( 376 ) 



1° 

 may without difficulty be effected accurate to — if the requisite 



precautions are taken. Though it is not certain that the determina- 

 tions in liquid hydrogen of the last series come up to this accuracy, 

 as there a systematic error caused by the filling may show its influ- 

 ence, which does not yet make its appearance at — 217°, yet it 

 lies to hand to suppose, that, at least with the thermometer filled 

 with distilled hydrogen, also these temperatures may be determined 

 with the same degree of accuracy. 



^ 9. Vapour tension of liquid hydrogen at the melting point. 



By sufficiently lowering the pressure over the bath of liquid 

 hydrogen the temperature was reached at which the hydrogen in the 

 bath becomes solid. This temperature indicates the limit below which 

 accurate determinations are no longer possible by the method discus- 

 sed in this Communication. 



It could be accurately determined by a sudden change in the 

 sound which the valves of the stirrer in the bath bring about. (See 

 Comm. W. 94/, XII § 3). 



It appeared from the indication of the resistance thermometer that 

 the gas in the hydrogen thermometer had partly deposited. Hence 

 the pressure in the hydrogen thermometer gives the vapour tension 

 of liquid hydrogen near the melting point. For this we found : 



H5«= 53.82 m.m.^). 



§ 10. Reduction on the absolute scale. 



The reduction of the readings of the hydrogen thermometer on 

 the absolute scale by means of the results of determinations of the 

 isotherms will be discussed in a following Communication. 



§ 11 . Variations of the zero point pressure of the thermometer. 



It is noteworthy that the pressure in the thermometer in determi- 

 nations of the zero point slowly decreases. This change is strongest 

 when the thermometer has just been put together and becomes Jess 

 in course of time. This is very evident when the results of the 

 determinations made at the beginning of every new period of obser- 

 vation are compared, so after the thermometer has been left unused 

 for some time under excess of pressure. 



Thus on the fifth of July '05 shortly after the thermometer had 



1) For this Travers, Senter and Jaquerod (loc cit., p. 170) find a value lying 

 between 49 and 50 m.m. The great difference is probably owing to the inferior 

 accuracy of these last determinations. 



