1. A series of 705 fractures of the tibia is reviewed, 674 of which were treated conservatively.
2. The factors most conducive to delayed or non-union are initial displacement, comminution, associated soft-tissue wound and infection. The extent to which these are combined in any fracture determines its "personality" and its inherent propensity for union.
3. Eight fracture types are differentiated based on the above "personality rating." The incidence of delayed union or non-union varies from 9 per cent in the most favourable type to 39 per cent in the least favourable. Infection raises the incidence to 60 per cent. Comparative statistics which fail to recognise these differences can be entirely misleading.
4. Continuous traction does not retard union.
5. The results of conservative treatment are analysed with regard to union, deformity, stiff joints and contractures and the conclusion is reached that no case has yet been made out for internal fixation as the method of choice in the treatment of this fracture.