The New Jersey Highlands and contiguous Hudson (New York) Highlands host hundreds of small, worked-out magnetite mines, and the major zinc-oxide deposits of Franklin and Sterling Hill. The origin of the magnetite ore remains controversial. Two temporally distinct genetic models have been proposed for magnetite: (1) a pre-Ottawan, sedimentary exhalative model in which ores were deposited on the seafloor as precipitates from iron-enriched hydrothermal fluids; (2) a late-Ottawan, fluid alteration model in which the current mineral composition of ores was derived from Fe-rich, alkaline fluids, associated with late episodes of granitic plutonism (low-Ti Kiruna-type deposits), or in which deposits derived from metamorphogenic fluids circulating in a regional shear zone leached metals from host rocks and precipitated them in veins and faults. Detailed mapping of ore deposits and host rocks near Wanaque and Ringwood, New Jersey, and Warwick, New York, reveal that ore bodies are hosted by supracrustal rocks deposited in a back-arc basin. At Wanaque and Warwick, ores are strataform and stratabound. Discordant ore bodies at Ringwood are associated with a feeder zone, presumed to have intersected the paleosurface at a steep angle. Sharply layered quartz-diopside gneiss is spatially associated with ore at Wanaque and Warwick, and massive diopsidite is spatially associated with ore at all three locations. Diopsidite in these rocks is nonaluminous, iron enriched, and shares these characteristics with modern metalliferous sediments. Massive diopsidite and quartz-diopside gneiss are interpreted to be metamorphosed carbonate facies iron formation, and associated magnetite to be metamorphosed oxide-facies iron formation.