The Michaelis constant values for substrates of transketolase from human tissues were determined over a wide range of substrate concentrations. It is shown that K(m) values determined by other authors are significantly overestimated and explained why this is so.