Some experiments have been recently published on the use of bottles of oxygen in which to burn fragments of magnesium ribbon, to take portraits by the magnesium light. An objection is the trouble and cost of making the oxygen, and that means have to be adopted to diffuse the light near its source. The objection to burning the ribbon in air is that the exposure is longer, and the burnt ash has a tendency to drop off and put out the light before the selected length of metal is consumed.

Some years ago, W. H. Harrison overcame the difficulties by the adoption of one part of the principle of Larkin's magnesium lamp. The method may be thus explained. On the top of a firm and solid table is placed a base-board supporting a wooden upright some 7-8 ft. long, consequently reaching nearly to the ceiling of the room. At the top is a kind of Tittle brass funnel with no neck, and a large opening where the neck should be. This brass inverted cone was no larger than an egg-cup,and its lower opening was about as large as a fourpenny piece. A tin spirit-lamp, with a horizontal neck and flame, is placed below, so that anything falling through the funnel must pass through the flame, yet just miss touching the wick. The combustible substance was magnesium powder mixed with sand. One thimbleful of the powder, for instance, is mixed with two of sand with a spoon or bit of wood on a sheet of paper; then, the sitter being in position and all being ready, the mixture is poured all at once into the funnel. A long sheet of flame of one or two seconds' duration is the result. If the picture prove over-exposed, the proportion of sand has to be increased in the next trial, or that of the magnesium powder correspondingly diminished.

When once the right proportions are known, portrait after portrait can be produced, properly exposed with dead certainty. No cap is necessary to the lens, provided a candle only is used for the normal illumination of the room, and it is not placed so that an image of it can be thrown by the lens into the camera. The proportion of magnesium powder regulates the proper exposure. At first there was a difficulty, and one only, with this simple apparatus, and that was that the aqueous vapour from the flame condensed on the lower parts of the brass cone, and the powder stuck to the neck of the wet funnel, sometimes blocking it and arresting the flow of illuminating material altogether. The neck of the funnel was therefore cut off, and the lower opening of the cone made rather large. It would be an improvement to use something on which aqueous vapour has less tendency to condense than upon cold brass. The brass, however, was subsequently kept hot by fixing the spirit-lamp nearer to it than in the first experiments.

The larger the proportion of magnesium in relation to the sand, the longer the flame and the shorter its duration. With a very rich proportion, it is possible to have a flame 7-8 ft. long. The long flames give the necessary diffusion of light, and a white sheet on the opposite side of the sitter improves the shadows. The finest artistic effects of light and shade on the countenance of the sitter can be obtained by this method when the lamp is placed in the proper position.

The sitter need not be fatigued with focussing operations. A candle placed where his face will come will do to focus upon. In fact, it is a capital method for the comfort of the sitter, who must not look in the direction of the coming flame, lest his eyes be dazzled by its magnificent flash, which, however, lasts but a second or two. {Brit. Jour. Photog.)