The incidence of craniotomy infections, usually less than 5%, is dependent on many factors, such as how the information is collected and how the percentage is calculated. Because these factors may vary from report to report, incidence figures should be read with skepticism. It is difficult to prove that a given factor contributes to infection. Most routines are based more on personal convictions than on solid evidence. CSF leak is one factor known to have great impact; it should be avoided with painstaking technique and, if it occurs, it should be treated promptly. Solid evidence favoring prophylactic antibiotics for persistent CSF leak is not available; but, until a well-designed randomized study tells otherwise, the high risk of meningitis justifies prophylaxis. Penicillin is adequate for leaks through the nose or the ear. For leaks through the skin, the antibiotic should be effective against staphylococci. The infection register should provide information about prevailing bacteria. In many hospitals, the prophylaxis should cover gram-negative bacilli. CRP is a useful diagnostic aid for detecting postoperative infections. The operation, however, also causes a CRP rise. Daily CRP monitoring, at least for patients with elevated temperature, is recommended. The third-generation cephalosporins are a welcome contribution to the treatment of bacterial meningitis. To avoid side effects, and to keep them potent when they are really needed, they should be used with caution. Most postoperative cases of meningitis are in fact aseptic. If the patient is moderately ill, chloramphenicol is still eligible as the first choice antibiotic. When the bacterial culture is negative, the antibiotic should be stopped. The standard treatment for bone flap infection is removal of the bone flap. The bone flap is essentially devascularized and comparable to a foreign body. The justification of vancomycin prophylaxis has been shown in a randomized study.