An epidemiological study of the prevalence of mycoplasmal conjunctivitis in the house finch (Carpodacus mexicanus) was conducted in Auburn (Alabama, USA) between March 1998 and February 1999. Clinical disease was observed in 4% of the 1,214 finches trapped and examined. This rate is comparable to the average annual prevalence observed in this population since 1996, although the prevalence of clinical disease observed in the peak months of September through November was lower than in previous years. Clinically ill birds were observed in all months of the study. To estimate the prevalence of recovering and asymptomatic, infected birds, we tested a subset of 334 house finches serologically for exposure to Mycoplasma gallisepticum (MG) using the serum plate agglutination (SPA) assay. The prevalence of clinical disease in this subsample was slightly higher (7%) than in the entire sample, reflecting the fact that the serological survey was initiated in the late summer when the prevalence of MG infection peaks in our study population and a sampling bias for symptomatic birds. The serological survey indicated that 13% of this subpopulation had been exposed to MG. We also tested 46 of 334 finches by polymerase chain reaction (PCR) to detect MG in seropositive, asymptomatic birds. Use of the PCR in conjunction with the SPA detected six asymptomatic, infected birds that may represent potential carriers or birds in the early stages of infection. The decreasing prevalence of clinical disease observed during the peak months suggests a changing host-parasite relationship. Continued surveillance of this population, employing both clinical observation and serological analysis will be useful in characterizing these changes over time.