The findings indicate that the published regression equations should be used only with adult long bones in identifying and distinguishing human bone fragments from those of other animals. The analysis conducted on human samples yielded an identification success rate of approximately 70 percent on adult long bones; however, the percentage of erroneous identifications was particularly high for newborns and older subadults, as well as for human flat bones in general. Application of the regression formula in animal species other than humans did not yield differences in error rate between long bones, flat bones, and juvenile bones; species determination was correct in most cases, with an average percentage of wrong answers being 0.9 percent. Only in the human samples did age and type of bone significantly influence the results. The test included 25 flat bones and 18 long bones of humans and varying numbers of flat and long bones from adults and subadults of the other animals. The long bone samples were obtained from femur and humerus, and the flat bone samples were obtained from the cranial vault bones and the scapula. The methods used for attempts at species identification are described in detail. 1 table, 1 figure, and 13 references