It is now widely accepted that the clinical diagnosis of deep venous thrombosis (D.V.T.) is unreliable. Many venous thrombi are nonobstructive and not associated with vessel wall inflammation or inflammation of the surrounding tissues and consequently have no detectable clinical manifestations. Moreover, none of the symptoms or signs of venous thrombosis are unique to this condition and all can be caused by non thrombotic disorders. On the other hand, in most of the D.V.T., the calf is the site in the legs where a thrombus starts. This thrombus begins commonly in valve pockets throughout various deep veins of the leg and in saccules of soleal veins. Several non invasive techniques have been developed for diagnosing D.V.T.: 125I, fibrinogen, impedance plethysmography, Doppler ultrasound, Duplex scanning. Many publications document the correlation between venography and these non invasive tests for D.V.T. Unfortunately it appears that, excepted 125I. Fg, these techniques are poorly reliable at the level of the calf. Moreover the diagnosis of D.V.T. may occur in particular and difficult situations such as a recurrent deep vein thrombosis. Considering all above the authors believes that contrast venography remains the standard and that it is less dangerous to do unnecessary venography than not to recognize a deep vein thrombosis.