AbstractExperimental data on the photoionization of liquid water are reviewed and reanalyzed. It is suggested that, near to the ionization threshold at 6.5 eV and up to about 8 eV, an optical Charge Transfer (CT) mechanism leads to the formation of hydrated electrons with a low quantum yield. The absorption band of liquid water at the longest UV wavelengths is assigned to this CT transition superimposed on an intramolecular transition into the R3s state. At excitation energies where the R3p states become attainable, i.e. somewhere between 6.5 and 8 eV, a photo‐induced electron transfer mechanism sets in which is characterized by high quantum yields of hydrated electrons and by average electron‐ion separations below those expected for intermediately “quasi‐free” electrons. This mechanism is dominant up to at least 10 eV, i.e. up to the highest excitation energies for which subpicosecond studies are currently available. In this energy range, autoionization generates only a minor part of all hydrated electrons produced. Only at excitation energies approaching the gas‐phase ionization potential, autoionization and direct photoionization are supposed to become dominant. If any autoionization channels are open below excitation energies of 10 eV, then the band gap energy of water, which is currently believed to be about 8.9 eV, has to be located as high as 10–12 eV.