This study was designed to assess the reproducibility of the computer-assisted semen analysis (CASA) system. Semen specimens from 41 suspected subfertile patients were analysed on two identical Hamilton-Thorn Motility analysers (HTM, model 2030, version 7). Specimens were analysed after liquefaction at 37 degrees C for 30 min on two HTMs adjusted to the same gate settings and by using the same semen droplet loaded on a Makler counting chamber prewarmed at 37 degrees C. Significant differences were seen in the total sperm count (P < 0.0004), motile sperm count (P < 0.004), amplitude of lateral head displacement (P < 0.0001), linearity (P < 0.01), and beat cross frequency (P < 0.0001) between the two HTMs. No significant differences between the two machines was seen in the case of other sperm motion parameters: percentage motility, average path velocity, straightline velocity, curvilinear velocity, straightness, and progressive motility. Our results indicate that the reproducibility of several semen parameters analysed by HTM is poor, and steps to remedy this problem are necessary.