We have studied the influence of several manoeuvres modifying inspiratory and expiratory flow, with and without an external obstruction, on measured single-breath diffusing capacity (DLCO). By this method, we could evaluate the effects of increased inspiration or expiration time, independently of the influence of high (negative or positive) mouth pressures. Ten tests were performed in each of 11 trained normal subjects. The sequence of repeated manoeuvres was as follows: (1) basal; (2) inspiratory obstruction (IO); (3) expiratory obstruction (EO); (4) slow inspiration without obstruction (SI), and (5), slow expiration without obstruction (SE). A second, inverted sequence was then performed. The obstructive manoeuvres were forced by a unidirectional valve with a circular hole of 5 mm in diameter. Mouth pressures were registered. The slow flows were voluntarily controlled to be similar to the corresponding flows during the obstructive manoeuvre (approximately 15% of each vital capacity/second). Determinations of breath-holding time (tBH) were calculated by the methods of Jones-Meade, Ogilvie and the American Epidemiology Standardization Project (ESP), but using the same expiratory sample. We found that (DLCO) obtained with the three methods were statistically different (p less than 0.0001) from each other for all manoeuvres. Regarding the basal manoeuvre, there were statistical differences (p less than 0.01) with SI on the Jones-Meade, with IO on the ESP and with SI and IO on the Ogilvie timing methods. Between IO and SI, statistical differences (p less than 0.01) were observed in the three methods, but there was no statistical difference between SE and EO by any method.(ABSTRACT TRUNCATED AT 250 WORDS)