No prospective randomized clinical studies are available to determine exactly how much time should be spent on investigation before initiating antibiotherapy in a patient with presumed bacterial meningitis. Experimental models show that antibiotics should be administered before the inflammatory response, but at this time the patient's symptoms are often unspecific. Models also demonstrate that a gain of time is beneficial at any time, in terms of inflammation, magnitude of bacteremia, or loss of hearing. Very few clinical studies address the acceptable delay between admission and administration of antibiotics and two of these show a correlation with outcome in adult meningitis. The available data supports the recommendation that hospital investigation of a patient with presumed bacterial meningitis should be conducted in such a way that efficient antimicrobial chemotherapy will be initiated within one hour after arrival.