The relationship between pet bird keeping and lung cancer according to exposure to tobacco smoking was investigated in a case-control study in hospitals of New York City and Washington, DC, USA. Newly diagnosed lung cancer cases (n = 887) aged 40-79 years were compared with 1350 controls with diseases not related to smoking, of the same age, gender and date of admission as the cases. The prevalence of pet bird keeping was 12.5% in men and 19.1% in women. There was no association between ever keeping a pet bird and lung cancer in never smokers (men adjusted odds ratio (OR) = 0.70, 95% confidence interval (CI) 0.15-3.17; women, 1.32, 95% CI 0.65-2.70), or in smokers and non-smokers combined, after adjustment for ever smoking (men: 1.28, 95% CI 0.88-1.86; women: 1.17, 95% CI 0.83-1.64; all: 1.21, 95% CI 0.95-1.56). Risk did not increase in relation to duration of pet bird keeping. Cases and controls kept similar types of birds. There was a tenfold increase of lung cancer risk associated with smoking among non-bird keepers (adjusted OR = 9.15). There was no indication of a synergism, either additive or multiplicative, between smoking and pet bird keeping with respect to lung cancer risk. Either alone or in conjunction with smoking, keeping parakeets, canaries, finches or parrots is not a risk factor for lung cancer among hospital patients in New York and in Washington, DC.