Correlated Changes in Reproductive Components Accompanying 10 Generations of Selection for Improved Sow Productivity Index' S. M. Neal2 and K. M. Irvin
ABSTRACT: Reproductive components were compared between a line of sows selected (SI for improved sow productivity index (SPI = 6.5 x number born alive + adjusted 21-d litter weight1 and sows from a n unselected control (C) line. Generation 9 and 10, second-parity, Landrace sows were chosen from both the S (n = 35) and C (n = 331 line. Sows were slaughtered at a commercial slaughter plant at approximately 75 d of gestation and their reproductive tracts were recovered. Reproductive tracts were evaluated for uterine weight (UTWT),uterine horn length (UTLN),ovulation rate (OR),number of fully formed fetuses (NF), number of mummified fetuses (NMI, percentage of fetal survival (FS = NF/OR), fetal space (FSPACE = UTLNANF + NMI), and fetal position, sex, and weight. Select-line sows had greater NF (P < .lo)
and higher FS (P .lo) than C-line sows. Selectline sows had longer (P c .05),and heavier ( P c ,011 uteri than C-line sows. However, uterine length adjusted for NF was not different between the two lines. Uterine weight adjusted for NF was greater in S-line sows (P < .05). Select-line sows had greater total fetal weight (TFWT)( P c .051than did C-line sows. Female fetuses positioned between two male fetuses were lighter in weight than all other female fetuses (P < .011. Male fetuses positioned between two female fetuses did not differ in weight from all other male fetuses. In all cases, average select-line fetuses were heavier than average control-line fetuses ( P < .01). Selection for improved SPI changed major components of reproduction.
Key Words: Pigs, Reproduction, Selection, Sow Production, Uterus, Fetal Development
J. Anim. Sci. 1992. 70:2322-2327
Introduction Improvement in litter size in swine has been attempted by selection for the components (ovulation rate, embryonic survival, and uterine capacity) and by selection for the composite trait itself or in combination with other traits. Selection for ovulation rate in swine resulted in a correlated increase in litter size, and this difference was maintained during a period of relaxed selection
'Salaries and research support provided by State and Federal Funds appropriated to the Ohio Agric. Res. and Dev. Center, The Ohio State Univ. Journal Article no. 107-91. 2Present address: Dept. of Anim.Ind., Agric. Tech. Inst., The Ohio State Wniv., Wooster 44691. Received July 24, 1991. Accepted March 17, 1992.
(Zimmerman and Cunningham, 1975; Cunningham et al., 1979; Johnson et al., 1984; Lamberson et al., 19911. Selection for a n index of ovulation rate and embryo survival increased ovulation rate and number of fetuses (Neal et al., 1989). Selection for litter size in swine has been generally unsuccessful (Rutledge, 1980; Ollivier, 1982). Bichard and Seidel(1982) suggested a hyperprolific selection scheme for improving sow productivity. Lamberson et al, (19911 reported increased litter size in a line that had been selected for ovulation rate and subsequently selected for litter size. Avalos and Smith (19871 proposed selection for litter size based on indexes including information on relatives, and the expected response indicated that indexes may be more effective in increasing litter size. Irvin (1975) developed several selection indexes to improve swine reproduction. The indexes included one that was based on the number of pigs born alive and 21-d litter
2322
Downloaded from https://academic.oup.com/jas/article-abstract/70/8/2322/4705031 by Iowa State University user on 19 January 2019
Department of Animal Science, The Ohio State University, Columbus 43210 and Ohio Agricultural Research and Development Center, Wooster 44691
SELECTION FOR IMPROVED SOW PRODUCTIVITY INDEX
Materials and Methods Selection for improved sow productivity index (SPI = 6.5 x number born alive + 21-d litter weight adjusted for age and number nursed) was applied for 10 generations in a closed population (SI of Landrace swine (approximately 40 sows and eight boars per generation). Replacement gilts were selected from those first-parity litters having the highest SPI value. Eight sire lines were maintained and replacement boars were selected on SPI within sire line. In addition, a contemporary randomly mated control line (C)of similar size was maintained. Replacement gilts for the C line were chosen from the litters that deviated least from average SPI and replacement boars were chosen from within each of eight sire lines from those litters that deviated least from the C line average SPI value. Mating for both lines was at random, except to avoid sib matings. Five gilts were assigned to each boar. Generation 9 and 10 first-parity sows from both the S and C lines were retained and mated (spring 1989 and 1990) to produce second-parity offspring and all sows that conceived in both lines were included in this study. Sows were slaughtered at approximately 75 d of gestation at a commercial slaughter plant and gravid reproductive tracts were collected and measured immediately (number of sows slaughtered [nl, Generation 9, S n = 18, C n = 12; Generation 10, S n = 17, C n = 211. Intact reproductive tracts were weighed wet (UTWTI and observed for abnormalities. Ovulation rate (OR) was then determined by counting the number of corpora lutea on each ovary. Uterine length (UTLN) of each horn was determined by laying the intact uterus on a flat surface and measuring from the bifurcation to the tip of the horn along the outer surface. Each fully formed or mummified fetus was removed sequentially from the uterus after tying the umbilical cord with string at the body junction and cutting the cord at approximately 2.5 cm from the body. Numbers of normal (NF) and mummified (NMI fetuses, individual fetal weight, sex, and position were recorded.
Percentage of fetal survival (FS) was defined as the percentage of ova that were present as normalappearing fetuses at 75 d of gestation. Fetal space (FSPACEI is a gross estimate of the amount of uterine length available for each developing fetus (FSPACE = UTLNANF + NMI). Data were analyzed using the GLM procedure of SAS (1987). For the traits OR, NF, NM, FS, FSPACE, UTLN, and UTWT, the model included the effects of year, line, the year x line interaction, and a covariate that adjusted for the day of gestation on which the sows were slaughtered (DAYSPREG). The traits UTLN and UTWT were further analyzed with the additional covariate of NF added to the previous model. Data were also analyzed to determine whether differences existed between uterine horns for the traits measured. The traits OR, NF, NM, and FSPACE were analyzed using a model that included the effects of year, line, horn, line x horn interaction, and the covariate DAYSPREG. Total fetus weight within a uterine horn (TFWT) and UTLN were analyzed with the same model as described above with the additional effect of NF included as a covariate. Data on fetuses were analyzed to characterize line differences for fetal weight, sex, and position. Uterine horns that contained no fetuses (one C-line uterine horn) or contained mummified fetuses (three S-line uterine horns and four C-line uterine horns) were omitted from these analyses. However, all sows were represented in the analysis with a t least one uterine horn. Individual fetus data were analyzed both within and across line. The effects of sex of adjacent fetuses and position within horn were compared. Weights of female fetuses positioned between two male fetuses IFBM) were compared to those of all other female fetuses (AOF), and, conversely, weights of male fetuses positioned between two female fetuses (MBFI were compared to those of all other male fetuses (AOM). The model for comparison of fetal weights based on sex of adjacent fetuses included the effects of year, line, year x line interaction, horn, DAYSPREG, and position (Le., FBM vs AOF; MBF vs AOM). Fetal weights were also compared for the fetus a t the tip (nearest the ovary) of the horn (TIP) vs the average of all other fetuses in the same uterine horn (REST). Additionally, the average of the fetus weights at the tip, nearest the ovary, and at the end, nearest the bifurcation (ENDS), was compared to the average of all other fetus weights in the same horn (MIDDLE). The model for the analysis of these data included the effects of year, line, year x line interaction, sow nested within year-line, horn, and position (Le., TIP vs REST: ENDS vs MIDDLE).
Downloaded from https://academic.oup.com/jas/article-abstract/70/8/2322/4705031 by Iowa State University user on 19 January 2019
weight (SPI). Selection for SPI was applied in a closed population of Landrace swine for 10 generations and pigs from this population were used in this experiment. Peterson and Irvin (1989)reported that selection for SPI was effective through six generations with a realized heritability of SPI of .38 k .14. The objectives of this experiment were to determine correlated changes in reproductive components of sows selected for SPI and to determine the effects of fetal position, sex, and orientation within the uterus on fetal weight.
2323
2324
NEAL AND IRVIN
Table 1. Select- and control-line means and standard deviations for sow productivity traits in Generations 9 and 10 first-parity Landrace sowsa Select
Control
Generation
Mean
SD
Mean
SD
Difference
TNB~
9 10 9 10 9 10 9 10 9
10.21 8.87 9.46 7.97 14.42 11.80 45.88 38.21 170.88 155.21
2.57 2.67 2.25 2.46 3.44 3.17 9.16 14.81 23.19 31.20
9.56 8.14 8.48 6.54 12.85 9.92 38.54 27.64 149.35 134.60
2.12 2.45 1.85 2.32 3.56 3.18 10.38 13.81 23.26 26.25
.65 .73 .98 1.43 1.57 1.88 9.34 10.57 21.53 20.61
NBAC
LBW~ LW 21 de SPIf
10 ~
&Generation 9 Select, n = 27; Control, n = 28; Generation 10 Select, n = 30; Control, n = 28. bTotal number born. CNumber born alive. dLitter birth weight, kilograms. eLitter 21-d weight, kilograms, adjusted for age a t weighing. fSow productivity index.
Results and Discussion Sow Data First-parity litter data for the sows included in this study are given in Table 1. Select-line sows had higher mean performance in total number born, number born alive, litter birth weight, number of pigs at 21 d, litter 21-d weight, and SPI in each generation. The difference between lines increased from Generation 9 to 10 for all traits except the number of pigs at 21 d and SPI. These differences indicated continued effectiveness of
selection and that S-line sows tended to have superior mothering ability and superior milk production, which resulted in greater pig survival and greater litter weight a t 21 d. Peterson and Irvin (1988)reported results of the same selection experiment discussed here through Generation 6. They found a per-generation improvement of $8pigs per litter born alive, .2O f .07 pigs at 21 d, and 1.2 k .4 kg in 21-dlitter weight relative to the control line. Second-parity data from sows slaughtered at approximately 75 d of gestation are shown in Table 2. Differences between the S and C lines were not detected for OR, NM, or FSPACE. Select-
Table 2. Least squares means and standard errors for select- and control-line Generations 9 and 10 second-parity sows for reproductive traits Trait
ORCh NFC~ NMC~ FSdh FSPACE, cmch UTLN, cmfh UTWT, kggh UTLN adj., cmfi UTWT adj., k g a
Selecta 14.11 10.21 .09 74.00 40.31 388.30 13.09 377.67 12.50
.41 .39 .07 2.83 2.02 f 11.31 f .58 f 9.17 f .35
f f f f f
Controlb 13.95 9.06 .17 65.00 40.37 347.68 10.64 359.23 11.42
f
P-value .44
f .42 f .08 f 3.04 f 2.19 f 12.25 f .63 f 10.00 f .38
NS < .10
NS < .10
NS < .05 < .01 NS .05
&Number of sows = .35. bNumber of sows = 33. COR = number of corpora lutea, NF = number of fully formed fetuses, NM = number of mummified fetuses. dFS = (NF/OR) x 100. eFSPACE = UTLNANF + NMI. fUTLN = uterine length, measured on intact uterus from tip of uterine horn to bifurcation on each horn, combined length. gUTWT = uterine weight wet. hAdjusted for day of gestation. 'Adjusted for day of gestation and NF. jNot significant.
Downloaded from https://academic.oup.com/jas/article-abstract/70/8/2322/4705031 by Iowa State University user on 19 January 2019
Trait
FOR IMPROVED SOW PRODUCTIVITY INDEX
SELECTION
2325
Table 3 . Least squares means and standard errors for selectand control-line sows by uterine horn Selecta Trait
Controlb
Left 7.63 4.98 .03 43.19 1.80 188.72
f f
f f It f
.43 .27 .04 3.00 .04 5.27
6.46 f 5.24 f
.06 f 42.34 f 1.82 f 189.37 f
Left
.43 .27 .04 3.00 .04 5.30
7.19 4.73 .03 42.22 1.71 182.71
Right
.44 .28 .05 3.09 .04 5.41
f f
f f f f
8.89 4.58 .12 41.12 1.73 182.08
f .43 f .28 f .05 f 3.15 f .04 f 5.43
&Number of sows = 35. bNumber of sows = 33. COR = number of corpora lutea, NF = number of fetuses, NM = number of mummified fetuses. dTFWT = total fetus weight, does not include mummified fetuses. eUTLN = uterine length, measured on intact uterus from tip of uterine horn to bifurcation on each horn. fAdjusted for day of gestation. BAdjusted for number of fetuses. ~FSPACE = UTLNANF + NM).
line sows had 1.15 more fetuses (P < . l o ) and had 9% higher FS ( P < .lo). Figure 1 illustrates FS at various levels of OR in the S and C line. Bennett and Leymaster (1989) proposed a model for litter size based on OR, potential embryonic viability, and uterine capacity. Because OR was similar for the lines, and S sows had both greater FS and NF, this would seem to indicate that over the duration of the selection experiment S sows had improved uterine capacity or other positive uterine environmental changes resulting from selection from SPI. Neal et al. (1989) reported results of selection based on a n index of OR and embryo survival. They found a n increase of .57 k .11 in OR per generation and a decline in embryo survival of .013 f .0096per generation with a combined effect
I
19
-
-
n- 1
I n-4
n-6
-
15
-
E
14:
33
12
I n-3 n-5
I
Legend
n-4
n-7
~
Select
Control n No ofsows
n=9
1
n-3
m m
e
$
.2 .1
a
n-2
11,
1 n-3 n-1 __I n-1 n-1
10 9 _______________ i o 20 30 46
0
Legend
.5 Y m
n-I
I n-2
I
16
m
j n-3
-
n-1
17
m e
___
-
--
18
of a n increase in number of fetuses of .2O f .20 per generation. Selection for SPI prevented a decline in OR relative to the control and increased FS, which translated into greater NF. This may partially explain improved litter size and 21-d weight postnatally in the S line. Gross uterine measurements of length and weight (Table 2) were both greater for S-line sows (P < .05 and P < -01, respectively) than for C-line sows. However, when both uterine length and weight were adjusted for number of fetuses, only uterine weight was greater in the select line ( P < .05). Uteri from S-line sows averaged 1.08 kg heavier than uteri from C-line sows. W u et al. (1987) found that gilts with longer uterine horns had lower fetal mortality than gilts with shorter
1 - T
-
56
60 ' 70 ' 80 Percentage fetal survival
~, 6 7 8 9 Number of fetuses in the uterine horn -~-
0 1
2
3
4
~
5
,
1
, 0
1 n-1
90
io0
Figure 1. Percentage of fetuses surviving (FS) to d 75 of gestation for select- and control-line sows (n) at various levels of ovulation rate (OR).
2
6
5
12
14
13
9
4
Controln=3
1
9
17
16
8
5
1
Seledn-
2 1
Figure 2 . Select and control line average fetal weight within uterine horn containing various numbers of fetuses.
Downloaded from https://academic.oup.com/jas/article-abstract/70/8/2322/4705031 by Iowa State University user on 19 January 2019
ORCf NFC~ NMC~ FSPACE, cmfi TFWT, kgdfg UTLN, cmefg
Right
2326
NEAL AND IRVIN
Table 4. Least squares means and standard errors for female individual fetal weight as affected by sex of adjacent fetusesa Data set
n
FBMb
AOFC
P-value
300
,328 f ,008 .331 f ,012 328 f ,009
,351 f ,004
e .01
351 ,006 352 f ,005
NSd < .OS
Table 6 . Least squares means and standard errors for average individual fetal weight of fetuses positioned at the ends of the uterine horn compared to fetuses positioned in the middle of the uterine horna -
~
Combined Control Select
141 159
*
uterine horns. Differences were observed between the uterine length data in our study and that of Wu et al. (1987). Sows from the C line of our study had uteri that were approximately 80 cm longer a t a similar stage of gestation than was reported by Wu et al. (19871. Likewise, FSPACE was greater in our data. However, it was not clear how the determination of length of uterine horn per fetus was made in the data reported by Wu et al. (1987). Therefore, differences between the results of the two studies could simply be due to a difference in estimation or measurement procedures. Measurement of uterine length was slightly different in our study in that ligaments of reproductive tracts were not cut and the tracts were not laid out length wise and measured. In addition, the breed of swine differed between the two studies. If in fact uterine length and weight predict uterine capacity, then S-line sows seem to have greater uterine capacity. This observation was reinforced by the trends for greater FS, NF, total number of pigs born, number of pigs born alive, and litter birth weight in favor of S-line sows. Individual uterine horn data were analyzed to determine whether differences were present for reproductive traits between uterine horns. Ovulation rate was 7.41 k .31 on the left side compared
Table 5. Least squares means and standard errors for male individual fetal weight as affected by sex of adjacent fetusesa Data set Combined Control Select
n 353 165 188
MBFb 369 f ,008 ,348 f ,013 ,387 f ,010
AOMC ,365 f .003 ,355 f .005 ,376 f .005
&Data do not include fetuses in uterine horns containing mummified fetuses. bMale fetus positioned between two female fetuses. cAll other male fetuses.
n
Combined Control Select ~~
250 126 124
ENDS~ ,361 f ,003 ,349 f .004 373 f .005
~
MIDDLE ,362 f ,003 ,346 f ,004 378 f ,005
-
&Data do not include fetuses in uterine horns containing mummified fetuses. Analyzed on a within-uterine-horn basis. bAverage weight of fetuses positioned a t both ends of the uterine horn. CAverage weight of all other fetuses within the horn.
with 6.57 _+ .30 on the right side (P < .05). Total fetus weight per horn, adjusted for NF, was greater in S-line sows (1.81 k .03 kg) than in C-line sows (1.72 k .03 kg) (P e .05). Least squares means for left and right uterine horns of the two lines are presented in Table 3. No differences for the traits measured were detected. Sample sizes in this study, for characterizing components of reproduction, were small and, therefore, genetic drift may be a possible explanation for differences among lines. Changes due to drift were not accounted for in this analysis.
Fetus Data Mean fetal weight for both the S and C lines is illustrated in Figure 2. It is evident that mean fetal weight declined as litter size within uterine horn increased. Present in Tables 4, 5, 6, and 7 are results of data evaluated to compare sex of adjacent fetuses and location within the uterus on fetal weight. Female fetuses located between two male fetuses were smaller (P < .01) than all other female fetuses
Table 7. Least squares means and standard errors for individual fetal weight of fetuses positioned at the tip (nearest ovary) of the uterine horn compared to average fetal weight of remaining fetusesa Data set
n
Combined Control Select
250 126 124
TIP^ ,362 f ,004 .343 It ,005 ,374 f ,006
REST^
,359 f ,004 ,349 f ,005 ,376 f ,006
&Data do not include fetuses in uterine horns containing mummified fetuses. Analyzed on a within-uterine-horn basis. bWeight of the fetus positioned at the tip of the uterine horn nearest the ovary. CAverage weight of remaining fetuses within the horn.
Downloaded from https://academic.oup.com/jas/article-abstract/70/8/2322/4705031 by Iowa State University user on 19 January 2019
*Data do not include fetuses in uterine horns containing mummified fetuses. bFemale fetus positioned between two male fetuses. cAll other female fetuses. dNot significant.
Data set
SELECTION FOR IMPROVED SOW PRODUCTIVITY INDEX
Implications Improving sow productivity index did not change ovulation rate but did increase percentage of fetal survival, which combined to produce more fully formed fetuses a t d 75 of gestation and potentially more live pigs at birth. Results seem to indicate that continued improvement in litter size from selection for sow productivity index would be limited by ovulation rate. Further study is needed to determine whether female fetuses positioned between two male fetuses in the uterus are adversely affected and are subsequently hindered in performance.
Literature Cited Avalos, E., and C. Smith. 1987. Genetic improvement of litter size in pigs. Anim. Prod. 84:153. Bennett, G. L., and K. A. Leymaster. 1989. Integration of ovulation rate, potential embryonic viability and uterine capacity into a model of litter size in swine. J. Anim. Sci. 87:1230. Bichard, M., and C. M. Seidel. 1982. Selection for reproductive performance in maternal lines of pigs. 2nd World Congr. Genet. Appl. Livest. Prod. 3:585. Cunningham, P. J., M. E. England, L. D. Young, and D. R. Zimmerman. 1979. Selection for ovulation rate in swine: Correlated response in litter size and weight. J. Anim. Sci. 48:509.
Irvin, K. M. 1975. Genetic parameters and selection indexes for sow productivity. PbD. Dissertation. The Ohio State Univ., Columbus. Johnson, R. K., D. R. Zimmerman, and R. J. Kittok. 1984. Selection for components of reproduction in swine. Livest. Prod. Sci. 11:541. Lamberson, W. R., R. K. Johnson, D. R. Zimmerman, and T. E. Long. 1991. Direct responses to selection for increased litter size, decreased age a t puberty, or random selection following selection for ovulation rate in swine. J. Anim. Sci. 89: 3129.
Neal, S. M., R. K. Johnson, and R. J. Kittok. 1989. Index selection for components of litter size in swine: Response to five generations of selection. J. Anim. Sci. 87:1933. Ollivier, L. 1982. Selection for prolificacy in the pig. Pig News Info. 3:383. Peterson, G. A,, and K. M. Irvin. 1988. Responses due to selection for sow productivity in Landrace. Ohio State Res. and Ind. Report. Animal Science Dept. Ser. 88-1. Peterson, G. A,, and K. M. Irvin. 1989. Realized heritability estimates for sow productivity index in Landrace swine. J. Anim. Sci. 87:48 Gbstr.1. Rutledge, J. J. 1980. Fraternity size and swine reproduction. I. Effect on fecundity in gilts. J. Anim. Sci. 51:868. SAS. 1987. SAWSTAT@User’s Guide. SAS Inst. Inc., Cary, NC. Waldorf, D. P., W. C. Foote, H. L. Self, A. B. Chapman, and L. E. Casida. 1957. Factors affecting fetal pig weight late in gestation. J. h i m . Sci. 18:976. Wu, M. C., M. D. Hentzel, and P. J. Dzuik. 1987. Relationships between uterine length and number of fetuses and prenatal mortality in pigs. J. Anim. Sci. 65:762. Zimmerman, D. R., and P. J. Cunningham. 1975. Selection for ovulation rate in swine: Population, procedures and ovulation response. J. Anim. Sci. 4031.
Downloaded from https://academic.oup.com/jas/article-abstract/70/8/2322/4705031 by Iowa State University user on 19 January 2019
(Table 4). However, when the data were analyzed on a within-line basis, this phenomenon was only present in S-line sows IP c .05). Therefore, it may be that because of increased competition within the uterus of S sows, male fetuses overcrowd female fetuses located between them. Male fetuses located between two female fetuses were not different in weight than other male fetuses in either line (Table 5). Fetal weights were also compared for location of development. Fetuses that were located a t the ENDS position did not differ in weight from fetuses in the MIDDLE position (Table 6). There were no differences in fetal weight between fetuses developing in the TIP position and fetuses located in the REST position (Table 7). Waldorf et al. (1957) found that fetal weight was greater for those fetuses located near the uterotubal and cervical ends of the uterine horn during the later stages of pregnancy. It may be that fetal competition becomes greater during the time beyond 75 d of gestation and that in this study the measurement was not done late enough in pregnancy to detect differences in fetal weight as a result of location.
2327
ABSTRACT: Reproductive components were compared between a line of sows selected (SI for improved sow productivity index (SPI = 6.5 x number born alive + adjusted 21-d litter weight1 and sows from a n unselected control (C) line. Generation 9 and 10, second-parity, Landrace sows were chosen from both the S (n = 35) and C (n = 331 line. Sows were slaughtered at a commercial slaughter plant at approximately 75 d of gestation and their reproductive tracts were recovered. Reproductive tracts were evaluated for uterine weight (UTWT),uterine horn length (UTLN),ovulation rate (OR),number of fully formed fetuses (NF), number of mummified fetuses (NMI, percentage of fetal survival (FS = NF/OR), fetal space (FSPACE = UTLNANF + NMI), and fetal position, sex, and weight. Select-line sows had greater NF (P < .lo)
and higher FS (P .lo) than C-line sows. Selectline sows had longer (P c .05),and heavier ( P c ,011 uteri than C-line sows. However, uterine length adjusted for NF was not different between the two lines. Uterine weight adjusted for NF was greater in S-line sows (P < .05). Select-line sows had greater total fetal weight (TFWT)( P c .051than did C-line sows. Female fetuses positioned between two male fetuses were lighter in weight than all other female fetuses (P < .011. Male fetuses positioned between two female fetuses did not differ in weight from all other male fetuses. In all cases, average select-line fetuses were heavier than average control-line fetuses ( P < .01). Selection for improved SPI changed major components of reproduction.
Key Words: Pigs, Reproduction, Selection, Sow Production, Uterus, Fetal Development
J. Anim. Sci. 1992. 70:2322-2327
Introduction Improvement in litter size in swine has been attempted by selection for the components (ovulation rate, embryonic survival, and uterine capacity) and by selection for the composite trait itself or in combination with other traits. Selection for ovulation rate in swine resulted in a correlated increase in litter size, and this difference was maintained during a period of relaxed selection
'Salaries and research support provided by State and Federal Funds appropriated to the Ohio Agric. Res. and Dev. Center, The Ohio State Univ. Journal Article no. 107-91. 2Present address: Dept. of Anim.Ind., Agric. Tech. Inst., The Ohio State Wniv., Wooster 44691. Received July 24, 1991. Accepted March 17, 1992.
(Zimmerman and Cunningham, 1975; Cunningham et al., 1979; Johnson et al., 1984; Lamberson et al., 19911. Selection for a n index of ovulation rate and embryo survival increased ovulation rate and number of fetuses (Neal et al., 1989). Selection for litter size in swine has been generally unsuccessful (Rutledge, 1980; Ollivier, 1982). Bichard and Seidel(1982) suggested a hyperprolific selection scheme for improving sow productivity. Lamberson et al, (19911 reported increased litter size in a line that had been selected for ovulation rate and subsequently selected for litter size. Avalos and Smith (19871 proposed selection for litter size based on indexes including information on relatives, and the expected response indicated that indexes may be more effective in increasing litter size. Irvin (1975) developed several selection indexes to improve swine reproduction. The indexes included one that was based on the number of pigs born alive and 21-d litter
2322
Downloaded from https://academic.oup.com/jas/article-abstract/70/8/2322/4705031 by Iowa State University user on 19 January 2019
Department of Animal Science, The Ohio State University, Columbus 43210 and Ohio Agricultural Research and Development Center, Wooster 44691
SELECTION FOR IMPROVED SOW PRODUCTIVITY INDEX
Materials and Methods Selection for improved sow productivity index (SPI = 6.5 x number born alive + 21-d litter weight adjusted for age and number nursed) was applied for 10 generations in a closed population (SI of Landrace swine (approximately 40 sows and eight boars per generation). Replacement gilts were selected from those first-parity litters having the highest SPI value. Eight sire lines were maintained and replacement boars were selected on SPI within sire line. In addition, a contemporary randomly mated control line (C)of similar size was maintained. Replacement gilts for the C line were chosen from the litters that deviated least from average SPI and replacement boars were chosen from within each of eight sire lines from those litters that deviated least from the C line average SPI value. Mating for both lines was at random, except to avoid sib matings. Five gilts were assigned to each boar. Generation 9 and 10 first-parity sows from both the S and C lines were retained and mated (spring 1989 and 1990) to produce second-parity offspring and all sows that conceived in both lines were included in this study. Sows were slaughtered at approximately 75 d of gestation at a commercial slaughter plant and gravid reproductive tracts were collected and measured immediately (number of sows slaughtered [nl, Generation 9, S n = 18, C n = 12; Generation 10, S n = 17, C n = 211. Intact reproductive tracts were weighed wet (UTWTI and observed for abnormalities. Ovulation rate (OR) was then determined by counting the number of corpora lutea on each ovary. Uterine length (UTLN) of each horn was determined by laying the intact uterus on a flat surface and measuring from the bifurcation to the tip of the horn along the outer surface. Each fully formed or mummified fetus was removed sequentially from the uterus after tying the umbilical cord with string at the body junction and cutting the cord at approximately 2.5 cm from the body. Numbers of normal (NF) and mummified (NMI fetuses, individual fetal weight, sex, and position were recorded.
Percentage of fetal survival (FS) was defined as the percentage of ova that were present as normalappearing fetuses at 75 d of gestation. Fetal space (FSPACEI is a gross estimate of the amount of uterine length available for each developing fetus (FSPACE = UTLNANF + NMI). Data were analyzed using the GLM procedure of SAS (1987). For the traits OR, NF, NM, FS, FSPACE, UTLN, and UTWT, the model included the effects of year, line, the year x line interaction, and a covariate that adjusted for the day of gestation on which the sows were slaughtered (DAYSPREG). The traits UTLN and UTWT were further analyzed with the additional covariate of NF added to the previous model. Data were also analyzed to determine whether differences existed between uterine horns for the traits measured. The traits OR, NF, NM, and FSPACE were analyzed using a model that included the effects of year, line, horn, line x horn interaction, and the covariate DAYSPREG. Total fetus weight within a uterine horn (TFWT) and UTLN were analyzed with the same model as described above with the additional effect of NF included as a covariate. Data on fetuses were analyzed to characterize line differences for fetal weight, sex, and position. Uterine horns that contained no fetuses (one C-line uterine horn) or contained mummified fetuses (three S-line uterine horns and four C-line uterine horns) were omitted from these analyses. However, all sows were represented in the analysis with a t least one uterine horn. Individual fetus data were analyzed both within and across line. The effects of sex of adjacent fetuses and position within horn were compared. Weights of female fetuses positioned between two male fetuses IFBM) were compared to those of all other female fetuses (AOF), and, conversely, weights of male fetuses positioned between two female fetuses (MBFI were compared to those of all other male fetuses (AOM). The model for comparison of fetal weights based on sex of adjacent fetuses included the effects of year, line, year x line interaction, horn, DAYSPREG, and position (Le., FBM vs AOF; MBF vs AOM). Fetal weights were also compared for the fetus a t the tip (nearest the ovary) of the horn (TIP) vs the average of all other fetuses in the same uterine horn (REST). Additionally, the average of the fetus weights at the tip, nearest the ovary, and at the end, nearest the bifurcation (ENDS), was compared to the average of all other fetus weights in the same horn (MIDDLE). The model for the analysis of these data included the effects of year, line, year x line interaction, sow nested within year-line, horn, and position (Le., TIP vs REST: ENDS vs MIDDLE).
Downloaded from https://academic.oup.com/jas/article-abstract/70/8/2322/4705031 by Iowa State University user on 19 January 2019
weight (SPI). Selection for SPI was applied in a closed population of Landrace swine for 10 generations and pigs from this population were used in this experiment. Peterson and Irvin (1989)reported that selection for SPI was effective through six generations with a realized heritability of SPI of .38 k .14. The objectives of this experiment were to determine correlated changes in reproductive components of sows selected for SPI and to determine the effects of fetal position, sex, and orientation within the uterus on fetal weight.
2323
2324
NEAL AND IRVIN
Table 1. Select- and control-line means and standard deviations for sow productivity traits in Generations 9 and 10 first-parity Landrace sowsa Select
Control
Generation
Mean
SD
Mean
SD
Difference
TNB~
9 10 9 10 9 10 9 10 9
10.21 8.87 9.46 7.97 14.42 11.80 45.88 38.21 170.88 155.21
2.57 2.67 2.25 2.46 3.44 3.17 9.16 14.81 23.19 31.20
9.56 8.14 8.48 6.54 12.85 9.92 38.54 27.64 149.35 134.60
2.12 2.45 1.85 2.32 3.56 3.18 10.38 13.81 23.26 26.25
.65 .73 .98 1.43 1.57 1.88 9.34 10.57 21.53 20.61
NBAC
LBW~ LW 21 de SPIf
10 ~
&Generation 9 Select, n = 27; Control, n = 28; Generation 10 Select, n = 30; Control, n = 28. bTotal number born. CNumber born alive. dLitter birth weight, kilograms. eLitter 21-d weight, kilograms, adjusted for age a t weighing. fSow productivity index.
Results and Discussion Sow Data First-parity litter data for the sows included in this study are given in Table 1. Select-line sows had higher mean performance in total number born, number born alive, litter birth weight, number of pigs at 21 d, litter 21-d weight, and SPI in each generation. The difference between lines increased from Generation 9 to 10 for all traits except the number of pigs at 21 d and SPI. These differences indicated continued effectiveness of
selection and that S-line sows tended to have superior mothering ability and superior milk production, which resulted in greater pig survival and greater litter weight a t 21 d. Peterson and Irvin (1988)reported results of the same selection experiment discussed here through Generation 6. They found a per-generation improvement of $8pigs per litter born alive, .2O f .07 pigs at 21 d, and 1.2 k .4 kg in 21-dlitter weight relative to the control line. Second-parity data from sows slaughtered at approximately 75 d of gestation are shown in Table 2. Differences between the S and C lines were not detected for OR, NM, or FSPACE. Select-
Table 2. Least squares means and standard errors for select- and control-line Generations 9 and 10 second-parity sows for reproductive traits Trait
ORCh NFC~ NMC~ FSdh FSPACE, cmch UTLN, cmfh UTWT, kggh UTLN adj., cmfi UTWT adj., k g a
Selecta 14.11 10.21 .09 74.00 40.31 388.30 13.09 377.67 12.50
.41 .39 .07 2.83 2.02 f 11.31 f .58 f 9.17 f .35
f f f f f
Controlb 13.95 9.06 .17 65.00 40.37 347.68 10.64 359.23 11.42
f
P-value .44
f .42 f .08 f 3.04 f 2.19 f 12.25 f .63 f 10.00 f .38
NS < .10
NS < .10
NS < .05 < .01 NS .05
&Number of sows = .35. bNumber of sows = 33. COR = number of corpora lutea, NF = number of fully formed fetuses, NM = number of mummified fetuses. dFS = (NF/OR) x 100. eFSPACE = UTLNANF + NMI. fUTLN = uterine length, measured on intact uterus from tip of uterine horn to bifurcation on each horn, combined length. gUTWT = uterine weight wet. hAdjusted for day of gestation. 'Adjusted for day of gestation and NF. jNot significant.
Downloaded from https://academic.oup.com/jas/article-abstract/70/8/2322/4705031 by Iowa State University user on 19 January 2019
Trait
FOR IMPROVED SOW PRODUCTIVITY INDEX
SELECTION
2325
Table 3 . Least squares means and standard errors for selectand control-line sows by uterine horn Selecta Trait
Controlb
Left 7.63 4.98 .03 43.19 1.80 188.72
f f
f f It f
.43 .27 .04 3.00 .04 5.27
6.46 f 5.24 f
.06 f 42.34 f 1.82 f 189.37 f
Left
.43 .27 .04 3.00 .04 5.30
7.19 4.73 .03 42.22 1.71 182.71
Right
.44 .28 .05 3.09 .04 5.41
f f
f f f f
8.89 4.58 .12 41.12 1.73 182.08
f .43 f .28 f .05 f 3.15 f .04 f 5.43
&Number of sows = 35. bNumber of sows = 33. COR = number of corpora lutea, NF = number of fetuses, NM = number of mummified fetuses. dTFWT = total fetus weight, does not include mummified fetuses. eUTLN = uterine length, measured on intact uterus from tip of uterine horn to bifurcation on each horn. fAdjusted for day of gestation. BAdjusted for number of fetuses. ~FSPACE = UTLNANF + NM).
line sows had 1.15 more fetuses (P < . l o ) and had 9% higher FS ( P < .lo). Figure 1 illustrates FS at various levels of OR in the S and C line. Bennett and Leymaster (1989) proposed a model for litter size based on OR, potential embryonic viability, and uterine capacity. Because OR was similar for the lines, and S sows had both greater FS and NF, this would seem to indicate that over the duration of the selection experiment S sows had improved uterine capacity or other positive uterine environmental changes resulting from selection from SPI. Neal et al. (1989) reported results of selection based on a n index of OR and embryo survival. They found a n increase of .57 k .11 in OR per generation and a decline in embryo survival of .013 f .0096per generation with a combined effect
I
19
-
-
n- 1
I n-4
n-6
-
15
-
E
14:
33
12
I n-3 n-5
I
Legend
n-4
n-7
~
Select
Control n No ofsows
n=9
1
n-3
m m
e
$
.2 .1
a
n-2
11,
1 n-3 n-1 __I n-1 n-1
10 9 _______________ i o 20 30 46
0
Legend
.5 Y m
n-I
I n-2
I
16
m
j n-3
-
n-1
17
m e
___
-
--
18
of a n increase in number of fetuses of .2O f .20 per generation. Selection for SPI prevented a decline in OR relative to the control and increased FS, which translated into greater NF. This may partially explain improved litter size and 21-d weight postnatally in the S line. Gross uterine measurements of length and weight (Table 2) were both greater for S-line sows (P < .05 and P < -01, respectively) than for C-line sows. However, when both uterine length and weight were adjusted for number of fetuses, only uterine weight was greater in the select line ( P < .05). Uteri from S-line sows averaged 1.08 kg heavier than uteri from C-line sows. W u et al. (1987) found that gilts with longer uterine horns had lower fetal mortality than gilts with shorter
1 - T
-
56
60 ' 70 ' 80 Percentage fetal survival
~, 6 7 8 9 Number of fetuses in the uterine horn -~-
0 1
2
3
4
~
5
,
1
, 0
1 n-1
90
io0
Figure 1. Percentage of fetuses surviving (FS) to d 75 of gestation for select- and control-line sows (n) at various levels of ovulation rate (OR).
2
6
5
12
14
13
9
4
Controln=3
1
9
17
16
8
5
1
Seledn-
2 1
Figure 2 . Select and control line average fetal weight within uterine horn containing various numbers of fetuses.
Downloaded from https://academic.oup.com/jas/article-abstract/70/8/2322/4705031 by Iowa State University user on 19 January 2019
ORCf NFC~ NMC~ FSPACE, cmfi TFWT, kgdfg UTLN, cmefg
Right
2326
NEAL AND IRVIN
Table 4. Least squares means and standard errors for female individual fetal weight as affected by sex of adjacent fetusesa Data set
n
FBMb
AOFC
P-value
300
,328 f ,008 .331 f ,012 328 f ,009
,351 f ,004
e .01
351 ,006 352 f ,005
NSd < .OS
Table 6 . Least squares means and standard errors for average individual fetal weight of fetuses positioned at the ends of the uterine horn compared to fetuses positioned in the middle of the uterine horna -
~
Combined Control Select
141 159
*
uterine horns. Differences were observed between the uterine length data in our study and that of Wu et al. (1987). Sows from the C line of our study had uteri that were approximately 80 cm longer a t a similar stage of gestation than was reported by Wu et al. (19871. Likewise, FSPACE was greater in our data. However, it was not clear how the determination of length of uterine horn per fetus was made in the data reported by Wu et al. (1987). Therefore, differences between the results of the two studies could simply be due to a difference in estimation or measurement procedures. Measurement of uterine length was slightly different in our study in that ligaments of reproductive tracts were not cut and the tracts were not laid out length wise and measured. In addition, the breed of swine differed between the two studies. If in fact uterine length and weight predict uterine capacity, then S-line sows seem to have greater uterine capacity. This observation was reinforced by the trends for greater FS, NF, total number of pigs born, number of pigs born alive, and litter birth weight in favor of S-line sows. Individual uterine horn data were analyzed to determine whether differences were present for reproductive traits between uterine horns. Ovulation rate was 7.41 k .31 on the left side compared
Table 5. Least squares means and standard errors for male individual fetal weight as affected by sex of adjacent fetusesa Data set Combined Control Select
n 353 165 188
MBFb 369 f ,008 ,348 f ,013 ,387 f ,010
AOMC ,365 f .003 ,355 f .005 ,376 f .005
&Data do not include fetuses in uterine horns containing mummified fetuses. bMale fetus positioned between two female fetuses. cAll other male fetuses.
n
Combined Control Select ~~
250 126 124
ENDS~ ,361 f ,003 ,349 f .004 373 f .005
~
MIDDLE ,362 f ,003 ,346 f ,004 378 f ,005
-
&Data do not include fetuses in uterine horns containing mummified fetuses. Analyzed on a within-uterine-horn basis. bAverage weight of fetuses positioned a t both ends of the uterine horn. CAverage weight of all other fetuses within the horn.
with 6.57 _+ .30 on the right side (P < .05). Total fetus weight per horn, adjusted for NF, was greater in S-line sows (1.81 k .03 kg) than in C-line sows (1.72 k .03 kg) (P e .05). Least squares means for left and right uterine horns of the two lines are presented in Table 3. No differences for the traits measured were detected. Sample sizes in this study, for characterizing components of reproduction, were small and, therefore, genetic drift may be a possible explanation for differences among lines. Changes due to drift were not accounted for in this analysis.
Fetus Data Mean fetal weight for both the S and C lines is illustrated in Figure 2. It is evident that mean fetal weight declined as litter size within uterine horn increased. Present in Tables 4, 5, 6, and 7 are results of data evaluated to compare sex of adjacent fetuses and location within the uterus on fetal weight. Female fetuses located between two male fetuses were smaller (P < .01) than all other female fetuses
Table 7. Least squares means and standard errors for individual fetal weight of fetuses positioned at the tip (nearest ovary) of the uterine horn compared to average fetal weight of remaining fetusesa Data set
n
Combined Control Select
250 126 124
TIP^ ,362 f ,004 .343 It ,005 ,374 f ,006
REST^
,359 f ,004 ,349 f ,005 ,376 f ,006
&Data do not include fetuses in uterine horns containing mummified fetuses. Analyzed on a within-uterine-horn basis. bWeight of the fetus positioned at the tip of the uterine horn nearest the ovary. CAverage weight of remaining fetuses within the horn.
Downloaded from https://academic.oup.com/jas/article-abstract/70/8/2322/4705031 by Iowa State University user on 19 January 2019
*Data do not include fetuses in uterine horns containing mummified fetuses. bFemale fetus positioned between two male fetuses. cAll other female fetuses. dNot significant.
Data set
SELECTION FOR IMPROVED SOW PRODUCTIVITY INDEX
Implications Improving sow productivity index did not change ovulation rate but did increase percentage of fetal survival, which combined to produce more fully formed fetuses a t d 75 of gestation and potentially more live pigs at birth. Results seem to indicate that continued improvement in litter size from selection for sow productivity index would be limited by ovulation rate. Further study is needed to determine whether female fetuses positioned between two male fetuses in the uterus are adversely affected and are subsequently hindered in performance.
Literature Cited Avalos, E., and C. Smith. 1987. Genetic improvement of litter size in pigs. Anim. Prod. 84:153. Bennett, G. L., and K. A. Leymaster. 1989. Integration of ovulation rate, potential embryonic viability and uterine capacity into a model of litter size in swine. J. Anim. Sci. 87:1230. Bichard, M., and C. M. Seidel. 1982. Selection for reproductive performance in maternal lines of pigs. 2nd World Congr. Genet. Appl. Livest. Prod. 3:585. Cunningham, P. J., M. E. England, L. D. Young, and D. R. Zimmerman. 1979. Selection for ovulation rate in swine: Correlated response in litter size and weight. J. Anim. Sci. 48:509.
Irvin, K. M. 1975. Genetic parameters and selection indexes for sow productivity. PbD. Dissertation. The Ohio State Univ., Columbus. Johnson, R. K., D. R. Zimmerman, and R. J. Kittok. 1984. Selection for components of reproduction in swine. Livest. Prod. Sci. 11:541. Lamberson, W. R., R. K. Johnson, D. R. Zimmerman, and T. E. Long. 1991. Direct responses to selection for increased litter size, decreased age a t puberty, or random selection following selection for ovulation rate in swine. J. Anim. Sci. 89: 3129.
Neal, S. M., R. K. Johnson, and R. J. Kittok. 1989. Index selection for components of litter size in swine: Response to five generations of selection. J. Anim. Sci. 87:1933. Ollivier, L. 1982. Selection for prolificacy in the pig. Pig News Info. 3:383. Peterson, G. A,, and K. M. Irvin. 1988. Responses due to selection for sow productivity in Landrace. Ohio State Res. and Ind. Report. Animal Science Dept. Ser. 88-1. Peterson, G. A,, and K. M. Irvin. 1989. Realized heritability estimates for sow productivity index in Landrace swine. J. Anim. Sci. 87:48 Gbstr.1. Rutledge, J. J. 1980. Fraternity size and swine reproduction. I. Effect on fecundity in gilts. J. Anim. Sci. 51:868. SAS. 1987. SAWSTAT@User’s Guide. SAS Inst. Inc., Cary, NC. Waldorf, D. P., W. C. Foote, H. L. Self, A. B. Chapman, and L. E. Casida. 1957. Factors affecting fetal pig weight late in gestation. J. h i m . Sci. 18:976. Wu, M. C., M. D. Hentzel, and P. J. Dzuik. 1987. Relationships between uterine length and number of fetuses and prenatal mortality in pigs. J. Anim. Sci. 65:762. Zimmerman, D. R., and P. J. Cunningham. 1975. Selection for ovulation rate in swine: Population, procedures and ovulation response. J. Anim. Sci. 4031.
Downloaded from https://academic.oup.com/jas/article-abstract/70/8/2322/4705031 by Iowa State University user on 19 January 2019
(Table 4). However, when the data were analyzed on a within-line basis, this phenomenon was only present in S-line sows IP c .05). Therefore, it may be that because of increased competition within the uterus of S sows, male fetuses overcrowd female fetuses located between them. Male fetuses located between two female fetuses were not different in weight than other male fetuses in either line (Table 5). Fetal weights were also compared for location of development. Fetuses that were located a t the ENDS position did not differ in weight from fetuses in the MIDDLE position (Table 6). There were no differences in fetal weight between fetuses developing in the TIP position and fetuses located in the REST position (Table 7). Waldorf et al. (1957) found that fetal weight was greater for those fetuses located near the uterotubal and cervical ends of the uterine horn during the later stages of pregnancy. It may be that fetal competition becomes greater during the time beyond 75 d of gestation and that in this study the measurement was not done late enough in pregnancy to detect differences in fetal weight as a result of location.
2327
