498 SCIENCE PROGRESS 



thermometer having a bulb of platinum-iridium of over a metre 

 in length and a litre capacity. Chappuis found that the 

 differences between the temperature scales given by hydrogen, 

 air and nitrogen under various initial pressures could be con- 

 sidered as identical for all but the highest class work and that 

 with hydrogen, at all events, the scale thus obtained was probably 

 very close to Lord Kelvin's " absolute " scale. He made 

 elaborate comparisons of the relation subsisting between the 

 readings of a number of specially prepared mercury thermometers 

 made of French hard-glass and the gas-scale. The behaviour of 

 this kind of mercury thermometer had been the subject of long 

 study, and the type was a great advance on any previously 

 used. Over the range o° to loo^ the greatest difference between 

 the French hard-glass mercury and the hydrogen scales occurred 

 at 40°, at which point the mercury thermometer read higher than 

 the hydrogen by "112°. 



Thermometers calibrated at the Sevres laboratory under the 

 direction of Drs. Benoit and Guillaume, provided with a system 

 of correction tables and capable of being read to an accuracy of 

 about "002° C. were sent out with all the national metric standard 

 measures of length and mass, and a number were also issued to 

 science laboratories and kindred institutions. Thus the intro- 

 duction of the International Scale was rendered possible in 

 practice. 



The large Sevres gas-thermometer was not well adapted for 

 work mucft above 100^ C. and investigation showed that at 

 temperatures considerably below 200° C. some action occurred 

 between the bulb and the hydrogen it contained. A second gas- 

 thermometer suitable for much higher temperatures was there- 

 fore erected. This was fitted with bulbs of either hard glass or 

 porcelain. With this instrument it was also found impossible 

 to use hydrogen for accurate observations at the higher tempera- 

 tures, and therefore nitrogen or air was substituted for this 

 gas ; at high ranges these give a temperature scale only differing 

 from that of hydrogen by an amount probably within the limits 

 of experimental error. Observations with this gas-thermometer 

 were made up to about 600^ C. 



Among the most important of the more modern investiga- 

 tions using the gas-thermometer are those of Callendar and 

 of Callendar and Griffiths, who worked with a form of ap- 

 paratus differing from those of Chappuis in several important 



