sound waves. A large plate of glass 

 was rested on the table, and another plate inclined to it at an 

 angle of 152°, the whistle being 67 inches from the flame, 4 

 inches from the inclined mirror, and 13 inches above the table. 

 Six interference bands were detected by means of the flame, 

 their mean distance apart being 4 inches. By subsequent cal- 

 culation this result was found correct to within a tenth of 

 an inch. An important source of uncertainty, however, in 

 this experiment arises from the waves proceeding directly from 

 whistle to flame. Even if a screen is interposed, enough space 

 has to be left below it to allow for the passage of sound rays 

 reflected from the two mirrors. From the lower edge of the 

 screen, therefore, waves are diffracted and may interfere with 

 either or both sets of waves reflected from the mirrors. The 

 trouble from this source caused the abandonment of this plan 

 of experiment. 



A modification of the Fresnel experiment is that of using 



but a single mirror, which may be rested horizontally on the 



Am. Jour. Sci.— Third Series, Vol. XXXVII, No. 220.— April, 1889. 

 17 



