Within the framework of the inner-hospital activity of our cardiac resuscitation team, arterial blood gas analyses at 10-minute intervals were obtained in 32 resuscitations. In the first ten minutes, the usual measures of cardiopulmonary resuscitation were carried out, and, thereafter the first blood gas analysis (BGA) was performed. If the pH value was less than 7.30, 1 mmol/kg body weight of sodium bicarbonate was infused, and a second BGA performed. Following this - as after all the other analyses - 0.5 mmol/kg body weight was infused on each occasion that the pH value was less than 7.30. From the beginning, an attempt was made to preserve hyperventilation (PaCO2 approximately 30 mmHg). It proved possible to resuscitate 15 patients without the instillation of sodium bicarbonate (group 1); seven patients were buffered before their own circulations were re-established (group 2). With comparable pH values in both groups, and similar times elapsed to resuscitation, the average adrenalin consumption in group 1 was 0.49 mg/min, in group 2 1.2 mg/min. 18.5% of our patients were discharged in a healthy condition. On the basis of these results, we conclude that buffering of acidosis at the start of a resuscitation attempt is hardly necessary. The use of sodium bicarbonate should be reserved for long resuscitations in the presence of (as far as possible) proven metabolic acidosis, and for the stabilisation phase.