ORIGINAL ARTICLE Genetic Resources (2021), 2 (4), 72–84 DOI: 10.46265/genresj.NDFM2712 https://www.genresj.org ISSN: 2708-3764 Morphometric and morphological characterization of chicken resources adapted to pastoral and agropastoral areas of southern Ethiopia Amine Mustefa *,a, Hizkel Kenfo b, Teklewold Belayhun a, Abebe Hailu a and Abraham Assefa a a Ethiopian Biodiversity Institute, Addis Ababa, Ethiopia b Ethiopian Biodiversity Institute, Hawassa Biodiversity Center, Hawassa, Ethiopia Abstract: Thirteen qualitative and six quantitative variables taken from 303 adult chickens (95 cocks and 208 hens) from three locations/districts were used to phenotypically characterize the indigenous chicken populations in pastoral areas of South Omo Zone, Ethiopia. The studied traits were influenced by the effect of location and sex, where chicken populations from Hamer district and females of all districts were the smallest and lightest. Qualitative characteristics of the studied chicken populations such as normal feather morphology and distribution, plain plumage pattern, flat head shape, triangular body shape, and dominant red eye, earlobe and plumage colour suggest that they constitute previously undescribed populations. Chest circumference, wingspan and body length were the three most important morphometric traits used in discriminating the studied chicken populations. On average, 61% of the sampled populations were classified correctly into their respective locations. The multivariate analysis results discriminate the chicken populations into two groups: the Hamer group and the Omo group (chickens from Bena Tsemay and Male districts). However, such grouping should be confirmed and advanced to ecotype level using further genetic characterization studies as the observed phenotypic differences might be due to genetic or environmental variations. Such confirmation is important to design breeding programmes (for sustainable utilization) specific to each ecotype. Keywords: Characterization, Chicken, Discrimination, Ethiopia, Indigenous, Multivariate Citation: Mustefa, A., Kenfo, H., Belayhun, T., Hailu, A., Assefa, A. (2021). Morphometric and morphological characterization of chicken resources adapted to pastoral and agropastoral areas of southern Ethiopia. Genetic Resources 2 (4), 72–84. doi: 10.46265/genresj.NDFM2712. © Copyright 2021 the Authors. This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Introduction Ethiopia, which is believed to be the main gate for most of the indigenous animal genetic resources from Asia to Africa, is one of the countries that possess a large number of livestock populations across the globe (EBI, 2016). Chicken are the most widely distributed livestock species in Ethiopia (EBI, 2016) and worldwide (FAO, 2012). The estimated chicken population size of the country in 2018/2019 was 59.42 million (6 million cocks) among which 85.68% were indigenous (Central Statistical Agency, 2019). ∗Corresponding author: Amine Mustefa (aminemustefa32@gmail.com) Poultry production in Ethiopia is one of the key livestock subsectors which plays an important role in terms of creating employment, subsidizing women, and improving the nutritional content of food. Due to the relatively low investment costs and the small quantity of land required for starting and running poultry produc- tion, it has become a suitable business for low-income farmers (FAO, 2019). However, this subsector is not con- tributing with its full capacity due to poor productivity of and less attention given to indigenous chickens. There- fore, many exotic chicken breeds were disseminated in the country with the objective of increasing production and productivity, becoming the major threat to diversity due to the dilution of indigenous genetic resources and yet providing minimum effect on production and pro- ductivity. Received: 04.05.2021 Accepted: 11.11.2021 Published online: 20.12.2021 https://www.genresj.org https://www.dx.doi.org/10.46265/genresj.NDFM2712 https://www.genresj.org https://www.dx.doi.org/10.46265/genresj.NDFM2712 mailto:aminemustefa32@gmail.com Genetic Resources (2021), 2 (4), 72–84 Ethiopian chicken genetic resources characterization 73 Genetic improvement and conservation as well as sus- tainable utilization of the resources can also be achieved through the application of within-breed selection-based breeding programmes (McDowell, 1972). Identifica- tion, characterization and documentation of the rich Ethiopian chicken genetic resources are a prerequisite for genetic improvement, breed conservation and sus- tainable utilization through designing suitable breeding programmes (FAO, 2012). Outputs of such studies will provide insight into variation within and between popu- lations, which is one of the key contributions to conser- vation and genetic improvement programmes; it is more likely to maintain diversity and bring genetic improve- ment in a population with high variability. Accord- ing to EBI (2016), Ethiopia possesses seven indigenous chicken ecotypes: the Farta, Horro, Jarso, Konso, Man- dura, Tepi and Tillili. The small number of indigenous chicken ecotypes reported in the country can be attributed to the lack of comprehensive identification and characterization work on the diversity and potential of Ethiopian chicken populations. In addition, limitations in methodology and a lack of concrete conclusions in terms of population differentiation were observed in most of the earlier characterization studies. For example, the region-wide chicken characterization study by Melesse and Negesse (2011) did not measure the most important linear measurements (wingspan, body length, chest and shank circumference) and failed to provide concrete conclusions. Measuring, recording and reporting values without strong comparisons, discriminations and conclusions cannot be taken as an input in diversity studies of any given animal breed or ecotype. Therefore, there is an urgent need for continued characterization of indigenous chicken genetic resources to understand their potential and reveal the relationships within and among populations. A short pilot survey performed by a team of Ethiopian Biodiversity Institute (EBI) researchers and South Omo Zone and district livestock experts hinted at the presence of unique chicken populations, in South Omo Zone, Ethiopia, that was not described before. Thus, the present study was aimed at characterizing the chicken populations found in the area based on the FAO (2012) guidelines for characterization of animal genetic resources. Materials and methods Study areas This study was conducted in South Omo Zone, which is found in the Southern Nations Nationalities and Peoples Regional State (SNNPR) and situated in the southern part of Ethiopia (Figure 1). The capital of South Omo Zone, Jinka, is 750 km south of the capital city of the country (Addis Ababa). The zone has a total area of 2,241,731.598 ha. The majority of the land is fertile and cultivated, with trees and bushes covering the remaining area. The zone falls in the arid and semi-arid category of the agroecological zone. Chickens are the most important animals for South Omo Zone pastoralists and agropastoralists next to ruminant animals (Tesfahun et al, 2017). This study was conducted in three randomly selected pastoral and agropastoral districts (Bena Tsemay, Hamer and Male) of the eight districts found in the zone. Moreover, nine sampling sites (kebeles), three from each sampled district were selected based on the chicken population size data (Table 1). Two to three chicken were randomly selected for measurement from each randomly selected household within each kebele. Study ecotypes In literature, chicken populations of the current study area were generally classified as Konso chicken ecotypes (Dana et al, 2010). However, during the classification, samples were not taken from areas near the current study areas. Moreover, the results of our short pilot survey in South Omo Zone hinted at the presence of unique local chicken populations, which are different from Konso and were not described before. Furthermore, the studied chicken genetic resources are indigenous/local, producing, reproducing and surviving under the local environment since many years. Data collection The FAO guidelines for animal genetic resource char- acterization (FAO, 2012) were used to set the over- all data collection procedures including the male to female ratio within the sampled chickens. The sampled chickens were randomly selected from flocks of repre- sentative households. Eleven qualitative traits (feather morphology and distribution, head shape, comb type and size, body shape, and colours of plumage body, earlobe, skin, shank, and eye), and six quantitative measurements (wingspan, body length, chest circumfer- ence, shank length, shank circumference, and live body weight) (Table 2) were recorded from 303 adult chicken (95 cocks and 208 hens) under the effect of district and sex. During the measurements, animals were carefully handled by trained laborers and made to stand properly with parallel legs. The measurements were carried out by two researchers – one measuring and one recording data. At the same time, two other researchers handled the qualitative data recording. To minimize subjective error, all the measurements were taken by the same researcher throughout the study. Quantitative measurements were taken early in the morning before the animals were fed and watered using textile tape measures. Body weight was taken using sensitive digital weighing balances to the nearest of 0.05 g. Data analysis Data entry and management were performed using Microsoft Excel© worksheets. Analysis of the quantita- tive measurements was carried out separately for both sexes and sex aggregated by fitting district as class vari- able. The UNIVARIATE procedure of Statistical Analy- 74 Mustefa et al Genetic Resources (2021), 2 (4), 72–84 Figure 1. Map of the studied areas. The three study districts Bena Tsemay, Hamer and Male are located in South Omo, a zone within the Southern Nations Nationalities and Peoples Regional State (SNNP) in the southern part of Ethiopia. sis Software (SAS) 9.0 (SAS Institute, 2002) was used to detect outliers and test the normality of the quan- titative measurements data. Data on quantitative mea- surements and qualitative characteristics were analyzed using the General Linear Model (GLM) and the fre- quency (FREQ) procedures of SAS 9.0 software respec- tively. Least Square Means (LSM) of the linear measure- ments were separated using the adjusted Tukey-Kramer test (SAS Institute, 2002). Quantitative and qualitative data were analyzed using the following model: Yi = µ+Ai+ei where Yi is an observation, µ is the overall mean, Ai is the fixed effect of district and ei is the ran- dom error attributed to the nth observation. Forward selection procedure of the stepwise discrim- inant function analysis (STEPDISC) was used to find out the quantitative variables that better discriminate populations from different districts. The canonical dis- criminant function analysis (CANDISC) was also used to find out linear combination of quantitative variables that provide maximal separations between the districts. Pair- wise squared Mahalanobis distances between locations were computed as:D2 (i|j) = (xi − xj) ′ cov−1 (xi − xj). Where D2 (i|j) is the distances between locations i and j, cov−1 is the inverse of the covariance matrix of mea- sured variables, xi and xj are the means of variables in the ith and jth populations. The scored canonical variables were used to plot pairs of canonical variables Table 1. Land use, climatic factors and chicken population size of the three studied districts. Data from Hidosa et al (2020); Hidosa and Tesfaye (2018); Gezahegn and Bamud (2018); Derib (2017). Variables Bena Tsemay Hamer Male Land use Agropastoralism Pastoralism Agropastoralism Altitude (m) 500 – 1800 450 – 1765 600 – 1500 Temperature (oC) 17.3 – 28.9 29 – 38 18 – 35 Annual rainfall (mm) 1,167 400 800 – 1200 Chicken population size 94,056 54,288 226,904 Genetic Resources (2021), 2 (4), 72–84 Ethiopian chicken genetic resources characterization 75 Table 2. Six quantitative measurements definition and description of measuring devices used to characterize chicken populations. Adapted from FAO (2012). No. Linear traits Definition of quantitative morphological variables Unit Measuring device 1 Wingspan Length between tips of right and left wings after both are stretched out in full cm Measuring tape 2 Body length Length between the tip of the rostrum maxillare (beak) and that of the cauda (tail, without feathers); the bird’s body should be completely drawn throughout its length cm Measuring tape 3 Chest circumference Taken at the tip of the pectus (hind breast) cm Measuring tape 4 Shank length Length of the shank from the hock joint to the spur of either leg cm Measuring tape 5 Shank circumference Taken at the middle of the shank of either leg cm Measuring tape 6 Body weight Live body weight g Digital balance to get visual interpretation of district differences. Per- centage assignment of observations to known geograph- ical locations (districts) and probabilities of misclassifi- cations were evaluated by discriminant function analysis (DISCRIM). Results Quantitative measurements Level of significance (P values) outputs for the effect of district and sex on the quantitative measurements analyzed for the chicken populations as a whole and separately for each sex are presented in Table 3. Results of the overall analysis show a significant effect of sex on the studied traits with dominance of cocks on all measured traits. Similarly, all studied traits were affected significantly by district in the sex-aggregated analysis. However, effect of district on wingspan and shank circumference was variable due to individuals’ sex, where the cocks’ wingspan and the hens’ shank circumference were not significantly different across districts. Least square means with the respective standard errors (LSM±SE) for the effect of district on the quantitative measurements of the chicken populations as a whole and separately for each sex are presented in Table 4. Cocks from Hamer district have the lowest values except for wingspan, while their counterparts from Bena Tsemay and Male districts were not significantly different in the measured traits. Body length, chest circumference, shank length and body weight of the Hamer hens were also lower than their counterparts from the other districts. Qualitative characteristics The hypothesis whether the qualitative characteristics of the studied chicken populations differ across districts and sexes was tested using chi-squared analysis. The results presented in Table 5 showed a significant effect of district and sex on most of the qualitative characteristics. The colour-related traits except earlobe colour of the cocks were significantly affected by district. The percentages of qualitative characteristics in each district are presented in Table 6 and Table 7. Accordingly, silky feather morphology was observed on Bena Tsemay cocks. The majority of Bena Tsemay cocks possess single comb while one-fourth of the Male and Hamer cocks have a double comb type. The rectangular body shape was observed on almost half of the Bena Tsemay cocks, while the triangular body shape was dominant on cocks from other districts. All Hamer hens possess triangular body shapes while one-fourth of the Bena Tsemay hens had rectangular body shapes. The majority of the studied chicken populations have been characterized by normal feather morphology and distribution, plain plumage pattern, flat head shape, single comb type, and triangular body shape. The comb size of the studied chicken populations was sex-dependent; the majority of hens had small combs while cocks had medium-sized combs. The spur was also absent in the majority of hens from all districts. Red plumage, earlobe and eye colour combined with yellow skin and shank colour, were dominantly observed in cocks from all districts (Table 7). On the other hand, red earlobe and eye colour with yellow shank colour were observed in the majority of hens in all districts. The majority of chickens in the Bena Tsemay and Male districts had yellow shank colour, while chicken populations from Hamer district had a large proportion with grey shank colour. Similarly, the skin colour of most of the chicken population from this study were yellow and white, with grey skin colour observed only in Hamer hens (18.6%). The plumage body colour of Male and Hamer hens was dominated by grey colour while brown, red and black were observed in the majority of Bena Tsemay hens. Multivariate analysis for discrimination of chicken populations Stepwise discriminant analysis Six quantitative measurements for both sexes were separately subjected to the STEPDISC procedure of SAS 9.0. Accordingly, all measurements in cocks and five measurements in hens were identified as the best discriminatory variables. These results were confirmed by Wilk’s lambda test (Table 8) where all selected variables had highly significant (P<0.0001) contribution 76 Mustefa et al Genetic Resources (2021), 2 (4), 72–84 Table 3. Level of significance for the overallanalysis and separately for each sex. WS = Wingspan, BL = Body length, CC = Chest circumference, SL = Shank length, SC = Shank circumference, BW = Body weight, *p<0.05, **p<0.01, ***p<0.0001, NS = Not Significant, CV = Coefficient of variation. Traits Overall chickens Cocks Hens District Sex CV District CV District CV WS ** *** 8.35 NS 7.63 * 8.76 BL *** *** 6.88 *** 6.49 *** 7.08 CC *** *** 9.25 *** 9.85 *** 8.79 SL *** *** 10.16 ** 8.45 ** 11.05 SC ** *** 10.56 ** 12.19 NS 9.27 BW *** *** 21.60 ** 22.93 *** 20.34 Table 4. Pairwise mean comparison (least square means and standard errors (LSM±SE)) for the effect of district within each sex. Means within a column bearing different superscripts are significantly different; a is given to the highest value. N = number of observations, WS = Wingspan, BL = Body length, CC = Chest circumference, SL = Shank length, SC = Shank circumference, BW = Body weight. Traits District Bena Tsemay Male Hamer Cocks N 37 30 28 WS (cm) 41.14±0.52 41.13±0.58 42.86±0.60 BL (cm) 42.30±0.44a 42.00±0.49a 39.00±0.51b CC (cm) 30.05±0.48a 31.37±0.53a 27.36±0.55b SL (cm) 9.92±0.14a 10.37±0.15a 9.39±0.16b SC (cm) 4.92±0.10a 5.00±0.11a 4.50±0.11b BW (g) 1695.89±62.92a 1859.63±69.88a 1428.86±72.33b Hens N 95 70 43 WS (cm) 36.06±0.32ab 35.50±0.38b 37.23±0.48a BL (cm) 38.09±0.27a 36.94±0.31b 35.02±0.40c CC (cm) 27.91±0.25a 27.67±0.29a 25.79±0.37b SL (cm) 8.28±0.09a 8.26±0.10a 7.67±0.14b SC (cm) 4.17±0.04 4.20±0.05 4.04±0.06 BW (g) 1373.87±27.78a 1400.16±32.36a 1124.23±41.29b Both sexes N 132 100 71 WS (cm) 38.67±0.29b 38.27±0.33b 40.02±0.38a BL (cm) 40.24±0.24a 39.34±0.27b 37.06±0.32c CC (cm) 29.06±0.24a 29.28±0.27a 26.67±0.31b SL (cm) 9.14±0.08a 9.25±0.09a 8.54±0.11b SC (cm) 4.53±0.04a 4.58±0.05a 4.30±0.06b BW (g) 1543.70±28.31a 1610.42±31.98a 1282.62±37.06b in discriminating the chicken populations into separate groups. The stepwise discriminant function analysis also revealed that chest circumference, wingspan and body length were the three most important morphometric traits used in discriminating the chicken populations from different districts. However, body length, wingspan and body weight were the three most important variables used to discriminate the hens from different districts. Shank circumference was found to be less useful in discriminating the overall chicken populations due to its lowest discriminatory power (Table 8). Discriminant analysis The multivariate statistics and F approximations used in discriminating the studied chicken populations (cocks, hens and both sexes) are presented in Table 9. Accordingly, all the statistics used in discriminating the populations were significant. Higher F and Den DF values were calculated as the number of observations increased. Results of a discriminant function analysis (Table 10) show the overall classification of individuals into a known location/district. Accordingly, the correct Genetic Resources (2021), 2 (4), 72–84 Ethiopian chicken genetic resources characterization 77 Table 5. Chi-square values and probabilities for the effect of classes on qualitative characteristics for the overall analysis and separately for both sexes. X2 = chi-square value; P = probabilities; *p<0.05, **p<0.01, ***p<0.0001, NS = Not Significant. Qualitative traits Overall Cocks Hens District Sex District District X2 P X2 P X2 P X2 P Feather morphology 7.93 * 13.40 ** 10.04 ** - - Feather distribution 7.91 NS 3.00 NS 3.20 NS 5.05 NS Plumage pattern 5.25 NS 1.85 NS - - 4.85 NS Head shape 8.65 NS 3.19 NS 2.66 NS 6.79 NS Comb type 25.35 ** 22.53 ** 17.34 ** 12.99 NS Comb size 3.73 NS 135.2 *** 5.43 NS 7.48 NS Body shape 27.72 *** 8.41 * 11.06 * 24.14 *** Spur presence 3.16 NS 118.6 *** 2.51 NS 7.32 * Plumage colour 36.13 *** 50.70 *** 22.18 * 26.09 ** Earlobe colour 17.38 ** 20.24 ** 6.55 NS 16.57 * Skin colour 49.06 *** 17.87 *** 6.71 * 49.90 *** Shank colour 52.07 *** 17.73 ** 27.69 ** 40.17 *** Eye colour 80.30 *** 5.50 NS 30.03 *** 57.02 *** Table 6. Percentages for the qualitative characteristics of the chicken populations from different districts. Qualitative traits Cocks Hens Bena Tsemay Male Hamer Bena Tsemay Male Hamer Feather morphology Normal 83.8 100 100 100 100 100 Silky 16.2 0 0 0 0 0 Feather distribution Normal 94.6 100 100 91.6 94.3 97.7 Naked neck 5.4 0 0 6.3 1.4 2.3 Crest 0 0 0 2.1 4.3 0 Plumage pattern Plain 100 100 100 95.8 100 100 Barred 0 0 0 3.2 0 0 Barring 0 0 0 1.0 0 0 Head shape Flat 100 96.7 92.9 92.6 98.6 95.4 Crust 0 0 0 5.3 1.4 0 Snake 0 3.3 7.1 2.1 0 4.6 Comb type Single 86.5 63.3 53.6 75.8 58.6 62.8 Pea 2.7 0 14.3 8.4 18.6 16.3 Rose 10.8 16.7 10.7 12.6 20.0 18.6 V-shape 0 0 0 3.2 0 0 Double 0 20.0 21.4 0 2.8 2.3 Comb size Small 35.1 20.0 21.4 85.3 97.1 93.0 Medium 37.9 53.3 64.3 13.7 2.9 7.0 Large 27.0 26.7 14.3 1.0 0 0 Body shape Blocky 0 3.3 0 2.1 0 0 Rectangular 45.9 16.7 17.9 26.3 7.1 0 Triangular 54.1 80.0 82.1 71.6 92.9 100 Spur presence Absent 27.0 43.3 42.9 89.5 98.6 97.7 Present 73.0 56.7 57.1 10.5 1.4 2.3 78 Mustefa et al Genetic Resources (2021), 2 (4), 72–84 Table 7. Percentages of some colour-related qualitative traits of the chicken populations from different districts. Qualitative traits Cocks Hens Bena Tsemay Male Hamer Bena Tsemay Male Hamer Plumage colour Black 5.4 0 0 20.0 11.4 9.3 Brown 5.4 0 0 25.3 17.2 9.3 Grey 10.8 30.0 17.9 15.8 30.0 44.2 Red + White 8.1 10.0 25.0 1.0 7.1 11.6 Red 46.0 56.7 53.6 25.3 18.6 16.3 White 24.3 3.3 3.6 12.6 15.7 9.3 Earlobe colour White 5.6 6.7 3.6 15.0 18.6 32.6 Red 88.8 66.6 71.4 77.5 72.8 51.1 Red + White 5.6 26.7 25.0 3.8 4.3 16.3 Black 0 0 0 3.7 4.3 0 Skin colour Grey 0 0 0 0 0 18.6 White 43.2 16.7 21.4 63.2 31.4 48.8 Yellow 56.8 83.3 78.6 36.8 68.6 32.6 Shank colour Yellow 91.9 96.7 50.0 56.8 70.0 37.2 Black 2.7 0 10.7 20.0 12.9 16.3 White 5.4 3.3 17.9 20.0 4.3 9.3 Grey 0 0 21.4 3.2 12.8 37.2 Eye colour Red 62.2 100 92.9 73.7 95.7 72.1 Blue 0 0 7.1 0 0 7.0 Brown 0 0 0 1.0 2.9 18.6 White 5.4 0 0 3.2 0 2.3 Yellow 32.4 0 0 22.1 1.4 0 classification of cocks into their location/district ranged from 51% in Bena Tsemay to 75% in Hamer. The overall average error rate was 39%, and 61% of the individuals were classified correctly. An average of 64% and 58% of the sampled cocks and hens were classified correctly into their corresponding districts, respectively. A higher error rate of 55% was observed in Bena Tsemay hens, while a lower error rate (23%) was obtained from the classification of Hamer district hens. Canonical discriminant analysis Different multivariate statistics (including canonical correlation and eigenvalues), the coefficient values for each trait used, and class mean outputs using the first two canonical structures are shown in Table 11. The first canonical structure (Can 1) explains the majority (85.4%) of the variability among the three districts (84.5% for cocks and 84.3% for hens). Can 1 also produces the greatest multiple correlation (60.2%) with the classes that was achieved by using the linear combination of the quantitative traits; the values were 72.7% for cocks and 55.5% for hens. The results revealed that Can 1 separates the chicken populations (class means) from different districts better than Can 2. The pairwise squared Mahalanobis distances between locations/districts for analysis groups presented in Table 12 were highly significant (P<0.0001). The shortest distance (0.6) was calculated between the chicken populations of Bena Tsemay and Male districts; 1.36 for cocks and 0.46 for hens. On the other hand, the chicken populations from Hamer district were more distantly related to the others (2.87 and 3.71 from Bena Tsemay and Male districts, respectively). Discussion Quantitative measurements The quantitative measurements taken have produced reliable information on characterization, evaluation and differentiation of the studied chicken populations. In line with most reports, these measurements were significantly affected by the location sampled. The longest wings and smallest body size of the Hamer chickens make them significantly different from the chicken populations of the other two locations. These long wings stretched from the small body size can be recorded and reported as the unique characteristics Genetic Resources (2021), 2 (4), 72–84 Ethiopian chicken genetic resources characterization 79 Table 8. Summary of the stepwise discriminant function analysis; ascending order of traits used in discriminating the chicken populations from different districts. Group Step Variable entered Partial R-Square F value P>F Wilks’ Lambda PF Cocks Wilks’ Lambda 0.3912 8.68 12 174 <0.0001 Pillai’s Trace 0.6986 7.87 12 176 <0.0001 Hotelling-Lawley Trace 1.3265 9.54 12 132.28 <0.0001 Roy’s Greatest Root 1.1217 16.45 6 88 <0.0001 Hens Wilks’ Lambda 0.6340 8.53 12 400 <0.0001 Pillai’s Trace 0.3916 8.16 12 402 <0.0001 Hotelling-Lawley Trace 0.5367 8.91 12 308.04 <0.0001 Roy’s Greatest Root 0.4459 14.94 6 201 <0.0001 Both sexes Wilks’ Lambda 0.5778 15.51 12 590 <0.0001 Pillai’s Trace 0.4560 14.57 12 592 <0.0001 Hotelling-Lawley Trace 0.6722 16.48 12 455.81 <0.0001 Roy’s Greatest Root 0.5696 28.10 6 296 <0.0001 of Hamer chickens, which might be related to their mothering ability. A previous study by Dana et al (2010) which describes Konso as a chicken ecotype took only two quantitative measurements (the body weight and shank length). In terms of body weight both cocks and hens of the current study were heavier than the Konso chicken ecotype where the body weights of the Konso cocks and hens were 1,411 g and 1,011 g respectively (Dana et al, 2010). Such phenotypic variations in body weight hint at the presence of genetic diversity that needs to be conserved and can also be used as a base in attaining genetic improvement through selection. Therefore, in terms of body weight, the current chicken populations were different from the Konso chicken ecotype. On the other hand, their shank length measurements were comparable. According to Melesse and Negesse (2011), the shank length is considered a good indica- tor of adaptation to lowland areas and skeletal develop- ment, which is related to the amount of meat a chicken 80 Mustefa et al Genetic Resources (2021), 2 (4), 72–84 Table 10. Number and (percent) of observations classified into districts. Group From district Bena Tsemay Hamer Male Total Cocks Bena Tsemay 19 (51%) 7 (19%) 11 (30%) 37 (100%) Hamer 3 (11%) 21 (75%) 4 (14%) 28 (100%) Male 8 (27%) 2 (7%) 20 (66%) 30 (100%) Total 30 (32%) 30 (32%) 35 (36%) 95 (100%) Error rate 49% 25% 34% 36% Hens Bena Tsemay 43 (45%) 18 (19%) 34 (36%) 95 (100%) Hamer 5 (11%) 33 (77%) 5 (11%) 43 (100%) Male 24 (34%) 9 (13%) 37 (53%) 70 (100%) Total 72 (35%) 60 (29%) 76 (36%) 208 (100%) Error rate 55% 23% 47% 42% Both sexes Bena Tsemay 65 (49%) 22 (17%) 45 (34%) 132 (100%) Hamer 9 (13%) 52 (73%) 10 (14%) 71 (100%) Male 31 (31%) 8 (8%) 61 (61%) 100 (100%) Total 105 (35%) 82 (27%) 116 (38%) 303 (100%) Error rate 51% 27% 39% 39% Table 11. Multivariate statistics, canonical coefficients of the quantitative variables, and class means outputs of the two canonical structures separately for each sex. Can = Canonical structure. Cocks Hens Both sexes Can 1 Can 2 Can 1 Can 2 Can 1 Can 2 Multivariate statistics Canonical correlation 0.7271 0.4122 0.5553 0.2885 0.6024 0.3051 Eigenvalue 1.1217 0.2047 0.4459 0.0908 0.5696 0.1026 Proportion 0.8457 0.1543 0.8309 0.1691 0.8473 0.1527 Cumulative 0.8457 1.0000 0.8309 1.0000 0.8473 1.0000 F value 8.68 3.60 8.53 3.65 15.51 6.08 P>F <0.0001 0.0052 <0.0001 0.0035 <0.0001 <0.0001 Traits Wingspan -0.3191 0.0338 -0.2780 0.1110 -0.3280 0.0350 Body weight 0.0004 -0.0029 0.0019 -0.0031 0.0015 -0.0028 Body length 0.1247 0.5403 0.0915 0.4719 0.0560 0.5070 Chest circumference 0.1727 -0.0725 0.1483 0.0356 0.1828 0.0089 Shank length 0.5074 -0.8638 0.3943 -0.5278 0.3388 -0.7575 Shank circumference 0.4542 0.8747 -0.1134 0.0512 -0.0147 0.1737 Class (district) mean Bena Tsemay 0.4234 0.5273 0.3164 0.2933 0.2980 0.3400 Male 0.9488 -0.5150 0.3679 -0.3858 0.5594 -0.3871 Hamer -1.5761 -0.1450 -1.298 -0.0200 -1.3419 -0.0869 can carry. This shows their high adaptability to the low- land areas of the pastoral and agropastoral community. Comparable, higher and lower values of wingspan, body length, chest circumference, shank length and shank circumference measurements were reported in indigenous chickens in different parts of the country. However, surprisingly low values were also reported. For example, Negassa et al (2014) reported 7.35–8.17 cm of wingspan and 22.6–24.2 cm of body length for chickens in southeastern Ethiopia. Similarly, Halima et al (2007) reported 12.67–15.83 cm of wingspan and 0.53–0.93 cm of shank circumference for chickens in northern Ethiopia. Effect of sex Cocks were bigger and heavier than hens, which follows Rensch’s rule (Rensch, 1950) where the males of an individual species are generally larger than the females. Such differences between cocks and hens may be attributed to the differences in hormone secretion, Genetic Resources (2021), 2 (4), 72–84 Ethiopian chicken genetic resources characterization 81 Table 12. Squared Mahalanobis distance between districts; output of the multivariate analysis calculated using the quantitative measurements. *** indicates significance of the distance calculations at p<0.0001. Group From district Bena Tsemay Male Hamer Cocks Bena Tsemay 0 Male 1.36** 0 Hamer 4.45*** 6.51*** 0 Hens Bena Tsemay 0 Male 0.46** 0 Hamer 2.70*** 2.91*** 0 Both sexes Bena Tsemay 0 Male 0.60*** 0 Hamer 2.87*** 3.71*** 0 which leads to enlargement of muscle mass and skeletal development (Baneh and Hafezian, 2009). These results were in line with the reports of Bekele et al (2015), Getu et al (2014), Negassa et al (2014), and Melesse and Negesse (2011), Dana et al (2010) and Halima et al (2007) on different indigenous chicken populations of Ethiopia. Similarly, sexual dimorphism was reported for most traits with males having higher values when compared to the females in other species of Ethiopia, including goats (Mustefa et al, 2019), cattle (Mustefa et al (2020b)), donkeys (Mustefa et al (2020a)) and sheep (Hailu et al, 2020). Qualitative characteristics Observable qualitative characteristics, in addition to the quantitative measurements, have allowed us to characterize, identify and differentiate the studied chicken populations. Dana et al (2010), Bekele et al (2015), Melesse and Negesse (2011) reported normal feather morphology and distribution in the majority of indigenous chicken populations in Ethiopia’s southern region that are comparable with the results of the current study. Melesse (2000) described the naked- neck gene as one of the main genes responsible for heat tolerance of some Ethiopian indigenous chicken populations by improving and enhancing heat dissipation due to the reduction in feather coverage. However, the current study also revealed the adaptation of indigenous chickens to the local pastoral and agropastoral lowland areas with a lower frequency of naked-neck chickens. This lower frequency of naked- neck chicken in the current study area specifically, as well as in the country in general, may also be a factor for their poor productivity (Ajang et al, 1993). Similarly, Yunis and Cahaner (1999) reported the probability of attaining higher egg and/or meat production in chicken with reduced feathering due to the saved feather protein. Qualitative deviations were also observed between Konso ecotype and the current chicken populations. Firstly, the comb types, which vary greatly depending on the breed, are also a great indicator of a chicken’s health. Bright fleshy combs indicate physical strength and good health. Combs also help chickens regulate their body temperature – those with large combs are able to circulate blood faster through their combs, which helps to release body heat. This is why chickens adapted to warm environments have larger combs (like single combs and buttercup combs) while breeds from colder areas have smaller combs (like pea and rose combs) (Vaughn, 2019). The majority of chickens from the current study area were single-combed while Dana et al (2010) reported a higher frequency of pea-combed chickens in the Konso chicken ecotype. Dana et al (2010) also reported humid lowland to wet highland ecological zones as local areas of the Konso chicken ecotype. The comb size of the studied chicken populations was found to be sex-dependent where hens of a given population had smaller combs than the cocks. This was supported by Bell (2002) who reported the association of comb size with gonadal development and intensity of light. The earlobe colour of a chicken can determine the colour of the egg that it will lay. For example, if the chicken has a red earlobe, it will lay a brown-shelled egg and if it has a white earlobe, it will lay a white-shelled egg (Bell, 2002). Red earlobe colour was observed in the majority of chickens from the current study, while Dana et al (2010) reported equal frequency of red and white earlobe colours in Konso chicken ecotype. The observation of a higher proportion of hens with white earlobe colour than the cocks in each district was due to the sex-linked nature of the trait (Luo et al, 2018). Among the populations of the current study, a higher proportion of white earlobes was observed in Hamer hens than the others, which might be due to their distinction from the others as earlobe colour is a breed- specific trait, though it could also be affected by the nutritional status of the chicken populations (Melesse and Negesse, 2011). The triangular body shape was dominantly found in the current study while the results of Dana et al (2010) reported blocky body shape for the majority of the Konso chicken ecotype. On the other hand, similar qualitative characteristics, like flat head shape, were observed on both Konso chicken ecotypes (Dana et al, 2010) and the current chicken populations. Chicken populations, breeds and breeding groups can be identified by their plumage colours due to their decorative qualities. Plumage colour is also a key trait during the interaction of chickens due to their well- developed visual perception of the world (Makarova et al, 2019). In agreement with most Ethiopian reports, several plumage colours were found in the current study area with a higher frequency of red-coloured cocks. On the other hand, the predominance of grey (faded white and black) colour observed in hens is in line with the reports of Halima et al (2007). The diverse 82 Mustefa et al Genetic Resources (2021), 2 (4), 72–84 plumage colours in the current study, which might be due to either genetic or environmental factors, showed the presence of diverse genetic resources, which can be an input in conservation and genetic improvement programmes for sustainable use. Bekele et al (2015), Getu et al (2014), Halima et al (2007), Dana et al (2010) reported yellow as the most widespread shank and skin colour in different indigenous chicken populations of Ethiopia, which agrees with the current findings. According to Hammond and Harshaw (1941), a chicken’s shank and skin colour are influenced by breed, quantity of xanthophyll pigment in the diet, and quantity of a pigmentation- suppressing factor in the diet. Bell (2002) stated that yellow shank colour is due to nutritional carotenoid colourants in the epidermis when the melanic pigment is absent. Similarly, variable shades of black colour are the result of melanic pigment in the dermis and epidermis; if the black colourant is in the dermis and yellow in the epidermis, greenish shanks will appear. However, if both these pigments are completely absent, the shanks will be white (Bell, 2002). Commonly, red, yellow and black eye colours are due to the three pigments in the form of hemoglobin, carotenoids and melanin respectively. Red eye colour is influenced by the degree of iridic vessels injection and hemoglobin content of the blood (Nelson, 1947). According to Nelson (1947), when chickens suffer from general anemia, a smaller amount of red colouration will be noted due to loss of blood volume. Therefore, the dominant red eye colour observed in the current study is a confirmation that these chicken populations were free from such suffering. As per the same author, continual egg production can decrease the yellow factor in eye colour. Therefore, the yellow eye colour chicken from Bena Tsemay district (27%) might indicate restrictions in egg productivity. Multivariate analysis The use of more quantitative measurements during the discrimination analysis approaches the results towards reality. This is supported by the current results, where only one quantitative measurement in hens had low discriminatory power. In line with reports by Mustefa et al (2020b) on Raya cattle (61%) and Hailu et al (2020) on Tigray sheep (66%), the discriminant function analysis allowed the classification of an average 61% of the studied individuals into their respective locations. The observed highest classifications of individuals into their respective districts in Hamer cocks and hens (75% and 77% respectively) showed their distinctness from the others due to the unique characteristics they possess. However, the lowest classifications (45% and 53%) were recorded in hens from Bena Tsemay and Male district indicating strong similarities between them. The highly significant and longest pairwise Maha- lanobis distances between districts indicate the accuracy of the calculations and the distinctness of the popula- tions from each other in a measurable group difference for the considered quantitative measurements. Accord- ingly, the longest distances of Hamer chickens from the other two chicken populations in this study showed its distinctness. In conclusion, the main contribution of this diversity study is the provision of information and data for breed differentiation, conservation and sustainable utilization of the chicken ecotypes in the districts, as well as the collation of information and data available on chicken ecotypes of Ethiopia. Accordingly, the chicken populations can be categorized phenotypically into two groups: the Hamer group and the Omo group (which includes the indigenous chicken populations from Bena Tsemay and Male districts). However, such differences in phenotypic performances might be either due to genetic or environmental variations. Therefore, we cannot conclude that the differences are solely due to genetic variations. Thus, genetic characterization is recommended to understand their potential and the within- and among-population genetic diversity and population structures. Data availability As baseline data for further research and development work, these data will be made available in the country’s focal institute for indigenous animal genetic resources, the Ethiopian Biodiversity Institute (EBI) and the global Domestic Animal Diversity Information System (DAD- IS) databases. Acknowledgments The authors are highly grateful to the Ethiopian Biodiversity Institute (EBI) for funding the work. Our special appreciation also goes to the farmers for providing their animals for this work for free. We also take this opportunity to appreciate the animal science experts at zonal, district and kebele level for their endless help during the data collection. A special word also goes to our friend and work partner Mr. Tadesse Hunduma for mapping the study area. Author contributions All authors contributed to the study conception and design. Material preparation and data collection were performed by Amine Mustefa, Hizkel Kenfo, and Teklewold Belayhun. Data analysis and writing the first draft of the manuscript was performed by Amine Mustefa and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript. Conflict of interest statement The authors declare no conflict of interest. Genetic Resources (2021), 2 (4), 72–84 Ethiopian chicken genetic resources characterization 83 References Ajang, O. A., Prijono, S., and Smith, W. K. (1993). The effect of dietary protein level on growth and body composition of fast and slow feathering broiler chickens. British Poultry Science 34, 73–91. doi: https: //doi.org/10.1080/00071669308417564 Baneh, H. and Hafezian, S. H. (2009). Effect of envi- ronmental factor on growth traits in Ghezel sheep. African Journal of Biotechnology 8, 2903–2907. url: https://www.ajol.info/index.php/ajb/article/view/ 60943. Bekele, G., Kebede, K., Ameha, and N (2015). On-farm Phenotypic Characterization of Indigenous Chicken and their Production System in Bench Maji Zone. Science, Technology and Arts Research Journal 4, 68–73. doi: http://dx.doi.org/10.4314/star.v4i1.10 Bell, D. D. (2002). Anatomy of the Chicken. In Commercial chicken meat and egg production, ed. Bell, D., D., Weaver, and and, W. D., (USA: Springer, Boston, MA), 41-58. Central Statistical Agency (2019). Agricultural Sample Survey 2018/19. Report on Livestock and Livestock Characteristics. In Statistical Bulletin No. 588, volume II, Central Statistical Agency, Addis Ababa, Ethiopia. Dana, N., Dessie, T., Van Der Waaij, H. L., and Van Arendonk, A. M. J. (2010). Morphological features of indigenous chicken populations of Ethiopia. Animal Genetic Resource 46, 11–23. doi: https://doi.org/10. 1017/S2078633610000652 Derib, G. (2017). Production and Marketing Systems of Goat in Hamer Woreda, South Omo Zone, Southern Ethiopia. url: https://doi.org/10.20372/ nadre:1554185414.61. EBI (2016). Ethiopian National Strategy and Plan of Action for conservation and utiliza- tion of Animal Genetic Resources (Addis Ababa, Ethiopia: Ethiopian Biodiversity Institute). url: https://www.ebi.gov.et/wpcontent/uploads/2013/ 01/Final%20ENSAP final submitted%20(1).pdf. FAO (2012). Phenotypic characterization of animal genetic resources. In FAO Animal Production and Health Guidelines No. 11, FAO, Rome, Italy. FAO (2019). Poultry Sector Ethiopia. In FAO Animal Production and Health Livestock Country Reviews n.11, FAO, Rome. Getu, A., Alemayehu, K., and Wuletaw, Z. (2014). Phenotypic characterization of indigenous chicken ecotypes in the north Gondar zone. Ethiopia. Animal Genetic Resources 54, 43–51. doi: https://doi.org/10. 1017/S2078633614000113 Gezahegn, B. and Bamud, K. (2018). Identifi- cation of Major Crop Production Constraints in Bena-Tsemay District, South Omo Zone, Southern Ethiopia. . Food Science and Quality Management 74. url: https://www.iiste.org/Journals/ index.php/FSQM/article/view/41847. Hailu, A., Mustefa, A., Aseged, T., Assefa, A., Sinkie, S., and Tsewene, S. (2020). Phenotypic characterization of sheep populations in Tahtay Maichew district. Northern Ethiopia. Genetic Resources 1(2), 13–23. doi: https://doi.org/10.46265/genresj.SHBD3744. Halima, H., Neser, F. W. C., Van Marle-Koster, E., Dekock, and A (2007). Phenotypic variation of indigenous chicken populations in northwest Ethiopia. Tropical Animal Health and Production 39, 507–513. doi: https://doi.org/10.1007/ s11250-007-9032-2 Hammond, J. C. and Harshaw, H. M. (1941). Some Factors Influencing S hank a nd S kin C olor i n the Growing Chicken. Poultry Science 20, 437–444. doi: https://doi.org/10.3382/ps.0200437 Hidosa, D., Hailu, S., ’reagain, O., and J (2020). Goat Feed Inventory and Feed Balance in Hamer and Bena-Tsemay Woreda of South Omo Zone, South Western Ethiopia. Acta Scientific Veterinary Sciences 2(6), 28–43. url: https://actascientific.com/ASVS/ ASVS-02-0069.php. Hidosa, D. and Tesfaye, Y. (2018). Assessment Study on Livestock Feed Resource, Feed Availability and Production Constraints in Maale Woreda in South Omo Zone. Journal of Fisheries and Livestock Production 6(2). doi: https:// doi.org/10.4172/2332-2608.1000269 Luo, W., Xu, J., Li, Z., Xu, H., Lin, S., Wang, J., Ouyang, H., Nie, Q., and Zhang, X. (2018). Genome- Wide Association Study and Transcriptome Analysis Provide New Insights into the White/Red Earlobe Color Formation in Chicken. Cell Physiology and Biochemistry 46, 1768–1778. doi: https://doi.org/10. 1159/000489361 Makarova, A. V., Mitrofanova, O. V., Vakhrameev, A. B., and Dementeva, N. V. (2019). Molecular-genetic bases of plumage coloring in chicken. Vavilov Journal of Genetics and Breeding 23, 343–354. doi: https://doi. org/10.18699/VJ19.499 McDowell, R. E. (1972). Improvement of Livestock Production In Warm Climates. W.H. Freeman & Co., San Francisco. Melesse, A. (2000). Comparative studies on per- formance and physiological responses of Ethiopian indigenous (Angete-Melata) chickens and their F1- crosses to long-term heat exposure (Martin-Luther University of Halle-Wittenberg, Halle-Saale, Ger- many) . url: https://www.cabdirect.org/cabdirect/ abstract/20036794426. Melesse, A. and Negesse, T. (2011). Phenotypic and morphological characterization of indigenous chicken populations in southern region of Ethiopia. Animal Genetic Resources 49, 19–31. doi: https://doi.org/10. 1017/S2078633611000099 Mustefa, A., Assefa, A., Misganaw, M., Getachew, F., Abegaz, S., Hailu, A., and Emshaw, Y. (2020a). Phe- notypic Characterization of Donkeys in Benishangul Gumuz National Regional State. Online Journal of Animal and Feed Research 10(1), 25–35. doi: https: //dx.doi.org/10.36380/scil.2020.ojafr4 Mustefa, A., Belayhun, T., Melak, A., Hayelom, M., Tadesse, D., Hailu, A., and A, A. (2020b). Phenotypic https://doi.org/10.1080/00071669308417564 https://doi.org/10.1080/00071669308417564 https://www.ajol.info/index.php/ajb/article/view/60943 https://www.ajol.info/index.php/ajb/article/view/60943 http://dx.doi.org/10.4314/star.v4i1.10 https://doi.org/10.1017/S2078633610000652 https://doi.org/10.1017/S2078633610000652 https://doi.org/10.20372/nadre:1554185414.61 https://doi.org/10.20372/nadre:1554185414.61 https://www.ebi.gov.et/wpcontent/uploads/2013/01/Final%20ENSAP_final_submitted%20(1).pdf https://www.ebi.gov.et/wpcontent/uploads/2013/01/Final%20ENSAP_final_submitted%20(1).pdf https://doi.org/10.1017/S2078633614000113 https://doi.org/10.1017/S2078633614000113 https://www.iiste.org/Journals/index.php/FSQM/article/view/41847 https://www.iiste.org/Journals/index.php/FSQM/article/view/41847 https://doi.org/10.46265/genresj.SHBD3744 https://doi.org/10.1007/s11250-007-9032-2 https://doi.org/10.1007/s11250-007-9032-2 https://doi.org/10.3382/ps.0200437 https://actascientific.com/ASVS/ASVS-02-0069.php https://actascientific.com/ASVS/ASVS-02-0069.php https://doi.org/10.4172/2332-2608.1000269 https://doi.org/10.4172/2332-2608.1000269 https://doi.org/10.1159/000489361 https://doi.org/10.1159/000489361 https://doi.org/10.18699/VJ19.499 https://doi.org/10.18699/VJ19.499 https://www.cabdirect.org/cabdirect/abstract/20036794426 https://www.cabdirect.org/cabdirect/abstract/20036794426 https://doi.org/10.1017/S2078633611000099 https://doi.org/10.1017/S2078633611000099 https://dx.doi.org/10.36380/scil.2020.ojafr4 https://dx.doi.org/10.36380/scil.2020.ojafr4 84 Mustefa et al Genetic Resources (2021), 2 (4), 72–84 characterization of Raya cattle in northern Ethiopia. Tropical Animal Health and Production 53, 48–48. doi: https://doi.org/10.1007/s11250-020-02486-1 Mustefa, A., Gizaw, S., Banerjee, S., Abebe, A., Taye, M., Areaya, A., and Besufekad, S. (2019). Growth perfor- mance of Boer goats and their F1 and F2 crosses and backcrosses with Central Highland goats in Ethiopia. Livestock Research for Rural Development 31(89). url: http://www.lrrd.org/lrrd31/6/amine31089.html. Negassa, D., Melesse, A., and Banerjee, S. (2014). Phenotypic characterization of indigenous chicken populations in Southeastern Oromia Regional State of Ethiopia. Animal Genetic Resources 55, 101–113. doi: https://doi.org/10.1017/s2078633614000319 Nelson, N. M. (1947). Normal Eye Color in the Chicken. Poultry Science 26, 61–66. doi: https://doi.org/10. 3382/ps.0260061 Rensch, B. (1950). Die Abhangigkeit der relative sexual differenz von der korpergrosse. Bonner Zoologische Beitrage 1, 58–69. SAS Institute (2002). url: https://www.sas.com/enus/ software/stat.html. Tesfahun, B., Kebede, K., and Effa, K. (2017). Traditional goat husbandry practice under pastoral systems in South Omo zone, southern Ethiopia. Tropical Animal Health and Production 49, 625–632. doi: https://doi. org/10.1007/s11250-017-1240-9 Vaughn, K. (2019). url: https://www. communitychickens.com/chicken-comb-types- zbw1912ztil/. Yunis, R. and Cahaner, A. (1999). The effects of naked- neck (Na) and frizzle (F) genes on growth and meat yield of broilers, and their interactions with ambient temperatures and potential growth rate. Poultry Science 78, 1347–1352. doi: https://doi.org/10.1093/ ps/78.10.1347 https://doi.org/10.1007/s11250-020-02486-1 http://www.lrrd.org/lrrd31/6/amine31089.html https://doi.org/10.1017/s2078633614000319 https://doi.org/10.3382/ps.0260061 https://doi.org/10.3382/ps.0260061 https://www.sas.com/enus/software/stat.html https://www.sas.com/enus/software/stat.html https://doi.org/10.1007/s11250-017-1240-9 https://doi.org/10.1007/s11250-017-1240-9 https://www.communitychickens.com/chicken-comb-types-zbw1912ztil/ https://www.communitychickens.com/chicken-comb-types-zbw1912ztil/ https://www.communitychickens.com/chicken-comb-types-zbw1912ztil/ https://doi.org/10.1093/ps/78.10.1347 https://doi.org/10.1093/ps/78.10.1347 Introduction Materials and methods Study areas Study ecotypes Data collection Data analysis Results Quantitative measurements Qualitative characteristics Multivariate analysis for discrimination of chicken populations Stepwise discriminant analysis Discriminant analysis Canonical discriminant analysis Discussion Quantitative measurements Effect of sex Qualitative characteristics Multivariate analysis Data availability Acknowledgments Author contributions Conflict of interest statement