in Part 1 of this series. we considered the theoretical basis behind the pulsed-field gradient nuclear magnetic resonance method for measuring diffusion. In this article the experimental and practical aspects of conducting such experiments are considered, including technical problems involved in gradient production such as eddy currents, gradient calibration. internal gradients in heterogeneous samples, and temperature control. Furthermore, the means for recognizing and preventing or at least minimizing these problems are discussed. A number of representative pulse sequences are also reviewed. (C) 1998 John Wiley & Sons. Inc.