AT THE TIME OF WRITING THERE WERE NO RELIABLE AND VALID DATA ON THE CHARACTERISTICS AND CONSEQUENTLY, THE INCIDENCE, OF CHILD ABUSE. THE SPECTRUM OF DEFINITIONS IS VERY BROAD AND SIGNIFICANT DIFFERENCES EXIST IN THE AMERICAN POPULATION CONCERNING APPROPRIATE PHYSICAL PUNISHMENT. MOST CURRENT DEFINITIONS INCLUDE 'ACCIDENTAL' AND 'NONACCIDENTAL' INJURY. THE PURPOSE OF THIS PAPER INVOLVES THE PRESENTATION OF THE FOLLOWING ELEMENTS: (1) THE MAGNITUDE OF DISAGREEMENT ON THE CONCEPT OF CHILD ABUSE; (2) THE DEMONSTRATION OF UNTESTED HYPOTHESES REGARDED AS SPECULATIONS BECAUSE OF THE LACK OF SYSTEMATIC DATA; AND (3) A DEFINITION FOR SYSTEMATIC RESEARCH. THE FOLLOWING CHARACTERISTICS OF ABUSED CHILDREN ARE CONSIDERED: WITHDRAWAL, PSYCHOMOTOR RETARDATION IN INFANTS, AND SUPERFICIAL RELATIONSHIPS IN OLDER CHILDREN. ABUSING PARENTS FORM A HETEROGENOUS GROUP INCLUDING ALL TYPES OF PERSONALITY DISORDERS, NEUROSES, AND PSYCHOSES. THUS, CLASSIFICATION CAN BE MADE BY AGE, RELATIONSHIP OF THE ADULT TO THE CHILD, TYPE OF PHYSICAL INJURY, AND THE PSYCHOPATHOLOGY OF THE ADULT OR CHILD, WITHOUT RESORTING TO DETERMINATION OF THE PARENTS' INTENTION TO INJURE THE CHILD. THE ISSUES CONCERNED WITH ACCIDENTAL VERSUS INTENTIONAL CANNOT BE SEPARATED NOR CAN THEY BE RELIABLY MEASURED: THEREFORE, IT IS SUGGESTED THAT THEY BE REMOVED FROM ANY APPROACHES TO THE STUDY OF CHILD ABUSE AND THAT STUDIES INCLUDE ALL CHILDREN RECEIVING MEDICAL ATTENTION FOR PHYSICAL INJURIES ASSOCIATED WITH THEIR PARENTS' OR CARETAKERS' BEHAVIOR.