Several factors that influence the sensitivity of the syncytia infectivity assay for the bovine leukemia virus (BLV) and BLV-infected lymphocytes have been examined. The use of early-passage indicator bovine embryonic spleen (BESP) cells and their pretreatment with diethylamino-ethyl-dextran (DEAE-D) was essential for optimal sensitivity. Polybrene was less effective than DEAE-D. The combination of DEAE-D and polybrene was more effective than DEAE-D alone when BLV-infected leukocytes were used as the inoculum, but not when the inoculum was a cell-free BLV preparation. Using BESP cell passages 4 to 11 as indicators, reproducible titers were obtained when aliquots of the same virus stock were assayed at different times after freezer storage. When assaying peripheral blood lymphocytes from infected cattle, optimal syncytia responses were observed consistently by inoculating 5 X 10(6) viable lymphocytes per 60-mm Falcon dish. Centrifugation of peripheral blood leukocytes from BLV-infected cattle in discontinuous bovine serum albumin gradients can be used to separate a subpopulation of infected lymphocytes. Use of this subpopulation as the inoculum, rather than unseparated buffy-coat leukocytes, greatly increases the sensitivity of the syncytia infectivity assay.