The pyroclastic deposits of the 1300 B.P. eruption of Newberry Volcano, OR, USA, contain minor amounts of obsidian (1-6 wt.%). The volatile (H 2 O and CO 2 ) contents and textures of these clasts vary considerably. FTIR measurements of H 2 O in obsidian pyroclasts range from 0.1 to 1.5 wt.% indicating equilibration pressures ≤20 MPa. CO 2 contents are low (<10 ppm) except in clasts that also contain xenolith powder that provided a local CO 2 source. Obsidian clasts exhibit a range of color and textural types that changed in relative proportion as the eruption progressed. Together these data indicate that there were multiple origins of obsidian and that the dominant source changed during the eruption. Early in the eruption, obsidian was almost entirely black or grey (microlite-bearing) and probably derived from dikes or wall rock fractures filled with vanguard magma or tuffisite that, together with wall rocks, were eroded and incorporated into the eruption column as the vent widened. Later in the eruption, following a brief cessation of activity, the proportion of obsidian to wallrock lithic clasts increased and new types of obsidian dominated, types that represent remnants of a shallow conduit plug, welded fallback material from within the conduit, and sheared and degassed magma from near the conduit walls. Analysis of bubble shapes preserved within obsidian indicates that shear stresses and shear rates varied by over two orders of magnitude, with maxima of 88 kPa and 10 −2.3 s −1 , respectively, based on an assumed magma temperature of 850°C. Furthermore, the highest shear rates and stresses, and the shortest flow times (10-20 min), are preserved in clasts that also contain wall rock. The longest deformation times (5 and 8 h) correspond to two microlite-rich clasts, suggesting that the higher microlite content results from slower ascent rates and/or longer magma residence times at shallow levels. Differences between obsidian pyroclasts from the Newberry eruption and those of the Mono Craters may relate to the nature of the conduit feeding the two events. From this comparison, we conclude that obsidian can provide information on time scales and mechanisms of pre-fragmentation magma ascent.