150 Mr. J. T. Bottomley on a Practical 



defined, and high temperatures determined. The thermo- 

 dynamic researches of Sir William Thomson have furnished 

 an absolute thermodynamic definition of temperatures ; and 

 the experimental researches of Dr. Joule and Sir William 

 Thomson have established the practical agreement of Reg- 

 nault's air-thermometer with the thermodynamic scale of 

 temperatures. Lastly, the air-thermometer is the only in- 

 strument known at present, with the exception of a mercurial 

 thermometer which has been compared with an air-ther- 

 mometer, by means of which temperatures higher than, say, 

 150° C. or 200° C. can be determined within 3° C. or 4° C* 



In experimenting on the resistance of platinum and carbon 

 filaments at high temperatures, in connexion with a research 

 on thermal radiation with which I have been engaged, I have 

 used air-thermometers of various forms ; and I have recently 

 been using a constant-volume air-thermometer, which I first 

 described to Professor Gray, of University College, Bangor, 

 just two years ago (January 1886), and partially constructed 

 for him at that time. It is this instrument, greatly improved 

 as to practical details, which I now desire to bring before the 

 Royal Society. 



The best known constant- volume air-thermometer is that of 

 Jolly of Vienna. It is a convenient instrument, and is fairly 

 accurate for moderate temperatures ; but for high tempera- 

 tures a correction, which it is necessary to apply on account 

 of expulsion of air from the heated part of the thermometer, 

 becomes serious, at any rate with the dimensions commonly 

 given to the instrument. It has also some other defects, 

 among which may be mentioned difficulties as to the capillary 

 surfaces of the mercury, want of flexibility or adaptability for 

 various positions, and the proximity of the manometric column 

 to the heated regions. 



The modifications which I have made in the construction 

 of the air-thermometer have a threefold object, one part of 

 which is to improve on the accuracy of the instrument, and 

 reduce to the minimum that is practicable the correction above 

 referred to for the air expelled by heat from the thermometer- 

 bulb or air-reservoir. A second object is to increase the range 

 of the instrument by giving it a form in which the hard 

 Bohemian glass can be used in the construction of the part to 

 be heated. The third object is to make that part of the ther- 

 mometer which is to be heated, and which, in the use of the 



* Mr. H. L. Callendar has proposed to use the resistance of platinum 

 for thermometric purposes ; but in this case also the final standard of 

 reference is the air-thermometer. 



