Few empirical investigations permit systematic comparison of the impact of widely-varying delivery systems within a single population sample. This study provides such a comparison, describing patterns of ambulatory care among patients using five different systems in Washington, D.C. as a regular source of health care: solo practice, fee-for-service group practice, prepaid group practice, public clinics, and hospital outpatient departments or emergency rooms. Comparisons are adjusted statistically to account for major patient group variations, and the results reveal substantial differences among the five systems. Sources used primarily by the poor--hospital outpatient departments, emergency rooms, and public clinics--contained important structural and financial barriers, and had the lowest rates of patient-initiated use. The prepaid system, in contrast, maximized patient's access to both preventive care and symptomatic care, and did not seem to inhibit physician-controlled follow-up care. The results suggest some perverse effects of fee-for-service payment: patients, especially poor patients, appeared to be deterred from seeking preventive and symptomatic care, while physicians were encouraged to expand follow-up services. Moreover, services in fee-for-service systems were distributed less equitably relative to both income and medical need than in the prepaid system. These findings have direct implications for policy decisions concerning organizational and financial arrangements for the delivery of ambulatory care.