In this case this line sometimes coincides with the hydrogen line, sometimes it is shifted. Stark attributes the appearance of this line to the glow of the ion itself—in the given case, hydrogen: the displacement of the line serves as a clear indication (according to the Doppler principle) of the high velocity of this ion. Unfortunately, the spectrograms of the northern lights are so weak that an exact determination of the position of the lines, and consequently of the velocities, is as yet impossible. The change in the coloration of the northern lights must evidently be ascribed to the different nature and different velocities of the excited positive ions.

S. Vavilov.

A New Method for Determining the Temperature of Luminous Flames.

(Hermann Senftleben and Elisabeth Benedict. Eine Methode zur Temperaturbestimmung leuchtender Flammen. Phys. Ztsch. 19 p. 180, 1918).

The method developed by the authors consists in the following. Into a luminous flame (Bunsen burner, Hefner lamp, etc.) a thin platinum wire is introduced, through which a current is passed. The deposition of incandescent particles of carbon, which cause the luminosity of the flame, on the wire will obviously take place until the temperature of the latter is lower than the temperature of the flame. By selecting the strength of the current heating the wire, one can find the temperature at which the deposition of soot just ceases; at this moment the temperatures of the flame and of the wire are equal. On the basis of the laws of radiation and the known reflecting power of platinum, the temperature of the wire was determined with a Holborn–Lummer pyrometer. Into the results obtained it is necessary to introduce a correction for the radiation of the layer of luminous gas situated between the wire and the pyrometer, and also for its absorption. The magnitude of this correction, found pyrometrically, is very small (not more than 1%). The temperature of the Hefner lamp found by this method, \(1690^\circ\), is in excellent agreement with the figures found by other methods. The results of the measurement, within comparatively wide limits, do not depend on the thickness of the wire introduced into the flame.

S. Vavilov.

Interrupter for Strong Currents.

(W. Kasperowicz. Galvanischer Unterbrecher, Phys. Ztsch. XIX p. 187, 1918.)

The interrupter proposed by the author, in its simplest form, has the shape of a U-shaped tube, the knees of which are connected by a narrow channel. The tube is filled with mercury, and electrodes are introduced into the knees. With a sufficient current strength and with a corresponding diameter and length of the connecting channel, the mercury in the latter passes into the vapor state, an arc arises, and the current either is interrupted completely or weakens. After the interruption, the mercury vapors rapidly condense, and the phenomenon repeats itself again. Thanks to