The complexity of the structure of the dental surface of the first lower molar (m1) of mainland and island populations of the reed vole Alexandromys fortis (Büchner, 1889) was assessed using two different methods. Using the traditional approach and methods of geometric morphometry (GM), the morphological variability of the anteroconid m1 of this species of the mainland (the south of Primorsky Krai) and the islands of the Peter the Great Gulf was analyzed. The traditional approach was based on describing the folding of the occlusal surface using three features: 1) folding of the lingual side of the anterior unpaired anteroconid loop (AC); 2) the shape of the labial side of the AC; 3) the shape of the fourth lingual reentrant angle. On the mainland part of the species' range, the highest occurrence was shown by morphotype III-C-a, which also predominated in the Late Pleistocene samples of the species in Primorye. The research revealed increased morphotypic diversity in island populations of the reed vole, which is probably due to the phenotypic and genotypic characteristics of the founders of island populations, as well as inbreeding, which can cause the release of hidden diversity and an increase in the number of rare morphotypes. The most numerous morphotype on the De Livrona – Durnovo Islands was II-C-a, rare for the mainland. In general, the most complex forms are present on the islands. Using a geometric morphometry approach, the anteroconid was divided into two separately studied structures: the anterior unpaired loop of the anteroconid (AC) and the base of the anteroconid (AB). The distribution of AC configurations in the space of the two principal components took on an arcuate shape, revealing a “forbidden zone.” The main changes occur on the lingual side of the AC. Regression analysis showed a direct relationship between the length of the contour (mm) and the complication of the shape of the anterior unpaired loop. The base of the anteroconid (AC) also undergoes transformation: it either stretches or contracts, while if the anterior unpaired loop of the anteroconid (AB) is more complex, then simplification and compression of its base is observed (negative regression). Analysis of the results of using two approaches, traditional and GM methods, in the analysis of the variability of the anteroconid section m1 of A. fortis showed extreme subjectivity in the use of the first one and the wide possibilities of the second one for identifying different aspects of molare transformation using mathematical methods.
