xv , 5 Perkins: Rancidity of Coconut Oil 471 
that of the acids from the neutral fats in the same oil. It seems 
probable that the free fatty acids had suffered much more oxida- 
tion in the process of rancidity than had the unhydrolyzed 
glycerides. Hence Thum’s experimental data tend to disprove 
his conclusions. The subject requires further investigation. 
Light . — In view of the results shown in Table III, I cannot 
subscribe to either of the extreme opinions quoted from Lew- 
kowitsch and Walker, respectively, on the influence of light. 
On the one hand, the samples of series Od, though kept in dark- 
ness, all became rancid; on the other hand, exposure to light 
had a marked effect. It increased hydrolysis even in the sealed 
samples, and in two of these (2 SI and 5 SI), in conjunction 
with moisture, it caused a complete loss of fresh-coconut odor. 
In the open samples of oil A it doubled the hydrolysis, more 
than doubled the decrease in iodine number, and intensified the 
rancid odor. 
Enzymes . — The enzymes mentioned in the quotations from 
Walker, given early in this article, are those produced by molds. 
A calculation from his results 17 on the increase of acidity in a 
sample of oil from moldy copra indicates an enzyme effect in 
some cases. Fat-soluble enzymes seem to have been responsible 
for 0.5 per cent acidity increase during the first year and 0.2 
per cent during the second. This was in the samples sealed 
in full bottles. In the half-full bottles no effect can be attributed 
to fat-soluble enzymes. Practically no effect seems to have 
been produced by enzymes insoluble in fat. 
Lewkowitsch, 18 in placing emphasis on the action of enzymes, 
calls attention not so much to mold enzymes as to those fat- 
splitting enzymes which, he says, 
seem to occur in most, if not all, oleaginous seeds, and no doubt play 
an important part in the utilization of the fatty reserve products stored 
in the seeds. 
The evidence concerning the presence of enzymes in the 
coconut is somewhat conflicting, but it has been shown in the 
laboratory 19 of the Bureau of Science that normal coconut meat 
contains little or no lipase, zymogen, or oxidase. An oxidase 
was found, however, in coconut milk, which accounts for its 
occasional presence in the oil. 
In the present investigation the oils were prepared from 
fresh, dried coconut meat not appreciably attacked by micro- 
17 Philip. Joum. Sci. § A 3 (1908) 127. 
IS Op. cit. 49. 
19 Brill and Parker, Philip. Joum. Sci. § A 12 (1917) 109. 
