Heterosis retained in different generations of inter se mating between D’man and Sardi sheep

The improvement of sheep productivity requires efficient actions on different components, mainly on prolificacy. The use of prolific breeds in the creation of synthetic breeds allows this objective to be attained and the wishes of breeders who like to have one breed type which permits replacement from their own flock to be realized. Nevertheless, the efficiency of such a breeding system is related to the heterosis at the first generation and to the proportion of retained heterosis in the subsequent generations of inter se mating (Young et al . 1986). A programme for the creation of a synthetic breed of sheep having 50% D’man and 50% Sardi was initiated ( Boujenane & Bradford 1991), and preliminary results were reported (Boujenane & Chafik 1994). The objectives of this study were to analyse the reproduction, growth and survival performance of Sardi, D’man and F 1 to F 5 generations of crossbred sheep, and to calculate the heterosis and the heterosis retained at different generations. Material and methods Animals The study was undertaken at the Tadla Farm of the Institut Agronomique et Vétérinaire Hassan II. The farm is located at 32.5°N, 7°W at approximately 415 m elevation in an irrigated area of the interior of Morocco, 20 km north of the High Atlas mountain, 150 km east‐northeast of Marrakech and approximately 160 km from Casablanca on the Atlantic Coast. The analysis has concerned 1418 litter records of 751 ewes and 1702 growth records of 1996 lambs born from 97 sires. The mean age and weight at mating of the ewes were 34 months and 42 kg, respectively. The average number of lambings was 2.4. Data were collected during six lambing periods from 1990 to 1995 inclusive. The matings were made to establish a new synthetic breed. Ewes were of Sardi, D’man, F 1 , F 2 , F 3 and F 4 breed groups, whereas lambs were of Sardi, D’man, F 1 , F 2 , F 3 , F 4 and F 5 breed groups. In this study the F 1 is defined as the first generation that reflects the final breed composition of the synthetic breed which was established by using the same sires and dams used in the parental breeds. Animals of the F 1 to F 5 generations were produced by inter se mating. Crossbred ewes were mated to rams of their own breed group, but purebred D’man (D) and Sardi (S) ewes were mated to both D’man and Sardi rams in order to produce purebred and reciprocal F 1 (S × D and D × S) lambs. Details concerning the parental breeds and the origin of their samples were reported by Lahlou‐Kassi et al . (1989). Management Ewes were managed under an annual lambing system. They were mated for the first time at 16 months of age. Mating periods started on average at June 25 and lasted 40 days. Each year, ewes were placed at random in different pens, with an average of 20–25 ewes per ram. Mating was performed indoors at night. The rams were used for 1 or 2 years. Apart from the mating period, all ewes and their subsequent lambs were subjected to the same management. The ewes were kept on pasture (fallow and wheat stubble), except during the mating and lambing periods, when they were kept in confinement and fed on alfalfa hay, barley, sugar beet pulp and a mineral and vitamin mixture such as their dietary requirements were covered. When the lambs reached 1 month of age they were creep‐fed ad libitum on a concentrate composed of barley, sunflower and mineral and vitamin mixture. The ewes were vaccinated against enterotoxemia. The lambs were injected with 1 ml of a commercial product (Bioselenium; BCI, Rabat, Maroc) containing 1 mg/ml of selenium and 50 IU/ml of vitamin E to avoid white muscle disease, and they were vaccinated against enterotoxemia at weaning. At birth, the lambs were ear tagged and weighed. Subsequent weighings were taken every 2 weeks until weaning at 90 days. The weights at 30 days, at 60 days and at 90 days were calculated by linear interpolation. Sheep were shorn approximately 2 months before the mating season. The lambs were not docked and the ram lambs were kept intact. No selection was practised between generations for any trait, but the ewes were culled at approximately 6 years of age. Statistical analyses Traits studied The ewe traits studied were litter size at birth, litter weight at birth, litter size at 60 days per ewe lambing and litter weight at 60 days per ewe lambing. The litter weights were computed after the individual lamb weights at birth and at 60 days were corrected for sex, since the analysis showed that both of these weights were affected by the sex of the lamb. The lamb traits studied were weight at birth, at 30 days, at 60 days and at 90 days, as well as survival from birth to weaning. Analysis of variance Data were analysed by least‐squares mixed‐model procedures ( Harvey 1990). The model used to analyse litter traits included the effects of breed group of ewe (Sardi, D’man, F 1 , F 2 , F 3 and F 4 ), ewe within breed group of ewe (random effect), age of ewe at lambing (<2.5, 3.5, 4.5 and >4.5 years), period of lambing (1990, 1991, 1992, 1993, 1994 and 1995), and the interaction breed group × age of ewe. The other interactions were not significant (p > 0.05). For lamb weights, only those lambs that had performance on all weights studied were included in the analyses. The mixed model used to analyse lamb weights and survival included the effects of breed group of lamb (Sardi, D’man, F 1 , F 2 , F 3 , F 4 and F 5 ), sire within breed group (random effect), age of dam (<2.5, 3.5, 4.5 and >4.5 years), sex of lamb (male and female), type of birth and rearing (1–1, 2–2, 3–3, 2 or greater −1 and 3 or greater −2), and the period of birth of lamb (1990, 1991, 1992, 1993, 1994 and 1995). However, for birth weight, instead of the type of birth and rearing, only type of birth (singles, twins and triplets or greater) was considered. Due to the computer program limitations, two–way interactions that included the period of birth, were not tested, except for the interaction sex × period of birth. The two–way interactions tested that were not significant (p > 0.05) for lamb survival were deleted from the final analyses. For weights, interactions that were not significant for any trait were deleted from the final models. The mean squares for ewe within breed group of ewe and the mean squares for sire within breed group were used as the error term to test the significance of differences among breed groups ( F ‐test) for litter traits and preweaning lamb survival and weights, respectively. Estimation of heterosis and heterosis retained. If heterosis is primarily determined by dominance, then heterosis in advanced generations of crossbreeding programmes should be retained in the proportion of the retained heterozygosity. Since the Sardi and D’man breeds contributed equally to the synthetic breed, retention of initial (F 1 ) heterozygosity after crossing and subsequent random ( inter se ) mating within the crosses is equal to 0.5 (Dickerson 1969, 1973). This loss of heterozygosity occurs between the F 1 and F 2 generations, and, if inbreeding is avoided, further loss of heterozygosity in inter se mated populations does not occur (Dickerson 1969, 1973). Thus, the expectation for mean heterosis (individual heterosis H i , maternal heterosis H m and paternal heterosis H p ) was (1 H i  + 0 H m  + 0 H p ) for the F 1 generation (0.5 H i  + 1 H m  + 1 H p ) for the F 2 generation, and (0.5 H i  + 0.5 H m  + 0.5 H p ) for the F 3 and subsequent generations of the synthetic breed. However, since Boujenane & Bradford (1991) and Boujenane et al . (1991a,Boujenane & ;, –; Famula (1991b) showed that maternal and paternal heterosis were negligible for most traits from the crossbreeding between D’man and Sardi sheep, individual heterosis was estimated assuming H m  =  H p  = 0. Thus, linear functions of the means of parental breeds and crossbred generations were computed to estimate the heterosis. The heterosis was estimated by the difference between the performance of crossbred for the i th generation and the mean for the parental breeds. The mean heterosis was computed as the difference between the average performance of crossbred for different generations and the mean for the parental breeds. The heterosis retained was computed as the difference between one‐half of the initial (F 1 ) heterosis and the mean heterosis found for subsequent generations. Results and discussion Analysis of variance for litter traits Results for the analysis of variance and least squares breed group means for litter traits are presented in Tables 1 and 2. The effects of breed group were important for all litter traits studied. D’man ewes had the highest litter size at birth and at 60 days, Sardi ewes had the lowest performance, and the crossbred ewes were intermediate. For litter weight at birth and at 60 days, ewes of the F 1 generation had the heaviest litters and ewes of the F 4 generation had the lightest litters. The latter result is mainly due to the low individual weight at birth and at 60 days of lambs from F 4 dams. Except for litter size at birth, which was not affected by the age of ewe, all the other traits were influenced by both age of ewe and period of lambing. Ewes that were 2.5 years of age at lambing had the lowest performance, and those that were 4.5 years or older had the highest performance. These results are in agreement with those reported by Boujenane & Bradford (1991) and Boujenane et al . (1991b). On the other hand, the interaction age of ewe × period of lambing had a significant effect on litter weight at 60 days. 1 Analysis of variance for litter traitsLitter size at birthLitter size at
60 daysLitter wt at birth
(kg)Litter wt at 60 days 
(kg) Source of variationdfMSdfMSdfMSdfMSBreed group (B)516.3 ***57.6 ***57.7 ***5411.3 ***Ewes (B)7450.5 ***7450.4 *7451.8 *74544.6 **Age (A)30.831.936.7 **3344.7 ***Period of lambing50.9 *50.8 *538.6 ***5213.3 ***B × A––––––1477.6 *Residual6590.36590.36591.364537.2*p < 0.05; **p < 0.01; ***p < 0.0012 Least squares breed group means for litter traitsNumberLitter size at birthLitter size at
60 daysLitter wt at birth
(kg)Litter wt at 60 days 
(kg)Least squaresmeans (μ)14181.571.364.0216.6Breed groupSardi3081.10d1.04d3.87c15.1cD’man1551.90a1.52a4.18ab16.9bF 12431.68b1.48a4.32a19.0aF 22621.62bc1.44ab4.00bc17.2bF 31271.53bc1.35bc3.98bc16.8bF 43231.59c1.32c3.79c14.6ca–d Means within a column that does not have a common superscript differ (p < 0.05)Analysis of variance for lamb weights and survival Results for the analysis of variance and least squares breed group means for lamb weights and survival are presented in Tables 3 and 4. Except for lamb survival, which was not affected by the breed group of lamb and the period of birth, all weights and survival were significantly affected by the breed group of lamb, age of dam, sex of lamb, type of birth and rearing and period of birth of lamb. Sardi lambs had the highest weight at birth, at 30 days and at 60 days, lambs of the F 2 generation had the highest weight at 90 days, and lambs of the F 5 generation had the lowest weights at any age. This result may be explained by the fact that lambs of the F 5 generation were born from dams that were at their first or second lambing. In fact, lambs born from young dams had the lowest weights, and those produced by old dams had the heaviest weights. Also, ram lambs were heavier than ewe lambs at all ages. Single‐born lambs were consistently the heaviest at all ages, whereas those born as triplets or greater and raised as triplets or greater were the lightest. In general, lambs raised as twins excelled those raised as triplets or greater. Within the same type of rearing, the advantage was in favour of those born in small litters. These results are in agreement with those reported by Boujenane et al . (1991a). Moreover, breed group of lamb × sex, breed group of lamb × type of birth and rearing, and age of dam × type of birth and rearing interactions had significant effects (p < 0.05) for weight at 30 days and survival, whereas weight at birth and at 90 days was significantly (p < 0.05) affected by sex of lamb × period of birth of lamb interaction. 4 Least squares breed group means for lamb weights and survivalNumberBirth wt
(kg)30‐day wt
(kg)60‐day wt
(kg)90‐day wt 
(kg)Survival 0–90 daysLeast squaresmeans (μ)17022.646.8911.816.819960.82Breed groupSardi1603.31a7.96a12.7a17.6a1910.75D’man1592.42cd6.47de11.4c16.8bc1950.83F 12072.71b7.00bc12.0b17.0b2360.82F 23472.64b7.27b12.6a18.0a4020.87F 32902.48cd6.57d11.3cd16.3c3280.87F 43982.52c6.75cd11.6bc16.7bc4660.80F 51412.37d6.21e10.9d15.5d1780.79a–e Means within a column that does not have a common superscript differ (p < 0.05)Heterosis for litter traits Estimates of heterosis in crossbred ewes, resulting from crossing D’man and Sardi breeds, are presented in Table 5. The effects of heterosis were significant for all litter traits for F 1 generation, significant for litter size at birth (p < 0.05), litter size at 60 days (p < 0.01), and litter weight at 60 days (p < 0.05) for F 2 generation, and not significant (p > 0.05) for any trait for F 3 and F 4 generations. There was a tendency towards a decrease in heterosis for litter traits in advancing generations. On the other hand, mean heterosis effects for F 1 , F 2 , F 3 and F 4 generations were significant for all litter traits, except for litter weight at birth (p > 0.05). Percentage of mean heterosis was 6.7 (p < 0.01), 9.4 (p < 0.01), and 5.4% (p < 0.05) for litter size at birth, litter size at 60 days and litter weight at 60 days, respectively. These values are within the range reported (Nitter 1978). 5 Effects of heterosis on litter traitsItemLitter size at birthLitter size at
60 daysLitter wt at birth
(kg)Litter wt at 60 days 
(kg)Linear contrastsHeterosisF 1 minus purebreds0.18 ***0.20 *** 0.29 ** 2.96 ***F 2 minus purebreds0.12 *0.16 **−0.03 1.15 *F 3 minus purebreds0.030.06−0.05 0.80F 4 minus purebreds0.090.04−0.24−1.44F 1 , F 2 , F 3 & F 4 minuspurebreds0.10 **0.12 **−0.07 0.87 *Retained heterosis0.5HF 1 minus 0.33(HF 2 +HF 3 +HF 4 )0.010.01 0.25 *** 1.31 ***p < 0.05; **p < 0.01; ***p < 0.001Heterosis retained for litter size at birth and litter size at 60 days did not differ (p > 0.05) from expectation based on retained heterozygosity ( Table 5). However, for litter weight at 60 days, the retained heterosis was less (p < 0.01) than expectation based on retained heterozygosity. In a crossbreeding experiment including Romanov and Berrichon de Cher breeds, Ricordeau et al . (1982a) reported no heterosis for litter size at birth in F 1 nor any significant difference among the first four generations. Visscher (1987) reported that there was no apparent loss in reproductive traits of the synthetic ewes from a reciprocal cross between Finnsheep and Ile de France compared with F 1 ewes. Hight & Jury (1970) reported a decrease in most reproductive traits from the F 1 to the F 2 and from F 2 to the F 3 , such that the F 3 approached the performance of the least productive parental breed. Boylan (1985), analysing data from a crossbreeding experiment including Finnsheep and Suffolk breeds, as well as Finnsheep and Targhee breeds concluded that the recombination effects were large and negative for Finn–Suffolk crosses. Heterosis for lamb weights and survival The effects of heterosis for lamb survival and weights are presented in Table 6. The effects of heterosis were positive and not significant (p > 0.05) for lamb survival for each generation of the synthetic breed. Heterosis effects were negative and significant (p < 0.05) for birth weight for the F 2 , F 3 , F 4 and F 5 generations, negative and significant for weight at 30 days for the F 3 and F 5 generations, negative and significant for weight at 60 days and at 90 days for the F 5 generation. On the other hand, mean heterosis effects for the F 1 , F 2 , F 3 , F 4 and F 5 generations were negative and significant (p < 0.01) for birth weight, negative and not significant (p > 0.05) for the other weights, and positive and not significant (p > 0.05) for preweaning lamb survival. The percentage of heterosis for weight at birth was −11.2%. Rastogi et al . (1982) reported estimates of individual heterosis of 4.6% for birth weight but near zero for weaning weight, whereas Nitter (1978), reviewing heterosis for growth in sheep, reported average estimates of individual heterosis to be about 3.2 and 5% for birth and weaning weights, respectively. 6 Effects of heterosis on lamb weights and survivalItem
Birth wt
(kg)30‐day wt
(kg)60‐day wt
(kg)90‐day wt
(kg)Survival 
0–90 daysLinear contrastsHeterosisF 1 minus purebreds−0.16−0.22−0.01−0.22 0.02F 2 minus purebreds−0.22* 0.05 0.51 0.79 0.08F 3 minus purebreds−0.38***−0.64*−0.78−0.87 0.08F 4 minus purebreds−0.34***−0.46−0.42−0.48 0.00F 5 minus purebreds−0.50***−1.00**−1.15*−1.70** 0.00F 1 , F 2 . F 3 , F 4 & F 5minus purebreds−0.32**−0.45−0.37−0.49 0.04Retained heterosis0.5HF 1 minus 0.25(HF 2 +HF 3 +HF 4 +HF 5 ) 0.40*** 0.57* 0.59 0.64−0.04Heterosis retained for weight at birth and at 30 days was less than the expectation based on retained heterozygosity, and not significantly different (p > 0.05) from the expectation based on retained heterozygosity for the other weights and survival. These results are consistent with those of Ricordeau (1982b) who reported that there was no decline in lamb survival and weights from the F 1 to the F 4 generations of a crossbreeding between the Romanov and Berrichon de Cher breeds, and also with those of Rastogi et al . (1982) who reported small recombination effects for various growth traits from data involving the Columbia, Suffolk and Targhee breeds. However, during the development of the Romnelet breed, Peters et al . (1961) found reductions in birth weight and weaning weight from the first‐cross to the F 2 generation, and a further significant decline in birth weight from the F 2 to the F 7 generations, but an increase in weaning weight in the later generations. Similar results were reported by Hight & ; Jury (1970). General The results of the present study showed that the proportion of retained heterosis was not less than the proportion of retained heterozygosity for most traits. They are in agreement with the negligible epistatic recombination effects on litter traits, weights and survival from the F 1 and the F 2 generations D’man × Sardi crosses reported by Boujenane & Bradford (1991) and Boujenane et al . (1991a,b). Consequently, these results suggested that heterosis in sheep can be accounted for by the dominance effects of genes. The low performance of ewes for the F 4 generation and their lambs were mainly due to their younger age compared to ewes of the other breed groups. Therefore, the development of a synthetic breed from crosses among D’man and Sardi should be effective. Heterosis and heterosis retained for different generations of a crossbreeding between D’man and Sardi sheep were estimated from 1418 litter records of ewes, and 1702 lamb preweaning weights from 1996 lamb born. These data were collected from D’man, Sardi and their F 1 to F 5 lambs during six lambing seasons. Mean heterosis for the F 1 to F 4 generations were significant for all litter traits, except for litter weight at birth (p > 0.05). Heterosis retained for litter size at birth and at 60 days did not differ (p > 0.05) from the expectation based on retained heterozygosity, whereas it was less (p < 0.01) than the expectation based on retained heterozygosity for litter weight at 60 days. Mean heterosis effects for the F 1 to F 5 generations were negative and significant (p < 0.01) for birth weight, and not significant (p > 0.05) for weight at 30 days, at 60 days and at 90 days, as well as for lamb survival. For weight at birth, the heterosis retained was less than the expectation (p < 0.001) based on retained heterozygosity. It was concluded that for most traits the proportion of retained heterosis was not less than the proportion of retained heterozygosity, which suggested that heterosis in sheep can be accounted for by the dominance effects of genes.