Kinetic data from two submerged nitrifying filters in parallel used for maintaining the water quality in an experimental recycle smolt hatchery system is examined. The filter inlet total ammonia nitrogen concentrations recorded were up to 0.32 mg/l and correspond to those occurring in natural watercourses with superficial residence times ranging from 3 to 45 minutes. The smolts received a constant daily amount of nitrogen in their feed resulting in a constant ammonia load in the closed system. In addition to the system perturbations from feeding, further perturbations were introduced by altering the flowrate through the filters once every 24 hours. With the oxygen levels not limiting and with a constant pH the ammonia oxidation rate would be expected to be a function of the ammonia concentration alone. Excellent chi-square probability fits are found with liquid film diffusion limitations, but at much reduced rates from those anticipated. When comparing the rate constants of each of the two filters, half order kinetics is supported. Michaelis-Menten kinetics or the biological rate equation that accounts for internal diffusion are also supported, but only with significant differences in the number of enzyme sites in each filter. The use of an activator may well then be involved, but this contradicts literature mechanisms.