The first and most simple method consists in bringing the tube about half an inch beyond the ends of the reflectors. A circular piece of thin plane glass of the same diameter as the tube, is then pushed into the tube, so as to touch the reflectors. The pieces of coloured glass being laid upon this piece of glass when the tube is held in a vertical position, another similar disc of plane glass, having its outer surface ground with fine emery, is next placed above the glass fragments, being prevented from pressing upon them, or approaching too near the first plane glass by a ring of copper or brass; and is kept in its place by burnishing down the end of the tube. The eye being placed at the other end of the instrument, the observer turns the whole round in his hand, and perceives an infinite variety of beautiful figures and patterns, in consequence of the succession of new fragments, which are brought opposite the aperture by their own gravity, and by the rotatory motion of the tube. In this rude state, however, the instrument is by no means susceptible of affording very pleasing exhibitions. A very disagreeable effect is produced by bringing the darkest sectors, or those formed by the greatest number of reflexions, to the upper part of the circular field, and though the variety of patterns will be very great, yet the instrument is limited to the same series of fragments, and cannot be applied to the numerous objects which are perpetually presenting themselves to our notice. These evils can be removed only by adopting the construction shown in Fig. 22, in which the reflectors reach the very end of the tube. Upon the end of the tube _a b_, _c d_, Fig. 22, is placed a ring of brass, _m n_, which moves easily upon the tube _a b c d_, and is kept in its place by a shoulder of brass on each side of it. A brass cell, =M N=, is then made to slip tightly upon the moveable ring _m n_, so that when the cell is turned round by means of the milled end at =M N=, the ring _m n_ may move freely upon the tube. The fragments of coloured glass, etc., are now placed in a small object-box, as it may be called, consisting of two glasses, the innermost of which, _m n_, is transparent, and the other ground on the outside =P=, and kept at the distance of ⅛th or ⅒th of an inch by a brass rim: this brass rim generally consists of two pieces, which screw into one another, so that the object-plate can be opened by unscrewing it, and the fragments changed at pleasure. This object-box is placed at the bottom of the cell =M N=, as shown at =O P=, and the depth of the cell is such as to allow the side =O= to touch the end of the reflectors, when the cell is slipped upon the ring _m n_. When this is done, the instrument is held in one hand with the angular point =E=, Fig. 21, downwards, which is known by a mark on the upper side of the tube between _a_ and _b_, and the cell is turned round with the other hand, so as to present different fragments of the included