

181)1.] On a Compensated Air Thermometer. 249 



.standard constant pressure, the trouble of reading the barometer will 

 be saved, and the calculations will be considerably simplified. The 

 simplest, and at the same time the most accurate, method of securing a 

 standard constant pressure is to connect the outer limb of the 

 sulphuric acid gauge to a glass bulb filled with air of suitable 

 density, and kept in melting ice, preferably in the same receptacle 

 as the mercury bulb in which the dilatation of the air is measured. 

 The pressure in the thermometric bulb can easily be adjusted to 

 equality with the standard pressure to within 1 or 2 mm. of sulphuric 

 acid, and the small outstanding difference of pressure can be read 

 quickly and accurately by means of a kathetometer microscope. 



I set up an experimental instrument of this kind in October, 1890, 

 and satisfied myself that it was quite possible to read its indications 

 to the thousandth part of a degree C. at ordinary temperatures. This, 

 1 believe, to be a very much higher order of accuracy than has 

 hitherto been thought attainable with an air thermometer. My only 

 remaining difficulty was the slight uncertainty as to the mean tem- 

 perature of the connecting tubes. By the use of the sulphuric acid 

 gauge it is possible to make this correction comparatively small, but 

 it still remains uncertain, and varies slightly with the extent of 

 immersion of the stem of the thermometer. It is also a rather 

 troublesome correction to apply, and complicates all the calculations 

 very considerably. 



It has since occurred to me that this troublesome and uncertain 

 correction may be entirely eliminated, both from the observations and 

 from the calculations, with this particular form of instrument, by 

 making the standard pressure bulb communicate with a set of con- 

 necting tubes equal in volume and similarly situated to those of the 

 thermometric bulb itself. 



The method of determining the correction to be applied for that 

 part of the stem of which the temperature is variable, by means of 

 a similar compensating tube placed in close proximity to it, has 

 occasionally been applied by previous observers. It was first em- 

 ployed by Deville and Troost in 1864, in their experiments on the 

 expansion of porcelain at high temperatures. They connected the 

 compensating tube to a separate manometer, and by observing the 

 pressure or the amount of the air it contained were enabled to 

 eliminate the term representing the effect of the connecting tubes 

 from their equations. 



Other observers have used a similar device, but, so far as I am 

 aware, no one has hitherto noticed that, in the case of the differential 

 air thermometer, the compensation can be rendered automatic, so that 

 changes of temperature of the connecting tubes have no effect on the 

 readings, and need not be taken into account in the calculations. 



The conditions under which the compensation is perfect with the 



