The objective of these studies was to optimize conditions for computer-assisted sperm analysis (CASA) of rat epididymal spermatozoa. Methodologic issues addressed include sample collection technique, sampling region within the epididymis, type of diluent medium used, and sample chamber depth. In addition, sources of variation were identified and accuracy of the analysis was examined. All samples in the report were analyzed using a Hamilton Thorn Motility Analyzer (HTM-2000; Hamilton Thorn Research, Danvers, MA). The authors found that allowing the sperm to swim out from cuts made in the distal cauda epididymidis yielded samples with percentages of motile sperm 60% higher than samples collected using an aspiration method. Furthermore, sperm isolated from the distal cauda epididymidis exhibited slightly but significantly greater percentages of motile sperm and swimming speeds than sperm isolated from the proximal cauda epididymidis. Of the four motility media examined, all maintained a high percentage of motile sperm over an hour-long incubation period, but Medium 199 and modified Hanks' Balanced Salt supported substantially greater sperm velocity than Dulbecco's Phosphate Buffered Saline (with Ca++ and Mg++), with or without glucose. Motility and velocity endpoints were comparable in 200-, 100-, or 40- micrometers deep chambers, but significantly lower in 20- micrometers deep chambers. (Copyright (c) 1991 Pergamon Press plc.)