Propagation of Waves in Gaseous Media. 167 



mean limiting velocities in the three tubes — that is to say, those 

 velocities which correspond to a wave sufficiently enfeebled from 

 its commencement as no longer to affect the membranes. 

 These limiting velocities were found to be : — 



metre. metres. metres, 



diameter } ' 108 V' =326'66 Dist. traversed =4055 -9 



0-300 V' =328'96 „ =15240-0 



1-100 V' = 330-52 „ =19851-3. 



In these experiments the wave was produced by the same 

 charge of powder. The membranes are the same, and have 

 therefore the same sensitiveness ; in other words, they ought to 

 cease to indicate in the three tubes when the waves have acquired 

 the same feebleness. If the weakening of the wave arose entirely 

 from the loss of vis viva through the walls of the tube, the mean 

 limiting velocity should be the same in the three tubes, because 

 the wave has the same intensity in all three cases, both at its 

 origin and at the moment when it makes its last impression upon 

 the membrane. These limiting velocities being, on the contrary, 

 very different, we must conclude that the walls of the tube exer- 

 cise upon the enclosed air another action besides the action which 

 we have just pointed out, which considerably diminishes its elas- 

 ticity without sensibly changing its density. In consequence of 

 this the velocity of propagation of a wave of the same intensity in 

 straight tubes is less according as the section of the tube is less. 



It is probable that the nature of the wall of the tube (its more 

 or less perfect polish) exerts an influence upon this phenomenon. 

 I may mention a fact which proves it. In the Paris sewers of 

 great diameter the workmen are usually signalled to by sound of 

 trumpet. It is observed that the signals are carried incompa- 

 rably further in galleries whose walls are covered with smooth 

 cement than in those lined with rough stonework. 



In order that this action of the walls upon the elasticity of the 

 gaseous medium may be absolutely nothing, the diameters of the 

 tubes must be infinite ; in other words, the propagation of the 

 sound must take place in the open air. But the action of the 

 walls must be very small in the great tube of 1'10 metre diame- 

 ter. I have assumed that it is nothing; and I have concluded 

 from my numerous and very concordant experiments, that the 

 mean rate of propagation in dry air at zero of a wave produced 

 by the discharge of a pistol and reckoned from the muzzle of the 

 pistol to where it becomes so weak as to cease to affect the most 

 sensitive membranes, is 



V' = 330-6 metres. 



I have also endeavoured to determine the velocity of the weak- 



