The dose emitted from dry powder inhalers may be inhalation flow-dependent. Using an ex vivo method, the Electronic Lung, we have measured the aerodynamic characteristics of the emitted dose for both active constituents from Seretide Diskus (salmeterol xinafoate 50 mcg; fluticasone propionate 500 mcg) and Symbicort Turbuhaler (formoterol 6 mcg; budesonide 200 mcg). Electronic inhalation profiles were collected from 20 severe asthmatics (mean PEFR 53% predicted) when they inhaled using a placebo Seretide Diskus and a placebo Symbicort Turbuhaler. These were replayed in the Electronic Lung with the respective active inhaler in situ. Mean(S.D.) peak inhalation flow rates (PIFR) through the Diskus and Turbuhaler were 94.7(32.9) and 76.8(26.2) l min(-1), respectively. From the Electronic Lung the Diskus inhalation profiles provided a mean(S.D.) fine particle dose (FPD) for fluticasone propionate and salmeterol of 20.4(4.8) and 18.4(4.4)% labelled dose. For Turbuhaler inhalation profiles the FPD was 23.1(12.9) and 20.7(11.1)% labelled dose for budesonide and formoterol, respectively. The linear (p < 0.001) relationships between FPD against PIFR for budesonide and formoterol were 3 (p = 0.002) and 2.8 (p = 0.007) times steeper than fluticasone propionate and salmeterol, respectively. The results highlight a more significant effect of inspiratory flow on variable dosage emission when using the Symbicort Turbuhaler compared with the Seretide Diskus.