Nellore Brown ewes (n = 20) aged 1-4 years reared under semi intensive conditions, with standard management conditions fed with greens, concentrates and adlib fresh water and salt licks were selected for the study. The ewes were subjected to estrus synchronization protocol of seven days using vaginal sponges and 300 IU of PMSG (Folligon, Intervet International, Boxmeer, Netherlands) and 75 µg of PGF_2α (Cloprostenol; Pragma, Intas Pharmaceutica Limited, Matoda, Ahmedabad) intra muscularly at sponge withdrawl. Upon sponge withdrawl the synchronized ewes were kept with a ram, color painted, and each day ram was replaced with a new one with changed color paint at brisket region. The next day of receiving last mating mark was treated as day 1, start of pregnancy. Transabdominal ultrasonography was conducted using a real time B mode scanning (ALOKA SSD 500, Aloka co Ltd, Japan) equipped with 5 MHz convex transducer on mated ewes. Serial ultrasonographic examinations were carried out on weekly basis starting from day 8 of mating. Once embryonic vesicle was identified the pregnancy was confirmed and scanning was performed weekly till end of gestation. During scanning biparietal diameter and occipito nasal length were identified and images frozen, saved and measured with built in electronic callipers.

By day 36, head and trunk of the embryo could be visualized, by day 43, a clear organization of the fetus into head, neck and body was observed in this study and these findings were in agreement with Ali and Hayder (2007) and Valasi et al. (2017) who reported differentiation of embryo into head, body and limbs on 38 ± 3.2 and 35 days respectively. On the contrary, Kumar et al. (2015) reported organization of fetal body on day 42, later than recorded in this study while Santos et al. (2018) reported in 4^th week, earlier than observed in this study. In 1939, Cloette grossly observed these structures between days 28 and 35 after conception.

The Orbits of the eye observed as anechoic pockets, mandible, nasal bone ossification was noticed in few by day 43. Ossification of the head bones was observed by Day 50. Present findings were in corroboration with Ali and Hayder (2007) and Kumar et al. (2015) who reported skull ossification between 42 - 48 days of gestation and differed from Valasi et al. (2017), who reported mandible and orbit walls ossification on day 36 and head bones ossification on day 56. These differences could be attributed to breed variation (Kumar et al., 2015), individual differences and inability to establish precise timing of ovulation as well as fertilization (Valasi et al., 2017).

The BPD was detected from day 36 until end of pregnancy. However, measured between day 43 to 113, after that the BPD exceeded the ultrasound screen length and obtaining images became difficult. Even though, head was detected from day 36, BPD could be measured in 50% (10/20) of ewes only on day 43. However, by day 50 BPD was measurable in all ewes. Many authors (Sergeev et al., 1990; De Bulnes et al., 1998; Airina et al., 2011; Santos et al., 2018 and Haq et al., 2020) measured BPD between day 32 and 125 in ewes and goats. Haibal and Perkins (1989) stated that it was not possible to obtain symmetrical images of skull before 40-50 days of pregnancy. Sergeev et al. (1990) opined that fetal skull was clear and defined between 49 – 84 days and was easy to measure, however fetal positioning influenced the ease of obtaining an image suitable for measurement.

In present study, BPD could be measured in all ewes until day 99, however on day 106 and day 113, head width could not be measured in 20% and 45% ewes respectively either due to fetal positioning or fetal size. This was in agreement with findings of Haibel (1988) who stated that the BPD measurement becomes difficult after day 105 of gestation due to variability of fetal head location and posture which may be anywhere in the abdomen from the inguinal region to the xiphoid (Jones and Reed, 2017) and diminished uterine fluids as proportional to uterine contents often defies measurements of head in transabdominal position. Kelly and Newnham (1989) also experienced constraints in measurement of head width from about day 100-110 of gestation. It could be due to increasing calcification of bones and shadowing effects. Sergeev et al. (1990) were unable to measure head width in the later stages of their study between day 106 to 119, due to inadequate depth of penetration of the 5MHz transducer employed. Haq et al. (2020) detected head occasionally after day 125, in the last month of pregnancy. This difficulty was attributed to the increase in fetal size and compression of the head by other fetal parts (Amer, 2010).

A high positive correlation (r = 0.96867; R² = 0.938) was derived between GA and BPD in the present investigation. These results were in consistent with findings of Sergeev et al. (1990), De Bulnes et al. (1998) and Gunduz et al. (2010) who reported correlation as r = 0.961, 0.96 and 0.977 respectively between GA and BPD in sheep and goats. Kuru et al. (2018) also reported similar results in Abaza (R² = 0.925) and Gurku (R² = 0.928) goats.

However, higher correlation (0.97-0.99) was reported by Haibel (1988), Haibel and Perkins (1989), Raichle and Haibel (1991), Suguna et al. (2008) and Santos et al. (2018) between GA and BPD in sheep and goats than the present study. Lower correlation in the present study might be attributed to the fact that the scanning was extended to third trimester, the intervals between consecutive scanning and breed differences. BPD measurements at early gestational ages (day 40−50) had moderate variability due to depth of the uterus in the abdomen (Haibel, 1988). In contrast some authors like Abdelghafar et al. (2011), Venkato et al. (2013) Petrujkic et al. (2016) and Haq et al. (2020) reported lower coefficient of determination (0.75 – 0.91) between GA and BPD in sheep and goats.

In the present investigation, ONL was detected from day 36 until end of parturition however measurements could be recorded between days 43 and 113. This was in consistent with Santos et al. (2018) who measured ONL from 6^th week of gestation, while De Bulnes et al. (1998) and Yazici et al. (2018) recorded ONL earlier than this study. However, Nwaogu et al. (2010) and Petrujkic et al. (2016) obtained measurements in later gestational days (from day 57 and day 46 respectively) than current study.

Nwaogu et al. (2010) measured ONL up to day 124 which is later than this study, While Kelly and Newnham, (1989), De Bulnes et al. (1998), Petrujkic et al. (2016), Santos et al. (2018) and Yazici et al. (2018) took measurements up to day 95, 91, 63, 98 and 74 respectively in sheep and goats. The limited observations made for ONL was due to specific fetal posture (Khan et al., 2015) or fetal dimensions exceeding the width of transducer in many ewes (Kelly and Newnham, 1989). From 13 week onward it was difficult to measure nasal to snout length due to fetal size, ultrasonographic characteristics, positioning difficulty and fetal movement (Santos et al., 2018).

A high positive correlation (r = 0.9873 and R² = 0.974) was established between ONL and gestational age in the present study. The correlation developed between ONL and GA was in accordance with the findings of Petrujkic et al. (2016) and Yazici et al. (2018). In contrast, De Bulnes et al. (1998), Nwaogu et al. (2010), Khan et al. (2015) and Santos et al. (2018) reported lower correlation (r=0.7 to 0.97).

The mean BPD and ONL (Table 1) showed significant increase (p<0.05) during the measurement period.

In the farmers flocks in 21 pregnant ewes BPD measurements were taken for validation and it was found that 61.9 (13/21) and 71.4 (15/21) % of deliveries occurred within ±7, and ±14 days of expected lambing dates. Amle et al. (2014) evaluated the equation developed by Abdelghafar et al. (2007) for BPD in Sangamneri and Osmanabadi goats (n=10 each) and found GA as 85.99 ±1.29 and 81.03 ± 2.16 days when observed GA was 88 and 81 days respectively. However, Haq et al. (2020) validated regression equation for BPD and recorded difference between actual and estimated day of gestational stage was 6.8± 1.0 days in Beetal goats (n = 20). In human fetuses the measurement of BPD before gestational week 20 predict the gestational age with an accuracy of ±7-11 days and the precision tolerance of BPD dimension decreased during the third trimester of pregnancy (Hadlock et al., 1991).

The regression equation for BPD used in this study was applicable between day 43 – 113 of gestation only. Further, the BPD measurements taken in the third trimester may not be true to its actual value due to fetal head positions which might lead to incorrect estimation of gestational age. Gestational age is best estimated by measurements in early pregnancy since at later gestational ages these measurements reflects the growth characteristics of the fetus rather than its age (Kurtz and Needleman, 1988). Prediction of GA would be increasingly unreliable from day 80 onwards due to differtial growth rates of skull even if accurate measurements were made as opined by Kelly and Newnham (1989).

In farmers flocks in villages in 12 pregnant ewes ONL measurements were obtained and validated. 58.3 (7/12) and 100 (12/12) per cent of ewes lambed within ±3 and ±11 days of expected lambing dates, respectively, for ONL The highest correlation r = 0.9873 obtained for ONL in the present study justified more than 50% of deliveries occurring within ±3 days of expected dates. No previous literature for validation of ONL in field condition obtained.

It was concluded that the ultrasonic measurements of BPD and ONL of the fetus could be valuable in the estimation of gestational age and for the use of the regression formulae to predict the parturition under field conditions, however ONL measurements gave better prediction than BPD and it is recommended to conduct the study on large number of ewes.