534 



Prof. J. Dewar. On the [June 17 y 



bottle. One division on the arbitrary scale represented about of a 

 degree C. Two thermo- junctions, made of iron-copper wires, insu- 

 lated by a covering of marine glue, the junctions themselves being 

 covered with a thin layer of gutta-percha dissolved in benzol, were 

 employed in the experiments. 



One junction was placed in the iron bottle, to record the effect of 

 pressure under different circumstances, the other was kept in a beaker 

 outside at the constant temperature of melting ice. 



Series I. 



(Fig. 3 shows the general arrangement of bottle and thermo- junc- 

 tions.) 



Fig. 3. 



A junction, fixed in a brass flange with marine glue, was frozen in 

 a test-tube containing boiled water, and placed in the iron bottle of 

 the Cailletet apparatus, surrounded with water at 0° C, the bottle 

 itself being packed round with ice. 



The pressure was increased by 25 atmospheres at a time. The 

 galvanometer showed a deflection immediately after the pressure was 

 applied, rapidly coming to rest, and it remained perfectly steady at 

 the lower temperature as long as the pressure was maintained constant. 



The following table gives the results of one of many series of ex- 

 periments, which agreed very nearly. Pressures were recorded by a 



