The b-type haem groups of cytochrome o have generally been thought to be protohaems. However, we have recently shown that they are of a novel kind, for which we propose the name haem O. On the basis of its pyridine haemochrome spectrum, chromatographic behaviour and molecular mass of 839 Da, we suggest that haem O is a haem A-like molecule with the formyl group in position 8 of the latter replaced by methyl, but with retention of the 17-carbon hydroxyethylfarnesyl side chain in position 2, characteristic of haem A. This structure has now been unambiguously verified by resonance Raman, H-1-NMR, and infrared spectroscopy, as well as by mass spectroscopy of molecular fragments, in collaboration with C. K. Chang, G. T. Babcock and their co-workers at Michigan State University. Properties of haem O are compared with haem A and protohaem, and possible contributions of haem O structure to the cytochrome o oxidase function are discussed.