Correlated responses in growth, carcass, and meat quality traits to divergent selection for testosterone production in pigs 1

Published December 8, 2014 Correlated responses in growth, carcass, and meat quality traits to divergent selection for testosterone production in pigs 1 J. M. Bender,* M. T. See,* D. J. Hanson, T. E. Lawrence, and J. P. Cassady* 2 *Department of Animal Science, North Carolina State University, Raleigh 27695-7621; Department of Food Science, North Carolina State University, Raleigh 27695-7624; Division of Agriculture, West Texas A&M University, Canyon 79016 ABSTRACT: The objective of this project was to characterize changes in growth, carcass yield, and meat quality traits in castrates and gilts in response to divergent selection for testosterone production. In generation 21, endogenous testosterone concentrations in Duroc boars of the high (HTL) and low (LTL) testosterone lines averaged 49.0 and 27.8 ng/ml (P < 0.01), respectively. Eight LTL and 10 HTL boars were used to sire 29 LTL and 33 HTL litters. To remove the effects of inbreeding, these same boars were mated to females of a Large White Landrace composite (WC) to generate 11 WC by LTL litters (WLT) and 23 WC by HTL litters (WHT). Castrates and gilts were then allotted to LTL (n = 53), HTL (n = 61), WLT (n = 102), and WHT (n = 101) for testing. Growth and carcass traits analyzed included days to 114 kg (D114), ADG, backfat adjusted to 114 kg (ABF), LM area adjusted to 114 kg and predicted percent lean (PPL). Fat-O-Meater data collected were adjusted fat depth (AFD), adjusted loin depth, and percent lean. Meat quality traits characterized at 24 h postmortem included marbling score, percent lipid, ph, drip loss, color score, and Minolta L*, a*, and b*. Data were analyzed with a mixed model including fixed effects of line, mating type (purebred or crossbred), sex, and the random effect of sire nested within line. All possible interactions among fixed effects were tested. The HTL had fewer D114 (P < 0.05), greater ADG (P < 0.01), greater ABF (P < 0.01), and lower PPL (P < 0.01) than LTL. The WHT and WLT did not differ for D114, ADG, or ABF. The WHT had smaller LM area adjusted to 114 kg (P < 0.05) and greater drip loss (P < 0.05) than WLT. The WLT had lower adjusted loin depth (P < 0.05) than LTL and HTL. The LTL and HTL had greater subjective scores for marbling (P < 0.05) compared with WLT and WHT. The least squares mean for percent lipid for HTL and LTL was 4.00. The WHT had greater means for L*, a*, and b* (P < 0.05) than WLT. Pigs selected for increased testosterone production grew faster and produced fatter carcasses than pigs selected for decreased testosterone. Changes in growth, carcass yield, and meat quality traits were detected in castrates and gilts in response to divergent selection for testosterone production. Key words: growth, meat quality, pig, testosterone 2006 American Society of Animal Science. All rights reserved. J. Anim. Sci. 2006. 84:1331 1337 INTRODUCTION The swine industry strives to produce profitable pigs. Nutrition, management, and exogenous hormones can help a pig reach its maximum genetic potential. Genetic selection must be used to improve that maximum genetic potential. Selection is practiced by identifying and selecting superior animals for growth potential, carcass yield, and meat quality. Many of these traits are moder- 1 Funding for this project was provided through the North Carolina Agriculture Experiment Station and North Carolina Department of Agriculture. 2 Corresponding author: joe_cassady@ncsu.edu Received June 17, 2005. Accepted January 9, 2006. ately to highly heritable, and response to selection is dependent on variation in the population and selection intensity (Jensen et al., 1967; Falconer, 1989). Response to selection is also influenced by correlations among traits, which may be desirable or undesirable (Jensen et al., 1967; Lo et al., 1992). Detecting these associations can provide information that leads to greater insight with regard to biological interactions. It has been established that physiological differences exist among castrates, intact males, and females. Intact males grow faster and are leaner than females and castrates. In addition, castrates are fatter than females (Cassady et al., 2004). Differences are due to presence and absence of endogenous hormones, and positive correlations have been established between testosterone production and growth (Lubritz et al., 1991). Robison et al. (1994) found high testosterone boars had greater 1331

1332 Bender et al. Table 1. Distribution of animals available by line, mating type, sex, and sex within slaughter date Slaughter date Oct. 2, 2003 Nov. 6, 2003 Line 1 Mating type 2 n Castrates Gilts Castrates Gilts Low line Purebred (LTL) 53 1 1 20 31 Crossbred (WLT) 102 52 50 0 0 High line Purebred (HTL) 61 2 1 19 39 Crossbred (WHT) 101 22 21 28 30 1 Line = The low line was selected for decreased testosterone, and the high line was selected for increased testosterone. 2 Mating type = purebred or crossbred pigs [purebred (Large White Landrace composite)]. LTL = low testosterone purebreds; WLT = low testosterone crossbreds, HTL = high testosterone purebreds; and WHT = high testosterone crossbreds. ADG. In some species, supplying exogenous testosterone and estrogen to females and castrates improves their testosterone production to a level comparable with intact males (DeWilde and Lauwers, 1984; Montgomery et al., 2001). These sex steroids are similar in structure, share the same steroidogenic pathway, and are controlled by the same genes (Land, 1973). Consequently, it is possible that selection for endogenous testosterone in intact males will alter the physiology of females and castrates. The objective of this project was to characterize changes in growth, carcass yield, and meat quality in castrates and gilts in response to divergent selection for testosterone production. MATERIALS AND METHODS This project was approved by the NCSU Institute Animal Care and Use Committee. Population Structure Data were collected from 2 Duroc lines of pigs maintained at the North Carolina Department of Agriculture Tidewater Research Station, Plymouth. These lines were divergently selected for 10 generations for testosterone production and then maintained by random within line selection (Robison et al., 1994). In generation 21, the endogenous testosterone concentration in the high (HTL) and low (LTL) testosterone purebred lines averaged 49.0 and 27.8 ng/ml (P < 0.01), respectively (Walker et al., 2004). Eight LTL and 10 HTL boars were used to generate 29 LTL and 33 HTL litters. The same boars were mated to a common line of Large White Landrace composite (WC) females, which were previously selected for increased litter size (Holl and Robison, 2003) to create 11 WC LTL litters (WLT), representing 7 LTL sires and 23 WC HTL litters (WHT), representing 7 HTL sires. Available castrates and gilts, LTL (n = 53), HTL (n = 61), WLT (n = 102), and WHT (n = 101), were selected for this study (Table 1). Management Animals were reared in total confinement. At farrowing, pigs were crossfostered to standardize litter size. At approximately 28 d of age, pigs were weaned and transferred to the nursery. At 56 d of age, pigs were transferred to the finishing floor, and groups of 10 pigs per pen (0.93 m 2 /pig) were allocated according to size and sex. Pigs were allowed ad libitum access to diets that met or exceeded minimum NRC nutritional recommendations (NRC, 1998). Data Collection Pigs were weighed at birth, 21 d of age, and 48 h before slaughter. Average daily gain was calculated based on age and weight for the interval between 21 d age and 48 h before slaughter: (ADG = [{Actual wt 21 d wt}/{age at slaughter 21 d}]). Days to 114 kg (D114) was estimated using age, ADG, and final weight deviation from the 114 kg target: (D114 = Actual age + [{114 actual wt}/adg]). Pigs were scanned on the right side by a certified ultrasound technician 48 h before slaughter using real-time ultrasound (Aloka, Ithaca, NY). Adjusted backfat (ABF) and adjusted LM area (ALMA) were measured and adjusted to 114 kg, and percent lean (PPL; NSIF, 2003) was predicted: (PPL = [80.95 {16.44 ABF, in.} + {4.693 ALMA, in. 2 }] * 0.54). Pigs weighing a minimum of 90 kg were transported 333 km to a commercial slaughter facility, arriving at least 3 h before slaughter. The Fat-O-Meater optical probe (SFK Technology, AIS, Denmark) was used to measure fat depth (AFD) and loin depth (ALMD) within approximately 40 min of slaughter. Fat-O- Meater measures were taken through a section of LM between the 10th and 11th rib, inlet 7 cm from the midline and exit 4 cm from the mid-line split. In addition, HCW was recorded online and percent lean (%LEAN) was predicted. Carcasses proceeded through a snapchilling stage followed by an equilibration stage, resulting in a chilled muscle temperature of approxi-

Selection for testosterone in pigs 1333 mately 3 C at 24 h postmortem. Of the 117 purebred Duroc carcasses, 95 were diverted for hair by USDA inspectors during slaughter onto a side rail. For the crossbred pigs, which had white hair, only 37 of 203 carcasses were diverted. When carcasses were diverted, the interval from stunning to cooling was extended. The length of time by which the interval was extended was inconsistent. Fifty-two of the detained carcasses were sequestered for the remaining data collection. At 24 h postmortem loin chops (n = 237, 3.5 cm thick) located at the right side between the 9th and 10th rib were collected. Chops were exposed to 0 to 1.1 C air to allow for oxygenation of myoglobin for a minimum of 10 min, as recommended by the NPPC (2000). Loin muscle marbling score (MAR: scale 1 to 10; 1 to 10% intramuscular fat) and color score (CS: scale 1 to 6; 1 pale gray to 6 dark purplish red) were visually accessed by a trained meat scientist based on standards published by the NPPC (2000). A handheld Minolta CM-508d spectrometer (Azuchi-Machi, Chuo- Ku, Osaka, Japan) was used to measure L*, a*, and b* for each chop. Scores were based on the average of 3 replicate measures on each sample. In addition, ph was measured using a MPI-100 ph meter (Meat Probes, Inc., Topeka, KS). Upon completion of data collection, chops were packaged and transported to North Carolina State University. At NCSU, a 48-h drip loss test was conducted to determine drip loss percentage (DRIP; Honikel, 1987). Remaining loin samples were frozen at 27 C for approximately 7 mo. Sixty-five samples (LTL: n = 28; HTL: n = 37) were analyzed for percent lipid (%LIPID) using the Folch method (Folch et al., 1957). Statistical Analysis Traits were analyzed using the GLM and MIXED procedures of SAS (SAS Inst. Inc., Cary, NC). A preliminary analysis was conducted to determine the significance of slaughter date. Growth and carcass traits were not evaluated for the significance of slaughter date due to confounding effects of genetic makeup, which influenced when the pigs reached market weight. Meat quality traits, DRIP, ph, CS, L*, a*, and b*, were analyzed in a model including fixed effects of slaughter date and sex. Slaughter date was significant for objective color scores and these traits were adjusted to a second slaughter-date equivalent as follows: L*: Initial L* 2.00767035; a*: Initial a* 0.737419031; and b*: Initial b* 0.869719325. The remaining analyses encompassed variations of 2 models that accounted for all necessary comparisons. Tests for fixed effects of traits for all models are given in Table 2. Fixed effects with P > 0.05 were not included in the final model. The first model compared differences between LTL, HTL, WLT, and WHT pigs. The model included fixed effects for line (low and high), mating type (MT: purebred and crossbred), sex (females and castrates), all 2- way interactions, and the random effect of sire nested within line. To adjust the data to either an off-test Table 2. Test for fixed effects for growth, carcass, Fat-o- Meater, and meat quality traits Fixed effect 1 Trait 2 Line 3,4 MT 3 Sex 3,4 Line MT 3 D114, d 0.03 0.01 0.01 0.11 ADG, k/d 0.01 0.01 0.01 0.14 ABF, mm 0.01 0.04 0.01 0.01 ALMA, cm 2 0.15 0.01 0.01 0.13 PPL, % 0.01 0.01 0.01 0.02 AFD, mm 0.29 0.33 0.01 0.77 ALMD, mm 0.39 0.01 0.01 0.99 %LEAN 0.21 0.41 0.01 0.90 MAR 0.61 0.01 0.01 0.92 %LIPID 0.87 0.18 ph 0.85 0.09 DRIP, % 0.01 0.01 0.27 0.84 CS 0.25 0.03 L* 0.04 0.37 a* 0.01 0.22 b* 0.03 0.43 1 Line = Fixed effect of pooled high and low lines across purebreds and crossbreds, MT = fixed effect of mating type (purebred vs. crossbred); Sex = fixed effect of sex; and Line MT = interaction effect. 2 D114 = Days to 114 kg; ABF = real-time ultrasound backfat adjusted to 114 kg; ALMA = real-time ultrasound LM area adjusted to 114 kg; PPL = predicted percent lean; AFD = Fat-O-Meater fat depth; ALMD = Fat-O-Meater loin depth; %LEAN = Fat-O-Meater lean percentage; MAR = Marbling score (scale 1 to 10; 1 to 10% IMF); %LIPID = percent lipid determined by Folch method; ph = ph 24 h postmortem; DRIP = average collected purge from 40 to 50 g samples suspended 48 h (sample wt 100); CS = color scale (scale 1 to 6; 1 pale gray to 6 dark purplish red); L* = lightness color scale (range 0 to 100; 1 = pure black and 100 = pure white); a* = red to green color scale (greater value = redder color); and b* = blue to yellow color scale (greater value = more yellow). The MAR and DRIP include low and high line purebred and crossbred pigs; %LIPID includes low and high line purebred pigs; ph, CS, L*, a*, and b* include low and high line crossbred pigs. 3 Fixed effects included in model 1. 4 Fixed effects included in model 2. weight of 114 kg or an HCW of 80 kg, live weight and HCW were included as regression covariates for growth and Fat-O-Meater measurements, to obtain adjustment factors for unadjusted backfat, loin eye area, fat depth, and loin depth. Sex-specific adjustments were used for ABF. Adjustment equations were: ABF = unadjusted BF ([wt 114] * 0.26249149), (castrates); ABF = unadjusted BF ([wt 114] * 0.26219149) (wt 0.10391802), (gilts); ALMA = unadjusted LMA ([wt 114] * 0.17263813); AFD = ([80 HCW] * 1.30778547) + ([6400 HCW 2 ] 0.00613256) + unadjusted FD; and ALMD = ([80 HCW] * 0.32774811) + unadjusted LMD. Meat quality traits, including, ph, CS, L*, a*, b*, and %LIPID, were analyzed with the second model. The model included fixed effects of line (low and high), sex (females and castrates), and the random effect of sire nested within line. Detainment of carcasses affects meat quality traits of the LM (D Souza et al., 1998).

1334 Consequently, it was imperative that the traits be evaluated to determine if prolonged elevated temperature significantly affected DRIP, ph, CS, L*, a*, and b*. All traits except DRIP were influenced by prolonged elevated temperature. Because primarily LTL and HTL animals were retained, their observations were unduly altering the interpretation of the results. Removal of all LTL and HTL pigs eliminated the confounding effect; therefore only data from crossbreds were used. The %LIPID was only measured on LTL and HTL pigs. Interaction Effects RESULTS Tests for fixed effects are in Table 2. All growth traits were influenced by MT and sex (P < 0.01). Line influenced D114, ADG, ABF, PPL, DRIP, L*, a*, and b* (P < 0.05). Line MT significantly influenced ABF and PPL, and there was a tendency for it to influence D114. Line sex influenced ABF (P < 0.03). In all cases castrates were fatter than gilts; however, the extent to which they were fatter varied enough to be statistically significant. This interaction was not considered to be biologically relevant. In addition, MT sex only influenced D114 (P < 0.02). No significant interactions or line effects were observed for the Fat-O-Meater traits. In addition, MT did not influence AFD or %LEAN. Line was not significant for MAR but did influence DRIP, L*, a* and b* (P < 0.05). Sex did not affect DRIP, L*, a* or b*, but there was a tendency for sex to influence ph. Growth and Carcass Traits Least squares means and standard errors of the difference for all traits are in Table 3. The LTL had greater D114 and lower ADG than HTL. The WLT and WHT grew faster than LTL and HTL. Compared with all groups, HTL had the greatest ABF, followed by WHT, which were fatter than LTL but not significantly different from WLT. The WLT had larger ALMA than all others. In addition, LM area for WHT was larger than LTL and HTL. Predicted percent lean was greater in WLT than LTL and HTL. There was no significant difference between LTL and WHT for PPL, but both were greater than HTL. Castrates had greater ADG, but gilts had lower ABF, larger ALMA, and greater PPL. Purebred gilts (236 ± 2.2) had greater D114 than purebred castrates (220 ± 2.8), crossbred castrates (177 ± 1.9), or crossbred gilts (184 ± 1.9; P < 0.05). Fat-O-Meater Traits The ALMD of HTL and LTL was greater than WLT, and there were no significant differences among HTL, LTL, and WHT. Gilts had lower AFD, greater ALMD, and greater %LEAN. No significant differences were found for dressing percentage (HCW/live weight). Bender et al. Meat Quality Traits Greater MAR was recorded in LTL and HTL than WLT and WHT, and %LIPID was not significantly different between LTL and HTL pigs. In addition, DRIP did not differ significantly between HTL and LTL, but DRIP was lower in WLT than LTL, HTL, and WHT. Means were not different for ph or CS between WLT and WHT pigs. Minolta L*, a*, and b* values were lower in WLT than WHT. There were no differences between castrates and gilts for %LIPID, ph, DRIP, L*, a*, or b*. For castrates, MAR and CS were greater. DISCUSSION In this study, responses to divergent selection for testosterone production were evaluated. Only castrates and gilts were evaluated. Observed changes in growth and backfat are correlated responses to selection. Characterizing effects of divergent selection for testosterone provides a better understanding of growth, carcass yield, and meat quality. Further, it provides insight into effects of such selection across sexes. There were 3 significant interactions among fixed effects for growth traits. Rate of gain and type of gain were dependent on testosterone line, mating type, and sex. Combinations of these categories alter rank or magnitude of difference among groups. Testosterone production improves rate of growth (Lubritz et al., 1991). In addition, purebred pigs in this project were influenced by inbreeding depression, which often impedes performance. In this experiment, LTL pigs grew slower and had less backfat than HTL. This is in agreement with a previous study, which concluded that selection for increased testis size resulted in an increase in testosterone, body weight, and backfat (Schinckel et al., 1984). In addition, Lubritz et al. (1991) determined that growth and testosterone levels were positively correlated. Typically, testosterone production is associated with lean growth in boars. Also exogenous testosterone administered to animals alters carcass composition (Montgomery et al., 2001). Thus, increasing testosterone concentration should improve rate of muscle accretion (Snochowski et al., 1981). The specific biological mechanisms by which divergent selection for testosterone in males altered growth and backfat in females and castrates in this study has yet to be determined. However, testosterone and estrogen share the same steroidogenic pathway where testosterone is a precursor of estrogen. Testosterone is a steroid hormone synthesized from cholesterol through a series of 5 enzymatic reductions (Zubay, 1993). Therefore, it is possible that the steroidogenic pathway may have been altered in both males and females. This would be expected to result in an increase or decrease in estrogen. However, this does not explain observed changes in castrates. Exposure to different levels of testosterone during prenatal or early postnatal periods may have had an effect on growth and backfat at later ages.

Selection for testosterone in pigs 1335 Table 3. Least-squares means and SED for growth, carcass, Fat-o-Meater, and meat quality traits Line mating type interaction 1,2 Sex 3 Trait 4 LTL HTL WLT WHT SED 5 Castrate Gilt SED 5 D114, d 233 c 223 b 182 a 179 a 3.6 ADG, kg/d 0.52 a 0.55 b 0.67 c 0.69 c 0.012 0.62 y 0.59 x 0.006 ABF, mm 26.3 a 33.7 b 27.9 ac 29.2 c 1.39 38.2 y 20.4 x 0.59 ALMA, cm 2 35.8 a 35.6 a 39.8 c 38.3 b 0.78 36.3 x 38.4 y 0.40 PPL, % 48.5 b 45.9 a 49.6 c 48.6 bc 0.60 44.6 x 51.7 y 0.27 AFD, mm 31.1 33.3 30.2 31.6 2.61 34.2 y 29.0 x 0.77 ALMD, mm 53.8 b 55.0 b 50.0 a 51.2 ab 2.53 50.6 x 54.5 y 0.82 %LEAN 47.0 45.7 46.3 44.8 1.76 43.9 x 47.9 y 0.53 MAR 3.5 b 3.4 b 2.8 a 2.8 a 0.25 3.4 y 2.9 x 0.13 %LIPID 3.95 4.00 0.309 3.80 4.15 0.260 ph 5.91 5.91 0.029 5.93 5.88 0.028 DRIP, % 1.99 bc 2.57 c 1.22 a 1.87 b 0.303 1.83 2.00 0.155 CS 3.6 3.5 0.09 3.7 y 3.5 x 0.08 L* 50.22 f 52.10 b 0.768 51.33 50.99 0.378 a* 6.85 f 7.82 b 0.206 7.42 7.25 0.142 b* 1.02 f 1.90 b 0.352 1.53 1.39 0.182 a c Line mating type means without a common superscript letter differ (P < 0.05). x,y Sex means without a common superscript letter differ (P < 0.05). 1 LTL = low testosterone purebreds; HTL = high testosterone purebreds; WLT = (Large White Landrace composite) LTL; WHT = (Large White Landrace composite) HTL. 2 Means, within rows, for line mating type interaction effect lacking common superscript differ (P < 0.05). 3 Means, within rows, for sex effect lacking common superscript differ (P < 0.05). 4 D114 = days to 114 kg; ABF = real-time ultrasound backfat adjusted to 114 kg, ALMA = real-time ultrasound LM area adjusted to 114 kg; PPL = predicted percent lean; AFD = Fat-O-Meater fat depth; ALMD = Fat-O-Meater loin depth; %LEAN = Fat-O-Meater lean percentage; MAR = marbling score (scale 1 to 10; 1 to 10% i.m. fat); %LIPID = percent lipid determined by Folch method; ph = ph 24 h postmortem; DRIP = average collected purge from 40 to 50 g samples suspended 48 h (sample wt 100); CS = color scale (scale 1 to 6; 1 pale gray to 6 dark purplish red); L* = lightness color scale (range 0 to 100; 1 = pure black and 100 = pure white); a* = red to green color scale (greater value = redder color); b* = blue to yellow color scale (greater value = more yellow). 5 SED = standard error of the differences among least-squares means (most conservative). Finding that HTL was fatter than the WHT, LTL, and WLT lines was at first surprising. Compared among all groups HTL pigs had the greatest ABF and lowest PPL. However, synthesis of testosterone requires cholesterol, which is stored in fat, and it would be hypothesized that HTL boars would need an ample supply of cholesterol. Because cholesterol is stored in adipose tissue, animals with greater amounts of fat may have had a greater potential for producing testosterone and were more likely to be selected in the HTL line. As previously stated, crossbred pigs (WLT and WHT) grew significantly faster than purebreds (LTL and HTL). The 2 Duroc lines used in this study have been maintained as closed populations for 23 generations. Five to 10 sires were used in each line to sire each new generation. Thus, the lines were inbred. Unfortunately, due to lost pedigree information before 1990, actual inbreeding coefficients could not be estimated. Using the pedigree information that was available, the estimated average inbreeding coefficient was 17 and 15% in the LTL and HTL, respectively. It should be noted that these estimates do not include inbreeding accumulated during the time period when lines were undergoing selection. Mating the WC females to HTL and LTL males eliminated effects of inbreeding depression in the progeny. Rate of gain was slower in gilts than castrates, which is consistent with previous reports (Cisneros et al., 1996; Latorre et al., 2004). Also, castrates were fatter than gilts and had smaller ALMA. It is established that gilts are leaner (Cassady et al., 2004), and lean tissue is deposited at a much slower rate than fat. Thus, castrates grow faster, are fatter, and have smaller ALMA as expected. Data were collected at the plant to characterize loins and carcasses of these lines. Fat-O-Meater data must be collected rapidly to maintain chain speed. Techniques and instruments have been developed that measure traits in an acceptable and economically efficient manner, though ultrasound measures are generally more accurate. No differences were found among groups for AFD contrary to ABF. Fat-O-Meater AFD may be more variable. In addition, ABF and AFD are not measured at the same location. Comparing the ultrasound and Fat-O-Meater backfat indicates AFD typically has greater estimates. Contradictions were also observed between ALMA and ALMD. The WLT and WHT pigs had larger ALMA, whereas LTL and HTL had larger ALMD (Table 3). In addition, differences may be explained by breed. Derived from AFD and ALMD, %LEAN did not differ among the testosterone lines.

1336 Bender et al. Consistent with expectations, castrates had smaller ALMD than gilts, and gilts were leaner than castrates. High line crossbreds did not differ from WLT for ph. Castrates and gilts also had similar ph values. This is inconsistent with previous reports that females have lower ph (Leach et al., 1996; Latorre et al., 2004). The HTL animals had the greatest drip loss or lowest waterholding capacity. The values were not extremely high however. In fact, 2 to 3% is within the acceptable range. Color score is partially dependent upon the time from slaughter to chilling of the carcass. Most of the purebred pigs were detained on the slaughter line due to hair. As a result, color score measures were only analyzed on crossbred animals that were not detained. Subjective CS did not differ among crossbred high and low animals, but L*, a*, and b* were all lower in WLT pigs. The WLT meat appears to be darker, with green and blue tints. Darker meat coincided with greater waterholding capacity in WLT pigs. Castrates had a greater subjective color score, but there were no significant differences for any of the objective measures. Generally, castrates have paler meat color than intact males and darker muscle than gilts (Unruh et al., 1996; Cisneros et al., 1996). Because meat color was darker in castrates, it would be expected they would have lower DRIP, but this is not evident from this research. Marbling did not differ between high and low testosterone lines. However, purebreds (LTL and HTL) had greater marbling than crossbreds (WLT and WHT), and castrates had greater marbling than gilts. Castrates generally have more marbling because they are fatter and there is a moderate positive correlation between backfat and marbling (Candek-Potokar et al., 2002; Huff-Lonergan et al., 2002). In contrast to marbling scores, %LIPID did not differ between sexes. This contradiction is explained in 2 ways. First, marbling is subjective and %LIPID objectively quantifies all lipids including phospholipids. Second, only purebreds were analyzed for %LIPID, but marbling included both crossbred and purebred pigs. Despite discrepancies, the means for all classifications are within or above industry recommendations (Fernandez et al., 1999). Few objective reports of US pork industry meat quality traits are available. Using the NPPC genetic evaluation programs (NPPC, 1995, 1999) for comparison, the %LIPID recorded in the HTL and LTL lines are well above industry averages. The percentage of intramuscular fat in modern Durocs exceeded all other breeds in the NPPC studies and ranged from 3.2 to 3.3% (NPPC, 1995). The HTL and LTL lines were well above those averages. Overall the quality of the loins for all groups was acceptable. IMPLICATIONS Selection for endogenous testosterone production may be beneficial. High testosterone line animals grew significantly faster, whereas low line animals were significantly leaner. Both of the lines had acceptable meat quality. Color and moisture content were close to industry ideals. In addition, these purebred lines had exceptional marbling. Their marbling scores and percent lipid were well above industry average. LITERATURE CITED Candek-Potokar, M., G. Monin, and B. Žlender. 2002. Pork quality, processing and sensory characteristics of dry-cured hams as influenced by Duroc crossing and sex. J. Anim. Sci. 80:988 996. Cassady, J. P., O. W. Robison, R. K. Johnson, J. W. Mabry, L. L. Christian, M. D. Tokach, R. K. Miller, and R. N. Goodwin. 2004. National pork producers council maternal line genetic evaluation: A comparison of growth and carcass traits in terminal progeny. J. Anim. Sci. 82:3482 3485. Cisneros, F., M. Ellis, F. K. McKeith, J. McCaw, and R. L. Fernando. 1996. Influence of slaughter weight on growth and carcass characteristics, commercial cutting and curing yields, and meat quality of castrates and gilts from two genotypes. J. Anim. Sci. 74:925 933. DeWilde, R. O., and H. Lauwers. 1984. The effect of parental use of estradiol, progesterone, testosterone and trenbolone on growth and carcass composition in pigs. J. Anim. Sci. 59:1501 1509. D Souza, D. N., F. R. Dunshea, R. D. Warner, and B. J. Leury. 1998. The effect of handling pre-slaughter and carcass processing rate post-slaughter on pork quality. Meat Sci. 50:429 437. Falconer, D. S. 1989. Introduction to Quantitative Genetics. 3rd ed. John Wiley and Sons Inc., New York, NY. Fernandez, X., G. Monin, A. Talmant, J. Mourot, and B. Lebret. 1999. Influences of intramuscular fat content on the quality of pig meat 2. Consumer acceptability of m. longissimus lumborum. Meat Sci. 53:67 72. Folch, J., M. Lees, and G. H. Sloane Stanley. 1957. A simple method for the isolation and purification of total lipids from animal tissues. J. Biol. Chem. 226:497 504. Holl, J. W., and O. W. Robison. 2003. Results from nine generations of selection for increased litter size in swine. J. Anim. Sci. 81:624 629. Honikel, K. O. 1987. Critical evaluation of methods detecting waterholding capacity in meat. Page 225 in Accelerated Processing of Meat. A. Romita, C. Valin, and A. A. Taylor, ed. Elsevier Applied Sci., London, UK. Huff-Lonergan, E., T. J. Baas, M. Malek, J. C. M. Dekkers, K. Prusa, and M. F. Rothschild. 2002. Correlations among selected pork quality traits. J. Anim. Sci. 80:617 627. Jensen, P., H. B. Craig, and O. W. Robison. 1967. Phenotypic and genetics associations among carcass traits of swine. J. Anim. Sci. 26:1252 1260. Land, R. B. 1973. The expression of female sex-limited characters in males. Nature (Lond.) 241:208 209. Latorre, M. A., R. Lázaro, D. G. Valencia, P. Medel, and G. G. Mateos. 2004. The effects of sex and slaughter weight on the growth performance, carcass traits, and meat quality characteristics of heavy pigs. J. Anim. Sci. 82:526 533. Leach, L. M., M. Ellis, D. S. Sutton, F. K. McKeith, and E. R. Wilson. 1996. The growth performance, carcass characteristics, and meat quality of halothane carrier and negative pigs. J. Anim. Sci. 74:934 943. Lo, L. L., D. G. McLaren, F. K. McKeith, R. L. Fernando, and J. Novakofski. 1992. Genetic analyses of growth, real-time ultrasound, carcass and pork quality traits in Duroc and Landrace pigs: II. Heritabilities and correlations. J. Anim. Sci. 70:2387 2396. Lubritz, D., B. Johnson, and O. W. Robison. 1991. Genetic parameters for testosterone production in boars. J. Anim. Sci. 69:3220 3224. Montgomery, T. H., P. F. Dew, and M. S. Brown. 2001. Optimizing carcass value and the use of anabolic implants in beef cattle. J. Anim. Sci. 79(E. Suppl.):E296 E306. NPPC. 1995. Genetic evaluation: Terminal line program results. Natl. Pork Prod. Counc., Des Moines, IA.

Selection for testosterone in pigs 1337 NPPC. 1999. Genetic evaluation: Maternal line program results. Natl. Pork Prod. Counc., Des Moines, IA. NPPC. 2000. Pork composition and quality assessment procedures. Natl. Pork Prod. Counc., Des Moines, IA. NRC. 1998. Nutrient Requirements of Swine. 10th ed. Natl. Acad. Press, Washington, DC. NSIF. 2003. Guidelines for uniform swine improvement programs. Natl. Swine Improv. Fed., West Lafayette, IN. Robison, O. W., D. Lubritz and B. Johnson. 1994. Realized heritability estimates in boars divergently selected for testosterone levels. J. Anim. Breed. Genet. 111:35 42. Schinckel, A. P., R. K. Johnson, and R. J. Kittok. 1984. Testicular development and endocrine characteristics of boars selected for either high or low testis size. J. Anim. Sci. 58:675 685. Snochowski, M., K. Lundström, E. Dahlberg, H. Petersson, and L. E. Edqvist. 1981. Androgen and glucocorticoid receptors in porcine skeletal muscle. J. Anim. Sci. 53:80 90. Unruh, J. A., K. G. Friesen, S. R. Stuewe, B. L. Dunn, J. L. Nelssen, R. D. Goodband, and M. D. Tokach. 1996. The influence of genotype, sex, and dietary lysine on pork subprimal cut yields and carcass quality of pigs fed to either 104 or 127 kilograms. J. Anim. Sci. 74:1274 1283. Walker, S., O. W. Robison, C. S. Whisnant, and J. P. Cassady. 2004. Effect of divergent selection for testosterone production on testicular morphology and daily sperm production in boars. J. Anim. Sci. 82:2259 2263. Zubay, G. 1993. Biochemistry, 3rd ed. Wm. C. Brown Publishers, Dubuque, IA.