Sulfur isotope analyses have been performed on 64 monomineralic concentrates from 37 samples that are representative of mineralization in time and space at El Salvador. The hypogene sulfates (mean +10.7ppm; range +7.3 to +17.0ppm) are enriched in 34 S relative to supergene sulfates (-0.7ppm>; -4.6 to +3.6ppm) and to hypogene sulfides (-3.0ppm; -10.1 to -0.3ppm). Coexisting hypogene sulfides are increasingly depleted in 34 S in the order molybdenite, pyrite, chalcopyrite, and bornite. The isotopic data suggest that sulfur in the supergene sulfates was largely derived from the oxidation of hypogene sulfides and that supergene "chalcocite" probably replaced hypogene chalcopyrite or bornite, but not pyrite. Isotopic temperature estimates from sulfate-sulfide fractionation pairs range from 400 degrees to 570 degrees C and are only in crude agreement with temperatures (<300 degrees to >600 degrees C) indicated by other geologic evidence. Those estimated from pyrite-chalcopyrite fractionation pairs (95 degrees to 185 degrees C) are much too low. Fractionation between 13 coexisting hypogene sulfate-sulfide assemblages (21 mineral pairs) defines a rather narrow band in f (sub O 2 ) -pH-T space and suggests that f (sub O 2 ) and pH acted as internally controlled variables throughout mineralization. Mass balance estimates of delta 34 S (sub Sigma S) indicate a value of about +6 per mil for the sulfate zone and a value probably significantly heavier than 0 per mil for the entire deposit as presently exposed. The delta 34 S per mil values of coexisting hypogene sulfate and sulfide pairs approximate linear trends when plotted against their respective delta (delta ) values. These trends suggest that Early anhydrite-chalcopyrite-bornite assemblages were formed from a sulfur reservoir having delta 34 S (sub Sigma S) of approximately +1.6 per mil whereas Late anhydrite-pyrite-chalcopyrite assemblages formed from a reservoir +6.8 per mil delta 34 (sub Sigma S) . Speculative interpretation suggests that Late sulfur was derived either from remobilization of Early assemblages below the deepest levels of exposure or from volcanic wall rocks surrounding the deposit, rather than from continued emanations from the underlying magma chamber that was the source of Early mineralization. However, at least one totally different interpretation of these data is possible. Recent experimental work by Ohmoto and Rye (1975, written and oral commun.) indicates that our delta 34 S per mil values for pyrite may require a correction factor, which would reduce both Early and Late sulfate-sulfide assemblages to approximately single linear trends. This would imply that the underlying magma chamber continued to be the predominant source of sulfur (delta 34 S (sub Sigma S) [asymp] +2ppm) throughout the entire sequence of alteration-mineralization. The isotopic data do not show any consistent trends of 34 S depletion with either paragenesis or zoning that would suggest a restricted reservoir of sulfur in the hydrothermal system. More questions than answers are provided by these data.