246 THE rOPULAR SCIENCE MONTHLY
10 to 45(10)"® sees., depending on the line. The velocity of the cen- ters of Mg A 4481 was found to be about 2.5(10)^ cm. per sec. near the electrodes, dropping to 1.7(10)^ cm. about a millimeter from the electrodes. Air line centers have an existence of about 7(10)"'^ sec. The emission centers in flames and arcs have been studied by Lenard and others. The results obtained do not agree with those found by Stark working with canal rays.
Negative Electroxs as Emission axd Absorption Centers The Zeeman effect produced b}^ the action of a magnetic field upon the emission or absorption light centers shows that for many spectrum lines of gases and vapors the light center consists of a negative electron and the ratio of the charge to the mass of the electron obtained in this way agrees very well with the value obtained by other methods. The more accurate experiments give e/m = 1.775 while direct experiments give 1.772.
The Positive Electron The positive electron has never been isolated in any experiment with vacuum-tube discharges, radiations from radioactive materials, etc. The Zeeman effect of certain band spectra of chlorides and fluorides of some of the alkaline earth elements studied by Dufour and of the absorption spectra of neodymium and erbium compounds as studied by Becquerel indicate the existence of positive electrons. These Zeeman effects may be explained, however, as being due to induced magnetic fields being set up in the region of the light centers, magnetic fields whose intensities are very different from the field impressed from without.
Absorption Centers of Solutions of the Eare Elements Many solutions of salts of elements such as uranium, neodymium, erbium, somarium, etc., show a banded absorption spectrum. Many of these bands are very narrow. Jones, xA.nderson and the writer have found that the absorption centers of many of these salts (e. g., uranous chloride) consist of centers containing the salt and an " atmosphere " of the solvent, the whole center apparently acting as a compound. Thus in the above case it is possible to have " water and alcohol centers " of uranous chloride in a solution of uranous chloride in water and alcohol. Increasing the amount of one solvent appears to increase the relative number of the centers of that solvent without apparently changing their composition. The different solvent centers have different degrees of persistency. The water and alcohol bands of neod}Tnium chloride are of about equal intensity when the salt is dissolved in a solution con- taining about 3 per cent, water and 97 per cent, alcohol. Changes of temperature change the relative persistency of the liglit centers.
In the case of some uranyl salts the addition of free acid of the salt causes a shift of the bands. This has been explained by the writer as being due to the fact that the light centers consisted of " aggregates " of