The fixed light, on the other hand, presents to the eye a steady and unchanging appearance; and the chief object to be obtained in its construction, is to unite the greatest brilliancy with an equal distribution of the light in every direction. The condition of perfect distribution, as already said, is most rigorously fulfilled by the use of refracting zones or belts, which form, by their union, a cylinder enveloping the flame placed in its centre, and possess the property of refracting the light in the vertical direction only, without affecting its natural divergence horizontally. The light from the focus which is incident on the inner surface of the belt is therefore projected forwards in the shape of a flat ring of equal brilliancy all round the horizon. This repetition may seem needless, but it is hoped it will be found useful in rendering intelligible the following outline of the plan proposed by Captain HALL for the improvement of fixed lights, and the account of the trials that were made with that object in view. The familiar experiment of whirling a burning stick quickly round the head, so as to produce a ribbon of light, proves the possibility of causing a continuous impression on the retina by intermittent images succeeding each other with a certain rapidity. From the moderate velocity at which this continuity of impression is obtained, we should be warranted in concluding, _a priori_, that the time required to make an impression on the retina is considerably less than the duration of the impression itself; for the continuity of effect must, of course, be caused by fresh impulses succeeding each other before the preceding ones have entirely faded. If it were otherwise, and the time required to make the impression were equal to the duration of the sensation, it would obviously be impossible to obtain a series of impulses so close or continuous in their effects as to run into and overlap each other, and thus throw out the intervals of darkness; because the same velocity which would tend to shorten the dark intervals, would also curtail the bright flashes, and thus prevent their acting on the eye long enough to cause an impression. Accordingly, we find that the duration of an impression is in reality much greater than the time required for producing the effect on the retina. It is stated by Professor WHEATSTONE, in the London Transactions for 1834, that only about _one millionth part_ of a second is required for making a distinct impression on the eye; and it appears, from a statement made by Lamé, at p. 425 of his Cours de Physique, that M. PLATEAU found that an impression on the retina preserved its intensity unabated during _one hundredth_ of a second, so that, however small those times may be in themselves, the one is yet 10,000 greater than the other.