The majority of lower respiratory tract infections (LRTI) are treated "blindly" because the establishment of an aetiological diagnosis is not possible in most cases. The rational choice of therapy mainly rests upon the knowledge of the microbiological epidemiology of LRTI, and on the possible host-parasite relationship. In community-acquired pneumonia, there is general concensus that penicillin maintains its position as the first drug of choice, and that therapy can be changed to erythromycin or tetracycline in cases of therapeutic failure. Treatment of nosocomial pneumonia, and LRTI in immunocompromised patients, calls for antibiotics with a broader antimicrobial spectrum. Clindamycin has an antimicrobial spectrum which makes this antibiotic a possible alternative in community-acquired pneumonia, and its efficacy in pneumococcal pneumonia has been documented. However, as first choice therapy it should be reserved for cases of penicillin allergy, or cases of strongly suspected staphylococcal pneumonia. In aspiration pneumonia--nearly always caused by anaerobic bacteria--penicillin has long been the preferred therapy, even in cases with Bacteroides fragilis. However, recent publications have clearly documented that in primary lung abscess, clindamycin is superior to penicillin. These results are especially important since metronidazole has been shown to be less effective in such cases.