To test the hypothesis that fecal magnesium excretion is related to fecal volume, fecal specimens were collected from healthy persons who drank 32 single doses of poorly absorbed carbohydrate (mannitol, lactulose or raffinose): the concentration of Mg(++) (Y) in fecal water (X) was linearly and inversely related to fecal volume-that is, Y = 42-0.03 X. In contrast, after drinking magnesium sulfate, the concentration of Mg(++) in fecal water rose and, except after the lowest ingested dose of 10 mmol, the points relating Mg(++) concentration and fecal volume were outside the upper 95% confidence limit of the linear regression line. These findings could be useful when physicians are analyzing stool specimens for supportive evidence of magnesium misuse in baffling cases of diarrhea.