Factors affecting the formation of an archaeological wood charcoal assemblage
Three major stages affect the composition of an archaeological wood charcoal assemblage, from the source area (past vegetation) to the laboratory space (charcoal identification and interpretation). They are the following:
a. Environmental and cultural parameters (species availability, firewood selection and context of fuel consumption).
b. Physical parameters (burning, depositional environments and post-depositional transformations).
c. Field and laboratory techniques (recovery, sampling, sub-sampling and identification).
The environmental and cultural parameters influencing firewood collection and consumption (alongside a detailed review on the quantification debate) are discussed in detail in Asouti and Austin (in press). A major influence on charcoal interpretation has derived from the “Principle of Least Effort” (PLE). According to this paradigm, firewood collection in the past took place in those woodland catchments situated closest to the settlement and all species were collected in direct proportion to their occurrence in the site surroundings. Therefore, the frequencies of individual taxa in any given assemblage, allowing for potential biases introduced by differential preservation, rates of charcoal deposition and sampling strategies, should be considered as an accurate reflection of their abundance in the natural environment at the time of human habitation. Ultimately, the plotting of individual taxon frequencies should allow a precise reconstruction of vegetation shifts that could in turn be interpreted as a reflection of environmental (including climatic) change in the past. Other researchers have also pointed out that species availability cannot be considered strictly as a function of net species abundance in past vegetation. Other parameters, such as for example the presence of deadwood (easily collectable from the forest floor), may have a significant influence in species selection. There follows that species which tend to produce higher quantities of deadwood (e.g. through frequent shedding of their branches) are more likely to be gathered by prehistoric fuel collectors. Other species may be preferentially selected because they regenerate fast as a response to woodcutting and clearance. Another parameter affecting species selection may be cultural perceptions of preference and avoidance that determine what can and cannot be collected in a given socio-cultural context.
I chose the following quote to emphasize this point and make therefore clear that the scope of the analysis should extend (where possible) beyond mere vegetation and climate reconstruction, and encompass fundamental issues relating to the “rationale for the exploitation” of fuel and woodland resources:
‘From an anthropological perspective, the rationale for the exploitation of environmental resources is embodied in the social relations that govern their appropriation ... therefore an understanding of these relations must constitute a starting point for the analysis of economic behaviour. Prehistorians, on the other hand, are inclined to reduce the economy to the ecological dimension of population - resource balances, treating the constituents of the social domain as mere ephemera, super-imposed on a more fundamental and enduring set of biological imperatives’ (Ingold 1984: 3)
The impact of physical and cultural parameters: burning, depositional environments and post-depositional transformations
Burning
Charcoal forms as a result of the thermal decomposition of wood when burned in an inadequate supply of oxygen. It is possible to distinguish four successive stages in the combustion process, corresponding to different temperature environments (Beall 1972): dehydration (up to 200º C), char formation (200º-280º C), pyrolysis or carbonisation (280º-500º C) and ignition (above 500º C). The first two phases can be described as endothermic. At this stage, wood looses some 35% of its total weight in the form of vapour, non-combustible gases and organic compounds. During the stage of pyrolysis, the chemical degradation of cellulose and lignines produces flammable gases and aromatic compounds (generally classed as tars). Thermal decomposition that up to this point was dependent on a heating medium (e.g. kindling) becomes exothermic: temperature rises spontaneously and wood is set into flames. The transition from carbonisation to ignition can be very quick. At this stage, charcoal glows and may turn into ash (the inorganic by-products of charcoal combustion) if enough oxygen is available.
Carbonisation of wood causes a number of changes in its physical properties of which mass reduction, discolouration and shrinkage are the most apparent (Beall et al. 1974). Generally, it has been observed that two-thirds of the mass loss will occur between the temperatures of 200º to 400º C, whereas the total reduction has been estimated to represent approximately 80% of the original mass (ibid.). Shrinkage occurs in all three surfaces (tangential, longitudinal and radial) with longitudinal contraction being the most severe. Carbonisation results in volumetric shrinkage too that increases in proportion to the length of the log’s exposure to fire (McGinnes et al. 1971). Release of volatiles may also generate cracks (e.g. radial and longitudinal fissures) due to mechanical stresses caused by the uncontrolled drying of wood when heated (Zicherman and Williamson 1981). Despite these deformations, the gross anatomical structure of wood, as well as most of its micro-structural elements, remain largely unaffected (ibid.).
Such differences as they may arise in burning conditions (temperature, intensity of fire, length of exposure, heating environment) and wood properties (size, moisture content, taxon anatomical structure) have a direct effect on taxonomic representation within a wood charcoal assemblage. Small-sized woods such as shrubs, which may also be used as kindling, are more likely to be consumed entirely in lower temperatures, whereas pieces of wood lying at the centre of the fire heat faster and thus can burn completely (Smart and Hoffman 1988). On the other hand, charcoal that is buried in the ash at the bottom of the hearth has a greater chance of preservation due to lack of oxygen (ibid.).
It has also been observed that, as a general rule, soft woods such as willow and poplar, tend to conflagrate faster and reduce to ashes easier than dense ones such as oak, elm and chestnut (Rossen and Olsson 1985; for a corresponding classification of woods according to their value as fuels, see Boulton and Jay 1946: 112). However, density alone is not a secure indicator of the potential effects of carbonisation in terms of mass reduction and volume loss. Parameters such as the size of wood, its chemical composition and its moisture content, have themselves a critical impact on the rate of burning, and thus on the amount of mass decomposition different types of wood will undergo when subject to combustion.
Depositional environments
Wood charcoal macro-remains found in archaeological deposits are likely to represent either the remains of firewood or the burned vestiges of structural timber resulting from catastrophic conflagrations (the remains of charcoal used as fuel are not dealt with here; for further references, see Chabal et al. 1999). In the case of firewood, the type of fire installation and the associated discarding practices may influence in various ways the preservation of charred remains. For example, it has been observed through ethnoarchaeological studies that open-air hearths are rarely contained. Cooking of plant and animal foods may cause intermixing of deposits and considerable displacement of cinders, ash and fire-cracked stones, due to the constant searching in the ashes for roasted foods. Over time, as fires are re-kindled and the same processes repeated, the centre of the hearth tends to drift (Binford 1983: 157). On the other hand, hearths located inside habitation structures are usually lined with stones so as not to allow the spread of fire to flooring materials (Binford 1983: 156).
Furthermore, cooking habits may have a variable effect on the preservation potential of wood charcoal macro-remains. Covered hearths used for the preparation of meat or plant foods without direct exposure to fire, are more likely to retain wood charcoals in a good state of preservation than open fires (March 1992). Another ethnographic example, drawn from observations on the use of roasting pits by the Alyawara Australian Aborigines, serves to illustrate the point:
‘The burning wood is flamed up to a fast burn. Singeing the game as well as occasionally beating the burning wood results in the accumulation of a substantial bed of charcoal. Once it is judged that enough charcoal has been scaled off the burning wood, the remaining burning sticks are pulled out and tossed to the side, leaving only the charcoal in the pit and on the platform ... The kangaroo is nested in the charcoal within the pit, followed by the birds wrapped in leaves to hold in the juices formed during cooking. Once the hot sand and charcoal from the platform are shovelled into the pit to cover the meat, the cooking begins.’ (Binford 1983: 167)
In all cases, ash and hot charcoals may be regularly cleaned from the base of the fire installations and scattered around them or beyond the limits of the main activity area. Amongst the Hazda of northern Tanzania, ashes from domestic and open hearths alike were dumped along the edges of the campsite, a process that resulted in the formation of various types of secondary refuse deposits, from simple concentrations to distinct ash dumps (O’Connell et al. 1991). Similar patterns of refuse disposal are reported for the Efe Pygmy campsites in the Ituri forest of northeastern Zaire. The only time when hearth maintenance does not take place is in the event of camp abandonment, when ‘hot fires are left to burn out and nobody will be around later to sweep up and discard the ashes’ (Fisher and Strickland 1991). Bartram et al. (1991) in their description of the camps of the Kua San hunter-gatherer groups in east-central Kalahari, observe that fireplaces used for cooking were cleaned of their contents more frequently than small fires lit for other purposes (e.g., lighting tobacco pipes, straightening of bows, arrows and digging sticks, skin pegging, etc.) In most cases, ashes and hot coals would be swept away from the opening of the adjacent hut or windbreak structure. This resulted in the formation of ash scatters around one side of the hearths. Less often, the entire contents of the fireplace would be scooped onto dumps located at the edges of the camp area.
Maintenance of domestic spaces on permanent sites also entails the regular clearing of ashes from fire installations and their disposal in a spatially removed location, such as a midden, fill or abandoned structure. Only the smallest items will escape the sweeping of floors and hearths or those left as vestiges of the last phase of use. In fact, any sort of waste that is likely to obstruct indoors activities such as obsolete bulky items or those representing potential hazards to the house occupants (e.g., stone knapping debris) is rapidly disposed off as secondary refuse. An ethnographic example comes from the waste disposal routines at Hasanabad, a village in western Iran, whereby fireplaces are cleaned out on a daily basis and their contents are thrown into dump areas lying at the borders of the village or, more rarely, at a corner of the house courtyard (Watson 1979: 37). The same areas receive rubbish generated by floor sweepings, human and animal excrement, plus discarded household items and food processing waste (pieces of cloth, sticks, paper, bits of wool, broken utensils, bone fragments, goat and wild sheep horns and horn cores, etc.)
Such refuse deposits may display very complex depositional histories. Apart from the disturbance caused by dogs and chicken scavenging on their contents, they are frequently dug up by villagers for the extraction of earth (locally called “chineh”) to be used in the construction of walls for houses, stables and other buildings. Some of the organic debris may actually re-enter the household fireplace: in Hasanabad, “chineh” earth mixed with animal dung is used for the manufacture of dung cakes to fuel domestic fires (Watson 1979: 39).
Post-depositional transformations
Either as primary or as secondary refuse, all classes of archaeological remains are subject to further distortion under the impact of various post-depositional processes that will transform their original patterning, in both qualitative (spatial distribution and contextual associations) and quantitative terms. Trampling, variations in surface exposure and sediment moisture, reheating and freeze-thaw may result in further breakdown of wood charcoals. Moreover, vertical and horizontal displacement of charred remains may occur as a result of bioturbation (burrowing, earthworm activity, root penetration) and later erosion of archaeological deposits due to eolian and fluvial action.
Another factor influencing charcoal preservation in archaeological strata are the chemical conditions of the sediment matrix. Due to their large porous surfaces, wood charcoals are susceptible to the accumulation of mineral inclusions and precipitates, which can in turn decrease fragment porosity and increase density (Greenlee 1992). Therefore, and despite the lack of a universally agreed scheme for assessing wood charcoal taphonomy (as it happens for example in animal bone studies) an overall appraisal of the nature of post-depositional conditions is essential for understanding their impact on the preservation of wood charcoal macro-remains.
The impact of field and laboratory techniques: recovery of wood charcoal macro-remains, sampling, subsampling and identification
Recovery of charcoal macro-remains
Nowadays, most of the charred plant remains are retrieved from archaeological deposits by using some system of flotation. Indeed, with the exception of fine-grained sandy or ashy sediments, other methods of retrieval such as dry screening will cause excessive damage of charcoal fragments. Furthermore, the mesh sizes usually employed for both dry and wet sieving (in the range of 5mm) let too much material pass through, hence introducing a further source of bias in taxon representation.
A biased picture of sample composition also arises when charcoal fragments are manually collected from the archaeological strata. The only instance when handpicking should take priority is in the case of burned structural timber, whereby it is important to maintain the integrity of individual specimens for studying technological aspects of wood use or for dating purposes (i.e. dendrochronology and radiocarbon dating). Otherwise, manual retrieval invariably results in the selective choice of larger fragments thus leading to samples of small size and the subsequent recovery of only the most commonly present taxa (Keepax 1988: 43, Chabal et al. 1999: 65).
Returning to flotation, experimental work on its potential effects on the rates of recovery and preservation of wood charcoals (Keepax 1988: 70-79, Brady 1989, Greenlee 1992) has demonstrated that charcoal fragments incur a variety of destructive mechanical stresses that may lead to re-fragmentation and subsequent loss of charred plant material. These are of two types: a. Impact stresses, that potentially affect charcoal fragments during the flotation process, and b. Internal static stresses that may cause further breakage as charcoal dries out and moisture gradients develop from the outer layers towards the wet inner core, leading to differential compression and tension stresses. The obvious result of these processes is the accumulation of greater numbers of fragments in smaller size fractions (Keepax 1988: 76, Brady 1989: 210).
Aside from fragmentation, there have also been noted significant disparities and variation between pre- and post-flotation total charcoal weights. These appear to correlate with differences in anatomical properties and, even more so, sedimentary conditions which ultimately control the amount of soil chemical compounds precipitated as residues in the pores of charcoal fragments (Greenlee 1992: 279-280). It follows that the use of total charcoal weights as a means to evaluate relative taxon abundance, occupation intensity and preservation conditions between deposits runs the risk of failing to account for similar influences on charcoal recovery and loss rates.
Sampling
When choosing samples in the field, the archaeobotanist needs to understand in advance the function and provenance of the wood from which the charcoal remains retrieved on site originated. In other words, he or she should be able to tell whether wood charcoals represent the remains of fuel (domestic or otherwise) or structural wood, by exploring their contextual associations. Do they derive from domestic fire installations, open hearths, cremation fires and in situ burnt structural features or they were found scattered in external spaces? Not all deposits have the same interpretive potential for addressing questions relating to the uses of wood, the site environment and the modes of woodland exploitation (Chabal et al. 1999: 61). In practice, this objective can be achieved through the detailed consideration of the excavation records and the finds inventory for each sampled location (e.g. animal bone, seeds, etc.)
Once function and provenance have been established, attention must be paid to the duration of the activities represented in the archaeological record. In their effort to isolate those botanical assemblages that are most likely to represent the product of intentional human action (e.g. fuel consumption) instead of random events associated with post-depositional disturbances, archaeobotanists have to prioritise samples that are judged, on archaeological grounds, to stand for a certain duration of activities. Chabal et al. (1999: 62-63) distinguish two main types of deposits for which such predictions are feasible:
a. Short-term deposits
Typical examples offer contexts holding primary in situ refuse, such as hearths and fire installations. Wood charcoals found scattered in them are likely to represent the remains of their last use prior to abandonment, possibly heavily transformed by divers post-depositional factors and, furthermore, containing a limited number of taxa. Even if substantial quantities of charcoal are retrieved from fire installations and a high degree of taxonomic diversity is established, the probability that these are related to the specific circumstances of the last firing event and do not represent a long-term trend cannot be eliminated. However, such assemblages may furnish important information on the structure and function of particular hearth types. Equally, destruction levels can provide evidence on aspects of wood use (e.g., choice of building materials and woodworking), but little nonetheless on the duration of the activities represented in the archaeological record.
b. Long-term deposits
Into this group fall archaeobotanical assemblages that represent mainly discarded refuse, such as those deriving from external, non-domestic areas (middens, fills, etc.). Pending on the predicted frequency of the disposal events (day-to-day or at longer intervals) they are better suited to characterize lasting patterns of firewood selection and consumption. Such deposits are also most likely to produce a high diversity of woody taxa and thus maximise the potential of the analysis for palaeoenvironmental reconstruction. Finally, charcoal assemblages retrieved from such areas can be expected to have been subject to broadly the same range of post-depositional alterations, thus allowing for a more precise evaluation of the effect sedimentary conditions have imparted upon taxon representation.
Sub-sampling
Another
major consideration when sampling for charred plant remains concerns
the size and number of samples that are likely to provide statistically
meaningful results. Optimal sample size (the quantity of fragments per
sample that should ideally be examined by the charcoal analyst) varies
following sample properties and the degree of accuracy required (detailed
discussion of archaeobotanical sampling requirements can be found in van der
Veen and Fieller 1982).
Several authors have observed that taxonomic recovery follows an exponential curve: the number of the taxa present in a sample rises sharply as the first few charcoal specimens are examined and then settles down as more fragments have been identified (Keepax 1988: 44, Smart and Hoffman 1988, Chabal et al. 1999: 67). Keepax (1988: 120-124) has suggested that a minimum number of 100 fragments per sample should be examined, which may actually extend up to 300-400 fragments depending on the diversity observed in the charcoal assemblage. Chabal et al. (1999: 66) raise this lower limit to 250 fragments, with 400-500 fragments considered as the optimal subsample size per excavated level.
Provided that results are for their most part replicated across a certain number of samples from each excavated stratum, the size of the subsample can be more realistically set to 150-250 fragments per sample. It has been observed that the point when recovery curves tend to level off is not solely a function of the number of examined fragments but also depends on the spatial extent of the sample population across the excavated level (Badal Garcia 1992). Indeed, by maximising the spatial coverage of sampling it may be possible to compensate for temporary, and for that reason mostly unpredictable, “levelling-off” sometimes observed in individual recovery curves (Figueiral 1992). Keepax makes a similar point when she states that
‘Over-identification of individual samples does not compensate for insufficient sample number ... A certain number of samples must always be identified to account for between-sample variation’ (Keepax 1988: 45)
As to the number of samples to be analysed (depending on the research objectives and the available resources) recovery from twenty-five to fifty samples on average is considered as a reasonable minimum, whereas for more complex archaeological sites that have a greater variety of depositional contexts, one hundred or more samples may be required. For multi-period settlements, intra-site comparisons between contexts and/or excavated levels necessitate similar provisions (Keepax 1988: 45-47).
Equally important in terms of subsample selection is the size range of the fragments chosen for analysis. Opting for the larger fragments alone runs the risk of overlooking naturally small-sized taxa (e.g. shrubs) or those procured mainly in the form of twigs and small branches, all of which are likely to be better represented in smaller size ranges. Such a selection can be achieved through splitting the sample and randomly choosing a portion of it (Willcox 1974), by “grab-sampling” fragments of different size and shapes (Miller 1985) or passing the dry flot through a stack of sieves of graded mesh sizes and subsampling each size fraction (Zalucha 1982: 79). Of all three methods, dry sieving is by far the most efficient for this purpose. “Grab-sampling” suffers from a lack of standardization (it is inherently subjective), whereas splitting the sample with a riffle-box will invariably result in further re-fragmentations. Finally, the use of grid systems can prove very time-consuming without also being altogether “free” of similar subjective elements (cf. van der Veen and Fieller 1982).
Identification
Several factors may inhibit the precise identification of wood charcoal macro-remains. In many cases, it can prove very difficult to identify individual specimens to species level, due to the similarities in anatomical structure exhibited amongst members of the same family and/or genus (Hather 2000: 11-12). In addition to this, variation in anatomical characters can occur even amongst specimens belonging to the same taxon, due to differences in genetic stock, habitats, growing conditions, age and part (bark, stem, twig, branch, root) of individual plants and the exposure to occasional hazards such as fire, frost and pest outbreaks (Dimbleby 1967: 107-108, Wilson and White 1986: 198-199).
One remedy to this situation is the use of wood anatomical descriptions and extensive comparative collections covering particular geographical regions. However, such collections and/or descriptions usually comprise only trunk wood specimens and are mostly assembled from thin sections of fresh wood. For the purpose of charcoal identification, this can be problematic. Characters such as the size and dimensions of pores, vessel elements and rays that may be of diagnostic value in fresh specimens (e.g. Fahn et al. 1985) in charred specimens are either seriously deformed, due to shrinkage and cracking, or missing altogether as is the case with certain types of parenchyma and also septate fibres and crystals. Other features as well (e.g. spiral thickenings, intervascular pits) can be difficult to locate and describe with any precision, due to variations of lighting on charcoal surfaces during microscopic examination, or if fragments are not studied under sufficiently high magnifications (Western 1969: 112-113, 115).
Difficulties may also arise due to the small size of individual fragments. The required size will vary between taxa, pending on the relative frequency of the diagnostic features preserved within the charred specimen and the uniqueness of these features amongst the woody plants of the region (Smart and Hoffman 1988). Some anatomical characters that occur infrequently may be absent from specimens smaller than 4mm, whereas others can be quite distinctive even in small fragments if not shared between many taxa (e.g. the size and structure of multiseriate rays in oak).
All these problems can be partly overcome by using modern charred specimens as reference material, by examining in detail all three anatomical surfaces (transverse, radial and tangential) at least for the more “problematic” taxa, and through the appropriate adjustment of identification criteria. However, as it so often happens, specific identifications apart from ascribing a family or genus label are in most cases unattainable.
References cited
Asouti, E., and P. Austin (2005) Reconstructing woodland vegetation and its relation to human societies, based on the analysis and interpretation of archaeological wood charcoal macro-remains. Environmental Archaeology 10: 1-18.
Badal-Garcia, E. (1992) L' anthracologie préhistorique - À propos de certains problèmes méthodologiques. Bulletin de la Société Botanique de France 139: 167-189.
Bartram, L. E., E. M. Kroll, and H. T. Bunn (1991) “Variability in camp structure and bone food refuse patterning at Kua San hunter-gatherer camps,” in The interpretation of archaeological spatial patterning. Edited by E. M. Kroll and T. D. Price, pp. 77-148. New York and London: Plenum.
Binford, L. R. (1983) In pursuit of the past. London: Thames and Hudson.
Brady, T. J. (1989) The influence of flotation on the rate of recovery of wood charcoal from archaeological sites. Journal of Ethnobiology 9: 207-227.
Beall, F. C. (1972) Introduction to thermal analysis in the combustion of wood. Wood Science 5: 102-108.
Beall, F. C., P. R. Blankenhorn, and G. R. Moore 1974. Carbonised wood-Physical properties and use as an SEM preparation. Wood Science 6: 212-219.
Boulton, E. H. B., and B. A. Jay (1946) British timbers: their properties, uses and identification. London: Adam and Charles Black.
Chabal, L., L. Fabre, J.-F. Terral, and I. Théry-Parisot (1999) “L'anthracologie,” in La botanique. Edited by C. Bourquin-Mignot, J.-E. Brochier, L. Chabal, et al., pp. 43-104. Paris: Errance.
Dimbleby, G. W. (1967) Plants and archaeology. London: Baker.
Fahn, A., E. Werker, and P. Baas (1986) Wood anatomy and identification of trees and shrubs from Israel and adjacent regions. Jerusalem: The Israel Academy of Sciences and Humanities.
Figueiral, I. (1992) Méthodes en anthracologie - Étude des sites du Bronze Final et de l'Âge du Fer du Nord-Ouest du Portugal. Bulletin de la Société Botanique de France 139: 191-204.
Fisher, J. W., and H. C. Strickland (1991) “Dwellings and fireplaces: keys to Efe Pygmy campsite structure,” in Ethnoarchaeological approaches to mobile campsites: hunter-gatherer and pastoralist case studies. Edited by C. S. Gamble and W. A. Boismier, pp. 215-236. International Monographs in Prehistory-Ethnoarchaeological Series 1. Michigan: Ann Arbor.
Greenlee, D. (1992) “Effects of recovery techniques and postdepositional environment on archaeological wood charcoal assemblages,” in Deciphering a shell midden. Edited by J. K. Stein, pp. 261-282. London: Academic Press.
Hather, J. G. (2000) The identification of northern European woods. London: Archetype.
Ingold, T. (1984) “Time, social relationships and the exploitation of animals: anthropological reflections on prehistory,” in Animals and archaeology: 3. Early herders and their flocks. Edited by J. Clutton-Brock and C. Grigson, pp. 3-12. BAR (IS) 202.
Keepax, C. A. (1988) Charcoal analysis with particular reference to archaeological sites in Britain. Ph.D. Dissertation, University of London.
McGinnes, J. E. A., S. A. Kandeel, and P. S. Szopa (1971) Some structural changes observed in the transformation of wood into charcoal. Wood and Fiber 3: 77-83.
Miller, N. F. (1985) Palaeoethnobotanical evidence for deforestation in ancient Iran: A case study of urban Malyan. Journal of Ethnobiology 5: 1-19.
O’Connell, J. F., K. Hawkes, and N. B. Jones (1991) “Distribution of refuse-producing activities at Hadza residential base camps: implications for analyses of archaeological site structure,” in The interpretation of archaeological spatial patterning. Edited by E. M. Kroll and T. D. Price, pp. 61-76. New York and London: Plenum.
Rossen, J., and J. Olson (1985) The controlled carbonisation and archaeological analysis of SE U.S. wood charcoals. Journal of Field Archaeology 12: 445-456.
Smart, T. L., and E. S. Hoffman (1988) “Environmental interpretation of archaeological charcoal,” in Current palaeoethnobotany. Edited by C. A. Hastorf and V. S. Popper, pp. 165-205. Chicago & London: University of Chicago Press.
van der Veen, M., and N. Fieller (1982) Sampling seeds. Journal of Archaeological Science 9: 287-298.
Watson, P. J. (1979) Archaeological ethnography in western Iran. Viking Fund Publications in Anthropology 57. Tuscon: University of Arizona Press.
Western, A. C. (1969) An attempt at the ecological interpretation of charcoals with special reference to material from Jericho. B.Sc. Dissertation, University of Oxford.
Willcox, G. (1974) A history of deforestation as indicated by charcoal analysis of four sites in eastern Anatolia. Anatolian Studies 24: 117-133.
Wilson, K., and D. J. B. White (1986) The anatomy of wood: its diversity and variability. London: Stobart & Son.
Zalucha, L. A. (1982) Methodology in palaeoethnobotany: A study in vegetational reconstruction dealing with the Mill Creek culture of northwestern Iowa. PhD Dissertation, University of Wisconsin.
Zicherman, J. B., and R. B. Williamson (1981) Microstructure of wood char. Wood Science and Technology 15: 237-249.
© Eleni Asouti, 2006