Defining species boundaries, due to morphological variation, often represents a significant challenge in paleozoology. In this paper we report results from multi− and univariate data analyses, such as enhanced clustering techniques, principal coordinates or− dination method, kernel density estimations and finite mixture model analyses, revealing some morphometric patterns within the Eocene Antarctic representatives of Palaeeudyptes penguins. These large−sized birds were represented by two species, P. gunnari and P. klekowskii, known mainly from numerous isolated bones. Investigations focused on tarso− metatarsi, crucial bones in paleontology of early penguins, resulted in a probability−based framework allowing for the "fuzzy" partitioning the studied specimens into two taxa with partly overlapping size distributions. Such a number of species was supported by outcomes from both multi− and univariate studies. In our opinion, more reliance should be placed on the quantitative analysis of form when distinguishing between species within the Antarctic Palaeeudyptes.