RELATIONSHIP BETWEEN GROWTH OF SUFFOLK RAMS ON CENTRAL PERFORMANCE TEST AND GROWTH OF THEIR PROGENY D. F. WALDRON, D. L. THOMAS, J. M. STOOKEY and R. L. FERNANDO, USA University of Illin o is, Department of Animal Sciences, 1207 W. Gregory Dr., Urbana, Ilinois 61801, U.S.A. SUMMARY Eighteen Suffolk rams, purchased from three central test stations, were singlesire mated to both Suffolk and 1/2 Dorset, 1/4 Suffolk, 1/4 Rambouillet ewes at the University of Illin o is, Dixon Springs Agricultural Center, Simpson, Illin o is, U.S.A. Rams had been tested for growth over a 63-day period starting when they were approximately two months of age. Progeny lambs were weaned at 56+4 days of age and males were le ft intact. Lambs were fed a high energy ration ad libitum in confinement postweaning. Lamb body weights were taken at birth, 28+4 days of age, at weaning, and fortnightly postweaning until reaching 50 kg live weight. Progeny performance (birth, 60 and 120 day weights and gains between these ages) was regressed on sire's central test performance (average daily gain on test and weight per day of age at end of test). None of the partial regression co efficients were significantly different from zero. These data suggest that central test performance of Suffolk rams, measured on a 63 day test starting when rams are approximately two months of age, is not an accurate indicator of breeding value for growth. INTRODUCTION Central performance testing of rams in the United States was initiated at the Texas Agricultural Experiment Station, Sonora, Texas in 1948 (Shelton, 1979). This station tests primarily finewooled Rambouillet rams and is s t ill in operation today. Hescock (1981) reported that in 1981, 13 states in the U. S. had active central performace tests of rams with major differences among tests as to type of ration fed, length of test, age of rams at start of test, breeds tested, tra its recorded, etc. Shelton (1959, 1968 and 1979) has shown positive phenotypic trends over time for performance of Rambouillet rams on central test at Sonora, Texas and significant positive relationships between growth and fleece production of the rams on central test and growth and fleece production of their progeny within a flock. The accuracy of other central tests in estimating genetic differences among tested rams for economically important traits has not been estimated. A review of the contributions of ram test stations by Ahlschwede (1975) opened with the statement, "Specific data relating to the contribution of ram test stations to breed improvement programs is not readily available." The purpose of this study was to determine the relationship between the growth of Suffolk ram lambs on a 63-day central test and progeny growth within a flock. Central Tests. MATERIAL AND METHODS The Suffolk is the most popular pure breed of sheep in the United States; accounting for 49% of a ll purebred sheep registrations in 1984 (Niemeier, 1985) and the breed represented in the largest numbers in central ram tests in the mid- 639
western United States, Due to this popularity and availability, rams of the Suffolk breed were utilized in this study. Six rams were purchased from each of three central tests: the 1982 Illin o is, 1983 Iowa and 1984 Iowa tests. The three tests included 37, 92 and 76 Suffolk rams from 21, 30 and 29 different breeders, respectively, with an overall total of 205 rams from 59 breeders. Rams to be centrally tested are selected by individual breeders and brought to the central testing location near the firs t of April. Rams are group fed a high energy ration ad libitum in confinement. A six day adjustment period in the testing fa c ility is given the rams before an on-test weight is taken. Rams are approximately 60 days of age at the start of the test. Rams are weighed every 21 days with the off-test weight taken 63 days after the start of the test. Central test average daily gain (ADG) is off-test weight minus on-test weight divided by 63. Weight per day of age (WDA) is also calculated for each ram by dividing off-test weight by off-test age in days. ADG and WDA are not adjusted for any environmental effects such as type of birth, age of dam, date of birth, etc. A fter visual screening by a committee of breeders and elimination of rams with major faults, the top performing rams for ADG are sold at public auction with rams with higher ADG selling before rams with lower ADG. Ram Selection. A fter the completion of the test and before the public auction, rams from the 1982 and 1983 tests were listed from highest to lowest on ADG and divided into six groups as equal in number as possible. The firs t ram in each group was identified as the desired ram for purchase. In 1984 a similar procedure was used except that the rams were divided into nine groups with a ram identified for purchase from each of the top three and bottom three groups. If the identified ram sold for an excessive amount of money, it was not purchased, and the next ram in the group was iden tified for purchase. A m ajority of the o rig in a lly iden tified rams were purchased for the study, and in all cases, one of the rams from each group was purchased. Rams with very low ADG were not allowed in the public auction and were purchased through private negotiations with the appropriate breeder. One of the rams purchased in 1982 died before siring any progeny. A commerc ia l producer who had purchased a ram from the 1982 test allowed the use of his ram fo r the study. The borrowed ram was not from the same group as the ram that had died so only five of the six groups were represented in the sample of rams used from the 1982 test. The 18 rams used in the study were from 16 different breeders. The performance of each of the 18 rams used and the mean performance of a ll rams on test each year are presented in Table 1. Progeny Evaluation. Each of the rams was mated to approximately 15 purebred Suffolk and 15 crossbred 1/2 Dorset, 1/4 Suffolk, 1/4 Rambouillet ewes at the University of Illin o is, Dixon Springs Agricultural Center located in southern Illin ois near the town of Simpson. Breeding began in September when the rams were approximately seven months of age. Progeny were born in February and March and weighed at birth and at 28+4 days of age while nursing their dams. Lambs had continuous access to a creep ration while with their dams. Ram lambs were le ft intact. At 56+4 days of age, lambs were weighed and weaned from their dams. Postweaning, lambs were fed a high energy, ground ration ad libitum in a confinement barn on a 640
slotted floor and weighed fortnightly until reaching 50 kg live body weight. TABLE 1. PERFORMANCE OF CENTRAL TEST EACH YEAR EACH SELECTED RAM AND MEAN OF ALL SUFFOLK RAMS ON 1982 u 1983 1984 Item ADG(g) WDA(g) ADG(g) WDA(g) ADG(g) WDA(g) Growth of 604 527 622 613 636 568 individual 577 518 568 522 577 590 selected rams 554 495 549 490 563 499 504 477 518 481 481 504 504 590 481 490 468 454 445 513 427 395 395 445 Mean of a ll rams 536 490 522 504 518 518 faverage daily gain (in grams) while on central test. Weight per day of age (in grams) at end of central test. Five hundred ninety- four lambs were born over the three years (33 per sire; range of 22-48) with 420 lambs having adequate growth information to 120 days of age (23 per sire; range of 18-35). Sixty and 120 day weights were predicted for each lamb from the unique regression equation generated for each lamb by regressing a ll weights recorded for each lamb on age at weighing and age at. weighing squared. The mean performance of the lambs included in the study for each of the six progeny traits is presented in Table 2. TABLE 2. PERFORMANCE OF PROGENY OF CENTRALLY TESTED SUFFOLK RAMS Trait No. of progeny Mean + std. dev. (kg) Birth wt. 594 4.50 + 1.19 60 d wt. 420 21.60 + 4.53 120 d wt. 420 39.93 + 8.02 Gain, birth to 60 d 420 16.94 + 3.95 Gain, birth to 120 d 420 35.27 + 7.43 Gain, 60 to 120 d 420 18.33 + 4.55 Statistical Analyses. The six progeny traits analyzed were birth weight, predicted 60 and 120 day weight and gain from birth to 60 days, 60 to 120 days and birth to 120 days. The mixed linear model used to analyze the six dependent progeny variables included year of birth (Y ), age of dam (A), type of birth (T), breed of dam (B), sex (S) and the sig n ific a n t interactions: Y x T, Y x B, Y x S and B x S as fixed e ffe c ts ; sire and residual error as random effects and date of birth within year and sire's performance on central test (expressed as a deviation from the mean performance of a ll rams on test in the same year) as covariates. The partial regression of progeny performance on sire's central test performance is the simple phenotypic offspring-parent regression. The random effect of sire in the 641
above model takes into account genetic differences among sires that are not accounted for by the covariate of sire's performance on central test. Estimates of variance components of the random effects were obtained using the restricted maximum likelihood (REML) procedure (Patterson and Thompson, 1971) via the expectation maximization (EM) algorithm (Dempster et al., 1977). The best linear unbiased estimates (BLUE'S) of estimable functions of fixed effects and covariates were obtained with Henderson's mixed model equations (Henderson, 1973), using the REML estimates of the variance components. RESULTS AND DISCUSSION Presented in Table 3 are the partial regressions of progeny performance on sire's central test performance. None of the estimated regression coefficients TABLE 3. PARTIAL REGRESSION COEFFICIENTS (b) + STANDARD ERRORS (s.e.) OF PROGENY PERFORMANCE ON SIRE'S CENTRAL TEST PERFORMANCE Progeny traits (Y) Sire's central test traits ADG3 (X) WDAb b + s.e. b + s.e. Birth wt. -.81 +.85-2.51 +.91 60 d wt. 3.62 + 3.48 1.00 + 4.60 120 d wt. 3.68 + 7.55 1.09 + 9.69 Gain, birth to 60 d 4.57 + 3.00 3.42 + 4.01 Gain, birth to 120 d 4.59 + 6.96 3.45 + 8.95 Gain, 60 to 120 d 0.05 + 4.55-0.03 + 5.79 ^Average daily gain while on central test. Weight per day of age at end of central test. were significan tly different from zero. These results suggest that central test performance of Suffolk rams, measured on a 63 day test beginning at approximately two months of age, is not an accurate indicator of breeding value for growth rate. These results are not due to a lack of genetic variation among the lambs for growth. For example, the paternal half-sib estimate of h eritability of lamb gain from 60 to 120 days of age was.30. However, sire ADG was not an accurate predictor of progeny gain from 60 to 120 days of age. If it is assumed that average daily gain on central test is the same trait as average daily gain from 60 to 120 days of age on the farm, an estimate of heritability of the trait can be calculated by multiplying the regression of progeny performance on sire performance by two. Using the appropriate value from Table 3, the estimate of h eritability of central test average daily gain is.0017 = ([.05 * 60] x 2) which is not d iffe r ent from zero. The large difference between these two estimates of heritability fo r postweaning gain highlights the inability of these three central tests to rank sires on their ability to produce fast-growing progeny. These results d iffer from those of Shelton (1959). Using progeny from 83 Rambouillet rams centrally tested at the Sonora, Texas station, he found a correlation of.18 between sires' gain on test and progeny weaning weight and a 642
h eritability of gain on test of.22. Two important differences exist between the Texas and the Illin o is and Iowa testing procedures. Rams in Texas are older at the initiation of the test and are tested for a longer period of time than rams in the Ilinois and Iowa tests. In itia l age of rams tested in Texas is approximately seven months and test length is now 140 days, but test length was as high as 308 days in some of the early years (Shelton et a l., 1954) which generated the data for the paper by Shelton (1959). This compares to in itia l ram ages of approximately two months and a test period of 63 days for the Illin ois and Iowa tests. Since weights and gains obtained at older ages (up to 1 year of age) are generally more heritable and relatively highly correlated with earlier weights and gains (Vesely and Sen, 1961; Vesely et al., 1979; Olson et a l., 1976), the testing at older ages in Texas than at Illin ois and Iowa may account for the d ifferen t results. Also, since rams in Texas are tested for over twice as long as rams in Illin o is and Iowa, the pre-test environment may have less of an effect on test performance of rams tested in Texas than of rams tested in Illin ois and Iowa. No studies have been found that compare the effectiveness of central tests conducted for different lengths of time. However, Shrestha et al. (1983) has reported significantly higher heritabilities for 35, 70 and 140 day weights of lambs reared a r tific ia lly from one day of age versus with their dams to 70 days of age. The a rtific ia l rearing lessened the environmental effect of dam's milk production on lamb growth allowing for the greater expression of genetic d ifferences among lambs for growth. Lengthening of the testing period (increasing the time the rams are exposed to a uniform environment) in central tests may have a similar positive effect on identifying genetic differences among rams as does a r t ific ia l rearing. Also, since it is known that type of rearing, age of dam, weaning age and other factors can affect postweaning gain, gain on central tests of short duration should, perhaps, be adjusted for these environmental effects. CONCLUSION Growth of two month old Suffolk rams on 63-day central tests, as currently measured in the Illin o is and Iowa tests, appears to have lit t le relationship with progeny growth. REFERENCES AHLSCHWEDE, G. 1975. Contributions of ram test stations. Sheep Breeding and Feeding for Profit Symp., Sheep Industry Dev. Program, Inc., Denver, Colorado, U.S.A. (mimeo) and Sheep Breeder and Sheepman. 95 (12), 8-12. DEMPSTER, A. P., LAIRD, N. M. and RUBIN, D. B. 1977. Maximum likelihood from incomplete data with EM algorithm. J. Royal Stat. Soc. B. 39, 1-38. HENDERSON, C. R. 1973. Sire evaluation and genetic trends. In: Proc. Anim. Breeding Symp. in Honor of Dr. Jay L. Lush. Amer. Soc. Anim. Sci. and Amer. Dairy Sci, Assoc., Champaign, Illin o is, 10-41. HESCOCK, T. 1981. Ram performance testing is alive and well. Progressive Sheep Breeder. Fall, 1981, 15-54. NIEMEIER, L. L. 1985. Purebred review. Sheep Breeder and Sheepman. 105(3), 118-126. 643
OLSON, L. W., DICKERSON, G. E. and GLIMP, H. A. 1976. Selection criteria for intensive market lamb production: growth traits. J. Anim. Sci. 43, 78-89. PATTERSON, H. D. and THOMPSON, R. 1971. Recovery of interblock information when block sizes are unequal. Biometrika. 58, 545-554. SHELTON, M. 1959. Selection of finewool rams based on record of performance data. J. Anim. Sci. 18, 925-930. SHELTON, M. 1968. Performance testing and sire evaluation techniques and their potential for improvement in U.S. sheep population. Proc. Symp. Genetic Imp. Wool and Lamb Prod., Sheep Industry Dev. Program, Inc., Denver, Colorado, U.S.A., 64-92. SHELTON, M. 1979. Estimation of genetic change in a performance testing program for sheep. J. Anim. Sci. 48, 26-31. SHELTON, M., MILLER, J. C., MAGEE, W. T. and HARDY, W. T. 1954. A summary of four years work in ram progeny and performance testing. J. Anim. Sci. 13, 215-228. SHRESTHA, J. N. B., PETERS, H. F. and HEANEY, D. P. 1983. Inheritance of weights of lambs reared a rtific ia lly or with their dams. Canadian J. Anim. Sci. 63, 263-268. VESELY, J. A. and SEN, S. B. 1961. H eritabilities of weaning weight, yearling weight, and clean fleece weight in range Romnelet sheep. Canadian J. Anim. Sci. 4J^, 109-114. VESELY, J. A., PETERS, H. F., SLEN, S. B. and ROBISON, 0. W. 1977. Heritabilities and genetic correlations in growth and wool traits of Rambouillet and Romnelet sheep. J. Anim, Sci. 30, 174-181. 644