Time-resolved resonance Raman (RR) spectra were observed for catalytic intermediates of bovine cytochrome c oxidase at room temperature by using an Artificial Cardiovascular System for Enzymatic Reactions and Raman/ Absorption Simultaneous Measurement Device. The isotope-frequency-shift data using an asymmetrically labeled dioxygen, 16O18O, established that the primary intermediate (compound A) is an end-on-type dioxygen adduct of cytochrome a3. Higher resolution RR experiments revealed that the ~800-cm−1 band assigned previously to the FeIV=O stretching mode of the subsequent intermediate (compound B) was actually composed of two bands that behaved differently upon deuteration. The 804/764-cm−1 pair for the 16O2/ 18O2 derivatives were insensitive to deuteration, and these bands became noticeably stronger at higher temperatures. In contrast, the 785/751-cm−1 pair for the 16O2/ 18O2 derivatives was shifted to 796/766 cm−1 in D2O, and its intensity became virtually zero at 30 °C. The corresponding bands for the 16O18O derivative appeared at nearly the same frequencies as those of the 16O2 and 18O2 derivatives but not at an intermediate frequency. Therefore, neither of these bands could be attributed to the O-O stretching mode. Normal coordinate calculations using a set of suitable force constants of the Urey-Bradley-Shimanouchi force field for an isolated Fe-O-O-H group could explain the 785/751-cm−1 bands in terms of the Fe-O stretching vibration, although this model assumed a relatively strengthened Fe-O bond and weakened O-O bond for the Fe-O-O-H group. The 804/764-cm−1 bands were assigned to the FeIV=O stretching mode of the ferryl intermediate. The 356/342-cm−1 pair for the 16O2/18O2 derivatives was insensitive to deuteration, and its frequencies were the same as those obtained with 16O18O. The intensities of the components of this pair were appreciable even when those of the 785/751-cm−1 pair were zero. Therefore, contrary to the previous assignment, the 356/342-cm−1 pair cannot be ascribed to the hydroperoxy intermediate, but its intensity behavior with temperature was also not coincident with that of the 804/764-cm−1 pair. Although the assignment of the 356/342-cm−1 pair is yet to be determined, the present experiments have conclusively established that the main pathway for dioxygen reduction by cytochrome c oxidase is FeII—O2 → FeIII—?—?—H → FeIV=0 → FeIII—OH and that their key bands for the 16O2/18O2 pair are at 571/544, 785/751, 804/764, and 450/425 cm−1, respectively.