Organometallics 2001, 20, 700], have unusual E•••E distances, leading to ambiguities in how to best describe their electronic structure. Three limiting possibilities are considered: case A, in which the compounds contain singly bonded E 2 2− units; case B, in which a threeelectron E∴E half-bond exists in a formal E 2 3− unit; case C, in which two E 2− ions exist with no formal E−E bond. One-electron reduction of 1 and 2 yields the new compounds [Cp* 2 Co][Cp′ 2 Ni 2 E 2 ] (1red: E = S, 2red: E = Se; Cp* = 1,2,3,4,5-pentamethylcyclopentadieyl). Evidence from X-ray crystallography, X-ray absorption spectroscopy, and X-ray photoelectron spectroscopy suggest that reduction of 1 and 2 is Ni-centered. Density functional theory (DFT) and ab initio multireference methods (CASSCF) have been used to investigate the electronic structures of 1−3 and indicate covalent bonding of an E 2 3− ligand with a mixed-valent Ni 2 (II,III) species. Thus, reduction of 1 and 2 yields Ni 2 (II,II) species 1red and 2red that bear unchanged E 2 3− ligands. We provide strong computational and experimental evidence, including results from a large survey of data from the Cambridge Structural Database, indicating that M 2 E 2 compounds occur in quantized E 2 oxidation states of (2 × E 2− ), E 2 3− , and E 2 2− , rather than displaying a continuum of variable E−E bonding interactions.