Microculture tetrazolium assays are being widely exploited to investigate the mechanisms of both cell activation and cell damage. They are colorimetric assays which are based upon the bioreduction of a tetrazolium salt to an intensely coloured formazan. We contrast the responses obtainable with two new tetrazolium salts, MTS and XTT, when used on the rat lymphoma cell line (Nb2 cells), which has been activated by human growth hormone. These tetrazolium salts, unlike the more commonly used MTT, form soluble formazans upon bioreduction by the activated cells. This has the advantage that it eliminates the error-prone solubilisation step which is required for the microculture tetrazolium assays which employ MTT. Bioreduction of XTT and MTS usually requires addition of an intermediate electron acceptor, phenazine methosulphate (PMS). We found that the XTT/PMS, but not the MTS/PMS, reagent mixture was unstable. Nucleation and crystal formation in the XTT/PMS reagent mixture, prepared in DPBS, could occur within 1-3 min. This resulted in a decline in XTT-formazan production and manifested itself in the microculture tetrazolium assay as both poor within-assay precision and serious assay drift. Several features of the system suggested that the formation of charge-transfer complexes between XTT and PMS accounted for this instability. No such instability was encountered when MTS and PMS were mixed. We demonstrate that MTS/PMS provides microculture tetrazolium assays for hGH which are free from these serious artefacts and which are uniquely precise. In conclusion we therefore advocate the use of MTS in preference to XTT for the new generation of microculture tetrazolium assays.