We studied the feasibility of clinical tests in the diagnosis of syndesmotic injury of the ankle. 9 investigators examined 12 persons twice, including 2 patients with an arthroscopically-confirmed syndesmotic injury. They sat behind a curtain that exposed only the lower legs. We found a statistically significant relation between the final arthroscopic diagnosis and the squeeze, fibula translation, Cotton, and external rotation tests as well as for limited dorsal flexion. None of the syndesmotic tests was uniformly positive in chronic syndesmotic injury. The external rotation test had the fewest false-positive results, the fibula translation test the most. The external rotation test had the smallest inter-observer variance. The physical diagnosis was missed in one fifth of all examinations. When in accordance with medical history and physical examination, positive stress tests should raise a high index of suspicion of syndesmotic instability. The final diagnosis of such instability, however, should be made by additional diagnostic imaging and/or arthroscopy.