The objective of this study was to determine the effect of removal of the smear layer on canal obturation as measured by penetration of bacteria from a coronal direction. Fifty-four extracted human teeth were decoronated and instrumented in a uniform manner. Following instrumentation the root canals of 20 teeth were rinsed with 17% EDTA and 5.25% NaOCl to remove the smear layer before obturation. A second group of 20 teeth were flushed with NaOCl alone. The teeth of both groups were obturated with Thermafil plastic carriers and Roth's sealer. The root canals of another 10 teeth, five rinsed with EDTA and five without, were obturated with Thermafil without sealer. Two teeth serving as positive controls were instrumented but not obturated, while another set of two were sealed coronally and apically to serve as negative controls. The root surface of each tooth was sealed with nail varnish. A small chamber was thoroughly sealed around the coronal aspect of each tooth so that bacteria placed therein could move only through the obturated canal space. Each tooth was placed in a test tube containing sterile trypticase soy broth (TSB). An inoculum of Proteus vulgaris in TSB was placed in each coronal chamber at five day intervals and daily observations were made for bacterial growth in the apical reservoir for 21 days. Both positive control teeth showed bacterial penetration after 24 h. Neither of the two negative control teeth demonstrated penetration for the duration of the study. The frequency of bacterial penetration through teeth obturated with intact smear layer (70%) was-significantly greater than that of teeth from which the smear layer had been removed (30%) P < 0.05. All but one tooth obturated without sealer exhibited bacterial penetration, irrespective of the presence or absence of smear layer. Removal of the smear layer enhanced sealability as evidenced by increased resistance to bacterial penetration.