Fertility obtained by cross-breeding mares (Equus caballus) with jackasses (Equus asinus) was evaluated. Two extenders, containing skim milk-glucose or egg yolk-glycine were used to study the fertility of mares inseminated with diluted jackass semen (T1 and T2) or diluted and cooled semen at 5°C for 12 hours (T3 and T4). A total of 272 cycles of 208 mares of undefined breeds were evaluated, being uniformly distributed between groups. The cycles were controlled by transrectal palpation and teasing, and mares were inseminated every Tuesday, Thursday and Saturday (three times/week), from the detection of a follicle with 3.0 to 3.5cm diameter in one of the ovaries until ovulation. Pregnancy was detected using transrectal palpation, teasing and ultrasound exams every 14 days. The extenders had no effect on fertility (P>0.05). Pregnancy rates for the first cycle were 64.52%, 61.11%, 50.72% and 54.17% and pregnancy rates/cycle were 63.64%, 54.55%, 52.69% and 47.06%, respectively, for T1, T2, T3 and T4. Differences in pregnancy loss rates between groups and effect of month of conception on fertility were found. Pregnancy loss rates were significantly higher (P<0.05) in January (38.46%) and in February and March (52.38%), with an average of 33.09%. The results indicate that mares conceiving at the end of the physiological reproduction time, carrying a mule embryo, are more susceptible to pregnancy loss.