Fresnel invented a very simple system. He took a central piece of glass, which may be described as a bull’s-eye, and around this disposed a number of concentric rings of glass. But these rings projected beyond one another. Each constituted the edge of a lens which, while its radius differed from that of its neighbour, owing to its position, yet was of the same focus in regard to the source of illumination. The parts were shaped with extreme care and were united in position by the aid of fish glue, the whole being mounted in a metal frame. The advantage of the system was apparent in the first demonstrations. The lenses being comparatively thin, only one-tenth of the light passing through was absorbed, whereas in the old parabolic reflectors one-half of the light was lost. [Illustration: FIG. 3.--SINGLE FLASHING APPARATUS (ONE PANEL AND MIRROR). (_By permission of Messrs. Chance Bros. and Co., Ltd._)] This revolutionary development was perfected in 1822, and in the following year it was submitted to its first practical application on the tower of Cordouan in the Gironde. Several modifications were made by the inventor for the purpose of adapting his system to varying conditions. One of the most important was the disposition of lenses and mirrors above the optical apparatus for the purpose of collecting and driving back the rays which were sent out vertically from the illuminant, so that they might be mingled with the horizontal beam, thereby reinforcing it. At a later date similar equiangular prisms were placed below the horizontal beam so as to catch the light thrown downwards from the luminous source, the result being that finally none, or very little, of the light emitted by the illuminant was lost, except by absorption in the process of bending the rays into the desired direction. [Illustration: FIG. 4.--A TWENTY-FOUR PANEL LIGHT, WHICH WAS INTRODUCED INTO CERTAIN FRENCH LIGHTHOUSES.] In this ingenious manner the circle of light is divided into sections, called “panels,” each of which comprises its bull’s-eye and its group of concentric rings and prisms. The extent of this division varies appreciably, as many as sixteen panels being utilized in some instances. In this direction, however, subdivision can be carried too far. Thus, in some of the French lighthouses no less than twenty-four panels were introduced. The disadvantage is obvious. The total volume of light emitted from the luminous source has to be divided into twenty-four parts, one for each panel. But the fewer the panels, the more light is thrown through each, and the correspondingly greater power of the beam. Thus, in a four-panel light each beam will be six times as powerful as that thrown from a twenty-four panel apparatus of the same type.