258 THE CHLOR-HYDBOGEN PHOTOMETER. [MEMOIR XVIII. 



charged with chlorine, an excess of that gas would ac- 

 cumulate in the sentient tube. 



It is absolutely necessary, therefore, that there should 

 be an equilibrium between the gaseous mixture and the 

 confining fluid. 



As has been said, when hydrochloric acid is decom- 

 posed by a voltaic current, all the chlorine is absorbed 

 by the liquid and accumulates therein ; the hydrogen 

 bubbles, however, as they rise withdraw a certain pro- 

 portion, and hence pure hydrogen passed up through the 

 photometric fluid becomes exceedingly sensitive to the 

 light. 



There are certain circumstances connected with the 

 constitution and use of the photometer which continu- 

 ally tend to change the nature of its liquid. The plat- 

 inum wires immersed in it by slow degrees give rise to 

 a chloride of platinum. It is true that this takes place 

 verygradually, and by far the most formidable difficulty 

 arises from a direct exhalation of chlorine from the nar- 

 row tube ef, for each time that the liquid descends, a 

 volume of air is introduced, which receives a certain 

 amount of chlorine which with it is expelled the next 

 time the battery raises the column to zero; and this 

 going on time after time finally impresses a marked 

 change on the liquid. I have tried to correct it in 

 various ways, as by terminating the end f with a bulb ; 

 but this entails great inconvenience, as may be discov- 

 ered by any one w r ho will reflect on its operation. 



When by the battery we have raised the index to its 

 zero point, if the gas and liquid are not in equilibrium 

 that zero is liable to a slight change. If there be hydro- 

 gen in excess, the zero will rise; if chlorine, the zero 

 will fall. 



In making what will be termed " interrupted experi- 

 ments," we must not too hastily determine the position 





