High-resolution isotopic data link settlement complexification to infant diets within the Roman Empire

Abstract

Our study explores the potential relationship between infant feeding practices and settlement complexity in the Roman Empire through high-resolution Bayesian-modeled stable isotope measurements from incremental dentine. We compiled isotopic data from permanent first molars of individuals from various Roman sites: five from Bainesse (UK), 30 from Thessaloniki (Greece), along with new carbon and nitrogen isotope analyses from four individuals from Pompeii and six from Ostia Via del Mare (AVM). Our results reveal significant inter-site variability in breastfeeding durations, ranging from 1.5 years to approximately 5 years. Notably, individuals from the highly complex urban centers of Pompeii and Thessaloniki ceased breastfeeding around or below the 2-year weaning threshold recommended by Roman physicians. In contrast, individuals from the rural site of Ostia AVM and the site of Bainesse, near the northern frontier of the Roman Empire, generally ceased breastfeeding after 2 years of age. The link between settlement complexity and duration of breastfeeding observed in our study may have resulted from adherence to medical guidelines, support infrastructures, and/or strategies to mitigate financial constraints within households.

Keywords: Bayesian modeling; Roman Empire; incremental dentine stable carbon and nitrogen isotope analysis; infant feeding practices; settlement complexification.

Fig. 1.

Map showing the distribution of sites analyzed in our study and from previous studies. Design by Michelle O’Reilly for MPI GEA 2024.

Fig. 2.

(A) Bayesian estimated 68 and 95% credibility intervals for weaning completion at Bainesse, Ostia AVM, Pompeii, and Thessaloniki. Individual OS1 from Ostia AVM is not included in the plot due to poor sample preservation. Individuals METi_197, METi_125 and METi_67 from Thessaloniki are also excluded because their credibility intervals could not be estimated. (B) Bayesian dietary estimates of primary food source protein intake for individuals from Pompeii and Ostia AVM, spanning from birth to approximately 9.5 years old. Gaps in the estimates (e.g. PO1 [3.0–3.5]) are due to excluded isotopic measurements that did not meet reliability criteria. See Supplementary Material S3 for detailed modeling methods and estimates. A larger version of Fig. 2A, complete with individual IDs, can be found in Supplementary Material S3.5. Figure design by Michelle O’Reilly for MPI GEA 2024.