Endotoxin, a kind of lipopolysaccharide (LPS), is the glycolipid composition of cell walls of all Gram negative bacteria. The main biological reaction triggered by LPS is inflammation reaction. The surface epithelium of intestinal mucosa contains a large number of goblet cells which function as secreting mucus to soak chime and form a protection for digestive tract. The purpose of this experiment was to study the variations of the mucus secretion of intestinal gland goblet cells and whether dimethylthiourea (DMTU), a hydroxyl radical scavenger, could exert an inhibitory effect at different time points after LPS application. Sprague-Dawley rats received an intravenous injection of LPS (15 mg/kg, over 5 seconds) through the femoral vein. India ink was used to label the leaky microvessels where the extravasated colloidal particles of dye were trapped between the endothelium and basement membrane. 5 and 30 minutes after LPS injection, strips of intestinal tissues were dehydrated and embedded with glycol methacrylate. Tissue sections 2 micrometers in thickness were histochemically stained with Alcian blue-PAS reagent. The number of goblet cells in intestinal glands of duodenum and ileum were counted. Mucus area and epithelial area of sampled glands were recorded with SimplePCI and statistically analyzed. 5 or 30 minutes after LPS, the percent (%) of the number of discharging goblet cells in intestinal glands significantly increased to 7-8 times as the saline control in duodenum. In the ileum, the number of discharging goblet cells in intestinal glands significantly increased to 5-18 times as the saline control. 30 min after LPS, the percent (%) of goblet cell mucus in epithelial area of intestinal glands decreased to half the value of control. DMTU prior to LPS, the mucus-discharging ability of goblet cells was inhibited. In duodenum, DMTU pretreatment 30 min earlier than LPS, the percent (%) of goblet cell mucus in epithelial area of intestinal glands increased to 1.4 times and the percent (%) of the number of discharging goblet cells in intestinal glands significantly was reduced by 90%. In ileum, DMTU pretreatment 5 min earlier than LPS, the percent (%)of the number of discharging goblet cells in intestinal glands decreased to one half as the control. From the experiment results, we could come to a conclusion that intravenous injection of a high dose of LPS induced an compound exocytotic release of mucus granules resulting in cavitation of goblet cell supranuclear cytoplasm and the released mucus temporarily accumulated in the lumen of glands. It is suggested that hydroxyl radicals were involved in LPS-induced mucus release from goblet cells.