1 28 WALKER. 



It is necessaiy for the interpretation of these results to consider the 

 many factors which may enter into tlie decomposition of a freshly pre- 

 pared, commercial coconut oil. 



First, we may have present fat-splitting molds, albuminoids, sugar and 

 water, which cause the turbid appearance of commercial oils. It has 

 been shown in a previous ]:)aper " that mold action on copra is' the 

 princijjal factor in determining the initial acidity of an oil, and that 

 these same molds in the presence of sufficient nutritive matter may effect 

 the rapid decomposition of even a pure oil. Second, soluble or insoluble 

 enzymes jn'oduced by these molds may be the cause of the rise in free 

 fatty acid. Third, surface oxidation by the air, possibl}' assisted by light, 

 may take part in the decomposition. This surface oxidation is always 

 accompanied by a pungent, disagreeable odor, and the formation of 

 aldehydes and peroxides. Fourth, simple hydrolysis by heat and moisture 

 may be a factor. 



Considering first sample number 1 of this series, it is evident that 1-b, 

 the sample in the large bottle, may be subject to any one or all of the 

 foregoing factors causing increase of acidity, wliile 1-a, being in a p>rac- 

 tically full bottle, may be affected by any influence except that of oxida- 

 tion by the air. Any marked increase in acidity, then, of 1-b over 1-a 

 would be due to surface oxidation. By reference to the table we find 

 that du.ring the first year there has been practically no difference in the 

 rate of increase of acidity between 1-a and 1-b, which show respectively 

 a gain of 3.6 and 3.4 per cent. The second year, however, indicates a 

 decidedly different condition. While 1-a has only gained 0.4 per cent 

 free acid, 1-b has increased 2.4 per cent, leaving a difference of 3 per cent 

 which can only be due, aside from experimental error and an individual 

 variation in the samples, that should at most amount to no more than 

 a few tenths of a per cent, to surface oxidation by the air. The question 

 naturally arises, why should a jjeriod of two years be required for this 

 difference to show itself? This can be accounted for by two theories: 

 First, the presence of molds and of nutritive bodies such as sugar and 

 albumen, which would be in the oil in larger quantity in the first year 

 than in the second, is not favorable to the formation of peroxides of 

 the fatty acids, a fact I have previously noted in testing pure and com- 

 mercial oils for peroxides and aldehydes ; ' second, it is piossible that this 

 inliibiting effect may be mutual, so that oxidation once started would tend 

 to kill off or check mold or enzyme activity in 1-b sooner than this 

 activity would naturally cease through lack of nutriment in 1-a. In 

 either case, the result would be the same and the combined effect of the 



"This Journal (1906), 1, 139. 

 ''hoc. cit., 1.39. 



