The following is an abridged translation of a paper on the osmium lamp read by Herr Fritz Blau before the Elektrotechnischer Verein, of Berlin, on January 24:

The total energy radiated by a heated body is proportional to the fourth power of the absolute temperature of the body. As the temperature increases, the percentage of the short waves emitted, as compared with those of long length radiated, increases also very rapidly ; but the absolute value of this percentage varies for different bodies, being smallest in so-called " black" bodies, but comparatively high in bodies with a metallic lustre and a smooth surface. Only these short waves, it need hardly be said, are observed as light. An efficient light-giving body must, therefore, withstand very high temperatures; its radiating surface, at a given efficiency, should be as large as possible; the substance of the filament should be transparent or white to minimize the effects of blackening; the surface of the filament should not become blacker or rough with use, and its resistance should increase with increasing temperature.

Osmium, generally in the form of osmium-iridium, is found in platinum ores and also in gold and silver. It is at present usually obtained by alloying the residue which remains when platinum is dissolved in aqua regia, with metals, such as zinc, lead or tin. The os-mium-iridium passes thereby into a finely divided state, and if it is now heated in a current of oxygen it will split up into osmium tetroxide and iridium. Osmium may then be obtained by treating the osmium tetroxide with any reducing agent.

When Herr Auer von Welsbach invented the osmium lamp some six years ago, he found it impossible to draw osmium in form of wire, and he first tried the expedient of coating platinum wire with osmium and evaporating subsequently the greater part of the platinum. It was observed, however, that a platinum osmium alloy was formed which - when containing more than four per cent of platinum - began to melt when the platinum was being evaporated. In order to keep percentage of platinum below this figure, the finished filament had to be made about five times thicker than its platinum core. As uniform platinum wire was not obtainable in thickness less than 0.02 millimetre, it will readily be seen that the osmium filaments could not advantageously be made in diameters below 0.1 millimetre. Such filaments, however, are not suitable for lamps consuming less than one ampere.

The next and final step was to mix most finely divided osmium with certain organic binding substances to a thick, tenacious paste. The paste is forced under high pressure through a diamond or sapphire die on to a card which is moved about in such a manner that the filament is deposited in loops. After drying, th9 organic binding substance is carbonized by heating the filament in vacuo. This porous filament, rich in carbon, is gradually heated to an intense incandes-ence by an electric current in an atmosphere containing much steam and a certain quantity of reducing gases. The carbon of the raw filament changes into carbon oxide and carbon dioxide, and after a relatively short time the filament consists of pure, or very nearly pure, osmium only, the percentage of carbon being insignifi-cant. Although the density of the filament has greatly increased during this process, the filament is still porous and its surface is far from smooth. After trying unsuccessfully several means of connecting the filament to the platinum leads, it was found that a very satisfactory method was to fuse them together by the electric arc. Osmium is brittle when cold, but soft when heated, and for this reason the lamps as made at present, are required to burn in a definite position.

When the manufacture of this lamp was taken up by the Deutche Gasgluhlicht Actiengesellschaft, the highest voltage that the lamp could stand was twenty-seven volts. After the lamps had been perfected to burn at thirty-seven volts they were put on the open market. The lamp voltage was soon after brought up to forty-four volts, and during the last few months a considerable number of fifty-five and seventy-three-lamps have been sold. Quite recently 110-volt lamps have appeared.

The filament of an osmium lamp for thirty-seven volts, twenty-five candle-power (Hefner nnits) consuming one and one-half watts per candle-power, has a diameter of 0.087 millimetre and a length of 280 millimetres. The radiating surface is from three square millimetres to3.2 square millimetres per candle-power. If an ordinary glow lamp be run at one and one-half watts per candle-power, its radiating surface per candle power is only 1.6 square millimetres. The carbon filament, therefore, must be both hotter and " blacker" than the osmium filament. With an increase of voltage of ten per cent the current increases 6.5 in the case of the osmium lamp and twelve per cent in the carbon filament lamp. The corresponding increase in illuminating power is forty per cent and eighty per cent, respectively.

It is not possible to state definitely the life of the osmium lamp, but 5000 hours have been repeatedly exceeded. Blackening of the bulb is very rare; (observed in ten per cent only of the lamps investigated so far. The surface of the lamp filament is at first somewhat rough, but it gradually becomes smoother, and this is the reason why the candle-power increases dur-insthe first 250 hours or so. Before the lamp is ready for sale it is burned in the factory for several hours in order to bring the lamp up to an approximately steady state.

If the lamp is run at less than one and one half watts per candle-power, the structural changes in the filament occur too rapidly for practical purposes. Nevertheless, lamps consuming one watt per per candle-power only have been observed to burn for several hundreds of hours without decrease in light. When exposed to hard shocks, the osmium lamp filament is somewhat more liable to break than the carbon filament, but if properly packed the loss through breakages in transit is small, being only about one and one-half per cent. These lamps stand the shocks of railway carriages and omnibuses very well. The osmium lamp is efficient also for very low voltages, and a large number of two-volt lamps for mines are in use in connection with portable accumulators. For a long time it was not possible to make thin filaments, but recently it has become feasible to reduce the diameter down to 0.03 milimetre - Electrician.