The phytolith archaeological record: strengths and weaknesses evaluated based on a quantitative modern reference collection from Greece

The phytolith record from archaeological strata is a powerful tool for reconstructing aspects of past human behaviour and ecology. Considerable insights into the problems and potentials of this record can be obtained by studying phytoliths in modern plants. We prepared a phytolith reference collection of modern plants mostly from northern Greece that includes quantitative information on phytolith concentrations (number of phytoliths per gram dry organic material) and morphological assemblages. Here we analyse this reference collection with the aim of evaluating the strengths and weaknesses of the archaeological phytolith record. The reference collection comprises 62 different plant taxa. Of them, 28 are trees, 1 conifer, 10 shrubs and 1 herb, 3 domesticated annual legumes and 19 monocotyledons. We demonstrate quantitatively the extent to which woody species, legumes and fruits of dicots contribute only small amounts of phytoliths to the sediments per unit tissue dry weight, while leaves of trees and shrubs produce significant amounts of phytoliths and grasses are prolific phytolith producers. We compare the data of this reference collection with a similar reference collection from another Mediterranean ecosystem (Israel). The comparison indicates that some aspects of phytolith production are probably genetically controlled, whereas others are environmentally controlled. We note that despite the fact that woody species produce few phytoliths per gram of tissue, their record can be most informative when taking into account other properties of the sediments. We also note that jigsaw puzzle-shaped phytoliths normally form in plants that grow in humid conditions, but may form in arid environments when the plants are irrigated. This study clearly shows the extent to which some plants can be under-represented and others over-represented in the phytolith record. Knowing the extent of this bias can greatly improve our interpretation of the phytolith record.