In order to improve fetal sexing in the Dorper sheep breed, the objective of the present study was to determine, by repeated ultrasonographic examinations, the migration period of the genital tubercle (GT) in sheep fetuses derived from natural mating or embryo transfer and to compare the accuracy of a single examination with repeated examinations at short intervals. For this purpose, transrectal ultrasound was performed, using a double-frequency linear transducer (6.0 and 8.0 MHz) for monitoring 51 sheep fetuses distributed in three experimental groups (EI, EII and EIII). The fetuses in EI (n = 23) and EII (n = 18) derived, respectively, from natural mating and embryo transfer were monitored at 48-h intervals from the 30th to 60th day of gestation and sexed based on the final location of the GT. The fetuses in EIII (n = 10), which originated from embryo transfer, were examined only once on the 65th day of gestation and sexed taking into consideration the final position of the GT and/or by identification of anatomical structures of external genitalia. The accuracy of fetal sexing was 91.3% (21 fetuses sexed/23 quantified) in EI, 88.9% (16 sexed/18 quantified) in EII and 100% (10 sexed/10 quantified) in EIII, without significant difference (P > 0.05) between experiments. Migration of the GT occurred earlier (P < 0.05) in fetuses produced by natural mating (43.0 +/- 2.8 days) than in those derived from embryo transfer (46.1 +/- 4.7 days). The results show that fetal sexing can be done from the 50th day onward in fetuses produced by natural mating and from the 60th day onward in fetuses derived from frozen embryos. It can also be concluded that repeated ultrasonographic exams in short time intervals do not maximise the accuracy of fetal sexing. In addition, real-time ultrasonography is a reliable tool for fetal sex determination in sheep after Day 50 of gestation, taking into account both the location of the GT and the identification of external genital structures.