RICHARDS AND MARK. — THERMAL EXPANSION OF HYDROGEN. 121 



below, was always adjusted so that the dividing line between the black 

 and white surfaces was exactly tangent to the mercury meniscus, as 

 viewed by the carefully levelled telescope. In this way an exceedingly 

 sharp definition and reading may be obtained, as is well known. Inade- 

 quate precautions concerning this point undoubtedly vitiated to some 

 extent the preliminary determinations, but the readings obtained in the 

 determinations discussed below are believed to have been very exact. 



A few preliminary trials with the new apparatus under its improved 

 conditions served to show that the accuracy of the determinations of the 

 pressure of the gas, as observed by means of the Lord Rayleigh barometer, 

 had been increased nearly tenfold. Since this measurement now over- 

 stepped in accuracy all the others involved in the experiment, it became 

 necessary o improve all the others to an equal degree of precision. 



It is obvious that of the three variables — volume, pressure, and tem- 

 perature — temperature alone requires reference to some generally 

 accepted standard in order that the results of this investigation may be 

 compared with those previously obtained and in order that they may 

 readily be made use of in other calculations. Since changes in volume 

 and pressure are purely relative, the use of the same standards through- 

 out is all that is required ; but for the sake of convenience and consistency 

 the international standards were used throughout. 



The Measurements of Volume. 



A volume is most conveniently measured by weighing the water or 

 mercury which fills it, and making the appropriate corrections for tem- 

 perature and pressure. For this reason the standard of volume, the 

 liter, is fixed as the space occupied by a kilogram of water at 4° cor- 

 rected for the weight of the displaced air, but not for the contraction of 

 0.000043 produced by the pressure of the atmosphere. This standard is 

 used throughout the present work. 



It becomes now a matter of importance to decide upon the liquid to be 

 used in the actual measurement of the volume. Into this decision several 

 factors enter. 



In the first place, the internal and external pressures upon the rather 

 thin large bulb must be during the measurement as nearly as possible 

 those under which it is actually used. For this reason mercury was 

 rejected as the standard liquid in this case, because its weight would 

 inevitably distort the bulb. Another objection to the use of mercury is 

 the fact that in order thoroughly to expel the air when a large bulb is 

 being filled, the mercury must be boiled in the bulb. This involves the 



