2-Methyl-1-(2-phenyl-2-propenyloxy)-pyridinium hexafluoroantimonate (III) and hexafluorophosphate (IV) were synthesized. Their capability to act as initiator or coinitiator for the cationic polymerization of oxiranes and vinyl ethers was examined. Monomers under investigation were cyclohexene oxide (CHO), vinyl cyclohexene dioxide (VCHD), and n-butyl vinyl ether (nBVE). These monomers turned out to be polymerizable in the presence of III or IV provided free radicals are generated thermally at 80 degrees C or photochemically at lambda > 360 nm with the aid of appropriate radical sources, e.g. 2,2'-azobisisobutyronitrile, phenylazotriphenylmethane (PAT), and dibenzoyl peroxide (thermal radical sources) and benzoin, PAT and trimethylbenzoyl diphenylphosphine oxide (TMDPO) (photochemical radical sources). The polymerization of CHO is also induced in the absence of a free radical source at lambda = 300 nm where III absorbs light and also thermally at 80 degrees C, but here at a low rate. The latter process is very likely to be due to self-initiated free radical generation similar to that known for neat styrene. Regarding sources generating electrophilic free radicals the mode of action of III concerning the promotion of the cationic polymerization seems to be based on the addition-fragmentation mechanism. Accordingly, a free radical adds to the carbon-carbon double bond and fragmentation of the adduct radical results in the formation of a reactive onium radical cation. Evidence for the oxidation of nucleophilic radicals such as the triphenylmethyl radical by the onium ion was not obtained in this work. In conclusion, a wide range of radical sources can be employed in conjunction with onium salt III or IV to tune conditions (temperature and wavelength, respectively) for the cationic polymerization of various monomers. (C) 1997 Elsevier Science Ltd.