Several studies have suggested that the PCR could be used in epidemiological mass-screening surveys to detect Leishmania (Viannia) spp. infection in human and animal hosts. Dogs from an area of Leishmania braziliensis and Leishmania peruviana endemicity were screened for American cutaneous leishmaniasis (ACL) infection by established PCR-based and enzyme-linked immunosorbent antibody test (ELISA) protocols. PCR detected Leishmania (Viannia) infection in a total of 90 of 1,066 (8.4%) dogs: 32 of 368 (8.7%), 65 of 769 (8.5%), and 7 of 42 (16.7%) dogs were PCR positive by testing of whole blood, buffy coat, and bone marrow aspirates, respectively. ELISA detected infection in 221 of 1,059 (20.9%) tested dogs. The high prevalence of Leishmania (Viannia) detected by PCR and ELISA in both asymptomatic (7.5 and 19.2%, respectively) and symptomatic (32 and 62.5%, respectively) dogs is further circumstantial evidence for their suspected role as reservoir hosts of ACL. However, the low sensitivity of PCR (31%) compared to ELISA (81%) indicates that PCR cannot be used for mass screening of samples in ACL epidemiological studies. Unless more-sensitive PCR protocols were to be developed, its use should be restricted to the diagnosis of active (canine and human) cases and to the parasitological monitoring of patients after chemotherapy.