Practical Thermometric Standard. 539 



and porcelain. My own original observations of 1886 were 

 made with a constant-pressure air-thermometer of hard-glass ; 

 and in 1890, in conjunction with Mr. Griffiths, the same 

 apparatus was used with a bulb of soda-glass. Hydrogen 

 was originally selected at the International Bureau as the 

 standard gas, because it is so perfect at ordinary tempe- 

 ratures that no difference can be detected between the scales 

 of the constant-pressure and constant-volume hydrogen- 

 thermometers. It was found necessary, however, to adopt 

 nitrogen for the higher temperatures, in the comparisons with 

 the platinum-thermometer, owing to the action of the hydrogen 

 on the walls of the reservoir. 



According to the equation of Van der Waals, the pressure 

 of any gas at constant volume should be a linear function of 

 the temperature. If this were true, it would be a very strong 

 argument for the adoption of the constant-volume type of 

 gas-thermometer as the standard. It would appear, how r ever, 

 from the equation of Clausius, which is here supported by 

 the experiments of Chappuis, that there are quite mea- 

 surable differences between the scales of different gases at 

 constant volume and between the scales furnished by the 

 game gas at different initial pressures. The other argument 

 commonly advanced in favour of the constant-volume ther- 

 mometer, that it gives a scale of equal parts, and permits 

 greater relative accuracy of reading at the higher points of 

 the scale, is more apparent than real. It is not possible to 

 make the scale accurately equal owing to the corrections for 

 stem-exposure and for the expansion of the bulb. It is no 

 advantage to be able to read the pressure equally to the 

 hundredth of a degree at 0° and at 500°, or 1000°, because 

 an error of a hundredth in the fundamental interval means 

 an error of several hundredths at the higher points, and all 

 the other sources of error are greatly increased. The large 

 increase of pressure at high temperatures is really a disad- 

 vantage, because it becomes necessary to diminish the initial 

 pressure, in order to keep the pressure within range and to 

 avoid straining the bulb. Lowering the initial pressure means 

 diminishing the accuracy of the fundamental interval; it also 

 means a discontinuity in the scale of the thermometer, if 

 different initial pressures give different scales. Among minor 

 difficulties attending the use of the constant-volume thermo- 

 meter, may be mentioned the impossibility of determining the 

 effect of change of pressure on the volume of the bulb at 

 high temperatures, and the difficulty of satisfactorily deter- 

 mining the mean temperature of the mercury columns. 

 These are questions of considerable importance in accurate 

 work and at high temperatures. 



