Egg Production Performance in a Nigerian Local Chicken Ecotype Subjected to Selection

This study aims to contribute towards the genetic improvement of Nigerian local chicken ecotype through selection. Genetic parameters for body weight at first egg (BWFE), egg number (EN) and egg weight (EW) till first 90 days of lay were estimated for both selected and control lines. Selection was based on an index using BWFE, EN and EW as the selection criterion traits. After three generations of index selection, BWFE, EN and EW all improved significantly (P<0.05) in the selected line. The heritability estimates for all traits in the three generations for both lines were moderate to high (BWFE, 0.33-0.56; EN, 0.19-0.28; EW, 0.25-0.44). Low to high positive genetic and phenotypic correlation was observed between BWFE and EW. The genetic and phenotypic correlation between BWFE and EN, and EW and EN were generally moderate to highly negative in both lines for all generations. However, in the second generation of the selected line a positive genetic correlation (0.33) was observed between EW and EN.


Introduction
Nigeria has rich chicken genetic resources.It has been reported by Resource Inventory management (RIM) (1992) that the local chicken contribute 80% of the 120 million poultry types raised in the rural areas of Nigeria.The local chicken, therefore, play a vital role in household food supply/income.It is known to be hardy and highly adapted to the harsh hot and humid environment, however, its productive potentials has not been fully improvement in egg production traits under improved management (Nwosu et al., 1979;Adebambo et al., 1999;Momoh et al., 2007).
The egg production of the local chicken is a result of many genes acting on a large number of biochemical processes, which in turn control a range of anatomical and physiological traits.With appropriate environmental conditions (nutrition, light, ambient temperature, water, sound health, etc.), the many genes controlling all the processes associated with egg production can act to allow the chicken to express fully its genetic potentials (Fairfull and Gowe, 1990).Altering and improving the environment, physiological situation or manipulation of these birds though contribute immensely towards improvement of their production qualities, the possibility remains that variation in their productivity exists after optimum non-hereditary conditions have been established.A more permanent approach towards a sustainable productivity is genetic improvement, which can be achieved through selection and crossbreeding (Szwaczkowski, 2003).
The first strategy requires extensive sampling of local chickens throughout the country to select desirable genetic material that will constitute the base population, which will subsequently be crossed either with themselves or with other populations.Success of this strategy will largely depend on the initial response to selection.Though within-breed selection gives slowest genetic improvement especially if the generation interval is long, however, this improvement is permanent and cumulative, which is not the case for cross-breeding programs.Crossbreds from purebred parents show heterosis to the extent that their gene frequencies differ unlike hybrids from similar lines that manifest total heterosis.Several studies have been reported on crossbreeding of the local chicken with exotic breeds (Nwosu and Omeje, 1985;Adedeji et al., 2008, Adebambo et al., 2009).There has been no detailed information on selective breeding of the Nigerian local chicken.This paper reports results pertaining to the first three generations of selection based on selection index for increased body weight at first egg, and short-term egg production traits (egg weight and egg number) in a Nigerian light ecotype chicken population.

Experimental Site
The study was carried out at the Teaching and Research Farm of the Department of Animal Science, Ebonyi State University, Abakaliki, Ebonyi State, Nigeria.Abakaliki is located between Latitude 06 o 4'N and Longitude 08 o 65'E in the derived savanna ecological zone of Nigeria.Naturally, the day length of Abakaliki ranges between 12-14 hours all year round, it has an annual mean rainfall range of between 1500-2250mm with mean daily temperature ranges of 27 o C and relative humidity of 85% (Nwakpu, 2005).The experiment lasted for four years from 2003 to 2007 in which data on egg production traits were collected.

The Base Population and Management of Experimental Birds
The base population for the selection experiment was obtained from the 142 Light Ecotype chicken -LEmaintained at the poultry farm of the Department of Animal Science, University of Nigeria, Nsukka as non-pedigreed, unselected and unimproved random mating population.The light ecotype represents the chicken type obtained from the swamp, rainforest and derived savannah agro-ecological zones whose mature body weight ranges from 0.68 to 1.5 kg (Momoh and Nwosu, 2008).Fifty-five LE (5 cocks and 50 hens) were randomly allotted to 5 breeding pens and allowed to mate naturally in a mating ratio of 1 cock to 10 hens.Eggs from the breeding population were collected over a period of 5-days and fumigated with 100ml of 20% formalin before they were set in the incubator.The eggs in each generation, prior to incubation, were identified according to pedigree (sire) using an indelible marker pen.A total number of 294 day-old chicks were hatched at Nsukka and transferred to Abakaliki for the study, this served as the foundation population (G 0 ).On arrival, all the chicks were weighed, wing-banded according to sire and brooded separately for eight weeks in deep litter pens.On the tenth week, sexing and separation of the males from the females were done using secondary sexual characteristics (comb size and shape of the tail feather).The pullets were reared in replicate pens until 18 weeks of age, when they were randomly assigned to individual battery cages measuring 17 x 32 x 32cm.
From 282 birds alive by the tenth week, 30 males and 30 females were randomly allotted into the control line from which the subsequent control generations (G 1 and G 2 ) were obtained.The control population spanned for three generations (each generation had its own control population of same age).These were used to monitor environmental changes and to estimate genetic change(s) due to selection.The remaining birds served as the whole population from which the selected line were obtained.The population size for the different generations of study is presented in Table 1.
In all generations and lines, feed and clean drinking water were given ad libitum throughout the experimental period.The birds were fed diets formulated by Bendel Feed and Flour Mill Ltd, Benin, Edo State Nigeria.The birds were fed chick ration during the first eight weeks (21%CP, 2750ME (kcal/kg); grower's diet (15%CP, 2300ME (kcal/kg) was fed between eight and eighteenth week of age and from the eighteenth week of age until after the breeding phase, layers' diet (17%CP, 2800ME (kcal/kg) was fed to the pullets (subsequently hens) whereas the males were maintained on the grower's diet until the breeding phase when they were fed similar diet as the females.The birds were vaccinated against the major poultry diseases prevalent in the study area, these included Newcastle disease (i/o, lasota, and komarov), infectious bursal disease, and fowl pox.Other medications such as antibiotics, coccidiostats, and vitamins were administered via drinking water as the need arose.Strict sanitary measures were adhered to.

Traits Measured
Data were collected and evaluated from the selected and control lines for three generations on the following egg production traits: Age at first egg (days) (AFE) -this was taken as the number of days from hatch to the day the first egg was laid provided the second egg was laid in the next ten days; Body weight at first egg (g) (BWFE) -this reflected the live weight of each pullet on the first day it laid egg; Weight of first egg (g) (WFE) -the weight of the first egg laid by each hen as obtained soon after lay; Egg weight (EW) -eggs laid by each hen was weighed on daily basis.The average egg weight obtained from individual hens for each week of lay for each line over the short-term period (90 days from first lay) was used in the data analysis.All weights were obtained using an electronic weighing balance (Mettler P1020N) having a sensitivity of 0.01g; Egg number (EN) -this was taken as the total number of eggs laid by individual layer over the short-term egg production period.

Estimation of Variations in Egg Production Traits in Different Generations and Lines
The data on all traits in different generations and lines were subjected to a one-way analysis of variance using the General Linear Model of the SPSS ( 2001), significant means were separated using Duncan's test at 5% significant level.The statistical model was: Where: X ijk = the record of the k th individual of the i th generation in the j th line; μ = the common mean for the trait being considered; g i = the effect of the i th generation (i=G 0 -G 2 ); l j = the effect of the j th line (j= selected and control); and e ijk = random error.

Estimation of Genetic Parameters
Variance and covariance components of genetic parameters (heritability, genetic and phenotypic correlations) for the selection criterion traits (BWFE, EW, and EN) were estimated using paternal half-sib model as stipulated by Becker (1984).The analysis done using the mixed model least squares and maximum likelihood computer program PC-1 (Harvey, 1990).The statistical model was as follows: Where: Y ij = the record of the j th progeny of the i th sire; μ = the overall population mean for the trait being considered; S i = random effect of the i th sire; and Є ij = uncontrolled environmental and genetic deviations attributable to the individuals within sire groups.

Selection Index
Selection was based on an index which combined data on individual BWFE, EN and EW.The direction of selection was basically positive for all traits.The selection index defined as: I = b 1 X 1 + b 2 X 2 + b 3 X 3 , was calculated for each generation of selection using the matrix notation Pb =Ga according to Becker (1984).
Where: I = total index score; b 1 , b 2 and b 3 = standard partial regression coefficients of the traits in the index; X 1 , X 2 , and X 3 = phenotypic values of the traits (BWFE, EN and EW respectively); P = phenotypic variance-covariance matrix; b = vector of partial regression coefficients (weights); G= genetic variance -covariance matrix; and a = vector of relative economic values.

Effect of Selection on the Various Egg Production Traits in Different Lines and Generations
The means of the egg production traits studied AFE, WFE, BWFE, EW, and EN for the selected and control lines in different generations of selection are presented in Table 2.There were significant differences (P<0.05) in AFE between the generations in the selected line.Such differences were not observed in the various generations of the control line (P>0.05).Birds in the selected line laid eggs later than the control line.Tule (2005) worked with a random-bred population of LE and reported an average AFE of 156.5 ± 0.70 days and 155.85 ± 0.62 days for LE hens raised in the deep litter and battery cage respectively.This is not so different from the result obtained in the control line.Invariably, the discrepancy between the AFE for birds in the selected line and those in the control line is most probably as a result of the application of selection in the former populations which affected their performance.Birds in the control line had the least WFE whereas the selected line had the highest (P<0.05).Omeje and Nwosu (1983) reported a mean WFE of 25.98g for the Nigerian local chicken.Tule (2005) worked with the grandparents of the foundation population of this study and reported a mean WFE of 25.70g.The findings of this study show that the mean WFE obtained was at least 15.53% greater than these reports.Presumably, this difference could be traced to effect of selection on this trait over the generations.BWFE differed significantly (P<0.05) between the two lines and between G 0 and the subsequent generations (G 1 and G 2 ) of the selected line.Such differences did not exist (P>0.05) in the control line.The mean BWFE of the LE over the three generations for the control line were generally within the range reported for unselected local chicken in southern Nigeria (Okpeku et al., 2003;Ogbu and Omeje, 2011) and by Ndofor-Foleng et al. (2010) for the LE.The increase in the BWFE, AFE, WFE and EW in subsequent generations of selection confirms the report of Barbato (1999) that body weight has been shown to be highly responsive to selection such that genetic improvement for growth has resulted in increase in the egg weight and age at sexual maturity.
The average short-term egg weight for the two lines was 36.82g with a range of 35.50 -37.74g.Least EW was observed in the control line while the highest was obtained from the selected line (P<0.05).The EW in the last generation of selection was quite close to that obtained by Momoh et al. (2010) for the Nigerian heavy ecotype.An average of 38.56 eggs was obtained as the short-term egg number for the LE in the present study.The total number of eggs laid by the selected line increased significantly (P<0.05)within three generations of selection.The results of the present study with regards to total egg number indicates that egg number of the local chicken after three generations of selection was an average of 47.13 eggs for the selected population for the short term (90 days from first lay) production period investigated.This is similar to the report of Abdou and Kolstad (1984) for Fayoumi (44 eggs) and white Baladi (41 eggs).Gowe (1970) in his work on long-term selection for high egg production in 2 strains of Leghorn reported that selection for part-period hen-housed egg production was effective in increasing the performance of the selected strains.Birds from the control line laid on average less number of eggs (35.61 eggs) than those from the selected line (41.51eggs).Furthermore,

Estimates of Genetic Parameters of the Selection Criterion Traits
The heritability estimates using paternal half-sib variance analysis for the selection criterion traits by line and generation are presented in Table 3.The estimates for BWFE in two lines and generations varied from moderate to high.The selected line had relatively higher heritability estimates (0.41 to 0.56) for all generations studied, though the estimate decreased with each generation of study.The h 2 for the control line ranged from 0.33 in G 1 to 0.44 in G 0 .Heritability estimates for EW were moderate to high ranging from 0.26 in the G 1 of the control line to 0.44 in the G 0 of the selected population.The heritability estimates of EW in the selected line showed a similar trend as that of the BWFE (decreasing with subsequent generations).Conversely, the h 2 for the control line increased with subsequent generations ranging from 0.25 in the G 0 to 0.32 in G 2 .As shown in Table 3, estimated heritability for EN was moderate for all generations of study irrespective of the line.Highest heritability estimates 0.28 and 0.26 were obtained in the selected line in G 0 and G 1 respectively.The estimate obtained in G 2 (0.20) of the selected line was the same with that of control line.
The heritability estimates especially those of the selected line are on average close to the findings of Lwelamira et al. (2009) who worked on two Tanzanian chicken ecotypes.They reported heritability estimate of 0.45±0.09and 0.43±0.07for body weight in Kuchi chicken and Medium ecotypes respectively.Momoh and Nwosu (2008) worked with the Nigerian heavy ecotype and reported that values of heritability estimates of body weight of the heavy ecotype increased from 0.18 at 4 weeks to 0.43 at 8 weeks and thereafter declined to 0.16 at the 16 th week to rise again to 0.30 at the 20 th week.EN showed a moderate heritability in both lines and generations of study.The estimates obtained affirms those of Dana et al. (2011), who reported heritabilities of monthly egg numbers ranging from 0.20 to 0.56 in Horro chicken of Ethiopia.Estimates of heritability for egg number for both lines and generations in this study falls within this range.The heritability estimates for egg weight obtained for the different lines and over three generations in the present study with LE using sire component ranged from 0.21 to 0.56 indicating a moderate to high additive genetic variance for egg weight in the LE.
Genetic and phenotypic correlation between the selection criterion traits are presented in Table 4.The pattern of variations in the genetic and phenotypic correlations between the various traits studied in the selected and control lines were not consistent as those observed in the case of the heritability estimates.This is in line with Fairfull and Gowe (1990) who attributed such variation partly due to the fact that genetic correlations are generally estimated with less precision than heritabilities.However, the correlation estimates appeared to be more with other reports in the control line.This was especially true for the genetic correlations involving BWFE and EW,BWFE and EN,and EW and EN in G 1 and G 2 .The results of this study affirm those of several reports stating negative correlations (phenotypic and/or genetic) between BWFE and EN, and between EW and EN or egg production (Jeyarubau et al., 1996;Francesch et al., 1997).The genetic correlation between EW and EN in the present study, however, was positive in the G 2 of the selected population.This change in magnitude could be attributed to the selection method applied.Here, selection was based on an index score where, the selection criterion traits, BWFE, EW and EN were all selected for in a positive direction.In other words, only hens which ranked above or were equal to the index score were selected as parents of the next generation.Such selection tends to have increased the gene frequency of the favoured genes, which in the course of recombination were probably transmitted together as linked genes and translated to the maximum performance observed in the selected line.This should be of much interest to the breeder of LE, for perhaps with continuous selection and breeding, the LE could become the White Leghorn of Nigeria.

Conclusion
It is evident that the simultaneous inclusion of BWFE, EW, and EN in the selection index generally improved the performance of the selected line over the three generations of selection in the LE.This virtually suggests that selection based on an index should be applied in breeding programmes for the development and/or improvement of egg production traits in the LE.Furthermore, the heritability estimates and genetic/phenotypic correlations of the selection criterion traits in this study were moderate to high; this is an encouraging factor for intense selection within the local chicken population before being crossbred with improved stocks in order to achieve new breed(s).-.44 a BW= Body Weight at First Egg; EW = Average Egg Weight; EN = Egg Number r g = Genetic correlation; r p = Phenotypic correlation

1
LE -Light Ecotype 2 G 0 = Base Population; G 1 = Generation One; G 2 = Generation Two 3 AFE-Age at First Egg; BWFE-Body Weight at First Egg; WFE-Weight of First Egg, EW-Egg Weight; EN-Egg Number

Table 1 .
Population Size, Effective Population Size and Change in Inbreeding Coefficient over the Three Generations of Selection

Table 2 .
Least-square means (± SE) of the egg traits 1 Means in the same row with different superscripts are significantly different (p<0.05).yz Means in the same column with different superscripts are significantly different (p<0.05).

Table 3 .
Heritability estimates of the selection criterion traits in the selected and control lines 1 Heritability estimates for each trait in the control line are shown in the parenthesis Table4.Genetic and phenotypic correlation of different generations in the selected and control lines 1