We have recently discovered several chromitite dykes that cut a sequence of alternating layers of dunite and chromitite in the lower part of the Monchegorsk Layered Intrusion, Russia. The chromitite dykes are up to 18 cm wide and several metres long, have sharp boundaries with the host dunite, are fine-grained and contain up to 80–85% chromite. The dykes are internally zoned, with chromite showing an inward decrease in FeO and an increase in MgO and Mg-number [Mg/(Mg + Fe 2+ )]. Most olivine crystals in dunite adjacent to chromitite dykes are also reversely zoned and their rims show an increase in Mg-number towards contacts with the chromitite dykes. The dykes and layers differ in their platinum-group element (PGE) patterns, with the former being relatively enriched in Os, Ir and Ru and depleted in Pt, Pd and Rh. These differences preclude the chromitite dykes being remobilized chromitite layers. Similarly, the dykes cannot have formed from chromite-rich slurries emplaced into the evolving magma chamber. Gravity-induced separation of chromite from such slurries in feeder conduits cannot result in such a high concentration of chromite. Also, the lack of chromitite layers higher up in the layered series of the intrusion is not compatible with this idea. We propose that chromitite dykes resulted from the prolonged flow of chromite-saturated magma that crystallized chromite in situ along the walls of conduits. The high abundance of chromite in the dykes can be explained by the ability of magma flowing through channels to displace interstitial liquid from growing chromite. The zoning in the dykes and reversely zoned wall-rock olivine can be attributed to postcumulus interaction between chromite in dykes and olivine in the adjacent dunite, a process not fully understood but thought to require a film of trapped liquid. We infer that in situ crystallization of chromite from replenishing chromite-saturated magma may be a viable origin of massive chromitite in layered intrusions.