Livestock Research for Rural Development 27 (10) 2015 Guide for preparation of papers LRRD Newsletter

Citation of this paper

Linear weight estimation tapes from predictive models for matured normal feathered Nigerian indigenous chickens

S T Vincent, A Yakubu1, O M Momoh and J OEgahi

Department of Animal Breeding and Physiology, College of Animal Science, University of Agriculture, Makurdi, Benue State, Nigeria
vincent.samuelter@gmail.com
1 Department of Animal Science, Faculty of Agriculture, Nasarawa State University, Keffi, Shabu-Lafia Campus, Nasarawa State, Nigeria

Abstract

This study was carried out to examine the relationship between body weight and linear body measurements in Nigerian normal feathered chickens; to develop predictive models and produce weight estimation tapes from the predictive models. Two-hundred adult normal feathered (94 males and 106 females) Nigerian indigenous chickens were randomly sampled between May and June, 2014. Traits measured were: body weight (BW), body length (BL), chest circumference (CC), neck length (NL), shank length (SL), beak length (BkL) and wingspan (WS). Data were subjected to ANOVA, bivariate correlation, simple and stepwise regression analysis.

Sex-associated differences were observed for body weight and all the zoometric traits, with higher values recorded for males. The coefficients of correlation between body weight and zoometrical traits were highly significant (P<0.01) in both sexes. For both sexes, the simple model regressing CC against BW had the highest coefficient of determination (R2 = 0.65 and 0.78 for females and males, respectively). CC, WS, SL and BkL were included as the best variables for predicting BW in females in the stepwise multiple regression models. However, in males, only CC and WS were included in the model. The existing relationship between BW and zoometrical traits was leveraged upon to develop a predictive model which was in turn used to produce the first BW estimation tapes for Nigerian matured normal feathered chickens. Rural farmers could easily comprehend and use such tapes than predictive models.

Keywords: correlation, dimorphism, stepwise regression, zoometrical traits


Introduction

In Nigeria, like in many other developing countries, the local chickens form the bed-rock of the poultry industry (Ikani and Annatte 2000). They constitute the majority of poultry types in Nigeria, with a population of about 103 million (RIM 1992) and more than 80% are raised in the rural areas (Momoh et al 2010). Their production and management is easier than broiler and pullet production and they are resistant to some diseases and can withstand stress conditions (Orajaka 2005). Ajayi (2010) opined that local chickens are adaptive to rural environments, survive on little or no input and adjust to fluctuation in feed availability.

Local chicken production contributes significantly to food security, poverty alleviation and ecologically sound management of natural resources. The scope for utilizing local chicken as a source of poultry meat is high because consumers prefer its hard meat. There is a growing demand for local chicken in restaurants because of its sizes, low prices and its palatable meat when compared to exotic breeds of poultry (Kperegbeyi et al 2009). African indigenous chickens also show higher dressing percentage in comparison with their exotic counterparts (Gueye et al 2001). Chickens are also used traditionally for carrying out rituals (Orajaka 2005), as a means of knowing the time, offered as gifts and in cementing marriages and friendship. In communities where food shortages are uncommon, chickens are kept to supplement the meals or to honour a guest (Nwagu 2002).

One major problem associated with local chicken production is marketing; due to its unorganized nature (Ikani and Annatte 2000). Prices of poultry products largely depend on body weight (or, indirectly, age) and poultry species (Gueye 2001). However, in Nigerian markets, birds are generally sold on face value and not by actual weight (Durosaro et al 2013) as is the practice in developed countries. This has been attributed to the rural nature of most poultry markets which are characterized by lack of measurement scales (Yakubu et al 2009). Hence, the need for the development of simple and affordable means for the estimation of weight with good accuracy. Relationships exist between body weight and linear body measurements (Ige 2013) and these have been leveraged upon for selection by local sellers and for research purpose (Olowofeso 2009). Linear body parameters such as chest circumference, body length, and thigh length, which are easy to measure, have also been used in the prediction of body weight in poultry (Momoh and Karshima 2008; Yakubu et al 2009).

More often than not, researchers only stop at regressing linear body measurements to produce predictive models. These models have not been able to achieve the overall objective of producing an alternative that is readily accessible and comprehensible even in species that have been extensively characterized, since most smallholder farmers can barely comprehend the models. The objective of this study, therefore, was to examine the relationship between body weight and linear body measurements in Nigerian indigenous chickens, to develop predictive models for precise estimation of body weight as well as weight estimation measurement tapes where a weight for a corresponding measurement can be read from the tape. This will eventually pave way for a better assessment of the market value of indigenous chickens and ultimately lead to more gains to the poultry farmers and marketers.


Materials and methods

Study location

The study was carried out in Lafia local government area of Nasarawa State, which falls within the guinea savanna zone of Nigeria. It is located between latitude 070 52’N and 08056’N and longitude 070 25’E and 09031’N, respectively.

Experimental birds

Two-hundred adult normal feathered (94 males and 106 females) Nigerian indigenous chickens were randomly sampled between May and June, 2014. The birds were reared under the traditional scavenging system.

Data collection

Body weight and six (6) morphometric traits were measured. The anatomical reference points were as described by Yakubu et al (2009), Durosaro et al (2013) and Adeleke et al (2011). The traits measured were:

Body weight (BW): Live weight when placed on a top loading measurement scale.

Body length (BL): Length of the body from the base of the beck to the base of the tail around the uropygial gland.

Chest circumference (CC): Taken under the wings at the edge of the sternum.

Neck length (NL): Distance between the occipital condyle and the cephalic border of the coracoids.

Shank length (SL): Distance from the shank joint to the extremity of the digitus pedis.

Beak length (BkL): Taken from the rectal apterium to the maxillary nail.

Wingspan (WS): Measured distance between the tips of the longest primary feather on each wings.

Zoometrical parameters (BL, CC, NL, SL, BkL and WS) were measured in centimeters with tailor’s tape while body weight of individual birds was measured using weighing balance with accuracy of 0.1kg.

Statistical analysis

Pearson’s coefficient of correlation (r) among body weight and all zoometrical traits were estimated using SPSS (2010). Zoometrical traits were regressed against body weight also using the same statistical package. The regression models used are depicted as:

Y = α + βX + e (1) Simple regression model

Y = α + β1X1 + β2X2 + … + βkXk + e (2) Multiple regression model

Where Y = dependent variable (BW),

α = the intercept,

β = regression coefficients associated with BL, CC, NL, SL, BL and BkL,

X = independent variables (BL, CC, NL, SL, BL, BkL),

e = error term.


Results and discussion

Descriptive statistics

Table 1 shows the means and standard error of means (SEM) of body weight and zoometric traits of sexes in normal feathered chickens. Sex associated differences were observed for body weight and all zoometric traits. Superior measurements recorded for males in all the parameters taken, was a demonstration of sexual dimorphism, which is presumed to reflect adaptive divergence in response to selection favoring different optimal character states in the two sexes (Blanckenhorn, 2005). This could partly be explained by male sex hormone which is responsible for greater muscle development in males than in females Semakula et al (2011). The observed female mean BW in the current work is lower than the value (1.45kg) reported by Ukwu et al (2014), but greater than the values (1.19kg) and (1.06kg) reported by Yakubu et al (2009) and Momoh and Karshima (2008), respectively. This could be attributed to the varying management systems and age of birds. Mean BW of male in the present study was lower than the value (1.37kg) reported by Yakubu et al (2009), but consistent with the findings of Fayeye et al (2006) and Ukwu et al (2014). Descriptive phenotypic identification is often needed to determine performance characteristics, morphology and purity of a breed (Sri Rachma et al 2013); and information on the structure of body morphometric traits is essential for understanding selection for phenotypic variability (Rosario et al 2008).

Table 1. Descriptive statistics of body weight and zoometrical traits of Nigerian normal feathered chickens

Trait

Male

Female

Mean (±SEM)

Mean (±SEM)

BW

1.32±0.01a

1.20±0.01b

CC

29.4±0.37a

26.1±0.26b

BL

30±02.32a

27.8±0.25b

WS

69.1±0.21a

67.9±0.15b

BkL

2.48±0.06

2.29±0.06b

SL

7.38±0.09a

6.92±0.09b

NL

9.05±0.13a

8.36±0.10b

Means on the same row bearing different superscripts differ significantly (P<0.05)

Phenotypic correlations

Coefficients of correlation between body weight and zoometrical traits of male and female sexes in normal feathered chickens are shown in Table 2. The correlation coefficients between body weight and zoometrical traits were highly significant (P<0.01) in both sexes. Highest correlation (r = 0.88 and 0.79 for males and females, respectively) was recorded between BW and CC. For both sexes, the least correlation coefficient was between BkL and NL (r = 0.26 and 0.21 for males and females, respectively). The association between BW and all the zoometrical traits is in agreement with the findings of earlier workers (Ige 2013; Sri Rachma et al 2013). Therefore, highly correlated traits are the basic indicators for estimation of the continuous prediction of body weight of chickens.

Table 2. Phenotypic correlation between body weight and zoometrical traits of male and female Nigerian normal feathered chickens

Trait

BW

CC

BL

WS

SL

NL

BkL

BW

-

0.80**

0.62**

0.79**

0.76**

0.42**

0.55**

CC

0.88**

-

0.74**

0.76**

0.72**

0.54**

0.44**

BL

0.71**

0.83**

-

0.61**

0.53**

0.39**

0.34**

WS

0.79**

0.72**

0.56**

-

0.68**

0.46**

0.42**

SL

0.66**

0.69**

0.65**

0.49**

-

0.39**

0.56**

NL

0.56**

0.60**

0.57**

0.44**

0.56**

-

0.21*

BkL

0.52**

0.49**

0.36**

0.41**

0.50**

0.26*

-

* = P<0.05, ** = P<0.01 The correlation coefficients above the diagonal represent those of females while the once below represent those of males

Simple predictive models

Table 3 shows the simple predictive models relating BW and zoometric traits of both sexes in normal feathered chickens. For both sexes, the model regressing CC against BW had the highest coefficient of determination (R2 = 0.65 and 0.78 for females and males, respectively). R2 signifies the goodness of fit of a regression (Durosaro et al 2013) and thus, the higher the value, the better the variance that the dependent variable is explained by the independent variable. From this study, CC could be used in both sexes to predict BW with greater precision since it explained above two-third variability in BW. The present results corroborate the findings of earlier workers (Saatci and Tulku 2007; Raji et al 2009; Ige 2013) who all reported highest R2 value for CC in a linear regression model.

Table 3. Predictive simple linear model of body weight from zoometrical traits of Nigerian normal feathered chickens

Sex

Parameters

Predictive model

SE

R2

p

Female

CC

BW = 0.23 + 0.04CC

0.00

0.65

0.00

BL

BW = 0.38 + 0.02BL

0.00

0.38

0.00

WS

BW = -3.10 + 0.07WS

0.00

0.63

0.00

SL

BW = 0.51 + 0.10SL

0.01

0.58

0.00

NL

BW = 0.79 + 0.04NL

0.01

0.17

0.00

BkL

BW = 0.93 + 0.12BkL

0.02

0.30

0.00

Male

CC

BW = 0.28 + 0.04CC

0.00

0.78

0.00

BL

BW = 0.32 + 0.03BL

0.00

0.51

0.00

WS

BW = -4.08 + 0.8WS

0.01

0.62

0.00

SL

BW = 0.48 + 0.12SL

0.01

0.43

0.00

NL

BW = 0.73 + 0.07NL

0.01

0.31

0.00

BkL

BW = 1.01 + 0.13BkL

0.02

0.27

0.00

SE = standard error, R2= coefficient of determination

Stepwise multiple predictive models

The stepwise multiple predictive models relating BW and Zoometrical traits are presented in Table 4. These revealed that CC solely accounted for 65% of variability in BW in females. The inclusion of WS increased the proportion of the explained variance to 72%. When SL and BkL were further included in the model, the precision of the model increased as the proportion of the explained variable increased to 78%. In males, CC solely accounted for 77% of the variability in BW. With the inclusion of WS, 82% of the variability was accounted for with a corresponding improvement in the accuracy of the model. Apart from CC which has widely be reported as a good weight estimator, Getu et al (2013) found WS as another parameter of importance for weight determination in both male and female chickens.

Table 4. Stepwise multiple regression of BW from zoometrical traits of Nigerian normal feathered chickens

Parameters

intercept

R

SE

R2

p

Female

0.000

1

CC

0.23

0.04

0.00

0.65

0.000

2

CC

-1.69

0.02

0.00

0.72

0.000

WS

0.03

0.01

0.000

3

CC

-1.41

0.01

0.00

0.76

0.000

WS

0.03

0.01

0.000

SL

0.04

0.01

0.000

4

CC

-1.38

0.04

0.00

0.78

0.000

WS

0.03

0.01

0.000

SL

0.03

0.01

0.001

BkL

0.03

0.01

0.014

Male

1

CC

0.28

0.04

0.00

0.77

0.000

2

CC

-1.61

0.03

0.00

0.82

0.000

WS

0.03

0.01

0.000

R = regression coefficient, SE = standard error, R2= coefficient of determination

Optimum weight estimation

Due to the difficulties associated with incorporating the different independent variables in a single weight estimation tape, the simple model was used to produce chicken weight estimation tapes for both sexes of Nigerian indigenous matured normal feathered chickens (Table 5). For both sexes, CC was regressed against BW with 0cm representing 0.23Kg in matured female chickens and 0cm representing 0.28Kg in matured male chickens.

Table 5. Estimation of BW using simple model for Nigerian normal feathered chickens

Sex

CC

Equation

Estimated weight

Female

0

BW = 0.23 + 0.04(0)

0.23

5

BW = 0.23 + 0.04(5)

0.43

10

BW = 0.23 + 0.04(10)

0.63

15

BW = 0.23 + 0.04(15)

0.83

20

BW = 0.23 + 0.04(20)

1.03

25

BW = 0.23 + 0.04(25)

1.23

30

BW = -3.25 + .04(30)

1.43

35

BW = 0.23 + 0.04(35)

1.63

Male

0

BW = 0.28 + 0.04(0)

0.28

5

BW = 0.28 + 0.04(5)

0.48

10

BW = 0.28 + 0.04(10)

0.68

15

BW = 0.28 + 0.04(15)

0.88

20

BW = 0.28 + 0.04(20)

1.08

25

BW = 0.28 + 0.04(25)

1.28

30

BW = 0.28 + 0.04(30)

1.48

35

BW = 0.28 + 0.04(35)

1.68

This Table above was used to produce the linear weight estimation tapes of the Nigerian matured normal feathered chickens (Figures 1 and 2), where “1” represent that of females and “2” that of males. On the scale, linear distances are recorded against corresponding weight which is more comprehensible than predictive models. These tapes can now be used by farmers in the absence of weighing scales which are costlier. With these, poultry farmers can make better gains from chicken rearing partly because there are standards for determining the economic value of their birds instead of subjective weight determination by hand. Semakula et al (2011) reported that farmers made more profits by estimating live body weight from body measurements.

 Figure 1. Linear weight estimation tape for matured normal feathered female chickens.  Figure 2. Linear weight estimation tape for matured normal feathered male chickens.


Conclusions


References

Adeleke M A, Peters S O, Ozoje M O, Ikeobi C O, Bamgbose A M and Adebambo O A 2011 Genetic parameter estimate for body weight and linear body measurements in pure and crossbred progenies of Nigerian indigenous chickens. Livestock Research for Rural Development 23(1) Article #19. Retrieved November 30, 2012, from http://www.lrrd.org/lrrd23/1/faye123019.htm

Ajayi F O 2010 Nigerian indigenous chicken: A valuable genetic resource for meat and egg production. Asian Journal of Poultry Science 4: 164-172.

Blanckenhorn W U 2005 Behavioral causes and consequences of sexual size dimorphism. Ethology 111: 977–1016.

Durosaro SO, Oyetade M S, Ilori B M, Adenaike A S, Olowofeso, O, Wheto M, Amusan S A, Osho S O and Ozoje M O 2013 Estimation of body weight of Nigerian local turkeys from zoometric measurements at 4, 8 and 12 weeks of age. Global Journal of Science Frontier Research (Bio-Tech and Genetics) 13(1):

Getu A Alemayehu K and Wuletaw Z 2014 Phenotypic characterization of indigenous chicken ecotypes in North Gondar Zone, Ethiopia. American-Eurasian Journal of Scientific Research 8 (6): 248-255

Gueye E F 2001 Marketing of family poultry products in Africa to be improved. World Poultry-Elsevier 17: 12-16.

Fayeye T R, Ayorinde K L, Ojo V and Adesina O M 2006 Frequency and influence of some major genes on body weight and size parameters of Nigerian local chickens. Livestock Research for Rural Development. Volume 18, Article #37.Retrieved April 14, 2015, from http://www.lrrd.org/lrrd18/3/faye18037.htm

Ige A O 2013 Estimation of genetic parameters in Yoruba and Fulani ecotypes indigenous chickens of Nigeria. Transnational Journal of Science and Technology 3: 1-28.

Ikanni E I and Annatte A I 2000 Improving the performance of local chickens. Extension Bulletin No 92. Poultry series No 6. National Agricultural Extension and Research Liaison Services, ABU Zaria. Pp 30

Kperegbeyi J I, Meye J A and Ogboli E 2009 Local chicken production: strategy of household poultry development in coastal region of Niger Delta, Nigeria. African Journal of General Agriculture 5: 17-20

Momoh O M, Nwosu C C and Adeyinka I A 2010 Comparative evaluation of two Nigerian local chicken ecotypes and their crosses for growth traits. International Journal of Poultry Science 9: 738-743.

Momoh O M and Karshima D E 2008 Linear body measurements as predictors of body weight in Nigerian local chickens. Asset Series A8 (2): 206-212.

Nwagu B I 2002 Production and Management of Indigenous Poultry Species. A Training Manual in National Training Workshop on Poultry Production at Nigerian National Animal Production Research Institute, Shika, Zaria. Pp. 10-26.

Olowofeso O 2009 Phenotypic correlations and prediction of body weight and size parameters in broiler chickens. Journal of Applied Agricultural Research 1: 71-76.

Orajaka L J E 2005 The role of local chicken in poultry production in Nigeria. Nigerian Veterinary Journal 26: 68-72

Raji A O, Igwebuike J U and Usman M T 2009 Zoometrical body measurements and their relation with live weight in matured local Muscovy ducks in Borno state Nigeria. ARPN Journal of Agricultural and Biological Science 4: 58-62.

RIM 1992 Nigerian livestock resources. Vol. II. National synthesis. Annex Publ. Resource Inventor Management Ltd.

Rosario M F, Silva M A N, Coelho A A D and Savino V J M 2008 Selection of traits in poultry breeding using cluster analysis. International Journal of Poultry Science 7: 374-378.

Saatci M and Tulku M 2007 Zoometrical body measurements and their relation with live weight in Native Turkish Geese. Turkish Journal of Veterinary and Animal Science 31: 47-53.

Semakula J, Lusembo P, Kugonza D R, Mutetikka D, Ssennyonjo J and Mwesigwa M 2011 Estimation of live body weight using zoometrical measurements for improved marketing of indigenous chicken in the Lake Victoria basin of Uganda. Livestock Research for Rural Development. Volume 23, Article #170. Retrieved July 4, 2015, from http://www.lrrd.org/lrrd23/8/sema23170.htm

SPSS 2010 Statistical Package for the Social Sciences. SPSS Inc., 444 Michigan Avenue, Chicago.

Sri Rachma A B, Harada H, Dagong M I A, Rahim L and Prahesti A I 2013 Study of body dimension of Gaga’s chicken, germplasm of local chickens from south Sulawesi-Indonesia. International Journal of Plant, Animal and Environmental Science 3: 204-209.

Ukwu H O, Okoro V M O and Nosike R J 2014 Statistical modeling of body weight and linear body measurements in Nigerian indigenous chicken. IOSR-Journal of Agriculture and Veterinary Science 7: 27-30.

Yakubu A, Kuje D and Okpeku M 2009 Principal components as measures of size and shape in Nigerian indigenous chickens. Thai Journal of Agricultural Science, 42:167-176.


Received 7 April 2015; Accepted 31 August 2015; Published 1 October 2015

Go to top