( 368 ) 



rature could be found ^) at least with the thermo-elenient '). The 

 section of the glass capillary forming the stem of the thermometer 

 was 0,0779 mM\ With regard to the temperature correction (see 

 § 4, conclusion), it is desirable that this section is small. It appears 

 both from calculation and from observation '), that the equilibrium of 

 pressure between the space near the steelpoint and reservoir is still very 

 quickly reached with these dimensions of the capillary ''), much more 

 quickly than the equilibrium of the mercury in the two legs of the mano- 

 meter, which is inter alia also confirmed by the rapidity with which 

 the thermometer follows fluctuations in the temperature of the bath*). 

 The determination of the pressure which is exerted on the gas, 

 may, when the determinatioji applies to very low temperatures, be 

 simplified and facilitated by following the example of Chappuis') 

 and making the manometer tube serve at the same time as baro- 

 meter tube. The modification applied for this purpose to the arran- 

 gement according to Comm. N". 60 PL VI, is represented on PI. I, 

 which must be substituted for part of PI. VI belonging to Comm. N". 60. 

 By means of an india-rubber tube and a T-piece /„ the thermometer 

 {a, h, c, d, e, h, k) is connected on one side with the manometer 

 /j, to which (see PL VI Comm. N". 60) at m^ the reservoir at 

 constant temperature is attached and at m^ the barometer, on the 

 other side with the barometer tube {ii^, n^ (airtrap) n^. Besides from 

 the manometer and the barometer joined at iii^, the pressure can 

 now also immediately be read from the difference in level of the 

 mercury in n^ and in g. We have not availed ourselves of this 

 means for the determinations discussed in this Communication. 



§ 3. The hydrogen. 



The filling took place in two different ways : 



1) Travers, Senter and Jaquerod, (Phil. Trans. Series A, Vol. 200, Part. II, § 6) 

 who met with greater difficulties when trying to keep the temperature constant in 

 their measurements, had to prefer a smaller reservoir. 



2) A resistance thermometer is more sensible (Gf. Comm. Nos. 95" and 95<^). 

 As soon as one of suitable dimensions will be ready, the experiment will be repeated. 



^) Calculation teaches that for reducing a pressure difference of 1 c.M. to one 

 of 0.01 m.M., the gas flowing through the capillary requires 0.1 sec, the mercury 

 in the manometer 4 sec. Experiment gives for this time 25 sec. This higher 

 amount must be due to the influence of the narrowing at the glass cock k. 



^) We must be very careful that no narrowings occur. 



6) A great deal of time must be given to exhausting the reservoir witli the 

 mercury airpump when filling it, as the equilibrium of reservoir and pump is 

 established much more slowly than that between dead space and reservoir. 



•) Travaux et Mémoires du Bureau International, Tome VI. 



