RESEARCHES ON EVArOBATION AND DISSOCIATION. 307 



by a constriction in the tube. The open end of the U-tube 

 was connected with a gauge, the U-tube having been par- 

 tially filled with mercury. The U-tube was jacketed with 

 the vapour of mercury boiling under reduced pressures, so 

 that its temperature was known. Air was first completely 

 expelled from the closed end of the U-tube fiirough the 

 mercury by the vapour of the chloride. A largo number of 

 determinations was made by this method, and it was found 

 in all cases that, after a certain time, the rate of increase of 

 pressure at constant volume Avas constant, oven for many 

 hours. It was subsequently proved that this increase was 

 due to slow action of hydrochloric acid gas on mercury 

 (Than, Annalen 131, p. 131). Having ascertained the rate 

 of increase, it was possible to calculnie the true vapour- 

 pressure for each temperature. It is worth noting here 

 that the combination of dry HCl with dry NH^ is very 

 slow, even in presence of great excess of solid chloride. 



With nitric peroxide, contact with mercury was inadmis- 

 sible. A very large, thin-walled bulb, sealed to a graduated 

 stem, was filled with mercury, and, in short, acted as an 

 extremely delicate thermometer. It was surrounded by an 

 outer, much thicker bulb, and the space between the two 

 bulba was filled with the vapour of peroxide ; this space 

 communicated with an external bulb at some distance con- 

 taining a little liquid peroxide. Keeping the temperature 

 of the pseudo-thermometer constant, by means of snow, any 

 change of pressure caused by alteration of the temperature 

 of the liquid nitrogen peroxide made the mercury rise or 

 fall in the graduated stem. Experiments were made to 

 ascertain to what pressures the readings on the graduated 

 stem corresponded. For temperatures higher than 0°, tJie 

 temperature of the psoudo-thermometor was kept constant 

 by running water. 



