In the geosphere, germanium (Ge) has a chemical behavior close to that of silicon (Si), and Ge commonly substitutes for Si (in small proportions) in silicates. Studying the evolution of the respective proportions of Ge and Si through time allows us to better constrain the global Si cycle. The marine inventory of Ge present as dissolved germanic acid is facing two main sinks known through the study of present sediments: 1) incorporation into diatom frustules and transfer to sediments by these “shuttles”, 2) capture of Ge released to pore water through frustule dissolution by authigenic mineral phases forming within reducing sediments. Our goals are to determine whether such a bio-induced transfer of Ge is also achieved by radiolarian and whether Ge could be trapped directly from seawater into authigenic phases with no intervention of opal-secreting organisms (shuttles). To this end, we studied two Paleozoic radiolarite formations and geological formations dated of Devonian, Jurassic and Cretaceous, deposited under more or less drastic redox conditions. Our results show that the Ge/Si values observed for these radiolarites are close to (slightly above) those measured from modern diatoms and sponges. In addition, our results confirm what is observed with some present-day reducing sediments: the ancient sediments that underwent reducing depositional conditions are authigenically enriched in Ge. Furthermore, it is probable that at least a part of the authigenic Ge came directly from seawater. The recurrence and extent (through time and space) of anoxic conditions affecting sea bottoms have been quite important through the geological times; consequently, the capture of Ge by reducing sediments must have impacted Ge distribution and in turn, the evolution of the seawater Ge/Si ratio.