Download as pdf or txt
Download as pdf or txt
You are on page 1of 13

Journal of Archaeological Science 38 (2011) 1135e1147

Contents lists available at ScienceDirect

Journal of Archaeological Science


journal homepage: http://www.elsevier.com/locate/jas

Degradation of mud brick houses in an arid environment: a geoarchaeological


model
David Friesem a, b, Elisabetta Boaretto a, b, Adi Eliyahu-Behar b, Ruth Shahack-Gross a, b, *
a
Dept. of Land of Israel Studies and Archaeology, Bar-Ilan University, Ramat-Gan 52900, Israel
b
Kimmel Center for Archaeological Science, Weizmann Institute of Science, Rehovot 76100, Israel

a r t i c l e i n f o a b s t r a c t

Article history: A common assumption in Near Eastern tell archaeology is that the majority of sediments originate from
Received 4 November 2010 degraded mud bricks. Little is known about the mechanism of mud brick wall degradation. Here we
Received in revised form present a detailed macro- and microscopic ethnoarchaeological study of the degradation of a mud brick
19 December 2010
house and propose a comprehensive mechanism for tell formation processes in arid environments. The
Accepted 24 December 2010
study took place in southern Israel by trenching a ca. 60 year old abandoned mud brick house, followed
by extensive sediment sampling. Macroscopic observations showed that mud brick walls degrade by
Keywords:
collapse of single bricks and/or collapse of intact wall parts, either inwards or outwards. In addition, infill
Mud bricks
Site formation processes
sediments within the house and outside it, in close proximity to its walls, form alternating sedimentary
Degradation processes layers of various colors and textures. The degraded mud brick material lost its distinctive macroscopic
Fourier-transform infrared (FTIR) structure, which makes it impossible to accurately identify this material by field observations alone.
spectroscopy Mineralogical and elemental analyses established the sources of the house infill sediments, namely mud
X-ray Fluorescence (XRF) bricks and wind blown sediments. Alternating layers mostly originate from mixing between degraded
Micromorphology mud brick material and wind blown sediments. Micromorphological observations revealed microscopic
mechanisms of mud brick degradation and include processes of mud slurry gravity flows, sediment
coatings and infillings, wind abrasion of walls, small-scale puddling, and bioturbation. This study
provides a working scheme for site formation of abandoned mud brick structures in arid environments. It
provides a set of criteria by which it is possible to differentiate floors from post-abandonment sedi-
mentary features and thus improves the reliability of activity area research.
Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction a result of mud brick degradation (e.g., Davidson, 1973; Goldberg and
Macphail, 2006; Rosen, 1986) and the term “mud brick material” is
The study of human made mud bricks offers a wealth of archae- often used by field archaeologists to point out sediments that are
ological information. It contributes to our understanding of site assumed to be the degradation product of mud bricks. Field identi-
architecture, social issues related to buildings/houses, human selec- fication of this so-called “mud brick material” is based on intuition
tion and procurement of soil raw material, site formation, site and lacks standardization and definition. Criteria for the identifica-
stratigraphy, the influence of these activities on the landscape, and tion of decayed mud bricks therefore vary among excavated sites, and
post-abandonment processes (Goldberg, 1979; Goodman-Elgar, even among excavation areas at the same site. The study presented
2008; Matthews et al., 1997; Rosen, 1986; Shahack-Gross et al., below therefore aims at providing criteria for the identification of
2005; Stevanovic, 1997). Extracting archaeologically this wealth of mud brick degradation products, using macroscopic (field) and
information is hampered by the fact that unfired mud bricks are microscopic (laboratory based) characteristics. This was achieved by
prone to degradation. At historical tell sites in the southern Levant, for examining the degradation of the walls of an abandoned mud brick
example, mud brick walls are rarely preserved and architectural house in an arid environment, and studying the accumulation of infill
reconstructions are often based on presence of stone foundations sediments within and around the house.
only. In Near Eastern tells it is often assumed that most sediments are
1.1. Previous research

* Corresponding author. Dept. of Land of Israel Studies and Archaeology, Bar-Ilan


The most common construction with earthen materials in the
University, Ramat-Gan 52900, Israel. Near East is by using a box-shaped unit made of soil which is called
E-mail address: ruth.shahack@weizmann.ac.il (R. Shahack-Gross). a mud brick. While studies on the taphonomy of mud bricks are

0305-4403/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jas.2010.12.011
1136 D. Friesem et al. / Journal of Archaeological Science 38 (2011) 1135e1147

scant (see below), relatively many studies were conducted on the the Gvulot mud house e in order to gain more insights into the
preparation of bricks, building techniques, and sourcing of mud processes involved in mud brick degradation, the resulting “mud-
bricks at archaeological sites. brick material”, and the mixing of degraded brick material with
Construction with mud bricks in the Near East dates back to the regional sediments and the house floor. Overall we present a model
early Pre-Pottery Neolithic period, where in sites such as Netiv of mud brick house degradation, from walls to infill sedimentation.
Hagdud, Jericho and Gilgal (all located in the Jordan Valley) walls The study area is within an arid environment as such environments
were built of plano-convex mud bricks (Bar-Yosef, 1995). Ethno- prevail throughout the Near East. Future research will focus on
graphic and archaeological studies show that raw material for mud wetter environments.
brick preparation is mostly acquired from local soil exposures
around settlements (Canaan, 1932e33; Dalman, 1928e42; 1.2. The study area
Mcintosh, 1974). Once collected, the raw material is manipulated
by sorting, sieving, and mixing with materials such as sand, straw, The mud house chosen for this study is located near Kibbutz
dung or gravel in order to improve the mechanical strength of the Gvulot in the northern part of the western Negev, Israel (the Besor
mud brick (Rosen, 1986; Torraca et al., 1972). The resulting sedi- area; Fig. 1). The climate in this area is semi-arid, with an average of
ment is then mixed with water and put into a wooden mold which 150e200 mm of rain per year (Bitan-Buttenwieser, 1967). The rainy
is lifted and the wet brick is left to dry (Canaan, 1932e33; Koulidou, season is short (usually October to March) and rain showers are
1998; Spencer, 1979). usually brief but strong. The topography slopes gently from west to
Sun-dried mud bricks inevitably degrade because the earthen east at 60e160 m above sea level.
material of which they are composed reacts with the environment. The Gvulot mud house was built by Bedouins (nomadic pasto-
Previous research points out that the major agent in the degrada- ralists), who settled in the region during the 19th and early 20th
tion process of mud bricks is water from rain and/or rising damp centuries, under the Ottoman and the British mandatory regimes. In
and it introduces soluble salts into the earthen material which the first half of the 19th century large populations of Egyptian peas-
cause swelling and splitting of bricks upon dehydration (Carter and ants and Bedouins migrated from the Sinai Peninsula to the Besor
Pagliero, 1966; Mcintosh, 1974; Torraca et al., 1972). The degrada- region encouraged by the economical relationships between the
tion process of mud brick walls is also affected by wind erosion agricultural farms of the area with the city of Gaza. During the 19th
(Carter and Pagliero, 1966), biological activities such as penetration century the tribal conflicts in the area diminished under the Ottoman
of plant roots, bird nesting within mud brick walls, and burrowing regime contributing to stability of agricultural farms. The Ottoman
by rodents and insects (Goodman-Elgar, 2008; Torraca et al., 1972). Sultans encouraged the nomadic Bedouins to settle down, gave them
After abandonment, the degraded brick material accumulates on rights to cultivate the land and helped in designing their new
the regional sediment outside the house, or on the floor inside the settlements (Gazit, 2000). Due to the encouraged sedentarization of
house. This accumulated brick material usually cannot be distin- the Bedouins, their farm houses had an architectural plan similar to
guished by the naked eye from the regional sediments. In archae- that of the black tent. These houses included a living section
ology, this results in problems of field interpretation and inability to composed of two rooms, one for the men and the other for the
trace the contours of long-degraded mud walls. women (Gazit, 1986). They were built from mud bricks, did not have
Pioneering ethnographic observations on mud wall decay and an any windows, and the roof was flat and made from a braid of vegetal
excavation of modern decaying mud structures in West Africa were stems (e.g., dry reeds, palm leaves, thorn bushes, or cornstalks) and
conducted by McIntosh in the 1970s (1974; 1977). McIntosh (1974) wooden beams. This vegetal braid was covered by mud mixed with
observed that rain splash and capillary movement of salt-contain- straw (Canaan, 1932e33; Gazit, 1986). In front of the rooms a rect-
ing water at wall bases caused undercutting of the lower part of the angular yard was built surrounded by low walls of bricks or stone
mud walls. This process lead to outward wall collapse. In addition, (Gazit, 1986).
he observed that degraded brick material accumulated on both sides The mud house chosen for this study was identified by local
of the degrading walls sometimes forming thin films of sediment. archaeologist Dan Gazit, who also attributed it to the early 20th
Most other studies of mud bricks focused on particle-size analysis century based on the geographic history of the region, the archi-
(i.e., granulometry) in order to investigate the properties of various tecture of the house and the abundance of blackegray Gaza Ware
types of sediments, their composition, selection of raw material, and sherds typical of the local Bedouin material culture at the 19th and
sourcing of the mud brick material (Davidson, 1973; Emery and early 20th centuries (Gazit, 1986). The Gvulot mud house, like many
Morgenstein, 2007; Goldberg, 1979; Kemp, 2000; Nodarou et al., other Bedouin houses in the region, was abandoned in 1948 as the
2008). Besor area suffered from the outcomes of war between Egypt and
Only two geoarchaeological studies touch upon the issue of mud Israel. Since the early 1950s the region has been used by the Israeli
brick degradation. Koulidou (1998) excavated two abandoned army for training. The studied mud house was left to decay, with
modern mud structures in Northern Greece in order to investigate minor human activity in its vicinity.
the depositional patterns of mud wall degradation using particle-
size analysis. She noted a reduction of the grain sizes within the fill 2. Materials and methods
sediments toward the center of the studied room (Koulidou, 1998).
The second study was conducted by Goodman-Elgar (2008) who 2.1. Fieldwork strategy
examined the degradation of abandoned earthen dwellings in
Bolivia through micromorphological analysis. Goodman-Elgar’s The research included a small-scale excavation in the aban-
micromorphological work enabled her to identify accumulation of doned Gvulot mud house.
organic matter, burnt sediment, building material and to identify The house was mapped using a hand-held GPS (Garmin eTrex H)
post-depositional processes such as bioturbation (Goodman-Elgar, by UTM coordinates and height above sea level, accompanied by
2008). The studies by Koulidou (1998) and Goodman-Elgar (2008) measuring each feature noticeable in the field using a tape. The
were pioneering, but used only one technique each (particle-size study area was divided into a 1  1 m grid and two 1 m wide parallel
analysis and micromorphology, respectively). trenches were opened, traversing the house along its width (on
Here we present geoarchaeological research that was conducted a northesouth axis, Fig. 2a). Trench H traversed the roofed area
at a pre-modern abandoned mud house e hereafter referred to as while Trench K traversed the open courtyard area. The trenches
D. Friesem et al. / Journal of Archaeological Science 38 (2011) 1135e1147 1137

Fig. 1. Map showing the general location of Gvulot in southern Israel, and a photograph showing the mud house before excavation (April 2008). Length of wall in the background is
ca. 8 m.

were dug to a depth of a few centimeters below the presumed micromorphological analysis were sampled selectively based on
house floor. contextual considerations.
Continuous systematic sediment sampling was conducted, in
which samples were taken from a column, one above the other.
2.2. Laboratory techniques
Selective sampling was employed on specific features of interest
when a specific question arose in the field (Courty et al., 1989). The
Bulk samples were analyzed using FTIR spectroscopy and X-Ray
samples were collected according to sediment color, texture and
Fluorescence (XRF) spectrometry, the former in order to identify
hardness. Control sediments (i.e., regional sediments) were sampled
organic and mineral components and the latter in order to search
outside the site. The local soil was sampled in a pit excavated ca.
for patterns using chemical elements. Representative FTIR spectra
50 m east of the site. Wind blown sand dune sediments were
were obtained from all samples (n ¼ 112) by grinding a few tens of
sampled in excavated pits ca. 200 m south and ca. 500 m west of
micrograms of sample using an agate mortar and pestle. About
the site. Bulk samples were registered with running numbers and
0.1 mg or less of the sample was mixed with about 80 mg of KBr (IR-
put in plastic vials using a metal spoon. Undisturbed blocks for
grade). A 7 mm pellet was then made using a hand press. The
spectra were collected between 4000 and 250 cm1 at 4 cm1
resolution using a Thermo Nicolet 380 spectrometer and inter-
preted using an internal library of infrared spectra of archaeological
materials (Weiner, 2010).
Quantitative bulk chemical compositions were obtained by Ene-
rgy Dispersive X-Ray Fluorescence (ED-XRF), using a Spectro-XEPOS
bench-top instrument, with a Palladium (Pd) anode equipped with
a series of secondary targets. Evaluation of results was done using the
XEPOS method for powder materials. Fifty-eight representative
sediment samples were analyzed. Approximately 3 g of sediment
were lightly homogenized using a mortar and pestle. The measure-
ments were taken in an air environment. Under these conditions,
elements with an atomic mass lower than aluminum could not be
measured. Results are presented as percentage by weight (wt%) of
the element’s oxide form. Measurements were conducted in batches
of 12, from which one sample was a Certified Reference Material
(CRM): GSS-1; Geological Soil Sample (also called GBW-07401)
produced by the Institute of Geophysical and Geochemical Explora-
tion (IGGE), Ministry of Geology, and the Institute of Rock and
Mineral Analysis (IRMA) in China. The elemental nominal values of
the GSS-1 standard were compared to the results obtained from the
standard in each batch of measurements conducted in this study. In
order to test the reproducibility and to estimate the error of the
measurements, 2 archaeological samples as well as the standard
(GSS-1) were measured in triplicates. Upon analysis of the results, it
Fig. 2. Drawing of the excavated mud house. (a) Above ground features before exca- became clear that only 4 of the measured elements were informative
vation, showing the placement of trenches, one within the roofed area (western part of
the structure) and one within the courtyard area (eastern part of the structure).
for this study (Al, Si, Ca, Fe e see Results section). The reproducibility
(b) Above and below ground features following the excavation, showing the depth of of the measurements for these 4 informative elements is better than
walls, erecting the fallen walls, and extrapolating the extent of the house floor. 10% (Table 1).
1138 D. Friesem et al. / Journal of Archaeological Science 38 (2011) 1135e1147

Table 1
XRF results of the standard, triplicate and sedimentological groups studied. Values and averages are reported in wt%. Coefficient of variation (CV) is calculated as the percentage
of the standard deviation from the average value. Note that the analytical sum is always lower than 100% due to presence of unmeasured light elements. The results are thus
reported without correction or normalization.

Sample Al2O3 SiO2 CaO Fe2O3 Sum of analysis


GSS-1 nominal values 14.18 62.6 1.72 6.6
GSS-1 measured values (n ¼ 6) Average 17.56 63.14 2.04 5.68 92.86
Std 0.81 2.45 0.01 0.02 3.31
CV 4.59 3.89 0.50 0.35 3.56
Archaeological sediments triplicates Mud brick Average 9.90 39.07 17.02 4.16 73.20
Std 0.10 1.64 0.99 0.24 0.73
CV 1.03 4.19 5.82 5.72 1.00
Infill sediment Average 9.35 59.80 5.10 1.65 77.98
Std 0.15 2.26 0.48 0.14 1.62
CV 1.65 3.78 9.36 8.67 2.08
Control e Wind blown sediment (n ¼ 5) Average 9.09 59.17 5.43 1.59 77.32
Std 0.49 7.35 1.65 0.39 5.24
CV 5.35 12.41 30.37 24.67 6.78
Control e Husmas soil (n ¼ 1) 7.71 30.96 17.20 3.34 61.26
Control e Nahal HaBesor sediment (n ¼ 1) 5.71 19.84 14.10 4.71 46.28
Control e Mud bricks (n ¼ 6) Average 9.81 37.40 12.75 4.25 66.86
Std 0.87 3.82 1.74 0.64 3.08
CV 8.87 10.22 13.62 15.01 4.60
White layer (house floor) (n ¼ 5) Average 3.61 21.78 26.11 1.79 54.87
Std 1.02 5.66 4.33 0.23 3.20
CV 28.35 25.99 16.58 12.71 5.83
Yellow infill sediment (n ¼ 12) Average 8.99 55.15 7.06 1.70 74.97
Std 0.73 8.54 3.18 0.25 5.83
CV 8.13 15.49 45.09 14.80 7.78
Yellowegray infill sediment (n ¼ 5) Average 9.18 51.19 7.56 2.43 72.53
Std 0.65 9.06 2.87 0.81 5.72
CV 7.08 17.70 38.01 33.24 7.89
Gray infill sediment (n ¼ 8) Average 9.08 40.21 11.26 3.27 66.21
Std 0.89 8.04 3.31 0.63 6.10
CV 9.85 19.99 29.37 19.26 9.22
Brown infill sediment (n ¼ 6) Average 9.86 36.75 12.41 4.49 54.98
Std 0.56 2.56 1.63 0.44 2.13
CV 5.71 6.97 13.13 9.81 3.87
Gray infill sediment above white layer (n ¼ 6) Average 6.79 28.57 13.73 3.79 54.98
Std 0.80 1.71 1.33 0.48 2.13
CV 11.85 5.98 9.67 12.78 3.87

Since the 58 sediment samples were measured only once, an 3. Results: macroscopic field observations
average and a standard deviation could not be obtained per sample,
but as samples generally belong to various sedimentological groups 3.1. House location, shape and size
(see results section below), such values can be calculated per sedi-
mentological group (i.e., average and standard deviation of each The Gvulot mud house is located at 36-640-135 E/34-510-049 N
sedimentological group). The standard deviations calculated per in UTM coordinates, 2.5 km south of Kibutz Gvulot, and 147 m above
sedimentological group are generally larger compared to the stan- sea level. The house has a rectangular shape, oriented on a north-
dard deviations of the experimental triplicates, and those between westesoutheast axis. For simplicity, cardinal directions will be used
the nominal and measured values of the standard, reflecting the throughout the text, whereas the longer walls (forming the length of
natural heterogeneity of the sediments (Table 1). Note that the the rectangle) are the north and south ones, and the shorter walls
analytical sum of measured elements in the various sedimentolog- (forming the width of the rectangle) are the east and west ones.
ical groups is less than 100%, but that the standard deviation of the The house includes two clear walled spaces. Based on Gazit (1986;
sum of analyses per sedimentological group is low (Table 1). These personal communication, 2009) and on the height of preserved
observations indicate that the low values of sum of analysis do not walls, the larger space (ca. 9  8.5 m) on the east is interpreted as
reflect the quality of the analyses but the presence of light elements a courtyard, and the smaller space (ca. 5  8.5 m) on the west is
in the sediments. Results are thus reported without correction or interpreted as the roofed living quarters. No doorways have been
normalization. identified. According to Gazit (1986) the roofed area should include
Undisturbed monolithic sediment blocks were prepared for mic- two separate rooms, however the excavation in Trench H did not cut
romorphological analysis following conventional procedures (Courty through a separating brick wall. It is possible that in the studied
et al., 1989). The blocks were dried in an oven at 50  C for three days house the division between men and women quarters was made by
and then impregnated using a 9:1 mixture of polyester resin and a screen such as cloth or vegetal material, which is not expected to
acetone. Pre-cut sample slices were prepared to 30 mm thickness, leave a visible trace after degradation. It is also possible that a thin
2  3 inch, thin sections by a commercial company (Quality Thin mud-brick wall existed, not traversed by the excavation if it parallels
Sections, Tucson, Arizona). The prepared thin sections (n ¼ 19) were one of the baulks that were left standing during the excavation.
studied using polarizing light microscopes (Nikon Labophot2-LOP The house and close area surrounding it is slightly elevated
and Nikon Eclipse 50iPOL) at various magnifications (20, 40,100, relative to the general surroundings, forming a minor tell. More
200 and 400). Micromorphological descriptions employ the sediment accumulated inside the walled area than outside the house,
terminology of Bullock et al. (1985) and Stoops (2003). i.e., the house serves as a “dust trap”. In addition, more sediment
D. Friesem et al. / Journal of Archaeological Science 38 (2011) 1135e1147 1139

accumulated in the roofed area relative to the courtyard, indicating


that more wall material was present in the roofed area.

3.2. Brick description and wall preservation

The mud bricks in the site are ca. 33  19  17 cm. They are
composed of brown (Munsell dry: 10YR 5/3e6/3) hard sediment,
including visible straw fragments and/or straw voids. The two
trenches opened during field work cut through and next to the
southern and northern walls at four points, revealing the depth to
which the walls were built. This depth accords with a horizontal
calcareous layer interpreted as the house floor, which further
allows extrapolating the height of walls below surface around the
house also in the unexcavated parts (Fig. 2b).
The preservation of the mud brick walls varies around the site.
Where some of the walls preserved to a considerable height (close to
2 m above surface), others had degraded so much that nothing is
visible on the surface except for a minor elevated sediment heap.
Wall disintegration seems to occur by two mechanisms: the first is
the disintegration of single bricks (Fig. 3a) and the second is collapse
of intact wall parts either outwards (Fig. 3b) or inwards (Fig. 3c). A
third, microscopic, degradation mechanism was observed within
the house infill sediments (see details below).

3.3. Regional sediments and soils

Three different types of regional sediments were identified and


sampled in the vicinity of the site:

A) Wind blown sediments: Yellow (Munsell dry: 10YR 7/4e7/6)


loose dune sand, covering the landscape surface and reaching
a depth of ca. 30 cm below surface. These sandy sediments
originate from stabilized Pleistocene dunes that are covered by
eolian shifting sands (Dan et al., 2007).
B) Husmas soil: Husmas is a local name for a red sandy soil with
calcareous concentrations (Calcic Rhodoxeralf) formed on
either calcareous eolianite sandstones (locally known as Kur-
kar) or on sand dunes, most frequently found in the coastal plain
of southern Israel (Singer, 2007). This paleosol was identified in
the study area in a test pit excavated 15 m east of the site. It is
a brown (Munsell dry: 10YR 7/3) hard loam with white calcar-
eous nodules whose concentration increases with depth.
C) Alluvial sediments: The closest alluvial channel to the study
area is Nahal HaBesor. It is a perennial stream that currently
drains sewage. About 6 km northeast of the site the stream is Fig. 3. Macroscopic mechanisms of mud brick wall degradation in the study area.
closest to the abandoned house. HaBesor riverbed sediment is (a) Disintegration and collapse of single bricks. (b) Collapse of intact wall parts
outwards. (c) Collapse of intact wall parts inwards, and their subsequent burial
brown (Munsell dry: 10YR 4/3) loam.
under infill sediments. Scale bar ¼ 20 cm.

3.4. House infill sediments They include artifacts that are assigned to the Israeli army training in
the area after 1948 (e.g., metal cans and barrels, plastic bags, etc.).
The excavation in both trenches revealed 4 main types of sedi- The infill sediment types are (Fig. 4):
ments (Fig. 4). The lowermost sediment in both trenches is white
(Munsell dry: 10YR 8/1e8/4) and its upper surface as exposed in the A) Yellow sandy loose sediment, similar to the local wind blown
excavated trenches is semi-horizontal. The local bedrock and soil do sediments. This sediment is found in the house infill as patches
not exhibit these characteristics. Therefore this sediment is either and layers between other sediment types in the excavated
local calcareous bedrock that we did not encounter outside the site areas and on the surface.
that was cut flat for the purposes of the house building, or non-local B) Gray (Munsell dry: 10YR 7/1e7/3) soft and hard, loamy sedi-
sediment that was brought to the site for construction purposes. In ments. These sediments accumulated inside the house as thin
any case, because its color and texture differ markedly from the laminae. Despite their gray color, no ash was observed associ-
regional wind blown sediment and because Bedouin artifacts (metal ated with these infill sediments.
objects as well as Gaza Ware pottery) were found lying just on top of C) Brown hard loamy sediment (Munsell dry: 10YR 5/3e6/3),
it, this white surface is interpreted as the house’s floor. mostly adjacent to the inner part of the walls. This sediment is
Three main types of infill sediments were identified above the similar macroscopically (based on color and compactness) to
white floor. The infill sediments are bedded and alternate randomly. the house’s mud bricks. During the excavation several bricks
1140 D. Friesem et al. / Journal of Archaeological Science 38 (2011) 1135e1147

Fig. 4. Photograph showing the main sedimentological groups at Gvulot mud house.
(a) White horizontal homogenous surface interpreted as the house floor. (b) Massive
brown. (c) Yellow soft sandy sediment interpreted as originating from wind blown
dust and (d) Gray laminated sediments, interpreted as originating from decayed mud
bricks. Scale bar ¼ 20 cm. (For interpretation of the references to color in this figure
legend, the reader is referred to the web version of this article.)

were exposed far from the walls after they fell and rolled
toward the center of the house.

Various shades of gray have been identified in the field and


described as yellowegray, grayebrown, and black sediments.

3.5. General infill sedimentation patterns Fig. 5. Plots showing sediment accumulation patterns within the trench that traverses
the roofed area. (a) Thickness of sediment types along three columns in the trench, one
The thickness of the three main types of infill sediments was close to the southern wall, one close to the north wall, and one in between the other
two in the center of the roofed area. Note that the total thickness of all sediments is
measured in the trench in the roofed area. Thicknesses were measured
similar in the three columns, forming a semi-horizontal surface. (b) Cumulative
along three columns arbitrarily chosen on the eastern section of the thickness in the three columns according to sedimentological groups showing that
trench, one close to the southern wall, one in the middle of the roofed brown sediments are thicker next to walls and thinnest in the central area while the
space and one close to the northern wall. All measurements were gray sediments are thickest in the center and thinner next to walls. At the same time,
the yellow sediments show a rather uniform thickness along the studied section. (For
conducted relative to the topmost level of the white layer (Fig. 5a).
interpretation of the references to color in this figure legend, the reader is referred to
The cumulative thickness of each infill sediment type, above the the web version of this article.)
floor level is presented in Fig. 5b. The yellow wind blown sediment
does not change significantly across the three columns, with a range
of 25e29 cm, indicating an even distribution across the space. An (i.e., control samples) include soft yellow (Munsell dry: 10YR
inverse relationship is found for the gray and brown infill sediments. 7/4e7/6) wind blown sandy sediment. The major mineral compo-
While near the walls the thickness of the gray infill sediment is low, nent in this sediment is quartz, followed by clay and small amounts
the thickness of the brown infill sediment is highest. of calcite (Fig. 6a). Compacted yellow sandy sediments found a few
cm below surface have a similar composition as the soft sandy
3.6. Biological activity within the sediments sediments. Crusts on top of sandy deposits are composed mainly of
clay and quartz. Below surface, compacted dune sands include
At the macroscopic scale, arthropods (e.g., beetles, spiders, scor- white nodules and these sediments are dominated by calcite. They
pions, etc.), rodents, snakes and gazelles were observed either on site seem to correspond with the local Husmas soil, having varying
or in its vicinity during the one week of excavation. In addition to the amounts of clay and calcite according to depth (Fig. 6b). Lastly, the
fauna, plants and their roots are abundant in and around the site. alluvial sediment from Nahal HaBesor riverbed is composed mainly
Another biological activity, that had a most significant impact on the of clay and calcite, and traces of quartz.
site’s integrity, is the post-abandonment soldier activity. The soldiers Controls from within the site are in the form of the mud bricks
seem to have camped at the site, based on the refuse related to food themselves. Based on the infrared analyses, mud bricks from the
and other activities. Even though it cannot be proven, the soldiers studied house are composed primarily of clay, followed by calcite and
may have deliberately brought down parts of walls and the roof. a small amount of quartz (Fig. 6c). Note the similarity in mineral-
ogical composition between the mud brick and local Husmas.
4. Results: laboratory-based analyses of the sediments The infill sediments within the site are generally yellow, gray and
brown (see Section 3.4 above). The yellow infill sediments are
4.1. Mineralogical characterization of bulk samples via FTIR dominated by quartz (Fig. 6d) and their mineralogy resembles that
spectroscopy of the wind blown regional sediments. The brown and gray infill
sediments are composed of high amounts of clay, followed by calcite
Based on their infrared spectra, sediments in the study area and a small amount of quartz (Fig. 6e). This mineralogical compo-
are rather similar mineralogically, being composed of quartz, clay sition is similar to that of the control mud bricks (c.f., Fig. 6c) and
and calcite in various ratios. Sediments sampled around the site Husmas soil (c.f., Fig. 6b). The white sediment layer that was
D. Friesem et al. / Journal of Archaeological Science 38 (2011) 1135e1147 1141

Fig. 6. Infrared spectra of representative sediment samples in the study area. (a) Control, yellow wind blown sediment, (b) Control, Husmas local soil, (c) Control, mud brick, (d)
Yellow infill sediment, (e) brown/gray infill sediment, (f) white floor sediment. Note the similarity between spectra (a) and (d) where the sediment is dominated by quartz
(indicative absorptions at ca. 1086, 517, 798, 778 and 695 cm1); the similarity between spectra (b), (c) and (e) where the sediment is dominated by clay (indicative absorptions at ca.
1035, 913, 798, 778 and 530 cm1); and the significantly different spectrum (f) where the dominant mineral is calcite (indicative absorptions at ca. 1430, 875 and 713 cm1).

identified as the house floor is composed primarily of calcite, fol- meaningful patterns. The concentrations of most elements were
lowed by clay and trace of quartz (Fig. 6f). In comparison to the very low, thus when considering the analytical error, no clear
control sample of the Husmas soil outside the site, it has a signifi- patterns were revealed using these elements. On the other hand, 4
cantly higher concentration of calcite and rather low amounts of clay major elements e aluminum (Al), silicon (Si), calcium (Ca) and iron
minerals. No infill sediments resemble the mineralogical fingerprint (Fe) e showed significant patterns in relation to the sedimento-
of the white floor layer. The grayeyellow infill sediments show logical groups at the site (Table 1).
either high quartz or high clay components, i.e., hinting for a mixture Fig. 7a shows a significant difference in the elemental composi-
of sedimentary sources. Overall, all the sediments in the archaeo- tion of the white floor layer compared to most other sediments. As
logical context coincide with three main mineralogical groups, seen from the infrared spectra the white layer is mainly composed of
differentiated by the ratios between quartz, calcite and clay, and calcite, with a low amount of clay and quartz, which explains its
mineralogical similarities seem to indicate the source of some of relative richness in lime (CaO) and relative low amounts of alumina
these infill sediments (Table 2). (Al2O3), silica (SiO2) and iron oxide (Fe2O3); the latter relate to clay
and quartz. The composition of the major elements in the Husmas
4.2. Elemental characterization of bulk samples via X-Ray soil seems to be close to the group of dark-colored (brown and gray)
Fluorescence (XRF) spectrometry infill sediments and mud bricks. This was also observed in the
infrared spectra of these infill and control sediments. The sediment
A total of 58 control and infill sediments were measured in 6 collected from Nahal HaBesor riverbed does not seem to correspond
different batches. The results were plotted in order to search for clearly to any of the sedimentological groups, having alumina and
silica concentrations most similar to the white layer and at the same
time lime and iron oxide concentrations most similar to the dark-
Table 2
The correspondence between mineralogical groups (based on infrared spectroscopy colored infill sediments and mud bricks. This may be related to the
results) and sedimentological groups (based on macroscopic field observations) in fact that this river currently drains sewage. Alternatively, it may
the studied site. The dominant mineral in each group is marked in bold, followed by indicate that this sediment does not play a significant role in the
the other minerals in order of dominance. Gvulot house sedimentological system. Its elemental composition
Mineralogical groups Sedimentological groups will thus not be discussed further.
Quartz > clay > calcite Regional wind blown sediment The difference between the various infill sediments is clearest in
Yellow infill sediment the concentration of silica, where yellow-colored sediments have
Clay > calcite > quartz Grayeyellow infill sedimenta higher values than darker-colored sediments. This difference seems
Gray infill sediment
to correspond to the presence of higher amounts of quartz in the
Brown infill sediment
Mud brick material yellow sediments as is also reflected in their infrared spectra. The
Husmas paleosol darker-colored sediments include more clay than the yellow-
Calcite > clay > quartz White floor colored sediments, based on their infrared spectra. This explains the
a
This sedimentological group has variable mineralogy, dependent on the amount difference in concentrations of alumina and iron oxide between
of clay (more gray) and quartz (more yellow). these general sedimentological groups, with the darker-colored
1142 D. Friesem et al. / Journal of Archaeological Science 38 (2011) 1135e1147

Fig. 7. Plots showing elemental compositions for the sedimentary groups in the study area. (a) Average concentrations (wt%) of the oxides of aluminum, silicon, calcium and iron.
Error bars indicate the standard deviation where more than one sample has been analyzed from the same sedimentological group. Note that in most cases there are distinctive
similarities between Husmas soil, mud bricks and brown/gray infill sediments; between yellow control and infill sediments, and that the yellow sediments are different significantly
from the Husmas and brick sediments. Note also that the white floor sediment is distinctively different from all other sediment types. (b) The relationship between concentrations
of alumina and iron oxide, showing two distinct mixing lines that explain the variability observed in the infill sediments. Here the gray infill sediment samples are divided into gray
sediment that was sampled along the profile (n ¼ 8) and gray sediment that was sampled directly above the white floor (n ¼ 6). The latter is shown here with white asterisk inside
a gray square. (1) mixing between mud brick material and yellow wind blown sediments, resulting in gray infill sediments; (2) mixing between mud brick material and the white
floor sediment, resulting in another type of gray infill sediments. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this
article.)

sediments generally having higher concentrations of alumina and is much lower than in the sampled Husmas, implying that if the
iron oxide relative to yellow-colored sediments. The concentration Husmas was used by the builders they probably lowered the
of lime also seems to distinguish the yellow-colored from the darker- amount of lime by removing large calcitic white nodules from this
colored sediments, with the latter having more lime. This is probably soil. If true, this would have contributed to the higher relative
related to the presence of calcite (also reflected in the infrared content of silica in the mud bricks.
spectra of these sediments) and also possibly to the presence of Overall the elemental analysis supports and amplifies the field
Ca-montmorillonite in these sediments. Indeed, the highly calcitic observations and the attribution of the sediments to the different
Husmas and white layer have the highest lime concentrations. sedimentological and mineralogical groups.
Based on field observations and the infrared results, three
materials seem to be the origins of the archaeological infill sedi- 4.3. Micromorphology
ments at the site; the yellow wind blown sediment, the lower white
layer and the mud bricks. This is also reflected in the elemental The basic components observed in thin section follow the same
composition of the sediments, best shown through the relationship trends identified in the studied sediments based on the FTIR and
between the concentrations of alumina and iron oxide (Fig. 7b). The XRF analyses. Table 3 presents the overall composition and sedi-
mud bricks, the control wind blown sediments and the white floor mentological attributes (e.g., grain sizes, roundness and sorting) of
layer seem to form end-members. The mixture of these end- the studied sedimentological groups. These attributes show the
members shows graphically as mixing lines (Fig. 7b): the yellow similarity between control wind blown and yellow infill sediments,
wind blown infill sediment cannot be distinguished from the between the Husmas control and mud bricks, and between the
control regional wind blown sediment, indicating that the yellow mud bricks and brown and gray infill sediments. The white floor
infill sediments originate from wind blown dust. The brown sedi- layer differs from all infill sediments.
ment shows high similarity to the mud bricks, indicating that it is Micromorphological differences between the sedimentological
a direct product of mud-brick degradation, with virtually no mixing groups are best expressed by the void types and overall micro-
with wind blown dust. The elemental composition of the gray infill structure. Wind blown sediments are composed of a coarse fraction
sediments sampled just above the white floor is in the range without groundmass, therefore their microstructure is single-grain
between the mud bricks and the lower white layer, indicating their and voids are simple packing voids (Fig. 8a). The white floor sedi-
origin is from degrading mud bricks mixed with the white sedi- ment is massive and includes vertical desiccation cracks (Fig. 8b). It
ment, while the gray and yellowegray infill sediments in other is difficult to determine, microscopically, whether this sediment is
contexts at the site show that they originate from a mixture undisturbed calcareous bedrock or crushed and compacted calcar-
between mud bricks and yellow wind blown sediments. eous rock brought to the site on purpose. The Husmas soil includes
The Husmas soil in the site’s area has low iron oxide concen- chamber, vesicle and channel voids within a complex microstruc-
tration relative to most mud bricks, and also slightly less alumina. ture including vughy and massive microstructures (Fig. 8c). The
This indicates that this soil is either not the source of the mud voids together with snail shell fragments and thin clay coatings on
bricks, or, that it is one of the sources. Support for the second coarse grains indicate bioturbation. Overall, the Husmas soil seems
possibility is found in the micromorphological results (see below to have developed mostly from accumulation of eolian sediments,
Section 4.3.1). In addition, it is possible that the source materials for with little pedogenic processes due to the relative aridity at the
the preparation of the mud bricks have been manipulated by the study area. The presence of quartz grains larger than medium sand
builders, thus changing the original composition of the local Hus- (Table 3), i.e., grains that could not have been deposited through
mas. For example, the concentration of lime in the mud bricks wind action, might be either inherited from the parent rock below
D. Friesem et al. / Journal of Archaeological Science 38 (2011) 1135e1147 1143

Table 3
Basic micromorphological attributes in the studied sediments (based on analysis of 19 thin sections).

Coarse fraction Fine fraction Related Void types


distributiona
Minerals/Grains Grain Sizes Groundmass Birefringence
Fabric
Wind blown controls Quartz, well sorted, subangular Silt-fine sand None None Monic Simple packing
Yellow infill sediment Calcite Fine sand
Shells Fine sand
Husmas control Quartz, poorly sorted, subangular Silt-coarse sand Calcitic-clay Crystallitic Monic Vughs, vesicles
Feldspar
Hornblende
Calcite nodules, orthic and anorthicb Very fine-medium sand
Shells Medium-coarse sand
Oxides Silt-very fine sand
Mud brick (control) Quartz, poorly sorted, subangular Silt-coarse sand Calcitic-clay Crystallitic Porphyric Vughs, vesicles, planar
Feldspar
Hornblende
Flint
Brown and Gray infill Calcite, sparitic and/or micritic, Silt-coarse sand Calcitic-clay Crystallitic Porphyric Vughs, vesicles
sediments poorly sorted, subrounded
Shells Coarse sand
Oxides Silt-fine sand
Clay-rich aggregates, poorly sorted, Fine-medium sand
subrounded
White floor layer Quartz, poorly sorted, subangular Silt-medium sand Calcite Porphyric Porphyric Desiccation cracks
a
Monic related distribution: only grains of coarse fraction are present; Porphyric related distribution: the coarse fraction occurs in a dense groundmass of fine fraction
(Stoops, 2003).
b
Nodular bodies inherited from the parent material are termed anorthic, while those formed in situ are termed orthic (Stoops, 2003).

Fig. 8. Microphotographs of representative control samples in thin section, all in plane polarized light. (a) Wind blown yellow sediments. Note the near absence of fine fraction
resulting in a single-grain microstructure. (b) White floor layer (in the lower part of the photograph) showing a massive microstructure and vertical desiccation cracks. The upper
part is wind blown sediment. (c) Husmas sandy soil with complex microstructure and evidence of bioturbation. (d) Control mud brick showing features generally similar to those in
the Husmas soil, but with planar voids with/without straw temper and with rounded clay-rich granules. (For interpretation of the references to color in this figure legend, the reader
is referred to the web version of this article.)
1144 D. Friesem et al. / Journal of Archaeological Science 38 (2011) 1135e1147

the Husmas (possibly stabilized coastal dunes) or from mild sheet- fine sand to silt grains (Fig. 9b). The final stage of the process shows
wash erosion in the study area. The description and interpretation of that these crusts break and their orientations change following
the Husmas is in accord with Singer (2007). bioturbation within the infill sediments (Fig. 9d).
The basic components in mud bricks (observed in 5 different Overall, the micromorphological observations enable us to
thin sections) are similar to those in the Husmas soil (Table 3) identify the source materials of the mud bricks to a certain degree,
except for presence of clay-rich aggregates with or without micritic to evaluate the role of bioturbation at the site, and most impor-
calcite and silt to fine sand quartz grains (Fig. 8d). These clay-rich tantly, to identify the processes involved in mud brick degradation
aggregates may have originated from alluvial sediments. They and formation of the house infill sediments.
clearly do not form an integral part of the soil fabric. The void types
are similar between the Husmas and mud bricks except for the 5. Discussion
presence of planar voids in mud bricks. The elongated planar
voids resemble the shape of straw and lie in different orientations. The overall aim of this study was to understand the processes
In some of these voids the straw is present in different stages of related to mud brick degradation and formation of infill sediments
decomposition. Phytoliths, parallelepiped elongated long cells within abandoned houses in an arid environment. The results of this
indicative of grass leaf/stem origin, were observed within these study will help in developing a method for accurate identification of
voids (Fig. 8d). The microstructure is complex, including massive “mud-brick material” in archaeological sites in the Near East.
and vughy microstructures. Some of the coarse fraction grains are
clay coated, in a thicker layer than the thin clay coating in the 5.1. Sedimentary sources at Gvulot mud house
Husmas paleosol. The overall micromorphological features of the
mud bricks indicate that the bricks might have been prepared from The sediments forming the infill at Gvulot mud house were
a mixture of local Husmas with alluvial clay-rich sediments identified based primarily on mineralogical (FTIR) and elemental
(possibly from the nearby Nahal HaBesor stream) and grass straw (XRF) analyses. The use of these techniques in conjunction enables
as temper. us to not only identify source materials, but also mixtures between
The basic components in the brown and gray infill sediments are them. The major sources of the infill sediments were degrading mud
similar to those of mud bricks (Table 3). The main differences bricks and wind blown sediment. The large volume of wind blown
between the original mud brick sediment and the brown and gray sediments is probably typical of arid environments, and should be
infill sediments are: (1) the absence of the vegetal temper (straw), or considered in any geoarchaeological study of formation processes in
its indicative planar voids, (2) infilling of voids with micritic calcite, similar sites. The house floor is a third, but minor, source material, as
(3) formation of dirty calcitic-clay crusts, and (4) formation of pure indicated by the mixing lines observed following the XRF analysis
clay crusts (Fig. 9a). All these features are attributed to the degra- (Fig. 7).
dation process of the mud bricks and this micromorphological The infill sediments include the source materials and also
sedimentary facies relates to the brown and, in a more progressed mixtures between them in various ratios, thus exhibiting a range of
stage of degradation, to the gray sediments macroscopically iden- colors, hardness, texture, mineralogy and elemental compositions
tified at the site. that are wider than the range of the sources themselves. Clearly,
due to differences in mud brick recipes in different sites, the values
4.3.1. Depositional processes reported in this study should not be taken as a general rule. What
Within a general regime of wind blown infill sedimentation, bands can be used as a general rule for future studies is the mechanism of
of medium quartz sand were observed associated with the gray infill site formation processes we identified in this study.
sediment (Fig. 9b). Such large grains could not have been deposited The processes of infill include wall disintegration and collapse.
through eolian processes, indicating that small-scale changes in the Walls in this study have collapsed both outwards and inwards. This
forces of deposition took place at the site, shifting between wind and observation is in contrast to McIntosh’s (1974) observations,
possibly water or gravity flow depositional patterns. however, in this study we do not know the exact reason for the wall
collapse as no undercutting was observed, thus the possibility of
4.3.2. Post-depositional processes collapse due to earthquakes and/or post-abandonment deliberate
In all infill sediments, channel voids indicative of faunal move- human destruction are also valid. Walls also disintegrate as single
ment through the sediment cut through layers of the different bricks collapses, either inwards or outwards. Therefore, single
sedimentological groups. Fresh organic excrements (identified bricks and/or intact wall parts may be associated with house
following photographs in Stoops, 2003) were identified in voids in abandonment. These features are then buried as sediments accu-
the infill sediments. Bioturbation, as well as water and wind action mulate within the abandoned house but also around its walls on
at the site resulted in the mixing of the different sedimentological the outside, slowly forming a minor mound.
groups. One observation is that at the contact between the lower Fig. 10 presents a schematic summary of the processes that
white layer and the sediments on top of it there are areas where the operated in the formation of the microscopic features of the infill
sharp contact that characterizes this sedimentological transition is sediments at the site. Once buried, brick fragments undergo further
sometimes obscured by penetration of either the white sediment degradation by percolating water evident by the presence of silty-
upwards into the gray or yellow infill, or of the gray/yellow infill clay crusts on their edges and within voids (Fig. 9c). Another
downwards along cracks and channels into the white layer (Fig. 8b). mechanism of mud brick degradation is the direct impact of rain
The early stages of mud brick degradation were observed in water, washing the mud down the walls and into the spaces
a specific area where mud brick fragments degrade in situ, showing between the walls, forming vast silty-clay microlaminae that fill
the formation of calcitic-clay crusts around the fragments’ edges internal house spaces from wall to wall (see archaeological impli-
(Fig. 9c). The later stage in mud brick degradation is characterized cations below). The various components in the bricks e the organic
by the formation of calcitic-clay crusts, occasionally with silt to fine straw temper, the coarse mineral fraction and the fine mineral
sand grains within them, related to slow water movement in the fraction e undergo different formation pathways (Fig. 10). When
form of a slurry gravity flow. In certain locations, weak graded the surface of a brick degrades, the organic material (i.e., the vegetal
bedding hints for the presence of occasional puddles where crusts material such as the straw) is exposed and rapidly degrades. During
of pure clay accumulated above a calcitic-clayey crust that includes a rain event, the fine and coarse mineral fractions of the brick will
D. Friesem et al. / Journal of Archaeological Science 38 (2011) 1135e1147 1145

Fig. 9. Microphotographs of representative infill sediments in thin section, all in plane polarized light. (a) Brown/gray infill sediment resulting from mud brick degradation. Note the
absence of planar voids after vegetal temper, void infillings and crust formations. (b) Bands of medium sand within finer grained wind blown sediments, possibly indicating gravity
fall of quartz grains from walls following wind abrasion. (c) Initial stages of mud brick degradation around brick fragments within the infill sediments. Note the formation of silty-
clay crusts around the edges of the mud brick fragment. (d) Microlaminae of graded bedding indicating cycles of rain and/or eolian events. Note that medium and coarse sand grains
are ’floating’ on silty-clay laminae indicative of gravity mud slurry flows during rain events. In addition, note the disruption of bedding due to bioturbation. (For interpretation of the
references to color in this figure legend, the reader is referred to the web version of this article.)

move away from the wall, depending on the energy of water. If the 5.2. Infilling rate and its implications for the study of activity areas
energy is high, both coarse and fine minerals will form mud slurries
that slowly move by gravity to form thin microlaminae of mud brick In general, the abandoned house serves as a trap for both mud
material (Fig. 9b and d). With the decrease of water energy as the brick material and for dust. The result is a shallow mound (an
rain event diminishes, only the fine fraction, composed mainly of incipient tell). The total thickness including all sediment groups is
calcitic-clay and silt grains, is moved away from the wall and 70e80 cm (Fig. 5), spanning ca. 60 years that passed since the house
accumulates as a crust. In cases of slight puddling, graded bedding had been abandoned. Clearly the rate of infilling was not constant
will be formed, with the topmost bed composed of pure, oriented, as the study area is located within shifting sand dunes and expe-
clay. At this stage, the microstructure of the mud brick material riences several sand storms annually. In addition, the area experi-
does not include straw fragments and will thus not include large ences major annual differences in precipitation, and deflation
planar voids (in contrast to un-degraded brick material). processes are also common. Thus the accumulation is not consis-
Infilling processes in an arid environment such as in this study tent over time, and microlaminae cannot be used as indicators for
also include the deposition of wind blown sediments and their seasonal changes.
mixing, to various degrees, with the degraded mud brick material
(Fig. 10). This study shows that such mixing is quite extensive indi- 5.3. Archaeological implications for field archaeology
cating that “mud-brick material” in arid environments is not expec-
ted to resemble the original mud brick composition. Mineralogical, Some of the results of this study bear important implications for
elemental and micromorphological analyses are thus essential tools future archaeological excavations of mud brick structures in arid
to evaluate and understand the extent of mixing processes of various environments. The gray microlaminated infill surfaces that formed
sedimentary sources in archaeological sites. An interesting observa- directly from mixing of mud brick material and wind blown sedi-
tion is the presence of coarse sand grains together with smaller wind ments may easily be mistaken as floors during an archaeological
blown grains, suggesting that the coarse sand is in fact introduced excavation. An important observation is that these pseudo-floor
from the degrading walls, possibly as single-grains that detach from surfaces are intimately associated with walls e they are inclined
wall surfaces possibly due to wind abrasion. The last process to act on closer to the walls and horizontal in the middle of the built area,
the forming mound, occurring during and after depositional forming a U-shaped macro-stratigraphy (Koulidou, 1998 has also
episodes, is bioturbation (Fig.10). As the studied site is abandoned for observed the same pattern). Such observations between walls and
only ca. 60 years, it is possible that in an archaeological site, where floors at archaeological sites are abundant (e.g., Albert et al., 2008 on
bioturbation continued to act within the infill sediments for phytolith-rich layers and their relation to walls at Iron Age Tell Dor).
millennia, the infill sediments may be mixed to a degree that no This clearly complicates archaeological interpretation, however,
layering could be observed macroscopically. We would expect, with the use of micromorphology these pseudo-floors are readily
though, that patches of intact layers will still be detected micro- identifiable as the result of mud slurry deposition using criteria such
morphologically and will help in identifying formation processes. as the presence of graded bedding and clay crusts (Fig. 9b, c and d).
1146 D. Friesem et al. / Journal of Archaeological Science 38 (2011) 1135e1147

Fig. 10. A geoarchaeological model for the microscopic degradation processes of mud bricks and mud brick walls in an arid environment. Note that the coarse fraction, fine fraction
and organic components simultaneously undergo different degradation pathways, resulting in the formation of banded sedimentary layers across the entire internal surface of the
abandoned house.

An important outcome of this study is the observations regarding raw material for brick preparation. We note that the elemental
placement of various infill sediments in relation to walls. The sedi- composition and micromorphological characteristics of the mud
mentation pattern (Fig. 5) indicates that the areas close to the bricks are not identical to those of the Husmas paleosol, indicating
northern and southern walls have thicker accumulation of un-mixed that the paleosol was manipulated by the builders while the mud
mud brick material relative to the central part of the roofed area. bricks have been prepared.
The wind blown sediments have quite a similar thickness across the
measured section. The gray mixed sediments are thicker in the 6. Conclusion
central part of the section, indicating greater mixing in this context.
Overall, this pattern indicates that most un-mixed mud brick The study of human made mud bricks enables to understand
material is expected to be found close to degraded mud brick walls in social aspects related to buildings/houses, human selection and
an archaeological context. Following this observation, we suggest procurement of soil raw material, architecture and the influence of
that the best localities within abandoned archaeological houses for these activities on the landscape (Goldberg, 1979; Goodman-Elgar,
finding pristine microscopic activity remains should be close to 2008; Matthews et al., 1997; Rosen, 1986; Shahack-Gross et al.,
walls, and the identification of the contact between the activity floor 2005; Stevanovic, 1997). A model of the degradation and infill
and degraded mud brick on it, will improve activity area interpre- processes in an abandoned mud brick house has been presented.
tation. In addition, localities where massive wall collapse is identi- This study was conducted in an arid environment where processes
fied should too be considered preferential for microscopic activity are greatly influenced by eolian deposition. We do not expect similar
remains research due to the fast burial at these localities. processes to act in more humid environments, such as Mediterra-
When conducting activity area and household archaeology nean and temperate climates. The next phases in our study will aim
research one needs to know the exact surface of floors in order to at studying mud brick house degradation in humid environments,
evaluate whether artifacts used for activity area reconstruction are and also testing possibilities for enhancing identification of activity
placed directly on the floor surface below mud brick wall degra- areas based on microscopic and chemical remains.
dation sediments, or are placed slightly above the floor surface
within mud brick degradation products. As often happens in tell Acknowledgments
archaeology, the exact surface of floors is not easy to determine,
especially when the same soil type had been used for flooring and We are indebted to Dan Gazit without whom this project would
wall construction. This study gives micromorphological criteria to not have been carried out as smoothly and successfully as it did. We
approach the distinction between floor and mud brick degrada- would also like to thank our colleagues, especially Aren Maeir for
tion material in an arid environment and thus to better interpret contributing excavation equipment, Lior Regev for help with FTIR
archaeological activity area reconstruction. analyses and Steve Weiner for commenting on an earlier draft of
Lastly, this study shows that in contrast to various previous this manuscript. We are also grateful to those who helped in the
studies that identified the sources of sediments for mud brick excavation: Noa Lavi, Efrat Bocher, Shira Gur-Arie, Dan Cabanes,
preparation in alluvial deposits in the vicinity of the sites (e.g., Maite Cabanes and Tomer Aharon. The project was funded by
Davidson, 1973; Goldberg, 1979; Kemp, 2000; Nodarou et al., 2008; a grant from the Israel Science Foundation (Bikura track, 527/09) to
Rosen, 1986), here we showed that a local paleosol was used as the RSG and was assisted by funding from the Kuschitzky Fund at the
D. Friesem et al. / Journal of Archaeological Science 38 (2011) 1135e1147 1147

Dept. of Land of Israel Studies and Archaeology (Bar Ilan Univer- the Land of Israel, Arie Kindler Volume. Ingeborg Rennert Center for Jerusalem
Studies and Eretz-Israel Museum, Tel-Aviv, pp. 183e186.
sity), the Kimmel Center for Archaeological Science at the Weiz-
Goldberg, P., 1979. Geology of the Late Bronze Age mudbrick from Tel Lachish.
mann Institute and a European Research Council under the Tel-Aviv 6, 60e67.
European Community’s Seventh Framework Programme (FP7/ Goldberg, P., Macphail, R., 2006. Practical and Theoretical Geoarchaeology. Black-
2007-2013)/ERC grant agreement no 229418 to Israel Finkelstein well, Oxford.
Goodman-Elgar, M., 2008. The devolution of mudbrick: ethnoarchaeology of aban-
and Steve Weiner. doned earthen dwellings in the Bolivian Andes. J. Archaeol. Sci. 35, 3057e3071.
Kemp, B., 2000. Soil (including mud-brick architecture). In: Nicholson, P., Shaw, I.
(Eds.), Ancient Egyptian Materials and Technology. Cambridge University Press,
References Cambridge, pp. 78e103.
Koulidou, S., 1998. Depositional patterns in abandoned modern mud-brick struc-
Albert, R.M., Shahack-Gross, R., Cabanes, D., Gilboa, A., Lev-Yadun, S., Portillo, M., tures. Archaeology and Prehistory. University of Sheffield, Sheffield, UK.
Sharon, I., Boaretto, E., Weiner, S., 2008. Phytolith-rich layers from the Late Matthews, W., French, C.A.I., Lawrence, T., Cutler, D.F., Jones, M.K., 1997. Micro-
Bronze and Iron Ages at Tel Dor (Israel): mode of formation and archaeological stratigraphic traces of site formation processes and human activities. World
significance. J. Archaeol. Sci. 35, 57e75. Archaeol. 29, 281e308.
Bar-Yosef, O., 1995. Earliest food producers e pre pottery neolithic (8000e5500). In: Mcintosh, R.J., 1974. Archaeology and mud wall decay in a West-African village.
Levy, T.E. (Ed.), The Archaeology of Society in the Holy Land. Leicester University World Archaeol. 6, 154e171.
Press, London, pp. 190e204. Mcintosh, R.J., 1977. Excavation of mud structures e experiment from West-Africa.
Bitan-Buttenwieser, A., 1967. A Topo-climatological Research in the “Besor” Region. World Archaeol. 9, 185e199.
The Hebrew University of Jerusalem, Department of Geography, Jerusalem. Nodarou, E., Frederick, C., Hein, A., 2008. Another (mud)brick in the wall: scientific
Bullock, P., Fedoroff, N., Jongerius, A., Stoops, G., Tursina, T., 1985. Handbook for Soil analysis of Bronze Age earthen construction materials from East Crete.
Thin Section Description. Waine Research Publications, Wolverhampton. J. Archaeol. Sci. 35, 2997e3015.
Canaan, T., 1932e33. The Palestinian Arab house: its architecture and folklore. Rosen, A.M., 1986. Cities of Clay: The Geoarchaeology of Tells. University of Chicago
J. Palestine Orient Soc. 12e13, 223e247. 221e283. Press, Chicago.
Carter, T.H., Pagliero, R., 1966. Notes on mud-brick preservation. Sumer 22, 65e76. Shahack-Gross, R., Albert, R.M., Gilboa, A., Nagar-Hilman, O., Sharon, I., Weiner, S.,
Courty, M.A., Goldberg, P., Macphail, R., 1989. Soils and Micromorphology in 2005. Geoarchaeology in an urban context: the uses of space in a Phoenician
Archaeology. Cambridge University Press, Cambridge; New York. monumental building at Tel Dor (Israel). J. Archaeol. Sci. 32, 1417e1431.
Dalman, G., 1928e42. Arbeit und Sitte in Palästina. Bertelsmann, Gütersloh. Singer, A., 2007. The Soils of Israel. Springer-Verlag, Berlin.
Dan, J., Fine, P., Lavee, H., 2007. The Soils of the Land of Israel. Eretz, Tel Aviv. Spencer, A.J., 1979. Brick architecture in Ancient Egypt. Aris and Phillips, Warminster.
Davidson, D.A., 1973. Particle size and phosphate analysis: evidence for the evolu- Stevanovic, M., 1997. The age of clay: the social dynamics of house destruction.
tion of a tell. Archaeometry 15, 143e152. J. Anthropol. Archaeol. 16, 334e395.
Emery, V.L., Morgenstein, M., 2007. Portable EDXRF analysis of a mud brick Stoops, G., 2003. Guidelines for Analysis and Description of Soil and Regolith Thin
necropolis enclosure: evidence of work organization, El Hibeh, Middle Egypt. Sections. Soil Science Society of America Inc., Madison, Wissconsin.
J. Archaeol. Sci. 34, 111e122. Torraca, G., Chiari, G., Gullini, G., 1972. Report on mud brick preservation. Meso-
Gazit, D., 1986. The Besor Region. Ministry of Education and Culture, Tel-Aviv. potamia 7, 259e286.
Gazit, D., 2000. Settlements processes in the Besor region in the days of the Sultan Weiner, S., 2010. Microarchaeology: Beyond the Visible Archaeological Record.
Abed Elhamid the 2nd. In: Schwartz, J., Amar, Z., Ziffer, I. (Eds.), Jerusalem and Cambridge University Press, Cambridge.

You might also like