The chief paints which were unaffected, at all events, so far as some specimens were concerned, by the exposure, are included in the following list: Yellow ochre, raw sienna, deep cadmium, Mars red, light red, Indian red, oxide of chromium, Leitch's blue, cobalt, artificial ultramarine, Prussian blue, raw umber, burnt umber. Naples yellow (true) became blackish in darkness, but was unaltered in light, while orange chrome showed dark patches. Further details must be given as to the pigments which were affected. In the tabular statements appended a selection from the results afforded by eleven sets of experiments is recorded, the letters C and M prefixed to the entries respectively denoting the employment of cake or moist colours. The numerical values representing the residual hues are rough approximations only, but in some instances they were determined by means of comparisons with standard coloured liquids, contained in glass cylinders graduated into ten equal parts. Hellige's colorimeter or the tintometer of Mr. J. W. Lovibond, of Salisbury, may be employed with advantage in these estimations.

Changes In Water-Colours

The pigments containing lead, such as the ordinary chromes, and those having a copper basis, like emerald green, had altered capriciously, losing part of their original colour, and becoming tarnished or embrowned in patches. Brown pink faded like the yellow lakes, but acquired a bluish-grey residual hue.

The testing of the madder colours is so important that I introduce here a few additional experiments selected from my own note-books. The washes of the moist-colour paints were, as far as possible, of the same depth of tone, and they were all exposed together in a glazed frame to one year's sunshine:

In contrast to the above results with madder carmine, the following experiment with the ordinary carmine (prepared from cochineal) is instructive. On a sheet of Whatman paper, a space of 10 inches in length by 4 inches in width was covered with a uniform wash of the moist paint, having a depth of tint about equal to that of the petals of the old China rose. This coloured strip was then subjected to summer sunshine in such a way that successive single inches of its length received the light (during the same hours of similarly bright days) for periods of 2, 4, 8, 12, 20, 26, 30, 40, and 100 hours, one single inch at one end being, however, protected completely from all access of light. The exposure of 100 hours sufficed to bleach the last breadth completely, but had the rate of fading been in a simple arithmetical progression, a much shorter exposure would have sufficed. In fact, the bleaching action was far more energetic during the first period of two hours than during the second, about 20 per cent. of the original colour having been destroyed during these two first hours, while during the second equal period the loss of depth did not exceed one-tenth of this amount.

Moreover, it was noticed that the change of hue consequent upon the first exposure was different in kind to that which occurred subsequently.

Professor O. N. Rood's Experiments. - In his 'Modern Chromatics,' pages 90 and 91, Professor Rood gives the results of a few trials which he made as to the effect on washes of water-colours laid on ordinary drawing-paper of three and a half months' exposure to summer sunlight.

These pigments were unaffected:

Cadmium yellow, yellow ochre, Roman ochre.

Indian red, light red, Jaune de Mars.

Terre verte.

Cobalt, French blue, smalt.

Burnt umber, burnt sienna.

The following pigments were all affected. The sequence represents the amount of alteration, the list commencing with those colours which suffered but little change:

Professor Rood adds that rose madder, brown madder, and purple madder were all a little affected by an exposure to sunshine for seventy hours, and that pale washes were completely obliterated by a much shorter exposure to sunshine in the case of carmine, dragon's blood, yellow lake, gall-stone, brown pink, Italian pink, and violet carmine.