496 



an absolute zero-point pressure according to tiie tiieory of quanta 

 ought not to be introduced. 



In both respects the difficulty migiit be sufficienllj' avoided by 

 simply taking the melting-point pressure of the tliermomeler sufficiently 

 small, but in that case, as already ()ointed out, the thermal molecular 

 pressure begins to give difficulties which ultimately exceed all the 



others. In fact this pressure depends upon the ratio — , where h 



is the radius of the capillary and ^. the mean free path. 



Whereas we know the condition of the pressure-equilibrium 

 between the bulb of the thermometer and the manometer, when 

 the temperatures of both are given for the two extreme cases 



2R \ 2R , . • J- 



— =r and — = CO, this is no longer the case tor intermediate 



values of this fraction. In ordinary gas-thermometers with a melt- 

 ing-point pressure of about the normal atmospheric pressure, 



■2R 

 the condition — = oo is very nearly satisfied and the pressures p^ 



at the top and j)„ at the bottom of the capillary, where the tem- 

 peratures are 2\ (normalj and 1\ (to be measured) respectively, 

 may be taken as equal. As we shall see, this is by no means allowed 

 when temperatures are to be measured at which the vapour-pressure 

 of helium is no more than a few millimeters. In thermometers which 

 are adapted to this object considerable corrections have to be dealt 

 with, as will appear in the measurements to be discussed in this 

 paper, indeed the question naturally arises, whether in this case it 

 is not preferable in the temperature-measurement to start from the 



1R 1 . V^T\ p, 

 condition ot eiiuilibrium tor — := — viz. ^ = — . 



X CO \/2\ p, 



3. Di'serlption of the tivo thermometers. 



The thermometer with mercury-manometer (fig. 1) was the improved 

 form of tiiat in Comm. 119 as described in Comm. 124/;. The bulb 

 7'A, had about three times greater capacity, 23.95 cc. and the 

 capillary consisted of three parts, the first starting from below 

 C„ — Cij 15.3 cms. of 0.0362 cm. radius, the next Ci — Cc 9.80 cms. 

 of 0.0783 cm., the third C — Cd 22.59 cms. of 0.0947 cm. (the 

 upper 5,25 cm. having 0,090 cm, radius). To the top of the glas-s 

 capillary (being another part of 5,55 cm. of 0,090 cm. radius) was 

 soldered (entering o^'er this same length) a copper capillary of 

 1.2 mm. diameter, which was connected to the mercury manometer. 



