68 G 



PROCEEDINGS OF SECTION I. 



Table IV. 



With regard to the second (Table IV), when the amount of 

 oxygen was below -3 all four samples decomposed, and when above 

 •9 none (77) went. Between these there were 20 samples, of which 

 seven decomposed, but the intermediate series presented great 

 irregularities. 



The relation between oxygen absorbed and decomposition 

 (Table III) showed similar irregularities. When the amount was 

 less than -7 no decomposition occurred (66 samples). Between '7 

 and I'O, 6 out of 29 samples went, and above TO, 5 out of 13. 

 Further, the samples with the highest figures (C2 and C4, 1.35 and 

 1-33 respectively) did not decompose, while others with much 

 lower figures, e.g., B35 and B36 (TOS and TIO) went the opposite 

 way. 



Whilst, therefore, one may broadly say that the probability 

 of decomposition is inversely proportional to the amount of oxygen 

 in nitrites and nitrates and directly to the oxygen absorbed figures,, 

 still a much closer approximation of this chance may be made by 

 estimating the ratio between these two factors. 



Reference has been made above to a paper by Johnston and 

 others. Those authors take the oxj'gen absorbed in three minutes 

 at 80°F. as a basis of calculation, but they arrive at this by doing 

 a boiling-for-five-minutes' test and by dividing the result by 5. 

 They consider this as giving substantially the same figure as the 

 ordinary three-minute test, but consider it better inasmuch as the 

 influence of such substances as nitrites is minimised thereby. 

 Their third conclusion states that when the consumed oxygen value 

 is less than the oxygen in nitrites and nitrates, under ordinary con- 

 ditions, the sample wiU not putrefy. 



In the following table I have shown the relation between these 

 two factors in 105 samples of 108 dealt with above. The three- 

 minute oxygen absorbed test was made in the usual way. 



