D O C U M E N T 3 0 0 J U L Y 1 9 2 2 2 3 3 collision of an excited atom leads to an extinction of the luminosity, then the middle of the line is mainly weakened and an (apparent) line broadening is the conse- quence. I do not know if this second thought makes any real sense. The first one, though, does have some experimental backing that Franck also already mentions in a publication.[8] If one excites Na-vapor to fluorescence just with the D1-line, then just the D1-line occurs in the resonance radiation as well. If one raises the pressure by adding hydrogen, then the D2-line also appears in the emission. Evidently, col- lisions between excited Na atoms and H2 molecules shift electrons from the D1- orbit to the D2-orbit (the opposite for primary excitation with D2 works just the same, incidentally). One could just as well imagine that as a result of such a colli- sion the electron is thrown down directly onto the normal orbit (1.5 S-orbit)—the question now is whether it occurs while undergoing light emission but if so, then surely under disturbed conditions.– I think this is essentially what we agreed on. It would certainly be nice of you if you would let me know sometime what you think of it. That the luminosity can be completely extinguished by collisions is also proved, by the way, precisely for Hg resonance, which vanishes at Hg pressures that are too high. Surely the basic con- cept that must absolutely be adhered to is that whatever is all right for the emission process must also be valid for absorption. On the contrary, it is questionable how large the percentage of such destructive collisions is compared to the collisions evincing emission or by conveying the energy to another similarly excitable atom (Klein-Rosseland).[9] I hope I have not written too unclearly, either in the literal or figurative sense. With best regards, Peter Pringsheim. I just discovered in an older paper by Wood some information[10] —if completely correct—perhaps not without some interest: (1) By directing 1 atm. air to cold Hg- vapor, the absorption line is strongly weakened in the center but is broadened over- all enough for the total absorption to remain constant. (2) 30 mm air already almost completely suppresses emission of the resonance line, i.e., most of the excited atoms experience the extinguishing collision before their lingering period has elapsed. (3) However, 30 mm air does not really reduce the width of the absorption line, that is, the absorption of a beam of resonance rays sent through the Hg-vapor cell is weakened just the same as if no air were there (hence with reemission) but with 30 mm air (without reemission). Does this mean: the timeless process of absorption is not noticeably [. . .ed] by the light molecules, which do not yet affect the configuration much likewise, of course, for the individual process of emission itself (the emission line has not been measured but is surely just as little broad- ened), only that often mostly becomes impossible before it occurs, owing to the