Preliminary data suggest that barium sulfate suspension is a potentially useful negative gastrointestinal contrast agent for MR imaging. To evaluate this hypothesis in a controlled fashion, abdominal and pelvic MR studies of 10 normal volunteers were performed before and after both oral (600-900 ml) and rectal (400 ml) administration of barium. Standard spin-echo coronal T1-, axial T1-, proton density-, and T2-weighted images were obtained at 1.5 T. Images obtained were randomized and interpreted by three observers, who evaluated bowel visualization and delineation of normal anatomy. Bowel segments evaluated were stomach, duodenum, proximal small bowel, proximal colon, distal colon, and rectum. Anatomic structures examined were pancreatic head, pancreatic body, pancreatic tail, retroperitoneum, spleen, liver, pelvic side walls, uterus, vagina, bladder, prostate, and seminal vesicles. Data concerning barium tolerance and safety were recorded. Descriptive, percent change, and kappa statistics were analyzed. Pairwise agreement techniques and repeated measures analysis of variance were performed. This statistical assessment showed a significant improvement in both bowel visualization (59-123% improvement, depending on the segment) and delineation of normal anatomy (23-68% improvement, depending on the structure) after barium administration, particularly on T1-weighted images. In addition, barium was a well-tolerated and safe contrast agent that did not produce artifacts. Our results show that barium sulfate is a useful negative gastrointestinal contrast agent for MR because it improves bowel visualization and delineation of abdominal anatomy, particularly on T1-weighted sequences.