It seems reasonable to conclude that the classic underfill hypothesis, as initiated by ascites formation, or the overflow hypothesis, as initiated by a hepatorenal reflex, cannot explain the well-described hemodynamic and neurohumoral characteristics associated with the progression of the functional stages of cirrhosis from the compensated to decompensated to the hepatorenal syndrome state. The peripheral arterial vasodilation hypothesis, however, can explain the majority of findings over the spectrum of these functional stages of cirrhosis. In the early compensated stage of cirrhosis, there is evidence in experimental models that both a peripheral arterial vasodilation, primarily occurring in the splanchnic circulation, and a hepatorenal reflex may contribute to the early sodium and water retention in cirrhosis. However, studies in humans with decompensated cirrhosis demonstrate that central blood volume expansion with head-out water immersion and prevention of vasodilation with exogenous norepinephrine normalize sodium and water excretion. Although not measured, there would be no reason to believe that this combined maneuver would correct the increased sinusoidal pressure in cirrhosis, or act as the initiator of any hepatorenal reflex. The role of increased intrahepatic pressure as an initiator of a sodium retaining hepatorenal reflex may therefore be applicable to bile duct-ligated cirrhotic dogs but not to humans. Moreover, in a recent study in compensated cirrhotic patients receiving exogenous mineralocorticoid hormone, ascites developed in those patients with significantly lower peripheral vascular resistance independent of any differences in intrahepatic pressures. Studies in humans thus favor the peripheral arterial vasodilation hypothesis for both the initiation of sodium and water retention in cirrhosis and the spectrum of disease, including the compensated and decompensated states and the hepatorenal syndrome.