Random urine samples were obtained from 31 patients with neuroblastoma (newborn to 8 yr of age) and from 100 children without this tumor (newborn to 10 yr). The urine samples were studied for the presence of sulfate and glucuronide conjugates of homovanillic (HVA), dihydroxyphenylacetic, vanilmandelic, and vanillactic acids. The urinary concentrations of these acids were determined by capillary gas-chromatography before and after enzymatic treatment with glucuronidase and sulfatase. Concentrations of the "free" fraction and "total" urinary content of these acids were determined using the results from untreated and treated urines respectively. Age-related reference values were established for children without neuroblastoma. Fractions of the total content of urinary HVA (18-39%) and dihydroxyphenylacetic acid (36-66%) were excreted as glucuronides and/or sulfates by the control group, with the highest conjugated fractions found in the urine of young infants (0-3 months). Vanilmandelic was excreted mainly as "free" acid (unconjugated), whereas vanillactic acid was undetectable in almost all control samples. Patients with neuroblastoma also excreted a fraction of these acids as glucuronide and/or sulfate conjugates, (25% of urinary HVA, 39% of dihydroxyphenylacetic acid and 45% of vanillactic acid) whereas vanilmandelic acid was excreted only as "free" in controls. Determination of "total" rather than "free" urinary HVA was diagnostic in one neuroblastoma patient with borderline "free" HVA levels, whereas determination of "free" or "total" dihydroxyphenylacetic acid and vanillactic acid did not improve the diagnostic sensitivity in the cases examined. We conclude that it may be clinically useful to determine "total" urinary HVA in patients with borderline "free" HVA levels who are suspected of having neuroblastoma.