The diagnostic sensitivity of the polymerase chain reaction (PCR) was compared with that of culture on conjunctival swabs over the course of infection in 4 doxycycline-treated and 4 untreated cats that were experimentally infected with feline Chlamydia psittaci. Treated cats were given 25 mg (5 mg/kg) of doxycycline orally twice daily for 3 weeks from day 6 after challenge. Clinical signs improved within 3 days of institution of treatment. Culture remained positive for 1 day and PCR remained positive for up to 5 days after treatment was commenced. No recurrence of clinical signs occurred and the organism could not be detected by either PCR or culture for 2 weeks after cessation of therapy. In the 4 untreated cats, conjunctival swabs were taken daily to day 14 and every 2nd weekday to day 64 after challenge. PCR was significantly more sensitive than culture in untreated cats overall (PCR 85.7%, culture 72.9%, P approximately 0) and for cats with clinical signs (PCR 89.2%, culture 79.2%, P = .008). PCR and culture had equivalent sensitivity (100%) for cats showing clinical signs in the 1st month of infection, whereas PCR was considerably more sensitive than culture for cats showing clinical signs in the 2nd month (PCR 72.9%, culture 47.9%, P = .028). Organisms were not detected by PCR in blood or any tissue collected from treated or untreated cats at postmortem. Thus, effective treatment of chlamydiosis in cats is possible with much shorter treatment regimens than currently recommended, and PCR is the more sensitive diagnostic method in chronically infected cats.