One of us [Engel, 1940] has recently worked out a method for determining directly the permeability of the capillaries in a given region. This method appeared suitable for reinvestigation of the problem of increased permeability of capillaries in 'traumatic shock'. We have therefore examined whether, after crushing a limb, there occurs an increase in the filtration through the capillaries either confined to the traumatized region and its immediate neighbourhood or generalized throughout the body.In traumatic shock a great quantity of plasma is assumed to escape from the blood vessels owing to their increased permeability. Evidence for this loss of plasma is of a twofold nature: the haemoconcentration found in man as well as in experimental animals during traumatic shock and the increase in weight of a traumatized limb. Both findings, however, are open to a different interpretation. Haemoconcentration may be due to escape of fluid constituents of blood, but it may also occur as a result of mobilization of blood corpuscles from the blood reservoirs, especially the spleen and the bone marrow. This possibility has not always been excluded as a cause of the haemoconcentration. The increase in weight of a traumatized limb, found by Blalock [1930] has been contested by some [Moon, 1938] but confirmed by others [Phemister, 1928;Harkins, 1941]. It results from loss into the crushed limb out of the vessels, of plasma as well as blood, in amounts sufficient to account for death. It is not possible to determine to what extent the increase in weight is due to escape of whole blood from broken or torn vessels or to escape of fluid through the capillary walls. Beard & Blalock [1931], by centrifuging crushed tissue in order to separate the extravascular fluid, attempted to show that the oedema fluid of a crushed limb has a similar composition as serum. This method, however, does not differentiate between the plasma of the extra-and intravascular blood in the tissue and the extravascular oedema fluid. Both will appear in the supernatant fluid on centrifugation. There is thus no direct evidence of an increased capillary permeability in shock.PH. CiI. 9