This substance - that is, the entire mass of the white-lead ground in oil which has become solid, tough, and hard - seems to contain a small percentage of a lead-soap, formed probably out of the free fatty acids of the linseed-oil. But whatever the complete explanation of this hardening action may prove to be, there can be no doubt that we must attribute to the simultaneous presence of oil and the hydrato-carbonate of lead the preservation of the continuity of surfaces of the whites, and of the pale tints into which white lead enters, in many an old picture. No other pigment in common use is capable of solidifying the admixed oil to anything like the extent that characterizes white lead. Now there are modern preparations of white lead made chiefly by precipitation or the 'wet way,' which produces a pigment containing little or no lead hydrate. Some writers on pigments advocate the use of these newer products. 'Why,' say they, 'should you carefully exclude from your pictures oils, and varnishes, and siccatives which contain lead in solution, and then introduce the same or a like substance in your white lead ground in oil?' Many years ago I tried to answer such a question as this by means of experiment.

I was actuated by a desire, based on theoretical considerations, of preventing altogether the formation of lead-soaps. I tried comparative experiments with zinc oxide, pure lead carbonate, and the Dutch-made lead hydrato-carbonate, or ordinary flake-white. The two lead pigments (with which alone we are now concerned) were washed thoroughly with distilled water and dried before being ground in linseed oil. The oil-paints thus prepared were spread in duplicate series upon glass, paper, and primed canvas; one set was kept in a dark box. the other was exposed to strong light. So decided was the superiority of the ordinary flake-white over the pure carbonate, when both series of specimens were examined after the lapse of various intervals of time, that I was reluctantly compelled to abandon my recommendation of the latter. Ease in working, solidity of body, and rapidity of drying, were not the only points of superiority; for the films of paint, after having been kept a year, showed differences in hardness and in smoothness of surface which were all in favour of the hydrated carbonate. No discoloration was observed in the specimens exposed to light, except in the case of the pair upon paper; the absorbent ground had withdrawn some of the protecting oil, and both specimens had equally darkened.

In darkness all the specimens had become of a somewhat greyish yellow, the discoloration being about equal in all the pairs, the pair spread on paper having, as in the previous case, become darker than the others. The late Mr. G. W. Wigner tried a somewhat similar series of experiments, and came to the same conclusions. I should add that these deductions were corroborated by the results of other trials, in which numerous permanent coloured pigments mixed in pale tints with these two lead whites were treated in the same way. If, however, we feel bound to recommend the ordinary flake-white instead of pure lead carbonate, that recommendation does not prevent us from excluding lead-containing oils from our pictures, seeing that we possess perfect substitutes for them, and that there is no reason for thus multiplying the causes of possible change.

Before commenting on some of the lessons to be drawn from individual pictures, it may be desirable to make a few observations on some of the changes frequently observable in old illuminated manuscripts and choral books. The tarnishing of lead and copper pigments laid on without any protection but that of gum is very frequently seen. The darkening of vermilion is apparently capricious,* but is really explicable in part by the substitution of red lead for vermilion, and in part by the molecular change which the latter pigment is known to suffer, and which has been already described. Ultramarine always stands out absolutely intact; sometimes it acquires extraordinary prominence by reason of every other pigment on a page having altered. The red cochineal and kermes lakes have either gone or become paler and brownish. Sometimes fruits painted in vermilion have been shaded or dotted with a crimson lake, but the latter has disappeared, leaving nothing but a slight gummy appearance upon the scarlet ground. Blue flowers painted in smalt and veined with indigo show scarcely a trace of the latter pigment. Verdigris, which is partly soluble in water, has run and discoloured the vellum, and at the same time has acquired a brownish hue. Sap-green, from buckthorn berries, has faded greatly.

Lilien-grün (of the seventeenth century), from the flowers of Iris germanica, has disappeared.

We now cite a few pictures, out of a large number which have been studied for the purpose of observing the present state of the materials which have been used in their production. We begin with some works in the National Gallery, Trafalgar Square.

* An initial letter in vermilion, painted in the fifteenth century, and perfectly unchanged, became black by one year's exposure to sunshine.

Margaritone Di Magnano (1216-1293)

No. 564. In the very limited palette of this early painter in tempera we note that the vermilion, a yellow earth, lamp-black and a puce colour are well preserved; the last-named pigment may be a form of iron oxide, and corresponds in hue to the artificially prepared oxide called 'Mars violet.'

Giovanni Da Milano (Late Fourteenth Century)

No. 579A. The crimson on the robes of two of the three figures which occupy these panels seems to be derived from madder, and is well preserved.

Jan Van Eyck (1390?-1441)

No. 186 (dated 1434). The green robe in this famous picture shows distinct cracks, which differ in character and are larger in size than any others in the work. The flesh-tints are perfectly preserved as to texture. I suspect that verdigris has formed a constituent of the green paint employed.