We investigated the effects of cold temperatures on microvascular protein permeability in the isolated constant-flow perfused cat hindlimb. The perfusates were 20% cat plasma-80% albumin-electrolyte solution (low-viscosity perfusate, approximately 1 cP) or whole blood (high-viscosity perfusate, approximately 4 cP). The time at low temperature (< 10-degrees-C) was < 3 h (short term) or > 5 h (long term). Decreases in the solvent drag reflection coefficient (sigma(f)) indicated increases in permeability. The sigma(f)'s were determined with the integral-mass balance method from measurement of changes in protein concentration and hematocrit induced by fluid filtration into the tissues. Short-term cold exposure did not increase permeability with either a low- or a high-viscosity perfusate, whereas long-term exposure with limb temperatures of approximately 5-degrees-C significantly increased permeability when the perfusate was whole blood. In addition, we verified our previous prediction that flow had to be reduced to 6-8 ml . min-1 . 100 g-1 to avoid the hydrostatic edema caused by short-term perfusion with whole blood at approximately 5-degrees-C. Also, we found that at approximately 3-degrees-C histamine's permeability-increasing effect was totally abolished, whereas at approximately 20-degrees-C this effect was partially inhibited. Hence, constant-flow perfusion at low temperature with whole blood can cause edema by a pressure-dependent mechanism, whereas long-term perfusion with this perfusate at low temperatures can cause a permeability increase that further compounds edema formation. Histamine is not responsible for this permeability increase.