450 Mr. E. S. Willows on the Variation of the 



small thickness used any irregularity in the supply of steam 

 was shown by an irregular rise in temperature ; in addition, 

 the temperature that the thermometer shows is only the mean 

 of the temperature over the whole bulb, and not that of a 

 small part of the amalgam. 



To get rid of these uncertainties the amalgam was used in 

 the form of a sphere of 1*5 cm. diameter. A radial hole was 

 bored to the centre, and in this was placed a copper-iron 

 junction. The space between the couple and the sphere 

 could be packed with asbestos. This packing stopped currents 

 of air from circulating round the junction, and hence it gave 

 the temperature at the centre of the sphere and not that of 

 the surrounding air merely, while any small irregularities in 

 the supply of steam had their effect lessened by conduction 

 from the surface to the centre of the sphere. For a rise in 

 temperature of GO the galvanometer gave 450 divisions 

 deflexion. The steam being turned on the deflexion of the 

 galvanometer was taken every 15 seconds. After the tem- 

 perature had reached about 90° the current of steam was 

 replaced by one of cold water, galvanometer deflexions being 

 noted as before. When the temperature was being raised 

 there was a point at 36° at which it remained stationary for 

 two seconds. This temperature was the same for all ihe zinc 

 amalgams tried, and at this point, therefore, there is an 

 absorption of heat. The galvanometer was stationary again 

 at a temperature very near to that corresponding to Q 

 (fig. 3) ; hence there is an absorption of heat at the tem- 

 perature at which the resistance suddenly falls. There was 

 an evolution of heat at the temperature corresponding to R 

 (fig. 3), affording further proof that the increase in resistance 

 at this point is not due to a sticking to the sides of the tube 

 as this would not involve thermal changes. 



The amalgam does not appear to have melted at all below 

 100°, but it becomes considerably softer at this temperature. 

 If we assume that its mean specific heat does not alter between 

 20° and 100° (which will be nearly true if there is no melting 

 in the mass surrounding the crystals of zinc), and that the 

 emission or absorption of heat is proportional to the difference 

 of temperature between the body and the inclosure ; then from 

 the observed times during which the galvanometer was 

 stationary, we can calculate the relative amounts of heat 

 absorbed or evolved, and we find that the heat emitted at B, 

 (fig. 3) is less than that absorbed jointly at P and Q. The 

 thermo-couple was standardized before and after the experi- 

 ment, and showed no alteration. 



