The action of dimethyl sulfoxide (DMSO) was investigated employing an experimental model of intestinal perfusion in vivo of isolated loops of dog ileum before, during, and following 1-h ischemia. DMSO was administered either into the intestinal lumen or by a continuous injection via a branch of the artery supplying the experimental loop. In the intact intestine, intraluminal DMSO significantly decreased the net movement of water, electrolytes, and glucose without affecting either the active transport of phenylalanine and beta-methylglucoside or morphology. This inhibition was irreversible since, following DMSO removal, there was only partial recovery of water absorption, while the other parameters remained significantly low. Inhibition by DMSO was delayed when the substance was injected intraarterially (i.a.). Regardless of its route of administration, DMSO did not reduce the extent of ischemic injury in comparison with non-treated "Controls": during ischemia, all functional parameters were practically non-existent. Following the re-establishment of circulation, a net loss of water and electrolytes ensued, and active transport did not improve. In both instances, the structural alterations were those associated with ischemia of the experimental model employed: short, broad, club-shaped villi which had completely lost their epithelium, but fairly intact crypts. It can be concluded from the data presented that in the dog DMSO inhibits absorption in the intact ileum and exerts no protection against ischemic lesion.