It has been generally assumed that release of antibiotic from methylmethacrylate occurs either from the surface, through pores in the matrix of the cement, or by diffusion through the matrix. In vitro and in vivo studies of the release of antibiotic from cement have produced variable and inconsistent results. In our laboratory, preliminary observations suggested that antibiotic is released from methylmethacrylate by flow through an interconnecting series of voids and cracks in the cement, rather than through diffusion after having been homogeneously distributed throughout the cement. Therefore, experiments were performed to answer the fundamental question of whether the matrix of methylmethacrylate bone cement is permeable to gentamicin. In vivo elution studies were performed on injection-molded rods of methylmethacrylate that had been loaded with two different amounts of gentamicin. The first group of rods contained 0.5 gram of gentamicin for each packet and the second, 1.5 grams for each packet. The rods were embedded subcutaneously in the subcostal region of sheep for three months. Bioassay of sections of the rods, using the tube-diffusion technique of Mitchison and Spicer, showed that the more highly loaded cement had released a significantly greater proportion of gentamicin. This occurred because the more highly loaded cement contained a greater number of defects that contained gentamicin (filled voids and interconnecting cracks). In vitro diffusion studies were also performed, using 0.8-millimeter-thick disks of methylmethacrylate that did not contain antibiotic. Test solutions of either gentamicin or methylene blue were placed in the inner compartments of diffusion chambers. The outer compartments contained tissue-culture medium 199, which was sampled monthly and assayed for gentamicin or methylene blue.(ABSTRACT TRUNCATED AT 250 WORDS)