The May 18, 1980, eruption of Mount St. Helens (MSH) produced an extensive ashfall deposit in Washington, Idaho, and Montana with a minimum volume of 0.55 km3 (tephra). An unusual feature of the deposit is the occurrence of a second thickness maximum 325 km ENE of MSH near Ritzville, Washington. Grain size and component abundance analysis of samples along the main is very fine grained (mean size, 2 μm), poorly sorted, polymodal, and rich in glass shards and pumice fragments. A computer simulation of ash fallout from an atmospherically dispersed eruption plume was developed to evaluate various hypotheses for the origin of the distal ash characteristics, particularly the thickness versus distance relationship. The model was constrained by observations of the eruption column height, elevation of major ash transport, lateral spreading of the eruption plume, and atmospheric wind structure in the vicinity of MSH. Results of different simulations indicate that the second thickness maximum cannot be attributed to either decreased wind velocities over central Washington or injection of fine ash above the horizontal wind velocity maximum near the tropopause. For the model to fit the observed characteristics of the deposit, significant particle aggregation of ash finer than 63 μm must be invoked. The best fit occurs when ash less than 63 μm is aggregated into particles several hundred microns in diameter with a settling velocity of 0.35 m/s. Support for this process comes from the observation and collection of fragile ash clusters of similar size which fell at Pullman, Washington, during the May 18 eruption (Sorem, 1982). The premature fallout of fine ash as particle aggregates is a fundamental process in the origin of the grain size characteristics, variations in component abundances, and thickness versus distance relationship of the May 18 MSH ash fall deposit.