The question of how CPMAS polarization-transfer experiments (CP, cross polarization; MAS, magic-angle spinning) should be conducted in order to distinguish between slow chemical exchange and spin diffusion in the solid state has been studied. Both contributions can be separated by performing different types of polarization-transfer experiments in the laboratory and the rotating frame, since dynamics of spin diffusion but not chemical exchange differs from one experiment to the next. Generally, if both processes are present, polarization transfer is expected to be nonexponential and chemical-exchange as well as spin-diffusion rate constants can be obtained in one series of experiments. If the exchange is symmetric, however, polarization transfer is single exponential and a combination of different pulse experiments is required for obtaining rate constants of both processes. The results for 15N CPMAS NMR polarization-transfer experiments on crystalline meso-tetratolylporphin-15N4 (TTP) are presented. Experiments in the laboratory frame show that spin diffusion between the 15N atoms of TTP is characterized by a temperature-independent rate constant. The nature of this process was established by 1H decoupling during the mixing time, which results in quenching of the polarization transfer. Thus, the role of the 1H spin reservoir for laboratory-frame spin diffusion among chemically inequivalent 15N spins in 15N-enriched material is confirmed. At higher temperatures, polarization transfer in the laboratory and the rotating frame is observed due to a symmetric exchange of the nitrogen atoms arising from a double proton transfer which has been previously established. The double proton transfer rates observed with the different polarization-transfer methods agree well with the values predicted from high-temperature lineshape analysis and are found to be very close to the solution data. © 1988.