32 OLIVK OIL AND ITS ST INSTITUTES. 



Table XXII is ;i compilation of the results of u large number of 

 analyses, and gives the limits which have been obtained for this oil. 

 Some of these figures are so extraordinary that the oils from which 

 they were obtained can hardly be considered otherwise than abnormal. 

 In this bulletin the term "normal oils" is used to indicate those which 

 are not so rancid as to be. unfit to be eaten or which have not been 

 oxidized by artificial means. On rancid or oxidized oils results can 

 be obtained which are entirely different from those which would be 

 found on fresh oils, and in this report salad oils only are considered. 

 Cotton-seed oils with a specific gravity as low as 0.9160 and as high as 

 ().l:>r>0 are most likety abnormal. Such figures can hardly be consid- 

 ered of any value unless further data as to free acids and condition 

 of the oils are known. In this connection samples Nos. 13235 and 

 23656 of peanut oil in Table XXXIII serve as examples. The first 

 oil was what is called a "blown" oil, and the results obtained with it 

 are entirely abnormal. It has a specific gravity of 0.9364, an iodin 

 number of 77, and saponification number of 199. The second has a 

 specific gravity of 0.9155, but has 13.51 per cent of free acid. Such 

 an oil could not be used under any circumstances as salad oil, and cer- 

 tainly should not be considered in establishing limits for peanut oil. 



\Vhen added in any considerable amounts to olive oil, cotton-seed 

 oil is not difficult to detect, on account of its high specific gravity, 

 iodin number. Maumene figure, and index of refraction. Moreover, 

 t lie re are the qualitative tests of Bechi, Halphen, and the nitric acid 

 reaction to detect it in small amounts. The Halphen test a is extremely 

 delicate, detecting 1 per cent or less of the unheated oil. This test is 

 characteristic of cotton-seed oil, and if obtained in an olive oil is proof 

 of its presence. But Holde and Pelgry 6 have shown that cotton-seed 

 oil if heated for ten minutes to 250 C. loses the power of giving both 

 the Halphen and Bechi tests. The writers found that it took twenty 

 .minutes to completely destroy the power to give the Halphen test, but 

 the reaction was greatly weakened on oils heated a much shorter time. 

 This heated oil if refined would lose the taste and odor due to heating 

 and could be mixed with olive oil in small amounts without giving the 

 Halphen, Bechi, or Milliau test. But the nitric acid reaction is 

 apparently not affected in any way by this heating. 



Tortelli and Ruggrri'' have made an extensive study of heated 

 cotton-seed oil in order to discover means of detecting it in olive oil. 

 They found that it was necessary to heat the cotton-seed oil twenty 

 minute- at L'.'.n ( '. in order to completely destroy its power of reacting 

 with Halphen or Bechi reagent-. The value- of the oil are not 



"V. <. Dc|t. A-*-.. Kin-ran of Chemistry linl. ,:,. ,,. 82. 



''Clu-iii. Key. 1'Vtt. M. Hurt/, In.l.. ls<), <>: (17; .!..ur. S..r. Clu-in. hid., 1SW, 18: 

 711. 



'. \miali del Lal. rhim. cent. delle < ialn-llr. IMOII, j,. L'l'.t. 



