The testing of pigments for genuineness and for purity-has been discussed incidentally in Chapters XIII. to XIX. of the present volume. Though genuineness and purity* have often an important bearing upon the question of permanence, this last quality must be ascertained by independent experiments of another order. The study of old paintings often furnishes us with results of some value, the results of unintentional testings. But for the purpose of securing complete and wholly trustworthy information, we must know precisely all the materials and all the conditions of our trials. Not only must the painting-grounds, the mediums, and the pigments, be chemically examined, but we must be in a position to state the character of the atmosphere in which they have been exposed, and the nature and amount of the solar or other radiations to which they have been subjected. In the great majority of these trials accurate data as to materials and conditions are wholly wanting; even the South Kensington report affords us no information as to the composition of the pigments employed, nothing more than their commercial names, so that we have to take on trust their genuineness and purity.

However, in this same most important series, quite unusual, if not unprecedented, care was taken in order to determine the conditions, physical and chemical, under which the pigments were tested. In my own experiments, carried on between 1856 and 1879, in somewhat desultory fashion, and extended and improved since 1880, the influence of purity of sample, of the presence of moisture and of oxygen, and of the nature of the 'light' has not been neglected; the full details of the methods adopted, and of the results obtained, could not be appropriately introduced into an elementary manual. Mention will be made of the chief conclusions reached in the present chapter; there are also numerous references to them in Chapters XIII. to XXII.

* The chromatic values of pigments - their approach in hue, etc. to recognised standards of excellence - are not here taken into account. Such data may be obtained by the use of Lovibond's Tintometer when once the chromatic elements of a pigment in terms of the degrees of the standard glasses employed in this instrument have been determined. But really exact determinations of this kind require complex scientific apparatus, which cannot be profitably used except by an expert manipulator.

In many early treatises on painting we find observations as to the degrees of stability possessed by various pigments, along with suggestions as to methods of treatment by which in certain cases greater permanency may be attained. To some of these observations and suggestions reference has been made in those chapters of the present volume which deal with pigments; many of the remainder are now without practical importance, referring as they do to pigments which have been rightly discarded.

It is only of recent years that trials of pigments have been made with any approach to exactness. But in the majority of cases no information has been secured concerning the purity and genuineness of the pigments with which the experiments have been made. I have not been able to find a single chemical analysis of any one of the pigments tested. The chromatic analysis of the light they severally reflect has, however, been recorded in the case of the water-colour paints examined by Dr. Russell and Sir W. Abney, who have likewise furnished some particulars as to the intensity of the solar radiations to which the pigments were subjected.

More than a century ago Sir Joshua Reynolds tested, in a rough way, the stability of some of the paints he employed. Two causes detract from the value of his results. In the first place, the information furnished concerning the nature of many of the pigments he tried is too imperfect to be of any use; secondly, we are not acquainted with the conditions under which his specimens have been kept during the whole period since they left his hands. Two of his trial canvases are preserved in the Royal Academy. Although the darkening and embrowning of the surface are general, and though the names of the pigments employed are often undecipherable or meaningless, yet something useful remains. From the experiments made in 1772 we may gather the following facts: The proper hues of several pigments remain in a measure - orpiment, or kings' yellow in crystals; yellow lake, with wax and drying oil; gamboge, with lake and Venice turpentine; gamboge, with (Venice) terpentine; prepared gamboge, with wax; and verditer, with varnish.

On the other hand, gamboge with oil, lake with oil, and many other pigments of organic origin, when unmixed with wax or varnish, are names only, or, at the most, brown discolorations.

M. J. Blockz, in his 'Peinture à l'Huile,' gives the results of a number of experiments made by M. J. Dyck-raan. He condemns, for various reasons, not only asphalt, but also terre verte, cobalt green, emerald green, raw sienna, raw umber, ivory brown, and all burnt madders. Cassel earth was slightly changed; brown ochre, burnt sienna, Mars brown, ivory black, and vine black, proved to be permanent. His lists of incompatible pigments are somewhat unnecessarily extended, not being justified, in all particulars, by further and more careful experiments. The experiments of the late F. W. Andrew, formerly of the Victoria and Albert Museum, have been carried on since 1876, but have been confined to water-colours, both moist and cake. His chief results will be found recorded along with my own, in Part III., in the paragraphs devoted to the consideration of the several pigments. His water-colour washes, generally spread on Whatman paper, or Whatman board, were exposed for periods varying from twenty-eight to eighty-two months, in a south window, to all the sunshine available; half of each wash was doubled back, and so far excluded from light. In some cases a third part of the coloured slip was exposed to the air and light without the protection of glass.