The 6-month study of "in progress" burglary, funded by the Home Office's Policing and Crime Reducing Unit, used questionnaire surveys of patrol officers, burglary site surveys, and police records. A sample of 441 cases was drawn from the 9 percent of burglaries reported while "in progress" between July and December 1996. This consisted of 116 cases in which one or more suspects were caught in the act and 291 cases in which nobody was caught. Data were collected by using a questionnaire survey of police patrol officers, surveys of burglary sites, and police records, so that the various aspects of the crimes and their investigation could be related on the basis of individual incidents. An officer from each patrol unit that attended an incident was asked to complete a questionnaire. Findings show that single-officer units were no worse than two-officer patrols at catching burglars in the act. Although two-officer units reached the burglary scene more quickly, this was primarily due to their circumstances prior to the burglary, their location, what they were doing at the time, and whether it was day or night. Two-officer units did have a small influence on response speeds, but this operated more strongly when patrols were farther away from the scene, at distances where fewer of them reached the scene first, or made successful arrests. Boosting patrol provision would increase the number of burglars caught at the scene, and this could be achieved by switching resources from double to single-crewed units. 6 tables and 14 references