Rat spermatozoa recovered from different regions of the excurrent ducts of 10 adult males (proximal cauda epididymidis [PC], distal cauda epididymidis [DC], and vas deferens [VD]) were assessed by in vitro fertilization (LVF) using limited sperm numbers, and by continuous evaluation of motility parameters during 5 hours of incubation in vitro with automated computer-aided sperm analysis (CASA). Spermatozoa from the PC region fertilized (68 +/- 6%) a significantly greater (P < or = 0.005) number of oocytes than those from the DC (44 + 5%) or VD (47 +/- 7%). For pooled samples from all three regions, the mean fertilization rate (51 +/- 14%) was less tan for spermatozoa from the PC (P < 0.05) but was not significantly different from spermatozoa from the DC or VD. For each time point and sample, 1,592 +/- 428 sperm tracks were analyzed. CASA was verified by comparison with manual still-frame analysis of video recordings, by repeated analysis of the same or different samples of spermatozoa, and by examination of computer tracks. The coefficients of variation for various motion parameters suggested that the CASA obtained a high degree of precision. There were no significant differences in motility parameters for spermatozoa recovered from equivalent regions of the left or right tract or in motility parameters for spermatozoa from different regions of the tract immediately after recovery. However, during incubation in vitro, spermatozoa from the DC or VD regions exhibited a marked decline in straight-line velocity (VSL) compared with spermatozoa from the PC region. The reduction in VSL (combined values from right and left tract) for DC or VD spermatozoa compared with PC spermatozoa was significant at 2.5 hours of incubation (P < or = 0.05) and highly significant (P < or = 0.005) by the end of the incubation period. Differences in average path velocity (VAP) were also apparent after 4 hours (p < or = 0.05), but no significant differences were observed for measurements of curvilinear velocity (VCL) or lateral bead displacement (ALH). Overall, the decline in VSL over 5 hours was highly correlated (P < or = 0.001) with the outcome of fertilization in vitro. In contrast, initial VSL and changes in VCL of spermatozoa were not correlated with fertilization rate. These results indicate that the in vitro fertilizing capacity of rat spermatozoa is correlated with 1) the decline in straight-line velocity (VSL) as measured by repeated CASA during incubation in vitro and 2) with the site of recovery of mature rat spermatozoa from the distal excurrent duct. It is suggested that the deterioration of the quality of rat spermatozoa in the distal epididymidis and vas deferens during storage may occur sooner than previously realized, and therefore care must be taken when recovering samples for fertility assessment. In keeping with findings in other species, immediate "snapshot" analysis of rat motility was a poor predictor of sperm fertility. In contrast, continuous CASA provided significant information for determining sperm fertilizing capacity and will be a useful technique for reproductive toxicology.