By means of a mailed questionnaire, a survey was made on how frequently analysands referred to the burglary of Daniel Ellsberg's psychoanalyst's office. That 62.25% of analysts surveyed responded, and that of the respondents 69.33% voluntarily identified themselves, is in itself an unexpected and significant result. It was found that only 13% of 5,074 analysands referred in any way to the burglary. Of these an overwhelming majority were patients of a proportionately small number of analysts. The author discusses his findings in terms of the role played by "perimeter events" in analysis. He raises the question of scotomatization on the part of some analysts, patients or both.