178 Dreiv, Green and SL John.- 



inoidentally, our own experiments on these stringybarks have shown 

 that each species yields its own characteristic oil, which varies less 

 than any of the other features usually relied on by morphologists for 

 the identification of botanical species. 



We shall therefore describe in some detail our experiments on 

 the oil. 



Foliage was collected from four different trees, and in each ca.se 

 one or more sacks of the leaves were distilled with steam in an ex- 

 perimental still at the University Agricultural Chemistry Laboratory. 

 The 400-gallon tanks usually employed in the commercial extraction 

 of eucalyptus oils hold about half a ton of leaves. Our sitill had a 

 capacity of some twenty pounds only, but the processes were the same 

 in principle in both cases, except that we were able to carry out our 

 operations quantitatively. 



The distillation was generally considered to be complete after three 

 and a-half hours. In one experiment the following measurements 

 were noted : — 



Time of distillation (hours) 12 3 4 



Oil distilled 49 cc. ... 18 cc. ... 9 cc. ... 1 cc. 



Eucalyptus oils, in general, are mixtures of pinene, phellandrene 

 and eucalyptol, liquids boiling at 156 deg., 172 deg. and 176 deg. 

 respectively. They may contain also small percentages of less vola- 

 tile complex aldehydes and a sesquiterpene (aromadendrene), with 

 traces of more volatile alcohols, aldehydes and esters. 



The physical properties of an oil will naturally depend on the 

 variety and proportions of the ingredients present, and are often, 

 though not always, a useful guide to its composition. 



A careful fractional distillation enables us, partially, at least, 

 to separate the different constituents and form an approximate 

 estimate of the amount of each present. It should also be a pre- 

 liminary to the (pmntitative determination of eucalyptol, aldehydes, 

 etc. 



We have for this purpose used a glass still of definite dimensions 

 sealed on to a rod and disc fractionating column so that a definite 

 degree of fractionation will always be obtained. If the quantity of 

 distillate passing over between definite temperature limits be plotted 

 in the form of a temperature-percentage curve, we have a very con- 

 venient method of gra])hically comparing different oils. 



In dealing with an unknown oil, refractionatimi is resorted to, and 

 derivatives of each coinponcni pi-epared. and identified by their 

 physical constants. 



