Respiratory motion is an important problem in magnetic resonance imaging (MRI) of the thorax and upper abdomen. This study assessed several approaches for practical respiratory gating. Methods of acquiring respiratory signals, gated sequencing methods, duration of examination, strategies for reducing examination time, diagnostic quality of gated images, and the influence of respiratory gating on relaxation time measurements were evaluated. Of three different devices for acquiring the respiratory signal, a belt containing a displacement transducer placed around the upper abdomen was found to be most effective and practical. Two pulse-gating modes were implemented, as well as a method for combining cardiac and respiratory gating. Gating methods were tested using phantoms and human volunteers. A spin-conditioned mode of respiratory gating was found to be superior to a more simply implemented triggered mode in which spin-echo (SE) sequencing was interrupted. The time penalty for respiratory gating is technique-dependent. Gated studies with uncontrolled tidal breathing took two to four times longer than nongated studies. When the time between respirations was voluntarily prolonged, gated studies could be only 30%-50% longer than nongated. The standard deviation of relaxation-time measurements for organs that are displaced during respirations was substantially reduced by respiratory gating. Gating acquisition without spin-conditioning gating. Respiratory gating is a practical and useful technique for improving the contrast and spatial resolution of SE images of the upper abdomen and chest. SE images produced with short repetition times were particularly improved by respiratory gating.