For over the past two decades it has been believed that the intense orange-red photoluminescence in Bismuth-doped materials originates from Bi[Formula: see text] ions. Based on the results from magnetic circular polarization experiments, we demonstrate that this hypothesis fails for Bismuth-doped silica glasses. Our findings contradict the generally accepted statement that the orange-red luminescence arises from [Formula: see text] [Formula: see text] [Formula: see text] transition in a divalent Bismuth ion. The degree of magnetic circular polarization of this luminescence exhibits non-monotonic temperature and field dependencies, as well as sign reversal. This complex behaviour cannot be explained under the assumption of a single Bi[Formula: see text] ion. The detailed analysis enables us to construct a consistent diagram of energy levels involved in the magneto-optical experiments and propose a new interpretation of the nature of orange-red luminescence in Bismuth-doped silica glass. A centre responsible for this notorious photoluminescence must be an even-electron system with an integer total spin, presumably a dimer of Bismuth ions or a complex consisting of Bi[Formula: see text] and an oxygen vacancy.