In an open prospective study, adrenaline administration in ten patients with eleven episodes of septic shock was studied. Appropriate supportive therapy (antibiotics, laparotomies, parenteral alimentation, ventilation) was given as needed. Haemoglobin was kept at or about 12 g%, pulmonary capillary wedge pressure kept at approximately 15 mmHg, and cardiac index at greater than 4.5 l/min/m2. Only when systemic vascular resistance (SVR) dropped below 600 dyn. s. cm-5 was adrenaline given to raise the latter to no higher than 800 dyn. s. cm-5 and the adrenaline was titrated to this end point. Adrenaline was used at doses up to 0.47 microgram/kg/min for up to nineteen days. There was no reliable dose response curve for adrenaline: each septic insult needed different dosages. However, if high enough doses were given, SVR eventually increased. There was no deterioration in cardiac index nor further increase in pulse rate and no renal damage was demonstrated. Only one patient died in septic shock. Two others died from causes not directly related to sepsis and another two while still in hospital, but again not septic. Five patients were eventually discharged from hospital. Adrenaline can thus be used as a vasoconstrictor in septic shock without adverse effects, but initial doses have to be high and the effects measured and titrated carefully. Used this way, adrenaline provides time for the eradication of sepsis.