Objective: The objective of this study was to compare the fracture resistance of non-vital teeth restored with three core materials. Methods: Sixty sound maxillary premolar teeth of similar size were randomly allocated to four groups (n=15) to be restored as follows: amalgam core (Group I), resin composite core (Group II), resin modified glass ionomer core (Group III) and sound teeth which acted as controls (Group IV). Teeth from Groups I-III were decoronated 1mm above the cemento-enamel junction prior to receiving endodontic therapy. Gutta percha was removed from the access chamber and to a depth of 3mm in each root canal. Each tooth was restored with the respective core material, which was placed according to the manufacturer’s instructions to give a core of similar anatomical dimensions. The teeth were loaded to failure with the mode of failure recorded. Results: The mean (mean, s.d.) fracture resistance (KN) of the restored teeth was as follows: Group I (1.94, 0.22), Group II (0.75, 0.11), Group III (1.05, 0.20) and Group IV (0.79, 0.20). Groups I and III had significantly greater mean fracture strengths (p=0.00) than Groups II and IV which were similar. A higher proportion (67%) of teeth in Group I had minimal modes of failure in comparison to other groups. Notably, teeth in Group II had significant modes of failure with up to 50% of the core lost. Conclusion: Within the limitations of the study it was concluded that amalgam cores placed in root treated premolar teeth had the highest fracture resistance when compared with the other materials tested. It was also concluded that resin modified glass ionomer cement may be a more suitable alternative to amalgam than resin composite as a core build-up material.