ON THE MEASUREMENT OF TEMPERATURE, 
141 
Ill the determination of the ice-point the pressure could be measured without 
difficulty to 0'02 millim., the temperature of the dead-space to 0’01° C. ; as has been 
pointed out in Part I. of this paper, the accuracy of this measurement was of the 
order of 1 part in 20,000. The errors are thus confined to the measurement of the 
temperature of the liquid oxygen. In this part of tire experiment the pressure 
could not he measured with certainty to 0‘05 millim., though the error certainly 
never exceeded OH millim., whicli is equivalent to 0’03° C. Such errors as exist 
must be attributed to the unecjual heating of thermometer bulb, and of the bulb 
containing the liquid oxygen. 
7. The Smaller Constant-Volume Thermometers. 
As has already been stated, the use of a thermometer with a large bulli for the 
measurement of low temjieratures is open to serious objections. In investigating the 
whole range of vapour pressure of liquid oxygen and of liquid hydrogen, three 
instruments were used with bulbs of capacities corresponding to 12, 26 and 27 cub. 
centims. These instruments (fig. 3) did not differ in any inqiortant jiarticular from 
the large thermometer already described, but as it was not necessary to determine 
the pressure with a degree of accuracy greater than OT millim., the manometer 
column g, which was 7 millims. in diameter, was not jacketed witli Avater, and the 
distance betAveen the surface of the mercury in g and in the dead-s})ace d Avas 
read directly on a glass scale h, placed behind tlie insti'ument, by means of tAvo 
telescopies; the readings Avere corrected to 0° 0. Tlie tempierature of the dead- 
space d, and of the stem as far as c, Avas measured by means of a thermometer 
placed close beside d. The tempierature of the manometer column Avas taken at 
the mean betAAmen the tempierature of the dead-space and the tempierature indicated 
by a thermometer placed at the leAml of the surface of the mercury in the manometer 
tube g. 
The thermometer bulb d, Avhich had a capacity of 12, 26 or 27 culi. centims. in 
the three thermometers, AA-as sealed to tlie stem as indicated in the figure, so tliat 
if any mercury were accidentally introduced into the ladb, it could be coinjiletely 
removed from it by simpily opiening the stopi-cock f and loAATring the reseiwoir con¬ 
nected AAUth it. 
In making a measurement of tempierature liy means of a constant-A'olume gas 
thermometer, the temperature of the bulb a and of the dead-space d are supiposed 
to be constant; and Avhen the latter is measured liy means' of a mercury thermometer 
placed beside it, the former can be calculated from the piressure exerted by the gas. 
The temperature of the vertical piortion of the stem c is, hoAvever, uncertain, and is 
usually taken to be the mean of the extreme tempieratures. In measuring very Ioav 
temperatures, the density of the gas in the dead-space becomes so Ioav that its mass 
becomes nearly negligible, Avhile the amount of gas in the stem becomes sufficiently 
