Adhering to a galactose-free diet by strictly avoiding dairy products and known hidden sources of galac-tose does not completely normalize galactose-1-phosphate (gal-1-P) in erythrocytes from patients with galactosemia. Major neurological complications, even in the best treated patients, are threatening a good clinical outcome and dictate a continuous search for leaks in the dietary regimen. Raffinose and stachyose, present in important amounts in various vegetables, contain alpha-1,4 linked galactose which is cleaved only by bacterial alpha-galactosidases, presumably in the lower part of the gut. In order to test the hypothesis whether galactose released from raffinose and stachyose could be a source of absorbed galactose and a cause of elevated gal-1-P six patients with galactosemia (aged 6-24 years), underwent a raffinose- and stachyose-poor dietary regimen for 2 weeks. Before, after, and during the test period, the daily intake of stachyose and raffinose as well of protein, carbohydrate, fat and minerals was calculated from food protocols obtained from the patients. Plasma galactose and erythrocyte gal-1-P were measured at the end of the three test phases. Stachyose and raffinose intake was reduced to 5%-10% during the experimental diet, which was well tolerated, except for constipation in some patients. In five of the six patients gal-1-P in erythrocytes was somewhat lower (statistically not significant) during the test phase than during regular diet while plasma galactose remained unchanged. Galactose released from raffinose and stachyose may be absorbed and contribute to elevated gal-1-P values in erythrocytes of galactosemic patients.