Inferences about total renal (venous and urinary) PGE2 output from determinations of urinary excretion rates (U PGE2 V) cannot be made unless the distribution of PGE2 between renal venous plasma and urine is known. Therefore, in the present study on intact kidneys of anesthetized dogs both urinary excretion of PGE2 and the renal venous output (the product of plasma flow and venous concentration of PGE2) was determined during low and high rates of renal PGE2 synthesis. PGE2 was measured in urine and arterial and renal venous plasma by radioimmunoassay during the following conditions: (1) Hydropenia. In the control condition U PGE2 V averaged 0.041 +/- 0.012 pmol/g . min and varied between 4 and 70% of the total PGE2 output. With infusion of arachidonic acid (AA, 160 micrograms/kg . min) into the renal artery total PGE2 output increased from 0.18 +/- 0.03 to 3.23 +/- 0.51 pmol/g . min, whereas arterial concentrations of PGE2 were unchanged. The urinary fraction still varied between 6 and 46% of total renal PGE2 output. (2) High urine flows caused by mannitol, saline or saline and ethacrynic acid (ECA) infusion. These procedures did not stimulate total renal PGE2 output and the urinary fraction varied between 4 and 49%. ECA combined with saline infusion increased the urinary fraction significantly to 34.7 +/- 4.0%. AA increased the total PGE2 output as during hydropenia, but the urinary fraction fell to 13% in 13 dogs and was unchanged at about 8% in six dogs. On average the urinary fraction of total PGE2 output was significantly lower than in hydropenia. Thus, the urinary fraction of total renal PGE2 output is not constant, and urinary excretion of PGE2 does not give reliable information about renal synthetic rates of prostaglandins.