Size and time-resolved aerosol compositional measurements conducted during the 2006 eruption of Augustine Volcano provide quantitative information on the size and concentration of the fine volcanic ash emitted during the eruption and carried and deposited downwind. These data can be used as a starting point to attempt to validate volcanic ash transport models. For the 2006 eruption of Augustine Volcano, an island volcano in south-central Alaska, size and time-resolved aerosol measurements were made using an eight-stage (0.09–0.26, 0.26–0.34, 0.34–0.56, 0.56–0.75, 0.75–1.15, 1.15–2.5, 2.5–5.0, and 5.0–35.0μm in aerodynamic diameter) Davis Rotating Unit for Monitoring (DRUM) aerosol impactor deployed near ground level in Homer, Alaska, approximately 110km east–northeast of the volcano. The aerosol samples collected by the DRUM impactor were analyzed for mass and elemental composition every 90min during a four-week sampling period from January 13 to February 11, 2006, that spanned several explosive episodes during the 2006 eruption. The collected aerosols showed that the size distribution of the volcanic ash fallout changed during this period of eruption. Ash had its highest concentrations in the largest size fraction (5.0–35.0μm) with no ash present in the less than 1.15μm size fractions during the short-lived explosive events. In contrast, during the continuous ash emission phase, concentrations of volcanic ash were more significant in the less than 1.15μm size fractions. Settling velocities dictate that the smaller size particles will transport far from the volcano and, unlike the larger particles, not be retained in the proximal stratigraphic record. These results show that volcanic ash transport and dispersion (VATD) model predictions based on massless tracer particles, such as the predictions from the PUFF VATD model, provide a good first-order approximation of the transport of both large and small volcanic ash particles. Unfortunately, the concentration of particles in different size fractions depends on eruptive style and ash generation processes, so predicting the actual behavior of the ash particles from an eruption requires additional information in real time.