42 SCIENCE PROGRESS 



hydrogen has been up to quite recently the standard, as its very 

 low critical point (30 K.) makes the corrections quite small until 

 temperatures only obtained by liquid hydrogen are reached. 

 Now that helium is available with a critical point of about 5*1 K. 

 and a small very simple molecule, which makes divergences 

 extremely small, there is no doubt that it is the most suitable 

 gas for low temperatures, as the corrections are even small at 

 the temperature of solid hydrogen, the lowest temperature 

 obtainable without the aid of helium itself. Thus quite shortly 

 we may expect standard gas thermometry to be confined to 

 helium thermometers up to 100 9 C. and argon thermometers from 

 about o° C. upwards, there being a region of 150 to 200° over 

 which the two scales can be compared. However, for practical 

 purposes the hydrogen and nitrogen scales will continue to 

 be used, and, if the absolute corrections are known, readings 

 made with them are as accurate as if made with a standard 

 thermometer with the same care. 



The evaluation of the absolute scale is due to Lord Kelvin in 

 1847 from the theory of heat engines. Heat is taken in at a 

 temperature T and given out at a temperature T ! ,and the theory 

 says that the amounts of heat are proportional to the absolute 

 temperatures with a perfect reversible engine. As the most 

 perfect working substance is a perfect gas and as certain actual 

 gases approach very nearly to the standard of perfection, they 

 are clearly the most suitable substances to determine the value 

 of the difference between the Kelvin and centigrade scales. 



It is rather remarkable that the original value of — 273-1 C, 

 which was derived from gases whose properties were observed at 

 considerable distances from the absolute zero, should be almost 

 exactly the value which the most recent and careful determina- 

 tions would indicate. From time to time lengthy papers have 

 been published making estimations of the absolute zero derived 

 from measurements on the Joule-Kelvin effect which are known 

 not to be very accurate. It is not to be wondered at that there 

 should have been a considerable discrepancy between the results 

 obtained, but they at least all indicated that the value of the 

 Kelvin zero on the centigrade scale would be more than —273 

 and less than —273*5. Much more accurate information is, 

 however, obtained from a strict investigation of accurate 

 isothermals, and it will only be necessary to consider the results 

 furnished by, say, nitrogen, hydrogen, and helium with critical 



