Besides these observations another will have been made - different specimens of spirit of turpentine will have been found to differ much as to the rate at which these changes have taken place. Some samples, even in half-full bottles, remain clear and limpid for long; others become thick, opaque, and sticky in a few weeks. Such changes are undesirable in a solvent, diluent, or painting medium, on many grounds. The resin formed is an unsatisfactory one - soft, sticky, and contractile. The liquid decreases so greatly in mobility, and increases so greatly in viscosity, that its utility in thinning oil pigments, and in making fine touches, is greatly impaired. And this thickening of the liquid is accompanied by the production of acid substances and of water, which affect injuriously the ease of working and the stability of the picture. Spirit of turpentine should disappear by evaporation quickly and completely from the painting into which it has been introduced. Now, if it be easily oxidizable, even if it be kept from experiencing change before it is actually employed, it will, during the very time in which it is being used, attract oxygen; so that though a great part of it will escape by evaporation, the remainder will resinify on the canvas itself, adding a sticky deposit to the drying oils and hard resins which may have been used as the painting medium.

It is clear, from all the above considerations, that the greatest care ought to be taken in selecting, in the first instance, such a sort of spirit of turpentine as will resist oxidation under ordinary conditions. Even an inferior spirit may be used, with a minimum of disadvantage, if immediately after distillation it be poured into a number of small bottles, so as to fill each of them completely; they should be at once closed with sound corks. In this way the contents of a bottle may be used up very soon after it has been opened. Another precaution may be taken: A few small fragments of hard quicklime may be placed in each bottle to absorb any moisture produced by oxidation, and also the acid bodies which are formed at the same time. Even with the choicer samples of spirit of turpentine, which pass much less easily into resins, this precaution is desirable; but in this case the employment of many small bottles is unnecessary, and it will suffice to put a few hard pieces of lime, free from powder, into a pint or quart bottle, and then to fill it with the spirit.

The clear liquid may be poured off as required for use, any disintegrated particles of lime sinking readily to the bottom of the vessel.

Before giving details as to the sources and characteristics of the best terpenes, it may be useful to mention that commercial samples of spirit of turpentine may be tested and compared by means of a very simple experiment. Obtain the required number of small flat-bottomed, conical glass flasks with wide mouths, one flask for each sample; these flasks are known as Erlenmeyer's. Into the flasks pour enough of the several samples to cover the bottom to the depth of one-eighth of an inch; label each flask to correspond with the sample, and lightly close each mouth with a plug of carded cotton - the date of the experiment should be added on the label. Shake each flask so as to cause a number of bubbles to be formed in the liquid; the more rapidly these bubbles break, the better is the sample. Repeat the experiment of shaking the samples at short intervals for a few weeks - notable changes in the viscosity of the oils will be observed sooner or later. Any sample which after one month remains clear, and in which the bubbles formed on agitation break almost as quickly as at first, may be accepted as of good quality.

Another test for discriminating between the samples, so far as their state at the time of the experiment is concerned, is the very simple one of placing one drop of each oil upon a sheet of writing-paper, and gently warming the translucent stain it forms; with a good oil the mark completely disappears.

Two other obvious characteristics of different samples of spirit of turpentine may now be noticed - namely, odour and boiling-point. Some samples have a much more agreeable scent than others; the vapour of these seems to have a less marked tendency to produce headache than that of the pungent and cruder-smelling varieties. The range in boiling-point is not very extensive; but it may be taken as about 25° C., the figures ranging from 155° to 181.° Samples having lower boiling-points evaporate more quickly than those which enter into ebullition at higher temperatures. The solvent power on resins differs with different kinds; this is a property which is of importance in varnish-making, but very little accurate knowledge exists on this point. But there is one characteristic almost peculiar to the terpenes which must not be overlooked. When oxidizing they possess, unlike the alcohols and benzene and petroleum spirit, a power of starting or increasing the absorption of oxygen by linseed or other drying oils; in fact, they act as siccatives.

This property is constantly utilized in oil-painting; probably it is connected with the formation of hydrogen peroxide which passes on its surplus atom of oxygen to the oil present.

A few of the more important turpentine oils may now be named:

American oil, chiefly from Pinus palustris and P. Tæda.

Austrian oil, partly from Pinus Laricio, partly from P. Pumilio. Burmese oil, from Pinus Khasya. French oil, from Pinus Pinaster (= P. maritima). German oil, from Pinus sylvestris, P. Cembra, P. Abies, P. vulgaris, etc. Juniper oil, from Juniperus communis. Russian and Swedish oils, chiefly from Pinus sylvestris and P. Ledebourii. Strasburg oil, from Abies pectinata.

The above-named turpentine oils are accompanied by small amounts of various resins, camphor, and other oxygenated bodies, from which they may be separated by treatment with caustic potash, metallic sodium, and fractional distillation.