Daryl L. Kuhlers 3, Steve B. Jungst 3 and J. A. Little 4. Auburn University 3, AL ABSTRACT

Transcription

1 COMPARISONS OF SPECIFIC CROSSES FROM DUROC-LANDRACE, YORKSHIRE-LANDRACE AND HAMPSHIRE-LANDRACE SOWS MANAGED IN TWO TYPES OF GESTATION SYSTEMS: LITTER TRAITS AND SOW WEIGHTS 1,2 Daryl L. Kuhlers 3, Steve B. Jungst 3 and J. A. Little 4 Auburn University 3, AL ABSTRACT Two hundred fifty-six Duroc-Landrace (DL), Yorkshire-Landrace (YL) and Hampshire- Landrace (HL) sows, in two replicates, were randomly mated to purebred Duroc (D), Yorkshire (Y) or Hampshire (H) boars. The sows were kept either in pasture lots during gestation or in gestation stalls. These matings produced 844 litters, which were used to determine the effect of sire breed, dam breeding, gestation environment and the interactions of these effects on litter sizes and weights and sow weights after weaning the fourth litter. Gestation environment was not an important source of variation for any of the traits studied, nor were there any interactions of sire breed or dam breeding with gestation environment. Litter sizes at 21 d, 56 d and marketing were reduced by one pig/litter in H- sired litters compared with litters sired by D and Y boars. This resulted in H-sired litters weighing 2.7 and 20 kg less at 21 d and 56 d, respectively, and.7 kg less litter weight per day to 100 kg than D- and Y-sired litters. Differences in litter sizes and litter weights between D- and Y-sired litters were not significant. Litter sizes and weights among the three kinds of crossbred sows did not differ at any stage of production up to marketing at 100 kg. Three-breed cross litters were.6 to.7 pig/litter larger than backcross litters from 21 d to marketing. This resulted in three-breed cross litters weighing 3.8 kg and 17.1 kg more at 21 d and 56 d, respectively, and litter weight produced per day to 100 kg was.6 kg/d greater than in backcross litters. The DL sows had heavier weights after weaning the fourth litter than the YL and HL sows; these two latter breeds did not differ significantly. (Key Words: Pigs, Cross Breeding, Hybrids, Mothering Ability, Heterosis, Genotype Environment Interaction.) J. Anim. Sci : Introduction Crossbreeding is useful in increasing productivity for traits of economic importance such a litter sizes, litter weights and growth rates (Johnson, 1980, t981; Kuhlers et al., 1982; Gaugler et al., 1984; McLaren et al., 1Contribution from the Alabama Agric. Exp. Sta., Jo1.Lrnai Set. No P. ZThe authors are indebted to Waiter R. Harvey, Dept. of Dairy Sci., Ohio State Univ. for his advice and assistance in the data analysis. adept, of Anim. and Dairy Sci. 4Lower Coastal Plain Substation, Alabama Agric. Exp. Sta. Received June 24, Accepted September 28, ). However, further study is needed to determine which breeds to use on the female side of the pedigree and which breed of boar to use to maximize productivity. Johnson (1980) concluded that the two-breed crossbred sow with the largest 21-d litter size was Hampshire-Landrace (HL), with Yorkshire-Landrace (YL) ranking second. Kuhlers et al. (1981) found that Duroc-Landrace (DL) sows were approximately equal to YL sows in reproductive and maternal performance, as reflected by 21-d litter weights. Follow-up work indicated that HL sows were superior to DL and Spot- Landrace (SL) sows for 21-d litter weight, with SL sows ranking third (Kuhlers et al., 1982). With one exception (Holtmann et al., 1975), HL crossbred sows have not been evaluated contemporaneously with YL cross- 920

2 GENOTYPE-ENVIRONMENT INTERACTIONS IN SOWS 921 bred sows. One objective of this study was to evaluate the performance of these two crossbred sow types along with DL crossbred sows. Performance of YL sows has been inconsistent among studies (Johnson, 1980). They ranked first in two studies and last in another for litter size born. It was hypothesized that the gestation facilities used in the various studies may have altered the ranking of crossbred sow types, (i.e., a genotype x facility type interaction existed). Therefore, a second objective of this study was to compare the three crossbred sow types (DL, YL and HL) in two different gestation systems, either group-fed on pasture lots or confined in gestation stalls. Materials and Methods Two hundred fifty-six DL, YL and HL females, in two replicates, were allotted randomly to be mated to Duroc (D), Yorkshire (Y) and Hampshire (H) boars. Each sow had an opportunity to produce four litters. Each sow was mated to boars of the same breed throughout the experiment. A total of 844 litters was farrowed. The F1 sows were produced on the farm by mating purebred Landrace females to 23 boars (eight D, eight Y or seven H). The crossbred gilts were reared on concrete feeding floors to 100 kg using diets and management typical for market hogs at this station. After weighing 100 kg, the gilts were moved to drylots and fed 2.7 kg/d of a corn-soy diet prior to breeding at 8 mo of age. The litters were sired by 54 boars, 16 D, 19 Y and 19 H, that were selected from private breeders only on the basis of pedigree, general condition, conformation and strength of feet and legs. Sows that did not conceive during a 6-wk breeding period, that aborted during two consecutive parities or that suffered a severe disabling injury were culled from the study. Within each of the three farrowing groups of each replicate, a randomly allotted one-half of the sows of each cross (6 or 7 sows) were kept on bahia (Paspalum notatum) pasture during gestation (.4-ha lots/sow-cross). They were group-fed in feed troughs. The other one-half of the sows of each cross were kept in gestation stalls (.6 m x 4.2 m) during the gestation period. All sows remained in their assigned gestation system throughout the study and were fed, in both gestation environments, approximately 2.7 kg/d of a corn-soy diet during breeding and 1.8 kg/d (12.2% CP, 3, 076 kcal ME/kg) of a balanced diet after breeding for the duration of the gestation period. During the winter months, the amount of feed fed after breeding was increased to 2.7 kg/d. Farrowing took place in three groups, 2 mo apart, in a farrowing house with plasticcoated, expanded metal floors. Lactating sows were fed 1.8 kg/d of feed plus.45 kg/pig in the litter. Litters were weaned and boar pigs were castrated at 35 d or 42 d of age (confounded with farrowing group). After weaning, the pigs were moved to a fiat-deck, warm nursery with woven wire floors until they were approximately 10 wk of age. Pigs were not weighed at weaning, but rather at 56 d of age to include the effect of postweaning stress on litter sizes and weights. After leaving the nursery, the pigs were moved to open-fronted buildings with either solid concrete or slatted floors. The pigs remained in these buildings until they were marketed at approximately 100 kg. The litter traits studied included number of pigs born, number born alive, litter birth weight, number alive at 21 d, litter weight at 21 d, number alive at 56 d, litter weight at 56 d, number marketed and litter weight marketed per day. Litter weight marketed per day was defined as the weight per day of age to 100 kg for each pig in a litter and was summed over all pigs in the litter. Weight of the sow at weaning of her fourth litter was used as an indicator of salvage weight. The 21- and 56-d weights were adjusted, if the weights were not taken at 21 and 56 d of age, by the methods of Whatley and Quaife (1937). Days to 100 kg were adjusted using the recommendations of the National Swine Improvement Federation (NSIF, 1981). Statistical Analyses. The numbers of litters by sire breed, dam breed and gestation environment and the number of sow weights at weaning of the fourth litter are given in Table 1. The data were analyzed by least squares procedures using LSMLMW (Harvey, 1987). Preliminary analyses of the litter traits were completed with the following model: Yijklmnopq = I d, + Ri + Mj + F k + El+ Gm/Ri +RMij + RFik + REu + MFj~ +MEj] + FEkl + EGlm/Ri + Pn + RPin + ~jn + FPkn + EPln + GPmn]Ri + So/RMPij n + Dp/RMFEGijktm + Uq

3 922 KUHLERS ET AL. TABLE 1. NUMBER OF SOWS AT THE END OF THE STUDY AND LITTERS BY BREED OF SIRE, BREEDING OF DAM AND GESTATION ENVIRONMENT Gestation Environment Sire Dam No. of breed' breeding b sows c Pasture Stalls Totals D DL D YL D IlL Y DL Y YL Y HL H DL H YL H HL Totals D = Duroc, Y = Yorkshire and H = Hampshire. bdl = Duroc-Landrace, YL = Yorkshire-Landtace and HL = Hampshire-Landrace. csows with weights at the weaning of the fourth litter. where R, M, F, E, G, P, S, D and U are the effects of replicate, sire breed, dam breeding, gestation environment, farrowing group, parity-year-season, sire, dam and residual, respectively. Sire effects (S/RMP) were considered random and the mean square associated with them was the error term for R, M and its interaction. Dam effects (D/RMFEG) were considered random and the mean square associated with them was used as the error term to test the effects of F, E and G/R and the interaction of these effects with each other and with R and M. Sources of variation associated with P and interactions with P were tested with the residual mean square. On the basis of these preliminary analyses, ME, FE, EG/R, FP and EP were not important sources of variation (P >.25) and therefore were deleted from the model for the final analyses. Means were separated using Tukey's test. The sum of squares associated with the difference in performance of three-breed cross litters vs backcross litters was partitioned from the sire breed x dam breeding sum of squares. The contrast made was 1/6(D-YL + D.HL + Y.DL + Y.HL + H.DL + H.YL) - 1/3(D.DL + Y.YL + H.HL). Sow weight at weaning of the fourth litter was analyzed with a preliminary model that included effects of R, M, F, E, G/R, and the two-factor interactions among R, E, M and F. The effects of E and all of the two-factor interactions, except M x F, were not important sources of variation for this trait and were deleted from the model in the final analysis. Results and Discussion Litter Traits. Gestation environment was significant only for litter birth weight. Litters from sows assigned to the pasture gestation environment were 1.1 kg heavier than litters born to sows that were kept in stalls during the gestation periods (Table 2). The heavier litters produced by sows assigned to the pasture gestation environment may be due, in part, to the additional nuuients consumed by sows grazing the bahiagrass. Sire breed of the litter was not an important source of variation for litter size born or born alive or for litter weights at birth and 21 d (Tables 3 and 4). However, sire breed did influence the number of pigs in the litters at 21 d, 56 d and marketed. Sire breed also influenced weights of the litters at 56 d and litter weight marketed per day. In each case, H-sired litters had fewer pigs and weighed less than litters sired by D and Y boars. This agrees with previous work from this station (Kuhlers et al., 1981, 1982, 1988) that has shown that litter sizes and weights postfarrowing were reduced in H-sired litters compared with litters sired by other breeds of boars. These results also agree with work reported by Young et al. (1976), Nelson and Robison (1976) and Schneider et al. (1982). However, Drewry (1980) did not observe reduced litter sizes and weights in his work comparing litters sired by D, H and DH boars. The mechanisms by which litter sizes are reduced in H-sired litters have not been determined.

4 GENOTYPE-ENVIRONMENT INTERACTIONS IN SOWS 923 TABLE 2. LEAST SQUARES MEANS FOR GESTATION ENVIRONMENT Trait No. born No. born alive No. at21 d No. at 56 d No. marketed Litter birth wt, kg 21-d litter wt, kg 56-d litter wt, kg Litter wt per day, kg Sow wt, kg **P <.01. Gestation environment Pasture Stalls Difference :t: I "* :1: :l: :1: TABLE 3. LEAST SQUARES MEANS AND TESTS OF SIGNIFICANCE FOR THE LITTER SIZE TRAITS" Item NB NBA N21 N56 NM Sire breed Duroc (D) b 9.1 b 9.0 b Yorkshire (Y) *' 9.2 b 9.0 b Hampshire (H) " 8.2 ~ 8.0 ~ Dam breeding D-Landrace (L) YL IlL SE d "NB = number born; NBA = number born alive; N21 = number at 21 d; N56 = number at 56 d; NM = number marketed. bssire breeding least squares means with different superscripts differ (P <.10). daverage standard error for the least squares means in the column. TABLE 4. LEAST SQUARES MEANS AND TESTS OF SIGNIFICANCE FOR THE LITFER WEIGHT TRAITS Litter wt, kg LWPI~, Item Birth 21 d 56 d kg/d Sire breed Duroc (D) b~ b 5.3 b Yorkshire (Y) b b 5.1 b Hampshire (H) ~ r 4.5 ~ Dam breeding D-Landrace (L) 18.0 b YL 15.8 ~ HL 16.1 c SE d "LWPD = litter weight per day to 100 kg. b'gsite breeding or dam breeding least squares means with different su~ripts differ (P <.10). daverage standard error for the least squares means in the column. The three kinds of crossbred sows did not differ in their litter sizes at any of the stages of production up to market weight (Table 3). However, litter birth weights out of DL sows were significantly heavier than litter weights produced by YL and HL sows (Table 4). But

5 924 KUHLERS ET AL. TABLE 5. SIRE BREED x DAM BREEDING MEANS SQUARES FOR THE LITTER SIZE TRAITS a Source of variation df NB NBA N21 N56 NM Sire breed dam breeding * 47.4** 25.6* 27.5* 28.1" TBC-BC ~ ( 1 ) t 63.7* 74.4** 79.9* * Remainder (3) 44.1"* 49.2** Dam NB = number born; NBA = number born alive; N21 = number at 21 d; N56 = number at 56 d; NM = number marketed. ~I'BC = three-breed crosses, BC = backcrosses. 9 P < P < *P <.01. at 21 and 56 d, the differences among the three kinds of crossbred sows were small and not significant. Litter weight per day marketed also did not differ significantly among the three kinds of crossbred sows. The interaction effects of sire breed by dam breeding were significant for all traits except litter birth weight (Tables 5 and 6). Partitioning out the one degree of freedom contrast between three-breed cross litters and backcross litters, which estimates one-half of the individual heterosis, indicated that the comparison was significant for all the traits except litter size born and litter birth weight. The mean squares associated with the three remaining degrees of freedom were significant only for litter size born and born alive (Tables 5 and 6). These results would indicate that, except for litter size born and born alive, specific crosses of sire breeds and dam breeds do not affect litter sizes and weight in a non-additive fashion for the breeds of sires and breeding of dams used in this study. This, however, differed from the study by Kuhlers et al. (1982), who noted that litter sizes and weights at 21 and 42 d showed significant residual interaction effects after the removal of the sum of squares due to the three-breed cross vs backcross comparison. For most of the traits in the present study, the difference in individual heterosis between the three-breed cross litters and the backcross litters could explain the significance of the sire breed x dam breeding interaction source of variation. The heterosis estimates obtained from the differences in performance between the three-breed cross litters and the backcross litters, which estimates one-half of the individual heterosis (Tables 7 and 8), were similar to those reported by Wilson and Johnson (1981) and Kuhlers et al. (1982), but were larger than those reported by Sellier (1976), Johnson (1980) and Gaugler et al. (1984). One possible reason for the differences in these results was that smaller individual heterosis estimates were obtained using crossbred and purebred pigs TABLE 6. SIRE BREED x DAM BREEDING MEAN SQUARES FOR THE LITTER WEIGHT TRAITS Source of variation df Birth Litter wt, kg LWPIY, 21 d 56 d kg/d Sire breed x dam breeding * 13,179"* 18.7"* TBC-BC b (1) ,582* 50,918"* 65.7"* Remainder (3) Dam 193 (190) c , "LWPD = litter weight per day to marketing. btbc = three-breed crosses, BC = backcrosses. CThe number of degrees of freedom for LWPD. *P <.05. **P <.01.

6 GENOTYPE-ENVIRONMENT INTERACTIONS IN SOWS 925 TABLE 7. LEAST SQUARES MEANS FOR LITTER SIZE TRAITS a BY SIRE BREED AND DAM BREEDING Item NB NBA N21 N56 NM Sire breed x dam breeding Duroc (D) x D-Landrace (L) D x Yorkshire (Y)-L D x Hampshire (H)-L Y x DL Y x YL Y HL H DL H YL H x IlL SE b TBC-BC" 'NB = number born; NBA = number born alive; N21 = number at 21 d; N56 = number at 56 d; NM = number marketed. baverage standard error for the least squares means in the column. ~-TBC = three-breed crosses, BC = backcrosses. farrowed out of purebred sows, whereas the larger individual heterosis estimates were obtained from three-breed and backcross pigs that were out of two-breed crossbred sows. This would indicate that the amount of individual heterosis obtained may be influenced by the amount of maternal heterosis. Significant replicate x dam breeding interactions were noted for five of the nine litter traits studied (Tables 9 and 10). In replicate one, HL sows had larger litter sizes at 21 and 56 d and at marketing than DL and YL sows. However, in replicate two, significant differ- ences among the three kinds of crossbred sows did not exist. A similar result was found for 21-d litter weight. In addition, DL and YL sows had larger litter sizes born than HL sows in replicate two, but no significant differences were detected among the three sow breeds in replicate one. For litter birth weight the interaction was due to changes in the magnitude of the differences among the dam breeds, rather than to a lack of differences in one of the replicates. A significant replicate x dam breeding interaction was also found for litter weight per day marketed, but the mean separation technique employed could not dis- TABLE 8. LEAST SQUARES MEANS FOR LITTER WEIGHT TRAITS BY SIRE BREED AND DAM BREEDING Litter wt, kg LWPD a, Item Birth 21 d 56 d kg/d Sire breed x dam breeding Duroc (D) x D-Landrace (L) D x Yorkshire (Y)-L D x Hampshire (H)-L Y x DL Y x YL Y x IlL H x DL H x YL H x HL SE b TBC-BC ~ ~ = litter weight per day to 100 kg. baverage standard error for the least squares means in the column. CTBC = three-breed crosses, BC = backcrosses.

7 926 KUHLERS ET AL. TABLE 9. LEAST SQUARES MEANS FOR THE LITTER SIZE TRAITS" BY DAM BREEDING AND REPLICATE Item bib NBA N21 N56 NM Replicate x dam breeding 1 x Duroc-Landrace (DL) b 8.4 b 8.2 b 1 x York-Landrace (YL) b 8.3 b 8.0 b 1 Hamp-Landrace (HL) c 9.1 c 8.9 c 2 x DL 12.4 ~ x YL 12.8 b x IlL 11.9 ~ SE a NB = number born; NBA = number born alive; N21 = number at 21 d; N56 = number at 56 d; NM = number marketed. b'cdam breeding least squares means in a replicate with different superscripts differ (P <.10). ~Average standard error for the least squares means in the column. tinguish among the three dam breeds in either of the replicates. Sow Weight at Weaning of Fourth Litter. Weights of the sows differed among the three kinds of crossbred sows. The DL sows were significantly heavier at the end of four parities than were YL and HL sows, which did not differ significantly from each other (Table 11). The service sire breed did not have an influence on the sow's weight at the weaning of the fourth litter. However, it was noted that those sows that farrowed and reared threebreed cross litters for the four parities were 5.8 kg lighter than those sows that farrowed and reared backcross litters. This might be expected, because the size of the litters reared by sows farrowing three-breed cross litters was.6 to.7 pig/litter larger than those farrowing backcross litters. The mean squares associated with the remaining three degrees of freedom of the sire breed x dam breeding interaction were not significant. The three kinds of crossbred sows produced similar litter sizes and weights. However, length of life in the herd may be reduced in the case of YL sows, which could affect the total productivity of the sows originally brought into the herd (Jungst et al., 1988). The breed of sire significandy influenced the size and weight of the litter to marketing. The H sire breed reduced litter sizes by 1.0 pig/litter at marketing, which allowed D and Y sire breeds to produce 13 to 18% more litter weight per day to marketing. This would suggest that the H breed is the least desirable of the three sire breeds for siring crossbred litters. The gestation environment did not affect litter size and weight at any stage of production, nor did sow breeding interact with gestation environment. Although gestation en- TABLE 10. LEAST SQUARES MEANS FOR THE LITTER WEIGHT TRAITS BY DAM BREEDING AND REPLICATE Litter wt, kg LWPD', Item Birth 21 d 56 d kg/d Replicate dam breeding 1 Duroc-Landrace (DL) 16.2 b 42.0 bc x York-Landraee (YL) 14.8 c 41.1 b x Hamp-Landrace (HL) 15.4 c 45.4 c x DL 19.9 b x YL 16.7 ~ x IIL 16.9 ~ SEa "LWPD = litter weight per day to 100 kg. b'~dam breeding least squares means in a replicate with different superscripts differ (P <.10). ~Average standard error for the least squares means in the column.

8 GENOTYPE-ENVIRONMENT INTERACTIONS IN SOWS 927 Item TABLE 11. LEAST SQUARES MEANS FOR SOW WEIGHT AT WEANING OF FOURTH LITTER BY SIRE BREED AND DAM BREEDING OF THE LITTER Sow wt Sire breed Duroc (D) Yorkshire (y) Hampshire (H) Dam breeding D-Landraee (L) a YL b IlL b SE c 2.8 Sire breed x dam breeding D x DL D x YL D x HL Y x DL Y x YL Y x HL H x DL H x YL H x HL SE a 4.4 TBC-BC* * *'bdam breeding least squares means with different superscripts differ (P <.10). CAverage standard error for the sire breed and dam breeding least squares means. eaverage standard error for the sire breed x dam breeding least squares means. ~TBC = three-breed crosses, BC = backcrosses. *P <.10. vironment did not affect the performance of those sows on a litter farrowed basis, it did affect the total productivity of the three kinds of crossbred sows over four lactations (Jungst et al., 1988). The DL sows produced more total weight in four litters when kept in the pasture gestation system than when confined in gestation stalls. However, YL sows produced more total litter weight when they were kept in the gestation stall system rather than in the pasture system. A partial explanation was that longevity of the sows differed by breeding of sow; this is especially true for the DL sows, which had greater longevity when kept on pasture during gestation compared to being kept in confinement stalls. Additional studies need to be conducted with different crossbred sow types evaluated under different management systems. This may help explain some of the inconsistency of the results from crossbreeding studies previously completed. There was a suggestion that the amount of individual heterosis was influenced by the degree of maternal heterosis; therefore, studies need to be done to ascertain the importance of this interaction. Literature Cited Drewry, K. J Sow productivity traits of crossbred sows. L Anim. Sci. 50:242. Gaugler, H. R., D. S. Buchanan, R. L. Hintz and R. K. Johnson Sow productivity comparisons for four breeds of swine: Purebred and crossbred litters. J. Anita. Sci. 59:941. Harvey, W. R User's Guide for LSMLMW PC-1 Version. Ohio State Univ., Columbus. Holtmann, W. B., M. H. Fahmy, T. M. McIntyre and J. E. Moxley Evaluation of female reproductive performance of 28 one-way crosses produced from eight breeds of swine. Anita. Prod. 21:199. Johnson, R. K Heterosis and breed effects in swine. North Central Reg. Publ. No Johnson, R. K Crossbreeding in swine: Experimental results. J. Anim. Sci. 52:906. Jungst, S. B., D. L. Kuhlers and J. A. Little Longevity and maternal productivity of F 1 crossbred Landrace sows managed in two different gestation systems. Livest. Prod. Sci. 19:499. Kuhlers, D. L., S. B. Jungst, R. L. Edwards and J. A. Little Comparisons of specific crosses from Landrace, Duroc-Landrace and Yorkshire-Landrace sows. J. Anim. Sci. 53:40. Kuhlers, D. L., S. B. Jungst, R. L. Edwards and J. A. Little Comparisons of specific crosses of Duroc- Landrance, Spot-Landrace and Hampshire-Landrace sows. J. Anim. Sci. 55:236. Kuhlers, D. L., S. B. Jungst and R. A. Moore, Jr Comparisons of specific crosses from Yorkshire- Landrace, Chester White-Landrace and Chester White-Yorkshire sows. J. Anita. Sci. 66:1132. McLaren, D. G., D. S. Buchanan and R. K. Johnson Individual heterosis and breed effects for postweaning performance and carcass traits in four breeds of swine. J. Anita. Sci. 64:83. Nelson, R. E. and O. W. Robison Comparisons of specific two- and three-way crosses of swine. J. Anita. Sci. 42:1150. NSIF Guidelines for uniform swine improvement programs. USDA Ext. Service Program Aid 1157, Washington, DC. Schneider, J. F., L. L. Christian and D. L. Ktthlers Crossbreeding in swine: Genetic effects on litter performance. J. Anim. Sci. 54:739. SeUier, P The basis of crossbreeding in pigs; a review. Livest. Prod. Sci. 3:203. Whatley, J. A., Jr. and E. L. Quaife Adjusting weights of pigs to a standard age of 56 days. Proc. Am. Soc. Anim. Prod. 1937:126. Wilson, E. R. and R. K. Johnson Comparison of threebreed and backcross swine for litter productivity and postweaning performance. J. Anita. Sei. 52:18. Young, L. D., R. K. Johnson and I. T. Omtvedt Reproductive performance of swine bred to produce purebred and two-breed cross litters. J. Atom, Sei. 42:1133,