HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Free Via Open Access Abstract Full Text (PDF) All Versions of this Article: jas.2008-1065v1 87/7/2196    most recent Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Google Scholar Articles by Cloete, S. W. P. Articles by Olivier, J. J. PubMed PubMed Citation Articles by Cloete, S. W. P. Articles by Olivier, J. J.

Genetic parameters and trends for lamb survival and birth weight in a Merino flock divergently selected for multiple rearing ability¹

S. W. P. Cloete^*,,2, I. Misztal and J. J. Olivier

^* Department of Animal Sciences, Stellenbosch University, Private Bag X1, Matieland 7602, South Africa; and Institute for Animal Production: Elsenburg, Private Bag X1, Elsenburg 7607, South Africa; and Department of Animal and Dairy Science, University of Georgia, Athens 30605; and Agricultural Research Council: Livestock Business Division, Private Bag X5013, Stellenbosch 7599, South Africa

	Abstract

Top Abstract INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
Data of 5,390 Merino lambs born from 1986 to 2007 (~6.9 generations) were used to derive genetic parameters and trends for age-specific and overall lamb survival on the underlying scale, as well as for lamb birth weight, using Gibbs sampling. The majority of lambs were descended from lines that were divergently selected for the ability of ewes to rear multiples. The line selected in the upward direction was denoted as the high line (H line), whereas the line selected in the downward direction was the low line (L line). Analyses included the covariance between direct and maternal genetic effects, except where it was not estimable owing to small direct additive variance components, a high incidence of lambs surviving, or both. Direct heritability estimates were 0.02 for lamb survival at birth, 0.12 for lamb survival from birth to tail docking, 0.39 for lamb survival from docking to weaning, 0.28 for overall lamb survival, and 0.17 for birth weight. Corresponding estimates for the maternal genetic effect were 0.26, 0.14, 0.16, 0.14, and 0.29. Dam permanent environmental variance ratios were, respectively, 0.14, 0.09, 0.05, 0.07, and 0.07. Estimates of the direct-maternal genetic correlation were –0.60 for lamb survival from docking to weaning, –0.61 for overall lamb survival, and –0.15 to –0.23 for birth weight. Genetic, maternal genetic, and dam permanent environmental correlations between lamb birth weight and overall or age-specific lamb survival were not different from zero. Expressed relative to overall means, annual direct genetic change in the H line amounted to –0.01% per annum for lamb survival at birth, 0.52% per annum for lamb survival from birth to docking, and 1.3% per annum for lamb survival from docking to weaning. Corresponding trends in the L line were, respectively, –0.01, –0.42, and 0.042% per annum. Maternal genetic trends amounted to, respectively, 0.23, –0.11, and –0.24% per annum in the H line and –0.78, –0.50, and –0.16% per annum in the L line. It was concluded that sustained genetic progress in lamb survival is feasible if directed selection is applied to a correlated trait such as the ability of ewes to rear multiples.

Key Words: correlation • heritability • lamb survival • linear-threshold model • sheep

	INTRODUCTION

Top Abstract INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
Decreased and variable lamb survival is generally considered a major constraint to efficient sheep production (Haughey, 1991). A core level of lamb losses remains, even if managerial inputs are optimized (Alexander, 1984). Apart from the direct loss of production and income to farmers, lamb losses have also become an emotional issue involving animal welfare.

Studies involving the possible genetic improvement of lamb survival have found limited genetic variation (Olivier et al., 1998; Snyman et al., 1998; Lopez-Villalobos and Garrick, 1999). The lack of genetic variation has led to recommendations that the improvement of lamb survival should rather be based on modification of the environment to create conditions suitable for survival of lambs (Morris et al., 2000; Everett-Hincks et al., 2005). In contrast, some previous studies suggested differences between lines within breeds that were brought about by selective breeding (Haughey, 1983; Knight et al., 1988). Gudex et al. (2005) more recently reported significant variation in survival of the progeny of respective sires. However, Gudex et al. (2005) also suggested that it would be fruitless to study lamb mortality without considering lamb birth weight. The latter sentiments were shared by Morris et al. (2000).

Previous studies at the Institute for Animal Production: Elsenburg have shown that the mortality of Merino lambs born in a divergent selection experiment for the ability of ewes to rear multiples per mating differed (Cloete and Scholtz, 1998; Cloete et al., 2005). The objectives of this study were to derive genetic parameters (direct and maternal heritability and genetic correlations) for age-specific and overall lamb survival, as well as for lamb birth weight; and to obtain direct and maternal genetic trends resulting from divergent selection for the ability of ewes to rear multiples, or alternatively defined as number of lambs weaned per mating.

	MATERIALS AND METHODS

Top Abstract INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
The bulk of the data recording for the study was done before the formation of a formal Departmental Ethical Committee for Research on Animals in 2004. Based on information supplied by the scientists involved and the flock management staff, as well as on procedures witnessed during routine flock visits at lambing, post facto ethical approval was granted for the study.

Animals, Selection Procedures, and Locations

Two lines of Merino sheep were divergently selected from the same base population from 1986 to 2007, using maternal ranking values for number of lambs reared per mating (i.e., per year), as described by Cloete et al. (2004). The trait could also be defined as number of lambs weaned per mating (also referred to as net reproduction rate), but it has been termed multiple rearing ability previously (Cloete et al., 2004). The lines were derived from ewes descended from a Merino line selected for an increased wool secondary:primary follicle ratio (Heydenrych et al., 1984).

Ewe and ram progeny of ewes rearing more than 1 lamb per mating (i.e., reared twins at least once) were preferred as replacements in the high (H) line. Replacements in the low (L) line were preferably descended from ewes rearing less than 1 lamb per mating (i.e., barren or lost all lambs born at least once). Depending on the average reproduction of the lines and the replacement needs, progeny of ewes that reared 1 lamb per mating were occasionally accepted in both lines. Initially, when the average number of lambs weaned per mating in the H line was <1, such ewes had to be included to ensure an acceptable replacement rate. At present, selection in the H line can be strict, and all replacements in the H line have increased breeding values and are descended from ewes that reared >1 lamb per mating. In contrast, very few female progeny are often available in the L line at present, and progeny from ewes rearing 1 lamb per mating often have to be selected. Selection decisions were mostly based on 3 maternal matings in the case of rams. Owing to the greater replacement needs in females, progeny from ewes with fewer records were also selected as replacements. In the progeny groups up to 2002, only the maternal phenotype was considered during the selection of an individual, and no additional information was used. Once selected, ewes remained in the breeding flock for at least 5 matings, except for cases of death and teeth or udder malfunction. No selection on reproduction was therefore directed at the current flock. From the progeny group born in 2003, information on maternal ranking values used for selection was augmented by breeding values for number of lambs weaned per ewe joined. These breeding values were derived from a single-trait repeatability model, as described by Cloete et al. (2004).

At the onset of the experiment, approximately 120 breeding ewes represented each line. The H line was gradually allowed to increase to 130 to 140 breeding ewes, whereas the relatively poor reproduction in the L line resulted in breeding ewe numbers that decreased to 40 to 80. The H line was augmented by 28 ewes derived from a multiple ovulation and embryo transfer program. These ewes were born during 1991 and 1992 (detailed by Cloete et al., 2004). During the 1998 to 2002 lambing years, 6.5-yr-old dams with 5 lambing opportunities were screened into both the H and L lines from other selection lines maintained at the Tygerhoek experimental farm (Cloete et al., 1998) to augment numbers in the L line in particular. The number of ewes introduced in this way ranged from 8 to 18 per year, and totalled 74 (Cloete et al., 2004). Ewes rearing 7+ lambs from 5 opportunities and rearing at least 1 lamb per opportunity were screened into the H line (n = 28). Ewes rearing 1 to 3 lambs from 5 opportunities were selected in the L line (n = 46). Ewes rearing 4 to 6 lambs were not considered, as only those with extremely good or extremely poor performance in terms of reproduction were envisaged to contribute to the experiment.

During the first 3 yr of the current experiment, rams were selected on maternal performance from a line selected for clean fleece weight and the control line maintained alongside. The study on selection for clean fleece weight is well documented in the literature (Heydenrych et al., 1984; Cloete et al., 1998). Initially, 5 rams represented each line. Until 1992, rams were used for 1 breeding season only. During subsequent years, 1 to 2 rams in each line were carried over to the next year to provide sire links across years. From the mid 1990s, the number of rams used in the H line was increased to 4 to 6, whereas only 2 to 4 rams were used in the L line. This arrangement was needed to ensure that the number of ewes joined to each ram remained fairly constant. Ram replacements were selected to represent all of the sires present in their progeny group wherever possible. In rare cases, when no suitable replacement for a specific ram was available, a candidate sire descended from another sire group was selected.

Since 2003, part of the breeding flock was subjected to reciprocal crossbreeding between the 2 lines, with the intention of forming a genetic resource population for possible future genomic projects (Naidoo et al., 2005). Crossbred progeny were available from the 2003 year of birth onwards, whereas backcrosses became available from 2005. When the data used for the present study were sourced after the 2007 year of birth, purebred H and L line lambs were still the vast majority of the available data, with a total of 5,024 out of 5,390 lambs born in the present experiment (i.e., ~93% of the data).

The resource flock studied was maintained at the Tygerhoek experimental farm at first (1986 to 1992), and at Elsenburg subsequently. The climate, pastures grown, and managerial practices at the respective sites are adequately described by Cloete et al. (2004). The management of the experimental animals was similar on both farms, except for the timing of lambing and shearing. At Tygerhoek, ewes were hand-mated in spring to lamb during autumn (March to April) of the following year. At Elsenburg, mating took place in summer, for lambing during the winter (June to July) of the same year. At Tygerhoek, ewes were shorn annually during August to September, before mating. At Elsenburg ewes were shorn before lambing each year. Ewes thus lambed with fewer than 6 wk of wool growth. Lambing and reproduction practices at both sites were described by Cloete and Scholtz (1998) and Cloete et al. (2004).

Recordings

Data were recorded for 5,390 lambs born during the period from 1986 to 2007. These lambs were the progeny of 181 sires and 1,278 dams contained in the pedigree file. The data also contained 469 base parents that entered the breeding flock without pedigree information. An average generation interval of 3.2 yr (2.3 yr for rams and 4.1 yr for ewes) was derived, suggesting that the experiment spanned ~6.9 generations. Information recorded on individuals included year of birth, sex, dam age, and birth type. Ninety-nine triplets were pooled with twins and were denoted as multiples.

Individual lamb birth weights were obtained within 24 h of birth to the nearest 0.1 kg, when lambs were identified with their dams to ensure complete pedigree information (Cloete et al., 2004). Four birth weights of <1 kg recorded at the birth of mummified lambs being born in multiple litters and dying before birth were excluded. These lambs were fairly uncommon and constituted 0.07% of the overall number of lambs born.

Age-specific lamb survival was categorized as follows: survival at birth (lambs succumbing before or during parturition were regarded as deaths at this stage), survival from birth to tail docking at ~4 wk, and survival from tail docking to weaning (Table 1). Proportions of 0.957 (5,159 out of 5,390) lambs survived birth, 0.888 (4,579 out of 5,159) lambs survived from birth to tail docking and 0.888 (4,060 out of 4,579) survived from tail docking to weaning. Group specific proportions of lambs that survived were 0.964, 0.895, and 0.899 in the H line and 0.943, 0.867, and 0.844 in the L line, respectively. Overall lamb survival was also considered, lambs dying from before birth until weaning, which coincided with the recording of 100-d BW at an average (SD) age of 99 (14) d, being recorded as dead. Lambs present at weaning, with a 100-d BW, were considered to be alive. A proportion of 0.753 of the lambs that were born survived (4,060 out of 5,390). In the H line, 2,624 out of 3,384 lambs survived (a proportion of 0.776). In contrast, 1,133 out of 1,642 L line lambs survived (a proportion of 0.690). A proportion of 0.598 (2,024 out of 3,384) of H line lambs were born as multiples. Only 726 out of 1,642 L line lambs were born as multiples, constituting a proportion of 0.442.

A 2-trait animal model was fitted with age-specific or overall lamb survival and birth weight as the dependent variables. For analysis, lamb survival was defined as a binary trait with 2 categories (1 for lambs dying in the particular category and 2 for those surviving), whereas birth weight data were treated as continuous. Several preliminary analyses were conducted for age-specific or overall lamb survival. These included selection line concatenated with year of birth and birth type as fixed effects, whereas a reduced data set excluding crosses between the H and L lines was also considered. However, these adaptations seemed to have a small influence on the derived parameter estimates, and the full data set was analyzed without including selection line as fixed. Data are not shown for these preliminary analyses, but key outcomes will be provided where pertinent to the reasoning followed. The fixed effects included year of birth (1986 to 2007), sex (male and female), dam age (2 to 7+ yr), and birth type (single and multiple). The equation for the 2-trait model was the following:

In this model, y was a vector of observations for underlying values for age-specific or overall lamb survival and lamb birth weight; i was indicative of the respective traits (I = 1,2); f_ij was the fixed effect j for the ith trait; a_ik was the additive genetic effect of the kth animal for the ith trait; m_ik was the maternal genetic effect of the kth animal for the ith trait; c_il was the dam permanent environmental effect of the lth dam for the ith trait; and e_ijklm was the vector of randomly distributed residual effects. The genetic correlation between direct and maternal effects was considered for the first run on all 4 data sets (survival at birth with birth weight, survival from birth to tail docking with birth weight, survival from tail docking to weaning with birth weight, as well as overall survival with birth weight). The direct-maternal covariance for survival at birth and survival from birth to tail docking was excluded for the final run on the joint analyses on these traits and birth weight, as pointed out later (see Posterior Distributions in the Results section). However, the direct additive variance for survival at birth was retained as a default in these analyses.

The software used was THRGIBBSF90 (Misztal et al., 2002). This software is suitable for estimation of variance components and genetic parameters in threshold animal mixed models for any combination of categorical and continuous traits (Lee et al., 2002). The program POSTGIBBSF90 was used for post-Gibbs analysis (Misztal et al., 2002). The software allows for prediction of solutions for fixed and random effects, as described later.

In all cases, a single chain of 200,000 cycles was run, with the first 50,000 cycles used as the burn-in period (Donoghue et al., 2004). When the sampled values were plotted against the iterations, a stationary stage could be confirmed at this stage by graphical inspection. Every 10th sample was stored after 50,000 iterations, giving a total of 15,000 samples for the computation of posterior means and posterior SD, as well as 95% highest posterior density (HPD) confidence intervals. Based on results from the analysis, 90% HPD confidence intervals were also computed for the (co)variance components when applicable. Point estimates were calculated as the posterior mean of the specific variance component, using the results from the final 15,000 samples as set out above. Direct genetic, maternal genetic, dam permanent environmental, and environmental (residual) correlations were derived from these analyses.

Genetic Trends

Solutions for all fixed and random effects were based on the average of the last 15,000 samples, involving the 2-trait model described above for the respective analyses. Animal solutions pertaining to direct and maternal breeding values for both traits were obtained from the post-Gibbs analysis. Solutions for lamb survival were on the underlying scale. Individual breeding values on the underlying scale were regressed on year of birth, with 1986 taken as the base year for the analyses. All regressions were forced through the origin. Genetic trends were tested for divergence between the lines, using SE obtained for the regression coefficients. The absolute values of the regression coefficients were also compared between the H and L line, to test for possible asymmetric responses in the upward or downward directions.

	RESULTS

Top Abstract INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
Descriptive Statistics

Age-specific lamb survival on the underlying scale averaged 1.96 for survival at birth, 1.89 for survival from birth to tail docking, and 1.89 for survival from tail docking to weaning (Table 1). Overall lamb survival on the underlying scale averaged 1.75. Survival at birth and overall survival were also assessed in those lambs that died before parturition. Lamb birth weight averaged 3.8 kg, with a range between 1.0 and 7.0 kg. Apart from the BW of the mummified fetuses mentioned previously, birth weight was not recorded on a small number of lambs mutilated by vermin, hence the slightly decreased number of lambs recorded for birth weight in comparison with survival at birth and overall lamb survival. All traits were variable, with CV exceeding 10%.

Posterior Distributions

Posterior distributions for the genetic components (direct and maternal variances, the direct-maternal covariance, as well as additive covariance and the maternal covariance between traits) for overall lamb survival and birth weight are provided in Figure 1. The difference in the skewed distributions for lamb survival (a threshold trait) and the largely symmetric distributions for birth weight (a continuous trait) is evident. The direct-maternal covariance components were not symmetric about zero, and most values were negative. The 95% HPD confidence interval just included zero for both traits, the derived parameters obviously approaching significance. Further appraisal of the data indicated that the 90% HPD confidence interval excluded zero for both traits (–0.381 to –0.005 for overall lamb survival on the underlying scale, and –0.050 to –0.001 for birth weight).

View larger version (32K): [in this window] [in a new window]   Figure 1. Posterior density distributions for the genetic variance components pertaining to overall lamb survival (panel a), birth weight (panel b), and the genetic covariance components between traits (panel c). Direct and maternal variances, the direct-maternal covariance components (RAM) for both lamb survival on the underlying scale and birth weight, as well as the direct and maternal covariance components between lamb survival and birth weight, are depicted.

Genetic Parameters

The estimate of h² for survival at birth on the underlying scale was small and not significant at 0.02 ± 0.05 (P > 0.10; Table 2). In contrast, both maternal components were significant at 0.26 for m² and 0.14 for c². The direct additive variance component for survival from birth to tail docking (Table 3) on the underlying scale approached significance (P < 0.10), with 90% HPD confidence intervals of 0.025 and 0.350. Estimates of m² and c² were greater than double the corresponding SE at 0.14 and 0.09, respectively. Derived 95% HPD confidence intervals for variance ratios for lamb survival from tail docking to weaning included zero in all instances (Table 4). However, 90% HPD confidence intervals excluded zero in all cases, being 0.007 and 1.415 for additive effects, 0.007 and 0.565 for maternal genetic effects, and 0.002 and 0.169 for dam permanent environmental effects. The absolute value of the estimate of h² was fairly high, at 0.39. Comparatively smaller maternal variance ratios were found, namely 0.16 for m² and 0.05 for c². The direct-maternal correlation was high at –0.60, but the 90% HPD confidence intervals for the direct-maternal covariance component included zero and were –0.574 and 0.024, respectively.

Genetic parameters for birth weight were generally similar in the analyses reported in Tables 3 and 4. However, the joint analysis with survival at birth on the underlying scale differed somewhat. In this analysis, h² was similar to the other estimates, but a proportionally greater part of the maternal variance was partitioned toward the maternal genetic variance component, at the expense of the dam permanent environmental variance. When the results pertaining to birth weight were considered, h² amounted to 0.16 to 0.17, m² to 0.28 to 0.37, c² to 0.04 to 0.08, and the direct-maternal correlation to –0.15 to –0.23 (Tables 2 to 4).

Outcomes from analyses on overall lamb survival and birth weight involving the reduced data set and analyses including or excluding the selection line by year concatenation, as well as the selection line by birth type concatenation, in the models were in close correspondence (data not shown). All variance ratios (h², m², and c²) differed by <0.05 between analyses when equivalent random effects models were fitted. For example, (co) variance ratios (±SE) for lamb survival from tail docking to weaning were 0.41 ± 0.40 for h², 0.17 ± 0.15 for m², 0.05 ± 0.03 for c², and –0.60 ± 0.74 for the direct-maternal correlation when selection line was included in the model. Corresponding estimates for birth weight were, respectively, 0.15 ± 0.03 for h², 0.28 ± 0.05 for m², 0.08 ± 0.03 for c², and –0.18 ± 0.15 for the direct-maternal correlation.

Genetic correlations of age-specific lamb survival with birth weight ranged from 0.04 to –0.23 and were not significant (P > 0.20; Tables 2 to 4). Corresponding maternal genetic and dam permanent environmental correlations ranged from 0.08 to 0.34 and were positive in absolute terms, but not significant (P > 0.20). Environmental correlations of birth weight with age-specific lamb mortality were positive in sign and significant (P < 0.05). This result seems to suggest that sheep in an environment sustaining increased birth weight should also have improved age-specific lamb survival rate.

Genetic Trends

Given that h² for survival at birth on the underlying scale was very small (Table 2), it is not surprising that direct genetic trends were almost negligible compared with the other genetic trends involving lamb survival. When expressed relative to the overall mean, breeding values declined at 0.01% per annum in the H line, and at 0.01% per annum in the L line (P < 0.01 in both instances). Maternal breeding values increased by 0.23% per annum for survival at birth on the underlying scale in the H line, whereas the maternal genetic trend amounted to –0.78% per annum in the L line. When the SE of the regression coefficients were considered, it was evident that the responses were asymmetric, the upward response in the H line being smaller in magnitude than the downward response in the L line.

When direct responses in lamb survival from birth to tail docking on the underlying scale in Table 5 were expressed relative to the overall mean, they amounted to 0.52% per annum in the H line and to –0.42% per annum in the L line. When the absolute values of these regression coefficients were considered, it was shown that the upward trend in the H line was somewhat faster than the downward trend in the L line (P < 0.05). The corresponding maternal genetic trends were negative in both instances, amounting to –0.11 and –0.50% per annum in the H line and the L line, respectively. In this instance, the downward trend in the H line was slower (P < 0.05) compared with that in the L line.

Preliminary analyses involved the inclusion of the selection line x year and the selection line x birth type concatenations as systemic effects. This adaptation resulted in direct genetic trends for age-specific lamb survival on the underlying scale in the L line being elevated to roughly the same levels observed in the H line (data not shown).

All trends involving birth weight were small in magnitude, and genetic change amounted to <0.12% of the overall mean per year when regressions were significant (P < 0.01). The fastest rate of response was recorded in the maternal genetic trend for birth weight in the H line, where maternal breeding values increased at 0.11% per annum when expressed relative to the overall mean (Table 5). The downward direct genetic trend for birth weight in the L line amounted to 0.04% per annum.

	DISCUSSION

Top Abstract INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
Descriptive Statistics

Sawalha et al. (2007) reported that a proportion of 0.947 of Scottish Blackface lambs survived the period during or shortly after birth. Riggio et al. (2008) accordingly found a proportion of 0.08 lambs that were dead at birth. Survival at birth was slightly better in the present study, with a proportion of approximately 0.96 lambs surviving birth. Postnatal survival was increased in Sawalha et al. (2007), namely a proportion of 0.979 surviving from 1 to 14 d and 0.963 surviving from 15 d to 120 d of age. Cumulative survival in Riggio et al. (2008) was 0.868 after 4 wk and 0.846 after 12 wk. Postnatal survival was less in the present study.

The frequency of lamb deaths from birth to weaning ranged from 20.0 to 26.3% in the study of Gama et al. (1991). Overall lamb survival averaged from 0.776 to 0.801 in Morris et al. (2000). Means from Matos et al. (2000) were 0.82 for the Rambouillet and 0.75 for Finnsheep, with respective litter sizes of 1.47 and 2.49. Means for preweaning survival ranged from 0.78 to 0.86 (Lopez-Villalobos et al., 1999; Matos et al., 2000; Morris et al., 2000; Cloete et al., 2001; Matika et al., 2003). The average overall lamb survival in Welsh et al. (2006) was 0.81. Results from the present study (~75% of the lambs that were born survived to weaning) are in the lower range of these values. Coefficients of variation for lamb survival range from 0.48 to 0.57 in the literature (Matos et al., 2000; Safari and Fogarty, 2003; Safari et al., 2005).

Lamb birth weight averaged 3.8 kg in the present study. This mean is less than birth weight means of 4.2 to 4.3 kg reported for non-Merino breeds (Yazdi et al., 1999; Morris et al., 2000). However, it is in accordance with the means of 3.4 to 3.8 kg based on a review of the literature for Merino sheep by Safari and Fogarty (2003). At 3.85 kg, the previous mean reported for a smaller data set of the same resource population (Cloete et al., 2003a) was also in agreement with the present study. The variability of the birth weight data used in the present study seemed to be in the upper range relative to the literature, where CV ranging from 10 to 21% were reported (Yazdi et al., 1999; Morris et al., 2000; Safari and Fogarty, 2003; Safari et al., 2007). Literature averages for CV of birth weight derived by Safari et al. (2005) amounted to 14.2% for 4 wool sheep studies, 16.5% for 23 studies on dual-purpose breeds, and 19.2% for 6 meat sheep studies.

Genetic Parameters

When age-specific lamb survival was considered, relatively few estimates were found in the literature to compare with the present estimates of 0.02 for h², 0.26 for m², and 0.14 for c² for survival at birth. Sawalha et al. (2007) analyzed lamb viability (defined as survival at birth and up to 24 h), and obtained estimates of 0.05 for h² and 0.10 for m², with a litter effect of 0.19. The direct-maternal genetic correlation was estimated at 0.31, but with an SE larger than the estimate. Morris et al. (2000) and Riggio et al. (2008) termed an identical trait as perinatal survival. Using the logit link function, Morris et al. (2000) estimated h² at 0.01 to 0.02 in the Tokanui and Woodlands populations and 0.02 to 0.03 in the Rotomahana population. These estimates closely corresponded to those obtained for survival at birth in the present study. Corresponding estimates for m² ranged from 0.02 to 0.03 in the Tokanui and Woodlands populations and were constant at ~0.11 in the Rotomahana population. The latter estimates suggest that maternal effects are more important than direct effects at this stage, as was also seen in the present study and in Sawalha et al. (2007). Riggio et al. (2008) estimated h² at 0.33, using a sire model and the probit link function. These results suggest that survival at birth (and the period shortly after birth) could be determined genetically, as suggested by the m² estimate of 0.26 for survival at birth on the underlying scale in the present study. Support is also provided by Smith (1977), who suggested that dystocia may be improved by selective breeding.

Present estimates for survival from birth to tail docking were 0.12 for h², 0.14 for m², and 0.09 for c². Comparable estimates in the literature for lamb survival from 1 to 14 d of age were 0.13 for h² and 0.14 for m², with 0.25 for the litter effect (Sawalha et al., 2006). Total heritability (h²_T) can be calculated as h_T² = (_a² + 0.5_m²)/_p² in the absence of a significant direct-maternal covariance component (Willham, 1972). Applying this formula to the present data, h²_T was estimated at 0.19, which is in good agreement with a sire model h² estimate of 0.17 reported by Riggio et al. (2008).

Present estimates for lamb survival from tail docking to weaning were 0.39 for h², 0.16 for m², and 0.05 for c², with a direct-maternal correlation of –0.60. At close to 0.40, the h² estimate for survival from tail docking to weaning on the underlying scale was substantially greater than an estimate of 0.05 reported by Sawalha et al. (2006) for lamb survival from 14 d to weaning. However, estimates of the latter authors for m² (0.10) and the direct-maternal correlation (–0.91) were in closer correspondence. It therefore seems that useful genetic variation occurs in age-specific lamb survival. These estimates lend support to the significant genetic trends for the H and the L lines.

When overall lamb survival was considered in the present study, parameter estimates were 0.28 for h², 0.14 for m², 0.07 for c², and –0.61 for the direct-maternal correlation. These estimates are considerably greater than the bulk of comparable literature values. Sixteen literature estimates for h² of overall lamb survival averaged 0.03 (Safari et al., 2005). The corresponding mean for 24 estimates of m² averaged 0.04. Accordingly, Barwick et al. (1990) reported estimates of 0.03 for h², 0.02 for m², and 0.10 for c² for overall lamb survival for Suffolk sheep in the United States. Many of the previous estimates were obtained using linear models, which are not really suitable for the analysis of binomial lamb survival data. Link functions, like logit and probit, were commonly used to link the normal distribution to binomially distributed lamb survival data. However, this approach only marginally improved h² estimates for preweaning lamb survival in many cases. The direct h² estimate for overall lamb survival was 0.01 in Romney lambs when the logit link function was used (Lopez-Villalobos and Garrick, 1999). Corresponding h² estimates reported by Morris et al. (2000) on Romney sheep ranged from 0 to 0.02 using the logit-transformation. Estimates of m² amounted to 0.04 in Romney lambs (Lopez-Villalobos and Garrick, 1999), whereas estimates of 0.09 to 0.10 were reported by Morris et al. (2000) for logit-transformed data. Comparatively greater estimates of 0.14 for h² and 0.11 for m² were reported for overall lamb survival of Coopworth lambs by Everett-Hincks et al. (2005), using the logit-transformation.

The near-zero estimates of additive genetic variation for lamb survival have resulted in recommendations that this trait is unlikely to respond to selection, and that it could be improved by management (Olivier et al., 1998; Morris et al., 2000; Everett-Hincks et al., 2005). However, there is mounting evidence that some earlier studies underestimated the true genetic variation that is available for the improvement of lamb survival on the underlying scale. Matos et al. (2000) used linear and threshold models to estimate the h² for overall lamb survival in less fecund Rambouillet and very prolific Finnsheep. Linear models yielded h² and m² estimates of 0.03 and 0.03, respectively in Rambouillet lambs and 0.09 and 0.19 in Finnsheep lambs, respectively. Threshold models yielded corresponding estimates that were greater: 0.06 and 0.04 for Rambouillet lambs and 0.17 and 0.26 for Finnsheep lambs, respectively. Welsh et al. (2006), using a threshold model, reported estimates of 0.11, 0.08, and 0.10 for h², m², and c², respectively for overall survival of New Zealand Romney lambs. Casellas et al. (2007) reported an h² estimate of 0.14 for overall survival of Ripollesa lambs using Gibbs sampling, with 95% HPD confidence intervals of 0.057 and 0.226. The corresponding value for c² was 0.09 with 95% HPD confidence intervals of 0.05 and 0.14. These estimates are broadly consistent with parameters reported in the text for overall lamb survival and for age-specific lamb survival in the present study. It can be deduced from the literature that threshold animal model analyses detected mentionable levels of additive and maternal genetic variation in lamb survival. It follows that overall lamb survival is heritable and variable and may therefore respond to direct selection. This viewpoint is also supported by recent molecular research, suggesting that polymorphisms at the ovine β₃-adrenergic receptor locus were associated with lamb mortality in New Zealand flocks (Forrest et al., 2007).

Direct-maternal genetic correlations for lamb survival ranged from –0.03 to –0.34 (Lopez-Villalobos and Garrick, 1999; Morris et al., 2000). However, a highly unfavorable direct-maternal correlation of –0.74 was reported in Everett-Hincks et al. (2005). They also suggested that genes that regulate physiological and biochemical processes for survival may interact with those genes that facilitate ewe-lamb bonding, resulting in this unfavorable correlation. A similar estimate of –0.75 was reported by Welsh et al. (2006), based on a threshold model analysis of New Zealand Romney data. The present direct-maternal genetic correlation for survival from docking to weaning (–0.60) and for overall lamb survival (–0.61) was consistent with this range of values.

Our estimates of 0.16 to 0.17 for h², 0.28 to 0.37 for m², and 0.04 to 0.08 for c² for birth weight were comparable with those of a previous REML analysis (excluding the direct-maternal covariance) on a smaller data set of 4,235 records from the same resource population (0.12 for h², 0.23 for m², and 0.08 for c²; Cloete et al., 2003a). In comparison, averaged genetic parameters for birth weight of wool sheep were 0.21 for h², 0.21 for m², and 0.10 for c², as derived from 5 to 6 literature estimates in the review of Safari et al. (2005). Corresponding literature values were 0.19, 0.18, and 0.09 for dual-purpose sheep (21 estimates), respectively, and 0.15, 0.24, and 0.19 for 5 to 6 estimates involving meat sheep, respectively. Estimates for Suffolk sheep in the United States were 0.08 for h², 0.21 for m², and 0.09 for c² in the study of Barwick et al. (1990), whereas corresponding estimates of 0.15, 0.18, and 0.16 were reported for Scottish Blackface sheep (Riggio et al., 2008). Genetic parameters for birth weight are relatively scarce for Merino lambs. In this breed, previous estimates of h² ranged from 0.05 to 0.23 (Mortimer and Atkins, 1995; Analla and Serradilla, 1998; Duguma et al., 2002; Safari et al., 2007). Corresponding estimates of m² ranged from 0.14 to 0.29, and those for c² from 0.07 to 0.12. The correspondence of the present estimates with those cited above is good.

The direct-maternal correlation for birth weight of lambs ranged from –0.15 to –0.26 in the present study. These estimates are consistent with comparable estimates ranging from –0.16 to –0.73 in the literature (Barwick et al., 1990; Analla and Serradilla, 1998; Duguma et al., 2002; Safari et al., 2007).

No genetic correlations between birth weight and lamb survival were found in the comprehensive set of genetic parameters cited from the literature by Safari and Fogarty (2003). The summary of these data also did not make provision for correlations between birth weight and overall lamb survival (Safari et al., 2005). Previous estimates that were identified in the literature were those of Barwick et al. (1990) and Riggio et al. (2008). A small direct correlation of –0.02 was reported by Barwick et al. (1990), whereas the maternal genetic correlation went beyond the parameter space at 1.24. The corresponding environmental correlation was 0.26. The latter correlation was consistent with that found in the present study. Riggio et al. (2008) reported genetic correlations of 0.24 and 0.45, respectively, for perinatal survival and survival to 4 wk with birth weight. These values are not directly comparable with the direct genetic correlation of –0.20 and the maternal genetic correlation of 0.28 obtained from the present study. The phenotypic correlation of perinatal survival with birth weight amounted to 0.17 in the study of Riggio et al. (2008). Problems with fitting threshold-linear covariance components doubtlessly hampered research and development in this respect. However, the feasibility of estimating linear-threshold genetic parameters by the usage of Gibbs sampling was demonstrated by several authors, including Arango et al. (2005). Lamb viability (coded as 0 for survivors and 1 for dead animals) was related to birth weight by Sawalha et al. (2007). The derived genetic correlation was unfavorable at 0.21, whereas the environmental correlation was favorable at –0.25. The sign and magnitude of these correlations were consistent with those reported in the present study (–0.20 and 0.28, keeping in mind that the sign of the correlations has to be reversed because of different coding). Reasoning pertaining to this relationship is complicated by the fact that a well-established nonlinear relationship exists between lamb survival and birth weight (Sawalha et al., 2007). Both very light and very heavy lambs are at risk of dying before weaning, whereas lambs with an intermediate birth weight have a better chance of survival. Small lambs are likely to succumb to hypothermia and starvation, whereas heavy lambs are at a bigger risk of dying because of dystocia. Morris et al. (2000) reported that lambs weighing 0.4 to 1.2 SD below the mean birth weight were particularly at risk of dying before weaning. Knight et al. (1988) accordingly reported that improved lamb survival in the Marshall Romney ewe line was associated with an increase in birth weight relative to dam weight.

Genetic Trends

To our knowledge, the genetic progress made in direct and maternal breeding values for age-specific lamb survival on the underlying scale in the H line is the first evidence of sustained progress for lamb survival in a designed experiment. It should be considered that this genetic change was a correlated response to selection for the ability of ewes to bear and rear multiple lambs, as detailed by Cloete et al. (2004). Haughey (1983) demonstrated substantial differences in lamb survival of the progeny of lambed ewes that were screened from a larger population to have a good or a poor rearing record. Donnelly (1982) and Knight et al. (1988) accordingly reported differences in lamb survival between populations where selection was directed at lamb survival or ewe rearing ability. These results coincide with sire-line variation in the occurrence of deaths associated with starvation-exposure, as detected in 5 on-farm trials in New-Zealand (Gudex et al., 2005). Back-transformation to the phenotypic probability scale suggested that the lambs produced by the best Romney sire had a 14.2% better chance of survival compared with progeny of the poorest sire (Welsh et al., 2006). These reports contradict conventional wisdom and indicate that selection for improved lamb survival is possible.

However, none of the previous studies were designed to demonstrate sustained genetic improvement. When survival at birth is considered, Haughey (1983) and Knight et al. (1988) attributed improved overall lamb survival to fewer parturient deaths in their studies. The divergent maternal genetic trends for the H and L lines in the present study support the reasoning that a smaller likelihood of difficult births may contribute to improved overall lamb survival. It is also worthy of consideration that length of parturition showed significant maternal genetic variation (Cloete et al., 2002, 2006) and that length of parturition (analyzed as a trait of the ewe) was demonstrated to become shorter in the H line and longer in the L line at a genetic level (Cloete et al., 2003b). The asymmetric maternal genetic response in survival at birth (smaller in the H line) could possibly be related to the fact that the H line is fairly near to the boundary of 100% survival at birth; less than 3% of H line lambs were dead at birth in Cloete and Scholtz (1998).

Genetic responses in direct breeding values for early (birth to tail docking) and late (tail docking to weaning) postnatal survival were quite substantial in the H line. Early postnatal lamb survival in the L line was compromised both at the direct and maternal levels. These trends could be related to behavioral adaptations conducive to lamb survival in the H line, when compared with the L line (Cloete and Scholtz, 1998; Cloete et al., 2003b, 2005). Breeding values for maternal cooperation with the first suckling attempts of the neonate and the period that recently lambed ewes stayed on or near the birth site were demonstrated to improve in the H line, with opposite trends in the L line (Cloete et al., 2003b). Ewes in H line were less likely to circle and back away from the first suckling attempts of the neonate than L line ewes (Cloete and Scholtz, 1998). Lambs in the H line were also quicker to progress from standing to suckling, even after accounting for the better cooperation of H line ewes (Cloete and Scholtz, 1998). When 1-d-old lambs were tethered 20 m away from the dam, H line ewes were quicker to reach the tethered lamb, whereas H line lambs tended to be more active and to follow their dams more closely at this stage (Cloete et al., 2005). The better following behavior of H line lambs was confirmed with 3-d-old lambs. All of these behavior attributes were listed by Alexander (1988) as factors conducive to lamb survival. The genetic trends for both lines thus make a contribution to the existing knowledge by proving that cumulative gains are possible when sustained selection pressure is applied for a trait closely related to lamb survival. In Ercanbrack and Knight (1998), sustained improvement in total BW of lamb weaned was reported to be partially driven by improved lamb survival. When the components contributing to improved reproduction in the Trangie Fertility Flock were considered, it was shown that lamb survival was improved despite an increased multiple birth rate in the flock selected for net reproduction (Atkins, 1980). Experience with the H and L lines supports these findings. The present study contributes to the existing knowledge by proving that cumulative gains are possible when sustained selection pressure is applied for a trait closely related to lamb survival, even in the presence of sizable and unfavorable direct-maternal genetic correlations for survival from tail docking to weaning and for overall lamb survival.

Fairly small unfavorable genetic changes were observed in maternal breeding values for early and late postnatal lamb survival on the underlying scale in the H line. These trends would have cancelled some of the gains made by direct genetic gains in the aggregate genotype, as was speculated by Everett-Hincks et al. (2005) when discussing the unfavorable direct-maternal covariance component for lamb survival in their study.

The regression of individual breeding values on year of birth suggested significant genetic change in maternal breeding values for birth weight in the H line. However, the magnitude of the change was small. The total maternal genetic gain after 22 yr of selection amounted to ~90 g. This correlated response in maternal breeding values is consistent with the observed direction of the maternal genetic correlation, as well as the contention of Knight et al. (1988) noted above. The direct genetic trend in the L line suggested an overall reduction of only ~30 g in birth weight over the duration of the experiment. Van Wyk et al. (1993) reported increases of, respectively, 0.023 and 0.004 kg per annum in direct and maternal breeding values for birth weight in a South African Dormer flock. Both direct and maternal breeding values for birth weight increased at ~0.003 kg per annum in the 29-yr study of Duguma et al. (2002). Slow, but positive, genetic change in birth weight was present in the latter studies, as well as in the H line in the present study. Heavy birth weights are often associated with birth difficulties and parturient deaths (Grommers et al., 1985; Dwyer, 2003). The general recommendation is to curb genetic increases in birth weight in sheep because excessive increases may lead to an unprecedented increase in incidence of dystocia. However, it is anticipated that selection on a trait such as the ability of ewes to rear multiples will ensure that birth weight remains in the optimal range for survival under paddock conditions.

When expressed relative to overall means, marked and sustainable direct genetic progress was observed for age-specific lamb survival on the underlying scale in the H line. These trends ranged from 0.23% per annum in maternal breeding values for survival at birth to 1.3% per annum in direct breeding values for survival from docking to weaning. This is in contrast with earlier contentions in the literature that such progress in sheep flocks is unlikely.

Unfavorable maternal genetic trends (–0.11 to –0.24% per annum) were observed for postnatal lamb survival on the underlying scale in the H line. This trend may be attributed to a sizable unfavorable direct-maternal genetic correlation for late postnatal lamb survival, which warrants further investigation. However, there is little evidence that the existence of this correlation entirely cancels direct genetic progress.

The responses attained in lamb survival could be attributed to a correlated response to divergent selection for the ability of ewes to rear multiples in this particular resource flock. Commercial lamb production is seen to potentially benefit from such selection because Merino lines capable of rearing their lambs with minimal external efforts would be desirable from an economic, as well as from an animal welfare, perspective.

	Footnotes

 
¹ We are grateful to all those involved in the maintenance and recording of the research flock (A. C. M. Kruger, A. J. Scholtz, J. J. E. Cloete, D. Marang, and Z. Stentyi; Institute for Animal Production, Elsenburg, South Africa). The maintenance of the breeding flock would have been impossible without the financial support of the South African Wool Industry and the Technology and Human Resources for Industry Program program of the South African Department of Trade and Industry.

² Corresponding author: schalkc{at}elsenburg.com

Received for publication March 28, 2008. Accepted for publication March 6, 2009.

	LITERATURE CITED

Top Abstract INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 


Alexander, G. 1984. Constraints to lamb survival. Pages 199–209 in Reproduction in sheep. D. R. Lindsay and D. T. Pearce, ed. Australian Academy of Science in conjunction with the Australian Wool Corporation, Canberra, Australia.

Alexander, G. 1988. What makes a good mother? Proc. Aust. Soc. Anim. Prod. 17:25–41.

Analla, M., and J. M. Serradilla. 1998. Estimation of correlations between ewe litter size and maternal effects on lamb weights in Merino sheep. Genet. Sel. Evol. 30:493–501.[CrossRef]

Arango, J., I. Misztal, S. Tsuruta, M. Culbertson, and W. Herring. 2005. Threshold-linear estimation of genetic parameters for farrowing mortality, litter size, and test performance of Large White sows. J. Anim. Sci. 83:499–506.[Abstract/Free Full Text]

Atkins, K. D. 1980. Selection for skin folds and fertility. Proc. Aust. Soc. Anim. Prod. 13:174–176.

Barwick, S. A., K. A. Leymaster, J. W. Keele, and W. R. Harvey. 1990. Estimates of genetic parameters for lamb survival to weaning and birth weight in the U.S. Suffolk. Proc. Aust. Soc. Anim. Breed. Gen. 8:331–334.

Casellas, J., G. Caja, and J. Piedrafita. 2007. Genetic determinism for within-litter birth weight variation and its relationship with litter weight and litter size in the Ripollesa ewe breed. Animal 1:637–644.[CrossRef]

Cloete, S. W. P., A. R. Gilmour, J. J. Olivier, and J. B. van Wyk. 2004. Genetic and phenotypic trends and parameters in reproduction, greasy fleece weight and live weight in Merino lines divergently selected for multiple rearing ability. Aust. J. Exp. Agric. 44:745–754.[CrossRef]

Cloete, S. W. P., J. C. Greeff, and R. P. Lewer. 2001. Environmental and genetic aspects of survival and early liveweight in Western Australian Merino sheep. S. Afr. J. Anim. Sci. 31:123–130.

Cloete, S. W. P., J. J. Olivier, M. A. Snyman, and E. du Toit. 1998. Genetic parameters and trends in a selection experiment for increased clean fleece weight involving South African Merinos. Aust. J. Exp. Agric. 38:427–432.[CrossRef]

Cloete, S. W. P., J. J. Olivier, J. B. van Wyk, G. J. Erasmus, and S. J. Schoeman. 2003a. Genetic parameters and trends for birth weight, birth coat score and weaning weight in Merino lines divergently selected for ewe multiple rearing ability. S. Afr. J. Anim. Sci. 33:248–256.

Cloete, S. W. P., and A. J. Scholtz. 1998. Lamb survival in relation to lambing and neonatal behaviour in medium wool Merino lines divergently selected for multiple rearing ability. Aust. J. Exp. Agric. 38:801–811.[CrossRef]

Cloete, S. W. P., A. J. Scholtz, J. J. E. Cloete, and J. B. van Wyk. 2005. The ability of Merino ewes and lambs to reunite after separation, as affected by divergent selection for ewe multiple rearing capacity. Aust. J. Exp. Agric. 45:1131–1137.[CrossRef]

Cloete, S. W. P., A. J. Scholtz, A. R. Gilmour, and J. J. Olivier. 2002. Genetic and environmental effects on lambing and neonatal behaviour of Dormer and South African Mutton Merino lambs. Livest. Prod. Sci. 78:183–193.[CrossRef]

Cloete, S. W. P., J. B. van Wyk, A. J. Scholtz, and A. R. Gilmour. 2003b. Genetic and environmental parameters for lambing behaviour in Merino lines divergently selected for ewe maternal rearing ability. Proc. N. Z. Soc. Anim. Prod. 63:169–172.

Cloete, S. W. P., J. B. van Wyk, A. J. Scholtz, and A. R. Gilmour. 2006. Genetic parameters and trends for lamb behaviour. Proc. 8th World Congr. Gen. Appl. Livest. Prod., August 13–16, 2006. Belo Horizonte, Brazil. CD communication 17–07.

Donnelly, F. B. 1982. A practical attempt to breed for better lamb survival. Proc. Aust. Soc. Anim. Prod. 14:30–32.

Donoghue, K. A., R. Rekaya, J. K. Bertrand, and I. Misztal. 2004. Threshold-linear analysis of measures of fertility in artificial insemination data and days to calving in beef cattle. J. Anim. Sci. 82:987–993.[Abstract/Free Full Text]

Duguma, G. J., S. J. Schoeman, S. W. P. Cloete, and G. F. Jordaan. 2002. Genetic parameter estimates of early growth traits in the Tygerhoek Merino flock. S. Afr. J. Anim. Sci. 32:66–75.

Dwyer, C. M. 2003. Behavioural development in the new-born lamb: Effect of maternal and birth related factors. Theriogenology 59:1027–1050.[CrossRef][Medline]

Ercanbrack, S. K., and A. D. Knight. 1998. Responses to various selection protocols for lamb production in Rambouillet, Targhee, Columbia and Polypay sheep. J. Anim. Sci. 76:1311–1325.[Abstract/Free Full Text]

Everett-Hincks, J. M., N. Lopez-Villalobos, H. T. Blair, and K. J. Stafford. 2005. The effect of ewe maternal behaviour score on lamb and litter survival. Livest. Prod. Sci. 93:51–61.[CrossRef]

Forrest, R. H., J. G. H. Hickford, and C. M. Frampton. 2007. Polymorphism at the ovine β₃-adrenergic receptor locus (ADRB3) and its association with lamb mortality. J. Anim. Sci. 85:2801–2806.[Abstract/Free Full Text]

Gama, L. T., G. E. Dickerson, L. D. Young, and K. A. Leymaster. 1991. Genetic and phenotypic variation in sources of preweaning lamb mortality. J. Anim. Sci. 69:2744–2753.[Abstract]

Grommers, F. J., L. Elving, and P. van Eldik. 1985. Parturition difficulties in sheep. Anim. Reprod. Sci. 9:365–374.[CrossRef]

Gudex, B. W., J. G. H. Hickford, and C. M. Frampton. 2005. Sire-line differences in lamb losses from starvation-exposure. Proc. N. Z. Soc. Anim. Prod. 65:186–190.

Haughey, K. G. 1983. Selective breeding for rearing ability as an aid to improving lamb survival. Aust. Vet. J. 60:361–363.[Medline]

Haughey, K. G. 1991. Perinatal lamb mortality—Its investigation, causes and control. J. S. Afr. Vet. Assoc. 62:78–91.[Medline]

Heydenrych, H. J., J. J. Du Plessis, and S. W. P. Cloete. 1984. Increasing the wool production of Merino sheep in the South Western Districts of South Africa by direct and indirect selection. Pages 399–412 in Proc. 2nd World Congr. Sheep & Beef Cattle Breed., 16–19 April, Pretoria. J. H. Hofmeyr and E. H. H. Meyer, ed. South African Stud Book and Livest. Improve. Assoc., Bloemfontein, South Africa.

Knight, T. W., P. R. Lynch, D. R. H. Hall, and H.-U. P. Hockey. 1988. Identification of factors contributing to the improved lamb survival in Marshall Romney sheep. N. Z. J. Agric. Res. 31:259–271.

Lee, D., I. Misztal, J. K. Bertrand, and R. Rekaya. 2002. National evaluation for calving ease, gestation length and birth weight by linear and threshold model methodologies. J. Appl. Genet. 43:209–216.[Medline]

Lopez-Villalobos, N., and D. J. Garrick. 1999. Genetic parameter estimates for lamb survival in Romney sheep. Proc. N. Z. Soc. Anim. Prod. 59:121–124.

Matika, O., J. B. van Wyk, G. J. Erasmus, and R. L. Baker. 2003. Genetic parameter estimates in Sabi sheep. Livest. Prod. Sci. 79:17–28.[CrossRef]

Matos, C. A. P., D. L. Thomas, L. D. Young, and D. Gianola. 2000. Genetic analyses of lamb survival in Rambouillet and Finnsheep flocks by linear and threshold models. Anim. Sci. 71:227–234.

Misztal, I., S. Tsuruta, T. Strabel, B. Auvray, T. Druet, and D. H. Lee. 2002. BLUPF90 and related programs (BGF90). Proc. 7th World Congr. Genet. Appl. Livest. Prod. 33:743–744. Montpellier, France.

Morris, C. A., S. M. Hickey, and J. N. Clarke. 2000. Genetic and environmental factors affecting lamb survival at birth and through to weaning. N. Z. J. Agric. Res. 43:515–524.

Mortimer, S. I., and K. D. Atkins. 1995. Maternal effects influence growth traits in Merino sheep. Proc. Aust. Assoc. Anim. Breed. Genet. 11:421–424.

Naidoo, P., S. W. P. Cloete, and A. Fossey. 2005. South African Merinos divergently selected for multiple rearing ability: A preliminary study of divergence based on RAPD markers. Proc. Assoc. Adv. Anim. Breed. Gen. 16:254–257.

Olivier, W. J., M. A. Snyman, J. B. van Wyk, and G. J. Erasmus. 1998. Genetic parameters for fitness traits in Merino sheep. Livest. Prod. Sci. 56:71–77.[CrossRef]

Riggio, V., R. Finocchiaro, and S. C. Bishop. 2008. Genetic parameters for early lamb survival and growth in Scottish Blackface sheep. J. Anim. Sci. 86:1758–1764.[Abstract/Free Full Text]

Safari, A., and N. M. Fogarty. 2003. Genetic parameters for sheep production traits: Estimates from the literature. Tech. Bull. No 49, NSW Agriculture, Orange NSW, Australia.

Safari, E., N. M. Fogarty, and A. R. Gilmour. 2005. A review of genetic parameter estimates for wool, growth, meat and reproduction traits in sheep. Livest. Prod. Sci. 92:271–289.[CrossRef]

Safari, E., N. M. Fogarty, A. R. Gilmour, K. D. Atkins, S. I. Mortimer, A. A. Swan, F. D. Brien, J. C. Greeff, and J. H. J. van der Werf. 2007. Across population genetic parameters for wool, growth, and reproduction traits in Australian Merino sheep. 2. Estimates of heritability and variance components. Aust. J. Agric. Res. 58:177–184.

Sawalha, R. M., J. Conington, S. Brotherstone, and B. Villanueva. 2006. Estimation of genetic and non-genetic parameters for survival traits of Scottish Blackface lambs. Proc. 8th World Congr. Genet. Appl. Livest. Prod., August 13–16, 2006. Belo Horizonte, Brazil. CD communication 04–18.

Sawalha, R. M., J. Conington, S. Brotherstone, and B. Villanueva. 2007. Analysis of lamb survival of Scottish Blackface sheep. Animal 1:151–157.[CrossRef]

Smith, G. M. 1977. Factors affecting birth weight, dystocia and preweaning survival in sheep. J. Anim. Sci. 44:745–753.[Abstract/Free Full Text]

Snyman, M. A., G. J. Erasmus, and J. B. van Wyk. 1998. The possible improvement of reproduction and survival rate in Afrino sheep using a threshold model. S. Afr. J. Anim. Sci. 28:120–124.

Van Wyk, J. B., G. J. Erasmus, and K. V. Konstantinov. 1993. Genetic and environmental trends of early growth traits in the Elsenburg Dormer stud. S. Afr. J. Anim. Sci. 23:85–87.

Welsh, C. S., D. J. Garrick, R. M. Ens, and G. B. Nicoll. 2006. Threshold model analysis of lamb survivability of Romney sheep. N. Z. J. Agric. Res. 49:411–418.

Willham, R. L. 1972. The role of maternal effects in animal breeding: III. Biometrical aspects of maternal effects in animals. J. Anim. Sci. 35:1288–1293.[Abstract/Free Full Text]

Yazdi, M. H., K. Johansson, P. Gates, A. Näsholm, H. Jorjani, and L.-E. Liljedahl. 1999. Bayesian analysis of birth weight and litter size in Baluchi sheep using Gibbs sampling. J. Anim. Sci. 77:533–540.[Abstract/Free Full Text]

This Article Free Via Open Access Abstract Full Text (PDF) All Versions of this Article: jas.2008-1065v1 87/7/2196    most recent Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Google Scholar Articles by Cloete, S. W. P. Articles by Olivier, J. J. PubMed PubMed Citation Articles by Cloete, S. W. P. Articles by Olivier, J. J.