We shall limit ourselves here to pointing out the simplest and most practical of the numerous models of such apparatus that have been constructed up to the present time. All those that we shall describe are based on the incandescence of a platinum wire. A few have been constructed based on the induction spark, but they are more complicated and expensive, and have not entered into practical use. Before commencing to describe these apparatus, we shall make a remark in regard to the piles for working them, and that is that we prefer for this purpose Leclanché elements with agglomerated plates and a large surface of zinc. In order to bring about combustion in any given substance, it is necessary to bring near it an incandescent body raised to a certain temperature, which varies with the nature of the said substance, and which is quite low for illuminating gas, higher for petroleum, and a white heat for a wax taper or a candle. We have said that we make use exclusively of a platinum wire raised momentarily to incandescence by the passage of an electric current. The temperature of such wire will depend especially upon the intensity of the current traversing it; and, if this is too great, the platinum (chosen because of its inoxidizability and its elevated melting point) will rapidly melt; while, if the intensity is too little, the temperature reached by the wire will itself be too low, and no inflammation will be brought about. Practice soon indicates a means of obviating these two inconveniences, and teaches how each apparatus may be placed under such conditions that the wire will hardly ever melt, and that the lighting will always be effected. For the same intensity of current that traverses the wire, the temperature of the latter might be made to vary by diminishing or increasing its diameter. A very fine wire will attain a red heat through a very weak current, but it would be very brittle, and subject to break at the least accident. For this reason it becomes necessary to employ wires a little stronger, and varying generally from one to two-tenths of a millimeter in diameter. The current then requires to be a little intenser. The requisite intensity is easily obtained with elements of large surface, which have a much feebler internal resistance than porous-cup elements; and since, for a given number of elements, the intensity of the current decreases in measure as the internal resistance of the elements increases, it becomes of interest to diminish such internal resistance as much as possible. The platinum wires are usually rolled spirally, with the object in view of concentrating the heat into a small space, in order to raise the temperature of the wire as much as possible. There is thus need of a less intense current to produce the inflammation than with a wire simply stretched out. In fact, the same wire traversed by a current of constant intensity scarcely reaches a _red_ heat when it is straight, while it attains a _white_ heat when it is wound spirally, because, in the latter case, the cooling surface is less.