вход по аккаунту


Geophysical surveys at King Lobengula's Palace KoBulawayo Zimbabwe.

код для вставкиСкачать
Archaeological Prospection
Archaeol. Prospect. 12, 31–49 (2005)
Published online 17 December 2004 in Wiley InterScience ( DOI: 10.1002/arp.241
Geophysical Surveys at King
Lobengula’s Palace KoBulawayo,
GSB Prospection Ltd, Cowburn Farm, Market Street, Thornton, Bradford, West Yorkshire
BD13 3HW, UK
Cambria Archaeology (Dyfed Archaeological Trust), The Shire Hall, Carmarthen Street,
Llandeilo, Carmarthenshire,Wales SA19 6AF, UK
This report covers the application of magnetic survey, primarily using a magnetic susceptibility field
instrument, at the historically attested site of KoBulawayo, Zimbabwe. The approximate position of
the site was known before the geophysical survey took place; it was believed to comprise a Royal Enclosure, a surrounding open space possiblyused asa militaryparade groundand the Commoner/Peripheral Settlement of Lobengula, King of the Ndebele.Occupation at the site was short lived and after
only11yearsthe capital of the Ndebele state was destroyed by fire in1881. A pilot survey was undertakenin1994 to assessthe suitabilityof survey techniques.Consequently, a second, more extensive survey was carried out in late 1996 and early 1997 with the intention of delimiting the Royal Enclosure.
Furtherperiodsofdatacollectiontookplacelaterin1997 andin1998.Thisreportdescribesthemethods
used and theinterpretation ofthe geophysicalresultsinthe context ofthe understanding andmanagement of this important historical site. Additionally, some of the results of the geophysical work have
beentested byexcavation anda discussion ofthe correlation betweenthese data setsis also reported
in this article.Copyright 2004 JohnWiley & Sons,Ltd.
Key words: magnetic susceptibility; Zimbabwe; KoBulawayo; Ndebele; King Lobengula
During the short period AD 1870–1881 a low oval
hill, approximately 16.5 km to the southeast
of Bulawayo, was ultized as the capital of
the Ndebele, or Matabele state under King
Lobengula (Figure 1). Although the Ndebele
state could be described as multi-ethnic, the
military and political organization is derived, in
part, from its Zulu heritage. For reasons that are
far from clear the king moved his palace and
* Correspondence to: C. Gaffney, GSB Prospection Ltd,
Cowburn Farm, Market Street, Thornton, Bradford, West
Yorkshire BD13 3HW, UK.
Copyright # 2004 John Wiley & Sons, Ltd.
entourage to modern Bulawayo in 1881. He
ordered the destruction of KoBulawayo and the
site was razed to the ground by fire. Through this
act of deliberate destruction Lobengula signified
the end of the palace and entered the site into the
archaeological record. Significantly, the act of
burning can create dramatic increases in the
magnetic properties of the buried soil, which
are important for location using geophysical
means and the archaeological interpretation of
the site (see Linford and Canti, 2001; Peters et al.,
2001; Weston, 2002).
The layout and general architecture of
KoBulawayo can be reconstructed from three
principal types of source: archaeological remains,
pictorial sources and written descriptions (the
latter two by missionaries, traders and other
Received 23 August 2004
Accepted 28 August 2004
C. Gaffney, G. Hughes and J. Gater
Figure1. General location of KoBulawayo, Zimbabwe.
European visitors). The basic layout of the settlement was recorded during a conventional survey
carried out in December 1993 by a team from
The National Museums and Monuments of
Zimbabwe (Figure 2). They defined the area of
the whole settlement as a circular ring of structures, primarily clay round house floors, of approximately 500 m in diameter. This has been
termed the Peripheral Settlement and surrounds
an open central area or Main Enclosure that was
possibly reserved for cattle and used as a military
parade ground. Within the Main Enclosure and
slightly to the west of centre, is the Royal Enclosure, which was believed to be separated from
the Main Enclosure by a wooden palisade. At the
time of the original survey in 1993 a number of
buildings were visible within the Royal Enclosure,
including three stone or brick built structures and
three clay hut floors. A number of written and
pictorial depictions of the Royal Enclosure exist
from the time of the settlement but they provide a
highly varied view of the detail.
Copyright # 2004 John Wiley & Sons, Ltd.
The geophysical survey formed part of a programme of archaeological research and investigation that was intended to produce information
to enhance the development of KoBulawayo as a
heritage centre by the National Museums and
Monuments of Zimbabwe. The work was jointly
undertaken by a team from GSB Prospection,
Birmingham Archaeology and the National
Museums and Monuments of Zimbabwe; the
geophysical survey was supported by grants
from the Arts and Humanities Research Board,
the British High Commission in Harare and the
British Academy. Non-destructive survey techniques were regarded as an important first stage
in designing a targeted programme of excavation
while at the same time ensuring that as much of
the archaeology as possible was preserved in situ.
The definition of the undisturbed and unseen
archaeology is important for testing the accuracy
of the sources of information, for identifying
zones of archaeology for preservation and to
target any subsequent excavation.
Archaeol. Prospect. 12, 31–49 (2005)
King Lobengula’s Palace
Figure 2. Information recorded by conventional survey prior to the geophysical work.
Magnetic susceptibility survey
The major investigative technique used at this
site was magnetic susceptibility (MS) survey.
Magnetic susceptibility is described as an
‘active’ method as it requires a magnetic field
Copyright # 2004 John Wiley & Sons, Ltd.
to be induced in the soil by a pulse of low
frequency electromagnetic energy (Thompson
and Oldfield, 1986; Clark, 1996; Gaffney and
Gater, 2003). The instrument is ‘zeroed’ holding
the instrument above the head and measurements are taken with the coil pressed against
the ground (Figure 3: see Dearing (1999) for a
description of field practice). The size of the
Archaeol. Prospect. 12, 31–49 (2005)
C. Gaffney, G. Hughes and J. Gater
Figure 3. Magnetic susceptibility data collection at KoBulawayo.
induced response is related to the amount and
type of magnetic minerals within the volume of
the search coil and it is often described as the
ability of the soil to be magnetized, i.e. the
magnetic susceptibility. The penetration depth
for the instrument used is approximately 0.2 m. It
is known that MS can be increased anthropogenically, either through depositional activity or
through burning. The latter is particularly important if the temperature is elevated beyond the
Curie point, which is about 600 C for typical soil
minerals, as the enhancement then can be
regarded as permanent.
Although the equipment, a Bartington Magnetic Susceptibility Meter with a MS2D field coil,
is a standard geophysical instrument, the collection strategy was unusual owing to the high
sample density. Traditionally MS data have
been collected over large areas with the intention
of locating unknown buried archaeology. If the
sample interval is of the order of 10 m then it is
possible to cover in excess of 5 ha in a single day.
However, the data that are collected are very
coarse and the patterning that is produced often
reflects modern and/or agricultural activity
rather than archaeology. At KoBulawayo the
sample density was increased to 1 per m2, which
Copyright # 2004 John Wiley & Sons, Ltd.
it was hoped would allow for detailed analysis of
individual features. Although this is normally
beyond the depth penetration/resolution of the
MS2D field coil, the shallow depth of the soil at
KoBulawayo allowed some hope for this aim
(Figure 4).
At the time of the project, magnetic susceptibility survey at 1 m intervals was virtually
unheard of, except for small-scale surveys of
a few hundred square metres. The only previous
work known to have been undertaken in
Zimbabwe was by a Swedish group who were
working at 50 m intervals.
Fluxgate magnetometer survey
The second geophysical technique used at
KoBulawayo was fluxgate magnetometry (FM)
(or gradiometry). This is the most widely used
non-invasive technique for investigating shallow
archaeological remains (Gaffney and Gater,
2003). The popularity of the technique is due to
the speed with which large areas can be investigated. It was understood that previous attempts
to undertake fluxgate survey in Zimbabwe by the
University of Zimbabwe had not been particulary successful.
Archaeol. Prospect. 12, 31–49 (2005)
King Lobengula’s Palace
Figure 4. Excavation of a burnt clay house floor in the Northern Periphery. Note the shallow depth of soil above the floor.
The system uses two ‘fluxgate’ sensors to
measure modifications of the vertical component
of Earth’s magnetic field that are brought
about by a buried magnetic ‘object’. In this case
Geoscan Research FM36 instruments were used
with sensors separated by a vertical distance
of 0.5 m.
A fluxgate system effectively measures
changes in magnetic susceptibility and therefore
the enhancement of soil discussed above is similarly important here. Where this technique differs in practice from MS survey is that a FM
system measures down to a depth of ca. 1 m.
Although it is clear that this instrument can
detect archaeology buried at a greater depth
then the MS system, it is also true that the FM
data can potentially contain much higher levels
of noise owing to near-surface geological/pedological variation. There is also a variation in
response to similar features depending on the
position on Earth. In practice, as the instruments
measure local differences in the vertical component of the Earth’s magnetic field, the magnitude
of an anomaly is reduced near the Equator,
where this component is small. Fortunately this
is not too much of a problem as the shape of
response is still similar to those found at higher
latitudes (Heathcote, 1983).
Copyright # 2004 John Wiley & Sons, Ltd.
Common survey strategy for MS and FM
During each survey episode the area to be surveyed was subdivided into 20 20 m blocks. The
blocks were then subdivided further to allow
marked lines to be placed at 1 m intervals and
oriented east–west along the grid. The FM data
were captured directly into the on-board memory and transferred to a notebook computer
using the Geoscan Research software (Geoplot
3). The MS data were captured via a Psion
Organiser during the Pilot Survey. However,
the excessive heat during the survey resulted in
the failure of the automated logging system and
therefore for subsequent fieldwork the data were
hand-logged and typed daily into Geoplot. The
data were subsequently converted into format
for other display packages, most notably Surfer,
which is where the images used in this report
were created.
Initial survey (1994)
A pilot survey was undertaken toward the end of
1994. The aim of the survey was to assess the two
Archaeol. Prospect. 12, 31–49 (2005)
C. Gaffney, G. Hughes and J. Gater
Figure 5. Data collected during the initial survey in 1994.The box on the magnetometer image shows the position of the susceptibility survey.
magnetic techniques (FM and MS Survey) and
assess the suitability of the land for additional
geophysical techniques. The iron content of the
local soils is highly variable, but to some extent,
termite activity and erosion has probably had a
homogenizing effect (Limbrey, 1995). In the
weathered rocks and the red and reddish brown
soils, the predominant iron oxide is haematite, a
product of a lengthy soil development under
high temperatures and a seasonally dry climate.
The main reason for using magnetic techniques
was the nature of the abandonment of the site, i.e.
the reported use of fire to destroy the enclosure
and the accompanying settlement. Fortunately, a
number of exposed floors, perhaps parts of
dwellings or storage areas, were visible on the
surface and in areas devoid of scrub. These were
targeted using both techniques, along with
apparently ‘blank’ areas. Significant anomalies
were found (Figure 5).
In fact the MS data were particularly clear, with
impressive correlation with the observed position
of the evidently burnt floors. Background susceptibility measurements were found to be in the order
of 50–150 SI units, and the maximum values over
two of the huts were 675 and 924 SI units. Additional considerations, such as the shallow soils and
the fact that the equipment is relatively robust, led
to the use of the MS as the primary investigative
tool. The FM survey produced far less convincing
results—this was believed to be partly the result of
modern notice boards and the nature of the
archaeology itself (Gater, 1994).
Copyright # 2004 John Wiley & Sons, Ltd.
The second survey (November 1996
to March 1997)
The aim of this phase of the geophysical study
was to investigate the central part of the site, i.e.
the Royal Enclosure (Figure 6). It was hoped that
the surface features mapped by topographic
means would be supplemented by the location
of buried clay house floors, fence and palisade
lines, ditches and other potential archaeological
features (Gaffney and Hughes, 1997). Together
these two strands of evidence would be valuable
in assessing the pictorial evidence for the site as
well as providing a firm base for subsequent
excavation. This second survey was undertaken
in two stages in November 1996 and March 1997.
Part of the area investigated was overgrown and
the vegetation had to be cleared in front of the
magnetic survey. This clearance also allowed a
topographic survey to be undertaken.
Magnetic susceptibility
The initial susceptibility survey was conducted on a 1 1 m grid and resampling at a
lower sample intensity suggested that significant
feature definition would be lost if a lower intensity was implemented. It was therefore decided
to continue the survey using the same sample
interval. The area covered during this data collection episode can be seen in Figure 6, although
the data are displayed alongside those from the
third and fourth surveys (Figures 7 and 8). Visible evidence on the ground indicated that a
Archaeol. Prospect. 12, 31–49 (2005)
King Lobengula’s Palace
Figure 6. Diagram showing the extent and phases of geophysical work at KoBulawayo.
number of contemporary buildings were located
within the postulated Royal Enclosure. These
structures included two brick buildings on stone
foundations, three clay surfaces representing the
location of former roundhouses and the upstanding remains of the King’s stone-built Wagon
Shed. Only the area of the Wagon Shed was
unavailable for survey. A fourth exposed clay
floor was located to the northwest of the Royal
Enclosure and was associated with an area of low
stone walling.
The data are remarkable in their clarity
giving an extraordinarily clear view of the Royal
Copyright # 2004 John Wiley & Sons, Ltd.
Enclosure (Gaffney and Hughes, 1997). Perhaps
the most striking element is the clear presence for
a curved boundary feature delimiting the enclosure. The route of this enclosure ditch is well
defined along its northern, western and southern
boundaries. Of great interest is the eastern edge
of the enclosure, the interpretation of which is
complex. From historical accounts it was
expected that the entrance would be in this
zone, although nothing was known of the precise
arrangement. However, there were strong indications that the entrance was associated with a
circular cattle byre. It seems that access into the
Archaeol. Prospect. 12, 31–49 (2005)
C. Gaffney, G. Hughes and J. Gater
Figure 7. Various images of the complete magnetic susceptibility data set.
Royal Enclosure was either through this compound or around it.
The interior of the enclosure contains a wealth
of detail. The most visually apparent is the
collection of subcircular anomalies, which must
be dwellings, i.e. houses or associated structures.
In addition to the three partially upstanding
rectangular structures and the three partially
exposed house floors, the MS survey has added
Copyright # 2004 John Wiley & Sons, Ltd.
a further 17 possible structures. Although at first
sight the arrangement seems to be random, we
know from the pictorial evidence of the enclosure and from oral tradition that Ndebele homesteads were highly organized. It is believed that
the weak sublinear anomalies shown in part in
the MS data and partially in the FM data indicate
minor, but symbolically significant, demarcations of distinct settlement units (Hughes and
Archaeol. Prospect. 12, 31–49 (2005)
King Lobengula’s Palace
Figure 8. Summary greyscale showing high-pass filtered data.
Muringaniza, 2003, p. 129). It is possible that they
relate to the space allocated to individual royal
A zone of low magnetic susceptibility can be
seen within the western boundary of the Royal
Enclosure, suggesting a specific or unique use in
this part of the site. In this area a ring of discrete
stone structures was mapped on the surface of
the ground and interpreted as piles for grain
stores adjacent to the enclosure ditch.
The dramatic change in magnetic susceptibility near the northern edge of the Royal Enclosure is believed to indicate a change in geology.
Superimposed on this raised background are a
number of discrete anomalies that are believed to
Copyright # 2004 John Wiley & Sons, Ltd.
be anthropogenic; it is this area, to the north and
east of the Royal Enclosure, that was surveyed
during the third and fourth stages of the geophysical investigation of KoBulawayo.
Fluxgate magnetometer data
The results of the gradiometer data (Figure 9)
are described in more detail in Gaffney and
Hughes (1997). The technique has provided vital
information on two issues. First, the data confirm
the presumed geological change identified in the
MS survey; in the case of the gradiometer
the data are more varied to the north of the
line, although it is fortunate that the main enclosure lies within the low background noise.
Archaeol. Prospect. 12, 31–49 (2005)
C. Gaffney, G. Hughes and J. Gater
Figure 9. Gradiometer data collected over the Royal enclosure.
Second, the data within the Royal Enclosure
appear to provide evidence for minor divisions
in the form of possible ditches or palisades.
Third and fourth surveys (July 1997
to August 1998)
The aim of this work was to extend the areas
covered to include a significant sample of the
peripheral areas of the settlement. Only MS
survey was undertaken during this phase of the
work. The survey now included the whole of the
northern sector of the site from an area of visible
low stone walling to the northwest of the Royal
Enclosure to a line of stone cairns (interpreted as
grain bin foundations) to the northeast of the
modern road leading from the site to the modern
town of Bulawayo. The total area of the MS
survey covered approximately one third of the
entire settlement; a survey area of approximately
53 600 m2 (Figure 6).
In the northern and western parts of the newly
surveyed areas a distinct arc of anomalies can be
identified, including a number of subcircular
anomalies comparable to those identified within
the area of the Royal Enclosure. It is noticeable
that this band of possible settlement features
correlates with the band of visible stone cairns
and grain bin foundations recorded by the topographic survey. Together, the geophysics and the
topographic survey appear to provide a clear
indication of the character and density of strucCopyright # 2004 John Wiley & Sons, Ltd.
tures within this northern peripheral area of
settlement. It is also interesting to note that the
two surveys complement each other. The geophysics appears to be identifying buried clay
floors, whereas the topographic survey is recording visible stone structures that are not so apparent in the magnetic data (Figures 10 and 11).
A noticeable contrast between this peripheral
area of settlement and the Royal Enclosure is the
apparent absence of any magnetic anomalies
indicating the presence of palisades, either surrounding the peripheral settlement or dividing
the peripheral settlement from the central area or
Main Enclosure. The presence of such boundary
features or fences is suggested by the historical
evidence. It is possible that the fence was not
burnt down, perhaps indicating that the destruction of the site was not as complete in the outer
part of the site as within the Royal Enclosure. It
also must be considered that at least some of the
dwellings were similarly left unburnt. This is in
line with reports that the site was left partially in
ruins rather than destroyed. Should this be the
case then the estimate for the number of
dwellings based on the MS data must be
regarded as low.
Despite this, a tentative boundary can be
drawn through the survey area to indicate the
division between the Peripheral Settlement and
the Main Enclosure, where fewer dwelling-type
responses can be seen in the high-pass filtered
data. However, part of the definition of this zone
Archaeol. Prospect. 12, 31–49 (2005)
King Lobengula’s Palace
Figure10. Interpretation of the geophysical data.
is confused owing to the presence of a modern
‘ritual’ hut. There is evidently an element of
‘noise’ in the area around the hut and it is not
possible to differentiate the recent from the historical. The areas either side of the modern road
suffer from similar ‘noise’.
Testing the interpretation of the geophysical data
A number of trenches have been excavated at the
site and, although a fuller publication is underway, some detail from the excavation can inform
on the interpretation of the magnetic data. As
Copyright # 2004 John Wiley & Sons, Ltd.
expected the excavations found that the topsoil
was very thin ( < 0.1 m in places) and that the
presumed structures were indeed burnt clay
floors. More surprising is the fact that the presumed palisade surrounding the Royal Enclosure was insubstantial, amounting to no more
than 0.4 m wide and 0.2 m in depth. Charred
wooden stakes have been found within the palisade, supporting the fire-induced destruction of
the site.
The following three excavated areas have been
selected for brief descriptions because they highlight the character of features identified by the
Archaeol. Prospect. 12, 31–49 (2005)
C. Gaffney, G. Hughes and J. Gater
Figure11. Map showing the earthworksinthe area ofthe geophysicalwork.The excavationsmentionedinthe text are also marked.
geophysical surveys (Figure 11). However, they
represent only a small part of the area in and
around the Royal Enclosure that was ultimately
excavated. In all cases the deposits were excavated using small hand tools and all were dry
sieved through a 1 mm steel mesh for the recovery of small artefacts. All finds, from both hand
excavation and sieving, were recorded by 1 m
grid square. Deposits and features were recorded
on pro-forma record cards supplemented by
plans (at 1:20 and 1:50), sections (at 1:10 and
1:20), and monochrome and colour photography.
Copyright # 2004 John Wiley & Sons, Ltd.
Area A—three clay house floors in the
southwestern area of the Royal Enclosure
Area A (Figure 12) comprised an area of
272 m2. The topsoil was up to 0.05 m thick and,
over much of the excavated area, directly overlay
the natural bedrock. The fragmentary remains of
three house floors were exposed, from north to
south structures 9, 10 and 11.
Structure 9 comprised a low mound of compact red-brown clay fragments up to 0.3 m thick
and 4.5 m in diameter. This was overlain by a
fragment of polished clay floor surface up to
Archaeol. Prospect. 12, 31–49 (2005)
King Lobengula’s Palace
Figure12. Detailed plan of the excavation of Area A.
Copyright # 2004 John Wiley & Sons, Ltd.
Archaeol. Prospect. 12, 31–49 (2005)
C. Gaffney, G. Hughes and J. Gater
Figure13. Excavation of structure10, a house floor situated in Area A. Reconstructed dwellings may be seen in the background.
Figure14. Detailed plan of the excavation of Area K.
1.3 m long and 0.3 m wide. Structure 10 was
associated with several fragments of polished
clay floor. An arc of eight stake holes defined
the western wall of this building and suggested
Copyright # 2004 John Wiley & Sons, Ltd.
that it had a diameter of 4.2 m. Approximately
1 m to the west of the structure was a short arc of
stones forming a kerb 4.5 m long (Figure 13).
Structure 11 was similarly represented by the
Archaeol. Prospect. 12, 31–49 (2005)
King Lobengula’s Palace
wooden stumps were recorded within the fill of
the gully. In some areas the bedrock had been
scorched red, presumably when the palisade was
burnt down in 1881. Numerous fragments of clay
anthropomorphic figurines were recovered from
the fill of the palisade. One of these figurines was
in direct association with a complex glass bead
and another with a shirt button.
Figure 15. Excavation of the palisade surrounding the Royal
Enclosure. A reconstruction of the palisade can be seen in the
fragmentary remains of a polished clay floor and
a small group of stakeholes suggesting a building
approximately 4 m in diameter.
All three buildings have been badly disturbed
by termite activity. Neither the entrances nor
traces of any internal arrangements could be
Area K—the southern area of the palisade
surrounding the Royal Enclosure
Area K (Figure 14) comprised an irregular area
of approximately 36 m2. The topsoil was up to
0.1 m thick and directly overlay the bedrock. The
palisade gully was orientated southwest–northeast across the excavated area and was cut into
the bedrock to a depth of 0.15 m. It was up to
0.5 m wide and it was filled by a dark grey ashy
deposit (Figure 15). In total, a 9 m long stretch of
the palisade was excavated. A series of charred
Copyright # 2004 John Wiley & Sons, Ltd.
Area V—a house floor and associated
palisade gully in the southern area of the
Royal Enclosure
The main area of excavation (Area V; Figures
16 and 17) measured 86 m2 and was targeted on
one of the circular anomalies identified by the
magnetic susceptibility survey. Three smaller
‘keyhole’ trenches (V(1), V(2) and V(3)) were
also excavated with the objective of locating the
feature responsible for one of the linear
responses (a possible internal palisade gully) on
the gradiometer survey. The topsoil was between
0.05 and 0.1 m deep and, in the southern area of
the main excavation, directly overlay an area of
weathered bedrock.
Several areas of floor surface (structure 14)
were exposed within the western part of the
main excavated area. These comprised areas of
compact burnt clay with a smoothed or burnished surface. A number of disturbed fragments
of clay floor were also recorded. These floor
fragments were concentrated in an area 5.2 m in
diameter, suggesting the original dimensions of
the roundhouse. There was no evidence for any
burnt stakeholes or for a gully marking the line of
the wall of the house, nor was there any evidence
for the entrance or any other internal features. A
lower grind stone was located in the southern
area of the house along with several upper grind
stones or ‘pounders’. Part of a crushed pottery
vessel was recovered from the northwestern area
of the house.
The palisade gully (1090) was recorded in the
eastern part of the main excavated area. It was
between 0.3 and 0.4 m wide and up to 0.3 m deep
with steep sides and a flat base. Numerous,
irregularly spaced charred post stumps were
recorded within the stony fill of the gully. The
majority of these stakes were no more than 40–
50 mm in diameter. The palisade gully was
also recorded in each of the ‘keyhole’ trenches
Archaeol. Prospect. 12, 31–49 (2005)
C. Gaffney, G. Hughes and J. Gater
Figure16. Detailed plan of the excavation of AreaV.
Copyright # 2004 John Wiley & Sons, Ltd.
Archaeol. Prospect. 12, 31–49 (2005)
King Lobengula’s Palace
Figure17. Detailed plan of the excavation of AreaV1^3.
excavated (V(1), V(2) and V(3)). In trenches V(1)
and V(2) the gully had similar dimensions and
fill (1099 and 1100) to the section recorded in the
main area of the excavation. However, in the
northern ‘keyhole’ trench, V(3), it was deeper,
up to 0.4 m, with a causeway 0.5 m wide. This
suggested a narrow entrance into the enclosure
marked by the gully. Possible stake impressions
were recorded in the base of each of the gully
terminals (1102 and 1097). An elongated pit
(1098) was recorded immediately to the north
of this possible entrance. An upper grind stone
was recovered from the fill of this pit.
The combined evidence of the excavation and
the geophysical survey suggested an oval enclosure, 30 m long and 17 m wide, which surrounded roundhouses S6 and S14 and with a
narrow entrance to the north. Evidence for a
similar enclosure surrounding the roundhouses
in the southwestern area of the Royal Enclosure
(S9–S11) was also suggested by the gradiometer
survey. The identification of these enclosures
suggests that space within the Royal Enclosure
was subdivided into smaller areas, perhaps
Copyright # 2004 John Wiley & Sons, Ltd.
based on the domestic space of individual family
units or Royal wives.
The clarity of the plan produced by the geophysical survey is of exceptional value in the overall
spatial interpretation of the settlement (Figure 7).
When combined with the information recorded
during the topographic survey, a very detailed
plan of the Royal Enclosure and the Northern
Peripheral Area of the settlement has been
obtained. In addition to the visible structures,
the geophysical survey has been able to demonstrate the presence of a large number other buried
structures and that a number of fences or palisades subdivided space around, between these
buildings. Many of the anomalies have now been
tested by excavation and the results of which
have further characterized the nature of the structures and deposits. The most substantial feature
that was suggested by the magnetic data was the
palisade that surrounded the Royal Enclosure. It
Archaeol. Prospect. 12, 31–49 (2005)
was surprising, therefore, that when this feature
was tested through excavation (Area K) it proved
to be a relatively insubstantial feature, no more
than 0.5 m across and 0.2 m deep. It is suggested
that the broad ‘signature’ given by the magnetic
data is largely a reflection of the burning of the
Royal Enclosure in 1881. This palisade was subsequently excavated at other locations, for example the western section to the rear of Lobengula’s
house. At this point the cut for the palisade was
barely visible in the excavated evidence. Similarly, the boundary of the subenclosure near to
the eastern entrance of the Royal Enclosure
(Figure 10), which is so clear in the magnetic
susceptibility data, left barely a trace when this
area was subsequently excavated. This area proved to contain large quantities of animal bone
(Murray, 2001) and is thought to have served
the dual function of a midden and a cattle byre.
It can be concluded that, in some areas, the
geophysics has proved more useful than the
ground plan that emerged during the excavation.
That is, the signature provided by the magnetic
C. Gaffney, G. Hughes and J. Gater
susceptibility survey is clearer than the actual
excavated archaeology.
One possible explanation is that the boundary
features surrounding the Royal Enclosure and
the cattle byre may have been relatively insubstantial timber structures but were perhaps covered in thick layers of thatch. This would be
consistent with the excavated evidence that produced verification for small charred stakes rather
than substantial post stumps (see e.g. Figure 14).
When this thatch was torched in 1881, it left a
marked magnetic signature in the topsoil that
has been clearly detected by the geophysics but
not so evidently by the excavation. Ironically,
this hypothetical reconstruction of the boundary
features is in marked contrast with the actual
reconstructions that have been undertaken onsite (Figure 15). The building of these massive
timber screens was heavily influenced by visits
that were made to the reconstructions, by the
Historic Monuments Council of Kwa-Zulu/
Natal, of Zulu settlements in South Africa. It
now appears that the archaeological evidence is
Figure18. Three-dimensional representation of the high-pass filtered magnetic susceptibility data.
Copyright # 2004 John Wiley & Sons, Ltd.
Archaeol. Prospect. 12, 31–49 (2005)
King Lobengula’s Palace
suggesting structures that were quite different in
their appearance.
By contrast, no similar anomaly could be identified where comparable structures were
expected in the peripheral areas of the settlement. It maybe that these outer areas were not
subjected to a similar burning episode. It is
hoped that further trial excavation will be undertaken in these areas in the future.
The geophysical survey has for the first time
given a clear indication of the full dimensions of
the Royal Enclosure. The plot of the surrounding
palisade also suggests an elaborate entrance on
the northeastern side of the enclosure. This is
possibly linked to the location of a major cattle
byre in this area. The survey also gives a clear
indication of the number and density of buildings that were located within the Royal Enclosure and within the peripheral areas of the
settlement (Figure 18).
The combination of the geophysical and topographic surveys has provided a highly useful
prelude to excavation. They demonstrate that
geophysical survey is a cost-effective means of
providing detailed information about the layout
and plan of the settlement. Furthermore, it
appears that the detailed MS survey has
recorded information from the shallow topsoil
that cannot be detected by other survey or excavation techniques. The information has provided
an invaluable basis on which to plan a detailed
excavation strategy, and to assist in the planning
of the long-term conservation of those parts of
the site that will remain unexcavated.
The authors would like to thank the funding
bodies mentioned in the text as well as Claire
Stephens and Emma Wood for finalizing the
diagrams. A special thanks to all the members
of staff from the National Museums and Monuments of Zimbabwe who assisted with the survey and especially to Joseph Muringaniza who
was the Project Manager in overall charge of the
KoBulawayo project.
Copyright # 2004 John Wiley & Sons, Ltd.
Clark AJ. 1996. Seeing Beneath the Soil, 2nd edn.
Batsford: London.
Dearing J. 1999. Magnetic Susceptibility. In Environmental Magnetism: A Practical Guide, Foldfield F,
Smith J, Walden J (eds). Quaternary Research
Association, Technical Guide Number 6: London;
Gater JA. 1994. Old Bulawayo, Zimbabwe. Unpublished Report, GSB No. 94/125, GBS Prospection:
Gaffney CF, Gater JA. 2003. Revealing the Buried Past.
Tempus: Stroud.
Gaffney CF, Hughes G. 1997. Old Bulawayo, Zimbabwe. A Topographic and Geophysical Survey of the
Royal Enclosure, November 1996 and March 1997.
Unpublished Report, Birmingham University
Field Archaeology Unit (BUFAU) Report No.
Heathcote C. 1983. Applications of magnetic and
pulsed induction methods to geophysical prospection
at shallow depth. Unpublished Archaeological
thesis, University of Bradford.
Hughes G, Muringaniza J. 2003. Kings and ancestors: interpreting the past at KoBulawayo,
Zimbabwe. In Researching Africa’s Past: New
Contributions from British Archaeologists, Mitchell
P, Haour A, Hobart J (eds). Oxford University
School of Archaeology Monograph 57: Oxford;
Linford NT, Canti MG. 2001. Geophysical evidence
for fires in antiquity: preliminary results from an
experimental study. Archaeological Prospection
8(4): 211–226.
Limbrey S. 1995. The Soils in BUFAU, NMMZ and
GSB 1995. An Archaeological Assessment and Research Proposal for Old Bulawayo, Zimbabwe. Unpublished Report, Birmingham University Field
Archaeology Unit (BUFAU) Report No. 332.1–4.
Murray E. 2001. Pangolins, Pythons and Panthera: An
Interim Report on the Animal Bones from Excavations
at KoBulawayo, Zimbabwe 1997–2000. Unpublished
Report, Birmingham University Field Archaeology Unit (BUFAU) Report No. 332.11.
Peters C, Church MJ, Mitchell C. 2001. Investigation
of fire ash residues using mineral magnetism.
Archaeological Prospection 8(4): 227–237.
Thompson R, Oldfield F. 1986. Environmental Magnetism. Allen and Unwin: London.
Weston DG. 2002. Soil and susceptibility: aspects of
thermally induced magnetism within the dynamic pedological system. Archaeological Prospection 9(4): 207–215.
Archaeol. Prospect. 12, 31–49 (2005)
Без категории
Размер файла
861 Кб
lobengula, survey, kobulawayo, king, palace, zimbabwes, geophysics
Пожаловаться на содержимое документа