Synthetic starch solution, such as hydroxyethyl starch (HES), has been used clinically to restore cardiovascular volume in patients with hemorrhagic shock. Several HES solutions are available clinically, but each HES has a broad range of molecular mass fractions. We performed comparative studies of extremely low and high molecular HES to evaluate the effects of these HES solutions on lung lymph filtration during resuscitation. We prepared awake sheep with vascular monitoring and lung lymph fistulas. After baseline measurements, animals were bled from an arterial line to maintain shock. After 2 h of hemorrhagic period, the following three solutions were infused over 1 h, respectively. Experiment (Exp) 1 (n=6); low molecular HES; (molecular weight (MW) 70000, substitution fractions 0.5-0.55, Exp 2 (n=6); high molecular HES; (MW 450000, substitution fractions 0.65). Exp 3 (n=6); normal saline (NS). The quantity of solution was determined as the same volume of blood lost to induce hemorrhagic situation in each animal (Exp 1; 940+/-36 ml, Exp 2; 910+/-50 ml, Exp 3; 920+/-42 ml). Both low and high molecular HES could restore the systemic artery pressure and cardiac output, and significantly increased pulmonary microvascular pressure equally, which were significantly higher than those in normal saline. However, actual oncotic pressure gradient (plasma-lymph) rose transiently during low molecular HES infusion, while high molecular HES widened the oncotic pressure gradient even after the cessation of the infusion. Lung lymph flow during and after resuscitation with low molecular HES and NS rose significantly from the pre-shock baseline. There was no significant difference in increased lung lymph flow between low molecular HES and NS. However, lung lymph flow after high molecular HES was significantly less than that after low molecular HES. These data suggest that low molecular HES is as useful a plasma substitute as high molecular HES, but has a possibility to increase lung lymph filtration during the early phase of resuscitation.