The peroxidase-NADH oscillator examined here initially consists of a well-mixed aqueous solution of native horseradish peroxidase, reduced beta-nicotinamide adenine dinucleotide (NADH), methylene blue (MBC), and dissolved oxygen combined in a semi-batch reactor under a set of standard conditions. Hydrogen peroxide and superoxide have been implicated as important chemical intermediates. A comprehensive model which includes such intermediates and all initial chemical species has appeared elsewhere. To experimentally explore the role of hydrogen peroxide in the oscillator, H2O2 was substituted for MB(+) as an initial ingredient. This substitution allows relatively small, quasi-sinusoidal oscillations sensitive to the oxygen mass transport constant, and predicted earlier in a theoretical model. The oscillations become much larger when MB(+) is added, suggesting that MB(+) might serve as a chemical mediator between the small oscillations seen when H2O2 is substituted for MB(+), and the relatively large oscillations observed when MB(+) is present. Catalase and superoxide dismutase are used as enzymatic scavengers for H2O2 and O-2(.-), respectively. The enzymes are added individually to a working oscillator at oxygen minima and maxima to examine the roles and approximate the concentrations of H2O2 and O-2(.-) For the enzyme addition experiments, a perturbation model for oxygen behavior is proposed and applied to the interpretation of experimental data. Two methods of analysis for the addition of the enzyme probes indicate a higher concentration of H2O2 and O-2(.-) at oxygen maxima than at minima. Comparison of experimental and simulated data indicate that the relatively simple model presented here is a reasonable, yet apparently incomplete, representation of oxygen dynamics for the addition of scavenger enzymes to this oscillator.