252 



Scientific Proceedings, Royal Dublin Society. 



Discussion of Eesults. 



The new method was standardised at first by comparing the results obtained 

 by it with those obtained by Winkler's iodometric method, using distilled water 

 saturated in a thermostat under definite conditions. At a later date it was 

 tested by comparing the results of analyses obtained by its use and by Ijoilmg out. 

 Many series of experiments, with modifications of apparatus and procedure, were 

 performed, and the results are given in the accompanying table. Bach figure 

 represents the mean of three determinations. 



Eesults in c.c. gas at IST.T.P. per 100 c.c. water. Each figure average of S 

 determinations. 



It is 'seen that the total oxygen and nitrogen by the new method is con- 

 sistently greater than the standard values. On the other hand, the figure for 

 oxygen recorded by the new method is always less than that given by the 

 standards. The following appear to be the possible sources of error : — 



(a) Gas contained in potash sticks. — The magnitude of this error was 

 determined by performing blank experiments with air free water and 

 different amounts of potash in the apparatus. Minute bubbles were evolved in 

 all eases, and measured in the fine capillarj^ tubing of the tap. The size of the 

 bubble was proportional to the quantity of potash dissolved. When 26-4 c.c. of 

 water were saturated with potash the bubble measured 001 to 0015 c.c. This 

 would only account for about 006 c.c. of the difference of 0-12 c.c. between the 

 experimental and calculated values for total gas in the first iDart of the table. It 

 would, however, just suffice to account for the difference in results in the second 

 part. 



(&) Defective oxygen absorption. — At first the low value for the oxygen 

 figure was put do-wn to imperfect evolution of oxygen. When, however, this 



