The kinetics of the reaction of fully reduced membrane bound cytochrome oxidase with CO following photolysis of the fully reduced cytochrome oxidase-CO complex habe been re-examined by re-analysing the data of Clore and Chance (1978) Biochem. J. 175, 709-725) at six temperatures in the 178-203 K range simultaneously at only a single wavelength pair, 444-463 nm. The choice of the 444-463 nm wavelength pair was based on the fact that the absorbance change produced at 444-463 nm on photolysis of the CO complex is sufficiently large and the separation between monitoring and reference wavelengths sufficiently small to render the effects of any possible time dependent scattering changes insignificant. On the basis of our analysis only a two step mechanism (Model 1 of Clore and Chance (1978) Biochem. J. 175, 709-725) satisfies the triple requirement of a S.D. within the standard error of the data, a random distribution of residuals and good determination of the optimized parameters. The single step mechanism of De Fonseka and Chance (1978) Biochem. J. 175, 1137-1138) fails to satisfy all three requirements. The pure difference spectra of species Ic minus E, E minus IIc and Ic minus IIc are calculated from the computed kinetics of the individual species and repetitive slow wavelength scanning difference spectra (reaction sample minus the CO complex) taken during the course of the reaction of fully reduced cytochrome oxidase with CO at 176 K.