Abstract Derivatives of heme a have been examined by optical, MCD and EPR spectroscopy [1]. Five- and six-coordinate high-spin ferric species exhibit optical spectra recently classified as ‘Type a’ by Quinn et al. [2] while a low-spin bis-imidazole ferric derivative exhibits a ‘Type b’ spectrum. On reduction the visible spectrum of the low-spin derivative intensifies markedly and exhibits a single maximum at 589 nm; the visible spectrum of the high-spin species changes shape but the intensity is only slightly changed. The ferric high-spin compounds exhibits a transition in the near-infrared which has absorbance and MCD characteristics similar to the 655 nm band [3] of the resting enzyme. Composite spectra obtained by the addition of the individual spectra of the ferric high- and low-spin models and of the ferrous high- and low-spin models reproduce the essential features of the spectra of oxidized and reduced enzyme, respectively. The relative contributions of the high- and low-spin derivatives to the spectral changes at 589 nm produced by reduction are identical with those deduced by Vanneste [4] from an analysis of the visible spectra of the enzyme. Potentiometric measurements on the two heme centers confirms of stoichiometric reductive titrations [5] which showed that these heme centers have essentially identical electron affinities. However, parallel measurements of low-spin a+3 and a+2 by MCD do not provide any evidence for a low-to-high-spin transition of this heme center. Such a transition had been deduced earlier [5] from a comparison of the redox state of a+2 gauged by room-temperature MCD and a+3 as measured by EPR at 12 °K. We have measured Mossbauer spectra of resting enzyme, the resting enzyme-cyanide derivative and the reduced enzyme using 57Fe present in natural abundance. The Mossbauer parameters of these species are very similar to those recently reported for the enzyme isolated from Thermus thermophilus [6]. The isomer shift and quadrupole splitting for a+3 are 0.47 and 1.14 mm−1 sec−1; these values are quite inconsistent with assigning an oxidation state of IV to this center, as has recently been proposed [7, 8]. The values are, however, quite typical for high-spin Fe(III), a conclusion which may also be drawn from a comparison of the resonance Raman oxidation and spin-state markers for a+3 [9] and a+33 [10]. The Mossbauer parameters of the resting enzyme—cyanide derivative appear to be identical with those of the bacterial protein. It would therefore seem that this species is also a S = 1 ferromagnet, as deduced earlier by variable temperature MCD measurements [11].