The theoretical efficacy of hypertonic saline (HS) resuscitation for hemorrhagic shock purportedly stems from the osmolar extraction of intracellular fluid into the plasma. This hypothesis presumes a concomitant expansion of the interstitial fluid space. Colloid resuscitation, in theory, expands the plasma volume by extracting interstitial fluid. These hypotheses were tested in a canine-modified Wigger's model of hemorrhagic shock. Forty, male, splenectomized dogs were anesthetized and instrumented. Animals underwent a baseline equilibration period followed by shock for 120 minutes. Each animal was randomized to one of four groups and received equal amounts of Na+ either as lactated Ringer's (LR) solution, 10% dextran 40 (Dex) in normal saline, 7.5% saline (HS), or 7.5% saline plus Dex (HSD). Parameters measured at baseline, shock, and at postresuscitation 30 minutes, 60 minutes, 90 minutes, and 120 minutes, included: mean pressure (MAP), output, pulmonary capillary wedge pressure, prenodal skin lymph flow, serum and lymph albumin, wet-to-dry skin ratios, and plasma volume. MAP, cardiac output, and plasma volume were most quickly restored with LR and Dex resuscitation (MAP = 106 and 118 mm Hg) compared to HS and HSD (MAP = 98 and 92 mm Hg). Lymph flow and lymph albumin flux were best restored with LR and HSD (mean = 85 and 48 microL/min) compared to Dex and HS (mean = 36 and 37 microL/min). Wet/dry skin ratios were greatest at 60 minutes in the LR group but similar at 120 minutes in all four groups. These data suggest that interstitial fluid space remains contracted during the first hour after HS, HSD, and Dex resuscitation compared with LR resuscitation, even though the restoration of plasma volume, MAP, and cardiac output is greatest with the Dex regimen. Further studies with total body water and intracellular water are needed in this model.