ANALYSIS, REFINING, AND COMPOSITION. 



27 



would then be all heavy oils. These heavy oils might be changed by 

 the heat, either more or less than the terpenes were in figures 11 to 13, 

 but there is really no necessity for determining this change accurately 

 because a small change in the physical properties does not effect the 

 interpretation of the curves. For instance, if in figure 2 the gravity 

 of the residue had been 0.930 instead of 0.945, the composition of the 

 turpentine as interpreted from the curves would still have been 95 per 

 cent pinene and 5 per cent heavy residue ; that is, a gravity of 0.93 is 

 so much higher than that of the terpenes that the presence of a con- 

 siderable amount of terpenes in the residue is excluded; and the only 

 possible interpretation of the composition of the residue is that it 

 does not contain ter- 

 penes in considerable 

 quantities. 



The change in com- 

 position of a turpen- 

 tine during analysis 

 by direct distillation 

 is, therefore, without 

 effect on the interpre- 

 tation of the results, 

 except in case there is 

 no high boiling and 

 heavy residue present 

 in the original mate- 

 rial; and in -this case 

 the approximate effect 

 is known and the re- 

 sults can be corrected. 



CHOICE OF A DE- 

 PHLEGMATOR. 



160 165 no ns ie&> eoiuNe POIN 



Kb XI J&& JB9 90 SPECIFIC CAVITY A 

 INDeX OF EEFEACT10N f M470 1412 WT4 1416 1418 U80 



FIG. 14. Comparison of fractionation with Le Bel-Hen- 

 ninger and Ilempel column dophlegmators. 



The Le Bel-Hennin- 

 ger dephlegmator has 

 been used with good 

 results in many turpentine analyses; it is efficient in fractionating 

 power, and when properly constructed gives no trouble by filling up 

 with liquid; the only objections to it are that it is expensive and 

 fragile. The simplest, least expensive, and least fragile dephleg- 

 mator is of the Hempel column type, and several different sizes of 

 Hempel columns were tried to find if a practicable sized Hempel 

 would give as good separations as the Le Bel-Henninger. It was 

 found that the Hempel column, shown in figure 1, gave slightly 

 better separations than the five-bulb Le Bel-Henninger and a com- 



