Development of a strategy for the investigation of pulmonary infectious disease is aimed at identifying the organism responsible for the infection in order to prescribe the appropriate antibiotic therapy. Factors involved in the choice of a method are the underlying condition (healthy, high risk or immunodepressed subject), the type of infection (primary, secondary), the technical abilities of the medical and bacteriological team and finally the value of the different techniques of isolation. The latter must provide a specimen which is not contaminated by the oropharyngeal flora. Their reliability involves definition of a reference method which can be used to test the other techniques in comparison, and requires comparison of the bacteriological results obtained with clinical and radiological data as well as the results of the resultant therapeutic decision. Indirect methods of investigation (blood cultures, serological studies) are relatively unfruitful. Criteria of value of direct methods of investigation are defined and applied to each method (expectoration, transtracheal aspiration, bronchial fibroscopy, transparietal puncture). On the basis of data in the literature and their own results, the authors undertake an analytical then comparative (in a given patient) study of the different methods, and identify their indications.