We will have to a**ume that in the ionization of a gas by ultraviolet light one quantum of light energy is used for the ionization of one molecule of gas. From this it follows that the work of ionization (i.e., the work theoretically required for ionization) of one molecule cannot be greater than the energy of one effective quantum of light observed. If J denotes the (theoretical) ionization work per gram-equivalent, we must have Rβv ≥ J However, according to measurements by Lenard, the largest effective wavelength for air is about 1.9 x 10-5cm, hence Rβv = 6.4 x 1012 erg ≥ J An upper limit for the work of ionization can also be obtained from the ionization potentials in rarefied gases. According to J. Stark the smallest measured ionization potential (at platinum anodes) for air is about 10 volts. This one obtains 9.6 x 1012 as the upper limit for J, which is almost equal to the value we have just found. There is still another consequence, whose verification by experiment seems to me of great importance. If each absorbed quantum of light energy ionizes one molecule, then the following relation must hold between the quantity of light absorbed L and the number j of gram-molecules ionized by it: J = L/(Rβv) If our conception corresponds to reality, this relation must apply to all gases that (at the relevant frequency) display no noticeable absorption that is not accompanied by ionization. Bern, 17 March 1905.