( 369 ) 



n. By means of hydrogen prepared in the apparatus of Comm. 

 N". 27 with the improvements described in Comm. N". 94« (June 1905) 

 § 2. After liaving beforehand ascertained whether all junctures of 

 the apparatus closed perfectly, we maintained moreover all the time 

 an excess of pressure in the generator, in order to exclude any 

 impurity from the gas. The expulsion of the air originally present 

 in the apparatus was continued till it could be present in the gas at 

 the utmost to an amount of 0.000001. 



h. By means of hydi'ogen prepared as described in Comm, 

 N". 94/ XIV. In order to apply this more effective mode of 

 preparation, we must have liquid hydrogen at our disposal. ^) In § 7 

 the equivalence of the first method with the last is demonstrated 

 for measurements down to — 217°. It is still to be examined whether 

 systematic errors maj result from the application of the first method 

 of filling, in measurements on liquid hydrogen by the deposition 

 of impurities, less volatile than hydrogen. *) 



^ 4. The measurements. 



The zero point of the thermometer is determined before and after 

 every set of observations. Both for the zero point and for every 

 determination of temperature, an average value is derived from 

 three or four observations. Each of these observations consists of a 

 reading of the barometer, preceded and followed by a reading of the 

 manometer. The thermometers, indicating the temperature of the 

 mercury, of the scale and of the gas in the manometer spaces are 

 read at the beginning and at the end of every observation. The tem- 

 perature of the room is kept as constant and uniform as possible. 



The temperature of the thermometer reservoir is taken equal to 

 that of the bath. This is permissible for the cryostats described in 

 Comm. N". 94^ and Comm. N". 94^^ and the treatment given there. 



The temperature of the bath is kept constant by means of the 

 resistance thermometer, described in Comm. N". 95<^. In order to 

 facilitate the survey of the observations, the resistance was adjusted as 

 accurately to the same value as possible, and by means of signals the 

 pressure in the cryostat was regulated in such a way, that in the very 

 sensible galvanometer the mirror made only slight oscillations about 



1) The hydrogen in the vacuum glass B (see Coram. N". 94 ƒ XIV, fig. 4) proved 

 to evaporate so slowly, that a period of two hours was left for filling and 

 exhausting the thermometer again, which previously had been kept exhausted for 

 a long time, being heated during part of the time (cf. footnote 5 § 2). 



2) In a former set of observations deviations were found, which in conjunction 

 with each other prove that the hydrogen must have been mixed with air. 



