Ungated and gated magnetic resonance images of the spinal cord acquired with the use of long repetition times (TRs) and long echo-delay times (TEs) were compared in 21 studies performed on a 1.5-T system. Both normal and abnormal spinal cord conditions were compared. All images were acquired in an identical fashion except that ungated studies had TRs of 2,000 or 2,500 msec, whereas in gated studies, TR was determined by the patient's heart rate. The effective TR of images gated to the cerebrospinal fluid (CSF) fell primarily in the range of 1,500-1,800 msec. Gating was accomplished using a peripheral pulse. Three image parameters were assessed: signal-to-noise ratio, object contrast, and resolving power. For each parameter, in both normal and abnormal spinal cords, the CSF-gated studies proved superior by eliminating spatially mismapped signal intensity from pulsatile CSF.