Understanding of the human musculoskeletal system and common clinical disorders of bones, joints and soft tissues has been enhanced by the use of experimental animal models. Articles reporting on the results of these biomedical experiments frequently include conclusions that are based on the assumption that the biology of the animal model is similar to that of a human being for the disease process under investigation. The purpose of this investigation was to study the criteria and the considerations for selection of an animal model in musculoskeletal research. Selected journals from the musculoskeletal literature published between January 1991 and November 1995 were scrutinized for the use of animal models, and several criteria used in the selection of the various animal models were investigated. The selection criteria analyzed in this study included the biologic characteristics of the model, budget issues, the reproducibility of a musculoskeletal disease, and animal handling factors. A computer-assisted search of the musculoskeletal literature published from 1965 to 1995 was also performed to screen for reports comparing mammals used as animal models in terms of these selection criteria. Our findings imply that the selection of animal models in research of the musculoskeletal system is based partly on non-standardized criteria that are not necessarily based on the biology of the disease process being studied. In addition, there are limited comparative data on the selection and use of different animals for musculoskeletal research. We believe the selection of models should be more standardized based on both biological and non-biological criteria. Researchers would then be able to put in a more meaningful perspective the results of research using animal models and their clinical implications.