— 125 — 



In a note on the determination of petroleum in oil of turpentine, 

 published as a supplement to a paper read before the Society of 

 Public Analysts, J. H. Coste 1 ) contradicts the statement of Turner 

 that Armstrong's method of testing oil of turpentine (consisting of 

 treating first with diluted and then with strong sulphuric acid) 2 ) gives 

 under certain circumstances values which are by 2O°/ too low. Turner 

 had made his experiments with a mixture of oil of turpentine and 

 kerosene, but Coste points out that kerosene is a mixture of petro- 

 leum spirit, which is volatile under steam, and of non-volatile petro- 

 leum oil, whereas Armstrong's method especially applies to petroleum 

 spirit — a fact overlooked by Turner. The latter's process of treating 

 the upper layer of petroleum with strong sulphuric acid is more thorough 

 than the method described above, and will probably only give accurate 

 results when sources of possible error are mutually compensated. Ex- 

 periments made according to Turner's method on mixtures of petro- 

 leum spirits and oil of turpentine, both of known physical properties 

 and in different proportions, left an intact oil -mixture of higher 

 refraction and density than those of the component parts; and it is 

 surmised that terpinene and cymene were present as conversion-products 

 of the terpenes. When the oil sample is freed from petroleum oil 

 by steam distillation, and when for the rest Armstrong's original pre- 

 scription is carefully followed, his process gives good results. 



In the following discussion, Nash maintained that he was able to as- 

 certain the quantity of petroleum spirits added by simply shaking known 

 quantities of the adulterated oil of turpentine in a graduated tube with a 

 definite quantity of air. According to his observations the height of the 

 layer of froth which forms on an oil containing petroleum during the 

 manipulation of shaking, stands in direct relation to the petroleum content. 

 Pure oil of turpentine gives no froth. Of 14 tests recorded 9 agreed, 

 the other 5 showed more or less pronounced deviations. Tatlock, who 

 was in the Chair, asked whether other additions, such as spirits of 

 rosin, could be determined by this method, to which Nash replied 

 that he had not examined mixtures with spirits of rosin but only 

 mixtures with petroleum spirits. Spirits of rosin, he thought, would 

 hardly be used as an adulterant. Coste expressed his surprise at the 

 agreement in the figures obtained by Nash. According to him the 

 method proposed by Herzfeld 3 ) of direct treatment with concentrated 

 sulphuric acid, although based upon false premises, apparently gives 

 approximately correct results. For accurate testing also Armstrong's 

 process is very useful. Adulterations with petroleum are only very 



x ) Analyst 34 (1909), 148. 



2 ) Gildemeister and Hoffmann, The Volatile Otis, p. 246. 

 8 ) Zeitschr. f. off. Chem. 9 (1903), 454; Report April 1904, 85; October 

 1906, 78. 



