Part II. Phytolith identification in thin sections from Palaeolithic hearths

Introduction

      Hearths, ash and charcoal are important elements of the archaeological record as they can provide information on the manner in which fire was used at a site. Wood ash from burned trees comprises on the average 98% calcite and the remainder is a mixture of siliceous aggregates and siliceous phytoliths (Schiegl et al., 1994, 1996). The siliceous aggregates are composed of soil minerals embedded in a biogenic matrix of silica that also contains abundant aluminum, iron, and potassium (Schiegl et al., 1994), whereas the phytoliths are almost pure amorphous silica (Piperno, 1988). Thus, at pH values of around 8 or below - the common pH of water in karstic environments (Stumm, & Morgan, 1970) - phytoliths are highly insoluble.

      Hearths in prehistoric sites are not always easy to identify. Post-depositional processes can seriously alter the chemical nature of the ash and the charcoal, making them difficult to identify (Schiegl et al., 1994, 1996; Karkanas et al., 2001). Calcite readily reacts with phosphate released from decomposing organic matter. Ash therefore undergoes a series of dissolution and reprecipitation steps in which a series of phosphate minerals are formed. After several such steps in which part of the phosphate minerals dissolve and are not reprecipitated, only the relatively stable component of ash, mainly siliceous aggregates and phytoliths, remain. With time the siliceous aggregates also undergo diagenesis forming different precipitation products (Schiegl et al., 1996). Thus under these conditions the phytoliths are probably the most durable of all ash components and they will therefore be among the most stable biogenic and authigenic minerals present in a cave environment. One of the main objectives of the present project is to assess the technological aspects of fire use principally during the Middle and the Upper Palaeolithic periods. Examination of various characteristics of combustion features from these periods – type and state of combustible used, temperature of burning, how the fire was constructed – reflect short-term, individual behavioral strategies, and it will thus be possible to reveal behavioral differences among them.

      Phytoliths are formed in cell walls and within specialized cells of many plants (Piperno, 1988; Pearsall, 1989). Their morphologies can be used to reconstruct the vegetal remains in an archaeological site. Their potential for contributing to the understanding of fire use has been demonstrated by the quantitative analyses of phytoliths in the caves of Kebara (Albert et al., 2000), Tabun (Albert et al., 1999), Grotte XVI in France (Karkanas et al., 2001) and Hayonim cave (Albert et al., 2003).

      We present here the preliminary results obtained through the identification of phytoliths in thin sections, in different ash layers from Kebara and Hayonim caves (Israel) and Pech de l’Azé IV (France).

Materials and Methods

      Phytoliths were observed directly in thin sections (30 μm thick) using a petrographic microscope Olympus BX41 at magnifications of 100x and 400x. [For the preparation of the thin sections, archaeological samples were consolidated by immersing the undisturbed samples with polyester resin diluted with styrene and hardened with Methyl Ethyl Ketone Peroxide (MEKP)].

      The presence of phytoliths was investigated and their morphology identified according to available modern plant reference collection from the same area (Albert and Weiner, 2001; Albert et al., 2003; Albert et al., 2000).

      Morphological analysis of phytoliths permits the identification of the types of vegetation (woods, grasses, and reeds), the vegetative stage of the plants (stem, leaves, inflorescence, etc.); in specific cases, the genus and species of the plants can be determined. This type of study can therefore reflect on the type of plants used as fuel and the season of collection of those plants. In fact, for the phytolith analysis to be fully significant it should be quantitative, so that the different types of phytoliths could be counted (number/g of sediment) and their relative quantity compared with the quantity in control sediments.

      To perform the quantitative analyses, the next step on our research, will be the study of loose sediment according to the method of Albert et al. (2000); these loose, bulk samples have been collected at the same time as the micromorphological samples, and will be processed this coming year.

Results

      The list of the samples studied is shown in Table 1. The most interesting results were observed in some of the samples from Kebara Cave (highlighted in blue in Table 1). All these samples had in common the presence of huge amounts of vegetal tissue. Single phytoliths were also noted, most of them corresponding to the leaves of grasses and few of them to the inflorescences of the same plants.

      The most impressive samples were Keb-96-12-a and Keb-96-14-a, which showed complete fragments of vegetal tissue, some of them corresponding to grasses (Figure 1a and b). Other plant remains probably ascribed to the leaves of grasses were also noted (Figure 1c). Other types of plant tissues can be noted in Figure 1d: parenchyma cells from non-woody plants (Figure 1d). The rest of samples show the same type of material but in lesser abundance than in the two samples mentioned above. These vegetal remains were associated with clay, quartz and burnt wood.

      None of these remains was identified in other samples from Kebara or in the samples from Pech de l’Azé IV and Hayonim.

      One of the problems observed through this preliminary study is the difficulty in identifying single phytoliths: they are either masked by the presence of the other material or cut in such a way that they cannot be observed. Most of the time, the identified phytoliths correspond to the same morphological types: long cells from grasses; other morphologies cannot be seen.

      Thus, there are important biases in the identification and interpretation of phytoliths occurrence:
   a) the fact that not all the morphologies can be identified, can lead to a misinterpretation of the results. For instance, grasses produce 20 times more phytoliths than dicotyledonous wood/bark and 16 times more phytoliths than dicot leaves. Thus, if one of the most common phytolith identified is the long cells (which corresponds to grasses), we can arrive at the false conclusion that grasses were the most common type of plant remain identified in the samples.

   b) A second problem is the fact that, with the use of thin sections, it is not possible to quantify the amount of phytoliths and therefore, we cannot know the proportion of the different plants deposited in the sediments.

      Two possible solutions could be attempted to solve this problem:
   i) to prepare thin sections that are thinner (e.g., 10 μm), and
   ii) to collect “twin” samples in which one block is exploited for thin sections and a second, adjoining block for phytoliths. The latter can then be analyzed following standard procedures (Albert et al., 1999), which would permit a detailed morphometric and quantitative study of the phytolith remains. The selection of “twin” samples will take into account the preliminary results obtained through the study of thin sections and at the same time will enable us to contextualize the presence of plant material in relation to the rest of the deposits in the sample.

Table 1 - List of Phytolith Samples

SITE	LAYER	APPROXIMATE AGE (ka)	PERIOD	SAMPLE #	DESCRIPTION
Hayonim Cave	E,F,G	130-250	Acheulean/Yabrudian/MP	Hay-96-3 Hay-98-202 Hay-98-218 Hay-99-335-a Hay-00-500	Lens-shaped features in bedded silts. Massive ash, mostly calcareous, with little charcoal (Dumps).
Kebara Cave	VII-XIII	60-48	MP	Keb-88-20 Keb-89-11??? Keb-89-8 Keb-96-1-a Keb-96-1-b Keb-96-2 Keb-96-3-a Keb-96-3-b Keb-96-4-b Keb-96-4-c Keb-96-5-a Keb-96-5-b Keb-96-6-a Keb-96-6-b Keb-96-8-1 Keb-96-8-2 Keb-96-10 Keb-96-11-a Keb-96-11-b??? Keb-96-12-a Keb-96-12-b Keb-96-13-a Keb-96-13-b Keb-96-14-a	Abundant features of various types
Pêche de l’Aze	X,Y,Z	??	MP	PDA-4-52 PDA-4-56 PDA-4-57-a PDA-4-57-b PDA-4-111-a PDA-4-111-b	Dark-brown organic rich sand, with different thin burnt features; locally massive and undifferentiating

Figure 1: Photomicrographs of thin sections showing fragments of vegetal tissue from Kebara Cave. Pictures taken at 100 X. a), b), c) and d): sample 96-14A e) sample 96-13B.

Boston University - 6/10/2007