A computer-based search was conducted of medical and necropsy records of horses admitted to the teaching hospital from Jan 1, 1979, to Dec 31, 1987, to obtain the records of all horses admitted to the hospital for colic and subsequently found to have gastric rupture. Fifty cases of gastric rupture were found. The records were reviewed to obtain data regarding peritoneal fluid analysis. Cell counts of these samples were often erroneous because debris and clumps of bacteria were counted when most WBC were lysed. A cross-sectional study of gastric rupture cases versus all other colic cases regarding season of admission revealed that there was no association between season and the occurrence of gastric rupture. There was also no increased risk associated with age, gender, breed, and the occurrence of gastric rupture. One hundred colic cases, matched with the gastric rupture cases by year of admission, were randomly selected via a table of random numbers. A questionnaire regarding age, breed, gender, use of the horse, housing, diet, water source, deworming schedule, and medical history was completed from the medical records and phone conversations with the horse owners. The results indicated that horses on a diet of grass hay or grass/alfalfa hay only or those that drank water from a bucket, stream, or pond were at increased risk for having gastric rupture. In contrast, horses fed grain had a reduced risk.