In this study, honey bee (Apis mellifera L.) genotypes from different regions of Turkey were comparatively analyzed using numerical classification methods based on certain morphological characteristics.

Since the morphological traits related to the external structures of honey bees have different measurement units, the standardized Z-scores of these values were used to obtain a similarity matrix based on Euclidean and squared Euclidean distances. Using the Cluster module, the Single Linkage, Complete Linkage, and Ward’s methods were compared, and it was determined that the Ward method provided the most appropriate classification.

As a result of this analysis, the clustering solutions formed by the similarity matrix among the genotypes were evaluated using Hotelling’s Lawley Trace and Wilks’ Lambda statistics.

Among all three classification methods, the Ward method yielded the most meaningful classification, and under this method, the honey bee genotypes were grouped into five distinct clusters (P<0.001).