When monitoring nitric oxide diffusing capacity (DL(NO)) in patients, it is necessary to distinguish natural biological variation from a real change in alveolar-membrane conductance. The short-term variability of single-breath DL(NO) has not been established. The aim was to determine the short-term variability DL(NO) in healthy subjects. Twelve healthy subjects performed single-breath hold diffusing capacity tests at rest over a 2-month period (eight separate sessions with 8+/-3 days between each session). Each subject inhaled 41+/-4 ppm NO and a standard diffusion mixture. DL(NO), which is a multiple of the membrane diffusing capacity for carbon monoxide (Dm(CO)), as well as carbon monoxide diffusing capacity (DL(CO)) and pulmonary capillary blood volume (V(c)) remained unaltered over the 2-month period (P>0.05). Reproducibility (calculated as 2.77 multiplied by the within-subject standard deviation) over eight sessions was 20, 5 and 8 mL min(-1)mmHg(-1) for DL(NO), DL(CO) and Dm(CO), respectively, and 19 mL for V(c) (when Dm(CO)=DL(NO)/2.42). DL(NO), DL(CO), Dm(CO) and V(c) remain unchanged over a period of 2 months. Since the inter-session variability is 20, 5 and 8 mL min(-1)mmHg(-1) for DL(NO), DL(CO) and Dm(CO), and 19 mL for V(c), a meaningful change should equal or exceed those values. While there is a small chance that week-to-week variation can also be partly due to mild pathophysiological changes, any differences that are below the reproducibility values are likely to be natural biological variation or technical variation of the equipment, rather than true physiological change.