Radiography is a commonly used diagnostic tool for diagnosis of tuberculosis. It is still widely believed that tuberculosis of the lung can be diagnosed by chest radiography alone. However, practical experience and number of studies have proved beyond doubt that no radiographic picture or pattern is absolutely diagnostic or typical of tuberculosis. There are many conditions of the lung which show similar radiographic appearance and can easily be mistaken for tuberculosis. On the otherhand, in a substantial percentage of bacteriologically confirmed cases, radiological picture either did not show a shadow or was judged as non- tuberculosis. Lesions being hidden behind other shadows e.g., of the ribs etc. may be one of the reasons. A comparison of large X-ray with MMR indicates that large X-rays are easily readable and have low radiation dose but are more expensive and occupy more space. The paper critically discusses the reliability of chest radiography also. The concluding remarks are as follows: 1) 20% of lung fields are hidden behind bones, heart, soft tissues, etc. 2) Activity of an X-ray lesions cannot be determined on the basis of single X-ray picture. Even sequential X-rays or repeated serial X-rays of the patient over a period of time have their limitations and is not fool proof. 3) Inter-individual and intra- individual variations in interpretations seriously affect the diagnosis. Both over reading and under-reading are substantial. The latter can be improved to some extent by constant practice and experience in X-ray reading but cannot be eliminated entirely.