Before we proceed to a further description of this phenomenon, we may point out the most convenient mode of exhibiting it. Place a large convex glass on a table near the window; upon this glass lay a plate of well-polished mirror-glass, about the size of a playing-card, and the mere weight of the plate will press sufficiently to produce one or other of the phenomena above described. So, also, by the different weight of plates of glass, by other accidental circumstances, for instance, by slipping the plate on the side of the convex glass where the pressure cannot be so strong as in the centre, all the gradations above described can be produced in succession. 442. In order to observe the phenomenon it is necessary to look obliquely on the surface where it appears. But, above all, it is to be remarked that by stooping still more, and looking at the appearance under a more acute angle, the circles not only grow larger but other circles are developed from the centre, of which no trace is to be discovered when we look perpendicularly, even through the strongest magnifiers. 443. In order to exhibit the phenomenon in its greatest beauty, the utmost attention should be paid to the cleanness of the glasses. If the experiment is made with plate-glass adapted for mirrors, the glass should be handled with gloves. The inner surfaces, which must come in contact with the utmost nicety, may be most conveniently cleaned before the experiment, and the outer surfaces should be kept clean while the pressure is increased. 444. From what has been said it will be seen that an exact contact of two smooth surfaces is necessary. Polished glasses are best adapted for the purpose. Plates of glass exhibit the most brilliant colours when they fit closely together, and for this reason the phenomenon will increase in beauty if exhibited under an air-pump, by exhausting the air. 445. The appearance of the coloured rings may be produced in the greatest perfection by placing a convex and concave glass together which have been ground on similar segments of circles. I have never seen the effect more brilliant than with the object-glass of an achromatic telescope, in which the crown-glass and flint-glass were necessarily in the closest contact. 446. A remarkable appearance takes place when dissimilar surfaces are pressed together; for example, a polished crystal and a plate of glass. The appearance does not at all exhibit itself in large flowing waves, as in the combination of glass with glass, but it is small and angular, and, as it were, disjointed: thus it appears that the surface of the polished crystal, which consists of infinitely small sections of lamellæ, does not come so uninterruptedly in contact with the glass as another glass-plate would. 447.