The determination of ∆33S values in conjunction with δ34S values provides an effective means of evaluating the importance of S derivation from Archean country rocks in magmatic Ni-Cu-PGE ore genesis, particularly where δ34S values of potential S source rocks are near 0‰ and similar to most S of mantle derivation. However, ∆33S values of sulfides in some Archean rocks may also be near 0‰; care should be taken in the evaluation of S derivation from Archean rocks based solely on this parameter. Although anomalous ∆33S values are found primarily in Archean rocks, rare occurrences of nonzero ∆33S values have been identified in sulfides from younger rocks, and hence knowledge of the multiple S isotope composition of potential contaminants is essential for an accurate assessment of S contamination of mantle-derived magmas. Restricted ranges in ∆33S values may be characteristic of particular Archean formations, and when used in conjunction with mixing calculations may fingerprint a distinct S source rock. Spatial variability in ∆33S values within different types of mineralized systems may be caused by the emplacement of isotopically distinct magma pulses, by locally variable mixing proportions, and the addition of secondary sulfide to the primary assemblage. In the case of conduit-type deposits in the Midcontinent rift system, three-component mixing calculations show that both high and low ∆33S values in some units that are similar to end-member values in the Archean country rocks must be a function of assimilation of primarily Archean S even though Proterozoic carbonaceous sedimentary country rocks contain in excess of 1 wt % S and are particularly viable S source rocks.