When five cytotoxicity methods compared the toxicity of eosinophil granule major basic protein (MBP) and melittin to K562 and HL-60 cells, strikingly discrepant results were noted. Trypan blue staining, propidium iodide/CellTrackerGreen staining and incorporation of 14C-leucine assays indicated MBP damages > 75% of cells by 1 h. In contrast, 51Cr and lactic dehydrogenase (LDH) release assays indicated MBP damages most cells only at 20 h. All methods indicated melittin damages nearly all cells by 1 h. Further studies showed that without cell transfer, dye staining methods indicated MBP produces < 10% cytotoxicity after 4 h. A modified 14C-leucine assay, employing sodium dodecyl sulfate solubilization and trichloroacetic acid precipitation, showed lower cytotoxicity, 48%, at 4 h. Modified 51Cr and LDH assays showed increased cytotoxicities at 4 h, 34% and 58%, respectively. Overall, MBP's ability to cause molecular and cellular adhesion systematically confounds standard cytotoxicity measurements. However, the modified 14C-leucine assay provides a valid measure of MBP's cytotoxicity and may be useful for analyses of 'sticky' cytotoxins.