The benefits of inhaled therapy in ventilated neonates are recognized, but the reliability of drug delivery in nebulizer-ventilator circuits is uncertain. We quantified the effect of changing variables. Twenty-three freshly killed rabbits (1.15-1.9 kg) were ventilated via a tracheostomy by a pressure-limited, time-cycled ventilator (Neovent). A radioaerosol of 99Tcm pertechnetate from an Ultravent nebulizer (Mallinkrodt) was fed into the proximal ventilator tubing. Two 3-minute nebulizations at "standard settings" were followed by 2 at altered pressure, frequency, gas flow, I:E ratio, or position of the nebulizer in the circuit. Each nebulization was followed by a 3-minute gamma camera image and total deposited radioactivity was measured in excised lungs and trachea. Images demonstrated good peripheral aerosol deposition. At standard settings, lung deposition averaged 2.8% of the aerosol released. This was decreased markedly by reducing tidal volume (ventilator pressures) and residence time of aerosol (I:E ratio). Reduced gas flow decreased deposition slightly, presumably by increased particle size and marginally reduced tidal volume. Deposition did not change with increased frequency; increased minute ventilation was offset by decreased residence time of the aerosol. We conclude that the Ultravent nebulizer can be used to nebulize drugs in a standard neonatal circuit, although the dose delivered is small. Tidal volume and aerosol residence time are important determinants of aerosol delivery.