Many years ago Geissler, Crookes, and other scientists studied the spectra of gases excited to incandescence by the electric discharge in so-called vacuum tubes. The gases are placed in transparent glass or quartz tubes at rather low pressures and a high voltage is impressed upon the ends of these tubes. When the pressure is sufficiently low, the gases will glow and emit light. Twenty-eight years ago, D. McFarlan Moore developed the nitrogen tube, which was actually installed in various places. But there is such a loss of energy near the cathode that short "vacuum" tubes of this character are very inefficient producers of light. Efficiency is greatly increased by lengthening the tubes, so Moore used tubes of great length and a rather high voltage. As a tube of this sort is used, the gas gradually disappears and it must be replenished. In order to replenish the gas, Moore devised a valve for feeding gas automatically. An advantage of this mode of light-production is that the color or quality of the light may be varied by varying the gas used. Nitrogen yields a pinkish light; neon an orange light; and carbon dioxide a white light. Moore's carbon-dioxide tube is an excellent substitute for daylight and has been used for the discrimination of colors where this is an important factor. However, for this purpose devices utilizing color-screens which alter the light from the tungsten lamp to a daylight quality are being used extensively. The vacuum-tube method of producing light has an advantage in the selection of a gas among a large number of possibilities, and some of the color effects of the future may be obtained by means of it. Claude has lately worked on light-production by vacuum tubes and has combined the neon tube with the mercury-vapor tube. The spectrum of neon to a large extent compensates for the absence of red light in the mercury spectrum, with a result that the mixture produces a more satisfactory light than that of either tube. However, this mode of light-production has not proved practicable in its present state of development. Fundamentally the limitations are those of the inherent spectral characteristics of gases. Doubtless the possibilities of the mechanisms of the tubes and of combining various gases have not been exhausted. Furthermore, if man ever becomes capable of controlling to some extent the structure of elements and of compounds, this method of light-production is perhaps more promising than others of the present day.