The liver is almost universally involved in systemic amyloidosis. Patterns of topographic distribution of amyloid within the liver lobule have been recognized, but the reliability of using these for classification of amyloid type is in question. We examined 286 livers from cases of systemic amyloidosis obtained from autopsies at Los Angeles County-University of Southern California Medical Center, classifying them as AL or AA type by means of the potassium permanganate Congo red-staining method along with a specific anti-AA antiserum. Prior publications have asserted that deposition of secondary (AA) amyloidosis is limited to the vessels in the portal tract, constituting a "vascular" pattern, and that in primary (AL) amyloidosis the deposits exhibit a "sinusoidal" pattern in that they are seen along hepatic sinusoids as well as in portal vessels. We confirmed that AL amyloid involves the portal vessels as frequently as AA amyloid and that deposition occurred significantly more frequently in the portal stroma, the central vein, and the "sinusoidal" areas. However, we also found a "sinusoidal" pattern in 29 of 78 cases of secondary (AA) amyloidosis; in 14 of these, more than half of the sinusoidal spaces were replaced by amyloid deposits. We also noted that in 23 of the 29 AA amyloidosis cases with "sinusoidal" involvement, a "sago" pattern of distribution of amyloid in the spleen was present. No consistent association of a specific chronic inflammatory disease with "sago" spleen and "sinusoidal" deposits could be documented. We conclude that topographic distribution of amyloid within the liver lobule is not a reliable method of distinguishing AA from AL amyloidosis and that specific staining methods must be used if the physician is to be able to attempt modern therapeutic modalities.