By several different methods, data may be obtained which enable us to classify pigments - roughly, it is true - in accordance with their varying degrees of stability. Such data are derived partly from the known chemical and physical constitution of the various substances; partly from a study of old paintings and drawings in which they have been used; and partly from special experimental tests of permanency to which they have been subjected. Selections from these data are given in Chapters XX., XXIV., and XXVI., of the present work; but much additional information has been furnished by other trials, conducted by the author and other experimenters, for which space could not be found in this volume. Tables constructed from such data must not be regarded as affording exact values, but merely approximations. From some minute and often obscure cause differences of deportment, under exposure to hostile influences, will occasionally be observed in the case of two specimens of the same pigment having the same hue. And, further, the grouping of pigments into a small number of classes is a conventional and convenient arrangement which cannot accurately represent the numerous degrees of stability or instability which characterize the several pigments under discussion.

For when we leave the practically unalterable mineral pigments, we have to deal with a number of preparations which fall by irregular and often barely recog. nisable steps from the almost permanent to the hopelessly fugitive. One example of this difficulty in classification must suffice: aureolin is almost worthy of a place in Class I., Indian yellow scarcely deserves inclusion in Class II. The action of mixed pigments upon one another, though not as frequent as it is supposed to be, creates another difficulty in our classification, so also does the medium employed in painting, which may either protect an alterable pigment from change or aid in its destruction. In fact, each method of painting, if really distinct, requires a special classification of the pigments to be employed in carrying it out.

In the annexed classification, a limit of three orders of stability has been adopted, the first class including the practically permanent pigments; the second class those which, though liable to a variable measure of change, may yet generally be allowed; and the third class those which should be definitely excluded from the palette:

Classified Table Of Pigments For Oil-Painting

In order to adapt the foregoing classified table to water-colours, some changes and additions must be made. Flake white, Naples yellow (true), cadmium (pale), and vermilion (artificial), must be removed from the Class (I.) of permanent pigments and placed in Class III., to which also must be relegated several pigments from Class II., namely, chrome yellow, malachite, and madder brown. Of course, it should be clearly understood that no pigment belonging to Class III. should be employed in artistic painting. One satisfactory addition, and one only, can be made to Class I. in the table. Indian ink is a pigment available for water-colour painting, and when it is free from a brownish hue may be safely used. Bistre and sepia are likewise used only as water-colours, but they are both fugitive, and must be placed in Class III. Almost the same modifications of the table are required in the case of tempera-painting as in water-colour painting. With fresco-painting the exclusion of many more pigments is an absolute necessity, as they are completely ruined by caustic lime.

Not only are all the chromates inadmissible, as well as all the pigments which cannot be trusted as water-colours, but likewise Prussian blue and Antwerp blue, while the madder colours are much altered in hue when used in this process. In stereochromy the number of available pigments is still further reduced.

It may not be uninstructive if we cite in this place the classification of pigments as used in oil which M. Decaux has published. The order followed by this experimenter is that of stability; the figures prefixed to the names of the individual pigments indicate the degree of permanence, 1 marking out the materials which are quite unchangeable, while 45 is the most fugitive of all:

Decaux's Table Of Pigments For Oil-Painting

Class I

1. Zinc white.

1. Flake white.

1. Yellow ochre.

1. Naples yellow.

1. Cadmium (deep).

1. Raw sienna.

1. Red ochre.

1. Mars red.

1. Venetian red.

1. Burnt Italian earth.

1. Green oxide of chromium. 1. Ivory black. 1. Terre verte. 1. Green ultramarine. 1. Cobalt blue. 1. Artif. ultramarine.

1. Ivory black.

2. Mars brown.

3. Burnt sienna.

4. Cobalt green.

5. Mars yellow.

6. Mars orange.

7. Burnt umber.

8. Viridian.

9. Indian red.

10. Mars violet.

11. Indian yellow.

12. Emerald green.

Class II

13. Malachite green.

14. Scheele's green.

15. Raw umber.

16. Vandyke brown.

17. Prussian blue.

18 to 23. Various madder lakes. 24. Madder carmine.

26. Madder 'rose dorée.'

27. Brown madder. 29. Cassel earth.

Class III

30. Pale chrome.

31. Zinc chromate.

32. Pale cadmium.

33. Orange chrome.

35. Asphalt.

36. Brown pink. 38. Vermilion.

42. Burnt carmine.

43. Yellow lake.

44. Carmine.

45. Crimson lake.

On comparing this classified list with that previously given a general accordance will be perceived, the low position given to raw umber and to vermilion, as well as the very high place assigned to Indian yellow and to terre verte, constituting the chief exceptions.