Lerts anp Brake—On Atmospheric Carbonic Anhydride. 448 
One other determination is necessary, namely, that of the 
carbonic anhydride which the absorbing solution contains before 
treatment with the air sample. This is done by measuring off 
50 c.c. from the stock-bottle, and proceeding as just described. 
Duplicate analyses are advisable ; but the amount having been once 
_ definitely ascertained for a given stock of the absorbent, it is obvious 
that no re-determinations are necessary until the stock is exhausted. 
Fic. 2. 
Testing the New Method.—In testing the new method, artificial 
mixtures were employed containing measured volumes of carbonic 
anhydride, and air previously freed from that gas, the mixture 
corresponding as far as possible in composition with ordinary air. 
Some difficulty was experienced in devising a method for 
obtaining a supply of air freed from carbonic anhydride in a dry 
vessel. Eventually the arrangement shown in fig. 2 was employed 
