The method of conservative (rim) resection of the mandible is now well established and provides good control of disease in the primary site. There are few audits of this technique in terms of margins of resection for both rim and segmental resection of the mandible. Consecutive previously untreated patients managed with resection of the mandible as part of their treatment for oral and oropharyngeal squamous cell carcinoma were recruited for the study. The presence and extent of tumour invasion of the mandible was recorded and a retrospective analysis made to establish the incidence of unnecessary segmental resections. The invasion rate was 33% (8/35) for rims and 83% (54/65) in segmental resections of the mandible, which compares favourably with previous studies and indicates a reasonable accuracy of resection. In between 6 and 11 of the 62 cases (10-17%) a rim resection could have achieved a satisfactory resection margin and retained a functioning lower border of the mandible. The accuracy of resection in terms of margins was greater for mandibular resections (94% clear margins) than soft tissues (62% clear margins). The number of compromised margins was significantly greater in the invaded rim resection group (P=0.018). This study indicates that a more conservative mandibular resection was possible in a few cases. This is unlikely to have an adverse effect on the close or involved margin rate, which depends mainly on the accuracy of the soft tissue resection. Angling the horizontal rim resection to take into account tumour entry at the point of contact will help to ensure a clear bone margin if a conservative approach to mandibular resection is an option.