We coupled a kinetic procedure to a selective inhibiting method for determining amylase isoenzymes in biological samples, using 4-nitrophenylmaltopentaoside plus 4-nitrophenylmaltohexaoside as substrate and a wheat-germ selective inhibitor with the Gilford S-III spectrophotometer. On plotting remaining amylase activities/total amylase activities (R/T) vs pancreatic amylase activities/salivary amylase activities (P/S) ratios, we found the curve to be linear for P/S ratios from 0.2 to 5. The inhibition rate of amylase inhibitor was constant in solutions having total amylase activities between 20 and (at least) 900 U/L. CVs were 3.1 to 7.1% for pancreatic amylase and 2.0 to 12.9% for salivary amylase in serum. Correlation with the Phadebas method was excellent (r = 0.99) for both pancreatic and salivary amylase. We also automated this procedure in an Hitachi 705 analyzer and correlated the results (r = 0.99) with those by our manual method.