J. Anim. Sci. 1990. 68:1034-1043
© 1990 American Society of Animal Science
THE ILLINOIS SPRING BARROW SHOW: CARCASS TRENDS FROM 1968 TO 1989 AND USE OF CARCASS CHARACTERISTICS TO PREDICT DRESSING PERCENTAGE
D. G. McLaren1,
F. K. McKeith1,
T. R. Carr1 and
J. Novakofski1
University of Illinois, Urbana 61801
 |
Abstract
|
|---|
The Illinois Spring Barrow Show, a large annual carcass show of performance-tested pigs, may provide an indication of phenotypic trends in the pork industry over time. Objectives of this study were 1) to investigate time trends in live and carcass weights (LWT, CWT), dressing percentage (DP), length (LEN), loin muscle area (LEA), backfat (BF), muscle color, firmness and marbling (C,F,M) scores, estimated fat-standardized lean yield (FSL), FSL/d of age (LDOA) and FSL/d on test (LDOT) and 2) to assess use of carcass characteristics to predict DP. Data on 5,983 barrows exhibited between 1968 and 1989 were analyzed. Linear regressions expressed as change/yr (1977 to 1989, all P < .001) were .56 kg (LWT), .53 kg (CWT), .11 % (DP), .50 mm (10th rib BF), –49 g (wt-adj. FSL), 2.0 g (wt-adj. LDOA, 1977 to 1983) and –2.2 g (wt-adj. LDOT, 1984 to 1989). There were only small changes in LEN, C, F and M. Between 1968 and 1976, LEA increased .42 cm2/yr, then decreased .23 cm2/yr until 1984, increased .34 cm2/yr to 1988 and fell 1.2 cm2 in 1989 (all P < .001). Models including year, breed and their interaction (YB), ± linear and quadratic regressions on LWT, CWT, 10th rib BF, LEA and LEN, were used to assess the amount of variability in DP that could be accounted for by these carcass measures. R2 values (model effects in parentheses) were .16 (YB ± LWT), .28 (YB + CWT ± BF ± LEA) and .31 (YB + CWT + LEN ± [BF + LEA]). In conclusion, performance tested carcass show barrows have become heavier and fatter over time. Lean growth rate has decreased over the past 5 yr. Tenth rib fat and loin muscle area had no value as predictors of dressing percentage. Dressing percentage was not useful as a predictor of lean growth rate or lean yield.
Key Words: Pigs • Carcasses • Time • Trends • Prediction • Dressing Percentage
|
Introduction
|
|---|
Consumer demand for leaner meat products has increased over the past decade (Breidenstein and Carpenter, 1983
; Burke Marketing Research, 1987). Changes in consumer demand might be expected to affect production practices. There are few reports of phenotypic trends in growth and carcass composition of U.S. market pigs over time, however, and the present study reports trends for performance-tested carcass show barrows, which may or may not reflect the market hog population. The absence of such time trends in the U.S. reflects the lack of large-scale, organized performance testing programs, in contrast to programs in countries such as Canada (Kennedy et al., 1986), France (Tixier and Sellier, 1986) and Great Britain (Mitchell et al., 1982), where phenotypic and genetic changes in swine populations are more closely monitored.
Dressing percentage (yield) is a component of pricing procedures for hogs sold on a live basis. According to conventional wisdom fat hogs have a higher dressing percentage than thin hogs of the same weight (Bodus, 1989). In contrast, research by Kauffman et al. (1973, 1976, 1988) has indicated that muscling is a better predictor of dressing percentage than measures of fat. In either case, changes in the composition of pigs marketed might be expected to alter dressing percentage.
Our objectives were 1) to investigate phenotypic time trends for pig carcasses using historic data from a large annual carcass show and 2) to analyze these data to determine the relationship of carcass muscle and fat measures to dressing percentage. Specifically, we studied changes over time in carcass weight, dressing percentage, length, backfat measurements, loin muscle area, color, firmness and marbling scores, estimated fat-standardized lean yield (FSL), FSL/d of age and FSL/d on-test and use of carcass weight, length, loin muscle area and 10th rib fat to predict dressing percentage.
|
Experimental Procedure
|
|---|
Management and Data Collection.
Data on 5,983 performance-tested ("Premiere") barrows exhibited at the Illinois Spring Barrow Show between 1968 and 1989 were analyzed. The Illinois Spring Barrow Show began as an annual carcass show (i.e., all awards based on carcass data) in 1967. In 1968, 55 barrows were entered in the first "Illini Pork Premier Contest" of performance-tested pigs. Between 34 and 78 barrows were exhibited annually in this class from 1968 to 1976. Numbers then increased from 94 Premiere barrows in 1977 to peak at 758 barrows in 1987, the 1st yr that all barrows in the show were performance-tested.
Five hundred twenty performance-tested barrows were exhibited between 1968 and 1976 (8.8% of the data set), 2,190 were exhibited between 1977 and 1984 (36.6%), and 3,273 barrows (54.6%) were exhibited between 1985 and 1989 (Figure 1). The majority of barrows exhibited were crossbred (n = 3,727, 62.3%). The remainder were either of unknown breeds (n = 209; breeds not recorded in 1968, 1970, 1971 or 1972); or were purebred Hampshire (n = 650), Duroc (n = 543), Yorkshire (n = 386), Chester White (n = 145), Spotted (n = 113), Poland China (n = 113), Berkshire (n = 64) or Landrace (n = 33).
The 55 performance-tested barrows entered in the first Premiere contest were nominated and their ears were tattooed before they weighed 9 kg. Each exhibitor could nominate up to five pigs (increased to eight in 1973) and show one of the five (eight). Awards were made on the basis of an index of carcass weight/d of age, percent ham and loin in chilled carcass weight and loin muscle area.
In 1973 a "Test Station Premiere" class was added for pigs performance-tested at the Western Illinois Test Station, Macomb. Average (first rib, last rib and last lumbar vertebra) backfat was added to the index for ranking barrows in 1974. Nomination rules changed in 1975, and up to 10 barrows under 30 d of age (and a maximum allowable wt/d of age of .5 kg/d) could be tattooed per exhibitor, but each exhibitor was still only allowed to show one pig.
As of 1977, each exhibitor has been allowed to enter two barrows in the Test Station and one in the regular Premiere class out of 10 barrows nominated. Carcasses were ranked based on estimated (from carcass weight, 10th rib fat, loin muscle area and age at slaughter) days to produce 38.6 kg fat-standardized (10% lipid) lean starting in 1977.
Exhibitors could nominate 15 barrows starting in 1978, and two barrows per family firm or farm could be entered in the regular Premiere class starting in 1979. This was increased to up to 5 of 15 nominated barrows per family by 1989. Estimated lean gain/d on test replaced days to produce 38.6 kg lean as the method for ranking carcasses in 1981.
From 1984, only 10th and last rib backfat measurements were recorded (no first rib or last lumbar vertebra measurements) and color, firmness and marbling quality scores were reduced from 5- to 3-point scales (NPPC, 1983). All barrows exhibited in 1987 and subsequent years were tattooed at 35 d of age or younger and were eligible for the Premiere class.
Barrows were weighed and their ears were tattooed on test in August and September each year and show entrants were slaughtered at a commercial plant in late January or early February for carcass evaluation. Live weight was recorded 3 d prior to slaughter. Pigs were exhibited and evaluated for the next 1.5 d and then shipped to an abattoir and housed overnight with access to water until just prior to slaughter, which began at 0630. Carcass weight was recorded immediately following slaughter. Length, backfat and loin muscle area measurements and color, firmness and marbling scores were recorded within 24 h where CWT = hot carcass wt, kg; BF = 10th rib backfat, cm; LEA = 10th rib loin muscle area, cm2; ONTWT = on test wt, kg; DOT = days on test ONTWT mean = 8.7 ± .1 kg.
On test weights were not included in the data recorded between 1984 and 1986, although LDOT was calculated at the time of the carcass show (using the NPPC [1983] formula, developed from the same experimental data as equation 3) and entered in the database for all pigs from 1984 to 1987. The correlation between LDOT calculated using the Grisdale et al. (1984) formula vs the NPPC (1983) equation for 1,996 barrows slaughtered between 1987 and 1989 was .99.
Statistical Analysis
All analyses were conducted using Statistical Analysis System software (SAS, 1985). Linear time trends were estimated assuming GLM with breed (10 classes) as a fixed effect and time as a covariable in all models. Weight-adjusted regressions were calculated by including the linear and quadratic effects of live weight in models. Year least squares means obtained using GLM with breed and year as fixed effects were plotted.
Multiple regression procedures were used to examine the variation in dressing percentage explained by live and carcass weight, carcass length, 10th rib fat thickness and 10th rib loin muscle area. Models were compared based on coefficient of variation (R2) values and residual variances. Partial correlations (adjusted for postmortem. All data were collected by an experienced committee of extension livestock and meat specialists affiliated with the University of Illinois under the direction of an Illinois Meat Scientist. Carcass data used for analysis were for all barrows exhibited, including carcasses disqualified from competition due to excessive fatness or quality problems.
Fat-standardized lean yield (FSL; i.e., lean adjusted to a 10% lipid content), fat-standardized lean gained/d of age (LDOA) and fat-standardized lean gained/d on test (LDOT) were estimated using equations developed by Grisdale et al. (1984). Equations were:
 | (1) |
 | (2) |
 | (3) |
breed, year and breed x year interaction effects) between carcass and estimated growth characteristics also were computed.
|
Results
|
|---|
Average live weight of show barrows from 1977 to 1989 was 108 kg, with a mean dressing percentage of 73.9% (Table 1). The average barrow had a lean growth rate (LDOT) of 286 g/d on test (from 8.7 kg live weight for 133.3 d) and had a carcass that was 79.8 cm long and yielded 42.2 kg fat-standardized lean (FSL). Mean loin muscle area was 32.7 cm2 and mean 10th rib fat (TRF) was 28.7 mm (Table 1).
Live weight of barrows increased at an average rate of .56 kg/yr between 1977 and 1989 (Table 2). Average live weight increased from 101 kg in 1977 to peak at 110 kg in 1986 and then decreased to 105 kg in 1989 (Figure 2). Carcass weights paralleled live weights. Dressing percentage increased approximately .1% per annum, from 72.9% in 1977 to average 74.1% between 1986 and 1989, virtually independent of changes in live weight (Table 2, Figure 2).
Estimated fat-standardized lean yield (Table 2
, Figure 3) increased in absolute terms between 1977 and 1989 but decreased slightly (–49 g/yr, approximately .1% of the mean value) on a constant live weight basis. Lean/d of age (Table 2, Figure 3), calculated for 1,802 barrows between 1977 and 1983 increased by 2 g·d–1·yr–1 (.8% of the mean) on average (adjusted for live weight differences), Lean/d on test, however, estimated for 3,648 barrows exhibited between 1984 and 1989, decreased by approximately 2 g·d–1·yr–1.
Adjusting linearly and quadratically for weight differences, 10th rib backfat (used as a predictor of carcass fat) showed a .34 mm (1.2% of the mean) linear increase per annum, from a mean of 24.7 mm in 1977 to 30.2 mm in 1989 (Table 2, Figure 4). Last rib backfat also was measured from 1984 and changed in a similar fashion to 10th rib fat (Figure 4). Last rib fat increased by .61 mm/yr on average, although it might be argued that backfat peaked in 1987 and is now decreasing. The regression of last rib fat on year was 1.49 mm/yr from 1984 to 1987 and –1.09 mm/yr from 1987 to 1989 (Table 2).
Carcass length increased over the period 1968 to 1989, but analysis of 1977 to 1989 data revealed only a small, nonsignificant rate of increase over this period (Table 3, Figure 4). Loin muscle area changed erratically over time (Table 3, Figure 4). Cross-sectional area of the longissimus muscle at the 10th rib increased from a mean of 30.2 cm2 in 1968 to 34.2 cm2 in 1976 but then fell to 31.0 cm2 in 1984. Mean loin muscle area increased annually from 1984 to 1988 (32.3 cm2), but it fell to 31.1 cm2 in 1989 (P < .01).
The 1968 to 1976 rate of increase in loin muscle area was approximately .42 cm2/yr (1.3% of the mean), not inconsistent with expectations for genetic progress assuming selection on an index of economically important traits (Smith, 1984). The sharp decline in the late 1970s coincides with concerns regarding porcine stress syndrome, an autosomal recessive condition associated with, although not causally related to, carcass leanness (Christian, 1972). This probably reflects selection against heavily muscled hogs, although the phenotypic trends reported here cannot be partitioned into genetic and environmental components. In 1975, a maximum allowable wt/d of age of .5 kg/d was imposed for nomination.
Loin muscle area increased by approximately 1% of the mean per year from 1984 to 1988 but decreased in 1989. The linear regression of loin muscle area on year from 1984 to 1989 was not different from zero (P > .05). Color, firmness and marbling scores, although showing some statistically significant trends, indicated that changes in pork quality over time were negligible (Table 3, Figure 5). The small changes in color firmness, despite dynamic changes in loin muscle area, indicate that improving muscling or leanness does not necessarily increase the incidence of porcine stress syndrome and PSE pork.
View larger version (15K):
[in this window]
[in a new window]
|
Figure 5. Mean color, firmness and marbling scores by year. Scores changed from a 5- to a 3-point scale in 1984.
|
|
Partial correlations among carcass characteristics are presented in Table 4. Dressing percentage was lowly (.1 to .2) correlated with carcass length, backfat or loin muscle area. Correlations between dressing percentage and lean yield and lean growth rate were approximately .3. Length and loin muscle area were moderately to highly (.5 to .7) correlated with estimates of lean yield and growth rate, whereas 10th rib fat was lowly (0 to –.1) correlated with the same estimates.
Muscle quality scores were moderately (.2 to .5) correlated with each other on both the 3- and 5-point scales (Table 5). Correlations between color, firmness or marbling scores and live weight, dressing percentage, carcass length, lean yield or lean growth rate generally were close to zero. Loin muscle area was lowly and negatively correlated (approximately –.1) with all quality scores. Tenth rib backfat was lowly and positively correlated with quality scores on the 5-point scale (. 1 to .2) but essentially uncorrelated with scores on the 3-point scale.
Prediction equations for dressing percentage, including the effects of year, breed and the year x breed interaction as well as linear and quadratic regressions on live or carcass weight, carcass length, 10th rib backfat and loin muscle area, had coefficient of variation (R2) values ranging from 16 to 31% (Table 6). Year and breed effects alone accounted for 16% of the variation in dressing percentage. Adding live weight to the model failed to significantly increase the R2 value; adding carcass weight increased the R2 to 28%. For the prediction model including live weight, loin muscle area was the independent variable that resulted in the greatest increase in R2 value, from 16 to 19%. Including both loin muscle area and backfat increased the R2 to 23%, and adding length increased it slightly more to 25%. For the model including carcass weight, length was the independent variable resulting in the greatest increase in R2 value, from 28 to 31%. Adding loin muscle area and backfat failed to significantly increase the R2 value for this model.
To further examine the relationship between dressing percentage and lean growth, linear and quadratic effects of dressing percentage were included in models to predict estimated lean yield and lean growth rate (Table 7). Dressing percentage was not particularly useful as a predictor of fat-standardized lean yield, increasing the R2 from 10 to 21% over a prediction model containing only year and breed effects. Similarly for lean growth rate, dressing percentage increased the R2 for lean/d of age from 16 to 24% and from 12 to 19% for lean/d on test.
|
Discussion
|
|---|
Historical data inherently are variable because different individuals make measurements over time and, in the case of carcass data, measurements may differ as methodology improves. These problems have been minimized in the data set used for this study. The measuring process has been supervised by only two individuals over the time span 1968 to 1989, resulting in consistency of procedures across years. The types of data collected for placing pigs based on carcass merit have allowed calculation of comparable indicators of performance and carcass composition over much of the period analyzed. Furthermore, the large number of animals for which data are available allows reliable estimation of time trends.
Failure of the pork industry to consistently improve economically important characteristics such as lean yield, lean growth rate and backfat thickness of market hogs is disappointing, particularly because rates of genetic improvement in these traits of approximately 2% of the mean per year have been achieved by industry programs elsewhere (Mitchell et al., 1982), and even higher rates are possible theoretically (Smith, 1984). For example, the current rate of genetic change in backfat thickness for pigs in Ontario, Canada, has been reported to be –.2 mm/yr (Kennedy, 1987) whereas weight-adjusted 10th rib fat increased at a rate of .3 mm/yr in our dataset.
Canadian commercial hog carcasses showed a 10% reduction in backfat, equivalent to an increase of .97 kg trimmed retail product per carcass, between 1969 and 1982 (Fredeen, 1984). These results were based on between 7.5 and 12.9 million carcasses per year. Carcass grading became mandatory in Canada in 1944, and in 1968 an indexing system using carcass weight and backfat measurements to predict yield of trimmed cuts was adopted (Fredeen, 1984). Fredeen (1984) speculated that the prolonged linear time trends for index and backfat might be explained by progressive genetic improvement.
The barrows shown at the Illinois carcass show are selected for the show and thus do not represent a random sample of market hogs. Exhibitors selected up to 5 of 15 nominated (tattooed) barrows to show in the Premiere class. With awards based on estimated lean gain/d on test, and with availability of ultrasonic probe and growth data to aid in selection, the barrows slaughtered for the contest might be expected to be better than "average" market hogs. The trends toward increased backfat and decreased lean/d on test might be greater in commercial pigs.
Analysis of the barrow show data clearly does not support the conventional wisdom that fatter hogs have a higher dressing percentage. Indicators of empty body size (carcass weight) or indicators of muscling (loin muscle area) appeared slightly indicative of dressing percentage, as previously suggested by Kauffman et al. (1973, 1976) for both pigs and cattle.
Kauffman et al. (1988), in an experiment involving 24 pigs representing "four distinct and extreme combinations of muscling and fatness," concluded that degree of fatness had little effect on dressing percentage adjusted for sex, weight and gut-fill differences. However, pigs classified as "mesomorphs" had 3% heavier carcasses, again adjusted for sex, weight and gut-fill, than did "ectomorph" pigs. Pigs in our study did represent extreme types, and, as such, our results might not be expected to apply to "typical" pigs.
In the present study neither muscling nor fatness was particularly useful for predicting dressing percentage. It is unlikely that the poor statistical relationship between these parameters is the result of inadequate variation in the data because both muscling and fatness were fairly variable (CV of 13 to 25%). Dressing percentage, however, had a CV of only 3.2% (Table 1). This, and the correlation between dressing percentage and live weight of only. 1 (Table 4), suggests that visceral size is proportional to body mass in pigs. This appears to be true in many other species (Schmidt-Nielsen, 1984) and makes physiological sense because a given weight of organ in the viscera likely is able to support only a fixed amount of body mass.
Proportional growth of viscera and body mass would also be suggested by the low R2 for lean gain using dressing percentage as a predictor variable. This would be the case if lean and viscera grew at similar rates and thus remained proportional. Kauffman et al. (1986) indicated that, in swine near market weight, viscera tend to grow slightly slower than the whole animal. This relationship does not appear to be true in cattle (Kauffman et al., 1976), possibly because of large variation in muscle mass relative to other tissues, or because mass of the ruminant digestive tract increases sub-proportionally to the rest of body mass.
Measurement of the amount of ingesta (gut-fill) was not included in the currently reported data set. However, the inability to explain a large proportion of dressing percentage variation by variation in carcass characteristics suggests that gut-fill is the most important factor in determining dressing percentage in swine.
In conclusion, performance-tested barrows exhibited at the Illinois Spring Barrow Show have become heavier and fatter over time. Loin muscle area increased between 1968 and 1976, decreased between 1976 and 1984 and increased again between 1984 and 1988, only to fall in 1989. Lean growth rate decreased at a rate of approximately .8% of the mean a year over the past 4 yr. Backfat thickness and loin muscle area had no value as predictors of dressing percentage, itself a poor predictor of lean yield and growth rate.
|
Implications
|
|---|
Time trend analysis of swine from a carcass show over the past 32 yr show no consistent carcass quality improvement but a trend toward fatter pigs and slightly slower lean growth rates. This may reflect the failure of packers to penalize fat hogs and the failure of seedstock suppliers to make consistent genetic improvement in these traits. Large-scale breeding programs are required for such improvement, as are accurate whole-herd performance records and a technically sound, consistent selection program for traits of economic importance. Such seedstock suppliers exist, but they request a premium for their breeding stock. The commercial producer fails to evaluate savings in feed from producing a leaner pig and buys inexpensive boars. Without packer incentives, as exist in countries making genetic progress, improvement in carcass composition is likely to be erratic.
|
Footnotes
|
|---|
1 Dept. of Anim. Sci. 
Received for publication May 1, 1989.
Accepted for publication July 27, 1989.
|
Literature Cited
|
|---|
Bodus, T. 1989. Dressing percentage: Fat and fiction. Pork’89. February. p 74.
Breidenstein, B. C. and Z. L. Carpenter. 1983. The red meat industry: Product and consumerism. J. Anim. Sci. 57 (Suppl. 2):119.
Burke Marketing Research. 1987. Consumer climate for meat. Executive summary prepared for the National Live Stock and Meat Board and the American Meat Institute.
Christian, L. L. 1972. A review of the role of genetics in animal stress susceptibility and meat quality. In: Cassens, R., F. Giesler and Q. Kolb (Ed.) Proc. Pork Quality Symp. pp 91–115. Univ. Wisconsin, Madison.
Fredeen, H. T. 1984. Changes in the characteristics of commercial hog carcasses in Canada. Can. J. Anim. Sci. 64:569.
Grisdale, B., L. L. Christian, H. R. Cross, D. J. Meisinger, M. F. Rothschild and R. G. Kauffman. 1984. Revised approaches to estimate lean of pork carcasses of known age or days on-test. J. Anim. Sci. 58:335.[Abstract/Free Full Text]
Kauffman, R. G., T. D. Crenshaw, J. J. Rutledge, D. H. Hull, B. S. Grisdale and J. Penalba. 1986. Porcine growth: Postnatal development of major components in the boan. Univ. of Wisconsin-Madison Coll. Agric. Life Sci. Res. Rep. R3355.
Kauffman, R. G., R. H. Grummer, R. E. Smith, R. A. Long and G. Shook. 1973. Does live-animal and carcass shape influence gross composition? J. Anim. Sci. 37:1112.
Kauffman, R. G., R. L. Russell and S. W. Farwell. 1988. Students learn how muscling and fatness affect swine dressing percentage. J. Anim. Sci. 66 (Suppl.1):121 (Abstr.).
Kauffman, R. G., M. D. Van Ess and R. A. Long. 1976. Bovine compositional interrelationships. J. Anim. Sci. 43: 102.[Abstract/Free Full Text]
Kennedy, B. W. 1987. Selection for and prediction of efficient lean tissue growth. In: Proc. Natl. Swine Improvement Fed. Meetings. NSIF 12:85.
Kennedy, B. W., G.F.S. Hudson and L. R. Schaeffer. 1986. Evaluation of genetic change in performance tested pigs in Canada. Proc. Third World Congr. Genet. Appl. Livest. Prod. X:149.
Mitchell, G., C. Smith, M. Makower and P.J.W.N. Bird. 1982. An economic appraisal of pig improvement in Great Britain. 1. Genetic and production aspects. Anim. Prod. 35:215.
NPPC. 1983. Procedures to evaluate market hog performance. Natl. Pork Prod. Council, Des Moines, 1A.
SAS. 1985 SAS User’s Guide: Statistics. SAS Inst., Inc., Cary, NC.
Smith, C. 1984. Rates of genetic change in farm livestock. J. Agric. Res. Dev. 1:79.
Schmidt-Nielson, K. 1984. Scaling. Why is animal size so important. Cambridge Univ. Press, New York.
Tixier, M. and P. Sellier. 1986. Estimated genetic trends for growth and carcass traits in two French pig breeds. Genet. Sel. Evol. 18:185.
Top
Abstract
Introduction
