The cladogram resultingfrom the UPGMA cluster analysis showed that populationsBirjand and Shiraz then Sari and Isfahan were classifiedtogether.
Songhor population was separated completely.
Cluster analyses divided populations into three main groups(Fig. 4).
Variance of coordinates of 18 landmarks (geometric morphometric method) on forewings as well as coordinates of landmarks of average shape were evaluated. Results indicated that the fourteenth landmark, the junction of the Rs and 2r-mveins, had the maximum variation (S2=0.00006) in Iran;landmarks 18 and 12 had the next greatest variations (S2
=0.00005 and 0.00004 respectively). Also, the eighth landmark,the junction of the 1m-cu and cu veins, had the minimum
variation (S2=0.00001) (Fig. 2).
Furthermore, results ofMANOVA of forewings (with 32 shape variables) in geometric morphometric method showed significantdifferences between populations of the five areas in Iran (F=4.16, p=0.000).
Allometry testing on forewings was performed. The results identified a significant difference between the size (centroid
size) of the front wings and 32 shape variables (Wilks’ lambda=0.54, p=0.00); therefore, the shape variations of the
forewings were not uniform. Then, with the increasing of size of the front wings, wing shape changed. Area populations
based on 32 shape variables of forewings were classified by canonical variate analysis (CVA). Cross-validation tests based
on CVA of fore wings correctly classified 77.5% of the colonies (Fig. 5).
Statistical analysis results indicated that the honeybee populations of Songhor, Shiraz and Birjand were separated 80, 82.5 and 85 % from populations of other areas (table 6).
Relations of geographical populations were evaluated using the UPGMA method.
The cladogram resulting from the UPGMA cluster analysis showed that populations of Sari and Isfahan were classified together.
Birjand and Shiraz populations were separated completely.
Cluster analyses divided populations into three main groups (Fig.6).