Using organic peroxide initiators in the melt grafting of monomers containing amino groups onto polyethylenes is known to cause serious discoloration of the resulting polymers. To eliminate the discoloration while preserving the controlled degree of grafting, a nonoxidative and hydrogen-abstracting initiator with appropriate thermal stability is needed. In this study, the hydrogen-abstracting capability of three azo initiators with suitable decomposition rates in the melt grafting temperature ranges was evaluated using polyethylene cross-linking and polypropylene degradation experiments (called ''hydrogen-abstraction experiments''). Among the three azo initiators tested, only a phenylazo initiator, 2-phenylazo-2,4-dimethyl-4-methoxyl-valeronitrile (V-19), demonstrated strong hydrogen-abstracting capability from polymer backbones. These three azo initiators were used in the melt grafting of 2-(N'N-dimethylamino)ethyl methacrylate (DMAEMA) and 2-(N-t-butylamino)ethyl methacrylate (tBAEMA) onto a linear low-density polyethylene (LLDPE). In agreement with the results obtained from the hydrogen-abstraction experiments, only V-19 led to a considerable amount of grafting of both DMAEMA and tBAEMA onto LLDPE. As expected, polymers grafted with V- 19 showed significantly reduced discoloration compared with those grafted with peroxide initiators. Further examination of the grafting results indicated that the initiator efficiency, defined as the number of grafted monomer units per radical generated from initiator decomposition, was higher with the phenylazo initiator than with peroxide initiators. The hydrogen-abstracting capability of the phenyl free radical generated from the decomposition of V- 19 was estimated to be higher than peroxide and methyl free radicals produced in the decomposition of peroxide initiators.