WEED PHYTOSOCIOLOGY AND CROP HUSBANDRY - IDENTIFYING A CONTRAST BETWEEN ANCIENT AND MODERN PRACTICE

The potential of crop weeds as indicators of farming practices is discussed and, in particular, the usefulness, in this respect, of indicator species, of Ellenberg's indicator values and ecological groups. and of phytosociology is assessed. The paper illustrates the use of the higher phytosociological syntaxa for the understanding of ancient (cereal and pulse) crop husbandry regimes. The use of crop weeds to infer husbandry practices is complicated by the fact that many weed seeds are selectively removed during the successive stages of crop cleaning. To explore this problem further, crop samples were collected from traditional (i.e. non-mechanised) farmers in Greece. The weed seeds present in these samples allowed different soil types and crop types to be distinguished. This distinction was much clearer when by products from the same stage of crop processing (winnowing, sieving etc.) were compared. Of particular significance to attempts to reconstruct husbandry practices was recognition of the effect of crop processing on the major weed classes, Secalinetea and Chenopodietea. The weed seeds present in the modern by-products of different processing stages were compared and the proportion of Chenopodietea to Secalinetea character-species was found to decline through the processing sequence. This trend probably reflects the adaptation of the Secalinetea to dispersal with the seed corn. The representation of different weed classes and alliances was then compared for these ethnographically collected by-products and for samples from comparable stages of processing at bronze age Assiros Toumba in Greek Macedonia. A greater proportion of seeds of Chenopodietea character-species was found in samples from the bronze age site. This contrast can be interpreted in a number of ways. Consideration of the particular phytosociological alliances represented and of the storage context of some of the samples indicates that this is unlikely to be a result of contamination by ruderal plants. Likewise, the spring cultivation of millet cannot account for the relative abundance of Chenopodietea in stored wheat and barley. The conclusion, therefore, is that a husbandry regime more akin to present-day intensive garden than extensive field cultivation is indicated. Archaeological weed assemblages from other parts of Europe suggest that such, an intensive regime may have been widespread among early farmers and that some aspects at least of this regime may have persisted locally into tater periods.