— 76 — 



5 cc. water and added after separation of the solution previously run off, 

 and the total contents of the flask titrated with seminormal soda liquor. 

 Di-ethyl orange is used as an indicator and in all cases the titration is 

 made for the brownish tint which precedes the pink coloration. As each 

 cc. of seminormal soda liquor corresponds to 0,076 g. of citral, the citral 

 content of the oil may be calculated from the amount of soda liquor 

 which has been used up. 



Kleber found that parallel-tests almost invariably produced exactly 

 corresponding results, and that in the case of mixtures of known com- 

 position the difference rarely exceeded 0,1 °/o. As the phenylhydrazine 

 solution decomposes very rapidly, it is best prepared fresh each time; and 

 in no case should a solution be employed which is more than 24 hours old. 

 If necessary, the phenylhydrazine should be rectified beforehand in vacuo. 



Kleber has ascertained that his method is also suitable for estimating 

 the citronellal-content of citronella oil, and that it may be specially 

 employed with advantage in the evaluation of Java citronella oil. He 

 found Ceylon citronella oils to contain from about 7 to 9% » and Java oils 

 from about 32 to 37% citronellal, but in view of the limited number of 

 estimations which have thus far been carried out these figures are of 

 course not to be taken as limits of value 1 ). Kleber uses 5 g. Ceylon oil 

 or 2 g. Java oil in his determinations. 



Kleber believes that his method may be found generally applicable 

 for the estimation of aldehydes and ketones and leaves it open to his 

 colleagues in research-chemistry to make further experiments in this 

 direction. 



We have already stated that Dr. Kleber communicated his method to 

 us privately some considerable time ago and that, as our own experience 

 has shown it to give really practicable results, we are able to recommend 

 it. We ourselves have somewhat modified the method and proceed as 

 follows: — 



About 2 cc. oil are mixed with 10 cc. of a freshly-prepared 2°/o alcoholic 

 phenylhydrazin solution, and allowed to remain undisturbed during 1 hour 

 in a glass-stoppered flask of about 50 cc. capacity. We then add 20 cc. 

 decinormal hydrochloric acid and mix the liquid by gently moving the 

 flask backwards and forwards. After adding 10 cc. benzene the mixture 

 is shaken vigorously and poured into a separating funnel. The acid layer, 

 amounting to about 30 cc, which separates out clearly after the mixture 

 has been left to settle a short time is then filtered in a small filter. # 



To 20 cc. of this filtrate 10 drops di-ethyl-orange solution (1 :2000) are 

 added and the mixture is titrated with decinormal potash liquor until a 

 distinct yellow colour appears. From this the quantity, in cc, of deci- 



*) We ourselves have previously found a higher citronellal-content in both these oils 

 by an indirect method of estimation. Further examinations are needed to clear up these con- 

 tradictory results (Report October 1899, 21, and April 1900, 14). 



