The thermolysis of the Co-C bond in adocobinamide (AdoCbi(+)BF(4)(-)) in anaerobic ethylene glycol has been studied as a function of a series of para-substituted pyridine axial bases using the TEMPO radical-trapping method. In contrast to the slower rates of Co-C cleavage previously found for benzylcobinamide, neopentylcobinamide, and the (alpha-phenylethyl)cobaloxime coenzyme B-12 models, for AdoCbi(+) the rate of total Co-C cleavage becomes faster as the para-substituted pyridines become more electron-donating, Specifically, the 110 degrees C k(obsd) for AdoCbi(+)BF(4)(-) total Co-C cleavage increased 23-fold on going from 1 M pyridine (py) to 1 M p-(dimethylamino)-pyridine (Me(2)N-py). However, HPLC product studies reveal that the percentage of abiological Co-C heterolysis increases (to a limiting value); that is, Co-C heterolysis is a major reason for the observed rate increase seen for Me(2)N-py. Deconvolution of the k(obsd) rate constant into its heterolysis and homolysis components yields values of the 110 degrees C k(heterolysis) and k(homolysis) for AdoCb(+). base for Me(2)N-py and pyridine. These data in turn reveal that the AdoCbi(+)BF(4)(-) base-on homolysis rate constant does not increase within experimental error as one goes from py to the more basic Me(2)N-py (k(on,h)=8(3)x10(-4) and 7(1)x10(-4) s(-1), respectively), but that the base-on heterolysis rate constant changes by 17-fold (k(on,het)=0.4(0.1)x10(-4) and 7(1)x10(-4) s(-1) for py and Me(2)N-py, respectively). The plausible biological significance of these results is then discussed, notably the heretofore unsubstantiated idea, first suggested by Mealli, Sabat, and Marzilli, that one major evolutionary pressure for selecting and appending the 5,6-dimethylbenzimidazole axial base in coenzyme B-12 is because this base limits Co-C bond heterolysis, thereby promoting the biologically relevant Co-C cleavage reaction, Co-C homolysis.