In the last twelve years, the overwhelming effectiveness of restraining children in the United States, Canada and Europe has been proven in reducing death and injury in automobile accidents. Despite the proven benefits of restraining children, one type of injury has not been prevented. This paper is an analysis of stretch injuries to the spinal cord in the upper thoracic or cervical spine.This paper discusses, in general, spinal cord injuries from a biomechanical point of view. The relationship between various loading conditions and the resulting types of spinal cord injuries is discussed. This paper also examines seven real world automobile accidents. Information for each case includes: vehicles involved, type of roadway, crash Delta-V, occupant direction of motion, restraint type, injuries to occupants, and anthropometry of child with spinal cord injury. A description and location of each spinal cord injury that occurred at the time of the accident is discussed. A discussion of a possible injury mechanism for spinal cord stretch injuries to children is also given.The results of this research indicate that infants should ride in rearward-facing child restraint systems as long as possible. It was noted that the crash Delta-V that resulted in serious spinal cord injury is proportional to the child's age. All but one spinal cord injury could have been prevented if the appropriate child restraint system had been used, or if the one that was used had been used properly. Finally, the spinal cord injuries in this study could not have been predicted by shear forces or bending moments at the top of the neck.