The portal vein wall typically is hyperechoic over a wide range of beam-vessel angles, whereas the hepatic vein wall is hyperechoic only when the inci· dent beam and the vessel are perpendicular. This has been attributed to marked discrepancies in mural thickness, collagen content, or perivascular fat between portal and hepatic veins. We evaluated histologically the walls of portal and hepatic veins using three cadaveric livers. For vessels with lumi· nal diameter above 2 to 3 mm, hepatic vein and portal vein wall thicknesses were similar such that portal vein walls were not more than 50% thicker than those of hepatic veins of comparable size. Hepatic vein walls were mostly composed of parallel, tightly packed collagen fibers. In contrast, portal vein walls were composed of loosely arrayed, non parallel connective tissue fibers which were separated by multiple intervening spaces and only a minority of which were collagenous. Perivascular fat was not identified adjacent to intrahepatic vessels beyond the liver hilus. The marked differences in echogenicity between portal vein and hepatic vein walls typically observed at ultrasonography thus cannot be attributed to differences in mural thickness, collagen content, or perivascular fat between these vessels. Rather, the distinct composition of the hepatic vein wall renders it a specular reflector, which is hyperechoic only when the angle between the ultrasound beam and the vessel wall is dose to 90 degrees, whereas the composition of the portal vein wall enables it to appear hyperechoic at a wide range of beam-vessel angles. KEY WORDS: Liver, ultrasonography; Abdomen, ultrasonography; Hepatic vein; Portal vein. I t has long been known that echoes reflected from HV walls typically are less bright than those reflected by PV walls. Chafetz and Filly noted that HV walls were hyperechoic when HVs were aligned perpendicular to the ultrasonographic beam, in contrast to PV walls, which were hyperechoic over a wide range of beam-vessel angles.l They suggested that this phenomenon was due to more abundant connective tissue in PV walls than in HV walls,l an extrapolation based on the wellknown marked discrepancy in thickness between HV and PV walls at the level of the microscopic hepatic lobule.z They also suggested that differences between PVs and HVs in the quantity of perivascular fat might contribute to observed differences in echogenicity.l The imaging literature has supported the idea that the discrepancy between PV and HV 0 1997 by the American Institute of Ultrasound in Medicine • J Ultrasound Med 16:807-810, 1997 • 0278-4297/ 97/ $3.50