Abstract Objectives The objectives of this experiment were to determine in vitro the permeability of teeth exhibiting a range of natural dentine caries and the effects of caries excavation and subsequent restoration on the dentine permeability of extracted teeth. Methods Forty-five human permanent molars with obvious occlusal caries were selected and coronal dentine permeability was measured before (baseline) and after caries excavation, as well as after acid etching the prepared cavity and finally, after restoration. Results The small permeability values for all teeth made statistical testing between the different stages of specimen preparation both meaningless and inappropriate. An alternative means of examining the data was to calculate the percentage of teeth which had exhibited permeability values greater than zero at each of the four specimen preparation stages. The percentage of teeth with permeability greater than zero remained unchanged before and after cavity preparation (23%). Removal of the smear layer, however, increased substantially the percentage of teeth demonstrating permeability greater than zero, to 72%. None of the restored teeth demonstrated measurable permeability. Conclusion The teeth chosen exhibited a range of dentine caries and, surprisingly, demonstrated either low or no measurable permeability. This experiment highlighted the need for more investigation into the quality of carious-affected dentine and the properties of the smear layer produced from such dentine.