| Sumario: | Objectives: Sexual dimorphism in primates reflects evolutionary, ecological, and social pressures and varies widely across species, complicating its analysis. This study builds on previous research to investigate sexual dimorphism in the long bones of great apes, aiming to improve sex estimation and evaluate the effectiveness of various methods in classifying unsexed specimens. Materials and Methods: External morphology of humeri and femora from modern great apes—including Homo, Pan, Gorilla, and Pongo—was analyzed using 3D anatomical landmarks and geometric morphometrics. Various statistical approaches and dimensionality reduction techniques were employed alongside classification methods, including supervised machine learning algorithms. Results: Size, rather than shape, emerged as the main factor distinguishing male and female long bones in great apes—except in Pan, where dimorphism is minimal and classification accuracy remains low. Incorporating size improved classification accuracy for Gorilla, Pongo, and Homo, with results indicating strong dimorphism in Gorilla and Pongo, moderate dimorphism in Homo sapiens, and minimal dimorphism in Pan. Discussion: This study shows that combining geometric morphometrics with machine learning can enhance sex classification of great ape long bones. Nonetheless, limitations such as small or imbalanced samples highlight the need for larger datasets and further research—including internal bone structure—to better understand skeletal dimorphism and its evolutionary drivers.
|
|---|