The mechanism of ascites formation has been a subject of interest for many years and remains a challenge to the investigator. In previous reports (1, 2) on 61 patients with cirrhosis of the liver, it was pointed out that those with ascites have lower values for serum albumin than do those without ascites and that the tendency to diuresis (and the loss of ascites) was correlated with a rise of the serum albumin level. The "critical" point of diuresis was found to be about 3.1 grams per cent. However, the slowness of diuresis, frequently taking place over a period of several months, did not provide a sharp endpoint for establishing a critical level. Although other factors, such as an increased portal capillary pressure, were considered to be contributory, it seemed likely that the colloid osmotic pressure of the plasma was the chief determinant for the presence or absence of ascites.There are several objections to this concept: the rise of serum albumin did not regularly precede a diuresis although it appeared to accompany the phenomenon; wider experience in a series of 115 cases with ascites revealed 7 cases in which diuresis failed to take place even though the serum albumin rose above 3.5 grams per cent; and, conversely, there were 3 cases in which diuresis took place when values of serum albumin were less than 3.0 grams per cent. Ralli and her associates (3) have cited instances in which diuresis with loss of ascites took place at low values for serum albumin in patients receiving intravenous liver extract. Furthermore, they presented evidence suggesting that anti-diuretic substances might play a role in the formation of ascites.Concentrated human serum albumin solution has provided a means for testing directly the relation of serum albumin and colloid osmotic pressure to ascites formation, for by injecting albumin, the colloid osmotic pressure of the plasma can be raised abruptly. Concentrated serum albumin has been employed by others with the aim of determining its place as a therapeutic agent for cirrhosis of the liver with ascites. Janeway and his associates (4) reported that standard concentrated human albumin solution administered to 6 patients with portal cirrhosis produced no apparent change in ascites formation. The total dosage in 2 instances was 350 and 950 grams. Thorn and his co-workers (5) administered a salt-poor albumin solution to 5 patients with various types of liver cirrhosis associated with ascites and edema. A transient diuresis occurred in all 5 patients after a single infusion of 50 grams and after 3 daily infusions of 50 grams of albumin. The fluid seemed to have been mobilized from the peripheral edema rather than from the ascitps. Two of these patients subsequently lost their ascites after receiving a total dosage of 450 and 500 grams in 27 and 10 days, respectively. Gibson (6) and Kunkel (7) also report that prompt diuresis and loss of ascites occurred in certain patients after the administration of standard concentrated human albumin solution.The present study was designed to make ...