In addition to the two acids already mentioned, viz., acetic acid and benzoic acid (m. p. 122°)8), salicylic acid9) has been found, combined with methyl alcohol (methyl iodide)8) and benzyl alcohol (m. p. of phenyl urethane 78°)s). They constitute valuable constituents of the oil. Hence of esters there are present: methyl benzoate, methyl salicylate1), benzyl acetate, benzyl benzoate, besides free benzyl alcohol-).

*) Bull. Soc. chim. III. 11 (1894), 407, 576 and 1045; 13 (1895), 140.

-) Compt. rend 76 (1873), 1482.

s) G. I. P. 142859.

4) Report of Schimmel 8j Co. October 1901, 53.

5) F. A. Fluckiger, Arch, der Pharm. 218 (1881), 24.

°) Report of Schimmel $ Co. October 1901, 53 and 54.

7) Ibidem April 1903, 79.

*) Ibidem April 1902, 67.

8) Ibidem April 1900, 48.

Fig. 37. Distillation of cananga flowers in Java.

Fig. 37. Distillation of cananga flowers in Java.

The presence of the following substances may be assumed, although it has not yet been definitely established; viz., anthra-nilic acid methyl ester which causes the slight fluorescence of the oil, and creosol (3-methyl ether of the homopyrocatechol)3). In the saponification liquid, G. Darzens4) found methyl alcohol and cresol (m. p. of benzoate 70 to 71°). Of the latter he thinks that it is contained in the original ylang-ylang oil as acetyl-p-cresol.

R. F. Bacon5) identified formic acid and either safrol or isosafrol as constituents of the oil. It has not been definitely decided which of these two phenol ethers is present. Inasmuch as the odor of heliotropin was produced during the oxidation with potassium bichromate, the presence of isosafrol may be assumed. Contrary to the observation of Fluckiger6), Bacon did not obtain a decided color reaction with ferric chloride. He draws the inference that upon ageing the phenol ethers may be hydrolized. However, with fuchsine sulphurous acid the characteristic color reaction for aldehydes was obtained, although it was not possible to isolate aldehyde-like constituents either with phenyl hydrazine or bisulphite. Hence, they can be present in traces only7), if at all. Possibly, small amounts of valeric acid are also present.

From Javanese cananga oil Elze8) has isolated about 0,2 p.c. of nerol (m. p. of diphenyl urethane 50 to 50,5°) and about 0,3 p.c. of farnesol (b. p. 145 to 146°; d15o 0,895; aD inactive) by means of the acid phthalic ester.

Properties. As already pointed out under "Production" on p. 396, in the distillation of the ylang-ylang oils of better quality the more fragrant constituents that come over first are collected separately. Inasmuch as this separation is controlled merely by the sense of smell, it is not surprising that the products differ greatly. Hence the physical constants fluctuate within wide limits. Consequently it is often impossible to place a certain oil merely by its physical constants into one of the two grades of commercial oils, viz., ylang-ylang oil and cananga oil.

1) Report of Schimmel & Co. April 1900, 48. 2) Ibidem April 1902, 67.

3) Ibidem October 1901, 57.

4) Bull. Soc. chim. III. 27 (1902), 83.

5) Philippine Journ. of Sc. 3 (1908), A, 65. 6) Arch, der Pharm. 218 (1881), 24.

7) These reactions are possibly due to traces of benzaldehyde. 8) Chem. Ztg. 34 (1910), 857.

Properties of Ylang-Ylang Oil numerous observations admit of a conclusion, the physical constants do not differ materially from those of the Manila oil1).

With Manila oils the following observations have been made in the laboratory of Schimmel & Co.: d15o 0,930 to 0,945; aD - 37 to -57°; nD20o 1,491 to 1,500; A. V. up to 1,8; E. V. 75 to 120; E. V. after acetylation 145 to 160°.

In alcohol the oil is difficultly soluble. Ordinarily it yields a clear solution with 1/2 to 2 vol. of 90 p.c. alcohol which, however, as a rule becomes turbid again upon the addition of more alcohol. Similar is the behavior of the oil toward 95 p.c. alcohol. With ferric chloride the alcoholic solution of the oil yields a more or less pronounced violet color reaction.

In connection with 23 samples of Manila oils, Bacon1) established the following limit values: d30/4o 0,911 to 0,958; aD30° - 27 to -49,7°; nD30oJ,4747 to 1,4940; E. V. 90 to 150, in one instance 169. He arrived at the conclusion that, as a rule, the ester value is 100 or more, nD is rarely over 1,4900 and aD more than - 45° and that, as a rule, the latter fluctuates between - 32 and - 45°. In a general way it may be said that oils with a low ester value are inferior and that a high ester value is an indication of a good oil. The other properties vary: specific gravity, optical rotation and index of refraction are independent of the ester value. At times they are high, at other times low and do not afford a criterion as to the value of the oil. However, a high specific gravity and a high ester content as a rule cause a high index of refraction and such oils may be regarded as specially good.

Qualitatively, the ylang-ylang oil from Reunion is greatly inferior to that from Manila2). So far as the hitherto not very d15o0,932 to 0,962; aD - 34 to - 64°; E. V. 96 to 134. An older oil2) had the high density of 0,974 at 15°.

1) Loc. cit and Philippine Journ. of Sc. 5 (1910), A, 265. 2) In their report of October 1910, 63 Roure-Bertrand Fils make the following statement: "It would be better by far if less but better oil were produced".

In connection with other oils, which, however, play no commercial role as yet, the following constants have been established:

Jamaica3) (from the botanical garden at Kingston): d16o9407; aD - 38°6'; nD20o1,50510; A. V. 1,0; E. V. 57,6. Madagascar4): d15o0,9577; aD - 49°55'; nD20o1,51254; A. V. 1,8; after acetylation 167 to 180,8.