We measured gluconeogenesis (GNG) in rats by mass isotopomer distribution analysis, which allows enrichment of the true biosynthetic precursor pool (hepatic cytosolic triose phosphates) to be determined. Fractional GNG from infused [3-13C]lactate, [1-13C]lactate, and [2-13C]glycerol was 88 +/- 2, 89 +/- 3, and 87 +/- 2%, respectively, after 48 h of fasting. [2-13C]Glycerol was the most efficient label and allowed measurement of rate of appearance of intrahepatic triose phosphate (Ra triose-P), by dilution. IV fructose (10-15 mg/kg/min) increased absolute GNG by 81-147%. Ra triose-P increased proportionately, but endogenous Ra triose-P was almost completely suppressed, suggesting feedback control. Interestingly, 15-17% of fructose was directly converted to glucose without entering hepatic triose-P. IV glucose reduced GNG and Ra triose-P. 24-h fasting reduced hepatic glucose production by half, but absolute GNG was unchanged due to increased fractional GNG (51-87%). Reduced hepatic glucose production was entirely due to decreased glycogen input, from 7.3 +/- 1.8 to 1.1 +/- 0.2 mg/kg/min. Ra triose-P fell during fasting, but efficiency of triose-P disposal into GNG increased, maintaining GNG constant. Secreted glucuronyl conjugates and plasma glucose results correlated closely. In summary, GNG and intrahepatic triose-P flux can be measured by mass isotopomer distribution analysis with [2-13C]glycerol.