1 34 WALKER. 



Ijeen able to detect some al)noi'iiiality in the conditions under wliicli the 

 samjile liad been kept. For instance, uuml)ers 1 and 2, samjiles of 

 identically the same oil, the otily difference between which lay in the 

 fact that number 2 had l^een treated for three hours witli a current of di'v 

 air to deodorize it, increased in free acid in practically the same ratio 

 during the first six months ; at the end of a year numlier 1 contains only 

 1 per cent more acid than number 2, but after three years shows 9.4 per 

 cent as against 3.7 jDer cent in, number 2. On examining the bottles In 

 which these samples were kept I noticed that number 1, of which a 

 considerable quantity had lieen prepared at the start, was still in the 

 original 250 cubic centimeter bottle in which it had first been placed. 

 About half of this oil had afterwards been taken out for the preparation 

 of numljer 2, and filled into a 100 cubic centimeter bottle. This dif- 

 ference in the size of bottle, then, with a corresponding difl^'erent surface 

 exposed to oxidation, must account for the 5.7 jDer cent excess of free acid 

 of number 1 over number 2. 



A still more striking exanifjle of the influence or surface oxidation is 

 aft'orded by sample B. This was an exceptionally pure oil, sterile and 

 freed as far as jDossible from impurities by repeated filtration. One por- 

 tion of about 25 cubic centimeters was transferred to a small bottle nearly, 

 but not quite, filled, shortly after preparation, while in tlie original 500 

 cubic centimeter bottle there remained at the end of three years about 

 20 cubic centimeters of oil. The latter sample has increased in acidity 

 8.8 per cent in three years, while the former, from which air was nearly 

 completely excluded has gained only 0.31 per cent free acid in the same 

 time. Sample number 5, a commercial oil which changed from 5.5 per 

 cent free acid to 28.2 per cent, had originally been treated with live 

 steam to remove its unpleasant odor, and decanted into a bottle without 

 filtering out- all the water of condensation from the steam. AMien the 

 final titration was made there remained between 5 and 10 cubic centi- 

 meters of oil, together with considerable water, so that hydrolysis by 

 water undoubtedly had much to do witli the large amount of free acid 

 developed. 



It will be noted that oxidation once started proceeds more rapidly in 

 oil already having a large free acid content than it does in those com- 

 paratively low in acidity (compare numbers G to 13 and 17 to 21 with 

 the commercial oils from numljer 23 on). Exceptions to this rule are, 

 as above stated, due to abnormal conditions of storage. 



The behavior of these oils during the period prior to the appearance 

 of oxidation has been discussed in a previous paper and no new data have 

 been brought out Ijy longer storage which do not tend to confirm the 

 conclusions drawn at that time. 



