This study attempts to classify scaphoid nonunions, identify the prognostic indicators of scaphoid nonunions and report the clinical, functional and radiographic results of scaphoid nonunions treated by Herbert screw and bone grafting. The patient population included 51 scaphoid nonunions in 50 patients treated by Herbert screw with or without bone graft. Group I was patients with delayed union; Group II was stable nonunions, Group IIIA consisted of unstable nonunions without DISI; Group IIIB included unstable nonunions with DISI deformity. Although the overall union rate was 84%, 87% stable nonunions (Groups I and II) united compared to only 60% unstable (Groups IIIA and B). This was statistically significant using the Chi-square test at p < 0.05. Thus, delayed unions and stable nonunions unite readily but unstable and displaced nonunions have significantly lower rates of union even after Herbert screw and bone grafting.