372 LECTURE xxxr. 



states are probably often produced in succession in wind instruments blown 

 by the mouth, the air within them being at first cold and damp, and after- 

 wards warm and moist. 



In pure hydrogen gas, the velocity of sound ought, from calculation, to be 

 "" more than threje times as great as in common air, but the difference does not 



appear to have been so great in any experiment hitherto made on, the sounds 

 of pipes in gases of different kinds. For such experiments, the comparative 

 specific gravity of the gas may be most conveniently ascertained by Mr. 

 Leslie's method of observing the time employed in emptying a vessel through 

 a small orifice, by means of the pressure of an equal column of water; ac- 

 cording to the simple theory, the velocities of the gas thus discharged ought 

 to be in the same proportion as the respective velocities with which sounds 

 would be transmitted by them: and if any variation from this proportion 

 were discovered, it must be attributed to the different degrees of heat pro^ 

 duced by condensation in the different fluids. Steam, at the temperature 

 of boiling water, is only one third as heavy as common air; consequently 

 the velocity of sound in steam must be nearly three fourths greater thaa 

 in air. 



It does not appear tlmt any direct experiments have been made on the 

 velocity with wliich an impulse is transmitted through a liquid, although 

 it is well known that liqaids are capable of conveying sound without dififi- 

 culty ; Professor Robison informs us, for example, that he heard the sound 

 of a bell transmitted by water at the distance of ]20{)feet. It is, however, 

 ^ pasy to calculate the velocity with which sound must be propagated in any 



^.•^,yft** liquid of which the compressibility has been measured. Mr. Canton has 

 „^ _. J ascertained that the elasticity of water is about 22 000 times as great as that 



of air; it is, therefore, measured by the height of a column which is in the 

 same proportion to 34 feet, that is 750- thousand feet, and tlie velocity cor- 

 responding to half this height is 4900 feet in a second. In mercury, also, 

 it appears from Mr. Canton's experiments, that the velocity must be nearly 

 the same as in water, in spirit of wine a little smaller. These ex])eriments 

 were made by filling the bulb of a thermometer with water, and observing 

 the effects of placing it in an exhausted receiver, and in condensed air; 

 taking care to avoid changes of tcn>perature, and other sources of error: 



