Compound-specific hydrogen isotopic analyses are gaining incredible interest in paleoenvironmental studies due to the intimate relationship that exists between the δD of sedimentary compounds and climatic parameters (temperature, precipitation, humidity...; Sachse et al., 2012). To contribute to this line of methods, the PalHydroMil project (www.palhydromil.eu, ANR JCJC 2010) aims at developing a new proxy of past climatic conditions based on the hydrogen isotopic composition (δD) of an original molecular biomarker, miliacin. Miliacin was detected in Lake le Bourget (French Alps) sediments, where it has a single plant source: broomcorn millet (Panicum miliaceum) (Jacob et al., 2008). The evolution of miliacin concentration in sediments during to the Bronze Age is presently interpreted in terms of varying amounts of millet produced in the catchment. Whether this reflects any direct or indirect climatic control is to be tested. δD values of time sequences of this molecule extracted from twin sedimentary cores in Lake le Bourget is expected to provide clues to test this hypothesis, assuming that the relationship between the δD values of miliacin and the controlling parameters are understood and quantified. We grew millet in controlled environment chambers in which treatments varied in the following parameters: (i) δD values of the hydroponic solution (HS) (ii) relative humidity; (iii) hydric stress as controlled by polyethylene glycol (PEG) concentration; (iv) subspecies of broomcorn millet. The isotopic compositions of water in the hydroponic solution and in different organs were determined after cryogenic distillation on an Aquaprep (Isoprime irMS coupled to a Gilson 222 XL) for δ18O and on a PyrOH (Isoprime irMS coupled to a Eurovector Elemental Analyser) for δD. At maturity seeds were collected, miliacin was extracted and purified before determining its δD values by GC-irMS. δD values of the non-transpiring organs (roots and stems) were similar to those of the HS water. In contrast, water in transpiring organs (leaves and panicles) were depleted relative to their HS where δD values of HS water were low and enriched where δD values of HS water were high. These results are consistent with the physical process of selective transpiration. δD values of miliacin strictly paralleled the δD values of leaf water with a constant biosynthetic fractionation of 120 ‰. Three different relative humidity applied to the millet plant (48, 61, 74%) impacted the leaf water δD values but had no effect on the biosynthetic fractionation between leaf water and miliacin (120 ‰). Miliacin δD values were only affected by a 10 ‰ enrichment when the amount of PEG increased from 0 to 198 g l-1 (0 to -5 bars of osmotic potential). Higher amounts of PEG induced poor development of the plant, excluding any conclusion on the isotopic impact of a larger hydric root stress. The δD values of miliacin extracted from two varieties of P. miliaceum (var. ruderal and a cultivated var. sunrise) growing under the same climatic conditions was exactly the same, thus implying that the biosynthetic fractionations between leaf water and miliacin were unaffected by genetic differences. Our results indicate that the δD values of miliacin are mainly controlled by leaf water δD values with a constant biosynthetic fractionation and limited or no effect of hydric stress and varieties. In turn, leaf water δD values are affected by δD values of soil water and transpiration. REFERENCES Jacob et al., 2008. Millet cultivation history in the French Alps as evidenced by a sedimentary molecule. Journal of Archaeological Sciences 35, 814-820. Sachse, D., et al., 2012. Molecular paleohydrology: Interpreting the hydrogen-isotopic composition of lipid biomarkers from photosynthezing organisms. Ann. Rev. of Earth and Planetary Sciences 40, 221-249.