A systematic sample of 300 cases was drawn from each county. Data, collected between November 1984 and April 1985, were obtained from the jail book, inmate files, and county court records. Ninety-two percent of those booked were released prior to trial, with those detained tending to have higher bail amounts and prior conviction rates. Over 77 percent of those released were under bond. The average length of pretrial detention for those released within the first week was 14 hours. Drunk driving defendants comprised 58 percent of the traffic cases and 22 percent of the total cases. Counties with overcrowded facilities required a higher percentage of cases to go to court to secure release and held defendants significantly longer than counties with noncrowded facilities. Crowded facilities released 12 percent of their cases on personal recognizance compared with 35 percent for noncrowded jails. There was little variation in urban and rural pretrial practices. Although counties with pretrial release programs generally required defendants to go to court to secure release, the detention period was not significantly longer for these counties than for those without pretrial release programs. An analysis of pretrial recidivism for 967 cases found that 16 defendants were rearrested for felonies prior to trial. Implications are drawn for policies to reduce jail overcrowding and improve jail management information. 39 tables.