Multivariate Discrimination of Czech Autochthonous Horses

Abstract

Abstract We have used a discriminant function analysis to compare morphology of five Czech autochthonous breeds (including two colour varieties as independent breeds) to test whether a small number of basic morphological variables (wither height, thoracic, nose and shin perimeter, length of head) can discriminate them. The breeds included Czech Warmblood, black and grey colour variety of the Old Kladruby horse, Czech-Moravian Belgian horse and Silesian Noriker. The tested individuals were assigned with overall 81.9% classification success to correct breed. The best classification result reached Czech Warmblood 95.7%, the black Old Kladruby horse 87.5% and Silesian Noriker, respectively, 85.7%. Czech-Moravian Belgian horse showed a poorer success of classification (60%). Discrimination analysis identified the most important variables related to their head (nose perimeter and length of the head). Based on discrimination model both colour varieties of the Old Kladruby horse clustered more closely. Similarly both cold-blooded breeds (Czech-Moravian Belgian and Silesian Noriker) grouped more together and locations of the Czech warmbloods were more apart from all others. Such result is in concordance with the origin history of these horses.