J. Anim. Sci. 2003. 81:1406-1413
© 2003 American Society of Animal Science
Evaluation of maternal performance of daughters from high and low milk EPD sires
J. F. Baker*,1,
M. E. Boyd
,
A. H. Brown
,
D. E. Franke
and
C. E. Thompson¶
* Department of Animal and Dairy Science, The University of Georgia, Tifton 31793;
and
Department of Animal and Dairy Sciences, Mississippi State University, Mississippi 39762;
and
Department of Animal Science, University of Arkansas, Fayetteville 72701;
and
Department of Animal Science, Louisiana State University Agricultural Center, Baton Rouge 70803; and
and
¶ Department of Animal and Vet Sciences, Clemson University, Clemson, SC 29634
1 Correspondence:
P.O. Box 748 (phone: 229-386-3364; fax: 229-386-3219; E-mail:
jfbaker{at}tifton.cpes.peachnet.edu).
 |
Abstract
|
|---|
Angus bulls (n = 24) were selected for either high or low milk EPD, but with similar growth EPD and mated within location (n = 6) at random to Angus cows. Daughters from these matings were bred to calve first at 2 yr of age to common reference sires across locations. Lactation records for 192 daughters were used to evaluate 12-h milk yield, percentage of milk fat and protein, and weaning weight of offspring. Milk production was measured four times during the lactation at regular intervals within location. Dams were separated from their calves the night before milking and milked with a portable milking machine the next morning to estimate 12-h milk yield. A sample of the milk was collected from each cow and analyzed for percentages of milk fat and protein. Data were analyzed as repeated records of the dam. Fixed effects were location, genetic line of sire, gender of calf within location, and milking period, with postpartum interval used as a covariate. Fixed effects and the random effects of sire of dam nested within line, sire of calf, and year were estimated by REML. Genetic line was an important source of variation for milk yield (P < 0.01) and percentage of milk fat (P = 0.03) but not for percentage of milk protein (P = 0.49). Location was significant for all three milk variables (P < 0.01), but the interactions between line and location were not significant. Gender of calf was significant for milk yield (P = 0.04) but not for percentage of milk fat or protein. Line (P = 0.02), location (P = 0.01), calf gender (P = 0.01), and age at weaning (P = 0.01) were significant sources of variation for weaning weight but the interaction of line and location was not (P = 0.69). The correlation coefficient between the sire’s milk EPD and 12-h milk yield was significantly different from zero (r = 0.56). The difference between the least squares means for high and low lines for milk yield was 0.66 kg/12 h and the difference was 15.3 kg for weaning weight. The results indicate that there was not evidence for a genotype by environment interaction in milk production for daughters from divergent sires selected for high or low milk EPD.
Key Words: Beef Cattle • Milk Yield • Predicted Difference • Progeny
|
Introduction
|
|---|
The use of milk EPD in beef cattle has been effective in changing weaning weights, as demonstrated by genetic trends of numerous breeds. Milk EPD in beef cattle are estimated by an evaluation of differences in weaning weights of offspring. Statistical models used to predict EPD attempt to account for herd, year, and season effects, nonrandom mating, gender of calf, genetic relationships, and other factors. It has been shown that milk yield is positively associated with calf growth and accounts for a significant portion of the variation in preweaning rate of gain in beef calves (Neville, 1962; Gregory et al., 1992; Fiss and Wilton, 1993). In addition Miller et al. (1999) concluded that cows that have higher yield have a greater potential for profit at raising a calf to slaughter.
Several studies have estimated the correlation coefficients between milk yield, sire milk EPD, and body weights before weaning for specific locations (Marshall and Long, 1993; Minick et al., 2001). However, these previous studies were not able to evaluate interactions between genetic line and production location. In addition, predicted differences in weaning weight due to milk EPD were not validated. Milk yield and milk components were measured during the first lactation of daughters of Angus sires selected for either high or low milk EPD (mEPD) at six locations in the southeastern United States. Body weight at weaning and correlations with milk variables were also measured. The objectives were to evaluate the magnitude of differences between mEPD lines for weaning weights, measure milk yield, fat, and protein component differences between mEPD line and location, and to determine the importance of interactions between location, mEPD line, and interval since calving on milk yield, fat, and protein.
|
Materials and Methods
|
|---|
Records for active sires from the American Angus Association were used to rank the highest and lowest 100 bulls for milk EPD. Bulls (n = 19) were identified that had accuracy values for milk EPD greater than 0.6 from both lists. In addition to the milk EPD value, a secondary selection criterion was to identify bulls, with semen available, that were similar in growth EPD so that the average of the two lines for birth, weaning, and yearling EPD were similar. One location used five whose accuracy values were below 0.5 when selected. It should be noted that it was more difficult to find semen from bulls that had large negative milk EPD and also had an acceptable accuracy. Average EPD values for birth, weaning, yearling, and milk at the time the bulls were selected are presented in Table 1
. In addition to the averages at the time of selection, an average weighted by the number of generation 1 (G1) daughters is also included, as well as the weighted averages from the Fall 2001 American Angus Association analyses. Since EPD can change as additional information is acquired and with the passing of time, the weighted averages based on the most recent analyses are also presented in Table 1. The difference between the high and low weighted averages for milk EPD would be the best estimate for the difference in weaning weight between offspring from daughters of the two lines.
View this table:
[in this window]
[in a new window]
|
Table 1. Unweighted and weighted EPD for high- and low-milk EPD lines based on the number of daughters milked within line
|
|
There were six distinct locations where G1 daughters were produced for the project. The following locations contributed records for analyses: Edisto Research and Education Center (Blackville, SC), Louisiana State University Agricultural Center, Mississippi State University, The University of Arkansas, and The Coastal Plain Experiment Station of the University of Georgia (two locations). Location effects were not quantified but do represent a wide spectrum of nutritional, seasonal, and management differences that are found in beef cattle operations in the south. It is known that the locations have different forages and different breeding/calving seasons. The Arkansas and Mississippi locations each had pastures of tall fescue with infestation of the toxic fungus. Mississippi, Louisiana, Arkansas, and one of the Georgia locations had pastures with common bermudagrass. The other Georgia location and South Carolina had pastures with improved varieties of bermudagrass available with very little or no common bermudagrass. Calving seasons were started in December or throughout January depending on location but did not change within a location for the duration of the study. Louisiana and both Georgia locations utilized cool-season annuals for supplemental grazing of lactating cows when available in the spring.
In addition to the differences noted above, location effects also included genetic differences in the generation 0 (G0) cows that were the result of previous breeding decisions at each location. There was no known genetic tie between G0 cows at these locations other than that they were all purebred Angus. The only exception was that the cows at the two Georgia locations might have had more similarity in their pedigrees due to historical use of sires at each location. Cows at one Georgia location were registered but, at the other location, were not registered and were considered purebred. The G0 cows at each location were randomly mated by AI to either a high- or low-mEPD sire. The EPD for G0 dams were not used in the mating decisions. Therefore, it was assumed that G0 dams represented a random sample of Angus females that were available and sires from the high and low lines each had a random sample of potential mates. The list of sires within each line, the number of G1 daughters milked from each sire, the number of locations milking daughters from each sire, the year in which daughters were milked, and the number of milk records per year are presented in Table 2.
Initial matings were in 1991 and continued for four additional years. Each location determined how many G0 cows to breed for the project. The research protocol and facilities were approved by the individual state’s Institutional Animal Care and Use Committees. Generation 1 heifers from the project sires were reared and managed together at each location. Heifers were developed according to traditional management practices at each location with the stipulation that all G1 heifers would be bred to have their first calf at approximately 2 yr of age. Initial G1 heifer mating was by AI to a common sire selected to represent an average Angus sire unrelated to the sires within the high and low lines. A second unrelated sire was selected to use for breeding to the G1 daughters beginning with those born in yr 3 of the project. Natural-service bulls were also used at each location following the AI period, the length of which varied by location. Each location was responsible for selection of their natural-service sire, which would have a comparable EPD to the AI reference sires. The AI sires with accuracy values greater than 0.6 for growth and milk EPD were selected from commercial sources.
Procedures for estimation of milk yield were standardized across locations. Each G1 cow was milked four times during her first lactation. First milking date occurred when calves averaged between 45 and 75 d of age. The fourth milking occurred just before weaning. Second and third milking dates were determined to create approximate equal intervals between the first and fourth milking dates. The day before milking, the cow and her calf (generation 2, G2) were weighed and separated at approximately 1300. Pairs were placed together again at about 2000 for approximately 30 min. The cows and calves had access to water and grass or hay while separated. The next day, milk yield was determined by use of a portable milking machine after the overnight (12-h) separation from their calves. Approximately 10 min before attachment of the milking machine, each cow was injected (i.m.) with 2 mL (10mg/mL) acepromazine (except at Louisiana). The udder was cleaned, if necessary, to remove excess dirt. Immediately before attachment of the milking unit, the cow received 100 USP units of oxytocin (i.v.) to facilitate milk flow and extraction. When milk flow ceased or appeared to be finished, the machine was removed and the teats were dipped with a standard dairy teat dip. Milk was weighed and approximately 100-mL samples were obtained and preserved with 2-bromo-2-nitropropane-1,3-diol tablets. One location was not able to sample milk for fat and protein analyses. Milk samples were submitted to Dairy Herd Improvement Association laboratories appropriate for each location for the percentage of protein and fat analyses.
Data collected for analyses included calving date, weaning weight, milk yield, milk fat percentage, and milk protein percentage. Mixed model procedures (SAS Inst., Inc., Cary, NC) were used to analyze the data. The model for weaning weight included mEPD line, location, and G2 gender of calf (bull, heifer, or steer) nested within location as fixed effects. None of the locations used creep feed nor were any calves implanted. Two locations (South Carolina and one location in Georgia) castrated bull calves at or near birth. Other locations reared bull calves to weaning as intact males. Calf age at weaning was used as a covariate. Deviation of G1 cow age at calving in months from the mean G1 cow age at calving (24.18 mo) was also included as a covariate for weaning weight. Random effects included in the analysis of weaning weight included year, sire of G1 dam nested within mEPD line, and sire of G2 calf nested within year. The model for milk yield, milk fat percentage, and milk protein percentage was a repeated records model with subject being individual G1 dams. The model included the following as fixed effects: mEPD line, location, and gender of calf (bull, heifer, or steer) nested within location and milk-yield period (1st, 2nd, 3rd, or 4th). The two-factor interactions included line x location, line x period, and period x location. The postpartum interval at each measurement period was included as a covariate. Random effects included in the analysis of the milk variables included year, sire of G1 dam nested within mEPD line, and sire of G2 calf nested within year. Estimation method was REML, with a heterogeneous compound symmetry structure for the variance/covariance components. Additional analyses included estimation of correlation coefficients between selected variables using SAS procedures PROC CORR (SAS Inst., Inc.). To generate the coefficients between a sire’s EPD and daughter’s milk yields and grandoffspring weaning weights, the model presented earlier was modified to include the sire nested within line as a fixed effect with option of least squares means for sire. Therefore, the correlation coefficients are a measure of the relationship between a sire’s EPD and least squares means for milk, yield, milk components, and weights.
|
Results and Discussion
|
|---|
Milk Yield Analysis
Line and location were important sources of variation (P < 0.01) for 12-h milk yield. The interaction between line and location was not important (P = 0.99). Gender of calf and milk measurement period were significant sources of variation for milk yield. Two-factor interactions of period x location and period x line were also significant. Postpartum interval was significant for milk yield. Milk yield least squares means and standard errors for the two EPD lines are presented in Table 3. High-mEPD line dams produced significantly more milk than the low-mEPD line dams. Milk yield least squares means and standard errors for location x milk measurement period are presented in Table 4. Dams at the Louisiana location yielded significantly more milk at each measurement than other locations. Louisiana calving season was in the fall and the dams had access to winter annuals, whereas other locations calved in winter and early spring. The period x location interaction was caused by rank and magnitude changes among other locations across periods. The line x period interaction least squares means and standard errors for milk yield are presented in Table 5. High-mEPD line dams produced significantly more milk in 12 h than low-mEPD line dams at the first three measurements, but the means were not different at measurement four (P = 0.05). The interaction results from a more rapid decline in yield across time for the high-mEPD line dams compared with the low-mEPD line dams. The gender of the calf being nursed by the dam influenced 12-h milk yield (P = 0.04). Dams that nursed male calves had greater milk yields than dams that nursed heifer calves (Table 6).
View this table:
[in this window]
[in a new window]
|
Table 3. Least squares means and standard errors for 12-h milk yield (kg), percentage of milk fat (%), percentage of milk protein (%), and weaning weight (kg) by milk EPD line (high or low)
|
|
Differences in milk yields associated with milk EPD were also reported by Marston et al. (1992), Marshall and Long (1993), and Minick et al. (2001). Freetly and Cundiff (1998) reported that pattern of yield differed between breeds. In their study, the Belgian Blue-sired females were highest yielding early on, but were not different from several breeds during the last measurement period. Nutritional treatment effects were not significant for milk yield (Freetly and Cundiff, 1998). However, in the present study, location effects were significant, and at least part of the effect would have been nutritional inputs due to the fact that the Louisiana location had a fall calving season, with the dams having access to winter annuals. Postweaning gain differences had carryover effects for subsequent milk yield in first lactation in other studies (Ferrell, 1982; Buskirk et al., 1995).
Weaning Weight Analysis
Line (P = 0.02) and location (P < 0.01) were important sources of variation for weaning weight. The interaction between line and location was not significant (P = 0.69). Gender of calf and age of calf at weaning were highly significant sources of variation for weaning weight. Age of dam (recorded in months) was not significant (P = 0.90), but since all dams in this study calved within a narrow range in age, at a mean age of 24 mo, this lack of significance is reasonable. Mean age at weaning across all locations was 209 d. Weaning age means at each location were Louisiana, 212 d; Mississippi, 203 d; South Carolina, 219 d; Arkansas, 187 d; Georgia 1, 217 d; and Georgia 2, 205 d. Calves nursing dams from the high-mEPD line were significantly heavier at weaning than calves nursing low-mEPD line dams (Table 3
). Gender of calf within location least squares means and standard errors for weaning weight are presented in Table 7. Male calves at all locations were heavier than the heifer calves. However, only at the two Georgia locations and the South Carolina location were these differences between the genders significant (Table 7).
The difference between the weaning weights of calves nursing dams from the two lines was 15.3 kg (Table 3), with a standard error for the difference estimated as 5.8 kg. The EPD in average weaning weights between the two lines was 18 kg based on EPD when the bulls were selected, or 17 kg using EPD from the Fall 2001 American Angus Association analysis (Table 1). Therefore, the difference in weaning weights observed in this study was as predicted by the mEPD of the maternal grandsires.
Increasing milk yield has increased weaning weights in other studies (Marshall et al., 1976; Freking and Marshall; 1992; Minick et al., 2001), but Buskirk et al. (1995) reported no significant difference in weaning weight. Buskirk et al. (1995) suggested that calves from the dams that gave less milk might have compensated late in the nursing period with increased forage consumption. Differences in body weights were observed at earlier measurements in the study by Buskirk et al. (1995). In the current study, all heifers were managed together within location, and it is not known if mEPD line affected gains before first lactation, but growth EPD were determined to be similar. Differences in weaning weights in this study should be more closely associated with milk yield. In the study by Minick et al. (2001), breed effects were also significant for weaning weight differences between high- and low-milk EPD lines.
Milk Fat Analysis
Milk EPD line (P = 0.03) and location (P < 0.01) were significant sources of variation for percentage of milk fat, but the interaction between line and location effect was not significant. Measurement period (P < 0.01) and the interaction between period and location (P < 0.01) were significant effects for percentage of milk fat. Gender of calf and the interaction between line and period were not significant for percent milk fat. The covariate, postpartum interval, was a significant source of variation for percentage of milk fat. Milk EPD line least squares means and standard errors for percentage of milk fat are presented in Table 3. High-mEPD line dams produced milk with significantly higher percentages of milk fat than the low-mEPD line dams. Percentage of milk fat least squares means and standard errors for the period x location interaction are presented in Table 8. The significant interaction is a result of changes in rank as well as changes in the magnitude of differences between locations and measurement period.
Measured as a percentage, Buskirk et al. (1995) reported no significant treatment effect for milk fat, but it was estimated at only one time during lactation. Milk fat in the current study was influenced by line, location, and measurement period. The interaction between location and period that was observed in the current study does not have a pattern, and the means are associated with relatively large standard errors.
Milk Protein Analysis
Milk EPD line was not a significant source of variation for percentage of milk protein. The effect of location was important (P < 0.01), but the interaction with line was not significant. Gender of calf did not influence milk protein (P = 0.76). Measurement period and the interaction of measurement period with location were significant sources of variation (P < 0.01). The line x period interaction was not significant. Percentage of milk protein least squares means and standard errors for the two lines are presented in Table 3. Table 9 presents the percentage of milk protein least squares means and standard errors for the measurement periods and locations. The significant interaction results from changes in rank and magnitude across periods and locations. Only at the Georgia 2 location did the measurement period means not differ from each other. At the other locations at least one mean was different from the other means, but a trend for percentage of milk protein across time is not apparent. In the analysis of variance for percentage of milk protein, the residual error variance was much smaller than the residual error variance for the analysis of variance for percentage of milk fat (0.047 vs. 0.561, respectively). The percentage of milk protein was more uniform across periods and locations than percentage of milk fat. Buskirk et al. (1995) reported no significant differences in postweaning rate of gain treatment effects for milk protein and solids-not-fat.
Pearson Correlation Coefficients Between Variables
The degree of association between variables can be assessed by calculation of Pearson correlation coefficients. Two sets of primary interest in this study were the association between milk yield at one measurement period with yield at other measurements and the association between yield at a measurement period and calf weaning weight. Milk yield at each period was positively correlated with yield at the other periods (P < 0.01). The correlation coefficients for yields ranged from 0.67 to 0.82. The greatest association was for yield at period 1 with yield at period 2. The smallest association was for yield at period 1 with yield at period 4. The correlation of average milk yield and calf weaning weight was 0.48 (P < 0.01). The correlations between average milk fat and milk protein with calf weaning weight were small (-0.02 for both) and were not different from zero. The correlation coefficients (P < 0.01) for weaning weight with milk yield at each measurement period were 0.49, 0.44, 0.58, and 0.50 for the 1st, 2nd, 3rd, and 4th periods, respectively. The measurements of associations between yield, milk fat, and milk protein within measurement periods were not very informative or consistent. There was a tendency for the correlations between yield and the milk components to be positive for the first period and then become negative by period four. The association between yield and milk fat went from r = 0.26 (P < 0.01) for period 1 to r = -0.11 (P > 0.20) for period 4. The correlation coefficients between yield and milk protein went from r = 0.18 (P > 0.06) to r = -0.34 (P < 0.01). The correlation coefficients between percentage of milk fat and percentage of milk protein within measurement period were always positive in sign but only differed from zero twice (0.39, 0.02, 0.02, and 0.46, respectively, for periods 1 to 4). Percentage of milk protein at each measurement was positively correlated (P < 0.002) with percentage of milk protein at the other measurement periods (ranged from 0.27 to 0.54). Percentage of milk fat at each measurement was significantly different from zero for four of the six coefficients (ranged from -0.04 to 0.27).
The correlation coefficient between the sire’s milk EPD and 12-h milk yield was significantly different from zero when estimated for the higher accuracy bulls (r = 0.56). The coefficient was lower when all bulls were included (r = 0.40, P = 0.07). Correlation coefficient between the sire’s milk EPD and least squares means weaning weights from G2 calves was 0.78 (P < 0.01) when estimated using only the high-accuracy bulls. The coefficient was lower (r = 0.58), but still different from zero, when all bulls were included. Diaz et al. (1992) reported a correlation of 0.26 between milk EPD and 12-h milk yield. Marston et al. (1992) reported correlation coefficients of 0.32 and 0.44, respectively, between milk yield of Angus and Simmental cows and their sire’s mEPD values. Marshall and Long (1993) reported a smaller correlation (r = 0.14) between a sire’s mEPD and daughter milk yield. Higher correlation coefficients in this study may in part be due to the use of sires with high accuracy values and the fact that all females were 2 yr old. Marston et al. (1992) also estimated correlation coefficients between the sire’s mEPD and adjusted weaning weights for Angus and Simmental of 0.38 and 0.39, respectively. These values would also appear to be lower than those estimated from the current study when the value is based only on sires with high-accuracy mEPD.
|
Implications
|
|---|
The primary focus of this multistate cooperative study was to validate the basic premise that milk EPD reflects differences in weaning weight due to milk yield. Results validated that expected differences were observed for weaning weight and, just as importantly, that no indication of an interaction between genetic line and location was detected for weaning weight or the milk variables. The absence of an interaction is of importance to breed associations and small producers. These results are based on good linkages across diverse environments with standardized milk yield estimation procedures.
Received for publication March 5, 2002.
Accepted for publication January 30, 2003.
|
Literature Cited
|
|---|
Buskirk, D. D., D. B. Faulker, and F. A. Ireland. 1995. Increased postweaning gain of beef heifers enhances fertility and milk production. J. Anim. Sci. 73:937–946.[Abstract]
Diaz, C., D. R. Notter, and W. E. Beal. 1992. Relationship between milk expected progeny differences of Polled Hereford sires and actual milk production of their crossbred daughters. J. Anim. Sci. 70:396–402.[Abstract]
Ferrell, C. L. 1982. Effects of postweaning rate of gain on onset of puberty and productive performance of heifers of different breeds. J. Anim. Sci. 55:1272–1283.
Fiss, C. F., and J. W. Wilton. 1993. Contribution of breed, cow weight, and milk yield to the preweaning, feedlot, and carcass traits of calves in three beef breeding systems. J. Anim. Sci. 71:2874–2884.[Abstract]
Freetly, H. C., and L. V. Cundiff. 1998. Reproductive performance, calf growth, and milk production of first-calf heifers sired by seven breeds and raised on different levels of nutrition. J. Anim. Sci. 76:1513–1522.[Abstract/Free Full Text]
Freking, B. A., and D. M. Marshall. 1992. Interrelationships of heifer milk production and other biological traits with production efficiency to weaning. J. Anim. Sci. 70:646–655.[Abstract]
Gregory, K. E., L. V. Cundiff, and R. M. Koch. 1992. Effects of breed and retained heterosis on milk yield and 200-day weight in advanced generations of composite populations of beef cattle. J. Anim. Sci. 70:2366–2372.
Marshall, D. M., and M. B. Long. 1993. Relationship of beef sire expected progeny difference to maternal performance of crossbred daughters. J. Anim. Sci. 71:2371–2374.[Abstract]
Marshall, D. A., W. R. Parker, and C. A. Dinkel. 1976. Factors affecting efficiency to weaning in Angus, Charolais and reciprocal cross cows. J. Anim. Sci. 43:1176–1187.[Abstract/Free Full Text]
Marston, T. T., D. D. Simms, R. R. Schalles, K. O. Zoellner, L. C. Martin, and G. M. Fink. 1992. Relationship of milk production, milk expected progeny difference, and calf weaning weight in Angus and Simmental cow-calf pairs. J. Anim. Sci. 70:3304–3310.[Abstract]
Miller, S. P., J. W. Wilton, and W. C. Pfeiffer. 1999. Effects of milk yield on biological efficiency and profit of beef production from birth to slaughter. J. Anim. Sci. 77:344–352.[Abstract/Free Full Text]
Minick, J. A., D. S. Buchanan, and S. D. Rupert. 2001. Milk production of crossbred daughters of high- and low-milk EPD Angus and Hereford bulls. J. Anim. Sci. 79:1386–1393.[Abstract/Free Full Text]
Neville Jr., W. E. 1962. Influence of dam’s milk production and other factors in 120- and 240-day weights of Hereford calves. J. Anim. Sci. 21:315–320.[Abstract/Free Full Text]
This article has been cited by other articles:
|
|
|
|
 
R. L. Sapp, R. Rekaya, and J. K. Bertrand
Teat scores in first-parity Gelbvieh cows: Relationship with suspensory score and calf growth traits
J Anim Sci,
August 1, 2004;
82(8):
2277 - 2284.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
