Pula: Botswana Journal of African Studies vol. 15 no. 1
An archaeological survey of Ntsweng in Molepolole
Princess Sekgarametso
   Ntsweng was the capital of Bakwena before they moved to Molepolole. The bulldozing of a pan of
  Ntsweng for a sports field in 1994 inspired the mapping of the site. At Ntsweng, several clusters of
   ruins seem to represent different wards. Analyses of the settlement layout and soil samples suggest
   cattle were often kept in the center of each ward. This seems to follow the "central cattle pattern ",
   a conceptual model for the organization of settlements among southern Bantu, where the central
  placement of the kraal reflects the central role of cattle in the culture and ideology of the people.
Ntsweng is a historic site located a few kilometers southeast of Molepolole in Kweneng
District (Fig. 1). The area was first occupied by Bakwena, led by Sechele I, in 1864. It was
abandoned in 1937 when Bakwena were forced to move to Molepolole by Kgari II, assisted
by the colonial administrators (Schapera, 1943).
   Ntsweng is one of Botswana's most important monuments. Contrary to the 1970 National
Monuments and Relics Act, 24,000 sq m of the site were bulldozed in February 1994 to
make a sports field for Sedumedi Community Junior Secondary School. A rescue project
was carried out by the Archaeology Unit of the University of Botswana, with assistance
from students of Sedumedi Community Junior Secondary School and Legae Academy, to
salvage the damaged area. A lot of artifacts like bones, pottery, ostrich eggshell beads,
metal earrings and other metal objects were recovered from the area (Borgesen et al.,
1994).
   The inadequacy of cultural resource management in Botswana can be partially blamed for
the bulldozing of Ntsweng, as it failed to provide proper delimitation of the site (Borgesen
et ai" 1994). This inadequacy in cultural resource management extends even to
archaeological sites in Dimawe, Manyana, Lose, Modipe and Domboshaba which have also
been damaged.
   The importance of these sites is that historically written information about Botswana's
distant past is very limited. Information mostly comes from archaeological sites which
need to be protected, Even though Ntsweng is documented historically, there is still more
to the site which needs to be exploited archaeologically in order to add to historical
information. If these sites are not preserved, historical information would be lost and we
would have no history to talk about.
    Ntsweng today consists of a large area covered with traces of occupation. Most notable
are the patterns of stones laid on their edges to form house foundations which are still
visible on the surface. The only building still standing is what used to be Sebele II's office
which is at present referred to as Mmakgosi' shouse,
    The main aim of this paper is to investigate the organization of the settlement at Ntsweng.
This was done by conducting an archaeological survey of the remains of structures at
Ntsweng. In particular, the survey was to assess the damage done by the bulldozer and to
reconstruct the damaged area. Information generated from the survey is used to observe the
 settlement pattern and the site's functional and symbolic layout. The plan produced is also
 checked to see whether it conforms to Huffman's central cattle pattern, a characteristic of
 southern Bantu settlements where cattle have a central place (Huffman, 1986).
    The results of this survey can be used to guide subsequent archaeological research, as
 what lies on the site's surface is often a general reflection of what exists beneath.
 Excavators can use this surface information to place key excavation units so as to test the
                                                    23


hypotheses generated from surface data (Knudson. 1978). In this paper. I aim to address
three issues. I intend to find out from the archaeological            record how the settlement at
Ntsweng was organized. I intend to find if the settlement plttern at Ntsweng <:<>nfo~ed to
the central cattle pattern. And with the aid of phosphate analysis and oral mtervIews, I
intend to find out how labour was organized and space was used at Ntsweng.
Methodology                                                                                      ..
Archaeological survey is taking stock of everything that remains on an archaeological                  site
in the form of plans, notes and photographs              in as much detail as possible without
 excavating the site (Sharer & Ashmore, 1979). The archaeological              survey of Ntsweng was
done first by walking around the site to familiarize myself with the features and extent of
 the site. This was also done to choose an area to focus my study. The criteria employed for
 choosing the area to work were good preservation of stone alignments, sparse bush to allow
 good visibility, and a size comparable to the bulldozed area. After choosing the area, I
 started mapping the stones using a dumpy level. The instrument provided the bearing of the
 stones from the temporary benchmark we had created. A measuring tape was then used to
 take the distance of the features from the instrument. These measurements and angles were
 then plotted on graph paper.
    In addition, soil samples were collected and taken to the Archaeology                    Unit in the
 University of Botswana to test their phosphate content. This was done because it is known
 that human activities produce phosphorus-rich        organic materials which after decaying are
 incorporated into the soil in the form of stable phosphates.             These tend to preserve for
 thousands of years and can be detected by chemical analysis. This technique dates back to
 the 1920s when fieldwork in Sweden revealed the close correlation                    between ancient
 settlements and high concentrations of phosphorus (Renfrew & Bahn, 1991).                    Phosphate
 analysis is a geo-chemical technique the main scope of which is
    t? localize ancient monwnents not visible in a surface survey, to assist in detennination       of site
    Iwnts and to give infonnation about the function, activity or duration of occupation. (Bakkevig,
     1980:74)
    Proudford was able to estimate that activities of a POpulation of 100 people would deposit
  124 kg of phosphorus annually (Clark, 1990). Recent phosphate analyses done in England
  have revealed that archaeological       features that remain undisturbed          in the subsoil are
  accurately reflected in the topsoil.
     !n  the ~~ .topsoil was considered to ~ unstratified and hence devoid of archaeological
     infonnatlOn;
     Bahn,1991:87)It was often removed mechanically and quickly without investigation . (Renfrew &
     Cl~     (l~?    says t~t phosphate tests respond mostly to middens                 cattle pits and
  domestic actiVity espeCIally around pits and buildings.         Soil samples fro~ Ntsweng were
  collect~ to try and ~~te such areas and also to assess the oral historical data collected
                         =~~~
  concemmg the OfganIzabon and use of space at Ntswen              In all                               ed
  plus two control samples from ff th          .    Th         g.        ., 75 samples were collect
  natural content of b            . 0       ~ Site.    e control samples were taken to detect the
   10 cm into the so8                the so~:'fi The collection of soil was done by digging about
   100 g.                             ng SOl om underground           These samples weighed about
      A method known as spot testing was em I ed
  This was done by placing 0.5 g of so'l fi Peach   OY to test the phosphate content of the soil.
                                           1    om        bag on a separate filter paper. This was
                                                    24


then followed by adding two drops of amonium molybdate to the soil and allowing it to
spread into the filter paper. Two drops of ascorbic acid were then added to the soil solution
and left for three minutes to dissolve on to filter paper. It is important to have the
chemicals reacting within a given time to avoid over-reaction. The soil was then removed
from the filter paper which was examined. Sixty two percent of the filter papers showed no
sign of blue colouring which means that the soil was <0.08 % P20S. Twenty three percent
had a trace of blue up to 2 mm from the soil, which means that the phosphate content was
0.08 - 0.15 %. Nine percent were weak (blue circle, with rays merging) marking phosphate
content of 0.15 - 0.8 %. And finally, 5 % were positive (distinct blue band 10-15 mm in
diameter) marking phosphate content of 0.8 - 0.9 %. The control samples showed no
phosphate content which showed that naturally the soil contained no phosphorus.
    After collecting the soil samples, oral interviews were conducted in Molepolole. These
were used to help interpret the archaeological information at Ntsweng. A prepared
questionnaire was used to guide the oral interviews (see Sekgarametso, 1995). A total of 21
informants were consulted in individual and group interviews. The group interview was
conducted amongst a group of women known as Matsosa-Ngwao. These women are
volunteers who have dedicated their spare time to the promotion of Kweneng culture. Thus
they are mostly found at the Kgosi Sechele I Museum or at the kgosi-e-kg% 's kgot/a
assisting where help is required. The information gathered from these women was checked
by conducting individual interviews.
    Literature consulted was found at the National Museum, Phutadikobo Museum, Kgosi
 Sechele I Museum, National Library and University of Botswana Library. These books
 gave historical information about Bakwena and the stages that led to the abandonment of
the site at Ntsweng. Other more general books about settlement studies conducted outside
Botswana were also consulted Problems encountered during the research were mainly that
the instrument I was using demanded a lot of labour. Thus, I spent more time in the field
 than I had anticipated. The more efficient Electronic Distance Measurer (EDM) was not
 working.
 Literature review
 Archaeological work done in Botswana is very limited, especially on historically
 documented sites. More historical archaeology has been done in neighbouring South Africa
 and it is information from these sites that would be most useful for comparsion with the
 settlement patterns at Ntsweng (e.g., Pistorius, 1992). This is because the occupants of
 Ntsweng are said to be part of the Sotho-Tswana group which came from South Africa
 (Ngcongco, 1982).
    More broadly, comparative studies on settlement patterns and use of space have been
 done worldwide and especially useful to my study is a book edited by Kent (199Oa) on
 domestic architecture and use of space. The contributors provide very important and
 original insights into the relationship between the built environment and the organization of
 space, though from different backgrounds. In their studies the fundamental role of culture
 as expressed in technology, symbolism, socio-political organization and economics is
 explored. This is done with the aid of case studies of the past and present drawn from
 various countries. Especially useful information in this book was Lawrence's study on
 urban housing in Switzerland where he argues that when studying settlement patterns one
  should not look only at specific buildings but also look at the diverse actors involved with
  their design (Lawrence, 1990). Kent's (199Ob) cross-cultural study of segmentation,
 architecture and the use of space saw Batswana as segmented into gender and age
                                                25


distinctions when observing the relationship between use of space and domestic built
environments.
    Information regarding survey techniques was available from a number of sources
including Sharer & Ashmore (1979). Renfrew & Bahn (1991) and Knu~n                               (1978). They
provided the necessary details including case studies in which archaeological surveys we~e
done and interpretation of the results was given. Literature consulted on phosphate analy.S1S
includes Nunez (1990), Shackley (1981), Bakkevig (1980) and Clark (1990). These prOVIde
different ways of conducting phosphate analysis and include case studies. Particular~y
useful was Bakkevig (1980) who points out some of the problems in phosphate analysts
which have been done in the past and attempts to correct the mistakes ..
     Huffman (1986) has proposed a pan-southern African settlement system to explam
 southern Bantu settlements. This has been termed the southern Bantu cattle pattern or the
 central cattle pattern. This pattern, according to Huffman (1986:289) is " ... found among
 Bantu speakers in southern Africa who are predominantly partrilineal and who exchange
 cattle for wives". He further says that
     attitudes about economy, politics and religion result in a specific arrangement whereby an outer arc
     of houses, arranged according to some alternating system of status, surround a central zone that
     contains cattle byres, grain storage facilities, elite burials and the men's court. (Huffman, 1986:289)
     The settlement, says Huffman, was arranged in such a way that the front was public and
  reserved for dangerous activities and the back was sacred and private. He further says that
 the settlement pattern is governed by attitudes towards status. Using this model, I will
  investigate whether Ntsweng corresponds to Huffman's central cattle pattern.
      Pistorius' (1992) study of Bakwena settlement pattern near Rustenburg was especially
  useful. He conducted an ethnoarchaeological study in which historical, ethnographic and
  archaeological evidence was combined to interpret the settlement pattern at Molokwane. He
  further attempted to explain the way of life of the inhabitants of the village. Based on
  ethnographic and archaeological information obtained, Pistorius (1992) says that the
  settl~~ent had once been that of Bakwena. He found numerous wards occupied by related
  pamlineages (masika) and also found that the settlement pattern conformed to the central
   cattle pattern defined by Huffman (1986).
      McDonald (1940) conducted an ethnographic study of Bakwena material culture and
   architecture immediately after the abandonment of Ntsweng. Although his study mainly
   ~m~           the Bakwena ~f Molepolole and those in South Africa, he provided useful
   information about the SOClal, economic and political organization of Bakwena at Ntsweng.
   Lm:ssc:'n& Larsson .(1?84) documented traditional Tswana housing in Botswana. Their
   mam mterest was buil~g methods and they observed the layout of dwellings together with
   th~ use of space. Then book provides useful information for interpreting the stone
   altgnments at Ntsweng.
       Arc~l~gi~         wo~ <i?neon historically documented sites in Botswana include Pahl's
    (1971) linu~ .mveStigation of Motsenekatse Hil~ ongoing research at Modipe Hill in
    K~tleng Distnct ~ the Archaeology Section of Phuthadikobo MuseUlll, work done at
    ~mawe by the National Museum Monuments and Art Gallery and work done on Kutswe
    hill by Segobye (1994). Inaddition, I looked at van Waarden's interpretati                   f        . sat
    Vumba (1989), and her excavations at Leeuki . th                               ..        on 0 ~e
    Waarden (1980 1989) e......1            structuralistOP~ .eNo~ East District. At these SItes van
                      ,          uay,oys a                  View 10 the mte         tt.     f    terial   I
    At Vwnba she agrees with Hn"---'                                         .}"        on 0 IDa        cu ture.
    "           ..                           "'1ll14ll S (1986) central       cattle pattern, saying that
      ... oppoSltions of male and female, pastoralism and agriculture, ancestors and descendants,
                                                           26


rulers and subjects, cold and hot. .." are represented in the spatial oppositions of the
settlement plans (van Waarden, 1989:133). She says such cognitive and conceptual
structures arc present in all cultural behaviour. She also says that division of labour at
Leeukop was along sex lines rather than those of age or status, thus suggesting differences
in male and female activity (van Waarden, (980).
   Caister (1980) says that Iron Age settlers had occupied the Kweneng area by the middle
or late fIrst millennium. The major phase of occupation of the Molepolole area probably
began early in the second millennium and is so far represented by 24 sites distributed from
Dithejwane in the west to at least as far as Gaborone in the east. Caister mentions the
presence of Cenchrus ciliaris and stone walls on hill top sites around Molepolole and
Gaborone. The stone walls are said to have been built by the Nakedi and Kgwatheng who
were found occupying the Molepolole area when Bakwena entered present day Botswana.
His observations deny the fact that Bakwena could have been involved the building of the
stone walls. He further says that sites from the 1820s onwards were chosen for political and
military advantage during the Mfecane wars. This is why we find most of the sites located
on hilltops. The historical background of Bakwena and the causes and consequences of
forced abandonment of Ntsweng in 1937 are well documented (Schapera, 1943, 1980;
Okihiro, 1973, 1976; Ngcongco, 1982). These sources also describe subsistence amongst
Bakwena who inhabited Ntsweng.
Historical background
Bakwena are part of the Sotho-Tswana groups that arc said to have emerged in the period
AD 1000-1500 (Ngcongco, 1982). According to oral traditions, Bakwena trace their origins
to Malope, the leader of the Baphotu Kingdom in the early 10th century. The Baphotu
Kingdom was situated near the source of the Odi or Crocodile River to the northwest of
modern Pretoria (Ngcongco, 1982). Malope is said to have had two children, Mohurutshe, a
daughter and Kwena, a son. These are said to have formed the Kwena-Hurutshe group but
eventually split with one group going with Kwena and assuming the name Bakwena, and
the other with Mohurutshe and assuming the name Bahurutshe.
   Bakwena of Molepolole today trace their origins to Bakwena ba-ga-Mogopa in the
Rustenburg District where they had lived until the death of Tebele, when his son Mogopa
broke away from his people and settled at Mabyanamatswana (Swartkoppies, near Brits).
Among Mogopa' s people there was a ward under Kgabo who was Mogopa' s son from his
junior house. Periods of drought and moving from one place to another resulted in
Mogopa's people scattering allover. When he heard that at Mabyanamatswana rain had
fallen and there was plenty of food, he decided to gather his people and go back. Kgabo is
said to have promised to follow later, but instead he is said to have declared himself
 independent and crossed the Marico river into what is now Botswana, settling in present
day Kweneng District (Schapera, 1980). A further split occurred amongst the remaining
Bakwena at Mabyanamatswana due to drought. A group of Bakwena Bamogopa and
Bakwena Bamodimosana moved away and it was this group who built and occupied the site
of Molokwane that has been investigated by Pistorius (1992).
   Bakwena ba-ga-Kgabo first settled at Dithejwane Hill where they encountered earlier
 inhabitants of the area, the San and Bakgalagadi. These were" ... expelled, driven west into
the Kgaiagari and maintained under a symbiotic client relationship by the baKwena."
 (Okihiro, 1973:104). During this period Bakwena are said to have been very rich owning
cattle and sheep. They grew pumpkins, beans and other vegetables. This state of affairs
 went on until the death of Motswasele II (Okihiro, 1973). After this Bakwena are said to
 have experienced a period of raids from Sebetwane and Mzilikazi who were at the time
                                               27


terrorizing Batswana by attacking them and stealing their cattle. These raids scatterd
Bakwena and they were regrouped by Sechele I after Mzilikazi had been defeated at
Lephephe. Sechele Ihad before been captured and released by Sebetwane (Sillery. 1952).
   English traders and missionaries entered Bakwena territory at the time of Sechele I's
reign and he was the first Bakwena chief to be baptized by Lhingstone. Under Sechele I
Bakwena are said to have increased, absorbing refugees of both Tswana and non-Tswana
origin and placing them directly under the authority of the royal ward By so doing Sechele
is said to have been
    able to strengthen his bogosi and use it to capita1ize on the growth in the trans-Kgalagadi trade.
   After 1850 the BaKwena became the most powerful and prosperous morafe in the region. (Morton
    et ai., 1989:64)
 From Lephephe, Bakwena moved to several areas. In 1852, whilst staying at Dimawe, they
 were attacked by a Boer commando (Ramsay, 1991). Having defeated the Boers, Sechele I
 moved his capital first to Dithubaruba and then to Ntsweng in 1863. He was thus the fust
 Bakwena ruler to settle at Ntsweng, and he was well established and prosperous.
    Sechele I had three sons: Kgari, Sebele and Tumagole. When he died there was a
 succession dispute amongst his people. Kgari was the first born child from the first house
 but his mother was a Mongwato. Sebele was the first born child from the second house and
 his mother was a Mokwena. This divided the Bakwena, with Sebele's supporters saying
 that they did not want to be ruled by Kgari as he was a child born of a non-Mokwena
 mother. For this they referred to him as set/ogolo-sa-dichaba (child born of a non-
 Mokwena mother) (Okihiro, 1976). Another faction supported Kgari as the rightful heir
 because he was born into the first house. Unfortunately for him, Sebele managed to take
 over and Kgari and his people were driven away from the capital (Okihiro, 1976). It was
 during Sebele I's reign that Britain wanted to declare a protectorate over the land of
 Batswana. Sebele is said to have been against the idea and throughout his reign he is said to
 have prevented British interference in Bakwena affairs. In 1896 and 1899
     rinderpest, drought, and disease destroyed stock and decimated the BaK wena. For much of this
     penod Sebele abdicated his responsibilities by retreating to his cattle post and rejecting British
     otTers of food. (Mortonetai., 1989:110)
     Se~le I was succeeded by Sechele II who is sometimes referred to as Kealeboga. A
  co~ ..ct r~sulted ~hen Sechele II sanctioned the Anglican Church. As a result. LMS
  nusslOnanes are saId to have
      su~r~e;t th~ ?'~osana in a campaign to discredit him before the British. The resulting conflict
      sp.     0 epo ? e mto the ~tsw~g    and Borakalalo factions and occasioned the first BaK wena
      Tnbal CO~CIl. The colll1cll, which was controlled by Seche1e's sibling rivals Kebohula and
      MOiteelasilo,.~n lapsed ~ Sechele n asserted his authority by promoting .uni~r headmen into
      powerful posItIons and revIVffigbogwera. (Morton et al., 1989:112)          J
  Sechele II moved his people back to Ntsweng from BOrakaIal h                           .
  remained behind At Borakalal th                           .         0 were most of his opponents
  government camp (BNB 9719:3). o ey were ad'acent         "       to the coI0nial a dminis' tratlon
                                                                                                  ., s
      At the time of Sechele II's death there              bi       " .
   Ntsweng and Borakalal0 This di                  was ~ tter diViSIonbetween the residents of
   Sebele II took over he ~ntinued :.the              ::Ign of Se~le I~ (BNB 9719:14). When
                                                cmg ogwera, rammaking and polygamy which
                                                    28


 offended the Christians and angered the colonial administrator (Morton & Ramsay, 1990).
 Sebele II is also said to have promoted commoners.                This threatened the position of the
 elite. These dikgosi accused Sebele II of being backward                  The colonial administrators
accepted these accusations and supported petitions brought forward by Sebele' s opponents.
 These opponents, who were mostly from Borakalalo, were questioning Sebele's ability to
judge cases and govern without their advice (Morton & Ramsay, 1990). They wanted
 Sebele II to move from Ntsweng.            Sebele, however, managed to defeat his opponents (for
 details see Morton & Ramsay, 1990).
    After this he built a new "Bakwena National Office" next to his kgotla. Although Sebele
 II was seen by his opponents as stubborn, during his reign he is said to have
     imposed strict standards on the building of houses and fences and had streets kept clean and well
    maintained.     Rubbish pits replaced garbage heaps. He told his people to bury their dead in the
    graveyard rather than in their lolwapa. New roads were built, and dam construction was attempted
    (though unsuccessfully). Sebele also sanctioned youth programmes in the form of Life Brigades for
    boys and girls. And while in power he welcomed the establislunent of the Scottish-Livingstone
    Medical Hospital. (Morton & Ramsay, 1990)
     In spite of his efforts his opponents continued to accuse him of being a neo-traditionalist.
 Where Sebele II's opponents failed, the colonial administration succeeded: he was detained
 without trial and exiled to Ghanzi.            His brother Kgari was then made chief and built a
 house, with Bakwena money, at the site of the present Bakwena main kgotla in Molepolole.
     Kgari met many problems during his reign. The colonial administrator reported to his
 seniors that Kgari
     complained that the people of the hill [Ntsweng] considered themselves as a separate tribe and with
     the present position of having two villages in Molepolole, he had no control. (BNA S.260/13)
 The people of Ntsweng did not recognize Kgari as chicf. The area became a centcr of
 resistance as a movement was established to have Sebele II back as the rightful heir to the
 throne (Ramsay, personal communication).                  The colonial administration     responded to
 Kgari ,s complaints by asking him to
     inform the brothers and other headmen that the chief would have the support of the government and
     it is expected that they will carry out his orders, at the same time warning them that if they do not,
     other measures will have to be considered. (BNA S.260/13)
     The people of Ntsweng still refused to move from the area and it is said that the only
 neighbours Kgari had were the tax collectors. The government then decided to use force
 and in 1937 houses were knocked down at Ntsweng including the house of Sechele I
  (Morton & Ramsay, 1990). Some of the Ntsweng residents still refused to move and
 having given a final warning the colonial government decided to take the matter into their
 own hands. In April of the same year those who refused to move out were jailed (BNA
  S.260/15/2).      This punishment is said to have had an immediate effect and in May the
 building of houses near Kgari was in full swing.
      Sebele lIs mother, Phetogo, is said to have refused to move and she remained in Sebele' s
  National Office which is the only building still standing at Ntsweng. It is now referred to as
  "ko ga Mmakgosi" by Bakwena. The building is situated next to an area which used to be
  the royal kraal at Ntsweng but is now the royal cemetery (Morton & Ramsay, 1990).
                                                        29


Ntsweng households                                                                     .
The basic architectural unit at Ntsweng was the house. These were recogmzed by stone
alignments that according to oral informants marked the foundation of houses. and
courtyards (malapa). In a few exceptional cases the remains of house walls could stIll be
seen standing. According to Schapera & Comaroff (1991 :33), a Kwena house~old ...... has
its own compound (/olwapa, lapa) consisting of one or more huts and grananes Situated
within a courtyard surrounded by ... a wooden palisade." These houses, as Kent (1990b)
says, are separated according to their activities. Within a household there is said to be the
front lelapa and back lelapa (segotIo) (Fig. 2). The front lelapa consists of ntlo ya lelapa,
the main house where the owner sleeps. This is sometimes shared with small children and
for cooking in bad weather. This house normally has a courtyard (Ielapa) in front where
visitors are received and cooking and eating may be done here. Schapera & Comaroff
 (1991:33) say:
     adolescent children of both sexes may live together in another hut, or share one with an older
     female relative, and unmarried adults usually also have separate huts, one for the people of each
      sex.
 This adolescents' house is normally found on the side of the ntlo ya lelapa. It is usually
 refered to as ntlwana ya lelapa because it is normally where children sleep and it
  sometimes has a lelapa. This lelapa may be adjacent to the nt/o ya lelapa. On the side may
 be a house which is referred to as the loobo. This house is mostly used as a cooking area
  and sometimes it is built with a small wall forming a circle and the roof is held up by
  supports leaving an open space of about a meter between the roof and the wall to allow
  smoke to go out easily.
       At the back lelapa (segotlo) is normally the sesoa. This house is used for grain storage.
  According to Matsosa-Ngwao other household equipment such as building materials may
  be stored in the sesoa and sometimes small children or the very old in the household may
  use it as a sleeping area. There is also said to be a serobe at the back lelapa (segotlo) where
   lambs and kids are kept during the day, when the adult sheep and goats have gone out to
   graze. In the evenings they are taken back to their mothers at the central kraal to sleep. A
   small garden may also be made where ntshe (sweet reed), mabele (sorghum), dinawa
   (beans), magapu (watermelons) and other crops were planted to be eaten by the family
    when produce at the lands was not yet ready. If there were any chickens in the household
    ~ey normally slept on wooden palisades that made up the fence, or in areas made for them
    m hut stoops.
        Building, as mentioned by McDonald (1940) and Larsson & Larsson (1984) usually
    co~en~s        after the harvest in winter. This is done to prepare for the rainy se~son and
    building IS done by both men and women. The dagga application and decoration is done by
    w~men ~ th~ fnunewo~ buih by men. According to Bakwena tradition, before a hut is
    ~lt, certam ntual practl~ are performed. This is done to protect the household from
     Witchcraft and theft. The n.tual practice is performed by traditional doctors and during the
     process .wooden ~gs or ammal horns smeared with medicine are buried in the area where
     the hut ISto be built.
         According to ~n         & Larsson (1984) early records have it that the circular hut is the
     oldest shape of traditional Tswana h             At
     the circular one. According t M Do°naluSCSd    .    Ntsweng the most common hut shape was
                                     o c           (1940:33), one starts by marking
                                                      30


   a circle with a string or rein, one end of which is attached to a peg knocked into the groWld and the
   other end attached to another peg which is drawn at full length aroWld the first, in this way
   describing a circle.
After marking the circle, a hoe or a sJXldeis used to mark clearly the circumference of the
circle. Stones about 40 x 30 cm were then arranged on the circle as the foundation of the
house. This allows for the building of the wall. All this is done by women and the wall is
made by mixing clay with cow dung. The mixing with cow dung, the people say is to
harden and strengthen the wall and to guard it against the effects of rain and the ravages of
termites (McDonald, 1940:34).
    After the cow dung is well mixed with the clay, bricks are made. These bricks are then
used to form the wall. They are not left long enough to dry so they are easily compressed
together to form a circle. This is all done using hands. When the wall has reached the
desired height, men go out into the bush to cut supports for the roof (maotwana) and
sometimes for the center pole (pinagare) of the hut, if it is to be used. After the wooden
framework (go tlhomesa) is constructed by men the house is then thatched with grass
collected by women. Thatching, however, can be done by both sexes. The same technique
of building a hut wall is applied to building the lelapa. The lelapa walls are not round but
angular. Inthis case the shape is just marked on the ground using a stick and then stones are
used to mark boundaries.
    At Ntsweng there were sometimes double lines of stones marking foundations.
 According to oral information the second line of stones served as a foundation for stoops
 (maribela I direpodi). Maribela are made to prevent erosion of the walls and they are also
 made for people to sit on. Sometimes in the foundation of huts square shapes on comers of
 malapa (plural of lelapa) were made to provide seats. Smooth round stones known as lite/o
 were found at Ntsweng. These stones are said to have been used to smoothen floors of
 houses and walls after they had been plastered. A JXlvementof potsherds was found at the
 entrance of one of the huts. These were placed to prevent the floor eroding when it rains
 and they also served as decoration.
    Ntsweng has three types of houses. There is the circular shape which is the most
 commonly found. This is a typical Sotho-Tswana type of house. The diameter of these
 houses ranged between 4.4 - 4.8 m. There was, however, one house that had a diameter of
 6.2 m. There is also the semi-circular type of house which according to one of my
 informants, Mr. R Pheto, was invented by Bakwena and later spread to Bakgatla. Semi-
 circular houses mostly had a radius of 4 m. The rectangular type of house was adopted
 after the coming of the European settlers (Schapera, 1943; Larsson & Larsson, 1984). It
 was not common at Ntsweng: according to Matsosa-Ngwao this type of house marked
  status. This type of house was built with a concrete foundation in front and on the sides.
  This building material was not commonly used on the site. According to oral information,
  the rectangular houses at Ntsweng were those of the kgosi-e-kgolo and a missionary by the
  name of Konobole (Grobler), who owned a store at Ntsweng. Circular structures about one
  meter across were found beside the house. These structures, according to oral information,
  are said to have been bases of difalana (granaries). Similar structures have been found by
  others across southern Africa (Maggs, 1976; van Waarden, 1989; Campbell, 1991;
  Pistorius, 1992; Segobye, 1994).
     In the houses from cluster 1 (Fig. 3), 75% had no sign of phosphates, 18% had trace
  signs, while only 7% were weak. Out of 24 malapa there was only one /e/apo (structure 5)
  that had a positive sign of phosphorus. Most malapa had a trace of phosphorus. There was,
  however, a distinct relationship between the houses and malapa. The higher phosphate
                                                       31


    content within malapa probably resulted from different activities that took place in them.
    In the houses there was less activity. Structures 32, 33 and 34 were rich in phosphate
    content. These were close to each oilier and on these structure was a recent goat pen that
   could have accounted for the high phosphate content in the area ..
      As recognized by van Waarden (1989), labour amongst Bantu-speaking peoples IS
   divided along sex lines rather than age or status. lltis is characteristic of Bakwena. 'f!te~
   were also servants (malata) who were not given equal status in society to Bakwena. WI~
   a Kwena household the man is the head of the household but most of the activities within
   the household are carried out by women. According to Matsosa-Ngwao women are
   responsible for feeding the family. They work in the fields and this produce was used to
   feed the family and maintain the children. Water was fetched by women at Ntsweng from
   the Molepolole river which flows through the modem settlement and is said to have been
   perennial in the I 920s. To supplement this water SOurce European settlers introduced
  boreholes and wells which the people of Ntsweng were given access to (McDonald, 1940).
      Men were responsible for looking after domestic animals and hunting. Domestic animals
  included cattle, donkeys, horses, pigs and goats. According to oral information pigs and
  goats were kept in small numbers. They were mostly looked after by small boys under the
  supervision of their parents. Political and legal affairs were attended to by men and fi~1
  decisions
  with        for the household are said to have been made by men, sometimes in consultation
         women.
  Subsistence activities
  A lot of cattle bones were found at Ntsweng from the bulldozed area. lltis is because
  Bakwena attached much importance to the keeping of cattle as they were the major si~ of
  wealth and they were used for transport, ploughing, bridewealth and for ritual practIceS
  such as sacrifices to the ancestors (hadimo) (McDonald, 1940). Bakwena kept cattle at
  cattle posts far from the villages. These cattle posts were sometimes over 50 kilometers
 away from the village (Larsson & Larsson, 1984). Few were kept in the settlement at the
  central kraa1 of the kgotla they belonged to. These cattle consisted mostly of heifers which
 were used for transport and a few female cows which provided milk for the kgotla.
     Donkeys w~n: penned with cattle at the central kraal. These were often kept in small
 numbers fo~ nding or when driving cattle. According to oral information, donkey carts
 were only mtroduced after the arrival of Europeans. Europeans introduced horses to
 Ntsweng and these were owned by the rich and were a sign of status. According to
 McDonald (~940) both goats and sheep were kept though goats were few. Chickens were
 also
 (Mortonkept.& Pigs        kept but for hygenic reasons were later prohibited bv law at Ntsweng
                    were1990)..
               Ramsay,
     Evidence for agn.culturaI producti..      I de    ...                                  d at
 a di                                    on .mc u s grmding stones which were mostly foun
       stance from the houses. According to oral information grinding took place at the
 segotlo, ~t when performed communally it took place at the kgotla When these stones
 we~ not m use they were stored in the segotlo leaning against woode~ palisades that made
 ~~u~:~~ ~ h~se~~lds.             According to Mr R. Pheto, grinding stones found outside
 found at              Sat to mark where WOodenpalisades were. Part of an old plough was
(940) a;:::'
resulted in
                      ~
                    Ie :.
                          ':1.   These ploughs were pulled by oxen. According to McDonald
                              were as much as 20-30 miles away from the settlement and this
agricultural :n.           vmg homesteads at the lands which were mostly used during the
     Oral information has it that agri ltural           .
trading ptlrposes but fi fi . cu               .production at the fields was not undertaken for
                           or Ceding the family (Matsosa-Ngwao, personal communication).
                                                 32


These people did not have large fields but they were normally big enough to feed the
family. Crops were normally all planted in the same field. These fields were used yearly
and the reason why they managed to produce yearly was probably because the planting of
crops together in one field resulted in minerals absorbed from the soil by crops such as
sorghum being retained by others crops such as beans. In this way the fertility of the field
was maintained and leaching of the soil's nutrients did not occur.
    Crops planted were sorghum (mabele) which was the staple crop, beans (dinawa) water
melons (magapu) and sweet reed (ntshe). Other crops like maize, pumpkins, tobacco and
tomatoes were introduced by Europeans. Bakwena are said to have had problems with such
crops as they were not suited to local climatic conditions (Nangati, 1980).
    The dominance of cattle bone in the collection of artifacts from the bulldozed area
indicates that hunting amongst the people of Ntsweng did not play a major role in their
subsistence. However, evidence of other bovids has also been found in the assemblage.
Hunting was carried out by men and it was done for subsistence. With the coming of
European hunters and traders this changed. Hunting was then done for trading. Such items
as ostrich feathers, ivory and skins were traded for such goods as tobacco, beads, guns,
wagons and European clothing. As found by Nangati (1980: 141) "In the absence of money,
 exchange took the form of barter, the rate of exchange being necessarily unequal."
    Nangati (1980) also found evidence of Bakwena's trade with the Indian Ocean coast prior
 to the 184Os. There also existed trade with other Batswana in the Transvaal. Nangati
 (1980:140) says that
    Although there is as yet little archaeological evidence available regarding this trade, it appears from
    the observations of early travelers that a 'hunting economy' for trade with neighbours already
    existed in Kweneng prior to world market trade, and that the later development of trade with the
    Cape Colony was merely an extension of existing trade patterns.
 One of the existing trade IEttems was that of Ba-Kgalagadi and Bakwena. Skins of
 animals were obtained from Ba-Kgalagadi people and distributed amongst specialized
 kaross makers by shopkeepers. After being converted to mats and karosses they were
 traded outside Bakwena territory (McDonald, 1940).
     Beads, pieces of metal pots, pieces of porcelain and metal saucers have been found at
 Ntsweng. These, according to oral traditions, were obtained from stores which had been
  established at Ntsweng and Borakalalo by the Europeans. Such items as cattle were
  collected by the people of Ntsweng and traded for beads, salt and sweets at the stores. The
  existence of these stores led to competition between foreign goods and indigenious
  manufacturers. As more and more foreign goods came in, indigenously manufactured
  goods were abandoned and ended up no longer being manufactured in large numbers.
  Wooden and clay pots were replaced by metal pots.
     After the arrival of the Europeans some form of trade with grain and cattle is said to have
  taken place. These white traders are said to have stored grain in large containers and when
  drought came they were able to sell it to the people of Ntsweng. TIris changed the idea
  from producing crops for subsistence at Ntsweng to producing for commercial purposes in
  order to obtain foreign goods. TIris type of trade was, however, monopolized by the whites
  and Asians.
     But in the Kweneng Sebele did manage to exercise some control by regulating cattle and grain
     sales, investigating cases of cheating, and getting the people to boycott goods of poor quality.
     (Morton & Ramsay, 1990:36)
                                                       33


  Settlementpa«erns
  In total an area of around 56,000 sq m was mapped at Ntsweng. The mapped area revealed
  seUlement clusters which could be distinguished by the direction in which the houses were
  facing. These houses were all built forming a circle. In the center of each circle there was
  no evidence of structures (Fig. 4). There were a total of five clusters mapped and they all
  conformed to the same ~Uem, except cluster 2 which fonned a semi-circle. A similar kind
  of ~Uern is apparent in historical photographs which were taken at Ntsweng before the
  settlement was abandoned (Fig. 5). Unfortunately, neither the orientation nor the dates of
 these photogrnphs are known. These makgotla which were photographed were not all of
  the same size as is also evident from the plan.
     Cluster I (Fig. 3) has thirty-six distinguishable houses. The shapes of the houses are
  predominantly circular with a few exceptions which are semicircular in shape. It was
 difficult to determine from the ground or from the recorded pattern what structures 32 and
  33 were, but behind these Was a hut with a lolwapa (34). The hut /olwapa had ~ of the
 wall still preserved which was made from the traditional mixture of clay and mud. 33 and
  32 in the southern ~ are bordered by a long line of stones which had some wall still
 standing on the eastern side of 34. Leaning against the line of stones was a heap of stones
 forming a rectangular shape.
    The thick lines mark the pattern made by big boulders that were outside the cluster. We
 assumed that they could have marked the boundary of the cluster on this side. Right behind
 this boundary was what looked like a watercourse. From the southern and western edge of
 the cluster there were natural lines of stones that joined the hills and all the structures found
 within these lines of stones were considered ~ of cluster I. Artefacts were quite scarce
 in the area. At 39 a metal rod and a few pieces of poUery were found on the surface. The
 soil at the central ~ce was quite ashy. Holes which were recently dug and left uncovered
 revealed pieces of poUery and ~ of a metal plough. This was near 7 and 10. Grinding
 stones were found on the western side of the cluster as marked on the map.
    Cluster 2 (Fig. 6) forms an arc around a central s~.        On the western side the cluster is
 separated from cluster I by a small water course. It Was on this side along the banks of the
 water course that ~ding          stones were found        A lot of potsherds were also foun.d
 accurnul~ed by mnmng water. It was not particularly clear from the archaeology why this
 clUster ~d not form a circular shape like others. Also the shape of cluster 2 houses were
 predornmantly semicircular.
    It was not ~~ly          clear what marked the boundary between cluster 3 and cluster 2,
but ~e could distingwsh the difference between the clusters by the different direction the
h~ m cluster 3 face (Fig. 7). Most of the stnlctures built were facing the house rnaJk:ed 64
which was square and had a concrete floor. There was also in this settlement the central
space and a .few square houses. At 7I and 92 were houses with malapa on both sides. One
lo/apa was m the !font facing the central space and one was at the back. Behind 53 was an
area that looked like a midden. There were broken pieces of porcelain, pieces of pouey,
:ken      bottles, broke.n metal POts and the soil was also ashy. Behind 89 was a large space
       looked as~y. This also could have been a midden. At 72 a metal saucer was found and
84 had part of Its wall still standing.
    Cluster 4 (Fig. 8) also had structure bu'            .       '.                            1b
walls of 95 96 97 9                        s i1t fo~~        .a CIrcle WIth a central space.      e
:::: 1:~:~hO~ , . ' ,8 and 99 were unusually buIlt jOlmng one another. This formed a very
and bones. Borden: a:~
                               mOstly semi~         in shape. Between 118 and the hill was a
                                     The so~ Was ashy with broken pieces of pots, cups, bottles
                        g
a I.me 0fb' 19natural boulders.se ernent UnIt on the eastern side was a sman water course and
                                                 34


     Cluster 5 also formed a circle of huts with the central sJllce (Fig. 9). This area was not
 completely mapped but was partly recorded to indicate where cluster 4 ended. It was,
 however, not clear why there was such a big gap between structure 127 and 128. In this gap
 there were no distinguishable structures, but on either side the structures were facing
 towards the same central sJllce.
     Settlement at Ntsweng was governed by a specific attitude towards things. Oral
 information has it that the central sJllce in the clusters was an area in which cattle of the
 settlement unit known as a kgot/a (ward) were kept. In .this area there was the central kraal
 where cattle of all households were kept: The kind of cattle kept in these kraals were
 diph% (oxen) and maradu (adult female). Diph% came in when they had a job to do but
 they spent most of their time at the cattle post. Maradu were kept as a source of milk for
 the kgotla. Attached to the central kraal was the kgot/a (forum, meeting place). According
 to Maggs (1976), the kgot/a was where men socialized as well as where legal and
 administrative matters were discussed. Visitors were entertained there and their news heard
 and discussed         Various types of manufacturing activities were carried out at kgot/a,
 especially the processing of skins and hides by men. The central area was mostly
 associated with men; they spent most of their time here and were ultimately buried there.
 At the periphery were households which were associated with women and where they were
buried.
    According to Mr S. Masimega, within the central kraal there is said to be a path that
 Jllsses across the ward through the central kgot/a. This path is used by cattle and people
 entering the kgotla. The building of the kraal is said to be associated with the kgosi's
 (chief, headman) house. When settling in an area, the first person to choose an area to have
 his house is the kgosi. It did not matter which direction the kgosi wanted to face. The next
persons to choose an area to settle were the uncles. This was done by the oldest uncle
 settling on the right and then younger ones on the left of the kgosi. According to Mr R
Pheto, the idea was to have the kgosi in the center of his uncles. After this, on the right, the
 sons in the kgotta could choose where to stay, starting with the eldest. If you were not of
close relation to the kgosi you settled on the left. When everybody was settled, the central
kgot/a and kraal were erected. The kraal was built facing the kgosi 's house. This was so
that when entering the village, if you knew how the settlement was structured it was always
easy for you to know where the kgosi 's house was.
    ...it is appropriateto restate that buildingsembodyhuman intentions;therefore,to Wlderstandthe
    developmentof residential enviromnentsit is inadequate to study only extant dwellingWlits. The
    scholarmust analyzenot only specificbuildingsbut also the valuesand intentionsof diverse actors
    involvedwith their design... (Lawrence, 1990:91 )
    According to Schapera (1943) members of the same ward or family group were expected
to live together in the settlement and no man could build a home where he pleased nor
change residence without the chief knowing. The layout of the kgot/a enabled the kgosi to
recognize immediately those who would not conform to the ~                  and necessary action
could easily be taken against such a person. Hence the pattern established a social order.
    The soil samples for phosphate analysis were collected from cluster I. The open areas
which se~ed          structures mostly showed no sign of phosphorus. An exception was for the
area marked x (Fig. 3), where a positive presence of phosphorus was found This area is
located at the central space which according to oral and historical data should form part of
the central kraal. In this area there was also known to have been a kgot/a were
administrative and judicial proceedings took place. A fire was normally made early in the
                                                   35


 morning and men would sit and wann themselves. This resulted in. hearths being made
 where ash would be fonned making the soil rich in phosphorus. The higher phosphate level
 of the samples taken here is consistent with this interpretation ..
     As mentioned earlier, grinding stones were placed leaning against wooden palisades thus
 marking the boundary of households. The grinding stones found near structures 15 and 16
 could have marked the boundary of the household. The dominance of circular huts could
 possibly mean that besides the preference for that shape, the cluster was old an~ thus the
 shapes of houses that came later did not influence the cluster. It was, however, difficult to
 determine what the pile of stones near 34 could have been. This, I think, can only be
 determined by other archaeological techniques like excavation
    Cluster 2 was a very small kgotla consisting of about 12 houses. As already mentioned,
 there was no distinct boundary between clusters 2 and 3, except that the houses faced
 different directions. According to oral information, if a kgosi had two wives, often the
 senior son of the second lelwapa could not build his house in the same kgotla as his father.
 In this way he was given permission to start his own kgotla. Looking at the settlement
 pattern, it is possible that cluster 2 was a continuation of cluster 3, as the area for the kgot/a
 had already been filled up, or that junior sons in the household had moved out of the kgot/a
 and established their own at cluster 2.
    According to Mr 1. Diago and Huffman (1986), middens are always located outside the
kgotla near the entrance to the kgotla. This is said to have been so that when cattle go into
 the kraal they could pass through the midden and have their legs covered with ash
 (Huffman, 1986). This ash is said to have aided in the killing of ticks found on cattle hoofs.
 At cluster 3 what looked like middens were not formed at the entrance to the kgotla. This, I
think, was not possible because of lack of space and as a result the middens were placed at
the back of households as is the case in this kgotla as well as with clusters 4 and I.
    The joined malapa in cluster 4 are not easy to understand        According to my informants,
wives of the same man could not stay in the same lelapa so we caDDOt say that they could
have represented houses ofthe wives of one man (Matsosa-Ngwao). It was, however, said
that they could have been joined to show how closely related these people were.
    Huffman (1986) interprets Bantu settlements in binary opposites.         He says that in such
 settlements ~e front was seen as public and thus reserved for dangerous activities whilst the
~ck was pnvate and sacred and used as an area where chiefs would stay. This apparently
 l~ not how hO~holds         were organized at Ntsweng. According to oral traditions the kgosi
did not have his house at the back in a household The back, as already stated, was like a
 storage area. As Larsson & Larsson (1984) say the front of the households were normally
kept clean whilst the back yard was not always laken good care of suggesting that the area
was not at all sacred                                                    '
th Huffman (1986) furthe~ ~es          that huts were arranged with the right half for male and
   e left half for female aCtiVities. According to my informants, arrangement of houses were
 such that wherever the door opened, behind it would be the sleeping area. The far end of
the house was used as. ~ storage area, and the hearth could be on the other end. The people
of Ntsweng were )mtrilineal and paid bride wealth (bogadi) in marriage transactions. For a
~       to be seen as Weal~y he was judged according to the number of cattle he had. So as
m: ~
:~~~~~f
         e
             (1986) says this settlement pattern was organized such that cattle were in the
                     ~:~~ntral
           as they were the principal form of wealth. This was the case at Ntsweng where
                                    kraal and where they were well protected as they were the
lin~e
       .
          ~:~c:    0
                      pattern at ~tsweng. can be seen to have been or~ed        along strUcturalist
                      were asSOClated With cattle, were buried at the center where cattle were
                                                 36


    kept. Women were associated with the periphery which was the household and so when
    they died were buried there. Also in households there was the back lelapa which was
   associated with dirt and the front lelapa which was associated with cleanliness.
       Settlement at Ntsweng was hierarchical. It was divided into makgot/a which were all
   under one leader, the kgosi-e-kgolo (paramount chief). The type of social organization that
   existed at Ntsweng is what Kent (l990b:139) classes as category IV society. These, she
   says, had
       hereditary chiefs with formal power and inherited sociopolitical stratification in the form of classes.
      They often also have hierarchical individual ranking. These societies are usually characterized by
      full-time economic and sociopolitical specialization. (Kent, I 990b: 139)
      According to Mr S. Masimega chieftainship amongst the Bakwena was hereditary and the
   heir to the throne was usually the first born son of the senior house. At Ntsweng four
   important makgotla besides the kgosi-e-kgolo's are known to have existed. These were
  Difetlhamolelo, Senyeima, Maribana and Mollale. These makgotla were said to have been
  started by Sechele when he first settled at Ntsweng and they were named after the four men
  who wandered with him in the wilderness when he was homeless.
      Besides these makgotla there were others namely, Maunatlala, Mokgalong, Segosebe and
  Ntloedibe. Ntloedibe, however, was said to be the most important of the four (recording of
  Ratlou Sebetlela). This is because Ntloedibe is said to have been older than Sechele, but
  because his mother was not married by traditional expectations he could not be the chief of
  Bakwena. Oral information has it that during the first harvest (go loma) Ntloedibe is said to
  have had first right to the tasting of the produce and from then on his kgotla was given this
  right and not the leadership right Because of this, his kgot/a was treated as the most
  important kgotla of the other makgotla by the kgosi-e-kgolo. The dikgosi of the four
 makgotla were said to have been the closest relations to the kgosi-e-kgolo. Each of these
 eight makgot/a according to oral information had small makgotla under them. When a
 message was passed from the kgosi-e-kgolo, it was the responsibility of the leaders of the
 eight makgotla to pass the message to the small makgot/a.
      The social organization at the family level was that the family was represented by the
 man at the family kgotla where matters were discussed and then the kgosana (headman) of
 the kgotla would take the message to the kgosi (chief). From here the kgosi would take the
 message to the kgosi-e-kgolo where it would finally rest This was also how court
 proceedings were carried out. The most difficult cases like that of murder were the ones
 that reached the kgosi-e-kgolo, but small disputes like that of the family were nonnally
 solved by the kgosana. If the kgosana had problems solving the case it was passed on to
the kgosi and finally the kgosi-e-kgolo if need be (Schapera & Comaroff, 1991).
     Kgosi-e-kgolo is the one who gave the final word as to when ploughing could begin (go
bolotsa letsema) and when crops could be harvested (/egwet/a). In each case, before
everybody could start ploughing their masimo (fields) they first had to plough and plant the
kgosi's fields which were known as masotla. Before the eating of produce at the fields
could begin, a ceremony had to be held whereby part of the people's produce would be
taken to the kgosi. This was known as go isa dikgafela. The grain would then be stored in
large grain bins and would be redistributed to the morale during times of famine.
     Amongst Bakwena there were the nobles, commoners and those who had migrated into
the group. They all recognized the position of the kgosi who also reinforced his power by
providing food for the poor and entrusting some of his cattle to them. This was known as
                                                    37


 the mafisa system so that the poor could look after these cattle and benefit from the milk
 and calves produced by the cattle.
     When we look at the settlement pattern at Ntsweng, the household with a house having a
 concrete floor was located in a kgotla that had more structures than all the other makgot/a.
 According to Me S. Masimega, the kgosi's kgotla was usually larger than the kgosana's and
 it could be that cluster 3 was a kgotla of a kgosi while clusters 2 and 4 were makgotla of
 dikgosana. This means cluster 3 could have been one of the eight main makgotla at
 Ntsweng with the other clusters being under the order of the kgosi at cluster 3.
 Conclusion
 The main objective behind mapping Ntsweng was to reconstruct the organization of the
 settlement and to assess the damage done by the bulldozer. The total area surveyed at
 Ntsweng covered 56,000 sq m whilst the bulldozed area covered 24,000 sq m. Phosphate
 analysis was only useful in that it produced a high concentration of phosphate in the central
 space and low results where structures were. These results, with other archaeological
 techniques like phytolith analysis can help identify kraals. From this it may be possible to
 determine how many related features (e.g., makgotla) had been destroyed by bulldozer.
    The mapping of Ntsweng enabled the creation of a map that recorded the organization of
 settlement at Ntsweng. In this it was possible to recognize houses and malapa which were
 organized to form settlement units. The pattern in these settlement units was such that
 structures were built forming a circular shape with a central space forming a kgotla. The
 circular house was the dominant form in the settlement. There were also semi-circular and
 a few square houses. The square houses were mostly found in the largest settlement unit.
 This shape of house was found to mark status as it was mostly associated with the elite.
     The organization of the settlement at Ntsweng was hierarchical. From the archaeological
 data, I found that there were small settlement units that surrounded a large one. This was
 supported by oral and historical data that said that within the settlement at Ntsweng there
 was the kgosi-e-kgolo. kgosi and kgosana. The kgosana in the archaeological record was
 represented by the small makgotla. The large kgot/a represented that of the kgosi, and the
 kgosi-e-kgolo's kgotla was where the national office was located.
    From .oral and historical data I found out that at Ntsweng division of labour was alon~
 gender Imes rather those of age or status. Men were associated with the central space and It
 is in this area that they were buried. This indicates that the use of space within the
 settlement was along gender lines. However, within a household this was not the case.
 S~~s         were organized according to back and front. The use of space also differed
WIthin structures. Here the use of space was governed by the direction in which the door
opened.
    ~t was ~sibl~ from the map to test the applicability of Huffman's central cattle pattern.
WIth the aId of Information gathered about Bakwena it was evident that the settlement at
N~eng conformed to Huffman's model although in some cases like the use of space
WIthinstructures it does not appear to be applicable .
        .                                           Notes
Ms. Princess ~ekgarametso is currently curator at the Kgosi Sechele I Museum in Molepolole.         Her
thesIS, supemsed by Andrew Rel<l, was completed in 1995. The original contains eight photographs
~ stmc~        at Nts~~      which were left out of this edited version. A few minor illustrations were
   so onutted. In addition, the text was pruned and lightly edited for grammar.
                                                    38


                                                 References
Oral Sources
Mr. T. Diago: farmer, 94 years, Botlhaganeng ward.
Mr. S. Masimega: farmer, 80 years, Goo Tshosa ward.
Ms. S. Mmopelwa: elder, 90+ years, Bokaa ward.
Mr. R Pheto: farmer, 77 years, Bokaa ward.
Ms. M Bodigelo: member of Matsosa-Ngwao, 93 years, Goo Kodisa ward.
Ms. L. Sebele: member of Matsosa-Ngwao, 63 years, Kgosieng ward.
Ms. S. Moitlamo: member of Matsosa-Ngwao, 61 years, Goo Kodisa ward.
Ms. N. Rantaolana: member of Matsosa-Ngwao, 63 years, Difhetlhamolelo ward.
Ms. M Modise: member of Matsosa-Ngwao, 62 years, Mokgalong ward.
Ms. N. Bodigelo: member of Matsosa-Ngwao, 62 years, Goo Kodisa ward.
Ms. N. Moleele: member of Matsosa-Ngwao, 61 years, Gamollale ward.
Ms. B. Senuku: member of Matsosa-Ngwao, 64 years, Gamollale ward.
Ms. K. Bolele: member of Matsosa-Ngwao, 65 years, Gamosima ward
Ms. S. Lebeko: member of Matsosa-Ngwao , 72 years, Mokgalong ward.
Ms. N. Kgosidintsi: member of Matsosa-Ngwao, 62 years, Mokgalong ward.
Ms. A. Makodumentsa: member of Matsosa-Ngwao, 82 years, Bakwena Senyana ward.
Mr. A. Pheto: farmer, 80 years, Bokaa ward
Ms. M Gabaitse: elder, 93 years, Sekamelo ward.
Ms. M Gabaraane: elder, 79 years, Sekamelo ward.
Ms. M. Gaaratwe: elder, 90+ years, Sekamelo ward.
Mr. J. Ramsay: teacher, Gaborone.
Botswana National Archives
BNA BNB 9719 ("Preliminary look at aspects of the reign of Kgosi E Kgolo Sebele II of Bakwena
   prior to the time of his exile" Jeff Ramsay March 1985).
BNA S.260/13 _ 260/15/2 (Removal of Borakalalo and hill villages to central site. Location of kgotla
   residence and water supply.)
Recordings
Ratlou Sebetlela, 26-07-93, recorded by S.J. Knudson, Museum Tape I, Kgosi Sechele 1 Museum,
   Molepolole.
Published Sources
Bakkevig, S. (1980) 'Phosphate analysis in archaeology: problems and recent progress' Norwegian
   Archaeological Review, 13(2): 73-100
Borgesen, I.M., P. Lane, J. Ramsay & A. Reid (1994) 'Rescuing Ntsweng' Zebra's Voice, 21 (2/3):
   7-8
Caister, D. (1980) 'Archaeological         perspectives on settlement patterns in South East Kweneng
   District', 87-91 in RR Hitchcock & MR Smith (eds.) Settlement in Botswana Gaborone:
    Botswana Society
 Clark, A. (1990) Seeing Beneath the Soil: prospecting methods in archaeology London: B.T.
    Batsford Ltd.
 Campbell, A. (1991) 'Distributions of prehistoric Kalanga villages in the lower Shashe and Letsibogo
    areas', 19-33 in C. van Waarden (ed) Kalanga Retrospect and Prospect Gaborone: Botswana
    Society
 Huffinan T.N. (1986) 'Archaeological evidence and conventional exp1anations of Southern Bantu
    settlement patterns' Africa 56: 280-297
 Kent, S. (ed.) (19908) Domestic architecture and the use of space: an interdisdplinary cross-cultural
    study Cambridge: Cambridge University Press.
 Kent, SA (l990b) 'A cross-cultural study of segmentation, architecture, and the use of space', 127-
    152 in S. Kent (ed) Domestic architecture and the use of space: an interdisciplinary cross-cultural
    study Cambridge: Cambridge University Press.
                                                       39


 Knudson, S.1. (1978) Culture in Retrospect: an introduction to archaeology Chicago: Rand Mc Nally
 Larsson, A & V. Larsson (1984) Traditional Tswana housing: a study In four VIllages In eastern
    Botswana Stockholm: Swedish Council for Building Research.
 Lawrence, RJ. (1990) 'Public collective and private space: a study of urban housing in Switzerland',
    73-91 in S. Kent (ed.) Domestic architecture and the use of space: an      interdisciplinary cross-
    cultural study Cambridge: Cambridge University Press.
 Maggs, T.M (1976) Iron Age communities of the southern highveld Pietermaritzburg: Natal Mu~um
 McDonald, C.A. (1940) The material culture of the Kwena tribe of the Tswana.               Unpublished
    M.A.Thesis, University of South Africa.
 Morton, F., A Murray & 1. Ramsay (1989) Historical dictionary of Botswana Metuchen: Scarecrow
    Press.
Morton, F. & 1. Ramsay (1990) The Binh of Botswana: a history of the Bechuanaland Protectorate
  from 1910 to 1966. Gaborone: Longman
Nangati, F. (1980) 'Early capitalist penetration: the impact of precolonial trade in Kweneng (1840-
    1876)', 140-147 in RR Hitchcock & MR Smith (eds.) Settlement in Botswana Gaborone:
    Botswana Society
Ngcongco,1. (1982) 'Precolonial migration in south-eastem Botswana', 23-80 in R.R Hitchcock &
    M.R Smith (eds.) Settlement in Botswana Gaborone: Botswana Society
NWlez,M (1990 ) 'Phosphate Survey' Norwegian Archaeological Review 23: 120-127
Okihiro G.Y (1973) 'Resistance and accommodation: baKwena - baga Sechele 1842-52' Botswana
    Notes and Records 5: 104-116
Okihiro, G.Y. (1976) 'Hunters, Herders, Cultivators and Traders: intemction and change in the
   Kgalakgadi nineteenth Century' unpublished Ph.D. dissertation, University of Carlifornia, Los
   Angeles.
Pahl, R (1971) 'Notes on an Iron Age settlement on Motsenekatse Hill' Botswana Notes and Records
   3: 27-31
Pistorius, 1.C.C. (1992) Molokwane: an Iron Age Bakwena Village. Johannesburg: Perskor
Ramsay, 1. (1991) 'The Batswana-Boer war of 1852-53: how the Batswana achieved victory'
   Botswana Notes and Records 23:193-208
Renfrew, C. & P. Bahn (1991) Archaeology:. theories. methods and practice London: Thames &
   Hudson
Schapera,1. (1943) Native Land Tenure in the Bechuanaland Protectorate Alice: Lovedale Press.
Schapera,1. (1980) 'Notes on the early history of the Kwena' Botswana Notes and Records 12: 83-89
Schapera,1. & J. Comaroff (1991) The Tswana. revised edition London: Kegan Paul International
Segobye, AK (1994) 'Farming communities in Botswana: an archaeological study of land use and
   settlement ill the Mokgware hills c. 10th-15th centuries AD' unpublished Ph. D. dissertation,
   Cambridge University
Sekgarametso, P. (1995) 'An archaeological survey of Ntsweng in Molepolole' unpublished B.A.
   thesIS,History Department, University of Botswana
Shackley. M (1981) EnvirOflmental archaeology London: Allen & Unwin
Sharer, R1. & W. Ashmore (1979) Fundamentals of Archaeology Menlo Park: Benjamin Cummings
Sillery, A (1952) The Bechuanaland Protectorate London: Oxford University Press
Tlou T. & A Campbell (1984) History of Botswana Gaborone: Macmillan
van Waarden, C. (1980) 'Archaeological investigation of Leeukop: a fimctional approach' Botswana
  Notes and Records 12:151-164
van Waarden, C. (1989) 'The granaries of Vumba: interpretation of a Khami period commoner site'
  Jou17lll1of Anthropological Archaeology 8:131-157
                                                  40


   Fig. 1. Map showing the location of Ntsweng in southeastern Botswana.
Front                                                                 Back
              Fig. 2. The organization of a Kwena household
                                      41


                    rl stones                                                                O~30
       ~.t15 #. •.....!I.-
                                                          ~
                                                                 "                 '1
                     -'
                             -""
                                                                                                    .•.            grinding
                                                ~. '.l-
     t1~~                                                                                      ..~i ·
                                                            \                                     15 ...           stones
                                                        \.    -~        1_
                                    1                               ... ../
        -.J" ~
     ("t
                                                                                             '-i                  16
    ,-..:r                    ....   "  ...
                                                                    x
                                                                                                     r '1
                                          I
       \.
                       ... ,.,'-k
                      ,.
                         4>
                                   ..., ,                                                                    ,...
                     .. ,.
                 ,.../,.,.- f                                      ,f\],
                                                                                                        l)~
                                            I
             .~                    ,,," 32         .... -\         _, r-
              :•
               :
                         •    •-
                                 ,   33
                                                       .    '1,.... ,)•                 r
                     ~     34
                                                C!' '"                7     ~
                                                                                                                "J r
                                                     " "J'
                                                                                                                                 N
Cluster 1                                                                             U
                             Fig. 3. Part of Ntswengshowing the ruins in cluster 1.
 Ntsweng: mapped portion                                                                                          m
                                                                                                      OL..JL.["l...J50
                                              I
                                                •
                                                 I
                                                   ."0 '•.-
                                                         ~
                                                                                     ~                  f1
                                                                                 z: t- .... t...i:"
                                                                                   .l...
                                                                                ,.~~'f.)               • '- ..:~.
                                                                                                                        .,
                                                                                                                           ~1
                                                                              ."          •..• -                    1 ~
                   Cluster 4                                                         . ...
                                                                                      ,. (,.'
                                                                                 .-::=':".:. t~
                                                                                                     Cluster
                                                                                                            .J'.
                                                                                                                           .':.•
                                                                                                                              ~
   '~    ;'1,   ~~I      ...                                                                      - ..     --1'            ,.
                r- ~~
                     •
                                                                             -'
                                                                                    -            ~-~        F     r
                                                                                                             - ... --
                                                                                                                     -:      )
 - - - -                                                                                                                         N
                                                                                                                               r-
                                        Fig. 4. The mapped portion of Ntsweng.
                                                             42


Fig. 5. Historical photographs ofNtsweng: date and orientation unknown.
                                     43


01         1     m               I 30
                                                                          •
                                                                   grinding
                                                               •   stones
          n
         ,        \
         -- -                                                     Cluster 2
             Fig. 6. Cluster 2 of the mapped portion of Ntsweng.
Cluster 3
                                                            53 l. '
                                                                o   <> ,v,..
                     Fig. 7. Cluster 3 of ruins at Ntsweng.
                                     44


   Cluster 4
                                                -J
                                                                     )
   ...      ...,
     r'..,I
      \
    o L-...IfilL--l   30
                 Fig. 8. Cluster 4 of the mapped portion of Ntsweng.
                                       Cluster 5
0~30
                         Fig. 9. Cluster 5 of the ruins at Ntsweng.
                                           45