The chemotherapy of septicaemia in newborns differs fundamentally from that in older children or adults because, although newborns have a fully developed immunological system, the system has not yet "learned" to operate completely. Ultimately, optimal chemotherapy can only be found empirically. In this respect a few basic guidelines can be given however: 1. The initial therapy must bring the pathogen under control with a high degree of certainty, since a correction in therapy following pathogen indentification is usually too late. 2. Since the pharmacokinetics of antibiotics in newborns vary considerably, the minimal peak serum concentration observed should exceed the MIC of the pathogen. 3. In rapidly maturing newborns and premature babies the pharmacokinetics of each antibiotic must be known precisely. 4. Since in the individual case there can never be absolute certainty with respect to the three above-mentioned problems, combination therapy should be given at all times.