We evaluated two kinetic methods for determining total amylase activity and isoenzyme composition in serum. Stability studies of reagents for measuring total activity indicate that reagents containing 4-nitrophenyl-alpha-glucosides or enzyme-linked reagents can be stored only for seven days at 4 degrees C. Methods based on 4-nitrophenyl-alpha-glucoside substrates cannot be used if the reagent absorbance at 405 nm exceeds 2. However, in the alpha-amylase EPS method (Boehringer Mannheim) an ethylidene-protected 4-nitrophenyl-alpha-D-maltoheptaoside substrate is stable for up to 28 days after reconstitution. Further studies indicated that the Amylase-DS (Beckman) and the alpha-Amylase EPS standard curves are linear to at least six times the upper limit of the reference interval. Within-batch imprecision (CV less than 1.1%) and between-batch imprecision (CV less than 3.3%) for these two methods are comparable with those for other kinetic methods, and there is excellent correlation (r2 = 0.983) between the two methods. The reference interval, determined by use of samples from 90 healthy blood donors, is 31 to 141 U/L for the amylase-DS method, 22 to 92 U/L for the alpha-Amylase EPS method. We also used these two methods to measure amylase isoenzymes after inhibiting the salivary isoenzyme with either a lectin or a monoclonal antibody. We found the monoclonal antibody method more specific than the lectin inhibition method for determining the isoenzymes.