1. Arterial pH oscillations have been monitored in vivo in patients with well defined chronic obstructive bronchitis, asthma and clinical emphysema. 2. the patients with clinical emphysema were shown to differ from those with chronic obstructive bronchitis on the basis of a number of clinical and physiological criteria. 3. Patients with asthma showed least attenuation of their pH oscillations as a group, in contrast to emphysematous patients who showed most attenuation. in patients with clinical emphysema the attenuation was relatively homogeneous. the patients with chronic obstructive bronchitis showed the full range from normal oscillations (zero attenuation) to zero (complete attenuation). 4. the amplitude and approximate rate of change of upslope of the P aco 2 oscillations in vivo were calculated, from measured pH oscillation amplitudes, using buffer slope values in vitro from Siggaard Anderson [(1962, 1963) Scandinavian Journal of Clinical and Laboratory Investigation , 14 , 598–604; 15 , 211–217], then dividing the P aco 2 amplitude by half the respiratory period. 5. Mean arterial P co 2 in vitro showed a very strong correlation with the downslope of the pH oscillation in vivo (calculated as for P aco 2 upslope) in patients without clinical emphysema. This correlation would be expected to some extent, owing to the logarithmic relationship of P aco 2 oscillations to pH oscillations. However, the mean arterial P co 2 also showed a very strong correlation with the upslope of the calculated Paco 2 oscillations, again excluding patients with clinical emphysema.