Magnetite (Fe₃O₄) is a well-recognized petrogenetic indicator and is a common accessory mineral in many ore deposits and their host rocks. Recent years have seen an increased interest in the use of hydrothermal magnetite for provenance studies and as a pathfinder for mineral exploration. A number of studies have investigated how specific formation conditions are reflected in the composition of the respective magnetite. Two fundamental questions underlie these efforts — (i) How can the composition of igneous and, more importantly, hydrothermal magnetite be used to discriminate mineralized areas from barren host rocks, and (ii) how can this assist exploration geologists to target ore deposits at greater and greater distances from the main mineralization? Similar to igneous magnetite, the most important factors that govern compositional variations in hydrothermal magnetite are (A) temperature, (B) fluid composition — element availability, (C) oxygen and sulfur fugacity, (D) silicate and sulfide activity, (E) host rock buffering, (F) re-equilibration processes, and (G) intrinsic crystallographic controls such as ionic radius and charge balance. We discuss how specific formation conditions are reflected in the composition of magnetite and review studies that investigate the chemistry of hydrothermal and igneous magnetite from various mineral deposits and their host rocks. Furthermore, we discuss the redox-related alteration of magnetite (martitization and mushketovitization) and mineral inclusions in magnetite and their effect on chemical analyses. Our database includes published and previously unpublished magnetite minor and trace element data for magnetite from (1) banded iron formations (BIF) and related high-grade iron ore deposits in Western Australia, India, and Brazil, (2) Ag–Pb–Zn veins of the Coeur d'Alene district, United States, (3) porphyry Cu–(Au)–(Mo) deposits and associated (4) calcic and magnesian skarn deposits in the southwestern United States and Indonesia, and (5) plutonic igneous rocks from the Henderson Climax-type Mo deposit, United States, and the un-mineralized Inner Zone Batholith granodiorite, Japan. These five settings represent a diverse suite of geological settings and cover a wide range of formation conditions. The main discriminator elements for magnetite are Mg, Al, Ti, V, Cr, Mn, Co, Ni, Zn, and Ga. These elements are commonly present at detectable levels (10 to > 1000 ppm) and display systematic variations. We propose a combination of Ni/(Cr + Mn) vs. Ti + V, Al + Mn vs. Ti + V, Ti/V and Sn/Ga discriminant plots and upper threshold concentrations to discriminate hydrothermal from igneous magnetite and to fingerprint different hydrothermal ore deposits. The overall trends in upper threshold values for the different settings can be summarized as follows: (I) BIF (hydrothermal) — low Al, Ti, V, Cr, Mn, Co, Ni, Zn, Ga and Sn; (II) Ag–Pb–Zn veins (hydrothermal) — high Mn and low Ga and Sn; (III) Mg-skarn (hydrothermal) — high Mg and Mn and low Al, Ti, Cr, Co, Ni and Ga; (IV) skarn (hydrothermal) — high Mg, Al, Cr, Mn, Co, Ni and Zn and low Sn; (V) porphyry (hydrothermal) — high Ti and V and low Sn; (VI) porphyry (igneous) — high Ti, V and Cr and low Mg; and (VII) Climax-Mo (igneous) — high Al, Ga and Sn and low Mg and Cr.