Colluvium, or hillwash, is both an erosive deposit and a preserving medium for buried surfaces. The term encompasses slope deposits moved by shallow surface flow (or slope wash) or by mass movement (or creep or slide). Colluvium is soil- and/or sediment-derived material that accumulates on lower slopes. It is poorly sorted and heterogeneous, composed of any size grade from clay to coarse sand plus rock rubble, and can be up to several meters thick (Waters, 1992, 230–232; Selby, 1993, 243). Bedding and stratification are often poor and particularly hard to identify in the field (as opposed to in thin section), and colluvium may also contain a variety of artifact inclusions brought down from upslope. Colluvium may occur any place that possesses more than two degrees of slope, even beneath woodland (Imeson et al., 1980).
Hillwash can be generated by a variety of processes, of which devegetation and agriculture are two of the main instigators. Consequently, colluvial sequences frequently contain a record of human activities and reflect anthropogenic impacts on landscapes. Colluvium also has the ability to distort, bury, and preserve past landscapes, especially over the last 10,000 years. How and why deposits are laid down and subsequently modified, and at what rate, are central issues in geoarchaeology. Detailed information can be gleaned from colluvial sequences through the use of good field recording, particle size analysis, associated soil micromorphological and molluscan studies, and appropriate use of dating techniques. Thus, hillwash is a valuable resource in archaeology and geoarchaeology, but an understanding of its formation, depositional processes, and chronology is required.
This entry examines and illustrates the formation and depositional factors causing hillwash in a variety of landscape settings in order to demonstrate its importance in the interpretation of past land use and human activities and exemplify these through a number of case studies from around the world.
Formation of hillwash
Different amounts and rates of saturation will also affect the threshold at which soil movement occurs as well as the speed and distance of travel. For example, if a soil requires only 50 % saturation to become mobile (or plastic), shear will occur more quickly and more often, dependent, of course, on the nature of the vegetation cover and degree of slope; if a soil becomes saturated at 90 % moisture content, however, it will be less susceptible to shear and colluvial displacement downslope (Selby, 1993, 56–63). Plowing and overcropping can lead to the depletion of not just essential minerals for plant or crop growth, but the destruction of soil structure, thus making a soil more prone to destabilization and downslope movement. Soil texture also affects its movement, with fine sandy and coarse silty soils being much more susceptible to destabilization than well-structured silty clay loam soils (Selby, 1993, 106–122). This will be compounded when vegetative cover is removed or the soil is physically disturbed by plowing or animal trampling.
In addition to these formation factors, it is also worth considering the ecological concepts of thresholds of stability/instability (Butzer, 1982; Allen, 1992). Change in just one factor in a landscape, or a combination of several factors, might be sufficient to cause instability in a soil system. This could be as simple as an individual thunderstorm event on a dry, bare soil, or it could be a combination of factors, such as an unstructured sandy soil fabric on overgrazed and degraded grassland with 20° of slope and a prolonged rainy spell that causes a catastrophic shear and slumping of soil downslope. Each response will be governed partly by climatic, environmental, human, and land-use factors as much as the diverse pattern and different magnitudes of individual valley responses.
Rates of colluvial movement
Examples of measured rates of soil erosion from experimental plots in the southeastern United States (After Kirkby, 1969)
Type of vegetation cover
Soil loss (mm per year)
Bare abandoned land
Cultivated, with rows along the contour
Cultivated, with rows downslope
Rates of colluviation have been observed in many modern landscapes, and these data provide a good idea of the speed and volumes of material that can be moved downslope under particular conditions. For example, under coppiced oak/beech woodland on loessic (or windblown) soils in Luxembourg, as much as 6 cm of soil per 100 years had accumulated (Kwaad and Mücher, 1979; Imeson et al., 1980). Observed rates of soil loss can vary from 0.0045 g per square meter per year for areas of moderate relief under natural conditions to 0.045 kg per square meter per year for steep relief and rates of 4.5–45 kg per square meter per year on agricultural land (Young, 1969). A major controlling factor is the angle of slope; the total transport caused by sheet erosion has been observed to increase sixfold as the slope angle increased from flat to 25° (Moseley, 1973).
Factors that influence soil erosion and its severity are rainfall, runoff, wind, soil type, slope angle, and the amount of vegetative cover, both on localized and regional scales. Erosion tends to reach a maximum in temperate, semiarid areas with a mean annual rainfall of c. 250–350 mm per year (Langbein and Schumm, 1958). In more humid areas, and as one moves from drier to wetter environments, rates of erosion initially decrease rapidly to the point (at c. 600 mm per year) at which total vegetation cover is established; rates change little thereafter. For example, in the northern Cambridgeshire region, where the annual rainfall varies between 500 and 600 mm (Burton, 1981), the greater vegetative cover tends to counteract the erosive effect of greater rainfall. In contrast, the Rio Puerco of New Mexico experiences very low rainfall of 200 mm per year that occurs during thunderstorms; land use associated with intensive livestock grazing has contributed to conditions resulting in massive river incision of up to 11–12 m over approximately the past 120 years (French et al., 2009).
Rainsplash is probably the most important detaching agent and contributes considerably to runoff. Splash back following raindrop impact on a level surface has been observed to move stones 4 mm in diameter up to 20 cm, while 2 mm sized stones can be displaced up to 40 cm, and even smaller stones up to 150 cm (Kirkby, 1969). Short-lived, intense, and prolonged storms of low intensity have the greatest erosive effect (Morgan, 1979; French et al., 2009).
Surface runoff or overland flow occurs on slopes when the soil’s infiltration capacity is exceeded (Kirkby, 1969). Overland flow transports soil particles detached by rainsplash, often creating distinct gullies or channels (Figure 2a). It has been suggested that overland flow covers two-thirds or more of hillsides in a drainage basin during the peak period of a storm (Horton, 1945). Grains of c. 0.5 mm in diameter (coarse silt or very fine- and fine-sized sand) are most easily moved, whereas both smaller and larger grains require a much higher threshold velocity. Clay tends to resist detachment (Farmer, 1973). Grains are not redeposited until very low flow velocities are reached (Morgan, 1979). Subsurface soil water flow erodes possibly only 1 % of the total material from a hillside (Roose, 1970).
The resistance of soil to detachment and transport depends on the steepness of slope, vegetative cover, and disturbance by humans. It also varies with soil texture, aggregate stability, shear strength, infiltration capacity, and the organic and chemical components (Selby, 1993, 106–122). The least resistant particles are silts and fine sands. Soils with a low shear strength or low cohesiveness are susceptible to mass movement, as are those with a low infiltration rate and low organic matter content (Morgan, 1979).
The effect of slope is to increase erosion with increasing slope angle (Kirkby, 1969). Numerous forces such as gravity, frost heave, rainsplash, soil texture, and the lack of vegetative cover all help to produce erosion on slopes. The amount of vegetative cover has a considerable effect on the susceptibility of soil to erosion. Its effectiveness in reducing downslope erosion depends on the height, density, and continuity of the canopy or ground/root cover. Both forest and dense grass are more or less equally effective at reducing erosion. Vegetation intercepts rainfall and reduces the velocity of runoff. Mean annual soil loss on bare ground can be as much as 100 times the volume of loss from dense grass-covered ground under the same environmental conditions (Morgan, 1979), and soil loss on vegetated slopes can be reduced by 10–30 % (Horton, 1945). Moreover, conversion of forest to arable land will dramatically increase erosion by up to 200-fold (Wolman, 1967). Thus, vegetation is a critical factor, and the removal of plant cover especially on slopes can considerably enhance the potential for erosion by overland flow, slope wash, or some other mode of transport.
Variable hillwash accumulations are common. For example, Bronze Age settlement and agriculture on the sand-dominated soils at Sutton Hoo have led to over 1.5 m of undifferentiated soil hillwash accumulation downslope in the adjacent Deben valley (Figure 3b). In contrast, two major episodes of hillwash accumulation, one predating and the other postdating the Romano-British period, were found within the center of the late Neolithic henge at Durrington Walls, Wiltshire (Figure 7b). Associated with this was a major erosive interval affecting the calcitic rendzina soil material that was found in old paleochannels of the Avon valley at Durrington Walls (French et al., 2012). Also, a mixture of thick deposits of soil and rock debris is known to aggrade episodically as colluvium in dry valleys of Sussex in southeastern England from at least the Beaker period, e.g., at Newbarn Combe on the Isle of Wight (Allen, 1992: Figure 4.2; Bell, 1992: Figure 3.3; Boardman, 1992). Near Szazhalombatta in the lower Benta tributary of the Danube, there were three episodic periods of hillwash accumulation – in the Bronze Age, Iron Age, and medieval times – each on an old land surface represented by a paleosol indicating a lengthy period of stability (French, 2010a) (Figure 7c). Several episodic phases of hillwash accumulation were also observed burying Neolithic paleosols in the Dhamar highlands of Yemen (French, 2003, 224–234; Wilkinson, 2005) (Figure 7d).
Nine-unit land-surface erosion model and catena sequence
The nine-unit land-surface model combines slope aspect, degree of slope, erosion, and soil forming processes (Dalrymple et al., 1968) (Figure 8). The uppermost unit (1) exhibits less than 1° of slope; it is characterized by pedogenic processes with vertical subsurface movement and is often associated with waterlogging or severe denudation. Below that, unit 2 exhibits 2–4° of slope, with both chemical and mechanical eluviation (or removal) by lateral subsurface water movement. Unit 3 is the upper part of the fall face with 35–45° of slope, and it is characterized by bare rock surfaces, sheet erosion, soil creep, and terracette formation. Unit 4 is the lower and steeper part of the fall face with 45–64+ degrees of slope, which is characterized by physical and chemical erosion leading to much bare rock, rock falls, and slides. Unit 5 is the mid-slope zone with 26–35° of slope, surface and subsurface water action, transport by mass movement, terracette formation, and both the removal and accumulation of soil and sediment material. Unit 6 is effectively the colluvial footslope zone, where there is subsurface water action, redeposition of material by mass movement, and some surface wash as colluvium, as well as transport further downslope in the form of hillwash and down valley in the form of alluvium. This colluvium may be laminated, nonlaminated, or massive, and it often accumulates on a buried soil that may or may not be truncated. Unit 7 in the floodplain is characterized by alluvial deposition as well as downstream water movement containing colluvially derived material as alluvium. Unit 8 in the active floodplain exhibits channel avulsion and erosion, bank slump, and fall. Unit 9 is the active channel itself, with bed transport down valley, as well as periodic aggradation and erosion. Obviously, not all units and slope angles will be applicable to every landscape encountered during fieldwork, but the model gives foreknowledge as to where areas of denudation, transformation, and burial may be occurring within a valley complex. As a corollary, it will also provide a good indicator of where buried soils and land surfaces may survive and where buried waterlogged deposits may be found intact for paleo-vegetational and land-use reconstruction. These data can also be used to inform a universal soil loss equation (USLE) to model landscape change through time (Wischmeier et al., 1971; Wischmeier and Smith, 1978; Ayala and French, 2005; French, 2010b) (see below).
Methods of analysis
Two main approaches have been used to study colluvial processes, namely, (1) observation and recording in the field backed up by (2) experimental studies in the field and laboratory. More recently, a much wider repertoire of techniques has emerged for routine use. Good field description and mapping of soil and colluvial deposits is an essential starting point for appraising soil erosion in any landscape. Aerial photography and remote sensing methods such as LIDAR are essential tools (Wilson, 2000; Donoghue, 2001; Bewley et al., 2005), perhaps even geophysical survey and the use of ground-penetrating radar (GPR) to determine sediment thicknesses (Clark, 2000; Gaffney and Gator, 2003) accompanied by systematic augering profiles and test pitting to ascertain depths and characteristics of the erosion/soil complex (French, 2003; French et al., 2007). Test pitting allows bulk sampling from the section face for molluscan studies (Allen, 1992). It also permits block samples to be taken for soil micromorphological analysis (Courty et al., 1989; French, 2003), as well as small bulk sampling for particle size analysis, pH, carbonates, loss on ignition, and magnetic susceptibility (Allen and Macphail, 1987; Canti, 1995; Bertran and Texier, 1999; English Heritage, 2004; Goldberg and Macphail, 2006), while permitting the recording of stone content and its orientation, recovery of artifacts for relative dating, and sampling and sieving to retrieve charcoal for radiocarbon assay (French, 2003; Goldberg and Macphail, 2006). If any clear horizon differences exist, or there are perceptible old land surfaces in evidence, optically stimulated luminescence (or OSL) dating can also be used effectively to provide sequence dating (Grün, 2001). Finally, both laboratory and field experimental data from known present-day landscape scenarios on different soil and geological types need to be compiled and used for comparison as analogues for potential rates of slope erosion under known conditions (Kwaad and Mücher, 1979; Evans, 1992).
Studies of molluscan assemblages may also be used to help define different localized environments within a valley catchment and to recognize cyclical land-use regimes (Allen, 1988; Allen, 1992). Usually, the molluscan faunas in hillwash reflect arable activity and are generally depleted assemblages with a narrow species range (Bell, 1983). Any temporary hiatuses or standstill horizons can reflect grassland or incipient soil formation (Kerney et al., 1964; Allen, 1992). Occasionally palynological (Scaife, 1984) and micromorphological analyses (Macphail et al., 1990; Macphail, 1992) can aid in the interpretation of land use and activities represented in the colluvial stratigraphy.
Once the formation, age, depths, and spatial extent of colluvial deposits have been determined, and the deposits have been mapped and set within their geological, soil, and vegetational contexts, it then becomes possible to model past and present landscape change. This involves using the tools of geographical information system (GIS) and the universal soil loss equation (USLE) and erodibility factor (K) (Wischmeier et al., 1971; Wischmeier and Smith, 1978) to model erosion in a landscape, as well as dynamic people-landscape interactions in the archaeological record (Kwan and Lee, 2004; Conolly and Lake, 2006; Wainwright, 2008; French, 2010b). For example, in the Troina valley of north-central Sicily, USLE and GIS modeling techniques were used to model the potential impact of Roman agriculture on the erosion record (Ayala and French, 2005). The study suggested that greater clearance for winter pasture land may have been the major driving force causing intensified soil erosion. Other recent studies (Barton et al., 2004; Barton et al.,2010) have effectively employed both computational applications (USLE and GIS) and developed them further using a Geographic Resource Analysis and Support System (GRASS) to model socio-ecological interactions. This type of approach will not only enable visualization of a landscape’s erosion record, but it will also allow (1) comparisons to be made between long-term landscape dynamics and the archaeological record as well as (2) further testing using different landscape settings and different scales of human activity.
Dating of colluvial sequences is rarely easy. Artifacts such as pottery and lithics are often included in eroded soils as hillwash deposits, but of course, they may be inverted or mixed in depositional terms, making them unreliable indicators of the age of any slope deposit. They should certainly not be used on their own without other corroborative data and dating methods. But, it is a different matter if standstill zones with in situ archaeological materials are found within a colluvial sequence, such as the Bronze Age and later sequence at Brean Down (Bell, 1990) (Figure 7a). This type of episodic deposition and standstill sequence allows dates “after which” (terminus post quem) and relative dating of the main phases in the profile based on their archaeological components. In addition, it is sometimes possible to make relative linkages to other regional events recorded in associated environmental, archaeological, and/or historical records that are already dated. For example, Nile flood records in pharaonic Egypt (Adamson et al., 1980; Brown, 1997, 6–7), extreme weather conditions such as those occurring during the Little Ice Age of northwestern Europe in the AD 1500s (Lamb, 1979), the huge eruption of Santorini in the Bronze Age (c. 1625–1645) (Aitken et al., 1988), biostratigraphic correlation with well-dated pollen records in the study area (Brown 1997, 47), and tree-ring series and ice core sequences can provide proxy dated records of major erosive and climatic events for some parts of the world.
Radiocarbon dating is an ideal method (Taylor, 2001), but it depends upon an organic component – such as charcoal, wood, bone, or other humic substance – being incorporated within the eroded material or stratified within any standstill horizons. There might nevertheless still be problems related to how the organic matter selected for dating was introduced into the eroded sediments, and uncontrolled taphonomic processes could produce inverted or otherwise erroneous dates. The reliability of any dated sequence is improved by obtaining numerous dates. For example, in the Dhamar Highlands of Yemen, sets of radiocarbon dates from the fifth millennium BC were obtained from the humic content of buried soils associated with several pre- and post-Neolithic hillwash sequences (Wilkinson, 2005), and in the central Rio Puerco valley of New Mexico, some 125 radiocarbon dates were obtained from charcoal in one reach of the Rio Puerco and its associated tributary Arroyo Tapia in order to date the 11 m thick colluvial/alluvial/buried soil sequence from about 5750 BC to the present (French et al., 2009).
Luminescence dating is another excellent technique, now increasingly being used because it yields much better accuracy than it has previously and is widely applicable (Aitken, 1997; Grün, 2001). Thermoluminescence dating (or TL) may be used on pottery and burnt flint, and optically stimulated luminescence (or OSL) works well for sedimentary sequences, especially where there are clear contacts of eroded material with buried land surfaces. For example, early Neolithic TL dates were obtained from pottery in hillwash deposits associated with two Neolithic sites in the Troina valley of north-central Sicily, and Bronze Age, Roman, and late medieval OSL dates were determined from the various stop/start alluvial profiles in the associated river floodplain (Ayala and French, 2005; R. Bailey, pers. comm.). Recent work on the Channel Island of Herm has successfully employed OSL dating techniques to chart the episodic deposition of windblown sands between about 1200 BC and AD 1600 on buried soils of Neolithic age (Bailiff et al., 2014; Scarre and French 2013).
In sum, the combination of dated archaeological associations and environmental sequences and the use of radiocarbon and/or OSL techniques are essential tools with which to date colluvial sequences. OSL dating should revolutionize future understanding of dynamic landscape change where there are repeated episodes of soil erosion and deposition.
Importance and conclusion
Colluvial deposits and the land surfaces and paleosols that they bury represent valuable resources for archaeological, paleoenvironmental, and landscape interpretation. They inform us not only about landscape development and land use in the past but also (1) the variability of human and natural impacts on valley systems, (2) the links between past and current processes of land degradation, and (3) possibly even associated climate change. The use of new techniques, such as micromorphology, OSL dating, and GIS-based modeling tools, is beginning to transform our understanding of the chronology of many colluvial sequences, as well as our ability to relate these deposits to human-land interactions through time.
- Allen, M. J., 1988. Archaeological and environmental aspects of colluviation in south-east England. In Groenman-van Waateringe, W., and Robinson, M. (eds.), Man-Made Soils. Oxford: British Archaeological Reports. International Series, Vol. 410, pp. 67–92.Google Scholar
- Allen, M. J., 1992. Products of erosion and the prehistoric land use of the Wessex Chalk. In Bell, M. G., and Boardman, J. (eds.), Past and Present Soil Erosion. Oxford: Oxbow. Oxbow Monograph, Vol. 22, pp. 37–52.Google Scholar
- Allen, M. J., and Macphail, R. I., 1987. Micromorphology and magnetic susceptibility studies: their combined role in interpreting archaeological soils and sediments. In Fedoroff, N., Bresson, L.-M., and Courty, M.-A. (eds.), Micromorphologie des sols/Soil Micromorphology. Plaisir: Association Française pour l’Etude du Sol, pp. 669–676.Google Scholar
- Bailiff, I. K., French, C., and Scarre, C. 2014. Application of luminescence dating and geomorphological analysis to the study of landscape evolution, settlement and climate change on the Channel island of Herm. Journal of Archaeological Science, 41, 890–903.Google Scholar
- Bell, M. G., 1990. Brean Down Excavations 1983–1987. London: English Heritage. Historic Buildings and Monuments Commission Archaeological Report, Vol. 15.Google Scholar
- Bell, M. G., 1992. The prehistory of soil erosion. In Bell, M. G., and Boardman, J. (eds.), Past and Present Soil Erosion. Oxford: Oxbow. Oxbow Monograph, Vol. 22, pp. 21–36.Google Scholar
- Bell, M. G., and Boardman, J. (eds.), 1992. Past and Present Soil Erosion. Oxford: Oxbow. Oxbow Monograph, Vol. 22.Google Scholar
- Bewley, R. H., Crutchley, S. P., and Shell, C. A., 2005. New light on an ancient landscape: lidar survey in the Stonehenge World Heritage Site. Antiquity, 79(305), 636–647.Google Scholar
- Boardman, J., 1992. Current erosion on the South Downs: implications for the past. In Bell, M. G., and Boardman, J. (eds.), Past and Present Soil Erosion. Oxford: Oxbow. Oxbow Monograph, Vol. 22, pp. 9–19.Google Scholar
- Bolt, A. J. J., Mücher, H. J., Sevink, J., and Verstraten, J. M., 1980. A study on loess-derived colluvia in southern Limbourg (The Netherlands). Netherlands Journal of Agricultural Science, 28(2), 110–126.Google Scholar
- Burton, R. G. O., 1981. Soils in Cambridgeshire II. Harpenden: Soil Survey. Soil Survey Record No. 69.Google Scholar
- Canti, M. G., 1995. A mixed-method approach to geoarchaeological analysis. In Barham, A. J., and Macphail, R. I. (eds.), Archaeological Sediments and Soils: Analysis, Interpretation and Management. London: Institute of Archaeology, University College, London, pp. 183–190.Google Scholar
- Carver, M. O. H., 1998. Sutton Hoo: Burial Ground of Kings? London: British Museum Press.Google Scholar
- Castro, P. V., Chapman, R. W., Gili, S., Lull, V., Micó, R., Rihuete, C., Risch, R., and Sanahuja, M. E., 1999. Agricultural production and social change in the Bronze Age of southeast Spain: the Gatas Project. Antiquity, 73(282), 846–856.Google Scholar
- Clark, A. J., 2000. Seeing Beneath the Soil: Prospecting Methods in Archaeology, 2nd edn. London: Routledge.Google Scholar
- Courty, M.-A., Goldberg, P., and Macphail, R. I., 1989. Soils and Micromorphology in Archaeology. Cambridge: Cambridge University Press.Google Scholar
- Dalrymple, J. B., Blong, R. J., and Conacher, A. J., 1968. A hypothetical nine unit landsurface model. Zeitschrift für Geomorphologie, 12, 60–76.Google Scholar
- Donoghue, D. N. M., 2001. Remote sensing. In Brothwell, D. R., and Pollard, A. M. (eds.), Handbook of Archaeological Sciences. Chichester: Wiley, pp. 555–563.Google Scholar
- English Heritage, 2004. Geoarchaeology: Using Earth Sciences to Understand the Archaeological Record. Swindon: English Heritage.Google Scholar
- Evans, J. G., 1972. Land Snails in Archaeology. London: Seminar Press.Google Scholar
- Evans, R., 1992. Erosion in England and Wales – the present the key to the past. In Bell, M. G., and Boardman, J. (eds.), Past and Present Soil Erosion. Oxford: Oxbow. Oxbow Monograph, Vol. 22, pp. 53–66.Google Scholar
- French, C. A. I., 2003. Geoarchaeology in Action: Studies in Soil Micromorphology and Landscape Evolution. London: Routledge.Google Scholar
- French, C. A. I., 2005. Micromorphological studies. In Carver, M. O. H., and Sutton, H. (eds.), A Seventh-Century Princely Burial Ground and Its Context. London: The British Museum Press, pp. 384–389. 363–379.Google Scholar
- French, C. A. I., Passmore, D., and Schulte, L., 1998. Geomorphology and edaphic factors. In Castro, P. V., Chapman, R. W., Gili, S., Lull, V., Micó, R., Rihuete, C., Risch, R., and Sanahuja, M. E. (eds.), Aguas Project: Palaeoclimatic Reconstruction and the Dynamics of Human Settlement and Land Use in the Area of the Middle Aguas (Almería), in the South-East of the Iberian Peninsula. Luxembourg: European Commission, pp. 45–52.Google Scholar
- French, C. A. I., Lewis, H., Allen, M. J., Green, M., Scaife, R., and Gardiner, J., 2007. Prehistoric Landscape Development and Human Impact in the Upper Allen Valley, Cranborne Chase, Dorset. Cambridge: McDonald Institute for Archaeological Research.Google Scholar
- Gaffney, C. F., and Gator, J., 2003. Revealing the Buried Past: Geophysics for Archaeologists. Stroud: Tempus.Google Scholar
- Goldberg, P., and Macphail, R. I., 2006. Practical and Theoretical Geoarchaeology. Oxford: Blackwell.Google Scholar
- Grün, R., 2001. Trapped charge dating (ESR, TL, OSL). In Brothwell, D. R., and Pollard, A. M. (eds.), Handbook of Archaeological Sciences. Chichester: Wiley, pp. 47–62.Google Scholar
- Imeson, A. C., Kwaad, F. J. P. M., and Mücher, H. J., 1980. Hillslope processes and deposits in forested areas of Luxembourg. In Cullingford, R. A., Davidson, D. A., and Lewin, J. (eds.), Timescales in Geomorphology. Chichester: Wiley, pp. 31–42.Google Scholar
- Kirkby, M. J., 1969. Erosion by water on hillslopes. In Chorley, R. J. (ed.), Water, Earth and Man. London: Methuen, pp. 229–238.Google Scholar
- Kwan, M.-P., and Lee, J., 2004. Geovisualization of human activity patterns using 3D GIS: a time-geographic approach. In Goodchild, M. F., and Janelle, D. G. (eds.), Spatially Integrated Social Science. Oxford: Oxford University Press, pp. 48–66.Google Scholar
- Limbrey, S., 1975. Soil Science and Archaeology. London: Academic.Google Scholar
- Macphail, R. I., 1992. Soil micromorphological evidence of ancient soil erosion. In Bell, M. G., and Boardman, J. (eds.), Past and Present Soil Erosion. Oxford: Oxbow. Oxbow Monograph, Vol. 22, pp. 197–215.Google Scholar
- Morgan, R. P. C., 1979. Soil Erosion. London: Longman.Google Scholar
- Moseley, M. P., 1973. Rainsplash and the convexity of badland divides. Zeitschrift für Geomorphologie Supplementband, 18, 10–25.Google Scholar
- Mücher, H. J., 1974. Micromorphology of slope deposits: the necessity of a classification. In Rutherford, G. K. (ed.), Soil Microscopy. Kingston: The Limestone Press, pp. 553–566.Google Scholar
- Mücher, H. J., and Morozova, T. D., 1983. The application of soil micromorphology in Quaternary geology and geomorphology. In Bullock, P., and Murphy, C. P. (eds.), Soil Micromorphology. Berkhamsted: A. B. Academic Publishers, pp. 151–194.Google Scholar
- Roose, E. J., 1970. Importance relative de l’érosion, du drainage oblique et vertical dans la pédogenèse actuelle d’un sol ferrallitique de moyenne Côte d’Ivoire. Cahiers ORSTROM, série Pedologie, 8(4), 469–482.Google Scholar
- Scaife, R. G., 1984. Gallibury Down, Isle of Wight – Pollen Analysis of a Bronze Age Downland Palaeosol. London: English Heritage. Ancient Monuments Laboratory Report, Vol. 4240.Google Scholar
- Selby, M. J., 1993. Hillslope Materials and Processes, 2nd edn. Oxford: Oxford University Press.Google Scholar
- Statham, I., 1979. Earth Surface Sediment Transport. Oxford: Clarendon.Google Scholar
- Statham, I., 1990. Slope processes. In Goudie, A. (ed.), Geomorphological Techniques, 2nd edn. London: Unwin Hyman, pp. 225–259.Google Scholar
- Taylor, R. E., 2001. Radiocarbon dating. In Brothwell, D. R., and Pollard, A. M. (eds.), Handbook of Archaeological Sciences. Chichester: Wiley, pp. 23–34.Google Scholar
- Thornes, J. B., and Gilman, P., 1983. Potential and actual erosion around archaeological sites in southeast Spain. In de Ploey, J. (ed.), Rainfall Simulation, Runoff and Soil Erosion. Cremlingen: Catena Verlag. Catena Supplement, Vol. 4, pp. 91–113.Google Scholar
- Waters, M. R., 1992. Principles of Geoarchaeology: A North American Perspective. Tuscon: The University of Arizona Press.Google Scholar
- Wilson, D. R., 2000. Air Photo Interpretation for Archaeologists. Stroud: Tempus.Google Scholar
- Wischmeier, W. H., and Smith, D. D., 1978. Predicting Rainfall Erosion Losses: A Guide to Conservation Planning. Washington, DC: US Department of Agriculture. Agricultural Handbook, Vol. 537.Google Scholar
- Wischmeier, W. H., Johnson, C. B., and Cross, B. V., 1971. A soil erodibility nomograph for farmland and construction sites. Journal of Soil and Water Conservation, 26(5), 189–193.Google Scholar