The previous chest -ray films, that already had abnormal shadows, of 45 lungs cancer cases were reviewed. Most of the overlooked shadows were tiy of faint and were overshadowed by ribs, vessels or central shadow, Careful observation of the whole film, recognition of defferent densities and knowledge of normal chest structure ere useful in reducing the frequency of errors. Some overlooked cases were thought to be due to tiredness or carelessness of the observers. Others were deu to lack of differentiation between normal and abnormal shadows. Double check observation and comparative observation were useful in reducing the frewuency of errors. We expect that computed radiography will be of use reducing perceptual errors in the future.