the Thermal Unit. 433 



Part I. 

 The Capacity for Heat of Water at Different Temperatures. 



From the time of Black to the present day no practical 

 thermal standard, except the capacity for heat of water, has 

 been adopted. Its capacity at 0°, 4°, 15°, and its mean 

 capacity from 0° to 100°, have been variously selected as 

 standards by different observers ; and since our knowledge 

 of the comparative value of these standards is vague, to say 

 the least of it, comparison of the results of the different inves- 

 tigators becomes impossible. A consequent evil is the growth 

 of a specious accuracy : books of reference give the changes 

 in the specific heat of various bodies, as for example mer- 

 cury, to 4 or 5 significant figures*, whereas our knowledge 

 of the standard in which they are expressed is far less exact. 

 I give the following illustration as an example. Last year I 

 communicated to this Section the results of an inquiry into 

 the changes in the specific heat of Aniline due to changes in 

 temperature. The conclusions there given depended on a 

 method which, whatever may have been its merits, was at all 

 events an absolute one in the sense that the results were 

 independent of the changes in the thermal properties of water. 

 In the following autumn Mr. C. Grreen made a series of careful 

 determinations of the density of aniline over the same range 

 of temperature. An interesting and unexpected relation then 

 became evident, viz., that the " volume heat" of aniline over 

 the temperature- range 20° to 50° might be regarded as con- 

 stant f. I was naturally anxious to see if this relation held 

 in the case of any other body ; but I have been unable, in 

 spite of diligent search, to find any determinations of specific 

 heat which are independent of some assumption with regard 

 to the changes in the capacity for heat of water. This 

 example is sufficient to show how inquiry is checked by our 

 present uncertainties. 



It appears to me that we have been singularly unfortunate 

 in the selection of the capacity for heat of water as our 

 standard. At the same time our whole system of thermal 

 measurements has become so intimately connected with the 

 heat-capacity of that liquid that it would require nothing 

 short of a revolution to abolish its use as a standard : the 

 laboratory might consent but the workshop would rebel. 

 Under such circumstances we must make the best of the 



* See Landolt. 



t Proc. Camb. Phil. Soc, 1895. 



