This work is a continuation ofthat of Kazda, who found the long wavelength limit of a mercury surface cleansed of impurities by means of a constant overflow.
Kazda's value of 2735A for the threshold of clean mercury is checked. When the surface flow is allowed to stop in a high vacuum, some impurity attacks the surface, quickly raising the threshold to 2850A. If liquid air is not used this impurity is present in larger amounts and attacks the running surface. Indications are that a surface film is formed and maintained in spite of the flow when liquid air is not used, or requires two hours or more of flow for removal if liquid air is used. This impurity cannot be one of the gases with extremely low melting points. It is not water, but may be a component of the stopcock grease. When the surface is left standing several days in a high vacuum its threshold falls to 2680A. If liquid air is not used the standing surface has a limit of 2560A. All of these values are closely reproducible. Pure hydrogen in contact with the surface does not change the photoelectric behavior. When the mercury is condensed in the presence of hydrogen, some of the gas is dissolved in the metal. This does not change the characteristic threshold of the mercury. It does, however, have the effect of greatly impeding the action of other impurities that form on the surface. This is indicated by the fact that over two hours is required for the change from the threshold of 2735A for the clean surface to the maximum of 2850A, as compared with 13 minutes for this change when hydrogen is not present.
