Past efforts to characterize the ability of antibiotics to penetrate into tissues is reviewed, particularly the results of models that reflect the concentration of antibiotics in interstitial fluid (i.e., skin blisters, tissue chambers, skin chambers, wound exudates, implanted fibrin clots, and peripheral lymph). The preferable sources of such fluid appear to be skin chambers, suction-induced skin blisters, and peripheral lymph. Extent of penetration of antibiotics into tissue fluid is related to the amount of antibiotic not bound to protein. Protein binding is particularly inhibitory to penetration when >80% of the antibiotic is bound. It is preferable to follow the extravascular concentrations of antibiotic for longer periods than is usually done. The ability of antibiotics to penetrate is best evaluated by use of the ratio of the area under the concentration curve (AUC) for antibiotic in the peripheral locus to the AUC for serum. Pharmacokinetic analysis should be done with data from each individual, not with data derived from curves using mean values. Penetration into fibrin, lymph, and chambers implanted in tissues follows Fick's law of diffusion. Lag times indicate the interval before penetration of antibiotic in serum to the peripheral locus may be observed. First-order, one-compartment, open models are applicable to the characterization of extravascular concentrations of antibiotics. The AUC for antibiotic in serum during a time interval is the factor determining the amount of agent that passes into an extravascular focus. Concentrations of drug in tissues are lower than those in serum; the peaks occur simultaneously or shortly after the maximal levels are reached in serum, and for most antibiotics the elimination of antibiotics from extravascular parts of the body is slower than from serum, particularly for agents with a half-life in serum of <4–5 hr.