NEW PHILIPPINE ESSENTIAL OILS. 



341 



potassium carbonate was filtered with suction and the alcohol distilled in vacuo 

 until almost all of it had passed over. On adding water and extracting with 

 ether, no oil, biit only a tar was obtained. The action of alkalies in promoting 

 the formation of tar was also noticed when the substance was tested for alcohol, 

 as described above. The first cubic centimeter of the distillate contained minute 

 drops of an oil with an odor like that of carvone, but the alkaline solution very 

 quickly darkened and oil drops ceased to pass over. 



Because of the results noted above, several attempts were made to sjDlit 

 the compound with dilute acids. 



Three grams of the ketone were warmed to 60° with 50 cubic centimeters of 

 60 per cent alcohol containing 2 drops of concentrated sulphuric acid. After 

 the crystals had dissolved, 0.2 gram of the polymerization substance were obtained 

 by filtering. The solution was neutralized with sodium carbonate, sodium acetate 

 added, and the whole divided into two portions which were treated respectively 

 with phenylhydrazine acetate and semicarbazid hydrochloride. The first solution 

 yielded less than 0.1 gram of a crystalline phenylhydrazone contaminated with 

 tar, and the second deposited about the same amount of a semicarbazone which, 

 when recrystalized from dilute alcohol, melted at from 186° to 187°. The 

 quantity obtained was too small for analj-sis and the yield was so poor that the 

 experiment was not repeated. 



The few drops of oil resulting from the several saponification experiments 

 M-ere tested with phenylisoeyanate and with 1, 3, 5-dinitrobenzoyl chloride, but 

 no crystalline product was obtained. 



It is clear from the foregoing results that the crystalline substance 

 contains one ketone group, no free hydroxyl radicals, and one succinic acid 

 group. One carboxyl group of the acid is combined as the ethyl ester 

 and the other attached to a group which is probably a ketone alcohol 

 of the formula CioHijOg, the latter radical evidently containing the 

 group • HC : CH • g : 0. 



These considerations suggest the constitution. 



CH3C 



CjH.C 



C • C2H, • CO2 • C^Hg 



However, this representation does not agree with the empirical formula. 



If, instead of an isopropyl group the group — C is assumed to be 



\CH3 

 present, all the conditions are satisfied. Further work will be necessary 

 before this can be settled definitely. 



