Printer Friendly

ACCURACY OF ESTIMATES FOR LIVE BODY WEIGHT USING SCHAEFFER'S FORMULA IN NON-DESCRIPT CATTLE (BOS INDICUS), NILI RAVI BUFFALOES (BUBALUS BUBALIS) AND THEIR CALVES USING LINEAR BODY MEASUREMENTS.

Byline: R. Riaz, M. N. Tahir, M. Waseem, M. Asif and M. A. Khan

Key words: Accuracy of estimation, Body weight, Cattle and Buffalo, Regression analysis

INTRODUCTION

Livestock owners and animal production professionals need to know the weight of the animals for several purposes including feeding, growth estimation, medication, breeding, as well as marketing age and provision of housing, keeping the welfare in mind. The most reliable information of body weight can be obtained by weighing equipment (e.g. weighing platforms, etc.). Large size commercial farms may harbor the facilities for live animal weighing, however, farmers with limited financial resources are unable to install modern weighing facilities. Live animal traders and farmers mainly use their visual appraisal for live weight estimation which can be biased due to human error (Machila et al., 2008; Abdelhadi et al., 2009).

The measuring tape method (Schaeffer's formula) is a cheap, easily accessible and more reliable than visual estimation, and it has been widely validated ( Otte et al., 1992; Heinrichs et al., 1992; Dingwell et al., 2006; Yan et al., 2009). The measuring tape method uses body measurements such as heart girth (HG), diagonal body length (DBL) and/or body height which are regressed against the formula to estimate live body weight (BW) in most species of livestock (Otte et al., 1992; Tariq et al., 2013; Siddiqui et al., 2015). Msangi et al. 1999) reported that a single linear body measurement such as HG can be used for estimating the BW of all age groups of dairy cattle with great accuracy. For various buffalo breeds a high correlation between HG and BW was reported by Tariq et al. (2013) and Francis et al. (2002).

Live BW is affected by the season of the year, species, age, sex, breed and the status of the animal (lactation, pregnancy, etc.) (Fitzhugh et al., 1967) resulting in variation between the estimated and observed BW. In the field, those small biases in the prediction of BW may increase or decrease the commercial value of the animals (price). Schaeffer's formula for live BW estimations has been widely validated across many breeds of dairy cattle and produced highly reliable results in exotic cattle breeds (Otte et al., 1992; Dingwell et al., 2006). However, this formula may present large errors in estimating live BW in local breeds of cattle and buffaloes in Pakistan. These breed to breed differences may question the reliability of Schaeffer's formula for the said purpose. Hence, it is important to validate the existing Schaeffer's formula for weight estimation to include more data pertaining to local breeds.

The objectives of the present study were to evaluate the effects of species and season on body weight estimates, investigate the relationship between observed BW and individual or combined linear body measurements, and assess the association between estimated BW (tape method) and observed BW (scale method) and develop an equation to improve the accuracy of live BW estimation for the use of farmers and other stakeholders based on body measurements for cattle, buffalo and calves.

MATERIALS AND METHODS

Description of animals and diets: The animals used were maintained at University Livestock Farm of University College of Veterinary and Animal Sciences, The Islamia University of Bahawalpur, Pakistan and cared according to the instructions of Animal Care and Management Committee (Bio-Ethics and Experimental Use Committee, 2015). Fifteen adult cows (>3 years), six female young cattle calves (1-3 years) belonging to local breeds (Bos indicus, non-descript) and four Nili-Ravi buffaloes (>3 years) were included at random. All animals included in this study were non-lactating and non-pregnant. All animals had free access to water and feed during the whole trial period. Animals were stall fed in groups consuming 20 and 30 kg fresh green fodder for cows and buffaloes, respectively, 2 kg wheat straw and 0.5 kg wheat bran per animal per day to fulfill the maintenance and growing requirements.

Weight estimation procedure and prediction of BW from linear body measurements: Measurements were taken when the animals were standing squarely with all four legs equally distanced under the body and head at its normal upright position. Heart girth was taken through a measuring tape, drawn from a point slightly behind the shoulder blade, down to the 6th rib position and under the body behind the elbow, as suggested by Pater et al. (2007) or as shown in Figure 1a. The average DBL was taken from the point of shoulder to the point of pin bone on each side as illustrated in Figure 1b to reduce bias in the measurements. The body weights were determined two times in a season e.g. winter (February-March) and summer (July-August) during consecutive days in morning and evening times, using cattle measuring tape (tape method) and digital weighing platform (scale method with measuring range 0-1000 kg, accuracy 0.1 kg).

The same person took measurements with both methods to exclude human factor related errors, and all measurements were taken before feeding. Body condition scoring was simultaneously performed on each animal by visual appearance on 1-5 scale with 0.5 score interval. The time required for weight estimation/measurement using both methods was about 20 minutes.

The data on morphometric measurements were used to predict the BW and subject to simple and multiple linear regression analyses. Akaike Information Criterion (AIC; Snipes and Taylor, 2014) was used to find the suitable number of predictors of BW from linear body measurements.

Association between estimated and observed body weight: To determine the association between the observed and estimated BW, data on buffaloes of this experiment and those of Tariq et al. (2013) from group 2 (>3-8 years, n=94) and 3 (>8 years, n=33) and Khan et al. (1978) for all age groups (n=350) were combined. For calves, the data of this experiment and those of Tariq et al. (2013) for age group 1 (1-3 years, n=84) were utilized.

Data treatment and statistical analysis: The measured parameters values for HG and DBL were used to estimate the BW using the following formula:

(Estimated BW) = (HG)2 X DBL/300 (Johnson Et al., 1939)

Whereas, HG = Heart girth, DBL = Diagonal body length

The collected data were checked for any outliers and were subject to General Linear Model. The relationships between all parameters were determined using simple and multiple linear regression analysis using MINITABA(r) statistical software version 13.1.1. The results were declared significant when P was a$?0.05. The following regression models were employed:

Observed BW = [alpha] + [beta] (HG) + (Equation 1)

Observed BW = [alpha] + [beta] (DBL) + (Equation 2)

Observed BW = [alpha] + [beta]1 (HG) + [beta]2 (DBL) + (Equation 3)

Observed BW = [alpha] + [beta] (estimated BW) + (Equation 4)

Whereas, [alpha] = intercept, [beta] = regression co efficient, BW = body weight, HG = Heart girth, DBL = Diagonal body length

RESULTS AND DISCUSSION

The observed and estimated body weight ranged from 261.0 to 422.5 kg and 232 to 415 kg for cattle, from 502.5 to 600.5 kg and 530 to 614.5 kg for buffaloes, from 116.0 to 149.0 kg and 125 to 136 kg for calves (table-1). The values of HG and DBL ranged from 154.9 to 179.9 cm and 127.0 to 145.1 cm for cows, from 195.6 to 210.8 cm and 40.9 to 152.4 cm for buffaloes, and from 117.8 to 122.1 cm and 97.3 to 98.8 cm for calves.

The animal species and season of the year significantly influenced (P<0.001) animal BW and body measurements. In general, for all species, the animals having an observed BW of 390 kg or more showed a weight change in the range of 2 to 24 kg in winter while those having an observed BW below 390 kg showed a difference of 1 to 16 kg (table-2). The average BW differences between species (adult animals) and seasons were 160 and 13 kg, respectively.

Average values of HG and DBL for buffalo are higher than reported by Tariq et al. (2013). The values of both variables from the present study are also higher than those of Milla and Mahjoub (2012) for Nilore cattle and calves but lower than those of Khan et al. (1978) for Nili Ravi buffaloes. These differences might be due to age and nutritional factors (Tariq et al., 2013). For individual animal noticeable differences can be seen for estimated and observed values using both methods.

It was observed that animals lost an average 13 kg weight during summer. Further, it was noticed that the animals having a live BW of 390 kg or above lost more than those animals with less than 390 kg BW. Contrary to these results, Spencer et al. (1988) reported that BW was higher during summer than in winter. They also described that when the sex of the animal was included as a factor with respect to BW estimation, the predictions were even poorer for females compared to males probably, because the females experience changes in body condition scoring as a result of many stress conditions like calving, lactation, different managment practices, etc.

Prediction of live body weight using linear body measurements: The single and multiple linear regressions between the observed BW, HG and DBL alone and in combination (HG plus DBL) were calculated (table-3). The results showed that the correlation coefficient between BW and HG varied from 0.80 to 0.98 and was the highest for buffaloes. All the values were statistically significant (P<0.001). The correlation coefficient between BW and DBL varied from -0.63 to 0.47 and all the values were non-significant except for cattle; the BW and DBL showed a negative correlation for buffalo. The co-efficients of determination for HG were higher (R2 = 0.65, 0.97 and 0.90, P<0.001) than that for DBL when HG was alone used to predict observed BW. These higher values explained that BW can be predicted with high accuracy using a single body measurement like HG.

Many published studies (Nesamvuni et al., 2000; Bagui and Valdez, 2007; Siddique et al., 2015) supported the observation of using HG as single linear measurement in adult cattle. The results of present study for adult buffalo also coincide with the results of Tariq et al. (2013) and Khan et al. (1978) who reported that highest accuracy of BW estimation was achieved from HG measurements. However, literature about the prediction of BW using HG in calves is quite equivocal. It was observed that HG measurements can equally accurately predict BW in calves as those in adult animals. Contrary to present study findings, results of many studies revealed that the highest accuracy in prediction of BW using single linear measurement can be achieved with DBL instead of HG. This is because there is highly positive correlation in DBL-estimated BW in growing animals (Naz et al., 2006). The same was observed by Davis et al. (1993).

They reported that R2 for DBL relative to BW decreases but that of HG relative to BW increases as animal matures because small animals have both a low BW and a small thoracic girth while the opposite is true for large animals as HG tends to increase significantly as the animal matures (Otte et al., 1992). The regression equation was highly significant (P<0.001) for all heart girth-estimated BW correlations for all species.

In order to improve the accuracy of prediction of observed BW from linear body measurement, multiple linear regression equations were developed. Interestingly, a combination of HG and DBL as predictors significantly improved R2, reduced SE of estimate and AIC in general for all species of animals. It is also apparent that the response in the above said parameters was more prominent in calves than in adult animals. Contrary to current findings, Milla and Mahjoub (2012) reported that HG and DBL show no predictable differences in growing calves.

While comparing accuracy of prediction of BW for same value of HG, we observed that the cows exhibited less accuracy compared to buffaloes. Many plausible reasons can be quoted here; the compact, wide and deep body of the buffalo as compared to that of cattle. Hump variation due to cross bred animals and age, brisket fat and dewlap may vary the results. Therefore, species to species effect on prediction of BW should be included in future research. But for the simplicity to the prediction of BW in mature animals, single linear body measurement i.e. HG should be considered because it explained more than 64% of the variation among all predictors and species.

The results of the present study indicated that the value of regression co-efficient significantly improved with multiple linear regression models for cattle and calves, but had no effect on buffalo. By close interpretation it was estimated that for every 1 cm increase in HG of cattle results in a corresponding increase of 4.24 kg whereas every 1 cm increase in each HG and DBL results in a corresponding increase of 3.89 Kg BW with constants shown in table-3. Similarly, a unit increase in HG of calves corresponds to a 5.88 Kg increase in animal BW while including both HG and DBL improves the regression co-efficient to a value of .68 Kg. However, the value of regression co-efficient reduced for buffalo when both body measurements were included in the model (4.37 Vs 4.79).

Association between estimated and observed body weight: Association between both methods varied as the data was interpreted (table-4). In the present experiment, estimated and observed BW were highly correlated (r = 0.83, 0.96 and 0.72: P<0.001) for cattle, buffalo and calves, respectively. The regression co-efficient valued 0.88, 0.78 and 0.93 for cattle, buffalo and calves, respectively when simple linear regression was used (R2 = 0.68, 0.92, 0.75: P<0.001).

Average weight differences (estimated-observed) for cattle were 24.7 Kg (range from -24.3 to 80.1), for buffaloes were 8.37 Kg (range from -77.1 to 96.5) and for calves were 17.8 Kg (range from 24.5 to 104.9) and the overall co-efficient of variation for observed vs. estimated BW were 8.0, 1.0, and 4.0% for cattle, buffalo and calves, respectively and these were slightly improved after the corrective regression equation was employed. However, Otte et al. (1992) showed that the variability between-method differences tended to increase as the animals became heavier. It was interesting to know that the co-efficient of variation was reduced to zero when the data were regressed against a live BW of 153 Kg for cattle, 453 Kg for buffalo and 109 Kg for calves. The Schaeffer's formula consistently overestimated live BW for smaller animals while the opposite was true for the heavier animals.

Although weight estimates by both methods were highly correlated, still a 100% accurate estimation of BW in live animals is not possible owing to limitation of current methods of data acquisition as well as factors already described responsible for weight changes. Luckily, a rough approximation to the nearest possible values can be achieved with the linear body measurements for various species of the animals. It should be noted that the considerable differences for individual animals can occur between weight estimates by the both methods. This between animal variations can be reduced by including more number of experimental animals. It is further advised that in trials where BW is an independent variable the large differences in estimated and observed BW may be misleading and reduce the usefulness of tape method.

For example, the classification of heifers on the basis of BW into groups with larger vs. small intervals from service to conception as suggested by Otte et al. (1992). While in the trials where the BW is dependent variable such as weaning weight of animals maintained at various dietary treatments, the larger differences between estimated and observed BW do not produce noticeable errors in interpreting data.

Table-1. Mean values and standard deviation (SD) of body measurements, estimated and observed body weight (BW) of cattle, buffalo and calves.

Variables###Cattle###SD###Buffalo###SD###Calves###SD

N###15###04###06

n###64###08###12

HG###165.5###8.64###200.5###7.57###121.3###2.59

DBL###136.6###4.35###147.3###3.90###91.6###10.57

Predicted BW1###347.3###41.75###547.6###33.00###131.9###17.08

Observed BW2###333.7###43.22###542.6###33.70###136.5###14.79

Table-2. Least squares means and standard error of the mean (SEM) for estimated and observed body weight (BW) of cattle, buffalo and calves as affected by species of the animal and season of the year.

Variables###Cattle###SEM###Buffalo###SEM###Calves###SEM###Significance

###Species###Season

HG###168.2###2.09###200.5###2.71###122.0###2.40###<0.001###<0.001

DBL###137.3###1.26###147.3###1.62###94.7###1.45###<0.001###0.014

Estimated BW1###347.3###10.21###547.6###14.09###131.8###11.73###<0.001###<0.001

Observed BW2###333.7###510.91###542.6###14.14###137.46###12.52###<0.001###<0.001

Table-3. Prediction of body weight from single linear body measurements for cattle, buffalo and calves using single and multiple linear regression equations (Equations 1 through 3).

Regression###HG vs. observed BW###DBL vs. observed BW###HG and DBL vs. observed

parameters###BW

###Cattle###Buffalo###Calves###Cattle###Buffalo###Calves###Cattle###Buffalo###Calves

Intercept ([alpha])###-377###-334###-580###-345###1348###-64###-530###-587###-598

Regression coefficient ([beta])

HG###4.24###4.37###5.88###---###---###---###3.89###4.79###5.68

DBL###----###---###---###4.95###-5.46###0.766###1.54###1.14###0.45

P-value###<0.001###<0.001###<0.001###<0.001###0.092###0.266###<0.001###<0.001###<0.001

R2###0.65###0.97###0.89###0.23###0.40###0.12###0.67###0.98###0.94

r###0.80###0.98###0.95###0.48###0.63###0.35###---###---###---

SE###25.88###6.64###4.95###38.25###28.19###14.53###25.43###6.25###4.06

AIC###418.4###31.9###40.2###468.4###55.1###66.0###417.1###31.6###36.2

Table-4. Association between estimated and observed body weight using simple linear regression model.

Parameters###Simple Linear Regression (Equation 4)

###Cattle###Buffalo###Calves

Descriptive statistics

N###15###477###90

n###64###485###96

HG###168.2###276.3###134.6

DBL###137.3###200.7###109.4

Estimated BW###347.3###491.0###192.4

Observed BW###333.7###482.6###179.0

Regression parameters

Intercept ([alpha])###18.29###98.25###19.94

Regression coefficient ([beta])###0.88###0.78###0.82

P-value###<0.001###<0.001###<0.001

R2###0.68###0.94###0.97

R###0.83###0.96###0.98

SE###24.62###23.2###11.43

COV###0.08###0.01###0.04

COV after###0.07###0.01###0.04

Conclusions: It was concluded that body weight in animals can be predicted with acceptable accuracy by linear body measurements such as heart girth as the sole predictor.

Acknowledgements: The authors wish to thank Dr. Muhammad Tariq and Bakht Baidar Khan for provision of their published data and the staff/personals working at the University Livestock Farm, The Islamia University of Bahawalpur for their continuous support and patience during the experiment.

Conflict of interest: No conflict of interest declared

REFERENCES

Abdelhadi, O.M.A. and S.A. Babiker (2009). Prediction of zebu cattle live weight using live animal measurements. Livest. Res. Rural. Dev. 21: Article No.133 from http://www.lrrd.org/lrrd21/8/abde21133.htm

Bagui, N.J.G. and C.A. Valdez (2007). Live weight estimation of locally raised adult purebred Brahman cattle using external body measurements. Philippines J. Vet. Med. 44: 36-42.

Davis, H.P., W.W Swet. and W.R. Harvey (1961). Relation of heart girth to weight in Holsteins and Jerseys, Historical Research. Bulletins of the Nebraska Agricultural Experiment Station (1913-1993), 112.http://digitalcommons.unl.edu/ardehistrb/112.

Dingwell, R.T., M.M. Wallace, C.J. Mclaren, C.F. Leslie and K.E. Leslie (2006). An evaluation of two indirect methods of estimating body weight in Holstein calves and heifers. J. Dairy Sci. 89: 3992-3998.

Fitzhugh Jr., H.A., T.C. Cartwright and R.S. Temple (1967). Genetic and environmental factors affecting weight of beef cows. J. Anim. Sci. 26: 991-1001.

Francis. J., J. Sibanda, J.E. Hemansen and T. Kristensen (2002). Estimating body weight of cattle using liner body measurements. Zimbabwe Vet. J. 23: 15-21.

Heinrichs, A.J., G.W. Rogers and J.B. Cooper (1992). Predicting body weight and wither height in Holstein heifers using body measurements. J. Dairy Sci. 75: 3576-3581.

Johnson, D.W. (1939). Livestock weights from measurements. Extension Folder 70, Minn Agri Ext.

Khan, B.B., M. Azhar, N. Ahmed and R.A. Chudhary (1978). Estimation of live weight of buffaloes from body measurements. Pak. J. Agric. Sci. 15: 89-94.

Machila, N., E.M. Fevre, I. Maudlin and M.C. Eisler (2008). Farmer estimation of live bodyweight of cattle: implications for veterinary drug dosing in East Africa. Prev. Vet. Med. 87: 394-403.

Milla, A.P. and M. Mahjoub (2012). Estimation of live body weight from heart girth, body length and condtion score in Nilotic Catlle -Southern Sudan. Anim. Sci. Adv. 2:453-457.

Msangi, B.S.J., M.J. Bryant, P.Y. Kavana and J.B. Kizima (1999). Body measurements as a management tool for crossbred dairy cattle at a smallholder farm condition. In: Proceedings of 26th Scientific Conference of Tanzania Society of Animal Production (Edited by Mbaga et al) 3rd-5th August 1999, LITI - Tengeru, Arusha, Tanzania. 168-175pp.

Naz, N.A. and M. Ahmad (2006). Genetic and phenotypic correlations for some sexual maturity traits in Nili Ravi buffalo heifers. Pak. Vet. J. 26: 141-143.

Nesamvuni, A.E., J. Mulaudzi, N.D. Ramanyimi and G. J. Taylor (2000). Estimation of body weight in Nguni-type cattle under communal management conditions. South Afric. J. Anim. Sci. 30 (Suppl. 1): 97-98.

Otte, M.J., A.J. Woods and Y. Abubara (1992). Live weight estimation of cattle by scale and by tape: a method comparison study. Trop. Anim. Health Prod. 24: 109-114.

Pater, S. (2007). How much does your animal weigh? Retrieved on April 02, 2013, from http://ag.arizona.edu/backyards/articles/winter.11-12 pp.

Siddiqui, M.U., M. Lateef, M.K. Bashir, M.Q. Bilal, G. Muhammad, M.I.Mustafa and S.U. Rehman (2015). Estimation of live weight using different body measurements in sahiwal cattle. Pak. J. Life Soc. Sci. 13: 12-15.

Snipes, M. and D.C. Taylor (2014) Model selection and Akaike Information Criteria: An example from wine ratings and prices. Wine Econom. Policy 3: 3-9.

Spencer, W., J. Eckert and P. Jakus (1988). Estimating live weight and carcass weight in Gambian N'dana cattle. World Anim. Rev. 65: 18-23.

Tariq, M., M. Younas, A.B. Khan and E. Schlecht (2013). Body measurements and body condition scoring as basis for estimation of live weight in Nili-Ravi buffaloes. Pak. Vet. J. 33: 325-329.

Yan, T., C.S. Mayne, D.C. Patterson and R.E. Agnew (2009). Prediction of body weight and empty body composition using body size measurements in lactating dairy cows. Livest. Sci. 124: 233-241.
COPYRIGHT 2018 Knowledge Bylanes
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2018 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Publication:Pakistan Journal of Science
Date:Sep 30, 2018
Words:4166
Previous Article:PREVALENCE OF ANAPLASMOSIS AMONG RISK GROUPS OF CATTLE AND BUFFALOES IN MIRPUR, AZAD JAMMU AND KASHMIR.
Next Article:ANTIMICROBIAL AND ANTIOXIDANT SCREENING OF Gardenia tetrasperma.
Topics:

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters |