Purpose: To evaluate the usefulness of various radiographic imaging modalities in the diagnosis and characterization of melorheostosis. Materials and Methods: We retrospectively evaluated the plain film (n=8), computed tomographic (CT) imaging (n=5) and magnetic resonance (MR) imaging (n=5) findings of eight patients with melorheostosis diagnosed by bone biopsy (n=4) and characteristic radiographic findings (n=8). MR images were obtained with a 1.5-T scanner focused on the region of maximal radiographic abnormality. Pulse sequences include T1-weighted SE, T2-weighted fast SE (n=5) and postcontrast imaging (n=4). In order to define subtle enhancement of the lesions, subtraction MR images were obtained in one case. Imaging findings were analyzed with particular emphasis on the distribution of lesions along the sclerotome, differential radiographic findings between diaphyseal and metaepiphyseal lesions of the long bones, as seen on plain radiographs, and the density and signal characteristics of hyperostotic, lesions, as seen on CT and MR images. Results: Characteristic distribution along the sclerotome was identified in five of eight cases mainly along C6 and 7 (n=2) and L3, 4 and 5 (n=3) sclerotomes. In diaphyseal melorherostosis (8/8), a characteristic finding, i.e., a wax flowing down from the candle, was identified on plain radiographs. In all three patients with metaepiphyseal melorheostosis (3/8), multiple round or oval hyperostotic lesions were seen in the epiphysis and metaphysis of the long bones. On CT, the marrow cavity was partly obliterated by hyperostotic lesions in all five patients with endosteal hyperostosis. Among these, central ground glass opacity with a sclerotic rim was seen in three patients. Periosteal hyperostosis was seen in two of five cases, being visualized as irregular excrescences in the periosteal region and surrounding soft tissue. Individual hyperostosis was visualized as hypointense on T1-weighted images and as a hyperintense center with a surrounding hypointense rim on T2-weighted images (5/5). On postcontrast images, central enhancement was noted in all four cases. In one of these, in which the degree of central enhancement was subtle, subtraction images (postcontrast SE- precontrast SE) also revealed a central signal increment. Central enhancement corresponded to the hyperintense center seen on T2-weighted images (4/4) and the ground-glass opacity seen on CT (2/2). Conclusion: Radiographic imaging plays a crucial role in the diagnosis of melorheostosis. The future role of gadolinium-enhanced MR imaging in the characterization of the lesion may be important though further evaluation and pathologic correlation is required.