Commercial notes and scientific information. 45 



(oximation method). For Ceylon oil we have so far found from 29,6 to 34,4 p. c. 

 geraniol, and, according to the method employed from 7,5 and 11,6 (Weber) or 6,5 and 

 8,0 p. c. (Dupont and Labaune) citronellal. 



It has been shown that for the phenylhydrazine method it is advisable to take 

 1 gram Ceylon oil or 0,5 gram Java oil (exactly weighed), for the rest the method is 

 applied as described in our Report of April 1912, p. 75, for lemon oil. The phenyl- 

 hydrazine must be allowed to react for at least one hour and the reaction may be 

 prolonged to lVa hours, but it is not advisable to let it proceed beyond that limit, 

 because then the phenylhydrazine solution acquires such an intense colour as to 

 make titration more difficult. 



As regards the special estimation of geraniol by means of phthalic anhydride, it 

 goes without saying that in this case the other primary alcohols which occur in citro- 

 nella oil (nerol and citronellol) also enter into reaction, but they are quantitatively of 

 so little importance that there is no need to pay any attention to them in carrying out 

 the estimation. 



In connection with the above we may mention that we have also tested the phenyl- 

 hydrazine method with a series of other aldehydes and ketones. These tests showed 

 that cuminic aldehyde, benzaldehyde, and methylnonylketone may be quantitatively 

 estimated by the same method, while fenchone, thujone, camphor, and menthone give 

 no reaction at all with phenylhydrazine. In each case the quantity used for estimating 

 was from 0,12 to 0,15 grams. The following table gives particulars of the results 

 obtained: — 



Cuminic aldehyde (cont.2,7 p.c. cuminic acid) found 95,7 and 96,2 p.c. cuminic aldehyde, 

 50 p.c. cuminic aldehyde +50 p.c. carvene 1 ) found 51,5 and 49,5 p.c. cuminic aldehyde, 

 Benzaldehyde (cont. 2,4 p.c. benzoic acid) . found 98,0 and 99,6 p.c. benzaldehyde, 

 Methylnonylketone found 99,7 and 98,6 p.c. methylnonylketone. 



It should be borne in mind that cuminic aldehyde and benzaldehyde may contain 

 free acid due to oxidation. This explains why otherwise pure aldehydes in certain 

 circumstances did not afford 100 p.c. By this method we found cumin oils to contain 

 between 35 and 42 p. c. of cuminic aldehyde. In the case of bitter almond oils con- 

 taining hydrocyanic acid (as more fully explained on p. 24 of the present Report), only 

 the free benzaldehyde and not that which is in combination with hydrocyanic acid, 

 enters into reaction with phenylhydrazine. 



As we have shown, methylnonylketone is also capable of being estimated quanti- 

 tatively, and the same probably applies to the methylheptylketone which has been 

 detected in oil of rue; hence the method may eventually be used for testing the last- 

 named oil. Here, however, it should be remembered that in rue oil both the ketones 

 are always present in varying proportions. Hence the quantity of phenylhydrazine 

 which has been used only affords an approximate clue to the quantity of the ketones, 

 and it must at the same time be ascertained, by determining the solidifying-point, 

 which of the two ketones predominates in the oil, and whether it will be more practi- 

 cable in estimating to use as a basis the molecular weight of methylnonylketone (170) 

 or that of methylheptylketone (142). It remains to be seen to what extent this will 

 afford practically useful results. We found the value for an oil of rue with about 



x ) The carvene used by us as a diluent had been repeatedly distilled over sodium, but nevertheless it 

 used up a quantity of phenylhydrazine corresponding to a proportion of from 0,4 to 0,6 p.c. cuminic aldehyde. 



