Prehospital guidelines for the treatment of penetrating trauma recommend rapid volume resuscitation to normal blood pressure. There is evidence, however, that fluid resuscitation to normal blood pressure in the setting of uncontrolled hemorrhagic shock (UHS) causes increased bleeding, hemodilution, and mortality. To test this hypothesis, we evaluated the effects of blood pressure and hemodilution on survival in a rat model of UHS. UHS was produced in rats by preliminary bleed of 3 mL/100 g followed by a 75% tail amputation. Experimental design consisted of three phases: a prehospital phase, with uncontrolled bleeding and resuscitation to either 40 or 80 mm Hg with lactated Ringer's solution (LR) or lactated Ringer's solution and whole blood (WB); followed by a hospital phase, with control of the bleeding and continued resuscitation to mean arterial pressure (MAP) > 80 mm Hg and hematocrit near 30%; followed by a 3-day observation phase. There were four treatment groups, n = 8 in each group: group I, MAP = 80 mm Hg with LR only; group II, MAP = 80 mm Hg with WB and LR; group III, MAP = 40 mm Hg with LR only; and group IV, MAP = 40 mm Hg with WB and LR. All group I rats died within 2.5 hours. There were no significant differences in survival among groups II, III, and IV. Base deficit, arterial pH, and lactate levels were significantly worse in the rats resuscitated to a MAP of 80 mm Hg with LR (group I). The effects of blood pressure alone, hemodilution alone, and their interaction were significantly related to base deficit and arterial pH. Hemodilution, but not blood pressure as an end point in resuscitation, was significantly related to lactate levels. The high mortality in this model of uncontrolled hemorrhage was attributable to the effects of blood pressure, hemodilution, and the interaction between the two variables, rather than simply continued blood loss from increased hydrostatic pressure.