The plasma concentrations of neuropeptides (neurotensin, substance P, motilin, somatostatin, vasoactive intestinal peptide and gastrin-releasing peptide), the urinary excretion of 5-hydroxyindoleacetic acid and serotonin, and the platelet concentration of serotonin were compared in 133 patients who could be assigned to one of four groups. These groups were as follows: carcinoid tumors present; history of carcinoid tumors; miscellaneous tumors present; and non-tumor diseases. The test with the most sensitivity (i.e., patients with carcinoid tumors labeled positive) and the test with the most specificity (i.e., patients without carcinoid tumors labeled negative) for the presence of carcinoid tumors was determined. Urinary 5-hydroxyindoleacetic acid excretion had a sensitivity of 73 percent and a specificity of 100 percent; the plasma concentration of substance P had a sensitivity of 32 percent and a specificity of 85 percent; and the plasma concentration of neurotensin had a sensitivity of 41 percent and a specificity of 60 percent. Even when basal plasma concentrations of substance P and neurotensin were elevated, there was no additional increase of these neuropeptides prior to ethanol-induced facial flushing. Although measurements of plasma neuropeptide levels may be helpful in occasional patients with carcinoid tumors, it is concluded that measurements of serotonin overproduction--such as 5-hydroxyindoleacetic acid excretion--are of more general value.