( 481 ) 



LXV.— ON SOME METHODS OF MEASURING THE DENSITIES 

 OF GASES. By GEORGE FRANCIS FITZGERALD, 

 M.A., F.R.S., &c. 



[Read, June 15, 1885.] 



The following methods depend on determining the amount by 

 which a body is buojed up when immersed in the gas. 



The first arrangement consists of a large thin glass ball sealed 

 up and suspended bj a bifilar suspension. The period of oscillation 

 of such a ball round a vertical axis depends on the distance apart 

 of the suspending fibres, on its moment of inertia, and on the 

 weight of the ball. This latter depends on the density of the gas 

 in which it is immersed, and by observing the time of oscillation 

 of the ball it is possible to determine the density of the gas. In 

 order to do this conveniently it is well to make observations of the 

 period of oscillation in two gases of known densities, as for in- 

 stance in air at the ordinary atmospheric pressure, and at a very 

 much lower pressure obtained by exhausting by an air pump the 

 vessel in which the ball is suspended. As the relation connecting 

 the number of oscillations per second (iV) of the ball with the 

 density (g) of the gas in which it is immersed is of the form 



two observations serve to determine A and B, and then any other 

 density can be determined in terms of iV, the number of oscilla- 

 tions performed by the ball. From experiments I made I believe 

 it is easy to arrange that the ball shall perform one hundred oscilla- 

 tions without their amplitude being either initially too large or 

 finally too small ; and if by making the length of the suspending 

 threads long (this is better than making their distance apart small, 

 as this latter introduces a correction depending on their torsion), 

 the period of oscillation be considerable, it is evidently possible to 

 determine this period very accurately. It is better to determine 

 the period of this oscillation round a vertical axis than its period of 

 oscillation as a pendulum, because, without a very long suspension, 

 this latter would be more rapid, and would besides be more damped 



2T2 



