Data are summarized from five experiments in which polyethylene glycol (PEG) was fed to 51 subjects as a quantitative fecal marker. In three experiments PEG was fed continuously for 35 to 106 days. Preexperimental diet was assumed to have been completely excreted when the ratio of fecal PEG to fecal dry solids became constant. This took more than 7 days to occur in 50% of the subjects. Fecal recovery of PEG averaged only 93% of that consumed. PEG was a valid fecal marker for nitrogen, calcium, magnesium, iron, and zinc, but was less valid for sodium. In one experiment where PEG and chromic oxide (Cr2O3) were fed simultaneously, PEG was excreted faster on average than Cr2O3. In most subjects, small amounts of Cr2O3 were excreted sporadically for 42 days after Cr2O3 feeding had ceased. Inclusion of quantitative fecal markers permitted identification of the time when preexperimental diet had been completely eliminated, subjects who showed marked pooling of intestinal contents, and changes in the amount of fecal dry solids. However, an average fecal recovery of only 93% of the PEG or Cr2O3 consumed suggests that further work is needed before the recovery of either of these markers can be used with confidence for the correction of nutrient balance data.