The vacuum calorimeter second form

as has already been remarked, this temperature coefficient varies at low temperatures in a manner not very favourable to accurate measurements there is a fairly sharp maximum in the neighbourhood of 8o° abs., and it falls very rapidly below 40° abs. In spite of the most careful calibration, which presents difficulties in the region from 25° to 45° abs. on account of the lack of fixed points, there are still uncertainties of a small amount. There is a further inconvenience in the fact that the platinum wire has its resistance considerably changed by the introduction of the electrical energy, and that, in consequence, the accurate determination of the latter, though quite possible, is nevertheless somewhat troublesome. 

These disadvantages may all be avoided, and are in fact considerably reduced by the use of lead instead of platinum wire, but a simpler procedure is still desirable. The problem of establishing the variation of specific heat for a large number of substances as completely as possible is extremely important, both in the light of any future theory of the solid state, which will apparently be developed from a consideration of low-temperature properties, and also in view of the applications of our Heat Theorem and the study of chemical affinity. 

Little had been achieved in this connection when I set myself the task of examining a few substances with the greatest possible accuracy but the arbitrariness which would have existed in their selection would have weighed 

To the above-mentioned lead block was soldered a copper sheath which surrounded the calorimeter the effect of this was that the latter was in a space having an absolutely uniform external temperature. This protection might have been supposed superfluous, as neither radiation nor conduction is appreciable when the vacuum is good and the temperature low but it was found that if it were omitted the deflections of the galvanometer connected to 

The vacuum given by this excellent pump was found, however, to be insufficient even when it was run continuously. This is explained in part by the known fact that hydrogen is not very well exhausted by this pump hydrogen was present in the apparatus of necessity, having been used to cool the calorimeter before the experiment proper. Washing out with helium was of some benefit, and neon would probably have worked even better—other gases are, of course, of no use at temperatures of 150 to 20° abs. on account of their small vapour pressure. 

---