A number of mathematical distributions have been proposed for the description of the particle size distribution of unconsolidated sediments. However, few studies have mathematically described aeolian dust particle size distributions. Recent work has shown mathematically how the sequential fragmentation of materials leads to a Weibull distribution. Since the breakage of aggregates by saltating grains is a primary mode of aeolian dust production, we test the hypothesis that the Weibull distribution may be used to describe airborne soil grains. Surface samples were collected from 52 sites including soils, dirt roads and roadside ditches in the Southern High Plains of west Texas. The samples were tumbled in the Lubbock dust generation, sampling and analysis system to generate a dust cloud. The particle size distribution of the dust was measured in situ by laser diffraction and PM 10 concentration was determined gravimetrically (PM 10 = airborne particulate matter with diameter 10 "m). This study demonstrated that the Weibull cumulative distribution function (c.d.f.) is an excellent choice to describe the particle size distribution of dust suspended from mineral sediment. A Weibull c.d.f. used to describe the dust cloud size distribution, accounted for 94 per cent of the variation in estimates of particles 50 "m diameter. The fraction of dust particles 10 "m diameter, as estimated using the Weibull, was not correlated to suspended PM 10 concentration. However, the fraction of particles 10 "m was correlated with properties of the sediment from which the airborne dust was derived. As clay content increased, the total amount of suspended dust increased and the fraction of suspended particles 10 "m in the dust cloud decreased. Analyses of variance showed no significant differences (P `0Á05) among sampling locations (roads vs ditches vs soils) for cumulative fraction values for 2Á5, 10, 25, 30 and 50 "m diameter particles. However, the PM 10 concentration values were significantly different among dust generated from these locations. The road samples produced about twice the amount of PM 10 (490 mg m À3 ) as the soil or ditch samples (235 mg m À3 ). Published in 1999 by John Wiley & Sons, Ltd.