Abstract Although widely endorsed for specific treatment of pneumococcal pneumonia, penicillin G is seldom used for this purpose in clinical practice for at least three reasons: (1) concern about penicillin-resistant Streptococcus pneumoniae (PRSP) strains; (2) the difficulty of making an early etiologic diagnosis of pneumonia; and (3) lack of a clear consensus about the optimum dosage. Continuous infusion of 20–24 million units of penicillin per day provides serum levels of 16–20 μg/mL in persons with normal renal function. These levels easily exceed the minimum inhibitory concentrations (MICs) of penicillin G against most PRSP strains (4 μg/mL), which are actually strains with reduced susceptibility to penicillin. High-dose penicillin G therapy has not been shown to be therapeutically ineffective against pneumonia due to PRSP strains. However, the extent of penicillin resistance warrants continued monitoring, because strains exhibiting extremely high-level resistance (MIC ≥8 μg/mL) would probably respond poorly if at all. Development and use of rapid, sensitive, specific ways to diagnose pneumococcal pneumonia could extend the usefulness of penicillin G, thus postponing the emergence of resistance to other antibiotics.