— 50 — 
for Y2 I minute. Whilst under these conditions, according to 
B amber's statements, when pure citronella oil is used, the latter 
remains in solution, and only the cocoanut oil is separated again 
quantitatively, there would be, if the citronella oil contained an adul- 
terant insoluble in alcohol, a corresponding increase in the volume of 
cocoanut oil, from which the quantity of the adulterant could be im- 
mediately ascertained. By multiplying with 50, the percentage of the 
adulterant is found. In order to obtain correct results, it is, according 
to B amber, especially necessary to use an alcohol of the strength 
mentioned by him, viz. 83 ^/q by weight. In doubtful cases he re- 
commends to compare the oil to be examined with an authentic pure 
oil for its behaviour towards Bamber's test. Bamber does not 
mention any results of tests whatever. 
With the view of forming an opinion on the value of Bamber's 
method, we have made a number of tests, partly with mixtures of 
Java citronella oiU) and petroleum, and partly with commercial Ceylon 
citronella oils. As at the same time we wished once more to verify 
the reliability of Schimmel's test 2), we have also submitted all the 
oils to the latter, and for the purpose of greater clearness have 
summarised the results obtained by both methods in the following 
tables. When Bamber's test was employed, the alcoholic mixture 
was in every case centrifuged for i minute, corresponding to about 
1500 revolutions. 
The results obtained show clearly that Bamber's method renders 
satisfactory services for the qualitative estimation of an oil, and in 
a few cases also shows an adulteration where the ordinary Schimmel's 
test fails. With regard to the quantitative estimation of the adulterant, 
however, the new test is a failure. This must be attributed on the 
one hand to the more or less ready solubility of the adulterant in 
the alcohol employed, and on the other, to the fact that with the 
closely graduated scales slight errors in reading easily occur, which, 
in consequence of the small quantities of citronella oil used for the 
estimation, produce fairly large differences in the percentages. 
d^.o = 0,84108; such an alcohol has (1^=0,8283. According to the tables 
referred to, the specific gravity d^ = 0,8273 mentioned by Bamber, corresponds 
to an alcohol of 83,39 ^/^^ hy weight. 
^) As at the present time the commercial Ceylon citronella oils are mostly 
untrustworthy, we have employed Java oils in the mixtures with petroleum, in order 
to be more certain. In consequence of the somewhat different composition of 
these oils, it is possible that the results may vary slightly if an authentic pure 
Ceylon citronella oil is used, but in our opinion the difference cannot amount to much. 
^) In some of these cases the "raised Schimmel's test" (test of the oil after 
adding 5 ^/^ Russian petroleum) has been employed. 
