To the Editor .—Johnson et al 1 determined the percentages of race and sex distribution found in a series of abdominal aortic aneurysms (AAA) detected among a group of patients who had autopsies or computed tomographic (CT) scans at a North Carolina hospital. They reported the "incidence" by race of AAA by projecting this distribution to the state population. We object to the use of the term incidence , which is generally defined as the number of new cases that occur in a population over a specified time period. The authors were attempting to determine prevalence . We are concerned about selection bias and the representativeness of the study group with respect to the general population of people with AAA. People who undergo autopsy or CT scan are not representative of all blacks and whites in North Carolina unless systematic postmortem examinations or CT scans are performed in this population. 2 Many