ON MINIMETKIC ANALYSIS. 193 



inhaling the air out of the bottle^ so that fresh may enter. 

 No way is more exact than this, if care is taken not to 

 breathe into the bottle. This care is not at all difl&cult to 

 take; and no amount of apparatus can be more accurate 

 than this method, if done intelligently. If the slightest 

 amount of breath goes into the bottle, the process of rubbing 

 clean and drying must be undertaken anew. 



When the bottle is filled with the air of the place to be 

 examined, add the half ounce of baryta- water, put on the 

 stopper, and shake. If there is no precipitate, the air is 

 not worse than 0*04 per cent. When it is desired to ascer- 

 tain if it really contains as much as 0-04, then a bottle 

 holding 7'o6 ounces must be used. 



Having ascertained that the air around contains no more 

 than 0*04, it may be decided that a sitting-room shall not 

 be allowed to contain more than 0"o6, 0*07, or cio per 

 cent. If the first, then a bottle holding 3*78 ounces is 

 taken; if the air does not contain above 0"o6 per cent., 

 there will not be any precipitate in the liquid. If it is 

 allowed to contain O'lo per cent, (i per thousand), and on 

 some evenings many houses will contain this, then a bottle 

 of 2*46 ounces is enough. 



If in workshops 0'25 per cent, is allowed, then a bottle 

 holding 1*29 ounce is enough. 



This plan does not enable us to make an analysis of air. 

 The person to whom the care of the atmosphere would be 

 committed would have only one bottle of the proper size, 

 and would only require to see that the air never gave any 

 precipitate with that size of bottle. The order might be 

 given for any required purity, and by this test an unedu- 

 cated man could tell when the amount of carbonic acid was 

 too great. 



Eor a private house the rule would be not to have the 

 air above 0*07 at most ; better to have less. 



The baryta-water need not be of any particular strength ; 



SER. III. VOL. III. o 



