For Part 1 see ibid., vol.10, p.219-30 (1989). The authors show that the technique of exciting a differential pressure transducer at one port to measure its differential gain yields incorrect results in the case of asymmetrical differential pressure transducers but is acceptable in the case of symmetrical transducers. They have measured the common-mode gain of a symmetrical and an asymmetrical differential pressure transducer from 1 Hz to 100 Hz, both directly and by computation from differential gain measurements. No compensation for common-mode error is necessary when a symmetrical transducer is used in mechanically ventilated neonates. The authors have also measured the frequency response of neonatal Fleisch and screen pneumotachographs connected to Validyne MP45 differential pressure transducers in air, 60% oxygen and 100% oxygen, and concluded that the effect of oxygen concentration is small below 40 Hz. However, the normalised frequency response of the flow transducer differs markedly from unity at frequencies inside the flow bandwidth generated by neonatal pressure cycled ventilators and dynamic correction is therefore necessary.