Livestock Production Science 88 (2004) 9 – 16 www.elsevier.com/locate/livprodsci
Selection response of US Holstein AI bulls for milk production in Chile and Argentina R.A. Verdugo a, A.A. Jara b, R.W. Everett c, N.R. Barrı´a Pe´rez b,* b
a Department of Animal Science, University of California Davis, Davis, CA 95616-8521, USA Facultad de Ciencias Veterinarias y Pecuarias, Universidad de Chile, Casilla 2, Correo 15 La Granja, Santiago, Chile c Department of Animal Science, Cornell University, Ithaca, NY 14853-4801, USA
Received 3 June 2003; received in revised form 23 October 2003; accepted 18 November 2003
Abstract The correlated responses to selection of US AI dairy bulls from Chilean (45,447) and Argentinean (193,193) primiparous daughters lactating between 1992 and 1998 were estimated. Direct response obtained in the Northeast of the US (341,981) was also estimated for comparison with the Latin American results. Chile presented a higher production mean for 305-day milk yield (7162 kg) and fat (263 kg) than Argentina (6857 and 235 kg, respectively). Chile showed also a higher variability than Argentina. The responses were estimated using a regression coefficient of the adjusted production on the predicted transmitting abilities of the sires. The selection response for milk production was higher in Chile (0.566) than in Argentina (0.442) but lower than in the Northeast US (1.164). Similarly, the response for fat production in Chile (0.543) was almost half of the response in the US (1.101), and was very low in Argentina (0.095). Protein response could only be estimated in the US (1.209) and Chile (0.590). Chile and Argentina showed heterogeneity of responses by herd-year-mean (HYM) and herd-year-standard deviation (HYSD) levels of classification. The responses were lower than the response obtained in the US at similar levels. The differences in response were better explained by the levels of variation. D 2003 Elsevier B.V. All rights reserved. Keywords: Correlated response; Milk production; US Holstein sire; Chile; Argentina
1. Introduction Holstein breeding in Chile and Argentina has been initiated differently but has had a common develop-
Abbreviations: AI, artificial insemination; GE, genotype – environment interaction; HYM, herd-year-means; HYSD, herdyear-standard deviations; ME, mature equivalent; PTA, predicted transmitting ability. * Corresponding author. Tel.: +56-2-678-5572; fax: +56-2-6785611. E-mail address:
[email protected] (N.R. Barrı´a Pe´rez). 0301-6226/$ - see front matter D 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.livprodsci.2003.11.001
ment during the last few decades. The Argentinean Holstein herd started with Friesian cows brought from Holland in 1880 by President Julio A. Roca. They settled in three central provinces, but now they are spread over the region called ‘‘Pampa Hu´meda’’ in the middle Argentina and in some minor areas in the north of the country (ACHA, 2002). In southern Chile the base-herd was introduced by the German colonists consisting of black and white and red and white Friesian cows. Since then the dairy population in both countries has been genetically modified by selection and by mating with other highly specialized
10
R.A. Verdugo et al. / Livestock Production Science 88 (2004) 9–16
breeds, primarily American Holstein Friesian. This breeding system has been applied by importing frozen semen from the US mainly (Barrı´a and Stolzenbach, 1992). However, given that the selection of imported bulls has been achieved in a different environment, the productive response of these bulls from Latin American daughters could not be the same as that achieved from US daughters, due to Genotype by Environmental Interaction (GE) giving correlated responses less than 1.00. The GE effect has been identified in some Latin American countries by observed correlated responses of 0.31 in Colombia, 0.32 in Puerto Rico, and 0.54 in Mexico for milk (Stanton et al., 1991a). Correlated responses of 0.66 and 0.55 for milk and fat, respectively, were observed in Brazil (Costa et al., 1998). 1.1. Correlated response A trait measured in different environments can be considered as different but correlated characters instead of just one character (Falconer and Mackay, 1996). Therefore, the selection response in one environment will depend on the genetic correlation between environments, i.e. it will be affected by the importance of genetic – environmental interactions which cause changes in ranking of animals between environments (Dickerson, 1962). This can be applied to Latin America, where the obtained genetic change would be a mainly consequence of the correlated response to selection of dairy bulls in other countries. The correlated response could be smaller than direct response due to the presence of GE between countries of selection and production. Nonetheless, Powell and Norman (1984) found that the obtained response in the US was also affected by the herd production level. Even when the average selection response throughout production levels was close to the unit, it was higher when obtained only from first lactation daughters and was found to be positively associated with the production level, with values from 0.75 to 1.49 in the lowest and higher level, respectively. These results suggested presence of differences in heritabilities between production levels. Heterogeneous heritabilities have been empirically found by many authors, for several dairy breeds
(Lofgren et al., 1985). The heritabilities usually showed a positive trend by herd mean and standard deviation. Nevertheless, Meinert et al. (1988) clearly demonstrated that the heterogeneity of heritabilities, and hence of selection responses, is primary due to the within-herd variance. Therefore, the effect of the herdyear-mean (HYM) on selection response is only explained by its correlation with the within-herdyear-standard deviation (HYSD). 1.2. Correlated response in Latin America Stanton et al. (1991a) estimated correlated responses to selection of US bulls for milk yield in Colombia, Mexico and Puerto Rico. They used a coefficient of regression of milk production on sire predicted difference. The obtained responses in these countries were 0.32 F 0.02 for Colombia, 0.31 F 0.02 for Puerto Rico, and 0.54 F 0.02 for Mexico. The responses increased with HYM and HYSD class levels, but were lower to those reported by Powell and Norman (1984) for similar production and variability levels in the US. Afterward, looking at the causes of this heterogeneity of responses, Stanton et al. (1991b) estimated the genetic correlation between the US and Colombia (0.78) and the US and Mexico (0.90). They concluded that lower heritabilities and genetic correlations were insufficient to explain the heterogeneity of responses. Presence of pseudointeraction was most likely to cause less response in Latin America than in the US and in low versus high variance Latin American herds. However, Cienfiegos-Rivas et al. (1999) found a genetic correlation less than unity and a correlation of sires’ PTA ranking between the US and Mexico of less than 0.7. This indicated that real change of order, and then GE, was present between both countries. Costa et al. (2000) estimated the correlated response obtained in Brazil by selecting sires for production performance of half-sisters in the US. The largest correlated response was obtained in herd-yearseason (HYS) classes with high-standard deviation when using information from half sisters in lowstandard deviation classes (0.77 and 0.63 for milk and fat yield, respectively). They concluded for Brazilian herds using US semen, that the computing of bulls’ genetic values from daughters only in low-
R.A. Verdugo et al. / Livestock Production Science 88 (2004) 9–16
standard deviation herd levels would improve the obtained selection response. Cienfiegos-Rivas et al. (1999) concluded the same for Mexico. 1.3. Genotype by environment interaction in Chile Jara and Barrı´a (2000) (Reunio´n Latinoamericana) evaluated the presence of GE and they estimated variance components and relative importance in four herd production levels using a test day model, for milk, fat, and protein. The sire production-level interaction effect was not significantly important explaining only 1.22%, 0.87%, and 0.86% of phenotypic variability for milk, fat, and protein day production, respectively. Furthermore, they found genetic correlation high and positive between environments, measured by correlation within classes, indicating that bulls keep the same ranking in these four production levels for these traits. Nevertheless, heterogeneous variances were detected between environments, which were not completely explained by scale effects. Thus, productive responses in Chile are expected to be higher and lower in more and less variable production levels, respectively. Since profitability from using US Holstein semen in Latin American dairy herds depends on daughter milk response (Holmann et al., 1990) it is interesting to evaluate the obtained correlated responses in Chile and Argentina. The primary objective of this study was to evaluate the selection response obtained in Chilean and Argentinean dairy cows for production traits using semen from US AI Holstein sires. A secondary objective was to estimate the presence of heterogeneous correlated responses through herd milk production and variability levels for milk, fat, and protein yields, in Chile and Argentina.
2. Materials and methods Data were obtained from primiparous Holstein cows calving in the US, Chile and Argentina from 1992 to 1998. Records were collected from officially milk-recorded cows. Later lactations were not considered to avoid the selection effect of culled cows.
11
2.1. US information The information corresponded to the Northeast of the US and was provided by the Department of Animal Science of Cornell University, New York. Production records were standardized to 305 days, mature equivalent (305 ME), and two milking per day. These lactations were collected from 775,858 first parity lactations in 5463 dairy herds. 2.2. Argentinean information The Argentinean information was obtained from official records maintained by the ‘‘Asociacio´n Criadores de Holando Argentino’’ (ACHA). Data of 263,969 first parity lactations from 2138 herds were available. Records were obtained from two-milking systems and the lactations were standardized to 305 ME for milk and fat. 2.3. Chilean information Chilean data were also obtained from official records maintained by Cooprinsem, Osorno in the IX and X Regions in the South of Chile. Data of 45,447 first parity lactations from 180 herds were collected from two-milking systems. Standardized yields to 305 days in milk and mature equivalent were obtained for milk, fat, and protein. Data from these three countries were restricted to cows with an identified sire and were grouped by HYS. Arbitrary seasons were established by climatic conditions. Two seasons were considered in the Northeast of the US, from December to April of the next year, and from May to November. Two seasons were also considered for Argentina, from September to March and from April to August. For Chile, three seasons were established from December to February, from March to July, and from August to November. Only those HYS’s grouping six or more daughters and two or more sires were considered. A statistical description for milk, fat, and protein data after editing is presented in Table 1 for the US, Argentina, and Chile. The US presented higher production and variability levels than Chile and Argentina, and Chile were slightly superior to Argentina for milk and fat yields.
12
R.A. Verdugo et al. / Livestock Production Science 88 (2004) 9–16
Table 1 Number of records, mean, and standard deviation (S.D.), for milk, fat, and protein production adjusted 305 ME in US, Chile, and Argentina Trait
Milk (kg) Fat (kg) Protein (kg)
US
Chile
Argentina
n
Mean
S.D.
n
Mean
S.D.
n
Mean
S.D.
341,981 341,981 341,981
11,571 419 361
2314.3 83.4 68.5
39,074 39,074 20,722
7162 263 233
1728.9 59.1 56.8
193,193 193,193
6857 235
1360.1 46.4
Milk production means and standard deviations were calculated by herd-year (HY) groups. Data were classified by six HY production levels (HYM), from < 7000 to z 15,000 kg, and by six levels of within-HY standard deviation (HYSD), from < 500 to z 2500 kg.
Summaries, 2000). Only genetic evaluations from records in the US were included in the analysis. The selection response was estimated by a linear regression of standardized yields on the sires’ PTA’s, for US, Chile, and Argentina using the following model:
2.4. Genetic levels of US Holstein bulls
yijk ¼ l þ HYSi þ b1 X1j þ b2 X2j þ eijk
The genetic levels for milk and milk components were estimated using the bull-PTA (kg) weighted by the number of daughters of each bull. A description of the used US Holstein bulls is presented in Table 2. These mean values showed a superiority of the US AI bulls used in the US over Chile and in Chile over Argentina for every measured trait. Protein was the exception that averaged a larger PTA in Chile than in the US. Difficult to explain from a breeding point of view are the average PTA values for milk and fat in Argentina since the imported semen was obtained from proved bulls in the US. 2.5. Correlated response to selection Milk and milk components PTA’s of US Holstein bulls were utilized to estimate correlated response in Chile and Argentina. PTA’s were obtained from the August 2000 Sires Summaries of the USDA (Sires
where yijk is the standardized production yield of the k-daughter of the j-bull in the i-HYS class, l the population mean, HYSi the effect of the i-herd-yearseason, b1 the linear regression of yijk on X1j, X1j the PTA in kg for j-sire, b2 linear regression coefficient of yijk on X2j, X2j the date entered in service for j-sire and eijk is the experimental error. Solutions were obtained with the PROC GLM procedure of the SAS System package v 6.12. The effect of HYS was absorbed using the ABSORB statement before solving the equations due to the large number of levels for this effect. The date entered in service of the bulls would account for an unbalanced distribution of the new and old bulls along the years of study. If this variable is not considered, deviations of cows between only daughters of old bulls (beginning of study) would be unfairly compared to deviations on mainly new bull’s daughters (end of study). The correlated responses for the US were calculated to have a comparison with the Latin American countries. Therefore, the maximum expected response
Table 2 Number of observations, mean, and S.D., for milk, fat, and protein PTA weighted by number of daughters with a production record, in the US, Chile, and Argentina Trait
US
Chile
Argentina
n
Mean
S.D.
n
Mean
S.D.
n
Mean
S.D.
Milk PTA (kg) Fat PTA (kg) Protein PTA (kg)
331,382 331,382 331,382
135.6 4.5 3.1
292.2 9.9 7.7
25,236 25,236 1911
99.1 6.5 4.9
299.2 11.5 7.6
68,046 68,046
19.8 3.0
290.4 11.8
R.A. Verdugo et al. / Livestock Production Science 88 (2004) 9–16
for Chile and Argentina was assumed to be the same to that obtained in the US. The coefficients of the fitted linear regressions were estimated within HYM and HYSD. Only classes grouping at least 150 lactations were used to calculate the regression coefficients. Fewer regression coefficients were calculated in Chile for protein than the other traits because fewer cows were registered for this trait.
3. Results 3.1. Production and variability levels As expected, the Northeast of the US, having a larger dairy cow population than both Latin American countries, was represented in all six HYM and HYSD levels. However, Chile had a wider distribution of cows across HYM classes than Argentina, even with less than half of the Argentinean cows population. Table 3 shows the distribution of HY groups and lactation records by HYM and HYSD class. Only classes grouping 150 records or more are shown and were included in the study. Correlation coefficients were calculated between HYM and HYSD in the US, Chile, and Argentina, and presented in Table 4. These correlations were all less Table 3 Number of HY, and records (Cows) by milk production and variability levels for the US, Chile, and Argentina US HY HYM (kg) < 7000 7000 to < 9000 9000 to < 11,000 11,000 to < 13,000 13,000 to < 15,000 z 15,000 HYSD (kg) < 500 500 to < 1000 1000 to < 1500 1500 to < 2000 2000 to < 2500 z 2500 Total
Cows
Chile
Argentina
HY Cows
HY
Table 4 Number of observations (n), coefficient of correlation (r), and Pvalue ( P) under Ho: r = 0, between HYM and HYSD in the US, Chile, and Argentina US n
Chile r
P
n
67 467 178 2217 1795 19,136 373 14,941 3431 114,565 6089 101,849 398 21,353 1516 72,918 5563 163,507 50 2563 106 3469 1433 54,459 163 2563 15,110 341,981 640 39,074 5234 193,193
Argentina r
P
n
r
P
15,110 0.474 0.0001 840 0.603 0.0001 5234 0.459 0.0001
than unity and significantly different ( P < 0.0001) from zero in the US, Chile, and Argentina. The highest correlation was found in Chile (0.603) being more similar between the US and Argentina (0.474 and 0.459, respectively). 3.2. Correlated response 3.2.1. Milk yield The observed overall selection response in the Northeast of the US was 1.164 F 0.01. The overall and within HYM and HYSD levels responses are shown in Table 5. The correlated responses in Chile and Argentina were lower than that obtained in the Northeast US at the period of the study. Chile obtained half of the response presented by the US, with 0.567 F 0.03 while the response of Argentinean daughters resulted Table 5 Selection response coefficient (b) and standard error (S.E. {b}) for milk by HYM and HYSD in the US, Chile, and Argentina US
Chile
Argentina
b
S.E. {b}
b
S.E. {b}
b
S.E. {b}
HYM < 7000 7000 to < 9000 9000 to < 11,000 11,000 to < 13,000 13,000 to < 15,000 z 15,000
0.562 0.760 1.010 1.239 1.370 1.177
0.18 0.04 0.02 0.02 0.03 0.14
0.490 0.575 0.914 0.501 1.542
0.04 0.04 0.15 0.30 0.51
0.389 0.469 0.545
0.03 0.03 0.12
HYSD < 500 500 to < 1000 1000 to < 1500 1500 to < 2000 2000 to < 2500 z 2500 Overall
0.181 0.434 0.880 1.263 1.543 2.426 1.164
0.09 0.03 0.02 0.02 0.04 0.22 0.01
0.450 0.596 0.775
0.04 0.04 0.15
0.137 0.373 0.522 0.840
0.07 0.02 0.04 0.21
0.566
0.03
0.442
0.02
Cows
50 450 441 19,183 3343 110,191 1730 19,743 327 17,083 1796 79,745 6581 106,502 50 1943 95 3257 5265 143,406 19 697 1428 68,707 3 168 56 3173
13
14
R.A. Verdugo et al. / Livestock Production Science 88 (2004) 9–16
in an even smaller figure of 0.442 F 0.02 (Table 3). The level of response appears directly correlated with the production and variability levels in the three countries under study, suggesting a persistent effect of the heterogeneous variances on the obtained selection response. 3.2.2. Fat yield An overall response of 1.101 F0.01 for fat yield in the Northeast US (Table 6). It was observed both Latin American countries showed a lower selection response than the US. Chile obtained near to half the response of the US, with 0.545 F 0.03. Extremely low was the response showed by Argentina with an estimate of 0.095 F 0.02, which resulted statistically different from zero ( < 0.0001). Fat showed a similar pattern to the milk selection response by HYM and HYSD in the US (Table 5). A positive relation was much more evident by HYSD classes. Moreover, no trend for Chile and a slight negative trend for Argentina by HYM classes can be observed. These could indicate that the positive association between herd production mean and standard deviation, reported in the US for fat yield (Meinert et al., 1988), is not clearly present in Argentina or in Chile. However, it has to be considered the lower
Table 6 Selection response coefficient (b) and standard error (S.E. {b}) for fat by HYM and HYSD in the US, Chile, and Argentina US
Chile
Argentina
b
S.E. {b}
b
S.E. {b}
b
S.E. {b}
HYM < 7000 7000 to < 9000 9000 to < 11,000 11,000 to < 13,000 13,000 to < 15,000 z 15,000
0.661 0.711 0.985 1.172 1.231 1.210
0.20 0.04 0.02 0.02 0.03 0.14
0.474 0.592 0.550 0.564 0.584
0.04 0.04 0.13 0.24 0.67
0.114 0.090 0.025
0.02 0.02 0.09
HYSD < 500 500 to < 1000 1000 to < 1500 1500 to < 2000 2000 to < 2500 z 2500 Overall
0.768 0.744 0.982 1.184 1.157 1.359 1.101
0.18 0.06 0.02 0.02 0.04 0.21 0.01
Table 7 Selection response coefficient (b) and standard error (S.E. {b}) for protein by HYM and HYSD in the US and Chile US
Chile
b
S.E. {b}
b
S.E. {b}
HYM < 7000 7000 to < 9000 9000 to < 11,000 11,000 to < 13,000 13,000 to < 15,000 z 15,000
0.378 0.813 1.076 1.299 1.342 1.239
0.22 0.04 0.02 0.02 0.03 0.15
0.401 0.677 0.870 0.006 2.423
0.08 0.08 0.25 0.44 0.84
HYSD < 500 500 to < 1000 1000 to < 1500 1500 to < 2000 2000 to < 2500 z 2500 Overall
0.323 0.621 0.990 1.323 1.408 2.030 1.209
0.16 0.03 0.02 0.02 0.04 0.25 0.01
0.309 0.695 0.759
0.08 0.07 0.27
0.590
0.06
number of observations per HYM and HYSD levels in these countries, what do not allow for a more clear comparison. 3.2.3. Protein yield The US showed a selection response of 1.209 F 0.01, the larger among milk, fat and protein. Again, Chile obtained almost half of the US response, with an estimated regression coefficient of 0.596 F0.06 (Table 7). The protein response for Chile was not smooth across. However, an average positive trend may be depicted from the graph. A clearer relationship could be observed by HYSD classes stating this classification as a better method to account for differences in protein selection responses in Chile.
4. Discussion 4.1. Production and variability levels
0.406 0.619 0.515
0.04 0.04 0.13
0.000 0.094 0.095 0.171
0.543
0.03
0.095
0.07 0.02 0.03 0.15
0.02
From comparing production and variability levels between countries, the Latin American dairy herds were shown to be equivalent to those with the lowest management conditions in the US, as previously found in Colombia, Puerto Rico and Mexico by
R.A. Verdugo et al. / Livestock Production Science 88 (2004) 9–16
Stanton et al. (1991a). It has been shown that using genetic evaluations based on cows producing in low US HYSD can help to predict the selection response obtained in Latin American countries having more restrictive environments (Cienfiegos-Rivas et al., 1999; Costa et al., 2000). Since the correlation between production and variability levels of the herds was less than one in the three regions under study, the mean production would not completely represent differences in the withinherd variation. This suggests that it could be more accurate to use a direct measure of variation, like the HYSD, to take account of heterogeneity of responses (Meinert et al., 1988). The correlation in the US in the 1992 – 1998 period was 0.474, larger than that reported by Everett et al. in 1982 (cited by Meinert et al., 1988) of 0.24 for this country, but similar to the coefficient obtained in Argentina (0.459). However, Chile presented a clearly higher correlation of 0.603 for the same period of time. 4.2. Correlated response 4.2.1. Milk yield The overall selection response obtained in the Northeast US (1.164 F 0.01) was similar to the 1.20 F 0.01 reported for first lactations in the US by Powell and Norman (1984). In that study, a response larger than one was explained as a consequence from regressing first lactation data on PTA’s estimated from all lactations. These results support our estimations and render 1.164 as the theoretically expected response value for Chilean and Argentinean herds. Chile and Argentina showed a lower correlated response than expected from selection of US AI bulls as previously reported for five other Latin American countries by several authors (Costa et al., 2000; Cienfiegos-Rivas et al., 1999; Stanton et al., 1991a). The selection response for milk yield in the US, Chile, and Argentina was found to be positively associated with the production (HYM) and variability (HYSD) levels (Table 3) as it was by Powell and Norman (1984) in the US, by Stanton et al. (1991a) in Colombia, Me´xico, and Puerto Rico, by CienfiegosRivas et al. (1999) in Me´xico, and by Costa et al. (1998) in Brazil. The response obtained in higher levels of HYM and HYSD is larger than in the lower levels, but the differences are much clearer by the
15
second classification. This is in agreement with the conclusions of Meinert et al. (1988), who showed that the herd variance is the actual cause of heterogeneous selection responses. 4.2.2. Fat yield Selection response for fat yield in the US (1.101 F 0.01, Table 6) was fairly close to the response encountered by Powell and Norman (1984) for first lactations in this country (1.13). The response observed in Chile (0.545 F 0.03) was close to the 0.55 reported in Brazil (Costa, et al., 2000). However, the estimated correlated response in Argentina (0.095 F 0.02) is lower than any response reported in Latin America to the knowledge of the authors. 4.2.3. Protein yield Although protein yield is still poorly paid to dairy producers in Chile, it is thought to be an important item to be considered in future pricing policies. These results suggest that the producers will be able to improve this character as fast as they did for milk yield, by importing US semen as a strategy. The effect of the date entered service of the bull was significant for most of the HYM and HYSD levels for milk, fat and protein production in the three studied countries (data not shown). This was expected since the unequal distribution of old and new bulls along the period of the study. Therefore, accounting for this effect was required to obtain better estimates of the selection response. For these three traits, the classification by HYSD was more efficient to account for differences in selection responses. This and a lower average of HYSD levels in Chile and Argentina, compared to the US, raises the presence of heterogeneity of variances, between the US and the Latin American countries, as a cause for the smaller responses in Chile and Argentina. Nevertheless, the analysis of responses at equivalent HYSD levels showed a consistently smaller response in Chile than the Northeast US and in Argentina than Chile. No clear causes for these effects can be drawn from this study. Errors in the collection of production data in the Chile and Argentina or misclassification of the US sires could produce additional sources of error decreasing the obtained response to selection. However the authors are not aware of any reason why these effects would
16
R.A. Verdugo et al. / Livestock Production Science 88 (2004) 9–16
be higher in these countries than in the US. Therefore, further studies are needed to determine the other causes of these lower responses, not due to heterogeneity of variances. The rate of selection response for milk, fat, and protein yields in Chile represented nearly half of the US response for these traits. This, in connection with a higher cost of US semen doses, renders the real profitability of this investment to a much lower range for Chile. Argentina presented even a smaller milk correlated response than Chile and it was minimal for fat yield, making the previous conclusion more critical for this country.
5. Conclusions Correlated selection responses, for milk, fat and protein yields obtained in Argentina and southern Chile were less, on average, than direct response obtained in the Northeast of the US. Selection response in Argentina was less than in Chile. The direct response in the US was double the response obtained in Chile for milk, fat, and protein yields. The selection responses were positively associated with production and variability herd levels in both countries. However, the association between fat yield and HYM classes was shown to be null in Chile and negative in Argentina. The classification by HYSD could better explain differences in selection responses between and within countries for milk, fat and protein yields. The regression coefficients estimated in this study could help to predict future average daughter deviations in Chilean and Argentinean herds, by HYM and HYSD classes; to make more judicious decisions on selecting bulls based on potential profitability from investing in US bulls’ semen.
Acknowledgements This study was supported by a grant from Fondo Nacional de Ciencias y Tecnologı´a, Chile. Project No. 1000794. The authors would like to thank Asociacio´n Criadores de Holando Argentino (ACHA), Argentina;
Dr. Francisco Santiban˜ez, Cooprinsem, Osorno, Chile, for providing the original data.
References ACHA, 2002. Asociacio´n Criadores de Holando Argentino. http:// www.viarural.com.ar/. Barrı´a, N., Stolzenbach, G., 1992. Comparacio´n de la raza Friso´n Negro con la cruza Holstein x Friso´n Negro para produccio´n de leche y carne en la X Regio´n (Chile). Av. Prod. Anim. 17, 95 – 102. Cienfiegos-Rivas, E.G., Oltenacu, P.A., Blake, R.W., Schwager, S.J., Castillo-Jua´rez, H., Ruiz, F.J., 1999. Interaction between milk yield of Holstein cows in Mexico and the United States. J. Dairy Sci. 82, 2218 – 2223. Costa, C.N., Blake, R.W., Pollak, E.J., Oltenacu, P.A., 1998. Genetic relationship for milk and fat yield between Holstein populations in Brazil and the United States. Proceedings of Sixth World Congress on Genetic and Applied Livest Production, Armidale, Australia XXIII, pp. 323 – 326. Costa, C.N., Blake, R.W., Pollak, E.J., Oltenacu, P.A., Quaas, R.L., Searle, S.R., 2000. Genetic analysis of Holstein cattle populations in Brazil and United States. J. Dairy Sci. 83, 2963 – 2974. Dickerson, G.E., 1962. Implications of genetic – environmental interaction in animal breeding. Anim. Prod. 4, 47 – 63. Falconer, D.S., Mackay, M.T.C., 1996. Introduction to Quantitative Genetics, 4th ed. Oliver and Boyd, Edinburgh, UK. Holmann, F., Blake, R.W., Milligan, R.A., Barker, R., Oltenacu, P.A., Hahn, M.U., 1990. Economic returns from United States artificial insemination sires in Holstein herds in Colombia, Mexico and Venezuela. J. Dairy Sci. 73, 2179 – 2189. Jara, A., Barrı´a, N., 2000. Interaccio´n gene´tico ambiental para caracterı´sticas de produccio´n y puntaje de ce´lulas soma´ticas en vacas lecheras. XVI Reunio´n de produccio´n Animal y III Congreso Uruguayo de Produccio´n Animal. Marzo de 2000, Montevideo, Uruguay. Lofgren, D.L, Vinson, W.E., Pearson, R.E., Powell, R.L., 1985. Heritability of milk yield at different herd means and variance for production. J. Dairy Sci. 68, 2737 – 2739. Meinert, T.R., Pearson, R.E., Vinson, W.E., Cassell, B.G., 1988. Effect of within-herd variance and herd mean production on response to selection within herd. J. Dairy Sci. 71, 3405 – 3414. Powell, R.L., Norman, H.D., 1984. Response within herd to sire selection. J. Dairy Sci. 67, 2021 – 2027. Sires Summaries, 2000. Animal Improvement Programs Laboratory. http://aipl.arsusda.gov. Stanton, T.L., Blake, R.W., Quaas, R.L., Van Vleck, L.D., 1991a. Response to selection of US Holstein sires in Latin America. J. Dairy Sci. 74, 651 – 664. Stanton, T.L., Blake, R.W., Quaas, R.L., Van Vleck, L.D., Caraban˜o, M.J., 1991b. Genotype by environment interaction for Holstein milk yield in Colombia, Mexico, and Puerto Rico. J. Dairy Sci. 74, 1700 – 1714.
