French Et Al-1999-Geoarchaeology
French Et Al-1999-Geoarchaeology
French Et Al-1999-Geoarchaeology
Soils and sediments of a terraced slope at an Early Bronze Age site on the Aegean island of
Amorgos were examined micromorphologically to determine the nature and amount of ero-
sion on the slope during the past 5000 years, and how this had affected the formation of the
surviving archaeological record. The deposits forming representative terraces were examined,
as was the postdepositional sequence overlying the site, and a palaeosol preserved beneath
terrace retaining walls at the break of slope. The buried, preterrace system “red soil” was a
reworked red palaeosol, much affected by downslope erosion processes, which probably
commenced with clearance associated with the Early Bronze Age occupation of the site.
Examination of this soil suggested that there were at least two premodern phases of use of
the hillside. 䉷 1999 John Wiley & Sons, Inc.
INTRODUCTION
During surface investigations at the Early Bronze Age site of Markiani on the
island of Amorgos in the Cyclades (Figure 1), it became apparent that downslope
erosion had drastically modified the surface distribution of artefactual materials, a
process itself affected by the extensive construction of agricultural terraces across
the site. It was anticipated that the application of micromorphological techniques
(after Courty et al., 1989) could significantly enhance our understanding of the
erosion processes involved, as well as the disturbance processes related to the
construction, use, and subsequent decay of the relict agricultural terracing. There-
fore, the principal focus of this study has been to attempt to identify the soil trans-
formations associated with the periods of use of the site, and with the creation,
use, and collapse of the agricultural terraces on the slope. The research has con-
tributed to an understanding of the creation and modification of the surface distri-
bution of artefacts at the site, and has produced information about the poorly short
understood prehistoric environment and land-use of the Cycladic islands. standard
short
Figure 1. (a) Location of Amorgos; (b) location of Markiani; (c) site topography and profile locations.
standard
Following discovery and initial investigation in 1985, the site at Markiani was
subjected to detailed surface investigation in 1987 and excavation during 1988 –
1990, by a collaborative team from the universities of Ioannina, Athens, and Cam-
bridge, from 1987 to 1990 (Marangou, 1994:470 – 471; Davis, 1992:752 – 753). The
final publication of those investigations is in progress (Marangou et al., in prepa-
ration), while the present article reports on subsequent research undertaken to
assist in understanding issues of site formation and postdepositional modifications
of the archaeological record.
The site at Markiani is one of many small Early Bronze Age sites known from
the island (Marangou 1984, 1994). It was first occupied late in the first phase of the
Cycladic Early Bronze Age (ca. 3000 – 2800 B.C.), documented by ceramics of the
later Grotta-Pelos Group (Renfrew, 1972:152 – 169, 526 – 528, 1984; Doumas, 1977:
15 – 18). In all, four prehistoric occupation phases have been identified at the site,
the latest characterized by material of the Kastri Group, late in the Early Bronze
Age (Renfrew, 1972:172, 533 – 534; Sotirakopoulou, 1993; Manning, 1995:51 – 63, 81 –
86), approximately 2350 – 2200 B.C.
Later use of the site is represented by material ranging from the Geometric
through the Classical, Hellenistic, and Roman periods (ca. 800 B.C. – A.D. 500),
principally recovered through surface collection, but also represented in the su-
perficial levels of many of the excavation trenches. After this, use of the site appears
to be limited to the system of agricultural terraces (cultivation ceased during the
1960s), while at present the slopes around the site are grazed by sheep and goats,
with small folds on the summit of the hill.
Figure 2. View looking west across the slope at Markiani: (A) gentle terraces at base of slope;
(B) summit of hill.
(Figure 1[c]). While the basic outline of the history of site use could be ascertained
from the available archaeological data, too little work has been done on the down-
slope erosion of archaeological materials to enable detailed inferences about the
processes responsible for the creation of the present surface distribution of ceram-
ics, or the original depositional patterns of archaeological materials on the site
(Kirkby and Kirkby, 1976; Davidson, 1976; Rick, 1976; Stein and Rapp, 1978; Thor-
nes and Gilman, 1983; Reid and Frostick, 1985; Rosen, 1986; Wilkinson and Duhon,
1990; James et al., 1994; Wainwright, 1992). The exploratory study of the site and
slope sediments reported here was undertaken to provide sedimentological infor-
mation to aid in the interpretation of the erosion history of the site, and to con-
tribute to our understanding of the nature and effects of terrace construction on
the slope deposits and the archaeological materials embedded within them.
METHODS
Detailed examination of the slope, undertaken during and after the surface col-
lection, enabled the identification of a number of exposures across the site where
archaeological excavation or terrace wall collapse had exposed sediments which
could be investigated without further damage to slope stability. Prospection of the
main area of terraces on the slope beneath the site was undertaken with the fol- short
lowing aims in mind: standard
The soil blocks were impregnated and made into “mammoth” thin sections (after
Murphy, 1986) at the Geo-Archaeology Laboratory, Department of Archaeology,
Cambridge. The thin sections were analyzed using a Leitz polarizing microscope
and described using the terminology of Bullock et al. (1985). Table I is a summary
of the results, and the detailed descriptions are given in the Appendix.
past also have been the most easily worked section of the slope. Cultivation could
have been undertaken without the construction of terraces, though when the rest
of the slope was terraced in the more recent past, this area was also, removing
stones from the fields as well as creating a level and more stable cultivation surface.
Relatively few Early Bronze Age sherds were recovered this far downslope, and
most are small and heavily abraded. The densities of obsidian, contemporary with
and eroding from the same deposits as the Early Bronze Age ceramics, are signif-
icantly higher, indicating, as on other early prehistoric sites in the Cyclades (Whi-
telaw, 1991b), the preferential survival of obsidian during weathering and down-
slope movement, whereas relatively low-fired Early Bronze Age ceramics
comminute, abrade, and eventually disintegrate with movement downslope. This
would have been exacerbated on these terraces, if the prehistoric sherds were
exposed to a period of early historic as well as more recent cultivation.
Very few archaeological materials are visible in the exposed sections where ter-
race walls have collapsed. While too few for any effective quantitative comparison,
the low density visible in vertical section is compatible with the low density of
material visible on the surface. The absence of any clear deposition or sedimen-
tation levels within the fills behind the terraces suggests that (at least at the front
of the terraces) the soil has been thoroughly mixed in the construction of the
terraces, rather than representing a complex erosion or build-up sequence. Helle-
nistic as well as Early Bronze Age sherds are found through the whole vertical
exposure.
At two points among the terraces, it was possible to identify exposed probable
palaeosols directly underlying the base of the terrace walls (Profiles 2 and 3). No
sherds were found in these exposures, though given the low density of sherds in
any of the lower slope fills, this cannot be taken as necessarily documenting a pre-
Early Bronze Age date for these palaeosols. These levels predate the terraces they
underlie, differ visually from the overlying terrace fills, and could potentially rep-
resent sediments accumulating before the significant deposition of cultural mate-
rials on the slope.
Profile Description
At the base of Profiles 2 and 3, a preterrace soil was found to be preserved and
exposed beneath the lowermost stones of the terrace retaining walls (Figure 4).
This soil consisted of two horizons. The upper, thinner (ca. 80 – 100 mm) horizon
was a pale to medium brown silty clay loam with common small (⬍20 mm) stones.
It exhibited a poorly developed, small, irregular blocky ped structure (Figures 5
and 6). The lower, thicker (ca. 200 – 350 mm) horizon was a pinkish-red to orangey-
red clay loam which exhibited a similar soil structure, with an even mix of small
stones (⬍10 mm) and occasional larger stone inclusions (⬍100 mm) (Figure 5).
All thin sections from the main, lower horizon of this soil exhibited similar char-
acteristics. The matrix is a clay loam containing almost equal proportions of sand short
(30%), silt (30%), and clay (40%). It exhibits a moderately well developed, small to standard
short
Figure 4. The in situ red soil below the retaining wall, at profile 3. Trowel is 15 cm. standard
Figure 5. Photomicrograph of the red soil of Profile 3, illustrating the irregular to subangular blocky
ped structure and the common occurrence of small limestone fragments (white, coarse sand size). Cross-
polarized light. Frame width ⫽ 4 mm.
medium, sub-angular blocky ped structure, which is evident both in the field as
well as in thin section (Figure 5). The soil is relatively porous with many biological
aggregates (20 – 45%) and contains very little organic matter (⬍1%), and the whole
matrix is reddened as a result of impregnation with amorphous iron (but not burnt)
(Figure 6). The clay fraction is dominated by nonlaminated “dusty” (or impure)
clay, which is evenly distributed throughout the whole fabric, both in the ground-
mass and in the voids.
Interpretation
The unsampled upper horizon is essentially a more friable and browner version
of the underlying horizon that was examined in thin section. It is probably more
organic, and therefore could be considered as an A-type horizon. This suggests that
there was no appreciable preterrace truncation of this soil.
The lower horizon, with its distinctive structure, reddening and enrichment with
clay and iron, probably represents a red Mediterranean soil. Red soils and in par-
ticular, terra rossa, are believed to have been characteristic of this part of the
Mediterranean prior to clearance and intensive agricultural practices (Bridges, short
1978:67 – 70; Zangger, 1992). standard
Figure 6. Photomicrograph of the fine fabric of the red soil of Profile 3, with a fragment of oriented
clay subsoil in the top left corner. Plane-polarized light. Frame width ⫽ 4 mm.
Thus, these two horizons would appear to represent an in situ soil preserved in
“pockets” on the slope and beneath the much later terrace construction, although
it is not necessarily the original and unmodified preagricultural soil.
The substantial impure clay fraction in this palaeosol deserves further comment.
In this soil profile, it occurs both within the groundmass as an integral part of the
soil fabric and in the interpedal void space. “Dusty” or impure clay is composed of
a mixture of microcontrasted silt particles and minute fragments of organic matter
(Macphail, 1987). This type of illuvial clay is particularly characteristic of disturbed
soils (Slager and van de Wetering, 1977). This may be indicative of an earlier phase
of incorporation of fines as a result of the disturbance of a bare soil, associated
rain splash and/or slope wash erosion and the consequent addition of fines, and
within-soil mass movement of fines. On the other hand, the dusty clay lining the
void space suggest a subsequent phase of illuviation of fines down-profile (Mac-
phail, 1992; Kwaad and Mücher, 1977). Consequently, there may be two preterrace-
construction phases of disturbance and illuviation of fines. In addition, the fact that
this soil exhibits no sign of any recent illuvial movement of fines suggests that this
soil ceased to receive fine material once it was buried by the sediments later re-
worked in the construction of the terraces.
Both phases of illuviation suggest that there was considerable erosion of open short
and often bare slopes prior to the establishment of the terrace system. It is probable standard
that this illuvial dusty clay deposition was caused by rainsplash and localised slo-
pewash erosion of fine textured, bare soils and within-soil, down-profile, mass
movement of the dislodged fine material (Kwaad and Mücher, 1979). A rare torren-
tial rainstorm would have been sufficient to cause mass-movement of soil down-
slope (Clark and Small, 1982:35 – 40), very much like a mudslide. By implication it
may be suggested that this bare soil was associated with past agricultural or pas-
toral use. Thus this soil was already much modified by gradual colluvial accumu-
lation by the time it was buried by terrace wall construction.
The origin and formation of terra rossa is an extremely complex issue. There
are three major types (after van Andel, 1998, 1996; Heilmann, 1972; Limbrey, 1975;
Spaargaren, 1979):
1. original terra rossa formed by the in situ dissolution of limestone in karst
landscapes and characterized by amorphous iron oxides and hydroxides, with
the clay fraction dominated by kaolinite;
2. reworked red soils, which are generally noncalcareous and with up to 25%
aeolian content, often formed as a result of hillwash processes; and
3. red fan deposits which are colluvial/alluvial deposits derived from terra rossa,
or uplifted and dissected from flysch deposits.
Limbrey (1975:205) generalizes that terra rossa soil profiles occur on hard lime-
stones and other calcareous rocks, and are composed of a reddish brown A horizon
and a strong reddish brown to red textural B (or Bt) horizon (Limbrey, 1975:205).
This Bt horizon exhibits a strong blocky ped structure with marked illuvial clay
coatings in the inter- and intra-aggregate pore space. The leaching in these soils
leads to intense clay migration, with the characteristic reddening or rubifaction,
indicating no more than seasonal drying sufficient to cause the dehydration of iron
oxides. Furthermore, Limbrey (1975:204 – 205, 211) regards these red soils as relict
survivals of probable interglacial age, formed under various types of deciduous and
evergreen forest.
There is no doubt that this red soil at Markiani formed as a result of leaching,
deposition of amorphous iron, and clay migration down profile. As most of the
illuvial clay in thin section is clearly of nonlaminated dusty type, there has been
much within-soil mass movement of fine material, probably associated with dis-
turbance and erosion of the original soil. Thus, although the soil present beneath
the terraces at the base of Profiles 2 and 3 would appear to exhibit the terra rossa
characteristics set out by Limbrey (1975), they bear a much closer resemblance to
van Andel’s definition of reworked red soils (van Andel, 1998). Moreover, there is
no reason to think that this kind of intense clay illuviation, erosion and slumping
could not have occurred within the postglacial period; there is no need to invoke
a much greater age for this soil. Unfortunately, it is difficult to ascertain the tem-
poral range of this reworked red soil, or indeed to link its formation directly with
the site’s occupation in the Early Bronze Age. It most probably developed subse-
quent to deforestation of the slope and the illuviation events were associated with short
preterrace, agricultural or pastoral phases of use of the hillside. standard
Archaeological Implications
It would seem plausible to suggest that the erosion documented in this profile
was initiated with cultivation and grazing on and immediately around the site at
Markiani, following its establishment late in the first phase of the Early Bronze Age,
ca. 2900 B.C. While systematic, intensive, regional surface survey has not been
undertaken on Amorgos, extensive exploration during the past century (Tsountas,
1898; Dümmler, 1886; Marangou, 1984, 1994) has documented widespread occu-
pation throughout the island during the Early Bronze Age, with several sites within
about 1 km of Markiani. It cannot be certain that these are the closest contemporary
sites to Markiani, but similar patterns of intersite spacing in the Early Bronze Age
have been revealed where intensive surveys have been undertaken on other Cy-
cladic islands (Cherry, 1982; Érard-Cerceau et al., 1993).
The relatively small population of the site at any time during the Early Bronze
Age is unlikely to have required cultivation (and therefore terrace construction) on
the steep slope below the site. Rather, fields are likely to have been situated on the
gentle slopes immediately to the north, leading toward what were, until recently,
well-watered valleys draining north (S. Giannakos, personal communication, 1992).
However, cultivation at the top of the slope, dumping of refuse downslope from
the site, and grazing in the vicinity of the site, would all be likely to cause clearance
and erosion on the slope below the site.
Excavated deposits with ceramics of the first two phases of occupation at Mar-
kiani are limited, and the small and abraded condition of the sherds may indicate
that occupation of the site during the earlier part of the Early Bronze Age was
episodic rather than continuous. This may have resulted in several short-term cy-
cles of clearance and partial regeneration of the natural vegetation on the slope.
Equally, the limited extent of the site in the early phases might also mean that the
earliest occupation extensive enough to have a significant impact on the local en-
vironment was not until later in the Early Bronze Age (phases III and IV: ca. 2700 –
2200 B.C.).
The preservation of an A-Bt horizon profile at the top of this exposure indicates
that little, if any, of the original profile has been truncated by the construction of
the terraces. The fills behind the terraces, comparable to the fills sampled in Profiles
4 and 6, probably represent a reworking of sediments which had accumulated
above the palaeosol. The degree of development of the red soil indicates that these
overlying sediments resulted from deposition episodes substantially later than the
stabilization of the earlier deposit. The two phases of illuviation identified micro-
scopically could either predate, or relate to these proposed later depositional epi-
sodes. Given the time which must have elapsed for the development of the structure
visible in the palaeosol, the proposed later episodes of sedimentation should relate
to later utilization or partial clearance and destabilization of the slope, potentially
tying in with the Geometric to Roman use of the site.
While the interpretation of this soil is somewhat speculative, given the present short
state of our understanding of the soil and vegetation history of most of the southern standard
Aegean, let alone our limited understanding of settlement history on Amorgos, the
sedimentary evidence of this profile, associated with a well-documented occupa-
tional sequence on the site, allows a more direct linkage between human behavioral
processes and their potential impacts on the landscape through vegetation change
and erosion, than is usually possible in regional environmental studies. In this
sense, the micromorphological analysis of this profile opens up possibilities for
environmental and land-use research which were not originally anticipated, but
which deserve to be explored further.
Archaeological Context
The midslope terraces below the site, but above the sediment trap of the shallow
terraces considered above, are all small, 1.5 – 4.0 m in width, usually retained by
walls 1.0 – 2.0 m high. Bedrock is sometimes exposed at the upslope edge of the
cultivation surface, and the retaining walls are sometimes, at least for part of their
length, built on bare rock. The terraces are all of relatively short lateral extent, 4.0
up to 20.0 m, broken up by exposures of bedrock or large detached boulders. Many
of these terraces will not have been accessible for animals to draw a plough, and
will most likely have been cultivated by hand using a mattock or spade. As with
plough cultivation, soil turning during cultivation will have been limited to the
upper 10 – 15 cm of the deposit.
The Archaeological Service Representative who assisted us in collecting the soil
samples, Mr Simos Giannakos, had, in the past, cultivated terraced fields several
hundred meters southwest of those under study. As he explained, the usual method short
of construction of a terrace was to cut back slightly into the slope, build a wall standard
Figure 7. Photomicrograph of the poorly developed structure of the terrace make-up of Profile 4,
with the peds and quartz grains outlined by amorphous iron oxides. Plane-polarized light. Frame
width ⫽ 4 mm.
against the small escarpment formed, and then fill up behind the wall with earth
cut from the rear of the terrace — or cut away to form the seating for the next
terrace wall above.
Description
The most striking visual characteristic of the sediment is the amount, size, and
variety of stone evenly mixed, but poorly sorted, throughout the matrix; the stone
content varies in quantity from about 40% to 70% of the whole matrix. The remain-
ing soil size fraction ranges in quantity from about 30% to 60% of the matrix, and
is composed of a porous (30%), orangey/reddish brown sandy (clay) loam to loam.
In general, the soil component contains more fine sand and coarse silt, and much
less clay than in the buried soil. It exhibits a poorly developed irregular to suban-
gular blocky ped microstructure (Figure 7). There is a relatively minor organic
matter component to the fine groundmass (Figure 8), but it is three to four times
greater than in the reworked red soil in Profiles 2 and 3. There are also a few soil
faunal excrements within the void space, and the whole fabric is highly biologically
reworked. short
All of the slides contain a variety of different clay coatings within the soil fraction. standard
Figure 8. Photomicrograph of the very fine organic component within the fine groundmass of the terrace
make-up of Profile 4. Plane-polarized light. Frame width ⫽ 4 mm.
Nonlaminated dusty clay (15%) with poor to moderate birefringence is the most
common, with relatively minor amounts (5%) of similar impure clay “lining” the
interpedal voids and channels (Figure 9). In addition, there were rare (⬍2%) frag-
ments of “clean” yellow clay within the groundmass exhibiting strong birefringence
which are probably eroded fragments of bedrock (schist/mudstone) rather than
pure clay coatings (R. Macphail, personal communication, 1995).
Interpretation
The makeup of the terraces exposed by Profiles 4 and 6 is dominated by one
fabric composed of poorly sorted stone and an organic sandy clay loam to loam
with poor soil structure. This is a mixture of soil material derived from the mass
movement downslope and illuviation of fines (silt and clay), as well as higher ve-
locity erosion responsible for the incorporation of limestone pebbles and rocks, all
subject to considerable mixing by soil faunal activity. There are also minor com-
ponents such as eroded subsoil material, and past and present material derived
from rainsplash and hill-wash.
The clay component suggests that there were three types of erosion occurring
at different times:
short
1. Minute, clay-sized, eroded fragments of the subsoil: These may be associated standard
Figure 9. Photomicrograph of successive, laminated, pure and dusty clay infills of an interpedal channel
typical of all the terrace make-up deposits examined, which has been subject to some removal. Plane-
polarized light. Frame width ⫽ 4 mm.
Archaeological Implications
Unlike the lower, broader terraces, there is nothing to suggest that the upper
slopes of the site were ever cultivated in antiquity, though there was sufficient soil
disturbance in places upslope to cause subsoil and soil erosion downslope. The
Early Bronze Age sherds on the surface and incorporated in the terrace fill, given short
their edge abrasion, are likely to have been eroded from the site above, and the standard
Figure 10. View of the scarp of Trench I.5 at the summit of the site, looking west. The retaining wall
has been mostly removed from the front of the terrace at the left, and the bedded plane of small stones
marking the top of the Early Bronze Age fills is visible descending to the left, halfway up the section.
Height of soil section ⫽ 108 cm.
same seems likely for the later historical material. The terrace walls themselves
reveal no evidence for earlier foundations or phases of rebuilding.
The absence of any clear preterrace deposits on the main slope may be a function
of the steeper slope and shallower soils, such that all deposits have been reworked
in the construction of the terraces, or those that survive do so only in small pockets
in the rock, buried beneath the terrace fills. In contrast, the sediment trap formed
by the lower, broader terraces, may have been much more effective at retaining
deep slope deposits that did not need to be reworked so completely in the con-
struction of the recent terraces.
Archaeological Context
These samples constitute a sequence within and above the structures represent-
ing the final phase of occupation at the site late in the Early Bronze Age. No higher
levels representing the historical use of the site were preserved in this trench, but
there is late material in the plough-zone and on the surface, primarily of Hellenistic
date, representing a disturbed superficial level.
Occupation of the site extended onto this natural rock-bounded terrace, below
and to the south of the summit of the hill, by the third phase of occupation, ca.
2700 – 2400 B.C. There is no sign of a sudden or violent destruction of the phase IV
structures, and no complete vessels were found in situ on the house floors. The
site appears to have been abandoned (ca. 2200 B.C.), and the structures left to
collapse naturally.
The sampled section (Figure 10) comes from within a room or perhaps a small
enclosed yard. The lowest levels sampled sit on top of the final occupation floor,
but below the preserved tops of the surrounding walls, and so represent collapse
debris and sediments washed into the ruins, probably from similar deposits on the
exposed summit of the hill, immediately to the north.
This fill appears to have stabilized with a layer of small- to medium-sized rocks
and sherds, lying bedded downslope to the south, at the level of the preserved tops
of the walls. This layer represents deflation of concentrated debris forming a lag
deposit which capped the debris of the abandoned site (Figure 10). This would
have provided an effective guard against further natural erosion, particularly coun-
tering the erosive effects of rainsplash. Bounded downslope by rock outcrops, the
deposits would also have been stabilized against mass downslope slump.
In a trench immediately to the east, a wall from the later reoccupation of the site
(Hellenistic or Roman) lay exposed on the surface of the terrace. The two courses
of the wall sit on, but do not cut into, the lag deposit that seals the Early Bronze
Age collapse fill. In the sampled exposure there is likewise no evidence that later
occupation interfered with the earlier deposits or exposed them to further erosion.
In the sampled exposure, the Early Bronze Age walls, fill, and deflated surface
are directly covered by an undifferentiated fill, retained by the wall fronting the
terrace, itself founded directly on bedrock (Figure 10).
Description
Within the excavated area, the main upstanding section which survived to a
height of 1.08 m in Trench I.5/space 1 appeared to have the best potential for soil/
sediment information directly related to the abandonment of the site. The surface
of the terrace is level, well-compacted, and slightly deflated. It was last cultivated
several decades ago and, until excavation, provided one of the main access routes
for sheep and goats from the slope to the shelters immediately above, on the sum-
mit of the hill.
The sampled profile exhibited the following stratigraphy from top to bottom short
(Figure 10): standard
0 – 0.45 terrace buildup of a pale brown silty clay loam with common, ir-
regularly oriented stones (⬍80 mm);
0.45 – 0.60 similar sediment, but with abundant, large (100 – 250 mm) slabs of
schist and other medium and small stones and sherds, bedded
downslope;
0.60 – 0.85 similar sediment with common, irregularly oriented medium stones
(⬍80 mm);
0.85 – 1.08 similar sediment with occasional small stones (⬍20 mm) exhibiting
some horizontal bedding; and
1.08⫹ (m) as yet unexcavated, but could reveal several cms of in situ floor
makeup and earlier occupation deposits resting directly on bed-
rock.
The profile was sampled at the following intervals: 0.10 – 0.20, 0.26 – 0.40, 0.66 – 0.80,
0.80 – 0.94, and 0.94 – 1.06 m (below present terrace surface).
Interpretation
Samples 1 and 2 (0.10 – 0.20 and 0.26 – 0.40 m): This material, largely under-
lying the superficial ploughzone, is a golden to reddish brown, dense and well
mixed, sandy clay loam to loam which exhibits a poorly developed irregular to
subangular blocky ped structure. It contains abundant nonlaminated dusty clay
throughout the groundmass and few very fine fragments of organic matter.
This illuvial dusty clay fabric is suggestive of an eroding bare soil, accumulating
by rainsplash and slopewash erosion. It exhibits some structural development that
suggests that it has been undisturbed for some time, even though it is undoubtedly
associated with relatively recent terrace construction, and in its upper levels, cul-
tivation in the recent past.
Sample 3 (0.66 – 0.80 m): This material is very similar to that observed in the
above two samples, although it is much more porous (or vughy). This profile could
be the result of the incorporation of relatively large amounts of organic matter
which has since been destroyed by oxidization. This horizon might represent the
decay of occupation debris or postabandonment vegetation, combined with sedi-
ment derived from the hill-top through downslope erosion processes.
Samples 4 and 5 (0.80 – 0.94 and 0.94 – 1.06 m): Both samples are a fine, clay
loam with an intergrain channel and very porous microstructure. The fine ground-
mass is dominated by very abundant nonlaminated dusty clay, as well as occasional
zones of amorphous calcium carbonate.
These features suggest that both samples represent a downslope accumulation
of eroded fine material and much organic debris, now largely oxidized, leaving a
very open, vughy structure. In summary, it is suggested that this profile represents
the following sequence of events:
1. postabandonment building collapse, sediment inwash, and organic accumu- short
lation (1.08 – 0.85 m); standard
2. accumulation of eroded soil material from the highest levels of the former site
(0.85 – 0.60 m);
3. further collapse of surviving stone walls of former structures, followed by
rainwash and deflation (0.45 – 0.60 m); and
4. the deliberate buildup of material in terrace construction (0.45 – 0 m).
Archaeological Implications
The observations on the sediments correspond well with the archaeological un-
derstanding of the contexts of the samples, and the processes responsible for the
formation of the deposits. Because of the limited horizontal exposure and incom-
plete preservation, it is not clear whether the samples come from within a roofed
space, or immediately adjacent to roofed rooms, in a small open court. The deposits
described under 1 and 2 above comprise the collapse debris from the structures,
probably including considerable soil incorporated as water-proofing layers in the
make-up of flat roofs. The high organic content of these levels may derive from
debris left on the occupation surface when the structures were abandoned,
branches, brush, or reeds from roof construction (Cameron, 1972; Shaw, 1973:221 –
222), material dumped from neighboring occupation areas, if this set of rooms went
out of use before the final abandonment of the community, or may represent the
accumulation of organic litter from vegetation colonizing the nutrient-rich deposits
of the abandoned site. Additional sediments would have washed in from similar
deposits within and between other abandoned and collapsing structures immedi-
ately above, on the summit of the hill. Where not disturbed by subsequent erosion,
such collapse deposits from single-storey structures may accumulate up to 0.60 –
0.70 m in depth (Warren, 1972: Figure 6; Coles, 1973:68; Ammerman et al., 1976:
41 – 46).
The overall similarity of all of the samples from this profile is not surprising,
given that the sediments were derived from comparable deposits on the hilltop
immediately above. Likewise, the similarity of the uppermost two samples, repre-
senting fill introduced during the construction of the agricultural terrace, suggests
that this fill was also derived from nearby on the hilltop, if not simply reworked in
situ.
No distinct strata could be identified between the stone level marking the top of
the Early Bronze Age collapse deposits, and the fills introduced as terrace makeup,
which would correspond to deposits contemporary with the historic (Hellenistic/
Roman) use of the site. Limited traces of late walls, and the considerable quantity
of sherds on the surface of the site and incorporated into the topsoil, point to a
significant occupation, perhaps a seasonal or year-round farmstead. The absence
of a clear depositional level associated with these walls, and the incorporation of
sherds of this date into the terrace fill deposits, suggest that the relevant levels, at
least on this terrace, have been completely reworked during the construction of
the existing terraces on the upper slopes of the hill. short
As with Profile 3, the upper samples of Profile 7 provide no direct evidence for standard
the dating of the existing terraces. While they incorporate historical sherds through-
out their depth, these could originally have been a late surface deposit, subse-
quently mixed during the construction of the terraces. It is, however, worth noting
that sample 2, well below the ploughzone, is essentially indistinguishable from
sample 1, which will have been disturbed regularly by ploughing until the past few
decades. While the rate of soil formation in these soils is not known, the lack of
any marked differentiation between these samples would suggest that the terraces
are a relatively recent construction. This is consistent with the lack of evidence for
rebuilding of their retaining walls. Given the similarities in construction style and
condition of the terrace walls across the entire slope, this suggested recent date
should apply equally to those lower on the slope.
DISCUSSION
The micromorphological study of selected profiles at Markiani has provided de-
tailed descriptions of the terrace and archaeological deposits, and has given some
idea of the nature of the site formation and erosion processes affecting the ar-
chaeological site, and generating and modifying the surface archaeological record.
In addition, it has provided information on the history of vegetation and land-use
change in the vicinity of the site. Five principal contributions of this investigation
can be summarized:
very localized pattern of interactions, the direct spatial association of the archaeo-
logically documented human activities on the site and slope and the sedimentolog-
ical features identified micromorphologically enable a reasonably secure recon-
struction of patterns of land-use at Markiani.
There is little doubt that a reworked terra rossa-like soil is preserved in places
beneath the terraces on the slope below the site at Markiani. The analysis of this
buried soil suggests a sequence of landscape development involving: (1) terra rossa
soil formation; (2) an initial phase of illuviation and colluviation, associated with
increasing leaching and oxidation over time, probably associated with agricultural/
pastoral use of the hillslope, leading to the formation of a reworked red soil; (3) a
period of relative slope stability and soil formation processes within the reworked
red soil, which involved the illuviation of fine material; (4) a further period of
erosion leading to the deposition of illuvial fines (silt and clay), probably resulting
from rainsplash erosion and mass movement of soil washing off a bare ground
surface uphill; and (5) burial by the agricultural terrace system.
While the initial period of erosion of this red soil cannot be determined directly,
and no archaeological materials have been noted incorporated within it, the most
likely period for the onset of such erosion would appear to be the first occupation
of the site, fairly early in the Early Bronze Age, when activity on the hilltop and
cultivation and grazing in the immediate vicinity of the site would almost certainly
have had an impact on the stability of the slope soils. Alternatively, this period of
erosion might highlight an intensification of activity in the second half of the Early
Bronze Age, when the population of the community increased and occupation ex-
panded onto the upper southern slope of the hilltop.
No physical or sedimentological evidence survives to indicate that there was any
agricultural terracing on the slope during the Early Bronze Age. While any such
evidence might have been completely removed during later erosion or terrace con-
struction, the illuviation phases indicate that the lower slope deposits were affected
by erosion from above, such that the slopes were not, at that time, effectively
stabilized by terraces.
Potentially up to 800 years after the initial occupation of the site and the probable
start of landscape modification in its vicinity, the site was abandoned. The profile
studied in the excavation trenches at the top of the slope provides clear documen-
tation of the disuse, collapse, and silting of the ruins, and subsequent stabilization
of the slope. These deposits were capped and preserved beneath a deflation level
consisting of stones and sherds, representing a stable slope surface. This deposit,
however, cannot be taken as representative of the entire site area, since the bed-
rock outcrops which surround this terrace have obviously help to stabilize and
preserve the archaeological levels. Elsewhere on the summit of the hill, the ar-
chaeological deposits have been severely eroded, and little survives in situ except
in pockets in the bedrock. That the later (predominantly Hellenistic) activities on
the site did not disturb the underlying Early Bronze Age deposits in the area sam-
pled probably relates to the depth of post-Early Bronze Age sediment accumulation short
on that specific shelf, and cannot be extrapolated to the entire site. We can there- standard
fore anticipate that there was erosion associated with the post-Early Bronze Age
use of the site, in large part responsible for the deposition of the sediments later
reworked into the terrace fills on all parts of the slope. Because of the mixing of
these deposits in the construction of the terraces, the period of deposition of these
sediments cannot be dated on the basis of the sherds incorporated in them. De-
positionally, we can only say that they should significantly postdate the abandon-
ment of the Early Bronze Age community, and predate the construction of the
terraces. However, given the quantity of sherd material down the entire slope, it
would be surprising if the Hellenistic phase of activity on the site did not have a
detectable effect on slope stability and sediments.
The final period of significant human alteration of the slope deposits is repre-
sented by the construction of the agricultural terraces. These cannot be dated di-
rectly, though the absence of structural evidence for rebuilding suggests that they
are of no great antiquity, a conclusion supported by the absence of significant soil
structure development in the sediments behind them.
less clear. The quantity of sherds on the surface and slope indicate a substantial
amount of activity, centered on a focus as small as the Early Bronze Age occupa-
tion, but the quantity of material deposited over a much shorter period suggests
fairly intensive use. The good preservation of the sherds on the lowest terraces
may indicate that this part of the slope, where terraces were not necessary, was
cultivated. While traces of terracing higher up the slope might have been completely
removed or obscured by more recent terrace construction, no hints of such con-
struction can be found either in the standing architecture of the present terrace
walls or in the many soil profiles examined, and those analyzed.
The argument for a recent date for the single phase of terrace construction pres-
ently visible on the slope is based on the relatively good condition of the standing
terrace walls and the absence of any clear indications of rebuilding, such as have
been noted in systems a century or more in age (Whitelaw, 1991a:405 – 410; Wells
et al., 1990:228). From the perspective of the soils behind the terraces, the absence
of a significant contrast in soil structure development between the upper cultivated
ploughzone, and the underlying terrace fill deposits, cannot be used to date the
terraces directly, but does not suggest that the fills have been in situ for a long
period of time.
While one cannot generalize from the terracing evidence at a single small rural
site, an approach combining archaeological study of the slope, with the analysis of
soil development in the associated terrace fills, holds out the promise of a more
effective approach to understanding and dating Aegean terrace systems. The study
of terrace soil development has been recognized as potentially useful (van Andel
et al., 1986:117; Wells et al., 1990:228), though it has not previously been pursued
systematically.
graphic evidence for the formation of a lag deposit on top of the abandoned Early
Bronze Age occupation area. While this may have stabilized particularly deep de-
posits on the protected rock shelf where the samples were taken, an identical
deposit was also found in an archaeological trench downslope 50 m to the south-
east. It is equally likely that similar stable deposits would have formed across the
rest of the slope, protecting pockets of slope soils. Corroboration for the episodic
nature of slope erosion comes from the formation processes of the preterrace soil,
which document a considerable period of slope stability before subsequent further
deposition of sediments.
Sherds and sediments are likely to have been dumped and to have eroded down-
slope from the settlement on the summit of the hill throughout the life of the
community, probably exacerbated during the expansion of population in the later
phases of occupation. Erosion is also likely to have increased as exploitation of
the surrounding slopes intensified, and vegetation was reduced by grazing. Thus,
there may have been a period of up to one-half a millennium during which Early
Bronze Age ceramics moved downslope, with new material constantly being added.
After the abandonment of the community, a considerable amount of erosion, par-
ticularly on the exposed summit of the hill, is documented by the fills inside the
collapsing structures on the shelf immediately to the south. Elsewhere, where such
natural terraces did not capture sediments, they would have eroded further down-
slope, until vegetation and deflation stablized the surface.
The second major phase of erosion may be attributed to the historical use of the
site, represented primarily by quantities of Hellenistic sherds on the site and slopes.
While some of the better-preserved sherds on the gentle terraces at the base of the
slope may have been deposited during cultivation of this section of the slope, some
of this material, and all of that on the upper slopes, probably derives from occu-
pation on the summit of the hill. Soil eroding with these sherds probably accounts
for the sediments which were later reworked as the fills for the terraces constructed
across the entire slope. While the underlying Early Bronze Age deposits in the
excavated trenches just below the summit of the site had not been significantly cut
into during this phase of use of the site, this need not apply to other areas, partic-
ularly shallower deposits on the summit or slope, and this may have been a period
in which previously stable sediments containing Early Bronze Age sherds were
destabilized, and new material added to the surface assemblage subject to erosion.
The duration of this phase of use of the site, and the erosion which almost certainly
accompanied it, remains unclear: ceramics (primarily coarse wares) ranging from
Geometric to Late Roman in date are present on the slope, though the vast majority
seem most consistent with a Hellenistic date. Therefore, the serious impact on
slope stability may have been confined to a fairly limited period of several gener-
ations.
Finally, we can anticipate some additional erosion in the recent past. This might
have been severe, if there was any significant utilization of the site or slope before
the construction of the existing terrace system, but this seems unlikely. There are short
very few recent ceramics either on the surface of the slope or in the superficial standard
excavated levels on the site. Furthermore, if serious soil erosion had been insti-
gated on an unterraced slope, there would have been little sediment out of which
to construct the existing terraces. Overall, while terrace construction will have
mixed and shifted sediments locally, it is unlikely to have induced large-scale ero-
sion, in that the slope would have been disturbed only to be immediately restabil-
ized by the terrace walls. While there has been some terrace collapse since the
abandonment of cultivation on the slope, exacerbated by the breakdown of some
terrace walls by sheep and goats, this appears, so far, to have been fairly limited
and localized.
Combining the archaeological and sedimentological observations, it appears that
the extensive surface sherd distribution at Markiani, relating to two principal pe-
riods of use of the site, results from two periods of disturbance and serious erosion,
during the deposition of each ceramic assemblage, and the period of time after
each occupation until the slope had stabilized. Recent observations suggest that
stabilization through colonization by vegetation might have been accomplished on
the order of decades, rather than centuries (Jameson et al., 1994:373; Rackham,
1990b).
conclusive, it does point to the potential for new and complementary insights based
on the micromorphological characteristics of dateable soil horizons.
A particular advantage of a site-specific study such as this is that, even without
direct dates for individual sedimentary events, a direct linkage with the on-site
archaeologically documented sequence of site occupation and use can be estab-
lished (Davidson, 1976; van Andel and Runnels, 1995). Such a direct linkage is rarely
possible with diachronic geomorphological or palaeobotanical investigations pur-
sued at the regional scale. In such studies, the linkage between cause and effect
can rarely be argued except through the broad synchronism of fairly low-resolution
behavioral and environmental sequences. While investigation at each scale can
make significant and complementary contributions, small-scale studies tied directly
to the occupation history of specific sites deserve greater attention.
manly induced landscape degradation suggests a new dynamic element which will
need to be explored more fully in attempts to understand the dramatic settlement
pattern changes throughout this period in the southern Aegean.
CONCLUSION
In general, the application of micromorphology to this archaeological site asso-
ciated with an intrusive terrace system has demonstrated its usefulness. Soil com-
position, and in particular, transformations of the soil/sediment regime, are directly
recognizable. Remnants of original soil types point to the existence of at least two
earlier types of landscape-stabilized open but vegetated slopes and eroding/utilized
slopes. Moving beyond the original objectives of this study, in future investigations
of southern Aegean landscape history, it will be crucial to establish closer temporal
control over the construction, use and degradation of terrace systems. There is an
obvious need to prospect for and to select better preserved sites for extensive
sampling and analysis, and to situate micromorphological research within a
broader program of palaeoenvironmental and geomorphological study.
Profile 4
Sample 1 (0.35 – 0.49 m)
In general, this sample is very similar to Profile 6, except for the following.
Organic Components: greater frequency of sesquioxidic plant pseudomorphs;
Mineral Components: up to 70% stone content, comprised of 30% of 10 – 30 mm,
10% at 5 – 10 mm and 30% at 2 – 5 mm in size; 30% soil fraction: as for Profile 6,
sample 1, with coarse/fine ratio ranging from 30 – 50/50 – 70; Pedofeatures: Textural:
nonlaminated dusty clay acts as “linings” of voids as link-bracing between grains
and aggregates of fine fraction; Amorphous: rare (⬍2%) rolled aggregate of dung-
like material, black and striated, ⬍250 m.
Profile 6
Sample 1 (0.20 – 0.30 m)
Structure: homogeneous; poorly sorted; poorly to moderately developed ped
structure, irregular to subrounded, 2 – 8 mm; Porosity: 30 – 32%; 10% interpedal
vughs, irregular, 1 – 4 mm, smooth to weakly serrated; 20% complex packing voids/
chambers, irregular, ⬍8 mm; 10% intrapedal channels, irregular, 100 – 500 m wide,
⬍10 mm long, walls partially accommodated, smooth to weakly serrated; 2% inter-
pedal channels, irregular, ⬍100 m wide, ⬍5 mm long, walls partially accommo-
dated, smooth to weakly serrated; Organic Components: very few (⬍2%) flecks of
charcoal, 50 – 200 m; ⬍4% very fine flecks of charcoal/organic matter in the
groundmass, ⬍50 m; very few (⬍2%) sesquioxide pseudomorphs of plant matter;
Mineral Components: 55% stone content, in size ranges 15 – 40 mm, 5 – 15 mm and
⬍10 mm, all subrounded to subsquare; 45% soil fraction; limit 100 m; coarse/fine short
ratio: 50/50; coarse fraction: 20% coarse and 10% medium mica, schist, and lime- standard
stone, and 20% fine quartz and mica, subangular to subrounded, 100 – 1000 m; fine
fraction: 10% very fine quartz and mica, 50 – 100 m, subrounded to subangular;
30% silt and 20% clay; very weakly speckled; orangey/reddish brown (CPL), reddish
brown (PPL), light orangey red (RL); Groundmass: fine: close porphyric; weakly
mosaic speckled to weakly reticulate striated in places; coarse: undifferentiated;
related: open porphyric; Pedofeatures: Textural: groundmass dominated by non-
laminated dusty clay (15%), poor to moderate birefringence, amber to orangey red
(CPL); interpedal channels and complex packing voids discontinuously “lined” with
nonlaminated dusty clay coatings (5%), ⬍50 m thick, poor to moderate birefrin-
gence, amber (CPL); Excrements: very few (⬍2%) rounded excrements in clusters
in void space, ⬍50 m; Amorphous: rare (⬍2%) very fine fragments of birefringent
clay or mudstone within groundmass, yellow (CPL), ⬍25 m; very rare (⬍1%)
phytoliths; whole fabric slightly impregnated with sesquioxides; few (2%) sesquiox-
ide nodules, subrounded, ⬍250 m, reddish black (RL).
Structure: intergrain channel (50%) or vughy (50%) structure; ca. 50% of the
groundmass is comprised of irregular aggregates and clusters of aggregates with
the groundmass having a “perforated look”; Mineral Components: ⬍5% stone frac-
tion; ⬎ 95% soil fraction; Pedofeatures: Textural: many to abundant (10 – 15%) non-
laminated limpid clay in groundmass, moderate to strong birefringence, yellow to
gold (CPL); abundant (15 – 20%) nonlaminated dusty clay mainly in groundmass
and coating grains with little in the voids, poor to moderate birefringence, gold to
amber (CPL); Amorphous: very rare (⬍1%) subrounded aggregates, ⬍500 m, com-
posed of very fine quartz, silt and dusty clay, dark brown mottles in golden orange
background (CPL); few (⬍2%) amorphous, irregular zones of amorphous calcium
carbonate within groundmass; rare (⬍1%) phytoliths in sesquioxidic pseudomorphs
of plant material; one small piece of pottery, subrounded, 6 mm.
We would like to thank Lila Marangou and Colin Renfrew for encouraging and facilitating this research.
Permission to take samples was granted by Lila Marangou and the KB’ Ephorate for Prehistoric and
Classical Antiquities, to whom we are also grateful. Simos Giannakos, Manoli Despotides, and Georgios
Gavalas have provided every assistance with the research at Markiani over many years. Simos Gian-
nakos, Georgios Gavalas, Natasha Angelopoulou, and Kasia Gdaniec assisted with sample collection in short
the field, while Julie Boast undertook thin section preparation in the laboratory, Lisa Nevett collaborated standard
in the identification of the post-Bronze Age ceramics from Markiani, and Richard Macphail and Tjeerd
van Andel provided advice on soils. Paul Goldberg, Lisa Nevett, Dave Passmore, Tjeerd van Andel, and
Malcolm Wagstaff kindly commented on an earlier draft of this article. The field and laboratory work
was funded by the British Academy and the McDonald Institute for Archaeological Research.
REFERENCES
Acheson, P.E. (1997). Does the “economic explanation” work? Settlement, agriculture and erosion in
the territory of Halieis in the Late Classical-Early Hellenistic period. Journal of Mediterranean Ar-
chaeology, 10, 165– 190.
Ammerman, A., Cavalli-Sforza, L., & Wagener, D. (1976). Toward the estimation of population growth
in Old World prehistory. In E.B. Zubrow (Ed.), Demographic Anthropology: Quantitative approaches
(pp. 27– 61). Albuquerque: University of New Mexico Press.
Ashton, N. (1991). Siphnos: Ancient towers B.C. Athens: Eptalophos Abee.
Barber, R.L.N. (1987). The Cyclades in the Bronze Age. London: Duckworth.
Betancourt, P.P., & Hope Simpson, R. (1992). The agricultural system of Bronze Age Pseira. Cretan
Studies, 3, 47– 54.
Betancourt, P.P., Goldberg, P., Hope Simpson, R., & Vitaliano, C.J. (1990). Excavations at Pseira: The
evidence for the Theran eruption. In D. Hardy (Ed.), Thera and the Aegean World III (Vol. 3, pp. 96–
99). London: The Thera Foundation.
Bintliff, J.L. (1977). Natural environment and human settlement in Prehistoric Greece. British Archae-
ological Reports, Supplementary Series 28. Oxford: British Archaeological Reports.
Bintliff, J.L. (1992). Erosion in Mediterranean lands: A reconsideration of pattern, process and meth-
odology. In M. Bell & J. Boardman (Eds.), Past and present soil erosion. Archaeological and geo-
graphical perspectives (pp. 125– 132). Oxbow Monograph 22, Oxford: Oxbow Books.
Blackman, D.J. (1997). Archaeology in Greece 1996– 97. Archaeological Reports, 43, 1– 125.
Bottema, S. (1985). Palynological investigations in Greece with special reference to pollen as an indi-
cator of human activity. Palaeohistoria, 24, 257– 289.
Bradford, J. (1956). Fieldwork on aerial discoveries in Attica and Rhodes. Antiquaries Journal, 36, 172–
180.
Bridges, E.M. (1978). World soils, Cambridge: Cambridge University Press.
Broodbank, C. (1989). The longboat and society in the Cyclades in the Keros-Syros culture. American
Journal of Archaeology, 93, 319– 337.
Brückner, H. (1990). Changes in the Mediterranean ecosystem during antiquity— a geomorphological
approach as seen in two examples. In S. Bottema, G. Entjes-Nieborg, & W. van Zeist (Eds.), Man’s
role in the shaping of the Eastern Mediterranean landscape (pp. 127– 138). Rotterdam: Balkema.
Brunet, M. (1990). Terrasses de cultures antiques: l’example de Délos, Cyclades. Mediterranée, 3, 5– 11.
Brunet, M. (1996). La campagne délienne. In R. Etienne & M.-D. Nenna (Eds.), L’espace grec. Cent
cinquante ans de fouilles de l’École française d’Athenes (pp. 59– 65). Baumes-les-Dames: Fayard.
Bullock, P., Fedoroff, N., Jongerius, A., Stoops, G., & Tursina, T. (1985). Handbook for soil thin section
description. Wolverhampton: Waine Research.
Cameron, M. (1972). Appendix IV. The plasters. In P.M. Warren (Ed.), Myrtos: An Early Bronze Age
settlement in Crete (pp. 305– 314). British School at Athens, Supplementary Volume 7. London:
Thames and Hudson.
Cavanagh, W., Crouwel, J., Catling, R., & Shipley, G. (1996). Continuity and change in a Greek rural
landscape: The Laconia Survey (Vol. II0, The British School at Athens, Supplementary Vol. 27).
London: British School at Athens.
Cherry, J.F. (1979). Four problems in Cycladic prehistory. In J.L.Davis & J.F.Cherry (Eds.), Papers in
Cycladic Prehistory (22– 47), Monograph 14. Los Angeles: Institute of Archaeology, University of
California. short
standard
Cherry, J.F. (1982). A preliminary definition of site distribution on Melos. In C. Renfrew & M. Wagstaff
(Eds.), An island polity: The archaeology of exploitation in Melos (pp. 10– 23). Cambridge: Cambridge
University Press.
Cherry, J.F. (1983). Frogs round the pond: Perspectives on current archaeological survey projects in
the Mediterranean region. In D. Keller & D. Rupp (Eds.), Archaeological survey in the Mediterranean
area (pp. 375– 416). British Archaeological Reports, International Series 155. Oxford: British Ar-
chaeological Reports.
Cherry, J.F. (1994). Regional survey in the Aegean: the “New Wave” (and after). In N. Kardulias (Ed.),
Beyond the site. Regional studies in the Aegean area (pp. 91– 112). London: University Press of
America.
Cherry, J.F., Davis, J.L., & Mantzourani, E. (1991a). Landscape archaeology as long-term history:
Northern Keos in the Cycladic Islands. Monumenta Archaeologica 16. Los Angeles: University of
California Press.
Cherry, J.F., Davis, J.L., & Mantzourani, E. (1991b). Prehistoric northern Keos: Analysis and interpre-
tation of the survey finds. In J.F. Cherry, J.L. Davis, & E. Mantzourani (Eds.), Landscape Archaeology
as long-term history: Northern Keos in the Cycladic Islands (pp. 217– 232). Monumenta Archaeolo-
gica 16. Los Angeles: University of California Press.
Clark, M., & Small, J. (1982). Slopes and weathering. Cambridge: Cambridge University Press.
Coles, J.M. (1973). Archaeology by experiment. London: Hutchinson.
Courty, M.A., Goldberg, P., & Macphail, R.I. (1989). Soils and micromorphology in archaeology. Cam-
bridge: Cambridge University Press.
Dalongeville, R., & Renault-Miskovsky, J. (1993). Paysages passés et actuels de l’ı̂le de Naxos. In R.
Dalongeville & G. Rougement (Eds.), Recherches dans les Cyclades (pp. 9– 57). Lyon: Maison de
l’Orient Méditerranéen.
Davidson, D. (1976). Processes of tell formation and erosion. In D. Davidson & M.L. Shackley (Eds.),
Geoarchaeology (pp. 255– 265). London: Duckworth.
Davidson, D. (1978). Aegean soils during the second millennium B.C. with reference to Thera. In C.
Doumas (Ed.), Thera and the Aegean world (Vol. 1, pp. 725– 739). London: The Thera Foundation.
Davidson, D. (1980). Erosion in Greece during the first and second millennia B.C. In R.A. Cullingford,
D. Davidson, & J. Lewin (Eds.), Timescales in geoarchaeology (pp. 143– 158). Chichester: Wiley.
Davidson, D., & Tasker, C. (1982). Geomorphological evolution during the late Holocene. In C. Renfrew
& M. Wagstaff (Eds.), An island polity. The archaeology of exploitation in Melos (pp. 82– 94). Cam-
bridge: Cambridge University Press.
Davis, J.L. (1992). Review of Aegean prehistory I: The islands of the Aegean. American Journal of
Archaeology, 96, 699– 756.
Doumas, C. (1977). Early Bronze Age burial habits in the Cyclades, Göteborg: Paul Astroms Förlag.
Doumas, C. (1988). EBA in the Cyclades: Continuity or discontinuity? In E.B. French & K.A. Wardle
(Eds.), Problems in Greek Prehistory (pp. 21– 30). Bristol: Bristol Classical Press.
Dümmler, F. (1886). Mitteilungen von den griechischen Inseln. Athenische Mitteilungen, 11, 15– 46.
Érard-Cerceau, I., Fotou, V., Psychoyos, O., & Treuil, R. (1993). Prospection archèologique a Naxos
(région nord-ouest). In R. Dalongeville & G. Rougement (Eds.), Recherches dans les Cyclades (pp.
59– 96). Lyon: Maison de l’Orient Méditerranéen.
Foxhall, L. (1996). Feeling the earth move: Cultivation techniques on steep slopes in classical antiquity.
In G. Shipley & J. Salmon (Eds.), Human landscapes in Classical Antiquity environment and culture
(pp. 44– 67). London: Routledge.
Gifford, J., & Reese, D. (1995). The physical geology of the western Mesara and Kommos. In J.W. Shaw
& M.C. Shaw (Eds.), Kommos I. The Kommos Region and houses of the Minoan town (pp. 30– 90).
Princeton: Princeton University Press.
Greig, J., & Turner, J. (1974). Some pollen diagrams from Greece and their archaeological significance.
Journal of Archaeological Science, 1, 177– 194.
Halstead, P.L.J. (1994). The north– south divide: Regional paths to complexity in prehistoric Greece. In short
standard
C. Mathers & S. Stoddart (Eds.), Development and decline in the Mediterranean Bronze Age (pp.
195– 219). Sheffield Archaeological Monograph 8. Sheffield: J.R. Collis Publications.
Heilmann, P.G.F. (1972). On the formation of red soils in the Lower Gati Basin (S. Italy). Publication of
the International Soil Museum (ISM) Utrecht. Enshede: International Institute for Aerial Survey and
Earth Sciences.
Isager, S., & Skydsgaard, J. (1992). Ancient Greek agriculture. An introduction. London: Routledge.
James, P., Mee, C., & Taylor, G. (1994). Soil erosion and the archaeological landscape of Methana,
Greece. Journal of Field Archaeology, 21, 395– 416.
Jameson, M., Runnels, C., & van Andel, T.H. (1994). A Greek countryside. The Southern Argolid from
Prehistory to the present day. Stanford: Stanford University Press.
Johnson, G.A. (1977). Aspects of regional analysis in archaeology. Annual Review of Anthropology, 6,
479– 508.
Keller, D. (1985). Archaeological survey in Southern Euboea, Greece: A reconstruction of human activity
from Neolithic times through the Byzantine Period, Ph.D. Dissertation, Bloomington: Indiana Uni-
versity.
Kirkby, A., & Kirkby, M. J. (1976). Geomorphic processes and the surface survey of archaeological sites
in semi-arid areas. In D. Davidson & M. L. Shackley (Eds.), Geoarchaeology (pp. 229– 253). London:
Duckworth.
Kwaad, F.J.P.M., & Mücher, H.J. (1977). The evolution of soils and slope deposits in the Luxembourg
Ardennes. Geoderma, 17, 1– 37.
Kwaad, F.J.P.M., & Mücher, H.J. (1979). The formation and evolution of colluvium on arable land in
northern Luxembourg. Geoderma, 22, 173– 192.
Laxton, R., & Cavanagh, W. (1995). The rank-size dimension and the history of site structure from survey
data. Journal of Quantitative Anthropology, 5, 327– 358.
Limbrey, S. (1975). Soil science and archaeology. London: Academic Press.
Lohmann, H. (1993). Atene. Forschungen zu Seidlungs- und Wirtschaftsstruktur des klassischen Attika.
Köln: Böhlau Verlag.
Macphail, R.I. (1987). A review of soil science in archaeology in England. In H.C.M. Keeley (Ed.), En-
vironmental archaeology: A regional review (Vol. II, pp. 332– 377). HBMC Occasional Paper No 1,
HBMC: London.
Macphail, R.I. (1992). Soil micromorphological evidence of ancient soil erosion. In J. Boardman & M.
Bell (Eds.), Past and present soil erosion (pp. 197– 215). Oxbow Monograph 22, Oxford: Oxbow
Books.
Manning, S. (1995). The absolute chronology of the Aegean Early Bronze Age. Archaeology, radiocarbon
and history. Sheffield: Sheffield Academic Press.
Marangou, L. (1984). Evidence for the Early Cycladic period on Amorgos. In L. Fitton (Ed.), CyclADica:
studies in memory of N.P. Goulandris (pp. 99– 115). London: British Museum.
Marangou, L. (1994). Nees martyries gia ton KyklADiko politismo stin Amorgo. In X. Tzouvara-Souli, A.
Vlachopoulou-Oikonomou, & K. Gravani-Katsiki (Eds.), Fhigos: Timitikos tomos gia ton kathigiti
sotiri dakari (pp. 467– 488). Ioannina: University of Ioannina.
Marangou, L., Renfrew, C., & Doumas, C. (in preparation). Markiani: An early fortified site on Amorgos.
Athens: Archaiologiki Etaireia.
Mee, C., et al. (1997). A rough and rocky place. The landscape and settlement history of the Methana
Peninsula, Greece. Liverpool: Liverpool University Press.
Moody, J., & Grove, A.T. (1990). Terraces and enclosure walls in the Cretan landscape. In S. Bottema,
G. Entjes-Nieborg, & W. van Zeist (Eds.), Man’s role in the shaping of the Eastern Mediterranean
landscape (pp. 183– 91). Rotterdam: Balkema.
Murphy, C.P. (1986). Thin section preparation of soils and sediments. Berkhamsted: AB Academic.
Nevett, L. (1988). An island’s polities: A diachronic investigation of settlement on Melos. M.Phil. Dis-
sertation, Cambridge: Cambridge University.
Parsons, M., & Gifford, J. (1995). Soil and land use studies at Kommos. In J.W. Shaw & M.C. Shaw (Eds.), short
standard
Kommos I. The Kommos region and houses of the Minoan town (pp. 292– 324). Princeton: Princeton
University Press.
Pope, K.O., & van Andel, T.H. (1984). Late Quaternary alluviation and soil formation in the Southern
Argolid: Its history, causes, and archaeological implications. Journal of Archaeological Science, 11,
281– 306.
Rackham, O. (1978). The flora and vegetation of Thera and Crete before and after the great eruption.
In C. Doumas (Ed.), Thera and the Aegean World (Vol. 1, pp. 755– 764). London: The Thera Foun-
dation.
Rackham, O. (1990a). Observations on the historical ecology of Santorini. In D. Hardy (Ed.), Thera and
the Aegean World III (Vol. 2, pp. 384– 391). London: The Thera Foundation.
Rackham, O. (1990b). The greening of Myrtos. In S. Bottema, G. Entjes-Nieborg, & W. van Zeist (Eds.),
Man’s role in the shaping of the Eastern Mediterranean landscape (pp. 341– 348). Rotterdam: Bal-
kema.
Rackham, O., & Moody, J. (1992). Terraces. In B. Wells (Ed.), Agriculture in Ancient Greece (pp. 123–
130). Stockholm: Aströms Förlag.
Reid, I., & Frostick, L. (1985). Arid zone slopes and their archaeological materials. In A. Pitty (Ed.),
Themes in geomorphology (pp. 141– 157). London: Croom Helm.
Renfrew, C. (1972). The Emergence of Civilisation: The Cyclades and the Aegean in the Third Millennium
B.C. London: Methuen.
Renfrew, C. (1984). From Pelos to Syros: Kapros grave D and the Kampos Group. In J.A. MacGillivray
& R.L.N. Barber (Eds.), The Prehistoric Cyclades. Contributions to a workshop on Cycladic chro-
nology (pp. 41– 54). Edinburgh: University of Edinburgh, Department of Classical Archaeology.
Rick, J.W. (1976). Downslope movement and archaeological intrasite spatial analysis. American Antiq-
uity, 41, 133– 144.
Rosen, A.M. (1986). Cities of clay. The geoarchaeology of Tells. London: University of Chicago Press.
Rutter, J. (1993). Review of Aegean prehistory II: The prepalatial Bronze Age of the southern and central
Greek mainland. American Journal of Archaeology, 97, 745– 797.
Shaw, J.W. (1973). Minoan architecture: Materials and techniques. Annuario della Scuola Archeologica
di Atene, 49.
Shay, C.T., & Shay, J.M. (1995). The modern flora and plant remains from Bronze Age deposits at
Kommos. In J.W. Shaw & M.C. Shaw (Eds.), Kommos I. The Kommos region and houses of the
Minoan town (pp. 91– 162). Princeton: Princeton University Press.
Shelmerdine, C. (1997). The palatial Bronze Age of the southern and central Greek mainland. American
Journal of Archaeology, 101, 537– 585.
Slager, S., & van de Wetering, H.T.J. (1977). Soil formation in archaeological pits and adjacent loess
soils in southern Germany. Journal of Archaeological Science, 4, 259– 267.
Sotirakopoulou, P. (1993). The chronology of the “Kastri group” reconsidered. Annual of the British
School at Athens, 88, 5– 20.
Spaargaren, O.C. (1979). Weathering and soil formation in a limestone area near Pastena (Fr., Italy),
Publicates van het Fysisch Geografisch en Bodemkundig Laboratorium van de Universiteit van Am-
sterdam, No. 30, Amsterdam: University of Amsterdam.
Stein, J., & Rapp, G., Jr. (1978). Archaeological geology of site. In G. Rapp, Jr. & S. Aschenbrenner
(Eds.), Excavations at Nichoria in Southwest Greece. Vol. I: Site, environs, and techniques (pp. 234–
257). Minneapolis: University of Minnesota Press.
Thornes, J., & Gilman, A. (1983). Potential and actual erosion around archaeological sites in south east
Spain. Catena Supplement, 4, 91– 113.
Tsountas, C. (1898). Kukladika. Archaiologiki Ephimeris, 1898, 137– 212.
Turner, J. (1978). The vegetation of Greece during prehistoric times: The palynological evidence. In C.
Doumas (Ed.), Thera and the Aegean World (Vol. 1 pp. 765– 773). London: The Thera Foundation.
van Andel, T.H. (1998). Palaesols, red sediments and the Old Stone Age in Greece. Geoarchaeology,
13(4), 361– 390.
van Andel, T.H., & Hansen, J. (1987). Evolution of the Franchthi landscape. In T.H. van Andel & S.B. short
standard
Sutton(Eds.), Landscape and people of the Franchthi Region (pp. 55– 62). Bloomington: University
of Indiana Press.
van Andel, T.H., & Runnels, C. (1995). The earliest farmers in Europe. Antiquity, 69, 481– 500.
van Andel, T.H., & Zangger, E. (1990). Landscape stability and destabilisation in the prehistory of Greece.
In S. Bottema, G. Entjes-Nieborg, & W. van Zeist (Eds.), Man’s role in the shaping of the Eastern
Mediterranean landscape (pp. 139– 157). Rotterdam: Balkema.
van Andel, T.H., Runnels, C., & Pope, K. (1986). Five thousand years of land use and abuse in the
Southern Argolid, Greece. Hesperia, 55, 103– 128.
Vita-Finzi, C. (1969). The Mediterranean valleys. Cambridge: Cambridge University Press.
Wagstaff, J.M. (1981). Buried assumptions: Some problems in the interpretation of the “Younger Fill”
raised by recent data from Greece. Journal of Archaeological Science, 8, 247– 264.
Wagstaff, M., & Cherry, J.F. (1982) Settlement and resources. In C. Renfrew & M. Wagstaff (Eds.), An
island polity: The archaeology of exploitation in Melos (pp. 246– 263). Cambridge: Cambridge Uni-
versity Press.
Wagstaff, M., & Gamble, C. (1982). Island resources and their limitations. In C. Renfrew & M. Wagstaff
(Eds.), An island polity: The archaeology of exploitation in Melos (pp. 95– 105). Cambridge: Cam-
bridge University Press.
Wagstaff, M., Augustson, S., & Gamble, C. (1982). Alternative subsistence strategies. In C. Renfrew &
M. Wagstaff (Eds.), An island polity: The archaeology of exploitation in Melos (pp. 172– 180). Cam-
bridge: Cambridge University Press.
Wainwright, J. (1992). Assessing the impact of erosion on semi-arid archaeological sites. In M. Bell & J.
Boardman (Eds.), Past and present soil erosion. Archaeological and geographical perspectives (pp.
227– 241). Oxbow Monograph 22. Oxford: Oxbow Books.
Warren, P.M. (1972). Myrtos: An Early Bronze Age settlement in Crete,. British School at Athens, Sup-
plementary Volume 7. London: Thames and Hudson.
Watrous, L.V., Hatzi-Vallianou, D., Pope, K., Mourtzas, N., Shay, J., Shay, C.T., Bennet, J., Tsoungarakis,
D., Angelomati-Tsoungarakis, E., Vallianos, C., & Blitzer, H. (1993). A survey of the western Mesara
plain in Crete. Preliminary report on the 1984, 1986, and 1987 field seasons. Hesperia, 62, 191– 248.
Wells, B., & Runnels, C. (1996). The Berbati-Limnes archaeological survey 1988– 1990. Stockholm: Skrif-
ter Athenensia.
Wells, B., Runnels, C., & Zangger, E. (1990). The Berbati-Limnes archaeological survey. The 1988 season.
Opuscula Atheniensia, 18, 207– 238.
Whitelaw, T. (1991a). The ethnoarchaeology of recent rural settlement and land use in northwest Keos.
In J. Cherry, J.L. Davis, & E. Mantzourani (Eds.), Landscape archaeology as long-term history:
Northern Keos in the Cycladic Islands (pp. 403– 454). Monumenta Archaeologica 16. Los Angeles:
University of California Press.
Whitelaw, T. (1991b). Investigations at the Neolithic sites of Kephala and Paoura. In J. Cherry, J.L. Davis,
& E. Mantzourani (Eds.), Landscape archaeology as long-term history: Northern Keos in the Cycladic
Islands (pp. 199– 216). Monumenta Archaeologica 16. Los Angeles: University of California Press.
Whitelaw, T. (1994). An ethnoarchaeological study of rural land-use in north-west Keos: Insights and
implications for the study of past Aegean landscapes. In P. Doukelis & L. Mendoni (Eds.), Structures
rurales et sociétés antiques (pp. 163– 186). Paris: Les Belles Lettres.
Whitelaw, T. (in press). Colonisation and competition: Settlement pattern changes in Northwest Keos.
In L. Mendoni & P. Doukelis (Eds.), Proceedings of the International Symposium: Keos and Kythnos:
Historical and Archaeological Research. Athens: National Hellenic Research Foundation.
Wilkinson, T.J., & Duhon, S.T. (1990). Franchthi Paralia. The sediments, stratigraphy, and offshore in-
vestigations. Bloomington: Indiana University Press.
Wright, H.E., Jr. (1972). Vegetation history. In W. McDonald & G. Rapp Jr. (Eds.), The Minnesota Mes-
senia Expedition (pp. 188– 199). Minneapolis: University of Minnesota Press.
Young, J. (1956). Ancient towers on the island of Siphnos. American Journal of Archaeology, 60, 51–
55. short
standard
Zangger, E. (1992). Neolithic to present soil erosion in Greece. In M. Bell & J. Boardman (Eds.), Past
and present soil erosion. Archaeological and geographical perspectives (pp. 133– 148). Oxbow Mono-
graph 22. Oxford: Oxbow Books.
Zangger, E. (1993). The geoarchaeology of the Argolid. Berlin: Gebr. Mann Verlag.
Zangger, E. (1994). The island of Asine: A palaeogeographic reconstruction. Opuscula Atheniensia, 20,
221– 239.
short
standard