Migration from can-coatings into retorted canned food simulants (canned oil and water, 121 degrees C, 30 min) was investigated through HPLC with a fluorescence detector and evaporative light scattering detector, and by measurements of residue on evaporation and consumption of potassium permanganate. HPLC analysis revealed that migration into the canned oil was hundreds of times more than that into n-heptane (25 degrees C, 60 min, the official test conditions according to the Japanese Food Sanitation Law), whereas it was similar to the migration into isooctane-butyl acetate mixtures (60 degrees C, 60 min), and that migration into the canned water was several times more than that into water (95 degrees C, 30 min, the official test conditions). Residue on evaporation for the n-heptane extract was several-fold lower than 30 ppm (the official limit), whereas that for the isooctane-butyl acetate mixtures exceeded 30 ppm. Consumption of potassium permanganate for the canned water was 30 times higher than that for the water extract (95 degrees C, 30 min). The official test conditions for can-coatings, in particular the use of n-heptane as an oil simulant, were suggested to lead to substantial underestimation of migration into canned food.