Phleborheography, a recently described noninvasive test for deep venous thrombosis, was compared with leg venography in 75 patients. Acute deep venous thrombosis was accurately diagnosed by phleborheography in 24 patients, with no false-positive diagnoses. External venous compression without thrombosis was diagnosed correctly in two patients. The remaining patients appeared normal or had chronic venous disease by phleborheography; however, 11 of these had acute deep venous thrombosis by venography, for a false-negative rate of 31%. Most undetected thrombi were in small calf veins. The specificity of phleborheography is thus 100%, but the sensitivity is only 69%. Similarly, its positive predictive value is 100% and the negative predictive value is 78%. When phleborheography shows acute deep venous thrombosis, this diagnosis may be accepted with confidence and therapy chosen accordingly, without venographic confirmation. Venography may still be required to withhold anticoagulation when phleborheography is negative.