OBJECTIVE To assess the prevalence of Clostridium perfringens enterotoxin in feces of dogs with and without diarrhea, and to compare the use of microbial cultures from fecal specimens and evaluation of stained fecal smears for endospores with the presence of enterotoxin as tools for diagnosing C perfringens-associated diarrhea. METHODS Prospective study. METHODS 144 dogs representing hospitalized dogs with (n = 41) or without (50) diarrhea, and clinically normal dogs treated as outpatients (53). METHODS Fresh fecal specimens from all dogs were examined as Gram-stained fecal smears to determine numbers of Gram-positive spore-forming rods/100x objective field. Enterotoxin was assayed directly by use of a reverse passive latex agglutination assay. Fecal specimens were plated directly to prereduced egg yolk agar plates and incubated overnight at 37 C in an anaerobic chamber. At 24 hours, up to 3 lecithinase-positive colonies were subcultured to Brucella blood agar to evaluate for double zone hemolysis. Colonies with double zone hemolysis were tested for aerotolerance and Gram-stained. RESULTS A significant difference was not detected among groups with respect to the presence of C perfringens as determined by culture, the presence of endospores, and the reaction patterns of fecal enterotoxin assays. An association was not found between number of endospores and the presence of fecal enterotoxin. CONCLUSIONS The presence of C perfringens enterotoxin in feces of dogs, as detected by the latex agglutination assay used in this study, correlates poorly with the number of fecal endospores, regardless of the dog's clinical status.