A 13-landmark morphometric system was used for 617 specimens to investigate the hypothesis of species differentiation
among Cobitis keyvani, Cobitis faridpaki and an unknown Cobitis sp. population from the River Babolrud in the southern Caspian Sea basin. Univariate analysis of variance showed significant differences among the means of the three groups for 41 out of 78 standardized morphometric measurements. In discriminant function analysis, the overall assignment of individuals into their original groups was 58.0 %. The proportions of individuals correctly classified into their original groups were 58.1 %, 64.7 % and 44.1 % for Cobitis sp., C. faridpaki and C. keyvani,respectively. Principal component analysis (PCA) and canonical variates analysis (CVA) confirmed the significant difference among these populations (λ de Wilks = 0.42, P = 0). The CVA scatter plot showed 617 specimens grouped into three distinct areas with some overlap. Clustering based on Euclidean distances among the groups of centroids using an UPGMA indicated segregation of the three populations into two distinct clusters: Cobitis sp. and C. faridpaki in one group and C. keyani in other group. Also the unknown population of spined loach from the River Babolrud was recognized as C. faridpaki and the river as a new habitat for the species.