After decades of ordinary scientific interest, fluid resuscitation of patients with septic and haemorrhagic shock took centre stage in intensive care research at the turn of the millennium. By that time, resuscitation fluids were the mainstay of haemodynamic stabilisation, avoidance of vasopressors and treatment of hypovolaemia in patients in shock, but were accompanied by adverse events such as excessive tissue oedema. With the spread of early goal-directed therapy research intensified and it was realised that type, volume and timing of resuscitation fluids might affect the course and outcome of critically ill patients. At the same time, the importance of microvascular blood flow as target of resuscitation was accepted. Today, once-forbidden albumin is the recommended colloid in severe sepsis and septic shock, and the European Medical Agency is considering the removal of starch solutions from the European market because of an increased incidence of acute kidney injury and mortality. This is unprecedented, especially because the administration of low-molecular-weight starches seems to have advantages in indications other than sepsis, and because practices in fluid resuscitation have changed fundamentally since the negative starch studies. Crystalloids are still the mainstay of hypovolaemia treatment in critically ill patients, but awareness is increasing that electrolyte composition, strong ion gap, tonicity and the bicarbonate-substituting anion may have an effect on adverse effects and outcome. In haemorrhagic shock, the utilisation of crystalloids and colloids is retreating, and plasma and erythrocyte concentrates are gaining more importance in the resuscitation of the patient with acute bleeding. However, there are still influential voices warning against the liberal usage of plasma concentrates and erythrocytes in trauma and haemorrhagic shock. This review describes the evidence relating to fluid resuscitation in sepsis, septic shock and massive haemorrhage. Beside the scientific evidence based on clinical trials, possible effects on the microcirculation and, therefore, organ function will be illustrated and areas of future research highlighted. The critical appraisal of the existing evidence should enable the reader to choose the optimal volume substitution for an individual patient.