A prospective study tested the hypothesis of Bell and Loop (1971) that 21 high yield findings are significantly associated with the presence of a skull fracture and can be used to decide which patients need skull radiography. In our study, 17 fractures were detected in 594 patients who had skull radiography because of trauma to the head. Six of these fractures occurred in children who had none of the 21 Bell and Loop high yield findings. These six patients were asymptomatic except for superficial injury at the site of trauma. Had the Bell and Loop criteria been used to decide which patients had skull radiography, 35% (all in children) of the fractures would have gone undetected. Thus, in our study population, the Bell and Loop high yield findings were not satisfactory criteria for the decision about obtaining skull films in children.