Accuracy of a calculating system in transcribing output from a reaction rate analyser

A system for the mechanized calculation and transcription of output from a reaction rate analyser has been evaluated in the performance of 8 enzyme assays, many of which show anomalous kinetic behaviour with regard to "lag" or "accelerated" phases or "blank" reactions. For all of these enzymes, a cycle of three consecutive 18-second readings per assay proved optimal, the result being based upon the final output value unless this showed marked discrepancy with the second value, in which circumstances it is helpful to refer to a strip chart recording of the progress curve. Excellent agreement was obtained with manually-derived results and the system functioned without downtime for more than one year. Marked descrepancies between manual and mechanised results were invariably traced to technician error.