Ii92 



SCIENCE. 



[N. S. Vol. XIV. No. 365. 



small number were heard. Tyndall long 

 afterwards, in 1875, explained this curious 

 phenomenon, attributing it to the existence 

 at Villejuif of a heterogeneous atmosphere, 

 caused by the heated air which came from 

 Paris. 



Since the memorable experiments of the 

 Bureau des Longitudes of Paris, various 

 individuals have from time to time under- 

 taken to solve the same problem. Among 

 these may be mentioned Moll and van Beck 

 (at Utrecht), Gregory Woolwich, Stone and 

 Captain Perry in his voyages to the polar 

 regions in 1822, 1824, and Kendall in the 

 Franklin expedition in 1825. In some of 

 these experiments the temperatures ranged 

 from 2° to — 40°, the results obtained ac- 

 cording with the theoretical values. In 

 1823 Stampfer and Myrback conducted ex- 

 periments between two stations in the 

 Tyrol at a difference of level of 1,364 m. ; 

 a similar experiment being undertaken in 

 1844 in Switzerland by Bravais and Martin 

 with a difference of level of 2,079 m. Both 

 experiments confirmed the law that the 

 velocity of sound in air is independent of 

 the pressure. 



In all these experiments the exactness of 

 the results was affected by the ditficulty of 

 estimating accurately the time between the 

 perception of the flash and that of the re- 

 port. Different observers of course gave 

 different estimates. This source of error 

 was first eliminated by Victor Kegnault, 

 who in his long series of researches be- 

 tween 1860 and 1870 made use of the 

 graphical method and electric signals to 

 measure time intervals. Eegnault's ex- 

 periments were conducted in seven tubes 

 (part of the Paris sewers) varying in length 

 from 70 m. to 4,900 m., and of diameters 

 from 0.11 m. to 1.10 m. Experiments were 

 also conducted in the open air by means of 

 reciprocal shots fired from two stations at a 

 distance of 2,445 meters. The number of 

 shots fired was 334. These researches 



of Regnault represent such an enormous 

 amount of work that I shall attempt to give 

 only the principal conclusions deducible 

 from them : 



1. In a cylindrical tube the intensity of 

 the wave varies, diminishing with the dis- 

 tance. The narrower the tube, the more 

 rapid is the diminution. 



2. The velocity of the sound decreases as 

 the intensity diminishes. 



3. The velocity approaches a limiting 

 value, which is higher, the greater the 

 diameter of the tube. The mean value in 

 dry air at 0° in a tube of diameter 1.10 m. 

 is 330.6 m. 



4. The velocity is not affected by the 

 mode of producing the sound wave. 



5. The velocity in a gas is independent 

 of the pressure. 



6. The ratio of the velocities in air and 



any other gas 



'^Jf 



where d is the density 



of the gas, supposed perfect. 



7. The average of the results of all the 

 experiments in the open air is 330.7 m. at 

 0°. 



Beguault was also the first to attempt 

 direct experiments for determining the 

 velocity of musical sounds. In this case, 

 however, the electric signals and the graph- 

 ical recording apparatus were not sensitive 

 enough to respond to the front of the wave, 

 and it became necessary to resort to the ear 

 alone. In these experiments Regnault had 

 the cooperation of Koenig as observer, with 

 whose assistance it was shown that : 



1. A note does not change sensibly 

 when it traverses long distances in tubes 

 of large diameter. 



2. When the sounds are observed by the 

 ear the velocity of high notes appears to be 

 less than that of low ones. This may be 

 due to the more ready response which the 

 tympanum makes in the case of low 

 notes. 



3. In traversing tubes of great length, a 



