Major element, trace element, and volatile concentrations in 187 glassy melt inclusions and 25 groundmass glasses from the 1980-86 eruption of Mount St. Helens are presented, together with 103 analyses of touching FE-Ti oxide pairs from the same samples. These data are used to evaluate the temporal evolution of the magmatic plumbing system beneath the volcano during 1980-86 and so provide a framework in which to interpret analyses of melt inclusions from the current (2004-2006) eruption. Major and trace element concentrations of all melt inclusions lie at the high SiO₂ end of the data array defined by eruptive products of the late Quaternary age from Mount St. Helens. For several major and trace elements, the glasses define a trend that is oblique to the whole-rock trend, indicating that different mineral assemblages were responsible for the two trends. The whole-rock trend can be ascribed to differentiation of hydrous basaltic parents in a deep-seated magma reservoir, probably at depths great enough to stabilize garnet. In contrast, the glass trends were generated by closed-system crystallization of the phenocryst and microlite mineral assemblages at low pressures. The dissolved H₂O content of the melt inclusions from 1980-86, as measured by the ion microprobe, ranges from 0 to 6.7 wt. percent, with the highest values obtained from the plinian phase of May 18, 1980. Water contents decrease with increasing SiO₂, consistent with decompression-driven crystallization. Preliminary data for dissolved CO₂ in melt inclusions from the May 18 plinian phase from August 7, 1980, indicate that X_H2O in a vapor phase was approximately constant at 0.80, irrespective of H₂O content, suggestive of closed-system degassing with a high bubble fraction or gas streaming through the subvolcanic system. Temperature and f _O2 estimates for touching Fe-Ti oxides show evidence for heating during crystallization owing to release of latent heat. Consequently, magmas with the highest microlite crystallinities record the highest temperatures. Magmas also become progressively reduced during ascent and degassing, probably as a result of redox equilibria between exsolving S-bearing gases and magmas. The lowest temperature oxides have f _O2 ≈ NNO, similar to high-temperature fumarole gases from the volcano. The temperature and f _O2 of the magma tapped by the plinian phase of May 18, 1980, are 870-875°C and NNO+0.8, respectively. The dissolved volatile contents of the melt inclusions have been used to calculate sealing pressures; that is, the pressure at which chemical exchange between inclusion and matrix melt ceased. These are greatest for the May 18 plinian magma (120 to 320 MPa); lower pressures are recorded by samples of the preplinian cryptodome and by all post-May 18 magmas. Magma crystallinity, calculated from melt-inclusion Rb contents, is negatively correlated with sealing pressure, consistent with decompression crystallization. Elevated contents of Li in melt inclusions from the cryptodome and post-May 18 samples are consistent with transfer of Li in a magmatic vapor phase from deeper parts of the magma system to magma stored at shallower levels. The Li enrichment attains its maximum extent at ~150 MPa, which is ascribed to separation of a single vapor phase into H₂ O-rich gas and dense Li-rich brine at the top of the magma column. There are striking correlations between melt-inclusion chemistry and monitoring data for the 1980-86 eruption. Dissolved SO₂ contents of melt inclusions from any given event, multiplied by the mass of magma erupted during that event, correlate with the measured flux of SO₂ at the surface, suggesting that magma degassing and melt-inclusion sealing are closely related in time and space. Textural and chemical evidence indicates that melt inclusions became effectively sealed (physically or kinetically) shortly before eruption. Thus by converting pressure to depth using a density model and edifice-loading algorithm for the volcano, changing depths of magma extraction with time can be tracked and compared to the seismic record. The plinian eruption of May 18, 1980, involved magma stored 5-11 km below sea level; this is inferred to be the subvolcanic magma chamber. The preceding eruptions, including the May 18, 1980, blast, involved magma withdrawal from the cryptodome and conduit down to 5 km below sea level. Subsequent 1980 eruptions tapped magma down to depths of ≤10 km below sea level. Tapping of magma stored deeper than 2 km below sea level stopped abruptly at the end of 1980, coincident with the onset of extensive shallow seismicity and a change from explosive to effusive eruption style from 1981 to 1986. Overall, the 1980-86 eruption is consistent with the evisceration of a thin, vertically extensive body of magma extending from 5 to at least 11 km below sea level and connected to the surface by a thin conduit. In the absence of sustained high magma-supply rates from depth, decompression crystallization of magma ascending through the system leads eventually to plugging of the conduit. The current eruption of Mount St. Helens shares some similarities with the 1981-86 dome-building phase of the previous eruption, in that there is extensive shallow seismicity and extrusion of highly crystalline material in the form of a sequence of flows and spines. Melt inclusions from the current eruption have low H₂ O contents, consistent with magma extraction from shallow depths. Highly enriched Li in melt inclusions suggests that vapor transport of Li is a characteristic feature of Mount St. Helens. Melt inclusions from the current eruption have subtly different trace-element chemistry from all but one of the 1980-86 melt inclusions, with steeper rareearth-element (REE) patterns and low U, Th, and high-fieldstrength elements (HFSE), indicating addition of a new melt component to the magma system. It is anticipated that increasing involvement of the new melt component will be evident as the current eruption proceeds.