The aim of this research was to estimate genetic parameters of male fertility after AI with three different types of AI doses obtained from ejaculates of bucks belonging to the Caldes line: 1) type 10: doses with 10 x 106 spermatozoa/ml and a period of 24 h of storage at 18ºC in a saline extender A. 2) type 40: doses with 40 x 106 spermatozoa/ml and the same storage conditions as type 10. 3) type X, doses prepared with semen diluted (1:5) with a saline extender B, but with unknown sperm dosage and no storage period. 3,628 AI were performed with the type 10 doses using crossbred females, 3,027 with the type 40 doses and the same population of females, and 5,779 with the type X doses, using purebred females from the Caldes line in a different farm. Fertility after AI with type 10 doses (F10), type 40 doses (F40) and type X doses (FX) was considered as three different binary traits. Data were analyzed under a three-trait threshold model. The mean of the marginal posterior distribution (MPD) for F10 minus F40 was estimated to be -0.13. This result indicates a clear effect of the sperm dosage on fertility, which could be non-linear. The mean of the MPD of F10 minus FX and F40 minus FX were respectively, -0.37 and -0.23 which indicates that the effect of the storage conditions on fertility could be even more important on fertility than sperm dosage, since FX was very close to fertility after NM and sperm dosage of this type of doses was in average lower than 50 x 106 spermatozoa/ml. Heritabilities seem to be similar for F10 and F40 and both of them could be higher than heritability of male fertility after NM and FX. Variance of the genotype x sperm dosage interaction was almost negligible since additive variances were similar for F10 and F40 and their genetic correlation was close to 1. However, this interaction could be more important between the genotype and the storage conditions.