Novel as the idea of using electricity for illuminating large spaces may appear to many, we have all of us been long familiar with the fact that electricity is capable of replacing the darkness of night by the light of broad day over areas far larger than those which our electricians hope to illuminate. The lightning flash makes in an instant every object visible on the darkest night, not only in the open air, but in the interior of carefully darkened rooms. Nay, even if the shutters of a room are carefully closed and the room strongly illuminated, the lightning flash can yet be clearly recognised. And it must be remembered that though the suddenness of the flash makes us the more readily perceive it (under such circumstances, for instance), yet its short duration diminishes its apparent intensity. This may appear a contradiction in terms, but is not so in reality. The perception that there has been a sudden lighting up of the sky or of a room, is distinct from the recognition of the actual intensity of the illumination thus momentarily produced. Now it is quite certain that the eye cannot assign less than a twenty-fifth of a second or so to the duration of the lightning flash, for, as Newton long since showed, the retina retains the sensation of light for at least this interval after the light has disappeared. It is equally certain, from Wheatstone's experiments, that the lightning flash does not actually endure for the 100,000th part of a second. Adopting this last number, though it falls far short of the truth--the actual duration being probably less than 1,000,000th of a second--we see that so far as the eye is concerned, an amount of light which was really emitted during the 100,000th part of a second is by the eye judged to have been emitted during an interval 4,000 times as long. It is certain, then, that the eye's estimate of the intensity of the illumination resulting from a lightning flash is far short of the truth. This is equally true even in those cases where lightning is said to be for awhile continuous. If the flashes for a time succeed each other at less intervals than a twenty-fifth of a second, the illumination will appear continuous. But it is not really so. To be so, the flashes should succeed each other at the rate of at least 100,000, and probably of more than 1,000,000 per second.