Scapolite from the Nickel Plate gold-skarn deposit, British Columbia, and the Archean Hemlo gold deposit, Ontario, has been studied in detail. Chlorine-carbonate scapolite is commonly present at Nickel Plate in both the mineralized and barren exoskarn and endoskarn, although it is generally more abundant within, and adjacent to, the sulfide-gold orebodies. Less commonly, scapolite also occurs in rare, vuggy cavities developed in a limestone-boulder-bearing unit (the Copperfield breccia) and in marble beyond the outer limits of the skarn envelope. Scapolite equilibria involving coexisting garnet-plagioclase-calcite-quartz in the CASCH system yielded X(CO 2 ) of 0.04+ or -0.02 and temperatures of 500+ or -20 degrees C at a pressure of 1 kbar. Fluid inclusions in scapolite from exoskarn, endoskarn and a vuggy cavity yielded similar temperatures of homogenization (250-470 degrees C, and mostly between 320 and 400 degrees C) and salinities (15.6 to 19 wt.% NaCl equivalent). At Nickel Plate, there is no apparent correlation between the Cl content of scapolite and the spatial proximity of the mineral to ore; instead, Cl content corresponds to EqAn content which, in turn, relates to the anorthite component of associated plagioclase. We suggest that the Cl content of scapolite is not directly proportional to the chlorinity of the coexisting fluid, but rather is strongly controlled by crystal structure (i.e., by charge-balance constraints imposed by the scapolite framework). The occurrences of Cl-bearing scapolite at the Nickel Plate deposit also provide further support for chloride complexes in the transport and deposition of gold in skarn systems. Sulfate-carbonate scapolite (EqAn 65 ), coexisting plagioclase (An (sub 30-35) ) and calcite at the Hemlo deposit crystallized immediately after the peak regional metamorphism and predated gold mineralization. The sulfate content of the scapolite at Hemlo is directly controlled by coexisting mineral assemblages (i.e., anhydrite-bearing and anhydrite-absent) and is consistent with the experimental calibration of Kotel'nikov (1987).