Fitzgerald— ilff^Aoffs of Measuring Densities of Gases. 483 



sharp splinters of glass which projected away from the stem in 

 such a way that the whole thing could be supported on the two 

 sharp points of the glass splinters. By bending the stem judi- 

 ciously, and attaching small pieces of glass to regulate the balance, 

 it is comparatively easy to arrange that the whole thing shall 

 balance on these two sharp points, and perform slow oscillations, 

 proving that its centre of gravity is close below the line joining 

 these points. The mirror should be vertical, and the stem nearly 

 horizontal. If it be now supported by the two points resting in 

 the concave side of a splinter of a fine glass tube, the whole 

 apparatus can bear to be moved about freely without shaking off 

 its supports. If it be now placed so as to reflect a spot of light 

 from the mirror to a vertical scale, or if an image of a vertical scale 

 in its mirror be viewed through a telescope, it is possible to observe 

 the position of the instrument very accurately. If the density of 

 the air in which the apparatus is immersed be changed its position 

 of equilibrium alters, and the spot of light is reflected to a different 

 point of the scale. I have constructed an instrument of this kind, 

 in which the alterations of position on a scale about one meter off 

 are as great as the alterations of the mercury barometer gauge of 

 the air-pump, showing that it is possible to determine pressures 

 with this instrument ' nearly as accurately as with a mercury 

 barometer gauge. This is of course the same as determining the 

 density of the air, for it is the density and not the pressure of the 

 gas present that affects the position of the instrument. I have 

 found that the most satisfactory way of using the instrument is to 

 have some means, as by an attached sprengle or other air-pump by 

 which to alter the pressure of the gas in which the instrument is 

 immersed, until the spot of light or image of the scale is in some 

 standard position, such as it takes up in air at a known temperature 

 and pressure. Under these circumstances the density of the gas 

 is the same as that of this air, and an observation of the tempera- 

 ture and pressure of the gas enables its density at any other 

 temperature and pressure to be calculated. As the whole thing 

 can be put inside a tube four centimeters long, and three in 

 diameter, it could evidently be easily jackitted so as to keep it at 

 a high temperature, and then be used to measure the densities of 

 va,pours, the only precaution required being that the temperature 

 should be sufficieutly high to prevent the gas from condensing as 



