The tetrazolium salt (MTT) method involving conversion of MTT to coloured formazan by cells serving as indirect measurements of cell growth/cell kill has been reported by several groups, although technical problems have been encountered. The present investigation was undertaken in order to delineate what laboratory variables have direct influence on the sensitivity and reproducibility of the method. The pH of the extraction buffer was of the utmost importance, since it was demonstrated that a pH greater than 5 would give rise to false signals. Furthermore, modifying the composition of the extraction buffer, all formazan dye grains were solubilised, totally. A direct comparison with published methods demonstrated that only the modified method would yield 100% higher signals without increasing the background. In contrast to previous reports, it was shown that phenol red does not interfere with the measurements and no washing steps are required since all ingredients can be added subsequently. Serum proteins at concentrations up to 25% have no influence on the result. All samples can be measured in an ELISA scanner at 570 nm with little intra-assay variation.