Evaluating macroscopic sex estimation methods using genetically sexed archaeological material: The medieval skeletal collection from St John's Divinity School, Cambridge.

Sarah Inskip,Christiana L Scheib,Toomas Kivisild,Xiangyu Ge,John Robb,Anthony Wilder Wohns

doi:10.1002/ajpa.23753

Abstract

ObjectivesIn tests on known individuals macroscopic sex estimation has between 70% and 98% accuracy. However, materials used to create and test these methods are overwhelming modern. As sexual dimorphism is dependent on multiple factors, it is unclear whether macroscopic methods have similar success on earlier materials, which differ in lifestyle and nutrition. This research aims to assess the accuracy of commonly used traits by comparing macroscopic sex estimates to genetic sex in medieval English material.Materials and MethodsSixty‐six individuals from the 13th to 16th century Hospital of St John the Evangelist, Cambridge, were assessed. Genetic sex was determined using a shotgun approach. Eighteen skeletal traits were examined, and macroscopic sex estimates were derived from the os coxae, skull, and os coxae and skull combined. Each trait was tested for accuracy to explore sex estimates errors.ResultsThe combined estimate (97.7%) outperformed the os coxae only estimate (95.7%), which outperformed the skull only estimate (90.4%). Accuracy rates for individual traits varied: Phenice traits were most accurate, whereas supraorbital margins, frontal bossing, and gonial flaring were least accurate. The preauricular sulcus and arc compose showed a bias in accuracy between sexes.DiscussionMacroscopic sex estimates are accurate when applied to medieval material from Cambridge. However, low trait accuracy rates may relate to differences in dimorphism between the method derivative sample and the St John's collection. Given the sex bias, the preauricular sulcus, frontal bossing, and arc compose should be reconsidered as appropriate traits for sex estimation for this group.

Highlights

Estimating biological sex from human skeletal remains is fundamental to most bioarchaeological research (Brickley & Buckberry, 2017)
Eighteen skeletal traits were selected for sex estimation based on their presence in widely used standards applied to remains from the British Isles
Comparison of macroscopic and ancient DNA (aDNA) sex estimates demonstrate that methods derived from postmedieval and modern collections are accurate when applied to the medieval skeletal collection from St John's Hospital, Cambridge

Summary

Introduction

Estimating biological sex from human skeletal remains is fundamental to most bioarchaeological research (Brickley & Buckberry, 2017). Concerns over the relevance of trends drawn from post-/industrial samples to material from other periods and locations has been raised (Ubelaker, 2008; Walker, 2008), because of the fact that sexual dimorphism varies between groups due to differences in growth and development, disease (Ubelaker & DeGaglia, 2017), activity patterns (Krishan et al, 2016), general secular trends (Godde, 2015), and genetic admixture This issue is highlighted in research where some methods or sexually dimorphic features appear accurate in one population, but not for others (e.g., Maat, Mastwijk, & Van der Velde, 1997; MacLaughlin & Bruce, 1990; Spradley & Jantz, 2001; Walker, 2005). It is not entirely clear how accurate commonly used methods/traits as assessed on recent material are for estimating sex for skeletal material predating the industrial period

Objectives

Results

Discussion

Conclusion