The instrumental conditions required for the determination of the extent of reaction in dimethacrylate-based dental composites using high-resolution solid-state C-13 nuclear magnetic resonance (n.m.r.) techniques have been investigated. In the absence of paramagnetic fillers, clearly resolved carbonyl resonances for reacted and unreacted methacrylate groups were observed. The two peaks showed considerably different relaxation times using cross-polarization excitation, and accurate relative intensities could only be obtained using direct excitation of the C-13 nuclei with a pulse interval of at least 30 s. Chemical shielding anisotropy spinning sidebands were observed for the reacted peak, but not for the unreacted. A spin-rate-dependent centreband intensity correction factor to account for this difference was evaluated using model triethyleneglycol dimethacrylate materials. The proportion of unreacted methacrylate groups determined using n.m.r. was essentially unaffected by spin rate, dipolar decoupling field strength and temperature, but was only about one-half that determined using infra-red spectroscopy. The origin of the discrepancy is as yet unclear.