Summary Differences in alkaline phosphatase activity of some commercial control sera were observed when comparing manual with AutoAnalyzer versions of the Kind and King procedure. This phenomenon was found to be associated with sera boosted with exogenous alkaline phosphatase and to be owing mainly to inadequate provision of buffer in some automated methods. It is suggested that a modification to the SMA procedure may lessen the problem, but the constraints of the system necessitate a compromise. There was better precision after the modification had been introduced. Awareness of this phenomenon is of paramount importance if commercial sera are to be used in any way in the calibration of automated systems for alkaline phosphatase assay.