For the detection of smaller amounts of petroleum, the presence of which is not revealed by the test just described, Schimmel & Co.1) have recommended their modified, more rigid Schimmel's test. This consists in adding 5 p. c. of Russian or American2) petroleum to the oil and subsequently testing it with 80 p. c. alcohol. Even when thus modified, pure citronella oil yields a clear solution with 1 to 2 vols, of 80 p. c. alcohol which remains clear also when the amount of solvent is increased to 10 vol. At most the solution reveals a slight opalescence but should not separate oily drops upon standing.

1) Report of Schimmel & Co. October 1912, 43.

2) Bericht von Schimmel & Co. October 18S9, 22. Report April 1896, 16.

Naturally, this sort of test affords but a general criterion as to the quality of an oil and leaves much to be desired. However, it has the advantage of being readily performed and may be expected to continue in practice in spite of the enemies it has made. Up to the present time at least, it has not been possible to replace this test by one equally simple and better. An attempt in this direction was made several years ago by M. K. Bamber3). He published directions for a test for which it was not only claimed that it was better than Schimmel's test, but also afforded the means of ascertaining the amount of adulterant insoluble in alcohol. The technique of the test known as "Bamber's Test" is as follows:

In a glass cylinder provided with the necessary graduation, a mixture of 2 cc. of pure, acid-free cocoanut oil and 2 cc. of citronella oil is shaken for a minute with 20 cc. of alcohol of 83 p. c. by weight (d30/15o 0,8273) at 29 to 30°. It is then centri-fuged for 1/2 to 1 minute. According to Bamber, pure citronella oil is completely dissolved in the alcohol, wheras the cocoanut oil remains behind quantitatively. If the citronella oil contains an adulterant insoluble in alcohol, this likewise is separated. Hence the volume for the cocoanut oil is increased and from the increase the amount of adulterant becomes apparent.

Schimmel & Co.4) have tested this method as to its practicability and have found that, whereas it may render good service for the qualitative testing of citronella oil, it fails as a quantitative test. Moreover it becomes apparent that unadulterated oils do not always comply with this test. This certainly is a disadvantage. It should be added further that Bamber's test demands greater exactness on the part of the operator than does Schimmel's test and that the apparatus is more complex. As a result, Bamber's test has not been generally adopted.

1) Report of Schimmel & Co. April 1904, 29; April 1910, 40. 2) Ibidem April 1911, 47.

3) Proceed, chem. Soc. 19 (1903), 292.

4) Report of Schimmel & Co. April 1904, 29.

Adulterations. As already pointed out, Ceylon citronella oil is adulterated principally with fatty oil and petroleum. Occasionally an oil is met with to which alcohol has been added. The addition of alcohol is indicated by a lowering of the specific gravity and can be shown by the reduction in volume when shaken with water. Inasmuch as alcohol is included in the geraniol assay when the oil is acetylated, its presence causes the total geraniol content to be seemingly increased. The true total geraniol content is ascertained only when the oil has been shaken repeatedly with water before it is acetylated. Details for the detection of alcohol will be found on p. 612 of the first volume of this treatise. The terpenes from lemon oil have also been observed as adulterant1).