An apparatus is described that can pneumatically move an aqueous liquid NMR sample sealed in a 5 or 8 mm standard tube, from the center of a standard Varian 500 MHz spectrometer and probe out to any position in its fringe field, and back, under computer control. Applications to low-field relaxation and other measurements are discussed briefly, particularly using 31 P, and 15 N and 13 C with proton detection, utilizing the full power of the commercial system for preparation before, and detection after, a relaxation or mixing interval at low field. Transverse relaxation, polarization transfer, and other transverse manipulations are virtually impossible with this apparatus while the sample is at the low (fringe) field, because of the tremendous inhomogeneity of that field. The spectrometer and probe are unmodified commercial products and the apparatus is expected to be usable in any other system or field without much modification. The apparatus is wheeled to the instrument and installed, and later removed without trace, in less than 1 h each way, including a probe change. A pneumatic glass shuttle tube, 22 mm inside diameter, held in an aluminum support tube, is temporarily inserted in place of the upper tube that is supplied by the manufacturer to be inserted into the top of the magnet. A standard thin-walled NMR tube is connected to a plastic piston shuttle that is moved up or down inside the shuttle tube, by low vacuum or pressure applied from the top. The low value of magnetic field, where the relaxation process measured occurs, is determined by a movable mechanical stop whose position is changed manually between runs. The downward path is initiated by a short higher pressure pulse, followed by a long lower pressure interval. The round trip time limit is about 0.2 s, suitable for relaxation rate measurements in the range up to about 10 s −1 .Variable-temperature operation uses the regulator of the host instrument. After an overview, a detailed description of all aspects of construction and use is provided. The shuttling apparatus itself is relatively straightforward, but development of a simple sample confinement method was difficult, as expected. Problems of bubble formation and of denaturation of a protein are partly solved, and sample loading and sealing is easy. Pulse programs for this apparatus are easily generated from otherwise standard pulse sequences by addition of a few simple instructions. Operation does not require any special knowledge, nor much extra training. A major result of our program is to show that the sensitivity using such a shuttle can be that of the host instrument, not affected by such possible problems as vibration, and only reduced by the predictable losses due to relaxation during the 200-300 ms round trip shuttle time, and the inability to utilize solvent flipback methods.