Having deduced the relations of colours regularly from white or light, through the primaries, secondaries, and tertiaries.

* Exp. xxvii. p. 247, 4to. edition of this, work.

To black or shade, we might have done the same inversely from black to white. On this plan the tertiaries, olive, russet, and citrine, take the place of the primaries, blue, red, and yellow; while the secondaries still retain their intermediate station and relation to both; thus russet and olive compose or unite in dark purple, citrine and olive in dark green, and russet and citrine in dork orange, as demonstrated, page 38. The tertiaries have, therefore, the same order of relation to black that the primaries have to white; and we have black primaries, secondaries, and tertiaries, inversely, as we have white primaries, secondaries, and tertiaries. directly; or, what is the same thing, we have light and dark colours of all clauses.

Theoretically, the tertiaries may be produced either by mixture of the primaries alone, the secondaries alone, or by the primaries and black; but in the latter mode the black must he perfectly neutral and the colours (rue, to do this practically, and none of our pigments ore perfect enough for this; the latter mode is, therefore, a bad practice, and applicable only to the production of shadow colours, distinguished by the term semi-neutrals. - P. 28. It is the imperfection or nnomaloiisness of pigments which renders these distinctions necessary, for had we pigments in the chromatic ami relative perfection which belongs to prismatic colour*, with also a perfectly transparent and neutral shade-colour with which to combine the whole, the inverse order of our classification would afford us the series from black to white; the contrary order adopted is, however, practically preferable, because we have white pigments sufficiently opaque and pure to compound ail tints without changing the denominations of colours; but, as before remarked, we have no black so transparent and neutral as to afford us equally perfect shades: both these orders are, however, represented by the definite scale and the scale of equivalents taken conversely, and the absolute completeness of the natural system of colours is demonstrated analytically ami synthetically, or rather antithetically.