To clarify the possible link between radicals and the cytotoxicity of eugenol-related compounds, 2-allyl-4-X-phenols (2-allyl-4-chlorophenol ( 1), 2-allyl-4-phenylphenol ( 2), 2-allyl-4-methoxyphenol ( 3), 2-allyl-4-acetylphenol ( 4), 2-allyl-4-nitrophenol ( 5), 2-allyl-4- t-butylphenol ( 6), 2-allyl-4-methyphenol ( 7), 2-allyl-4-bromophenol ( 8), 2,4-dimethoxyphenol ( 9)), and dimeric compounds from eugenol (4-allyl-2-methoxyphenol), BHA (2- t-butyl-4-methoxyphenol) or MMP (2-methoxy-4-methylphenol); bis-EUG (3,3′-dimethoxy-5,5′-di-2-propenyl-1, 1′-biphenyl-2,2′-diol) ( 10), bis-MMP (3,3′-dimethoxy-5,5′-dimethyl-1,1′-biphenyl-2,2′-diol) ( 11) bis-BHA (3,3′-di- t-butyl-5,5′-dimethoxy-1,1′-biphenyl-2,2′-diol) ( 12) were synthesized. The radical production, radical-scavenging activity and the cytotoxicity of these synthetic compounds and conventional antioxidants (i.e. butylhydroxytoluine, BHT; butylhydroxyanisole, BHA; α-tocopherol (α-Toc); eugenol, phenol) were studied. Erectron spin resonance (ESR) spectroscopy suggested that compounds of 3, 6, 9, eugenol and BHA, but not compounds of 10, 11, and 12 produced radicals in alkaline solutions (pH>9.5) and compounds, 3, eugenol and 9 most efficiently scavenged reactive oxygen species (ROS, O 2 −). The cytotoxic activity of 6 toward human submandibular gland carcinoma (HSG) cells was the highest and was 1000-fold greater than that of eugenol and 100-fold greater than that of BHA, possibly due to the high hydrophobicity and stable phenoxy radicals of this compound. The kinetic polymerization method in the presence of methyl methacrylate (MMA), an antioxidant, and 2,2′-azobisisobutyronitrile (AIBN) was developed for the measurements of the number of moles of peroxy radicals trapped by moles of the relative phenols (stoichiometric factors, n), the inhibition rate of polymerization ( R inh), and the inhibition rate constants ( k inh, the rate constants for scavenging of radicals by an antioxidant). The n values of conventional phenolic antioxidants decreased in the order: α-Toc>BHT>eugenol>phenol. Those for eugenol and phenol, less hindered phenols, were much less than two, whereas those for α-Toc and BHT, hindered phenols, were approximately two. The R inh of α-Toc significantly increased tcompared with that of BHT, eugenol and phenol. The k inh of the polymer radicals of the MMA reaction with conventional phenolic antioxidants was a low value of 1−2×10 2 M −1 s −1, suggesting that the antioxidants trapped radicals quickly. The comparative cytotoxicity of methoxyphenols against HSG cells was investigated. The cytotoxic activity of dimers of 10 and 12 was markedly lower than that of their corresponding monomers, whereas that of the dimer of MMP, 11 was not reduced even after the dimerization. In particular, visible-light (VL) exposure enhanced the cytotoxicity of 11 similar to the monomers of eugenol, BHA and MMP. Changes in BDE (ph(O–H)) (homolytic bond dissociation energy) for phenols is well known to be associated with the n and k inh values, and consequently the cytotoxic activity. Thus, the BDE was calculated using a PM3 semiempirical method. The n and k inh values for monophenols, but not for dimers were correlated to the BDE, possibly due to the steric hindrance of orthosubstituents of dimers. The quantitative structure–activity relationship (QSAR) of eugenol-related compounds was investigated, indicating that log P (octanol–water partition coefficients), the redox potential measured in culture medium, was effective as a term for QSAR. A parabolic relation between the cytotoxic activity and the log P or the redox potential, but not the BDE was observed with an optimum value. In conclusion, the cytotocity of eugenol-related compounds was significantly associated with the activity of the production of phenoxyl radicals, their stability of the subsequent quinonemethide (QM) and the hydrophobicity.