— 29 — 
trouble by using an ordinary refrigerator; only, the temperature 
should not exceed 20°. 
Moreover, when the test^) is properly carried out, it will but rarely 
be necessary to observe the solutions for any prolonged time, as in 
most cases the possible separation of oil occurs either very soon, or 
immediately. It might also be tried whether these objections cannot 
be removed by using in Ceylon a slightly weaker alcohol. 
Further, Schimmel's test is also blamed for being unreliable, as 
pure oils are said occasionally not to pass the test, whilst other oils 
containing up to 15% petroleum form a clear solution with 80^0 
alcohol. 
The first complaint is certainly without foundation, as really pure 
Ceylon citronella oils will always pass Schimmel's test. With regard 
to the oils adulterated with petroleum, our experience is, than even a 
small admixture of the difficultly soluble American petroleum can be 
detected by the solubility test, whilst the more readily soluble Russian 
petroleum 2) may remain undetected, if it is only a question of slight 
admixtures — up to at most io^/q. But in order to render Schimmel's 
test also useful in such cases, we recommend a method which, we 
are informed, is sometimes employed in Ceylon^) in a similar manner. 
To the oil to be examined, 5 ^/q Russian petroleum is added, and the 
behaviour of the mixture towards 80 per cent, alcohol is now watched. 
An oil which originally was pure will under these conditions pass 
Schimmel's test. By these means we have in many cases detected 
adulteration, even when it was only slight. 
Now K el way Bamber^) has recently published a method, which 
is said to be more serviceable for judging a citronella oil than Schimmel's 
test, and which is said to have moreover the advantage that it permits 
of ascertaining at the same time the quantity of adulterant (insoluble 
in alcohol) which has been employed. This method is as follows: — 
A mixture of exactly 2 cc. pure cocoanut oil free from acids, and 
2 cc. of the citronella oil to be examined, is shaken in a suitable 
graduated glass tube, at 29° to 30°, for one minute, with 20 cc. 
alcohol of 83 ^/q by weight (d^ = 0,82 73) 5), and is then centrifuged 
^) Comp. in this respect our Report October 1903, 26, note. In order to 
avoid all misunderstanding, we would add to what we said there, that even when 
less than 10 vol. alcohol are added, a separation of oil should not occur. 
^) According to reports from Ceylon it is said that Russian petroleum now 
controls the market. Comp. Chemist and Druggist 62 (1903), 630. 
°) In Ceylon the oil to be examined is mixed with 4 to 6 ^/^ petroleum, and 
the mixture tested for solubility in alcohol of 85 ^/^ by volume. 
^) Proc. Chem. Soc. 19 (1903), 292. 
^) The Imperial Commission for Standards mentions, on the strength of the 
formulae calculated by Mendel ejeff, for an alcohol of exactly 83 ^/^ by weight, 
