population in Lominat Forgena had correct classification of 76.67% in their sites and the incorrect classification (23.33%)
was explained in Gilal Wuha location. This revealed that males from these sites had common characteristics. The overall
correct and incorrect classification percentages of male populations in all the study sites were 75% and 25%, respectively.
Horro cattle (82.72%) and (Fasil, 2006) for Gojjam highland cattle (80.85%), respectively. Similarly, the correct
classification percentages of female sampled population for all sites were ranged from Gendawuha Birshign (42.5%) to
Mirt Gelegu (98.33%) sites. The remaining classification percentages of Gendawuha Birshign sites were incorrectly
classified in Dubaba (23.33%), Abrajira (33.33%) sites. Correspondingly, the correct and incorrect classification values of
female population from Lominat Forgena were 47.5% and 37.5%, respectively. This incorrect classification percentage
was explained in Gilal Wuha. This indicated that the two sites had similar phenotypic features. The overall correct
classification hit rates for female sample population were 60%. The overall error count estimation in male sample
population was better than female population (as the average hit rate was 75% than 60%).
Canonical discriminate analysis. It was performed by CANDISC procedures separately for male and female sample
population. As per Table 5, all mahalanobis distances obtained from the sites for male population were highly significant
(p<0.0001). Therefore, male population from each site has their own measurable differences. The nearest distances were
observed in Abrajira and Dubaba sites with the value of 0.76185. This might be due to the two sites had phenotypically
similar male cattle population. Whereas, the longest distances (37.88) was observed between Gendawuha Birshign and
Mirt Gelegu sites. Therefore, these two sites had their own distinct and different male cattle types. In case of female
sample population, the Mahalanobis distance between sites was slightly lower than that of male population. Mahalanobis
distances obtained for female population was significant for all sites. The nearest distance was observed in between
Gendawuha Birshign and Dubaba followed by Abrajira and Gendawuha Birshign with the values of 0.35833 and 0.51030,
respectively. The largest distance was observed between site 3 (Gendawuha Birshign) and site 2 (Mirt Gelegu) with the
value of 29.38. The narrower and wider differences were resulted from phenotypic similarities and disparities between the
sample populations.
Stepwise discriminate analysis. All 9 continuous variables for both sexes were separately subjected to the STEPDISC
procedure of (SAS, 9.4) and all 9 variables for males and 8 variables for females were identified as best discriminating
variables (Table 7). Variables used for discriminating male breed types were ordered based on their discriminating powers
namely cannon bone length, horn length, cannon bone circumferences and body length. Similarly, top variables, which
discriminate female population, were cannon bone length, horn length, mouth circumferences, body length, ear length,
and cannon bone circumferences. Whereas, height at wither was weak for separating female sample population and was
removed.
Correspondence analysis for qualitative variables. Twenty different categorical variables for all the study sites were
considered for correspondence analysis (Figure 1). Two dimensions were observed for the clear relationships of the traits
with study sites. Categorical variable lies on the middle line were predominantly found in all the population and
represented low variation among the population. Small hump size, and lateral ear orientation in (sites 1 and 2) and, erect
horn shape, convex facial profile, medium hump size and medium body condition both in site 3, 5 and site 4, 6 had very
close to the middle line. And, for all population large tail length, medium body condition score, sloppy rump profile, short
and curved horn shape and black white color variables represented very low variations. Whereas, categorical variables far
from the middle line were distinguished the 6 studied populations namely coat color type, navel flap, dewlap width, udder
size and teat length in one hand and the genetic groups were in the other hands.
Cluster analysis. Cluster analysis is a multivariate technique used to group individuals based on common
characteristics in to homogeneous and distinct groups (Peter, 1997). Due to this; the population in one cluster shared
certain common characters more than the population in other cluster. According to cluster analysis 3 similar clusters were
formed for both sexes in figure 2. Hence, the first cluster consisted of sample population of site 2 (Mirt Gelegu) which,
represent Fellata cattle type, came out distinctly far from the rest group breed types in both cases. The second cluster
contained three sites namely; site 1 (Dubaba), site 3 (Gendawuha Birshign) and site 5 (Abrajira). Moreover, the cattle type
from site 4 (Lominat Forgena) and site 6 (Gilal Wuha) grouped under the third clusters.
Cluster 1: Fellata cattle type. This cattle type is unique in many aspects from the rest of the population. They are
aggressive in their temperament and large in body size. Their coat colour pattern is plain (76.0%) and patchy (14.7%) and
spotted (9.3%). The hair types were shiny (69.3%) with dull (26.0%) and smooth (9.3%). Most of the animals have
exceptionally longer horns (50.95 cm in females and 47.06 cm in males) with white in colour and lyre (54%), curved
(26.7%) and straight (19.3%) in its shape. Mostly flat (58%) facial profiles were observed in both sexes and rarely concave
with thin neck in females and regularly convex in males. Ear is medium to big in size with mainly straight and pendulous
shapes. Humps and dewlaps are well developed in males with good height at wither but small humps with erected shapes
and almost non dewlaps in females (their unique features). Body condition scoring of this cattle was medium (78.7%),
lean (16.0%) and fat (5.3%) in both sexes. Tail is medium to long in females and long with bushy ends in males; important
for prevention of wild flies. Cows have longer navel length, large udder and teat size of all the breed types in the study
location. These characteristics were considered for selection criteria in the study location. Similarly, perpetual sheath and
testicular lengths are very long in males. These traits were also highly considered in the communities for selection
criteria’s of breeding bull. These cattle types mainly serve for the community through reproduction, milk production and
242
Citation: Emru E, Abegaz S, Getu A, Mengistu A, Fantahun T (2020). Phenotypic characterization of genetic resources of indigenous cattle in West Gondar zone of
Ethiopia. Online J. Anim. Feed Res., 10(5): 240-248.