Of course, exposure on a wall facing nearly south constituted a very severe test, yet the circulation of air In the tubes was more advantageous to the pigments than would have been the steamy heat of a closed vessel, or even of an ordinary paper-backed picture-frame. But on the other hand, this arrangement allowed the free access to the pigments of any noxious gases, such as sulphurous and sulphuric acids, and sulphuretted hydrogen, which might have been at any time present in the atmosphere. The general results of this first series of trials are gathered in the following table, the exposure in all cases asting from May, 1886, until March, 1888. The pigments are arranged in the order of instability, the most fugitive being placed first:

*Carmine. *Crimson lake. *Purple madder. *Scarlet lake. *Payne's grey. *Naples yellow. *Olive green. *Indigo.

*Brown madder. *Gamboge, *Vandyke brown. *Brown pink. *Indian yellow.

Cadmium yellow.

Leitch's blue. *Violet carmine. *Purple carmine. *Sepia.

Aureolin.

Rose madder.

Permanent blue.

Antwerp blue.

Madder lake.

Vermilion.

Emerald green.

Burnt umber.

Yellow ochre. Chrome yellow. Lemon yellow. Raw sienna. Indian red. Venetian red. Burnt sienna. Terre verte. Chromium oxide. Prussian blue. Cobalt. French blue. Ultramarine ash.

Show no change.

The pigments marked with an asterisk were found to have distinctly altered either in depth or hue by a much shorter exposure, from May to August 14,1886.

In a second series of trials the tinted papers were dried, and then introduced into the tubes, which had been previously heated; the specimens were then sealed up hermetically; as no moisture-absorbing material was enclosed with the papers, traces of water must have been present. My own much earlier results were abundantly confirmed by those obtained in this series, for the number of pigments which proved to be permanent under these conditions was double that of the first series. Brown madder and Prussian blue were, however, acted upon in this second series. Dr. Russell and Sir W. Abney make the remark that of the eight colours which remained unchanged in dry air, but were acted on in ordinary air, all, with the single exception of madder lake, are mineral colours. But this is not correct, for the pigments named are - madder lake, olive green, Payne's grey, sepia, Naples yellow, cadmium yellow, emerald green, and burnt umber, and of these the first four are either wholly or partly of organic origin, while the seventh contains an acetate.

In the next series of experiments, the pigments were exposed in the presence of moisture-laden air. Very few colours withstood this test - none of organic origin; both Prussian blue and Antwerp blue were entirely destroyed.

An atmosphere of moist hydrogen gas was employed in the fourth series. Under these conditions carmine, crimson lake, madder lake, brown madder, olive green, indigo, Payne's grey, sepia, and Vandyke brown, suffered no change.

When, as in the fifth series, both moisture and oxygen were excluded, scarcely any even of the most fugitive pigments were affected. Vermilion, however, as in all the other experiments, became black. We know that the reason for the change is physical, not chemical.

In the sixth series it was proved that the addition of ox-gall had no beneficial effect in lessening the change of hue and tone in fugitive pigments.

The remaining series, save the twelfth and last, were devised in order to learn what influence upon the stability of pigments might be exerted by admixture with Chinese white, by exposure to the light of the electric arc, by heat without light, by heat and light together, and by exposure to the light transmitted through coloured glasses. Amongst the results recorded, we may note the decided changes in several pigments caused by admixture with Chinese white, and by heating the prepared paper slips in sealed tubes for seven hours a day for three weeks, all light being excluded. In a twelfth series of trials, the pigments were exposed in a picture-frame under glass in such conditions, and to such an amount of light, as might be taken to represent the ordinary circumstances in which pictures are kept. The frame was exposed from August 6, 1886, until May 6, 1888, to very bright light, but not to sunshine. Gamboge, indigo, Naples yellow, brown pink, carmine, and Vandyke brown, had faded in varying degrees.

Some remarks on these results will be found further on in the present chapter; they are of extreme importance, considering the large use that has been made of these pigments by our water-colourists, and the mild treatment to which they were subjected during the short period of twenty-one months.

For the results obtained with mixtures of pigments under varied conditions of exposure, we must refer our readers to the report itself. It may, however, be remarked that, in the great majority of cases, the changes of tone and hue which occurred were such as might have been predicted from the known degrees of stability of the several constituents of the mixtures. Here, as elsewhere in the report, we find frequent mention of the strange, but long known, recovery in darkness of its colour by Prussian blue which has been bleached by sunlight.

In the fourth appendix to this report is an instructive list of the pigments employed by some of the most distinguished artists using water-colours. Forty-six painters replied to the invitation of the Science and Art Department; from their answers we learn that a large proportion of them include in their palettes many pigments which must be unhesitatingly condemned on account of their want of stability. Thus no less than seventeen out of the forty-six artists who responded to the appeal employ three of the most fugitive pigments in the series - namely, gamboge, brown madder, and indigo. Converted into percentages, we may say that 37 out of 100 painters in water-colours use these three untrustworthy pigments, besides others which are worse, and others which are little better; of course, they employ also certain colours as to the stability of which there is no question. The following tabular statement gives the proportion of artists, per 100, who use the eleven perishable pigments named below:

Gamboge ...

... 70 ...

Faded to 7.

1 represents the lightest tint, 8 the darkest. The washes of pigment were fully exposed for twentytwo months.

Indian yellow

... 24 ...

Faded to 6.

Vermilion ...

... 70 ...

Gone black.

Carmine ...

... 8 ...

Gone.

Crimson lake

... 22 ...

Gone.

Purple madder

... 28 ...

Faded to 8.

Brown madder

... 74 ...

Faded to 3.

Brown pink

... 11 ...

Faded to 7. Gone.

Vandyke brown

... 74 ...

Sepia ....

... 65 ...

Faded to 8.

Indigo ....

... 52 ...

Faded to 8.

We call these pigments perishable with good reason. For, according to the report under review, all of them were found to have faded, materially and conspicuously, after twenty-two months' full exposure in a south aspect, while three of them had entirely disappeared, and another (vermilion) had become black. But this is not all. For under a less severe trial (p. 45) - namely, exposure for the same time, not to direct sunlight, but to a very bright light from a window, 'under conditions approximating to those to which pictures are usually subjected' - six out of the eleven pigments had faded, though in varying degrees. With these figures and results before us, it is impossible to resist the conclusion that a life of 100 years is too much to allow to many of the water-colour drawings of the present day. What shall we say, then, as to the stability of the works of the earlier masters of the English water-colour school? How much care in the exclusion of 'the more fugitive colours' was taken by the water-colourists of 1780 to 1850? Could it be honestly said of any large number of such works, in which gamboge brown pink, crimson lake, sap-green, indigo, and sepia, were generally employed without stint, 'that about a century of exposure would have to be given to water-colour drawings in galleries lighted as are those at South Kensington before any marked deterioration would be visible in them'? * (Report, p. 46.)