In-vivo and in-vitro checks on measurements of relaxation times in MRI have been carried out as part of quality control. For the in-vitro experiments we used 13 reagents consisting of varying concentrations of copper sulphate manganese sulphate, gadolinium solutions and various oils as well as water. For in-vivo measurements we used liver parenchyma and fat in three subjects. Eight different sequences were performed; intra- and interindividual variations and comparison of three similar MRI units showed average variations of 2-3% for T1 and 3-4% variations for T2 relaxation times. This margin of error makes quantitative MRI acceptable for clinical use.