at tlie Royal Institution, 1 900-1 907, 



393 



dropping in a known weight of a snbstauce of high atomic weight, 

 such as lead, the specific heat of which varies but slightly with the 

 temperature ; the volume of gas given off is measured in the tube F. 

 The data in the following table illustrate the degree of sensitive- 

 ness of various liquefied gases as calorimetric agents when used in 

 such an apparatus. 



It is obvious that oxygen is about twice as sensitive as ethylene, 

 whilst hydrogen is between five and six times as sensitive as oxygen. 

 In practice, there is an advantage in using liquefied air, as air is the 

 substance surrounding us ; when liquefied hydrogen is used, it is 

 essential to prevent access of air to the apparatus, which may be 

 modified for the purpose in the manner shown in Fig. 13. 



Among the most interesting results obtained with these instru- 

 ments are those relating to carbon — in the forms of graphite and 

 diamond — and ice. In the case of carbon (diamond) previous deter- 

 minations pointed to the disappearance of all heat capacity at about 

 - 00". Sir James Dewar's observations show that although the heat 

 capacity diminishes greatly as the temperature falls, it is still per- 

 ceptible in amount even at - 220'. The same is true of ice, thus : 



Diamond 

 Graphite 

 Ice 



18° to - 78= 



0-0794 



0-1341 

 0-463 



rs° to 



0-0190 

 0-0599 



0-285 



188° to - 25-2^ 



0-0043 

 0-0133 

 0-140 



To judge from the determinations which have been made with 

 various substances, including those of the earlier period relating to 

 metals, it is clear that as the absolute zero is approached the inter- 

 molecular vibrations become more and more damped and perhaps 

 cease altogether at the zero. In the case of metals, electrical 

 conductivity is then at its maximum but at its minimum in the case 

 of non-metals and compound substances. To use a rough analogy in 



