A procedure is developed that allows particles inhaled with realistic breath patterns to be sized by cascade impaction at a constant flow rate. This procedure is then used to examine the difference between particle sizes obtained with constant flow rate (step profile) versus actual-subject breath patterns for two dry powder inhalers DPIs; the Ventodisk® and Spiros® inhalers (delivering salbutamol sulphate). Aerosol inhaled from the DPIs by a breath simulator was combined with make-up air to provide 300 l/min. to a pair of virtual impactors. These impactors separate out particles in the nominal diameter range of 1–10 μm for sizing at 30 l/min. by a MOUDI cascade impactor, with filter collection of particles outside this range. Breathing patterns of ten subjects ranging in age from 6 to 17 years of age were measured and recorded using whole-body plethysmography while these volunteers inhaled through Ventodisk® and Spiros® inhalers. Particle sizes with four of these breath patterns, as well as several constant flow rate step profiles, were then obtained using the sizing apparatus with a realistic mouth–throat intake. Our results show that as long as the constant flow rates were near typical values occurring in the actual-subject breaths, particle sizes obtained with constant flow rates were not significantly different ( P>0.01) from those occurring with actual-subject breath patterns. Significant differences are present if constant flow rates unrepresentative of those expected during particle uptake with the actual-subject patterns are used with the Ventodisk®. These results show that judiciously chosen constant flow rates give rise to inertial particle size measurements that are equivalent to those obtained during actual-subject inhalation for the two types of DPIs tested.