Practical Thermometric Standard. 541 



can be performed at leisure, whereas the pressure readings 

 must be taken at the time of the observation, and it is neces- 

 sary to observe and adjust two columns simultaneously. 

 (4) The small differences and variations of pressure between 

 the thermometric bulb and the standard pressure-bulb, which 

 need never exceed one or two millimetres, can be observed 

 by means of a cathetometer-microscope with a sulphuric-acid 

 gauge, which permits a higher order of accuracy than a 

 mercury manometer. (5) The troublesome and uncertain 

 correction for the connecting-tubes may be automatically 

 eliminated in a very simple manner by connecting to the 

 standard pressure-bulb a set of compensating tubes of equal 

 volume and similarly situated to the connecting tubes of 

 the thermometric bulb. (The same apparatus can be em- 

 ployed as a constant-volume thermometer, but the manipu- 

 lation is not so simple in that case.) For these and similar 

 reasons, the constant-pressure thermometer appeared to be 

 the most accurate and convenient type of instrument, espe- 

 cially for high-temperature work ; and I therefore adopted it 

 as the standard of comparison in my investigations. The 

 compensated type of instrument above referred to, which 

 permits the attainment of a much higher order of accuracy 

 than my original apparatus, was exhibited at a meeting 

 of the Royal Society in October 1891, and is described 

 in the ' Proceedings/ vol. 1. p. 247. The deviation of the 

 scale of the constant-pressure thermometer from the theoretical 

 scale is a little larger than that of the constant-volume type, 

 but is less uncertain, being more nearly the same on either 

 formula. The correction would in any case be practically 

 negligible if helium were used for the standard gas. The 

 essential point is to adopt for the standard that type of 

 instrument which is capable of giving the greatest accu- 

 racy of reading, and of avoiding the most uncertain cor- 

 rections. 



The advantages of helium over hydrogen for the definition 

 of the theoretical standard scale are such as to merit closer 

 attention. Although denser than hydrogen, it appears to be 

 certainly more difficult to liquefy. It shares with argon the 

 valuable property of extreme inertness, and would not attack 

 mercury, or reduce glass or porcelain. It does not diffuse 

 through platinum, and could therefore be safely used in 

 metallic bulbs at high temperatures. Being monatomic, it 

 would presumably be less liable to dissociation. I hope 

 shortly to be able, with the assistance of Mr. Eumorfopoulos, 

 to investigate the thermometric properties of both argon and 

 helium, prepared and purified by Professors Ramsay and 



