Single-crystal electron paramagnetic measurements have been carried out on an orthorhombic chlorotetrabromo salt of the mixed-valence μ-pyrazinc-decamminediruthenium(5 +) ion (Creutz-Taube ion) Below ≈ 200 K the spectrum consists of two lines arising from two magnetically inequivalent but crystallographically equivalent centres. The resonances can be described by a g tensor with principal axes aligned with ▪ and x 1. y 1: x 2. y 2 in the ab plane. At 113 K the comp of the g tensor are g x = 2.779(5). g y = 2.489(5). g z = 1.334(10). The point-group symmetry about the centre of unit is C 2h (m/2) with the c axis being the two-fold axis. The g z component is aligned perpendicular to the plane of the pyrazine ring and perpendicular to the Ru-Ru axis ( X). The remaining components should be aligned closely wilh X. Y. The angle between the X axes for the two sites as determined by X-ray measurements, is 85.8° compared with 82.5 ± 1° between the x 1. x 2 g tensor axes. Magnetic exchange coupling between the dimer ions is very small. The tetragonal approximation to the observed g tensor is g 1 = 2.632 ± 0.005. g 11 = 1.334 ± 0.010. The theory of the g tensor for this complex is discussed, it is shown that these values are within the range predicted for an ion with a symmetrical delocalized ground state, the orientation of the g 11 component normal to the plane of the pyrazine ring is also predicted for a delocalized ground state. The EPR results arc thus consistent with a stable single-minimum ground state for the Creutz-Taube ion Powder EPR experiments on a tosylate salt of this ion have previously been reported by Bunker et al. A g 11 signal was not observed and it was assumed to be very small (⩽ 0.6) and oriented along the Ru-Ru axis, however the absence of a signal at g ≈ 1.3 (in fact oriented perpendicular to the Ru-Ru axis) was presumably due to poor statistics for g 11 in EPR measurements on powder samples. The proposal of these authors that EPR measurements provide evidence for a trapped-valence ground state for the Creutz-Taube ion thus cannot be accepted.