The maximum deviation of experimental R value from its real value, which is defined as the ratio of tissue bound to tissue water tritium, has been calculated and verified experimentally by taking consideration of isotopic fractionation arised in the course of water separation. Experimental procedures examined for the purpose are the azeotropic distillation and lyophilization for the removal of tissue water and the oxidative combustion of organic residue either by thermal process or by low temperature plasma generation. Each procedure optimalized by obviating or correcting isotope effects as well as other sources of error has been tested with mixed standards and biological samples. By washing out the exchangeable tritium, also physically bound tritium, the precision and accuracy of R values are further improved.