After the March-April 1986 explosive eruption a comprehensive gas study at Augustine was undertaken in the summers of 1986 and 1987. Airborne COSPEC measurements indicate that passive SO2 emission rates declined exponentially during this period from 380??45 metric tons/day (T/D) on 7/24/86 to 27??6 T/D on 8/24/87. These data are consistent with the hypothesis that the Augustine magma reservoir has become more degassed as volcanic activity decreased after the spring 1986 eruption. Gas samples collected in 1987 from an 870??C fumarole on the andesitic lava dome show various degrees of disequilibrium due to oxidation of reduced gas species and condensation (and loss) of H2O in the intake tube of the sampling apparatus. Thermochemical restoration of the data permits removal of these effects to infer an equilibrium composition of the gases. Although not conclusive, this restoration is consistent with the idea that the gases were in equilibrium at 870??C with an oxygen fugacity near the Ni-NiO buffer. These restored gas compositions show that, relative to other convergent plate volcanoes, the Augustine gases are very HCl rich (5.3-6.0 mol% HCl), S rich (7.1 mol% total S), and H2O poor (83.9-84.8 mol% H2O). Values of ??D and ??18O suggest that the H2O in the dome gases is a mixture of primary magmatic water (PMW) and local seawater. Part of the Cl in the Augustine volcanic gases probably comes from this shallow seawater source. Additional Cl may come from subducted oceanic crust because data by Johnston (1978) show that Cl-rich glass inclusions in olivine crystals contain hornblende, which is evidence for a deep source (>25km) for part of the Cl. Gas samples collected in 1986 from 390??-642??C fumaroles on a ramp surrounding the inner summit crater have been oxidized so severely that restoration to an equilibrium composition is not possible. H and O isotope data suggest that these gases are variable mixtures of seawater, FMW, and meteoric steam. These samples are much more H2O-rich (92%-97% H2O) than the dome gases, possibly due to a larger meteoric steam component. The 1986 samples also have higher Cl/S, S/C, and F/Cl ratios, which imply that the magmatic component in these gases is from the more degassed 1976 magma. Thus, the 1987 samples from the lava dome are better indicators than the 1986 samples of degassing within the Augustine magma reservoir, even though they were collected a year later and contain a significant seawater component. Future gas studies at Augustine should emphasize fumaroles on active lava domes. Condensates collected from the same lava-dome fumarole have enrichments ot 107-102 in Cl, Br, F, B, Cd, As, S, Bi, Pb, Sb, Mo, Zn, Cu, K, Li, Na, Si, and Ni. Lower-temperature (200??-650??C) fumaroles around the volcano are generally less enriched in highly volatile elements. However, these lower-termperature fumaroles have higher concentration of rock-forming elements, probably derived from the wall rock. ?? 1990 Springer-Verlag.