Using Molotch and Lester's accident research methodology, this study coded each entry on the FBI for the period 1950-79 in the New York Times Index as an accident or scandal. The results confirmed the impression that the FBI was untouchable in the Hoover years, with 2.3 percent of all scandals and accidents occurring in the 1950's, 23.3 percent in the 1960's, and 74.4 percent in the 1970's. Prior to 1970, public disclosures about the FBI were disproportionately initiated by outsiders (accidents), whereas 84.6 percent of disclosures in 1975-79 were initiated by insiders (scandals). This shift suggests a weakening in the FBI's power to control information about itself and erosion of its semi-autonomous position within the Department of Justice (DOJ). The data also suggest that major disclosures, although initiated by outsiders, were used by insiders such as the DOJ and the White House to restructure relationships between the FBI and the executive branch. This view is confirmed by an analysis of six FBI organizational practices disclosed in the 1970's: its counterintelligence activities targeting political groups, Hoover's secret files on public officials, surreptitious entries, White House misuse of the FBI, internal financial corruption, and informants as agent provocateurs. Between 1973 and 1977, the DOJ was able to undermine the old guard Hoover-era veterans in the FBI by acknowledging abuses of that era. The FBI and the DOJ had achieved a more stable relationship by 1977, and both reverted to an attitude of disclosing as little information as possible. The reforms of the 1970's succeeded in making the FBI more accountable to the Executive and a more integral part of the DOJ and the intelligence community. Because the reforms did not address the limits of executive authority, the FBI's new investigative priorities are fragile and easily subject to change. Tables, 16 footnotes, and 32 references are included.