586 Professor Sir James Deirar [March 25. 



which can be bent so as to tilt a few of tbe small pieces of substance 

 •contained in C into the calorimeter, and which afterwards assumes a 

 position of rest somewhat like that in the diagram. 



With care one can manage to tilt a single piece at a time from (.' 

 into B, but an improved form of this receptacle is shown at CD'. 

 In it, P is a wire movable through the cork Q, fitted into the mouth 

 of the test-tube C, attached by a branch through the stiff rubbei- 

 tube D' to the end of G, as before. At the end of the wire P is a 

 hook, by which one piece of the substance at a time can be pulled u]t 

 and dropped into B'. When no other arrangements are made, the 

 portions of matter experimented on are at the temperature of the 

 room ; but when lower temperatures are required initially, a vacuum 

 vessel H containing either soUd carbonic acid, liquid ethylene, ah:, or 

 other gas, can be placed so as to envelop the test-tube C or Q> ; or if 

 higher temperatures are required, the surrounding vessel may \yQ filled 

 ^with the vapour of water or other liquids. 



Much study and handling of the instrument have brought out the 

 following matters as essential to high efficiency. I have already 

 pointed out that in the neighbourhood of - 180° C. to - 200° C. liquid 

 SLIT is the preferable substance to use, while liquid hydrogen enables 

 observations to be made as low as —250° 0. The value of the 

 •vacuum of the calorimeter itself is much enhanced by making it a 

 mercury vacuum ; and further hy having, previous to use, a good 

 ^mercury deposit over its surface. This is attained by putting some 

 liquid air into the calorimeter B and leaving it to stjind for some 

 time. When a quantity of liquid air has been undergoing volatihsa- 

 tion for a time, as the nitrogen evaporates more quickly than the 

 •oxygen, the boiling point rises slightly. Two points require attention 

 in consequence of this ; first, the maintenance of a constant tempera- 

 ture of the liquid air during any one series of experiments ; next, the 

 prevention of a tendency for the calorimeter B to " suck back " some 

 of the already volatilised gas. Hence the exterior vessel A should be 

 :filled with a large quantity^ — some two litres — of old liquid air, con- 

 taining a high percentage of oxygen, and the calorimeter itself should 

 be filled with some of the same fluid. This will maintain very closely 

 the constant temperature required. When any "sucking back" 

 seems to be taking place, the calorimeter should be emptied and tilled 

 anew from the larger flask A. The tube between the calorimeter and 

 the gas receiver should be of the size of wide quill tubing, and its 

 lower end should be so arranged below the surface of the liquid in 

 the collecting vessel, as to give no resultant pressure. With such 

 precautions, results may easily be obtained correct to within 

 2 per cent. 



The instrument having been set up and filled with liquid air, 

 according to the above directions and precautions, an experiment is 

 conducted by tilting up the little test-tube, previously cooled or 

 heated, thereby dropping into the calorimeter a portion of any snb- 



