74 SCIENCE PROGRESS. 



It will thus be seen that my independent investigation 

 agreed as to (a) with the conclusion arrived at by Callendar, 

 and was in direct contradiction to the report of the British 

 Association Committee with regard to the Siemens pyro- 

 meter in which the wire was insufficiently protected. Con- 

 clusion (/?) appeared discouraging, as it involved a separate 

 and difficult standardisation of each thermometer before 

 use. 



On comparing my results with Callendar's it was evident 

 that the upper portions of my t - - pt curves departed con- 

 siderably from the parabolic form, and we, therefore, ap- 

 peared to differ in our conclusions. The difference was so 

 marked that I consulted with him, and we decided to make 

 a thermometer similar to mine out of a portion of his origi- 

 nal coil (which he had fortunately preserved) and expose it 

 to sulphur vapour under the same conditions as those pre- 

 vailing during my own experiments. The result, assuming 

 formula (ci) and Callendar's previously found value of S, gave 

 the boiling point of sulphur as 442 '3°C, nearly 6°C. below 

 Regnault's number. It was therefore evident that either 

 Callendar's value of S was wrong, or that the thermometer 

 in my apparatus did not attain the temperature of sulphur 

 vapour, or else that Regnault's value of the boiling point of 

 sulphur was too high. The matter appeared so important 

 that the summer of 1891 was devoted by us to its investiga- 

 tion, and an account of the work will be found in (8). We 

 believe that we then established the following points : (a) 

 that Callendar's value of 8, as obtained during his experi- 

 ments in 1887, was practically correct ; (b) that the bulb of 

 the thermometer in my apparatus did not attain the tem- 

 perature of the sulphur vapour unless suitably screened and 

 that the error due to this cause might amount to 1 '05 C C. ; 

 (c) that Regnault's value of the boiling point of sulphur was 

 too high, the results by our air thermometer determinations 

 being 444*5^. as against Regnault's 448 '4°C. 



The curves obtained by me in 1890 were now re-drawn, 

 substituting our value of the boiling point of sulphur for 

 Regnault's. A study of those curves then led to the follow- 

 ing important conclusions : — 



