Issues of phytolith taphonomy at Palaeolithic sites: Investigation and results from Attirampakkam, India

International audienceInvestigating the taphonomy of phytoliths at open-air archaeological sites is essential for reconstructing past land cover and assessing the plant resources available to hominins, particularly when other indicators (e.g., pollen) are scarce. Here we analyse phytolith abundance and diversity encountered in distinct ‘reservoirs’ at and around the Palaeolithic site of Attirampakkam, South India: (1) in the present-day vegetation (database of 134 species), generating on that basis a fine classification of species-specific phytolith morphotypes; (2) in the topsoil (surface); and (3) at greater soil depths above the archaeological layers (subsurface). We then compare the data to results previously obtained from the deeper underlying archaeological layers. Comparison of the assemblages retrieved from the different reservoirs shows that morphotypes most likely to get carried over from the living plants to the oldest fossil specimens were subsets of herbaceous taxa (Poaceae and Cyperaceae) and a few woody monocotyledons (Arecaceae). The high fidelity of these taxa ranks them among the most reliable for conducting palaeoenvironmental studies at this site, and perhaps more widely in southern India. Provided those taxonomically traceable, non-redundant and non-multiple morphotypes are prioritised over herbaceous dicotyledons and most woody taxa, inferences about environmental conditions and changes at the time of hominin occupation during the Pleistocene can be made with a fair margin of confidence