246 Professor Sir James Dewar [Jan. 23, 



causes it to be driven from n to rn, and the mercury in the gauge 

 rises to p in the left Umb, owing to the diminished pressure, and is 

 depressed to o in the right limb. The difference in pressure is pro- 

 portional to the number of revolutions of the cylinder and the 

 internal friction of the gas. In practice the grooves are cut in the 

 cylinder, and the tongues, which project into the grooves, are fixed on 

 the casing. The speed of rotation of the central shaft carrying the 

 cylinder is about 8000 revolutions per minute. 



The efficiency of this pump is seen when it is used direct on a 

 large discharge tube 2 m. long and 6 cm. in diameter, provided with 

 electrodes in the manner of a De la Rue tube. In about a minute 

 from starting at atmospheric pressure all stages of discharge at low 

 pressures are traversed, and finally only a flickering phosphorescence 

 is seen. A further test on liquid oxygen shows the manner in which 

 a rapid evolution of gas is overcome. The melting pressure of oxygen 

 is too low to be realised by usual exhaust methods, as was shown 

 in the Discourse of January 20, 1911, when the solidification of 

 oxygen was shown to be possible only by the exhaust produced by 

 charcoal and liquid air. Formerly it was necessary to employ cooling 

 by liquid hydrogen to obtain oxygen in the solid state. If, however, 

 the fitting on the exhaust nozzle of the Gaede pump be directly 

 connected to a vacuum vessel containing some 10 c.c. of liquid 

 oxygen (isolated as usual by being surrounded with an outer vessel of 

 liquid air), three minutes' working of the pump results in the oxygen 

 beincr transformed into solid state. 



Low Te^iperature Calori3ietry. 



The application of vacuum tube methods has resulted in bringing 

 the law of Dulong and Petit into relation with the other laws 

 governing the behaviour of the atoms. Working at ordinary 

 laboratory temperatures Dulong and Petit had shown that the 

 atomic heats for the various elements (got by multiplying the specific 

 heat and atomic weight) was a constant of an approximate value six. 

 For working at 50'^ Absolute (-228° C), a liquid hydrogen calori- 

 meter has been constructed in which the specific heat of the various 

 elements can be easily determined in terms of the volume of gas 

 produced in cooling the substances. Under such conditions very 

 different results from those previously obtained have been deduced. 

 The atomic heats exhibit the same rough periodicity of function 

 with atomic weights as is shown by other properties of the elements. 

 A double curve demonstrates the complete paralleHsm between the 

 atomic volumes and the atomic heats of the principal elements 

 (see Fig. 7). In connection with speculations as to the constitution 

 of the atoms, physicists had been putting different interpretations 



