Background: Lung clearance index (LCI), obtained by multiple breath washout testing (MBW), is a sensitive measure of lung disease in infants. It has been identified as a particularly suitable endpoint for clinical trials in cystic fibrosis (CF), but has potential applications in many other conditions. However, MBW in infants presents a number of technical challenges. Conventional MBW is based on simultaneous measurement of flow and gas. These two signals are then aligned and combined to derive expired gas volumes and measures of ventilation inhomogeneity: this process becomes increasingly vulnerable to errors in gas signal alignment at rapid respiratory rates. At present, no existing system for infant MBW meets all the criteria set out in international guidelines, and there is no simple method of assessing lung function outside research laboratories in this population. This thesis describes an alternative method of performing MBW in infants. In this method, expired gas is collected and analysed to derive functional residual capacity (FRC) and LCI. There is no need to simultaneously measure flow, and therefore no need for the complicated step of integrating flow and gas signals. Dead space is also significantly reduced by removing the flowmeter. Methods: In the first phase of testing, an existing lung model was modified to generate realistic infant breathing parameters with high accuracy. The prototype system was modified to improve accuracy and subsequently tested at FRC of 100-250mls with respiratory rates of 20-60min-1. In the second phase, testing proceeded to an in vivo pilot study of the novel method in children with cystic fibrosis and healthy controls. Practical applicability of the system was determined by the number of successful duplicate tests, and within-subject repeatability. Comparison was made with LCI measurements obtained using a respiratory mass spectrometer, currently considered the gold standard for infant LCI. Results: In a total of 103 tests performed in the lung model, overall mean error (standard deviation) of FRC measurement was -1.0(3.3)%, with 90% of tests falling within +/-5%. 13 patients were excluded from the clinical study due to being unsedated or inadequately sedated and therefore failing to tolerate the test. A total of 25 patients (7 children with CF, 18 healthy control children) were deemed to be adequately sedated at the start of the test, of these 20 patients (7 with CF) successfully underwent duplicate testing (80% success rate). Mean FRC for healthy controls was 19.5ml/kg, and mean LCI 6.45. For children with CF, mean FRC was 21.8ml/kg and mean LCI 6.98. Mean within-subject coefficient of variation for FRC was 7.18% and for LCI 5.94%. Of 4 infants assessed with both the novel method and the respiratory mass spectrometer, there was good correlation in FRC measurement (mean difference -8.1%). Comparison of LCI with the mass spectrometer was affected by technical difficulties with the test; in those patients who underwent technically adequate tests with both methods, mean difference in LCI between the two methods was 1.65%. Discussion: FRC measurement using the novel method has superior accuracy in vitro than previously described systems. Data from the pilot study suggest that this is a feasible and reproducible method of performing LCI in infants and young children, as long as they are adequately sedated. Results in both children with CF and controls fall within the expected range, and well within accuracy limits set by international guidelines. However, the system and testing protocol could be further improved to reduce the number of technically inadequate tests having to be excluded. This could provide a more accessible alternative to previously described systems for infant MBW.