— 136 — 



On the other hand, Morrell 1 ), proposes an improvement on Armstrong's 

 test for detecting adulteration with petroleum (-distillates), for the detection 

 of which he relies chiefly upon the results of fractional distillation. This 

 consists in treating the reaction-mixture with steam after the action 

 of the sulphuric acid. The steam separates the unattacked petroleum 

 from the polymerised terpene-mixture and is said to be especially effec- 

 tive in the case of inseparable emulsions of oil and acid. The analytical 

 deviations observed in known mixtures with petroleum (d 0,793 to 0,809; 

 flashpoint 23,3 to 34,4°) did not exceed 2°/o. 



We are not quite clear as to what there is new in Morrell's mode 

 of applying Armstrong's test, for Armstrong himself has already proposed 

 steam-distillation of the reaction-products 2 ). 



According to Coste 3 ) the use of sulphuric acid of the strength indicated 

 by Armstrong is not admissible when testing turpentine oil for an admixture 

 of hydrocarbons, because sometimes the entire mixture remains unattacked. 

 The best method is to determine the constants and the iodine number of 

 the oil, distilling, if necessary, by steam, and separating the various com- 

 pounds by removing the terpenes and part of the olefines by polymerising 

 with dilute sulphuric acid. The remainder of the olefines is separated out 

 with strong sulphuric acid, and the aromatic portions are nitrated or 

 sulphonated with fuming sulphuric acid. 



The results 4 ) which have thus far been obtained with the sulphuric 

 acid method by different investigators (and which are for the greater part 

 unsatisfactory) have been brought together by Eibner and Hue 5 ), who have 

 tried to devise an improvement of the test. In order to attain this it 

 is first of all necessary always to state the strength of the fuming 

 sulphuric acid used, as the strength of commercial fuming sulphuric 

 acid varies greatly. The authors next found that the separation observed 

 by Herzfeld of only 1 to 2°/o residue from pure turpentine oils was 

 abnormally low and was never reached by them. Another source of 

 discrepancies was found in Herzfeld's method of water cooling, but 

 this could be overcome by adding the acids in a suitable manner. Eibner 

 and Hue next ascertained how much concentrated and, afterwards, 

 how much fuming acid (with an anhydride content of about 20°/o), is re- 

 quired in order that a definite percentage of residue should not be ex- 

 ceeded. It was found that the treatment of 10 cc. turpentine oil with 



2 ) Journ. Soc. chem. Industry 29 (1910), 241. 



2 ) Comp. Chem. Zentralbl. 1883, 206. 



3 ) Analyst 35 (1910), 112; quoted from Chem. Zentralbl. 1910, I. 1297. 



±) Comp. Report April 1904, 85; April 1905, 77; October 1905, 70; April 1906, 66; 

 October 1906, 75 to 77; November 1908, 122 to 125; April 1909, 90; October 1909, 125; April 

 1910, 109, 115, and the preceding abstracts. 



*) Chem. Ztg. 34 (1910), 643, 657. 



