Because of the difficulty of cryopreservation of boar semen, fresh ejaculated semen is still the main source of sperm used in pig IVF. However large variations among boars in terms of oocyte penetration and rates of polyspermy have been observed (Sirard et al., J Reprod Fertil Suppl 48, 3-16, 1993). Even ejaculates from the same boar may behave differently in vitro as inconsistent penetration rates were observed when the same boars were used within an experiment. The objectives of the present study were: 1) to determine the contribution of different fractions of single ejaculates to observed variability in quality of boar semen used for IVF, and 2) to standardize and improve techniques for evaluating fresh ejaculated semen from different boars using IVM and IVF systems. Ejaculates from 3 young boars were collected on 4 occassions as a series of separate 15 ml fractions. On the basis of sperm concentration, 3 fractions representing the first peak concentration (fraction 1), the lowest sperm concentration after fraction 1 (fraction 2) and the second peak concentration (fraction 3) were selected for in vitro analysis. Oocyte-cumulus-granulosa cell complexes were obtained by dissection from slaughterhouse ovaries and were co-cultured with 2 follicle shells as described by Ding et al (Mol Reprod Dev 33:59, 1992). In vitro matured oocytes were randomly assigned for fertilization by the 3 semen samples from each boar. Sperm concentrations were adjusted to 4 10 8 sperm cell/ml in all samples with Modified TCM-199 (pH 7.8) during a 90 min prefertilization incubation and the final concentration for fertilization was 5 10 5 sperm cell / ml. Data were analyzed using ANOVA for a split-plot design (4 replicates as blocks). In the presence of fraction effects, Student-Newman-Keuls (SNK) test was used for multiple comparison of treatment means. Oocyte penetration rate differed among fractions (P = 0.001) and varied from 69 to 100% (mean 95.7%) for fraction 1, from 0 to 100% (mean 53.3%) for fraction 2 and from 50 to 100% (mean 89.9%) for fraction 3. There were also differences in male pronuclear (MPN) formation rate (P = 0.028; mean 27.6, 9.3 and 16.4% for fraction 1, 2 and 3, respectively), in the rate of polyspermy (P = 0.0001; mean 92.3, 31.9 and 76.3% for fraction 1, 2 and 3, respectively) and in the number of penetrated sperm per oocyte (P = 0.002; mean 5.58, 1.94 and 4.07 for fraction 1, 2 and 3, respectively). These data confirm that the fraction of the ejaculate used for semen evaluation in the IVM-IVF systems had a significant effect on the results obtained. The first peak concentration of semen (fraction 1) showed superiority in fertilization ability and less variability in penetration rate from replicate to replicate compared with the other 2 fractions. By multiple comparisons for boar effects, boar 1 showed higher rates of penetration (P < 0.05), MPN formation (P < 0.05) and polyspermy (P < 0.05) than boar 2 or 3. In conclusion, the criteria used in this study allowed discrimination between boars in apparent semen quality. Because of the effect of ejaculate fraction, but no fraction by boar interaction, use of a standardized semen fraction is essential for optimizing comparisons between boars, and the first ejaculate fraction will provide the best estimate of semen quality in IVM-IVF systems.