Evolution of human sexual dimorphism: Using assigned resampling method to estimate sexual dimorphism when individual sex is unknown.

Abstract

Sexual dimorphism is a major component of variation within a species, and reflects many aspects of history, socioecology and adaptation. Understanding the modern human pattern of sexual dimorphism may benefit from understanding the pattern of its change through time. The evolutionary perspective takes fossils as the primary data source, which poses several problems. Fossil samples are often quite small in sample size, fragmentary, and sex is usually unknown. Although several methods have been proposed to measure sexual dimorphism in fossil samples, none has been fully successful in overcoming the limitations imposed. In order to study the evolution of sexual dimorphism, a reliable method has to be developed to estimate the sexual dimorphism in fossil samples. In this study, I developed a new method, the assigned resampling method (ARM), for estimating sexual dimorphism in a sample of small size and unknown sex. The performance of ARM was explored by comparing the ARM estimates in a number of skeletal variables with their true sexual dimorphism in samples of known sex. Samples of three closely related species were used: modern humans (n = 91), chimpanzees (n = 46), and gorillas (n = 56). Forty measurements of the cranium, mandible, lower canine, humerus, radius, femur, and tibia were taken. For each variable of each sample, the ARM estimate of sexual dimorphism was calculated, treating the sample as if the sex were unknown. Each estimate was then compared with the true value based on known sex. Results show that ARM provides reasonable approximations of the true sexual dimorphism. Moreover, the performance of ARM was shown to be stable under modification of the initial assumptions. ARM was used to test the null hypothesis that there is no systematic change in the variation of sexual dimorphism after the speciation event in the <italic>Homo</italic> lineage at about two million years ago. Six fossil samples were analyzed for the variables of cranial capacity, dentition, and the humerus. Results show no systematic pattern; therefore, it is concluded that the modern human pattern of sexual dimorphism originated shortly after the speciation event of the <italic>Homo</italic> lineage, and has undergone no systematic change since that event.