The use of threaded endodontic dowels is a controversial issue. The purpose of this study was to compare the potential for root fracture resulting from the cementation of nine threaded and three nonthreaded endodontic dowel systems. The clinical crowns of 140 extracted premolars were removed at the cementoenamel junction. The teeth were randomly assigned to the following treatment groups of 10 teeth each: Group 1, endodontically instrumented but not obturated; group 2, instrumented and obturated; group 3, instrumented, obturated, and restored with custom-cast gold dowel and cores; groups 4 and 5, instrumented, obturated, and restored with prefabricated, nonthreaded dowels; and groups 6 through 14, instrumented, obturated, and restored with one of nine prefabricated, threaded dowels. All dowels were inserted according to manufacturer's directions, removed, and cemented with vinyl polysiloxane impression material. Each specimen was demineralized and cleared. Photographs at 1:1 magnification were taken to assess dowel fractures. Fisher's test and chi square analysis were performed to evaluate the differences between post types, and between posted and nonposted controls (p less than 0.05). The results indicate no statistically significant differences between dowel types when compared with each other, regardless of dowel shape, taper, or presence or absence of threads, or when compared to instrumented, nonobturated controls. The amount of remaining dentin and existing root morphology may be a determining factor for endodontically treated teeth to resist fracture during dowel placement.