The accuracy of prediction of body weight from body measurements in beef cattle

The objective of this study was to determine the accuracy of prediction of body weight from body measurements in beef cattle. Wither height, chest girth, body length, chest depth, hip width and hip height measurements were obtained from Holstein, Brown Swiss and crossbred (n=140). Determination coefficients (R2) of regression equation that included all body measurements were higher in Brown Swiss and crossbred than Holstein (92.2, 95.0 and 68.2 %, respectively). However, it was found that chest girth was the best parameter of all for prediction of body weight in Brown Swiss (R2=91.1 %) and crossbred cattle (R2=88.8 %) in comparison to Holstein (R2=60.7 %). According to these results, the body weight estimation of Brown Swiss and crossbred cattle using the body measurements produced higher prediction accuracies than Holstein but chest girth was the best parameter to prediction of body weight among all body measurements. However, the prediction accuracy of prediction of body weight from body measurements and also chest girth was decreased when the animals frame size was increased.


Introduction
Body weight of animals is an important factor associated with several management practices including selection for slaughter or breeding, determining feeding levels and also it is a good indicator of animal condition (ULUTAS et al. 2001).
The relationship between body measurements and body weight depends upon breed, age, type, size, condition and fattening level of the animals (HEINRICHS et al. 1992, YANAR et al. 1995).VAN MARLE-KÖSTER et al. (2000) described body measurements as selection criteria for growth in cattle.HEINRICHS et al. (1992) indicated that body measurements can be used for prediction of body weight.GILBERT et al. (1993) reported that there is close correlation between body weight and body measurements.MSANGI et al. (1999), SLIPPERS et al. (2000), FOUIRE et al. (2002), WILLEKE and DÜRSCH (2002) and BOZKURT (2006) indicated that chest girth can be used to predict body weight that it is the best prediction parameter.CAGLAR and SEKERDEN (1993) declared that the regression equations must be determined for all beef breeds for different country and region.
The aim of this study was to determine the accuracy of prediction of body weight from metric body measurements in beef cattle.

Material and methods
This study were carried out in Isparta and Burdur provinces in the Mediterranean part of Turkey and 140 male animals in total were used and comprised of 56 Holstein (body weight ranging 337 to 677 kg), 30 Brown Swiss (body weight ranging 326 to 930 kg) and 54 Crossbred cattle (body weight ranging 326 to 677 kg).
Body weight of animals was determined by using a digital weighing scale prior to slaughter (Marmara 0580 MEB).The parameters such as body weight, chest girth, wither height, body length, chest depth, hip width and hip height were measured using measuring stick and tape (Hauptner, Germany) when animals were standing as described in OZKAYA and BOZKURT (2008).
The best prediction equations for body weight from other traits (chest girth, body length, wither height, chest depth, hip width and hip height) as independent variables were determined.Descriptive statistics and regression analysis of body weight on each of the independent variables were performed using the MINITAB, 13 Inc (2001).Comparisons between means were determined by Tukey test.
Correlation coefficients were also obtained from parameters.Linear, quadratic and cubic effects of independent variables on body weight were included in the following model: where is Yi the body weight observation of an i-th animal, b0 the intercept, b1, b2, b3 the corresponding linear, quadratic and cubic regression coefficients, xi the body measurement (chest girth, body length, wither height, chest depth, hip width and hip height) and ei the residual error term.

Results and discussion
Descriptive statistics of body weight and body traits are shown in Table 1.The parameters of Holstein were higher and statistically significant (P<0.05)than Brown Swiss and crossbreds.All parameters were found no significant between Brown Swiss and crossbred cattle but only wither height values were found statistically significant in all breeds (P<0.05).The best regression equations of body weight on various body measurements are shown in Table 2. Results of regressions of body weight on the linear, quadratic and cubic effects of each body measurements are presented in Table 3. Chest girth was statistically significant for all breeds (Table 2).The R 2 value for Holstein was obtained from the equation contained body length, wither height, chest depth and chest girth was found 66.1 %.This result was lower than findings of TUZEMEN et al. (1995) who reported R 2 =90.7 %.The R 2 value obtained from the equation contained body length, wither height and chest girth was found 66.1 % and this result was lower than findings of SEKERDEN et al. (1991) (R 2 =97.7 %).R 2 value was obtained from the equation contained only chest girth (R 2 =60.7 %) was lower than those findings of SEKERDEN et al. (1991) (R 2 =97.3 %) and TUZEMEN et al. (1995) The highest R 2 value for Brown Swiss were obtained from the equation contained all body measurements (R 2 =92.2 %) (Table 2).R 2 value was determined as 91.8 % which contained body length, wither height, chest depth and chest girth.This result was in line with findings of TUZEMEN et al. (1993) (R 2 =90.7 %) and BOZKURT (2006) (R 2 =93.6 %).R 2 value was obtained from the equation contained only chest girth was higher than those findings of TUZEMEN et al. (1993) and BOZKURT (2006) (R 2 =91.1, 86.9 and 89.9 %, respectively).The highest R 2 value for crossbred was obtained from equation contained all body measurements (R 2 =95.0 %).In addition, R 2 value was found 88.8 % which included only CG (Table 2).The highest R 2 value was obtained from chest girth for all breeds (Table 3).For Holstein, R 2 was found 61.5 % and this result was lower than those findings of HEINRICHS et al. (1992) and WILSON et al. (1997) (R 2 =95 and 97 %, respectively).In the present study, results showed that when the body weight increased to 500 kg, the prediction accuracy of body weight from chest girth was decreased for Holstein (R 2 =39.4 %).
The highest R 2 value was found in body length and chest girth for crossbreds (Table 3) (R 2 =82.2 and 88.8 %, respectively).The cubic term was statistically significant for body length and chest girth (P<0.05).
The correlation coefficients of traits are shown in Table 4.The highest correlation was obtained between body weight and chest girth.The correlation coefficient (r=0.78) between body weight and chest girth for Holstein was lower than those findings of SEKERDEN et al. (1991) andTUZEMEN et al. (1995) (r= 0.99 and 0.83, respectively).For Brown Swiss, correlation coefficient between body weight and chest girth was found 0.95.These result was in line with findings of BOZKURT ( 2006) but was higher than YANAR et al. (1995) (r=0.86).For crossbreds, the correlation coefficient between body weight and chest girth was similar with Brown Swiss but higher than Holstein (Table 4).
In conclusion, this study showed that prediction accuracy of body weight using metric body measurements in Brown Swiss and crossbred was higher than Holstein.However, the prediction accuracy of chest girth was higher than other traits for prediction of body weight.However, prediction accuracy of body weight using metric body measurements was decreased in big size animals.

Table 1
Descriptive statistics for body weight and body measurements Beschreibende Statistik für Körpergewicht und Körpermaße a,b,c means in a column bearing different superscript are significantly (P<0.05)different

Table 2
The best prediction equations of body weight Beste Vorhersagegleichungen des Körpergewichtes

Table 3
Regressions of body weight on linear, quadratic and cubic effects of each body measurements Effekte linearer, quadratischer und kubischer Regressionsmodelle bei einzelnen Körpermaßen zum Körpergewicht

Table 4
Correlation coefficients between body weight and body measurements Korrelationskoeffizienten zwischen Körpergewicht und Körpermaßen