RÉSUMÉ
Geographical discrimination of mulberry leaves is very important for their efficacy and quality as a traditional Chinese medicine. Stable hydrogen, oxygen, and carbon isotope ratios were measured in 292 mulberry leaves collected at 2 growth stages in 2 seasons from 8 regions of China. A stepwise linear discriminant analysis (LDA) approach were proposed to combine with stable isotope technology to tracing the origin of mulberry leaves. The results showed that leaves sampled in autumn were extremely depleted in 2H and 18O and slightly enriched in 13C compared with leaves sampled in summer, correlated with the effect of season, transpiration and photorespiration on stable isotopes. δ2H and δ18O of the leaves were enriched during the growth process. The overall discrimination accuracy of the autumn tender model was 81%, demonstrating that analysis of δ2H, δ18O, and δ13C is a promising technique for tracing the geographical origin of mulberry leaves, although season, growth stage and number of samples affect the accuracy of discrimination.