The genesis of late metamorphic gold lodes in the northwestern Alps has remained enigmatic. In order to trace the source of the ore-bearing fluid at the Brusson locality and to test current genetic models, 130 Rb-Sr isotope measurements have been performed on hydrothermal vein minerals, their fluid inclusions, and their unaltered country rocks. The data, corrected for in situ grown 87 Sr since vein emplacement at 31 Ma (Sr initial ), reveal Sr isotope disequilibrium between the wall rocks of the veins (Monte Rosa metagranites: Sr initial approximately 0.7580, and paragneisses: Sr initial 0.7107-0.7129; Piemonte metabasites: Sr initial approximately 0.7038) and the vein-forming fluids (Sr initial = 0.7291 - 0.7079). At most vein localities, Sr initial of the vein minerals and fluid inclusions evolves monotonically away from the host-rock signatures with progressive vein growth, due to decreasing local fluid-rock interaction upon wall-rock armoring. These analyses demonstrate that fluid flow in the hydrothermal system was upward and the Sr initial value of the deep source fluid was less than, but probably close to, 0.7100. Since the Monte Rosa nappe has Sr initial values higher than 0.7107 and a vertical thickness of more than 10 km, the ore-bearing fluid must have derived from at least 10 km below the present outcrop level. We conclude that hydrothermal systems which produce Tertiary lode gold deposits in orogenic belts may have large vertical extents, perhaps up to the range recently reported for equivalent Archean deposits. When combined with published fluid inclusion compositions and Pb and He/Ar isotope data, our results rule out three of the previously proposed sources for the Brusson fluids: (1) metamorphic devolatilization of paragneisses and metagranites of the Monte Rosa nappe (vein Sr and Pb are too primitive and vein fluid chemistry is inconsistent), (2) fluid escape from coeval lamprophyres (vein Sr and Pb are too primitive), and (3) incursion of meteoric water (vein 36 Ar/ 3 He values are orders of magnitude too low). The only potential source rocks consistent with the new constraints and known to be abundant beneath the Monte Rosa nappe are calc-schists (0.7093 < Sr initial < 0.7104; model 238 U/ 204 Pb ratio, mu 2 , approximately 9.72) and intercalated metabasites of the Mesozoic Piemonte ophiolite nappe (Sr initial approximately 0.7038; 9.20 < mu 2 < 9.51). The Periadriatic syenogranitic intrusions that crop out 18 km southeast of Brusson have similar Sr initial values and are coeval with the gold deposits, but there is no evidence for the presence of such intrusions below Brusson. Our genetic scenario, in accord with the timing of orogenic events, is that regional Meso-Alpine metamorphism liberated CO 2 -H 2 O-NaCl-KCl-H 2 S fluids from the deep-seated Piemonte calc-schists via devolatilization reactions (yielding fluid Sr initial approximately 0.710, mu 2 approximately 9.72), and that gold was extracted as the fluids reacted with the intercalated Piemonte metabasites prior to rising through the Monte Rosa nappe to the sites of gold deposition.