To identify the most efficient device for the delivery of bronchodilator aerosol to nonventilated preterm infants with chronic lung disease, we compared the metered dose inhaler (MDI) used in conjunction with a non-valved spacer, an ultrasonic nebulizer with a small medication cup, and two jet nebulizers. The subjects were enrolled in two double-blind randomized crossover studies. In study A (n=10), each infant was given a nominal dose of 200 microg of salbutamol by a MDI (Ventolin) at 4 h intervals, and in random sequence via an Aerochamber (Neonatal Aerochamber) with its one-way valve removed, an ultrasonic nebulizer with a small cone-shaped medication cup (Siemens Electronics), and a jet nebulizer (Side-stream). Their functional residual capacity (FRC) and static respiratory system mechanics were measured before, and at 15, 30, 60, and 120 min after aerosol delivery. Study B (n=10) was carried out in an identical manner, but with a different jet nebulizer (Hudson). In both studies, administration of salbutamol aerosol via the MDI Aerochamber or ultrasonic nebulizer resulted in a significantly greater reduction in respiratory system resistance than via jet nebulizers. Furthermore, the use of MDI Aerochamber or ultrasonic nebulizer was associated with a greater degree of post-treatment tachycardia and improvement in FRC. The bronchodilating effect of salbutamol delivered via the ultrasonic nebulizer appeared to be slightly greater than that via the MDI-Aerochamber, receiving significance only in Study B. We conclude that both the metered-dose inhaler used with a nonvalved Aerochamber and the ultrasonic nebulizer with a small medication cup are both more efficient than the jet nebulizers in preterm infants.