The identification of complete and fragmentary remains of E. ferox within waterlogged archaeological horizons at GBY necessitated study of the site's taphonomy, with particular attention to the degree to which hominins were the agents responsible for the accumulation, spatial patterning and physical condition of the remains.
Examination of site formation processes at GBY indicates that the archaeological horizons were sealed rapidly, demonstrating a high integrity of preservation and an excellent context for studies of spatial patterning (Alperson-Afil et al. 2009). Observations from several archaeological horizons at the site may be summarised as follows:
Table 1 presents the occurrence and frequencies of the two types of water nuts at GBY. Both E. ferox and T. natans nuts are present in both archaeological horizons and geological layers, the latter being devoid of lithic artefacts. The marked difference between archaeological and geological layers in quantities of E. ferox and T. natans arises primarily from different sampling strategies. Geological layers were minimally sampled along the walls of the trenches (Trenches II–VI) as compared to the extensive studies conducted on archaeological horizons. Within geological layers, there are differences in the occurrences of the two types of nuts. In Layer III-7, both nuts are well represented with E. ferox predominating. In other layers, T. natans is the dominant seed, with its highest occurrence being in Layer II-9, a gray mud sediment characteristically deposited under the highest water column and which represents the deepest part of the paleo-lake. Sedimentologically, a group of geological layers (Layers II-9, II-10, II-11, III-7) with a high occurrence of T. natans represent coquinas. The absence of lithics in these layers does not necessarily mean that they are archaeologically sterile. Although this may appear to be a contradiction, our experience at the site leads us to suggest that layers that are apparently archaeologically sterile may bear artefacts further along the strike, beyond the boundaries of excavated areas. In some cases (e.g. Layer II-12 in Goren-Inbar et al. 2002b, 81), an isolated microartefact indicated the presence of an as yet unknown and unexcavated archaeological horizon. Thus, the presence of seeds in archaeologically sterile layers may be explained in two ways; as a natural occurrence or as a result of human behaviour. In the absence of further excavations, neither possibility can be confirmed.
Remains of E. ferox are extremely abundant in the rich archaeological horizons (Table 1, rows marked by *). They occur in varying frequencies in these layers, as well as within the archaeological complex of Layer II-6. Hominin activities varied significantly between the 15 rich archaeological horizons, as reflected by differences in the presence and frequencies of finds (stone artefacts, fossil bones and organic remains). Here, we propose the hypothesis that the spatial patterning, physical condition and concentration of E. ferox nut fragments at the site is the result of intentional hominin gathering and processing rather than of natural factors. The key evidence for this is the co-occurrence of nuts and their fragments with phantom hearths and pitted stones (Goren-Inbar et al. 2002a), which occur in 10 of the 15 archaeological strata containing E. ferox remains (Figure 3). One possible use of the pitted stones was as 'anvils' for popping nuts. On the basis of ethnographic parallels discussed below, we believe that the association of many E. ferox fragments with pitted stones and phantom hearths is an indication of intentional roasting followed by popping to extract the nut from its hard shell and exploit its maximum nutritive value (Jha and Barat 2003). A new type of anvil, a component of the percussive tool assemblage, was recently identified at GBY (Goren-Inbar et al. in prep). These passive (dormant) percussive tools (de Beaune 2000) consist of thin natural basalt slabs with two flat surfaces that are sometimes broken, probably as a result of use. This is the first report of such thin anvils from an Acheulian site. They are characterised by the presence of pits and other abrasive damage marks on one or both flat surfaces and sometimes on the thin edges. Such anvils are not found at the lake's edge but were selected for their particular morphology and specific characteristics (e.g., hardness and non-vesicular nature) and brought from exposures of basalt flows that are currently unknown in the vicinity of the site. The total number of these anvils at GBY is 36, 26 of which are pitted. Along with pitted stones (the latter primarily of basalt but also of limestone), these anvils occur diachronically throughout the cultural sequence of the site (Table 1). The same is true for pitted stones, although their availability was most probably more extensive since their morphology is more varied (Goren-Inbar et al. 2002a). Together, these percussive tools form the bulk of the tools that we propose were used to pop the E. ferox seeds and crack other types of nuts.