Among the "basic colors" we search in vain, however, for a really fast dye on any fiber. Still, Magdala red, perhaps, appears faster than the rest on silk, and among the greens and blues we find a few dull blues on cotton, which, for this fiber, have been recommended as substitutes for indigo, viz., Indophenin, paraphenylene, blue, cinerein, Meldola's blue, etc. The azine greens, also, appear tolerably fast on cotton and on silk, but although possessing some body of color, after exposure, the original dark green has changed to a decided drab.

When we examine the "acid colors," however, we meet with a number of scarlets, crimsons, and clarets, possessing considerable fastness both on wool and on silk. Some, indeed, appear almost, if not entirely, as fast as cochineal scarlet, e.g., Biebriech scarlet, brilliant crocein, etc.

Among the "acid oranges and yellows," we also find a goodly number which are of medium fastness. About ten, either on wool or on silk, may even be accounted really fast, and are fit, apparently, to rank with alizarin colors. Note, for example, on wool: Crocein orange, aurantia, orange crystal, tartrazin, milling yellow, palatine orange; on silk, acid yellow D, brilliant yellow, azo acid yellow, metanil yellow, curcumin S, etc. I may remark that these are some of the fastest yellows on wool and silk with which we are acquainted. It is interesting to note the decided fugitive character, on silk, of tartrazin, aurantia, orange crystal, etc., compared with their great fastness on wool. Observe, also, how, on wool, the pale lemon yellow of picric acid has changed to a full reddish brown.

Among the "acid greens and blues," all the colors are fugitive, both on wool and on silk. Patent blue appears slightly better than the rest. Of the "acid blacks and violets," a few colors are of medium fastness, both on wool and silk, e.g., naphthol black, naphthylamine, black, resorcinol brown, fast brown, etc.

When we examine the Congo colors, amid a number of very fugitive colors, we find a few which are satisfactorily fast. Among the reds, for example, diamine fast red is quite remarkable for its fastness, both on wool and silk, and may certainly rank with alizarin; but on cotton, it is quite as fugitive as the rest. Of medium fastness on wool are brilliant Congo G and R, Congo G R; and on silk, diamine scarlet B, deltapurpurin 5 B, and brilliant Congo R.

Among the "Congo oranges and yellows," we find some of the fastest on cotton of this class of colors. Still they deserve only the rank of medium fastness. They are Mikado orange 4 R, R, G. Hessian yellow, curcumin S, chrysophenin. On wool, we have about half a dozen of medium fastness, viz., benzo-orange, Congo orange R, chrysophenin G, chrysamin R, brilliant yellow. On silk, however, we find in this group about a dozen of the fastest oranges and yellows with which we are acquainted for this fiber, viz., Congo orange R, chrysophenin G, diamine yellow N, brilliant yellow, curcumin W, benzo orange, Hessian yellow, chrysamin R and G, cresotin yellow R and G, cotton yellow G, and carbazol yellow.

Does it not appear somewhat remarkable that we should find among this generally fugitive group of coloring matters colors which are so eminently fast on silk, and which we entirely fail to meet with among those groups which usually furnish our fast colors, e.g., the alizarin group?

Passing on to the "Congo violets, blues, and purples," we find few colors worthy of particular notice for fastness. Diamine violet N appears, perhaps, of medium fastness on wool and silk, while sulphonazurin, benzo-black blue, and direct gray may claim the same distinction on silk.

In the small group of colors which are produced directly upon the fiber, none seems to call for special notice, except aniline black, which, notwithstanding its direct derivation from aniline, is probably the fastest color we have upon any fiber.

Now, in classifying the whole range of coal tar coloring matters into "mordant dyes" and "direct dyes," and the latter into acid, basic, Congo colors, etc., I have looked at them from the point of view of the dyer and arranged them according to color and mode of application. The chemist, however, classifies them quite differently, viz., according to their chemical constitution, i.e., the arrangement of the atoms of which they are composed, and thus we have nitro colors, phthaleins, azines, and so on.

In studying the action of light on the coal tar colors from this point of view, we find that whereas the members of some groups are for the most part fugitive, the members of other groups are nearly all fast, and it becomes at once apparent that the chemical constitution of a coloring matter exercises a profound influence upon its behavior toward light. Members of the rosaniline group are all similarly fugitive, while those of the alizarin group possess generally the quality of fastness. Particularly fugitive are the eosins, and yet some of these, by a slight modification of constitution, e.g., the introduction of an ethyl group, as in ethyl-eosin, are rendered distinctly faster.

In the azo group some colors are fugitive, others are moderately fast, and it is generally recognized that certain classes of the tetrazo compounds are distinctly faster than the ordinary diazo colors.

By a careful study of the influence of the atomic arrangement upon the stability of colors, information useful to the color manufacturer may possibly be gained, but at present my facts are not yet sufficiently tabulated to enable one to recognize any generally pervading law in this direction.

It is scarcely necessary to say that the fastness to light of a color is independent of its commercial value, this being mainly determined by the price of the raw material from which it is manufactured, the working expenses, and the profit desired by the manufacturer. Neither must we suppose that facility of application necessarily interferes with its fastness to light, for some of our fastest coal tar colors on wool, e.g., diamine fast red, tartrazin, etc., are applied in the simplest possible manner. On the other hand, the intensity or depth of a color has considerable influence on its fastness. Dark full shades invariably appear faster than pale ones produced from the same coloring matter, simply because of the larger body of pigment present. A pale shade of even a very fast color like indigo will fade with comparative rapidity. The fugitive character of many of the coal tar colors is, in my opinion, rendered more marked, because, owing to their intense coloring power, there is often such an infinitesimal amount of coloring matter on the dyed fiber. Hence it is that in the Gobelin tapestries pale shades on wool are frequently obtained by the use of more or less unchangeable metallic oxides and other mineral colors, to the exclusion of even fast vegetable dyes.