The effect of nitrogen-15 proton-driven spin diffusion on quantitative (15)N T(1) measurements in solid proteins is investigated, and the impact on the measurement of dynamic parameters is assessed. A simple model of exchange between neighboring nitrogens is used to reproduce the evolution of (15)N spin systems whose longitudinal relaxation rates and exchange rates are compatible with experimental measurements. We show that the induced error in the measured T(1) and its effect on the determination of dynamics parameters is likely to be less than the current experimental error. The use of deuterated protein samples is shown to have a small but sometimes visible effect, and may also considerably slow down or even suppress the exchange of magnetization due to spin diffusion.