The purpose of this study was to evaluate relationships between measured sperm velocity and in-vivo fertility, using donor semen samples from an artificial insemination (AID) programme. Seventy-one frozen semen samples were examined; measurements of sperm velocity were made immediately after thawing, upon a motile 'swim-up' fraction, and finally after 3.5 h incubation at 37 degrees C in the freezing mixture. Zona-free hamster egg penetration assays were performed upon all samples. Two groups of samples were identified; seven donors (11 samples) had failed to produce any pregnancies through AID from a range of 3 to 14 cycles tested, whilst the remaining samples (from 25 donors) had achieved at least one pregnancy each. The mean sperm velocity (+/- SEM) for the latter 'fertile' group was significantly higher than the corresponding value for the 'infertile' group; (i) after thawing, 65.9 +/- 1.8 versus 50.4 +/- 3.2 microns/s (P less than 0.001) and (ii) after 3.5 h incubation, 42.1 +/- 2.1 versus 24.7 +/- 5.7 microns/s (P less than 0.002). Using the maintenance of sperm velocity during incubation as an indicator of survival, life-table analyses were used to calculate monthly conception rates on various sub-groups of the semen samples. Poor survival (greater than 40% decline in velocity over 3.5 h) was associated with a monthly pregnancy rate of only 11.58% (362 cycles), whilst better survival (less than 40% decline) was associated with the significantly higher (P = 0.024) pregnancy rate of 16.87% (480 cycles).