This study was undertaken to evaluate the relative contribution of ischemia, bacteria, and luminal substrate, the pathogenetic components of necrotizing enterocolitis (NEC), to the development of intestinal necrosis. Sprague-Dawley rats, either germ-free (No. = 25) or conventionally colonized (No. = 20) underwent laparotomy. Isolated ileal segments were created, two per rat. Ischemia was produced in one segment by application of a microaneurysm clip; the other segment served as a control. Segments were injected with 1 mL of either normal saline, dilute Similac formula, or standard formula. Groups were as follows: Group I (germ-free), received saline; Group II (germ-free), dilute formula; Group III (germ-free), standard formula; Group IV (conventional), saline; Group V (conventional), dilute formula; Group VI (conventional), standard formula. At 48 hours, the rats were evaluated for survival, gross bowel integrity, histologic severity of necrosis (graded 0 to 4+), and bacteriology. Gross analysis of bowel integrity showed no lesions in the ischemic segments of the germ-free rats (Groups I, II, and III) and necrosis in 75% of conventionally colonized animals (Groups IV, V, and VI; P less than 0.001). Microscopic necrosis was more common (P less than 0.001) in ischemic segments of conventional rats than in ischemic segments of germ-free rats. There was no difference in necrosis attributable to ischemic time or to the presence of either standard or dilute formula. Of the three pathogenetic factors evaluated, the presence of bacteria was most crucial to the development of bowel necrosis in this model. Improved treatment and prevention of NEC may depend upon suppression and/or modification of the gut flora.