Substitution-inert complexes such as Fe(CN)(6)(4-) are usually considered to react by outer-sphere electron transfer (ET) with most electron acceptors, including aryl diazonium ions (ZC(6)H(4)N(2)(+), where Z denotes a substituent on the benzene ring). However, in contrast to the conclusion drawn in a previous report ( J. Am. Chem. Soc. 1987 , 109 , 1536 - 1540 ), kinetic studies and identification of products from the reactions of 4-nitro- and of 4-methoxybenzenediazonium with an excess of Fe(CN)(6)(4-) show that this is not the case and that the reactions actually go via the formation of an adduct, a diazoisocyanide complex [ZC(6)H(4)N(2)(+) + Fe(CN)(6)(4-) --> ZC(6)H(4)N(2)(NC)Fe(CN)(5)(3-)]. The adduct decomposes heterolytically by expulsion of nitrogen either to form an isocyanide complex [ZC(6)H(4)N(2)(NC)Fe(CN)(5)(3-) --> ZC(6)H(4)(NC)Fe(CN)(5)(3-) + N(2)] or the 4-substituted benzonitrile via a ligand exchange [ZC(6)H(4)N(2)(NC)Fe(CN)(5)(3-) --> ZC(6)H(4)CN + Fe(CN)(5)(3-) + N(2)]. A competing homolytic decomposition resulting in an overall ET reaction occurs only to a minor extent, giving small amounts of Fe(CN)(6)(3-), ZC(6)H(5), and various organic compounds. In oxygenated solutions ZC(6)H(4)N(2)(NC)Fe(CN)(5)(3-) decomposes to Fe(CN)(6)(3-) and ZC(6)H(4)OH. The measurements with Fe(CN)(6)(4-) were supplemented by the study of the analogous reactions of Os(CN)(6)(4-), Mo(CN)(8)(4-), and W(CN)(8)(4-). The observation that isocyanide and even short-lived diazoisocyanide complexes are formed is in accordance with an inner-sphere mechanism. Further support of this conclusion comes from the observation that the slope of the activation-free energy plots for the reactions of NO(2)C(6)H(4)N(2)(+) and MeOC(6)H(4)N(2)(+) with the four metal cyanides is higher than that expected for an outer-sphere ET mechanism. The implication of these results are discussed in the context of the previous report (vide supra) on the extraction of the self-exchange reorganization energies for substituted benzenediazonium salts from their reactions with Fe(CN)(6)(4-) and decamethylferrocene. Our conclusion is that Marcus theory is not applicable in the interpretation of the measured rate constants, thereby also precluding a determination of such energies.