Towards improved detection and identification of crop by-products: Morphometric analysis of bilobate leaf phytoliths of "Pennisetum glaucum" and "Sorghum bicolor"

Abstract Better detection and taxonomic identification of cereal leaves is expected to result in a better understanding of the presence and function of crop products at archaeological sites. Therefore, this paper focuses on bilobate phytoliths from leaves of Pennisetum glaucum (L.) R. Br and Sorghum bicolor ssp. bicolor (L.) Moench, which are two important crop plants that regularly co-occur at archaeological sites in Africa and the Indian subcontinent. These two taxa are further compared with Panicum miliaceum L. and Setaria italica (L.) P. Beauvois, which is of relevance for agricultural sites in prehistoric South Asia, Eastern Asia, Africa, and part of Eurasia where the four crops certainly or presumably co-occur. Leaves of Pennisetum glaucum and Sorghum bicolor were systematically sampled to explore the variation of short cells and to collect 27 morphometric variables of 3100 bilobate phytoliths with newly developed open-source software. This study provides new information on the occurrence of cross-like and notched (nodular) bilobate short cells in leaves of pearl millet and trilobates in sorghum, which is of relevance for taxonomic distinction. The morphometric variables of the bilobates phytoliths do not allow for taxonomic classification between P . glaucum and S . bicolor . Possibilities for taxonomic distinction between the leaves of these taxa should thus be searched in other directions than bilobate phytolith morphometry. The obtained morphometric data are nevertheless important since they allow for future comparison with other taxa. Indeed, morphometric analysis allows for distinction between Pennisetum glaucum / Sorghum bicolor , Panicum miliaceum and Setaria italica . Furthermore, one P . glaucum population that was grown in a rather different climate than the others also shows different bilobate morphometry results. This difference between P . glaucum populations points to phytolith morphometry possibly being influenced by environmental settings. Moreover, it has implications for sampling strategies of similar research and the validity of morphometric identification criteria based on data from few reference populations or reference populations from a single region.