In the present study, we undertook to examine the relationship between urinary sodium retention and systemic hemodynamics in dogs with experimental portal cirrhosis induced by the sporadic feeding of dimethylnitrosamine. Sodium handling was studied by blanace techniques; plasma volume was measured serially with Evan's blue; and CO, blood pressure, CVP, and PVR were monitored through indwelling catheters. Six dogs were studied while standing quietly in a Pavlov sling, in a serial fashion starting 4 weeks after drug administration and continuing for some 3 months thereafter, until all dogs developed cirrhosis and ascites. Urinary sodium retention commenced generally between the ninth to twelfth week following the initation of treatment, but renal perfusion remained normal. Plasma volume expansion was noted within 1 week following the onset of sodium retention. Ascites was generally detected about 2 weeks following the initiation of sodium retention. No alteration in CO or PVR could be detected until ascites was present in significant amount. At that time, CO rose and PVR fell by about 20%. ABP tended to fall during the period of observation, but this was not significant. The initiation of sodium retention in this canine model does not depend on antecedent changes in CO or PVR.