1902.] Expansion of Ice, etc., at Low Temperatures. 239 



this form salts and other materials similarly treated are especially 

 adapted for accurate specific gravity determinations. After such 

 treatment it was found that all the mechanically attached water was 

 got rid of in the case of hydrated salts, and also in such as did not 

 combine with water. In order to get cylindrical blocks of the salts 

 showing no porosity the presence of water, or rather the saturated salt 

 solution, was found to be essential during the application of pressure. 

 In the same way it was found to be an advantage in compressing such 

 a substance as solid carbonic acid, to moisten it with a fluid like ether 

 before applying the hydraulic pressure. 



Recalling the work of Playfair and Joule,* which originated in a 

 suggestion of Dalton's that the volume of a hydrated salt in solution 

 was simply the volume of the water of crystallisation, ice and some 

 hydrated salts were selected, as well as some other bodies whose 

 coefficients of expansion they had determined. Substances of special 

 interest were included in the list, like mercury, sulphur, iodine, and 

 solid carbonic acid, the latter being particularly important as an 

 example of a solidified gas. 



In the further conduct of an experiment the observations made on 

 a substance were three, namely : (A) the weight in grammes of the sub- 

 stance and suspending platinum wire, either in air of about 17° C. tem- 

 perature or in the gaseous air in the flask containing the liquid air ; 

 (B) the weight in grammes of the body and wire when immersed in the 

 liquid air ; and (C) the weight in grammes of the suspending platinum 

 wire in ordinary (17°) air. This wire was bent into the form of a 

 small cage, to hold the compressed cylinder or portion of the substance 

 to be examined. (C) was measured once for all for each cage, and in 

 the different cases varied from about 0*2 to - 3 gramme. The relative 

 portion of wire not immersed in liquid air was so small that no appre- 

 ciable error will be made in (B) by considering that the whole of the 

 wire was immersed. Further, no allowance is made for the (17°) air 

 displaced by the weights themselves. The observed weights A, B, C 

 are therefore connected with the weights and specific gravities of the 

 substances concerned by the very approximate relations 



A = W + w--~-d d, 



o s 



W w 

 B = W + w--^B--D, 

 ^ s 



= w — d. 



s 



* " Eesearches on Atomic Yolume and Specific Gravity," ' Chem. Soc. Jour., 

 vol. 1,121. 



