[ 328 ] 



XXXVI. Experiments on the Velocity of Sound in Air. 

 By D. J. Blaikley*. 



[Plate VII.] 



1HAVE to bring before you this afternoon the results of 

 a few experiments carried out in continuation of the 

 series brought to your notice last November. Before con- 

 sidering the velocities obtained, however, it may be well to 

 show by experiment the reason for the doubt in my mind as 

 to the absolute reliability to be placed upon records obtained 

 from membranes as used by Regnault and Le Roux; and also 

 to explain a little more fully my reason for discarding the use 

 of organ-pipes, speaking under a considerable pressure and 

 with good musical tone, as used by Dulong. 



If we take a short tube, say 5 inches long, with a short 

 inner sliding-piece for adjustment of length, and close one end 

 with a diaphragm of gold-beater's skin, against which a bead 

 hung by silk falls, we have a very sensitive resonator for the 

 pitch of about 512 v. Jf this tube were closed by a rigid 

 material instead of by the skin, its length to give the maxi- 

 mum resonance to the 512 v. fork would be about 6 J inches; 

 but on drawing the slide to make the tube of this length there 

 is no agitation of the bead : the maximum agitation of the 

 bead, at the present tension of the membrane, is found to be 

 when the tube is about 5 inches long. This is its most sen- 

 sitive position ; but when the fork is vibrating strongly the 

 membrane will record its action, although the length of the 

 tube, from mouth to membrane, is somewhat greater, or less, 

 than is necessary for best effect. Applying this to the case 

 of a membrane stretched across a tube along which a single 

 wave is travelling, as in Regnault's experiments, it would 

 appear that the last possible record of the enfeebled wave 

 would be obtained from its point of maximum condensation ; 

 but a record of the wave in the early part of its course, when 

 it is strong, would be obtained from a point in the wave far 

 short of its maximum condensation ; so that the length mea- 

 sured by means of the membranes, as travelled by the wave, 

 does not necessarily give the length passed over by a given 

 point in the wave in the measured time. The error, small as 

 it would be when distributed over a great length, would become 

 more appreciable as the tube-length became shorter, and it 

 would in every case lead to an under-estimation of the velocity. 



* Read before the Physical Society, June 14, 1884, in continuation of 

 paper read November 10, 1883 (Phil. Mag. 1883, xvi. p. 447). 



