Gooch and Kr eider — Ch lorine for Laboratory purposes. 169 



and yet it is plain that a considerable amount of the oxide 

 still accompanied the chlorine. We made the attempt, there- 

 fore, to accomplish the destruction of the oxide by passing the 

 mixed gases as they issued from the generator through a small 

 wash-bottle containing a hot saturated solution of manganous 

 chloride in strong hydrochloric acid, making use of a device 

 previously employed in this laboratory for the decomposition 

 of nitric acid and chloric acid.* The results of several experi- 

 ments conducted in this manner are recorded in the accom- 

 panying statement. The gas after leaving the manganous 

 chloride passed through a small wash-bottle filled with water, 

 was dried by calcium chloride, and collected in the pipettes 

 arranged in train, and analyzed by the method previously 

 described. 



Strength of 



hydrochloric 



acid. 



Temperature 

 of acid. 



Temperature 



of the solution 



of MnClo. 



Bubbles of gas 

 passing the 

 wash-bottle 

 per second. 



Per cent of 

 chlorine. 



Sp. gr. 1*1 



83° 



90° 



5-6 



87-5 



cc a 



83 



90 



4-6 



89-5 



a a 



70-80° 



90 



1-3 



96-4 



a a. 



87-5 



90 



3 



98-6 



it a 



83 



90 



3 



96-9 



It is evident from these figures that the efficiency of the man- 

 ganous chloride is considerable and naturally most manifest 

 when the current of gas is slow. Indeed, during the passage 

 of gas through the manganous salt small bubbles of chlorine 

 are evolved from the entire surface of the liquid. The washed 

 gas is pure enough for most laboratory purposes, but if chlo- 

 rine perfectly free from chlorine oxide is desired it may be 

 obtained by passing the washed gas through a hard glass tube 

 filled with asbestos and heated by a Bunsen flame. 



In the exceptional cases in which the ignition tube is used it 

 is well to keep in mind the fact that in starting the generator 

 the acid should not be thrown in too great quantity upon the 

 chlorate at first lest an unusually large proportion of the di- 

 oxide be liberated, which may cause a slight explosion in the 

 ignition tube. Should such be the case, however, the liquid 

 in the wash-bottle will prevent the extension of the explosion 

 to the gas in the generator proper, and the insertion of the 

 wash-bottle should never be omitted when the ignition tube is 

 to be put to use. So long as it is kept hot we have never 

 found the apparatus other than perfectly manageable, safe, and 

 automatic in action. 



Inasmuch as a single gram of potassium chlorate produces 

 approximately a half-liter of chlorine, and a cubic centimeter 



*This Journal, xliv. 117 ; xlvi, 231. 



