Reconstructing infant feeding practices through stable isotope analysis: applying intrabone sampling to a medieval–early modern Estonian dataset

• Intrabone sampling was tested on Estonian medieval to early modern non-adults. • Three data points from growing long bones were compared to assess feeding status. • Nineteen individuals showed possible feeding-related physiological changes. • Breastfeeding lasted until ∼1 year, weaning extended to 3 y, sometimes over 4. • Intrabone sampling aids identifying feeding trends and selection for further studies. The differential sampling of skeletal elements represents an appealing research avenue in the study of infant feeding practices in past populations. It provides fresh insights into enduring traditions in childrearing and feeding, dietary shifts, and disease trends while taking into account bone turnover dynamics. In this study, carbon and nitrogen stable isotope analysis was conducted on 120 bone samples from 40 individuals under the age of six years old, collected from nine urban and rural cemeteries in Southern Estonia (13th–17th centuries AD). An intrabone sampling approach was adopted, comparing three data points from growing long bones (proximal metaphysis, diaphysis, and distal metaphysis) to identify possible breastfeeding and weaning statuses at death and throughout life. Among the whole dataset, nineteen individuals showed isotopic evidence of possible feeding-related physiological processes: five indicated breastfeeding, six weaning, and eight were indeterminate. Breastfeeding individuals ranged from full-term/neonatal to around 10 months old, while weaning children were between 2.5 and 4.5 years. These results support bulk isotope estimates, confirming breastfeeding lasted up to one year and weaning extended beyond three, sometimes past four years. No significant difference between urban and rural individuals was detected in terms of age at death or skeletal pathology, possibly suggesting homogeneous feeding habits for individuals of similar ages and different health statuses in these two contexts. The results of this study allowed increasing resolution in the reconstruction of infant feeding strategies in medieval and early modern Estonia; additionally, they confirmed that intrabone sampling of non-adults is effective for identifying key trends in breastfeeding and weaning practices, particularly in commingled assemblages or when skulls and teeth are unavailable. This method also constitutes a valid preliminary tool for selecting individuals for more detailed analysis through incremental dentine sampling techniques.