The effect of water availability on the carbon and nitrogen isotope composition of a C4 plant (pearl millet, Pennisetum glaucum)

• Published in: Journal of archaeological science, reports reports, 2021-08, Vol.38, p.103047, Article 103047
• Main Authors: Sanborn, Lily H., Reid, Rachel E.B., Bradley, Alexander S., Liu, Xinyi
• Format: Article
• Language: English
• Subjects: African millet; Greenhouse experiment; Palaeoenvironmental reconstruction; Stable isotope analysis; Water restriction
• ISSN: 2352-409X
• DOI: 10.1016/j.jasrep.2021.103047

•Pearl millet δ13C values are positively correlated with water availability.•Regional origin influences pearl millet carbon isotope ratios to a similar degree.•Past water status cannot be reconstructed from pearl millet δ13C values alone.•Pearl millet δ15N values did not differ among water treatments. To evaluate the potential utility of isotope ratios in plant materials as an archaeological proxy for past crop water status, relationships between water availability and stable isotope ratios in C4 species must be established. This study quantified the isotopic values (δ13C and δ15N) of pearl millet (Pennisetum glaucum) seeds and leaves in response to varying degrees of water stress. Under greenhouse conditions, we exposed five strains of pearl millet to three different watering treatments. Pearl millet seed δ13C values (mean and SD = -13.9 ± 0.5‰, n = 48) and leaf δ13C values (mean and SD = -14.8 ± 0.7‰, n = 75) were positively correlated with water availability across 75 plants from five strains. The magnitude of the relationship for seeds (0.24 ± 0.04‰ per 0.1 m3 m−3 increase in soil moisture) and leaves (0.24 ± 0.06‰) was similar. The five strains showed differences in bulk carbon isotope ratios but had indistinguishable responses to water availability. Water availability had no discernible effect on δ15N in any of these strains. These results suggest that, while in some cases sensitive to water availability, the differences in the isotope ratios of pearl millet seeds and leaves across treatments were not of sufficient magnitude for reconstructing past crop water status without additional information.