A great deal has been learned about fluid resuscitation in the last several decades. The choice of an appropriate fluid for resuscitation in every given clinical situation has not yet been definitively determined but we can make some conclusions based on currently available data. It should again be emphasized that the goal of resuscitation of hypovolemic shock is quite clear regardless of the choice of fluid; to resuscitate the shock state as quickly as possible while at the same time minimizing the deleterious effects of fluid resuscitation on the pulmonary, renal, immunological and other systems. Animal experimental work reveals that use of colloid solutions to minimize pulmonary edema formation is ineffective, especially in instances where pulmonary capillary permeability is increased. Further, there are suggestions that colloid solutions may actually exacerbate pulmonary dysfunction following resuscitation by changing the characteristics of the pulmonary interstitium and the dynamics of fluid flux in the lung. This is entirely consistent with Starling's theory of a balance of hydrostatic and osmotic pressure, given what we now understand about the “other” or interstitial side of the Starling equation. Aside from the lungs, there are known side effects of various colloid solutions on other organs and body systems. In addition, questions remain about other possible associated short-term and long-term renal, coagulation, and immunological effects. Clinical studies using extravascular lung water as an objective parameter of pulmonary dysfunction show no correlation with fluid balance and no deleterious effects of crystalloid resuscitation. The relative cost of various resuscitation fluids should be a minor point when making therapeutic decisions. These relative costs, however, argue strongly for crystalloid therapy unless advantages for colloid fluids can be proved. Such proof is lacking to date. This is not to say that colloid solutions might never be called for. An occasional situation may arise in which sudden acute hypovolemia is associated with difficulties in obtaining good intravenous access. An example would be the victim of an automobile accident trapped for a prolonged period in a vehicle. Such situations impose practical limits on the amount of volume that can be infused. The use of colloid solutions to provide maximal intravascular volume restoration may be justified in such instances. In the vast majority of cases of hypovolemia, however, the balance of experimental, clinical, and practical considerations convincingly favor the use of a crystalloid solution for resuscitation in association with blood and clotting factors as needed.