Becker & Abbott — Method of Estimating Dissolved Oxygen, <^c. 2o6 



figure continued iow in further series of experiments, and, moreover, the total gas 

 figures remained higlier than the tabular value, defective oxygen absorption 

 appeared to be the cause. The literature revealed a widespread dissatisfaction 

 with alkaline pyrogallol as an oxygen absorbent for very accurate work, and very 

 varied recipes are given by different authors. In our work the difficulty inherent 

 in alkaline pyrogallol is no doubt greatly aggravated by the minute size of the 

 bubble to be tested. The fonnula we linally adopted approximates to that 

 recommended by Anderson.' It is, nevertheless, not wholly reliable, as is amply 

 proved by the unsatisfactory results recorded in tlie second last series in the table, 

 obtained when using a freshly prepared sample of tiiis reagent. It was further 

 noticed that the reagent does not work very well on the first absorption of a 

 series if it has remained idle for some time. Tliis difficulty, which is mentioned 

 by other workers, can he o\'ercome by absorbing an unknown volume of oxygen 

 from air before beginning a series of determinations. 



Chromous chloride was tried as an alternative absorbent, but had to be 

 abandoned owing to the precipitation which occurred when the acid solution met 

 with a trace of potash. 



(c) IncomijUie evolution of gas. — The facts discussed under (a) mdicate that 

 it is probable that the error introduced by this cause is negligible. 



Conclusions. 



The work so far completed shows that the use of a very soluble substance, 

 such as potash, to exi^el dissolved gases has many advantages. The value of 

 direct gasometric measurements is very generally recognised, and volumetric 

 methods are only resorted to in order to avoid the tedious process of boiling out 

 the gases and measuring them. This method gives all the advantages of the 

 gasometric method, and the further important one of rapidity. 



The moditications which must be introduced into the simple Winkler pi'ocess, 

 for instance, when dealing with hea\-ily polluted liquids, or in the presence of 

 chlorides, not only complicate the manipulation, but add very greatly to the 

 time required. On the other hand, there is no reason to suppose that the action 

 of the potash would be affected by the ordinary impurities in a water, and an 

 estimation could be done as easily as with distilled water. 



Where the quantity of the sample is limited, the method only requires 20 to 

 30 c.c. of the water, aw compared with the 250 c.c. used in the volumetric 

 processes. 



The disadvantage (from a commercial point of view) of the cost of the potash 

 may quite possibly be removed by a more minute search for a suitable substance 

 than the present authors ha^'e had opportunity to make. The method certainly 

 works much better than its simplicity ^vould lead one to expect, and it is hoped 

 that it may be possible to develop it still further in the future. 



Chemical Labokatoky, 



College of Science, Dublin. 



1 Jouru. Ind. Ens- Chem., 1915, p. 587. 



SOIENT, PKOO. R.D.S., VOL. XVII, NO. bOi 3 A 



