Natural abundances of stable nitrogen and carbon isotopes (δ 15 n and δ 13 C) can vary with both dietary intake and metabolic (specifically catabolic) state. In low-income countries, weaning is a period of dietary transition from milk to plant-based foods and a high-risk period for malnutrition. We explored how diet and malnutrition impact hair δ 15 n and δ 13 C in young children by an observational, crosssectional study in Cox's Bazar District, Bangladesh [255 children, 6-59 months with 19.6% wasted (7.1% severely) and 36% stunted (9.8% severely)]. Hair δ 15 n and δ 13 C exhibited exponential decreases with age, with the loss of one trophic level (3.3‰ and 0.8‰, respectively) from 6 to 48 months, which we associate with the shift from exclusive breastfeeding to complete weaning. After adjustment for age and breastfeeding status, hair isotopic values were unaffected by wasting but lower in severe stunting (−0.45‰ to −0.6‰, P < 0.01). In this population of young children, whose isotopic values in hair primarily depended on age, we failed to observe any effect of wasting, likely due to opposite, compensating effects between dietary and metabolic changes involved. In contrast, we evidenced low δ 15 n and δ 13 C values in severely stunted children that likely indicate chronic exposure to diets low in animal products. The natural abundances of stable carbon and nitrogen heavy isotopes (δ 13 C and δ 15 N) in body proteins reflect those of the diet, plus a small discrimination factor, known as the trophic step, which is due to the preferential elimination of light isotopes and is smaller for carbon (+1‰) than for nitrogen (+3-4‰) 1,2. Since δ 13 C and δ 15 N values in the body primarily reflect those of the diet, they have long been used in ecology and archaeology to reconstruct trophic webs and characterize ancient diets 3-6. More recently, they have been used in epidemiology as biomarkers to assess the consumption of fish, animal products, and sugar 7-13. In this regard, measuring isotopic natural abundances in hair has several advantages. Hair is easily and non-invasively sampled and its isotopic values correlate well with those in other body proteins. Moreover, hair keratin is an ideal isotopic archive because it is produced sequentially and is inert after synthesis 14,15 , provided that a sufficient number of follicles is sampled to limit the potential bias due to the presence of a small proportion of hair in the telogen phase, i.e. not growing 15-17. Considering an average hair growth rate of 1 cm per month 17,18 and a one-week delay between hair synthesis and bulb exit 19 , the 5 mm of hair closest to the scalp contains isotopic information for the period between −3 and −1 weeks before sampling.