246 



On the Specific Heat of Metals, &c. 



It seems probable that at absolute zero the values of the products of 

 specific heat, multiplied by atomic weight, would be identical, or differ 

 only by the very small amount due to experimental error.* 



Appendix. 



The following calculations have been made by E. R. Verity, Assoc. 

 R.C.S., and H. L. Mann, Assoc. RC.S. :— 



If cobalt and nickel were in the states of perfect gases, their specific 

 heats at constant volume would be 



== 0-04123 for cobalt. 

 ko = 0-04145 for nickel. 



This is on the assumption that ko multiplied by the atomic weight 

 is the same as for hydrogen. It is not unreasonable to assume that in no 

 state can the substance have a smaller specific heat than k^. It is easy 

 to show that for any of the metals the specific heat at atmospheric 

 pressure K is not more than 2 per cent, different from the specific heat 

 under any other condition, such as great hydrostatic pressure, to keep 

 its volume constant. Our ignorance of the molecular state of a solid is 

 so great, that we cannot even speculate on how it is that when 1 gramme 

 of cobalt (we have much the same figures for other metals) at 50° C. 

 rises in temperature to 51° C, whether it is allowed to expand 

 freely or is subjected to great hydrostatic pressure which prevents 

 expansion, the energy 0-041 enters it as what may be called the 

 real sensible heat, and the energy 0-062 enters it as some kind of 

 energy of ■ disgregation, necessary because of change of tempera- 

 ture, and having nothing to do with change of volume or pressure. 

 The facts are not explainable by assuming that the atomic weights 

 are wrong, because, as we see from cobalt and nickel, K approaches 

 the value ko at low temperatures. Indeed, if the following formula 

 is correctly deduced from the above measurements, we may say that 

 in the solid state the product of the atomic weight and specific heat 

 may be anything between what it is for hydrogen and 2-7 times this 

 amount. The formula is 



where t is the absolute temperature. 



* Note added March 3. — Further experiments made, since the date of communi- 

 tion, upon the metals silver, copper, iron, and aluminium show, however, that this 

 suggestion will not be realised. The mean specific heat of silver between 15° and 

 182-4°, for example, is 0-0519, while from 100° to 15° it is 0*0558. The decrease of 

 specific heat at the lower temperature is therefore much less than in the case of 

 cobalt and nickel. 



