The phenomenon appears most readily between the minute lines, which the action of passing the fingers leaves on the clear surface; a somewhat rough state of the surface of the glass is otherwise requisite. On some glass the appearance may be produced by merely breathing; in other cases the wiping with the fingers is necessary: I have even met with polished mirror-glasses, one side of which immediately showed the colours vividly; the other not. To judge from some remaining pieces, the former was originally the front of the glass, the latter the side which was covered with quicksilver. 458. These experiments may be best made in cold weather, because the glass may be more quickly and distinctly breathed upon, and the breath evaporates more suddenly. In severe frost the phenomenon may be observed on a large scale while travelling in a carriage; the glasses being well cleaned, and all closed. The breath of the persons within is very gently diffused over the glass, and immediately produces the most vivid play of colours. How far they may present a regular succession I have not been able to remark; but they appear particularly vivid when they have a dark object as a background. This alternation of colours does not, however, last long; for as soon as the breath gathers in drops, or freezes to points of ice, the appearance is at once at an end. 459. Third condition.--The two foregoing experiments of the pressure and breathing may be united; namely, by breathing on a plate of glass, and immediately after pressing the other upon it. The colours then appear as in the case of two glasses unbreathed upon, with this difference, that the moisture occasions here and there an interruption of the undulations. On pushing one glass away from the other the moisture appears iridescent as it evaporates. 460. It might, however, be asserted that this combined experiment exhibits no more than each single experiment; for it appears the colours excited by pressure disappear in proportion as the glasses are less in contact, and the moisture then evaporates with its own colours. 461. Fourth condition.--Iridescent appearances are observable in almost all bubbles; soap-bubbles are the most commonly known, and the effect in question is thus exhibited in the easiest mode; but it may be observed in wine, beer, in pure spirit, and again, especially, in the froth of chocolate. 462. As in the above cases we required an infinitely narrow space between two surfaces which are in contact, so we can consider the pellicle of the soap-bubble as an infinitely thin lamina between two elastic bodies; for the appearance in fact takes place between the air within, which distends the bubble, and the atmospheric air. 463. The bubble when first produced is colourless; then coloured stripes, like those in marble paper, begin to appear: these at length spread over the whole surface, or rather are driven round it as it is distended. 464.