Multivariate Analysis of Four Strains of Broiler Chickens at Four Weeks of Age

Abstract

Eight hundred broiler chickens of four different strains (200 each of Arbor Acre, Hubbard, Marshall and Ross 308) were used to investigate the variations between four strains of broiler chickens commonly reared in North central Nigeria  using their  body weight and some morphometric body parameters.  At four weeks of age, body weight and some morphometric body parameters such as body length, keel length, shank length, body height, drumstick length, thigh length, wing length, shank circumference, comb length, body girth, neck length and beak length were appropriately measured. Data obtained were used for principal component, stepwise discriminant and cluster analyses mainly to identify the traits that could best describe the broiler body conformation. Four (4) principal components were extracted in Arbor Acre broiler strains; this explained 56.68% of the total variation in the parameters sampled across the strains. Six (6) principal components (PC) were extracted for Hubbard and Ross strains, while three (3) principal components were extracted for Marshall and these explained 59.46%, in Hubbard, 58.67% in Ross and 55.52% of the total variance, respectively. The result also showed that PC1 had the largest share of the total variance which was highly significant in the studied morphometric traits and discriminate each of the strain studied. Canonical discriminant functions showed a distinct Hubbard under function1 and a distinct Marshall under function 2. Standardized canonical discriminant analysis indicated that wing length and body girth had highest value for discriminating under function1 and 2, respectively. This study indicates that available strains of broiler chickens can be distinguished through the use of some multivariate analysis.