Unlike the originally crystallised eudesmol, these sublimed composite needles did not polarise at all well, but if melted, and again allowed to crystallise, they showed the same peculiarities of crystallisation and polarisation as in the case of the original eudesmol under similar conditions. Radiated masses of crystals in all cases gave well defined dark crosses similar to those shown by uniaxial crystals when observed under crossed nicol prisms.

The melting point of the sublimed crystals, derived from the liquid form, was some degrees lower than that of the original eudesmol; but in this respect there was little uniformity, as those which sublimed at the top of the covering dish, or furthest from the subliming material, melted at 74-76° C, while those formed nearest to it melted at 65-72° C. The slight amount of impurity present in the sublimed crystals was thus sufficient to lower their melting point, because when the whole was purified by recrystallising from alcohol, the melting point was that of the original eudesmol, 79-80° C.

The liquid residue, separated from the acid bodies as described above, was therefore shown to be eudesmol, and this conclusion was also supported by the formation of the acetate in almost theoretical amount on acetylation, as well as by the character and behaviour of the crystals formed by sublimation, and their microscopic appearance.

That eudesmol occurs naturally in the liquid form in some Eucalyptus oils is perhaps best shown with that of E. Rossii, and in no case has crystallised eudesmol been detected in the oil of this species under natural conditions

In the first edition of this work, page 71, we describe the oil of E. Rossii from two localities in New South Wales (Bathurst and Bungendore), and show the presence of a highly dextro-rotatory constituent in the higher-boiling portions of both, that from Bathurst being aD + 19.19°, and the other almost as high. These samples were mixed and stored in the dark, and twenty years afterwards the oil was again analysed. The highly dextro-rotating constituent, when finally distilled under reduced pressure at 10 millimetres, slowly crystallised and was found to be eudesmol melting at 79-80°, and to have specific rotation [a]D + 33.58° in a 12 per cent. chloroform solution. It is thus seen that the liquid form of eudesmol is an exceedingly stable substance, particularly as it was somewhat readily separated in the crystalline condition after the lapse of twenty years.

We have so far detected either the liquid or crystallised eudesmol in the oils of forty-two species of Eucalyptus, and have prepared the crystallised form in the pure condition from several. It is a coincidence perhaps, that in twenty-one

of these it was associated with phellandrene, whilst the other twenty-one did not contain that terpene. In the following lists the two classes are separated for purposes of comparison: -

*Although E. globulus is included in this list, yet, Schimmel & Co. (semi-annual report, April, 1904, 51), suggest that this substance (eudesmol) differs in certain directions from ordinary eudesmol, and Semmler proposes for it the name globulol. It was thought that similar relations existed between eudesmol and globulol as between borneol and iso-borneol.

Much Eucalyptus oil, however, supplied commercially under the name of Eucalyptus globulus had not been distilled from that species, and this difficulty was early recognised by Bourchardat and Tardy during their investigations on the terpene of E. globulus. [Compt. Rend. 120 (1895), 1417.]

The oils of the following Eucalypts appear to contain only the liquid form of eudesmol under natural conditions: -