Fifth condition.--Very delicate pellicles may be formed in various ways: on these films we discover a very lively play of colours, either in the usual order, or more confusedly passing through each other. The water in which lime has been slaked soon skims over with a coloured pellicle: the same happens on the surface of stagnant water, especially if impregnated with iron. The lamellæ of the fine tartar which adheres to bottles, especially in red French wine, exhibit the most brilliant colours, on being exposed to the light, if carefully detached. Drops of oil on water, brandy, and other fluids, produce also similar circles and brilliant effects: but the most beautiful experiment that can be made is the following:--Let aqua fortis, not too strong, be poured into a flat saucer, and then with a brush drop on it some of the varnish used by engravers to cover certain portions during the process of biting their plates. After quick commotion there presently appears a film which spreads itself out in circles, and immediately produces the most vivid appearances of colour. 471. Sixth condition.--When metals are heated, colours rapidly succeeding each other appear on the surface: these colours can, however, be arrested at will. 472. If a piece of polished steel is heated, it will, at a certain degree of warmth, be overspread with yellow. If taken suddenly away from the fire, this yellow remains. 473. As the steel becomes hotter, the yellow appears darker, intenser, and presently passes into red. This is difficult to arrest, for it hastens very quickly to bright blue. 474. This beautiful blue is to be arrested if the steel is suddenly taken out of the heat and buried in ashes. The blue steel works are produced in this way. If, again, the steel is held longer over the fire, it soon becomes a light blue, and so it remains. 475. These colours pass like a breath over the plate of steel; each seems to fly before the other, but, in reality, each successive hue is constantly developed from the preceding one. 476. If we hold a penknife in the flame of a light, a coloured stripe will appear across the blade. The portion of the stripe which was nearest to the flame is light blue; this melts into blue-red; the red is in the centre; then follow yellow-red and yellow. 477. This phenomenon is deducible from the preceding ones; for the portion of the blade next the handle is less heated than the end which is in the flame, and thus all the colours which in other cases exhibited themselves in succession, must here appear at once, and may thus be permanently preserved. 478.