34fi Mr Willis on the 



Second, — In a tube of the form ACDFB the current exerts an out^ 

 ward pressure, which will act on the membrane DF, accelerating it in its 

 passage from F towards F, and retarding it during its return to F. There 

 is an intermediate position DH, in which the current exerts no pressure 

 on the sides of the tube, and therefore none on the membrane. 



We may easily conceive then that when the current was first admitted 

 into the tube, it might, from the first cause, occasion a superabundant 

 pressure upon the outer surface of DE, which would set it in motion 

 towards DF. By virtue of this motion it would pass the line of equili- 

 brium DH, and would then be soon brought to rest by the resistance 

 arising from the second cause and its own elastic force. From this posi- 

 tion its elasticity and the pressure of the second cause would return it, it 

 would pass DH, be again brought to rest by the resistance of the first 

 cause, again return, and so on ; in this way it woidd oscillate for some 

 time till the friction and resistance of the air, rigidity of the membrane, &c. 

 gradually reducing the extent of its vibrations, would bring it to rest in 

 the position DH. 



Experiment shews however on the contrary that as long as the current 

 is maintained the vibrations continue. This may perhaps be explained by 

 a closer examination of the nature of the force arising from the lateral 

 communication of motion. When the membrane is passing from DF to- 

 wards DG, and has got into the position DH, we suppose this phenomenon 

 to begin; but as the rarefaction it occasions proceeds from a motion gra- 

 dually imparted to the air, it is plain that it takes time to perfect it, and 

 hence at any given point K the rarefaction is not so great when the 

 membrane passes it in going towards DG, as it is when the membrane 

 returns from DG* Hence the retarding force at each point K in going 



* This effect is assisted too by the circumstance that when the membrane is receding 

 from the current, the circumambient air can more easily rush in to supply the deficiency 

 than it can when the membrane is returning. 



