J. Anim Sci. 2007. 85:1385-1392. doi:10.2527/jas.2006-631
© 2007 American Society of Animal Science
Selection for litter size at day five to improve litter size at weaning and piglet survival rate1
G. Su2,
M. S. Lund and
D. Sorensen
University of Aarhus, Faculty of Agricultural Sciences, Department of Genetics and Biotechnology, DK-8830, Tjele, Denmark
 |
Abstract
|
|---|
Selection for total number of piglets born (TNB) since 1992 has led to a significant increase in this trait in Danish Landrace and Danish Yorkshire but has also been accompanied by an increase in piglet mortality. The objective of this study was to estimate the genetic and phenotypic parameters for litter size and survival to find alternative selection criteria to improve litter size at weaning. Data from Landrace (9,300 litters) and Yorkshire (6,861 litters) were analyzed using REML based on a linear model including genetic effects of sow and service-sire. The estimates of heritability (based on the sow component) for TNB, number born alive (NBA), and number alive at d 5 after birth (N5D) and at weaning (about 3 wk, N3W) ranged from 0.066 to 0.090 in Landrace and 0.050 to 0.070 in Yorkshire. Genetic correlations between TNB and N3W were 0.289 in Landrace and 0.561 in Yorkshire, but between N5D and N3W the estimated genetic correlation was 0.995 in both populations. The approximate estimates of heritability for survival rate per litter at birth (SVB = NBA/TNB), from birth to d 5 (SV5 = N5D/NBA), and from d 5 to weaning (SVW = N3W/N5D) were 0.130, 0.131, and 0.023, respectively, in Landrace, and 0.095, 0.043, and 0.009, respectively, in Yorkshire. Genetic correlations between TNB and survival rates at different stages were negative. On the other hand, genetic correlations between N5D and survival rates and between N3W and survival rates were strongly or moderately positive, except for the correlations with SVW in Yorkshire. The results suggest that selection for N5D could be an interesting alternative to improve litter size at weaning and piglet survival for Danish Landrace and Danish Yorkshire.
Key Words: genetic correlation • heritability • litter size • mortality • survival
|
INTRODUCTION
|
|---|
Litter size at weaning is one of the most important traits in pig production. Direct selection for this trait is generally restricted in practice due to cross-fostering, which also makes it difficult to adequately estimate genetic parameters for litter size at weaning. Therefore, in practical pig breeding, improvement of litter size at weaning is achieved by selecting for litter size at birth. Recent studies indicate that this approach may increase piglet mortality (Lobke et al., 1983
; Johnson et al., 1999; Lund et al., 2002; Robinson and Quinton, 2002).
Selection for total number of piglets born since 1992 has led to an increase in this trait in Danish Landrace and Danish Yorkshire, but also to an increase in piglet mortality, and it has been further observed that most cases of death occurred at birth and during the following 5 d (A. Vernersen, DanBred International, Copenhagen, Denmark, personal communication). Thus, it is expected that selection for litter size at d 5 will capture a large part of the genetic variance for piglet survival, and thereby will be more effective than selection for total number born with regard to genetic improvement of litter size at weaning and piglet survival.
Litter size and piglet survival rate are usually treated as traits of the sow. However, the fertilization capacity of the service-sire and the genotype of the piglet could have an effect on these traits (Strang, 1970; Hill and Webb, 1982; See et al., 1993; Woodward et al., 1993; Van der Lende et al., 1999). It therefore seems reasonable to estimate genetic parameters for these traits based on a model that includes effects of sow and service-sire.
The objectives of this study were 1) to present estimates of genetic and phenotypic parameters for litter size, 2) to report preliminary results involving piglet survival, and 3) to test the hypothesis that expected selection response for litter size at weaning could be increased by changing the selection criterion from total number born to number alive at d 5 in Danish sow populations.
|
MATERIALS AND METHODS
|
|---|
Animal Care and Use Committee approval was not obtained for this study because the data were obtained from an existing database of performance records.
Populations and Data
Data were collected from 43 nuclear farms of Danish Landrace and Danish Yorkshire during the period from May 2002 to December 2004. Pedigrees for sows and service-sires were traced back 5 generations or more. The structure of the data set is shown in Table 1. Sows were kept under commercial conditions, and all matings took place using AI. At farrowing, the total number of piglets born and the number of dead piglets (regarded as stillbirth) were recorded. Piglets were weighed individually within 2 d (80% within l d) after farrowing. Piglet mortality, BW, and the cause and date of death were registered during the preweaning period. Litter weight was recorded at weaning (average 21.27 d; range 19 to 23 d).
Assignment of cross-fostering was done within the first 3 d after birth, and information concerning donor and recipient dams was recorded. In total, 10.9% of all Landrace piglets and 7.0% of all Yorkshire piglets born alive were cross-fostered. A large proportion of the litters were involved in cross-fostering; 66% percent of the litters in Landrace and 47% in Yorkshire were donors or recipients.
The traits under analysis were total number of piglets born (TNB), number of piglets born alive (NBA), litter size at d 5 (N5D), litter size at weaning (about 3 wk, N3W), survival rate at birth (SVB), survival rate during early preweaning [i.e., from birth to d 5 (SV5)], and survival rate during late preweaning [i.e., from d 5 to weaning (SVW)]. The NBA was defined as TNB minus the number that died at farrowing (stillborn piglets). The N5D and N3W were measured on the basis of the number of piglets of the biological sow, ignoring the cross-fostering effect. Survival rates were defined as: SVB = NBA/TNB, SV5 = N5D/NBA, and SVW = N3W/N5D. The SV5 and SVW were calculated on the basis of the piglets remaining in their own litter (i.e., the piglets involved in cross-fostering were excluded from the analysis of SV5 and SVW). For example, a litter had TNB = 16 and NBA = 14, 2 piglets were moved out at d 1 and 1 piglet from another litter was moved in at d 3; among the 12 piglets (14 – 2) staying in their own litter, 2 died before or at d 5, and 1 died after d 5. The survival rates for this litter were calculated as SVB = 14/16, SV5 = 10/12, and SVW = 9/10, without considering the mortality of the 2 piglets moved out and the one moved in.
Statistical Analysis
Litter size and survival rate at different stages were analyzed using multiple-trait models. The basic model to describe the observations of litter size was
and the basic model for survival rate was
where y is the vector of observations; b is the vector of fixed effects including farm, year-season, and parity; c is the vector of farm x year-season interaction effects treated as random to avoid confounding with other effects in the model due to very few observations in some combinations of farms and year-seasons; pd is the vector of sow permanent effects; ps is the vector of service-sire permanent effects; d is the vector of genetic effects of sow; s is the vector of genetic effects of service-sire; e is the vector of random residuals; and X, Wc, Wpd, Wps, Zd, and Zs are incidence matrices associating b, c, pd, ps, d, and s with y. The random effects were assumed to be independent of each other, except for d and s, which were assumed to be correlated.
The (co)variance components for litter size at different stages were estimated using a 4-trait model (TNB, NBA, N5D, and N3W), and for survival rates at different periods using a 3-trait model (SVB, SV5, and SVW). The correlations between litter size and survival rate were estimated pairwise using a 2-trait model. All random effects were assumed to be normally distributed. Thus,
where C0, D0, S0, G0, and R0 are covariance matrixes for herd x year-season effects, permanent effects of sow, permanent effects of service-sire, additive genetic effects of sow and service sire, and residuals, respectively, I is the identity matrix, and A is the matrix of additive genetic relationships among animals in the pedigree.
Phenotypic variance (
2p) was defined as 2p = 2pd + 2ps + 2d + 2s + 2e for litter size and 2p = 2pd + 2d + 2s + 2e for survival, where 2pd was the variance of permanent effects of sow, 2ps was the variance of permanent effects of service-sire, 2d was the variance of sow additive genetic effects, 2s was the variance of service-sire additive genetic effects, and 2e was the residual variance. Heritability was defined as the ratio of sow additive genetic variance to phenotypic variance (h2sow = 2d/2p; i.e., both litter size and survival rate were considered traits of the sow). Correspondingly, the genetic correlation between any pair of traits was defined as the correlation between the sow additive genetic effects on the traits.
Standard error for a function of estimated (co)variances (F = f(
)) was estimated using an expansion of the Taylor series:
where n was the number of variables in the function, Var(i) was the asymptotic variance of I, and Cov(I, j) was the asymptotic covariance between I and j. Thus, the square of SE for variance component k in proportion to phenotypic variance 
was
For a correlation coefficient 
The asymptotic (co)variances for the estimates of (co)-variance components were obtained from the approximated observed information matrix. The asymptotic (co)variances related to the estimated phenotypic (co)-variance were calculated as a linear combination of the asymptotic (co)variances of the components involved.
An additional analysis was conducted for survival rates in the arc-sine scale. The arcsine transformation, yt = arcsin(
), is often used for binomial proportions. In the transformation, a zero survival rate was counted as 1/4n and a 100% survival rate as (n – 
)/n, where n is litter size at the start point. In the current study, inferences from the transformed scale were almost identical to those based on the original scale; therefore, we presented results based on the original scale only.
The analyses were performed using the average information (AI) REML procedure (Jensen et al., 1997) with the DMU package (Madsen and Jensen, 2004).
|
RESULTS
|
|---|
The mean and coefficient of variation for litter size and survival rate at different stages are shown in Table 2. In both populations TNB and total mortality were high. Landrace had a larger TNB, but also a higher farrowing mortality than Yorkshire. As a result, NBA was similar in the 2 populations. Preweaning mortality, N5D, and N3W were almost the same in the 2 populations. Coefficients of variation for litter size ranged from 25 to 31% and had a tendency to increase with age. Coefficients of variation for SVB and SV5 ranged from 14 to 18%, whereas SVW showed less variation.
Most cases of death occurred at farrowing and during the first 5 d after farrowing (Figure 1). After d 5, the mortality was very low in both populations. The most common cause of death was stillbirth (Table 3). Among the dead piglets, on the average over the 2 populations, 54% were stillbirth, 9.9% died soon after birth (before the time when birth weight was measured), 11.2% were crushed by the sows, 8.5% were due to starvation, and the rest 16.4% died of other causes. In addition, the proportion of stillbirth was higher in Landrace than in Yorkshire, whereas the mortalities due to crushing and starvation were higher in Yorkshire than in Landrace.
Table 4 shows estimates of variance components for litter size at different stages as proportions of phenotypic variance. Heritability (based on the sow component, h2sow) ranged from 0.066 to 0.090 in Landrace and from 0.050 to 0.070 in Yorkshire. Estimates of heritabilities for N5D and N3W were larger than those for TNB and NBA in both populations, but the differences between them were not statistically significant. Sow additive genetic variance (h2sow x 2p) was almost constant for the 4 traits. Relative to the phenotypic variance, estimates for sire additive genetic effects (g2sire) ranged from 0.026 to 0.041 in Landrace and from 0.012 to 0.017 in Yorkshire. Furthermore, there was a tendency for sow and sire additive genetic components to increase with the age of the piglets. Estimates of the variance component associated with sow permanent effects (PEsow) for litter size were close to 0.10 at all 4 stages in both populations. Estimates of the variance component associated with sire permanent effects (PEsire) ranged from 0.017 to 0.030 in Landrace and from 0.023 to 0.049 in Yorkshire, with a tendency to decrease with increasing number of days from farrowing. Estimates of repeatability for sows ranged from 0.145 to 0.218, and for service sires ranged from 0.040 to 0.065. All correlations between sire and sow additive genetic effects were positive.
View this table:
[in this window]
[in a new window]
|
Table 4. Phenotypic variance ( ), and the proportion of phenotypic variance due to sow permanent effect (PEsow), service-sire permanent effect (PEsire), sow genetic effect (h2sow), and service-sire genetic effect (g2sire) for litter size traits1
|
|
As shown in Table 5, the estimate of the phenotypic correlation between TNB and N3W was 0.538 in Landrace and 0.591 in Yorkshire, whereas the estimate of the correlation between N5D and N3W was 0.974 in Landrace and 0.957 in Yorkshire. Genetic correlations were estimated from the sow genetic components. Estimates of the genetic correlation between TNB and N3W were 0.289 in Landrace and 0.531 in Yorkshire, whereas the estimate of the genetic correlation between N5D and N3W was close to 1 (0.995) in both populations.
View this table:
[in this window]
[in a new window]
|
Table 5. Genetic correlations (based on sow genetic component, above the diagonal) and phenotypic correlation (below the diagonal) among litter size traits1
|
|
Estimates of variance components for survival rates were different in the 2 breeds and in the various periods (Table 6). In Landrace, the sow and sire variance components were much larger than in Yorkshire. Both variance components for SVW were very small in the 2 populations. The variance component of sire additive genetic effects was small, but statistically significantly different from zero, for the 3 survival rates in Landrace, and for SV5 in Yorkshire.
View this table:
[in this window]
[in a new window]
|
Table 6. Phenotypic variance (), and the proportion of phenotypic variance due to sow permanent effect (PEsow), sow genetic effect (h2sow), and service-sire genetic effect (g2sire) for piglet survival traits1
|
|
Phenotypic and genetic correlations among survival rates during different periods were positive (Table 7). Phenotypic correlations among the 3 survival rates were weak, ranging from 0.067 to 0.128. Genetic correlations were relatively high, ranging from 0.100 to 0.495, but the estimates were not significantly different from 0, except for the correlations between SVB and SVW and between SV5 and SVW in Landrace.
View this table:
[in this window]
[in a new window]
|
Table 7. Genetic correlations (based on sow genetic component, above the diagonal), and phenotypic correlations (below the diagonal) among piglet survival traits1
|
|
Phenotypic and genetic correlations between litter size and survival rates are shown in Table 8. Phenotypic correlations between TNB and the 3 survival rates were all negative. In contrast, the phenotypic correlations between N5D and N3W on the 1 side, and the 3 survival rates on the other, were all positive. The NBA showed a positive phenotypic correlation with SVB and a negative correlation with SV5 and with SVW. The pattern of estimates of genetic correlations was in agreement with that of the phenotypic correlations, except for the correlations between NBA and SV5 and between N3W and SVW in Yorkshire. Thus, TNB had a negative genetic correlation with the 3 survival rates, NBA had a moderate positive genetic correlation with SVB, and N5D and N3W had a moderate positive genetic correlation with SVB and SV5.
View this table:
[in this window]
[in a new window]
|
Table 8. Phenotypic correlations (rp) and genetic correlations (rg, based on sow genetic component) between litter size and survival traits1
|
|
|
DISCUSSION
|
|---|
Results from the present analysis show that the genetic correlation between TNB and N3W is low or moderate, whereas the genetic correlation between N5D and N3W is close to 1. Moreover, there is a negative genetic correlation between TNB and piglet survival rate, but a positive genetic correlation between N5D and piglet survival. In addition, the magnitude of genetic variation of N5D is similar to that of TNB, and the estimate of heritability for N5D is higher than for TNB (not statistically significant, P = 0.16 for Landrace and P = 0.26 for Yorkshire). The results support the hypothesis that selection for N5D should be more effective than selection for TNB with regard to increasing N3W and piglet survival.
The estimates of heritability for TNB and NBA are lower than average figures reported in the literature. As reviewed by Haley et al. (1988), most estimates of heritability for litter size are close to 0.1. However, one can expect variation among populations. Long-term selection for this trait could also reduce genetic variation to some extent. The effect of selection on genetic variability has been demonstrated by some studies (e.g., Bulmer, 1971). In the current study, estimates of heritability for litter size in Danish Landrace were larger than those in Danish Yorkshire. This agrees with the observation that selection response for TNB in Landrace was larger than that in Yorkshire, despite the fact that similar selection pressures were applied. In addition, it was found that the estimates of heritability for TNB and NBA were smaller than those for N5D and N3W in both populations. This could be partly explained by the effect of long-term selection for TNB.
For litter size, the variance component associated with service-sire permanent effects ranged from 0.017 to 0.030 in Landrace and from 0.023 to 0.049 in Yorkshire. This component is not related to the genotype of the piglet, but accounts for variation of sire permanent effects on sire fertility. The estimated variance components were consistent with reports in previous studies (e.g., See et al., 1993; Woodward et al., 1993; Van der Lende et al., 1999; Hamann et al., 2004). On the other hand, the variance component of sire genetic effects on litter size reflects sire fertility and the contribution to embryo survival through the offspring’s genotype. The repeatability of sire for litter size ranged from 0.040 to 0.065. Therefore, it is reasonable to include service-sire effects in a model for genetic evaluation of litter size. In addition, the correlation between sow genetic effects and service-sire genetic effects was small, but positive. This result is in line with the findings by Hamann et al. (2004).
Previous studies have reported low estimates of heritability for piglet mortality or survival rate, with an average of 0.05 (at the level of the litter and as a trait of the sow), as reviewed by Rothschild and Bidanel (1998). However, there is rather large variation among estimates from different studies. For example, Lamberson and Johnson (1984) reported an estimate of heritability for preweaning survival of 0.03, whereas Ferguson et al. (1985) reported a value of 0.14 in Yorkshire and 0.18 in Duroc. Damgaard et al. (2003) reported a heritability of 0.13 for proportion of stillbirths in Swedish Yorkshire. Our preliminary analysis showed that heritabilities (based on the sow component) for SVB and SV5 were substantial in Danish Landrace (0.130 and 0.131) and Danish Yorkshire (0.095 and 0.043), suggesting that genetic improvement of piglet survival rate at farrowing and during the early sucking period could be effective. Heritability for SVW was close to zero in Landrace and Yorkshire, probably due to very low mortality during this period in the 2 populations. Moreover, phenotypic correlations among the 3 survival rates were very low (0.067 to 0.128), and genetic correlations ranged from 0.100 to 0.495. This indicates that the genetic backgrounds for piglet mortality at the 3 stages may be different.
The estimated proportions of the genetic variation due to service-sire for survival rates were low (ranged from 0.002 to 0.034), but statistically significantly different from zero, for the 3 survival rates in Landrace, and for SV5 in Yorkshire. If piglet survival is regarded as a trait of the piglet, the sire component of genetic variance is equal to 1/4 of the additive genetic variance, and the estimates of sire variance component correspond to direct heritabilities for survival rate (based on litter record) ranging from 0.008 to 0.136. A more thorough study is necessary to investigate direct additive genetic effects on piglet survival.
Because of the difficulties in evaluating breeding values for N3W due to cross-fostering, genetic improvement of this trait by direct selection is not feasible. Therefore, genetic improvement in litter size at weaning has often focused on selection for TNB. However, selection for TNB alone leads to an increase in perinatal and preweaning mortality. High genetic correlations (about 0.5) between TNB and number of dead piglets have been reported in previous studies (Lobke et al., 1983; Robinson and Quinton, 2002). In the current study, estimates of the genetic correlation between TNB and the 3 survival rates were all negative. Similarly, Lund et al. (2002) reported a negative maternal genetic correlation between TNB and perinatal and preweaning survival rate in Finnish Landrace (–0.16 and –0.39) and in Finnish Yorkshire (–0.14 and 0.01). An argument based on a Taylor series expansion indicates that the negative genetic correlation between TNB and piglet survival should lead to a reduction in genetic correlation between TNB and N3W. This has been confirmed by the current study (0.289 in Danish Landrace and 0.531 in Danish Yorkshire).
Selection for TNB since 1992 has led to a total response of 3.8 and 3.0 piglets per litter (calculated as the difference in estimated breeding values between 1992 and 2004) in Danish Landrace and Danish Yorkshire, respectively (Nielsen, 2004). However, during this period, there was also an increase in perinatal and preweaning mortality in the 2 populations. As shown in the current study, total mortality of piglets was 31.3% in Landrace and 26.6% in Yorkshire. The values for total mortality are similar to those reported in previous studies on dam lines under selection for TNB (Serenius et al., 2003; Arango et al., 2006). Moreover, Landrace has shown a higher selection response for TNB than Yorkshire (Nielsen, 2004), but as shown in the current study, also a higher stillbirth rate. Similarly, Johnson et al. (1999) and Petry and Johnson (2004) reported that selection based on an index including ovulation rate and embryonic survival had an unfavorable effect on number of stillborn piglets.
Contrary to TNB, genetic correlations between N5D and piglet survival rates (especially SVB and SV5) were moderately positive, and the genetic correlation between N5D and N3W was close to 1. Moreover, results from a pilot study (not shown) involving cross-fostering data show that the nursing sow has a small effect on piglet survival rate during the first 5 d after birth. This implies that breeding values for N5D can be predicted ignoring the influence of cross-fostering. All of these results suggest that selection for N5D could be a good alternative approach to genetic improvement of N3W and piglet survival.
In conclusion, the current study shows that total number of piglets born has a negative genetic correlation with piglet survival rate and a low to moderate positive genetic correlation with litter size at weaning. On the other hand, the genetic correlation between litter size at d 5 and piglet survival rate was moderately positive, and between litter size at d 5 and at weaning was close to 1. These results suggest changing the selection criterion from total number born to number of piglets alive at d 5 in pig breeding programs. This approach is expected to effectively improve litter size at weaning and piglet survival rate.
|
Footnotes
|
|---|
1 The research was supported by the Research Project SUPERSOW from The Danish National Committee for Pig Production. 
2 Corresponding author: guosheng.su{at}agrsci.dk
Received for publication September 15, 2006.
Accepted for publication February 26, 2007.
|
LITERATURE CITED
|
|---|
Arango, J., I. Misztal, S. Tsuruta, M. Culbertson, J. W. Holl, and W. Herring. 2006. Genetic study of individual preweaning mortality and birth weight in Large White piglets using threshold - linear models. Livest. Prod. Sci. 101:208–218.[CrossRef]
Bulmer, M. G. 1971. The effects of selection on genetic variability. Am. Nat. 105:201–211.[CrossRef]
Damgaard, L. H., L. Rydhmer, P. Lovendahl, and K. Grandinson. 2003. Genetic parameters for within-litter variation in piglet birth weight and change in within-litter variation during suckling. J. Anim. Sci. 81:604–610.[Abstract/Free Full Text]
Ferguson, P. W., W. R. Harvey, and K. M. Irvin. 1985. Genetic, phenotypic and environmental relationships between sow body weight and sow productivity traits. J. Anim. Sci. 60:375–384.[Abstract/Free Full Text]
Haley, C. S., E. Avalos, and C. Smith. 1988. Selection for litter size in the pig. Anim. Breed. Abstr. 56:317–332.
Hamann, H., R. Steinheuer, and O. Distl. 2004. Estimation of genetic parameters for litter size as a sow and boar trait in German herdbook Landrace and Pietrain swine. Livest. Prod. Sci. 85:201–207.[CrossRef]
Hill, W. G., and A. J. Webb. 1982. Genetics of reproduction in the pig. Pages 541–564 in Control of Pig Reproduction. D. J. A. Cole and G. R. Foxcroft, ed. Butterworths, London, UK.
Jensen, J., E. A. Mantysaari, P. Madsen, and R. Thompson. 1997. Residual maximum likelihood estimation of (co)variance components in multivariate mixed linear models using average information. J. Indian Agric. Stat. 49:215–236.
Johnson, R. K., M. K. Nielsen, and D. S. Casey. 1999. Responses in ovulation rate, embryonal survival and litter traits in swine to 14 generations of selection to increase litter size. J. Anim. Sci. 77:541–557.[Abstract/Free Full Text]
Lamberson, W. R., and R. K. Johnson. 1984. Preweaning survival in swine: Heritability of direct and maternal effects. J. Anim. Sci. 59:346–349.[Abstract/Free Full Text]
Lobke, A., H. Willeke, and F. Pirchner. 1983. Genetic parameters for reproductive traits estimated by different methods in German Landrace. Proc. 34th Annu. Meet. Eur. Assoc. Anim. Prod., Madrid, Spain. I:14–15.
Lund, M. S., M. Puonti, L. Rydhmer, and J. Jensen. 2002. Relationship between litter size and perinatal and pre-weaning survival in pigs. Anim. Sci. 74:217–222.
Madsen, P., and J. Jensen. 2004. A User’s Guide to DMU, version 6, release 4.5. Danish Inst. Agric. Sci., Tjele, Denmark.
Nielsen, B. 2004. LG5 – new goal for litter size. DanAvl Magasinet, May 2004:19–20.
Petry, D. B., and R. K. Johnson. 2004. Responses to 19 generations of litter size selection in the Nebraska Index line. I. Reproductive responses estimated in pure line and crossbred litters. J. Anim. Sci. 82:1000–1006.[Abstract/Free Full Text]
Robinson, J. A. B., and V. M. Quinton. 2002. Genetic parameters of early neonatal piglet survival and number of piglets born. In Proc. 7th World Congr. Genet. Appl. Livest. Prod., Montpellier, France, Session 3:1–4.
Rothschild, M. F., and J. P. Bidanel. 1998. Biology and Genetics of Reproduction. Pages 313–343 in The Genetics of the Pig. M. F. Rothschild and A. Ruvinskty, ed. CAB Int., Oxon, UK.
See, M. T., J. W. Mabry, and J. K. Bertrand. 1993. Restricted maximum likelihood estimation of variance components from field data for number of pigs born alive. J. Anim. Sci. 71:2905–2909.[Abstract]
Serenius, T., M. L. Sevon-Aimonen, A. Kause, E. A. Mäntysaari, and A. Mäki-Tanila. 2003. Selection potential of different prolificacy traits in the Finnish Landrace and Large White populations. Acta Agric. Scand. Sect. A. 54:36–43.
Strang, G. S. 1970. Litter productivity in Large White pigs. 1. The relative importance of some sources of variation. Anim. Prod. 12:225–233.
Van der Lende, T., M. H. A. Willemsen, J. A. M. Van Arendonk, and E. B. P. G. Van Haandel. 1999. Genetic analysis of the service sire effect on litter size in swine. Livest. Prod. Sci. 58:91–94.[CrossRef]
Woodward, B. W., J. W. Mabry, M. T. See, J. K. Bertrand, and L. L. Benyshek. 1993. Development of an animal model for across-herd genetic evaluation of number born alive in swine. J. Anim. Sci. 71:2040–2046.[Abstract]
Top
Abstract
INTRODUCTION
