NYAME AKUMA No. 45 June 1996

expansion of BaNgwato power and political authority during the last quarter of the nineteenth century (Parsons 1973). Along with Sebele I and Bathoen I, the respective rulers of the Kwena and Ngwatketse, he also travelled to Britain in 1895 so as to prevent the transfer of the Bechuanaland Protectorate to Cecil Rhodes' British South Africa Company. Preliminary Archaeological Investigations at Phalatswe (Old The period from c.1850 onwards also saw the Palapye), Central District, Botswana: expansion of European trade into the territories The 1995 Season occupied by the various northern Tswana commu- nities. Although Europeans had already ventured into the area by the early years of the 19th centu- p owe Borjeson ry, it was only after the opening up of a route Department of Human Geography around the north-eastem edge of the Kalahari to University of Stockholm Lake Ngami by David Livingstone, and his two S- 106 9 1 Stockholm companions Oswell and Murray in 1849 that trad- Sweden ing activities began to intensify. By the time Paul J. Lane Khama I11 moved his capital to Phalatswe, most Archaeology Unit of the routes north were well established, and var- Department of History ious settlements already supported a number of University of Botswana European or Indian trading stores. The construc- PB 0022, Gaborone tion of a railway line linking Mafeking in the Botswana south with Bulawayo in the north, beginning in 1895, further facilitated the expansion of European trade and by the time of its completion in 1897 there were several halts along its route Introduction within the territories of the various Tswana merafe (states). These included Palapye Road Siding, sit- The site of Phalatswe, or Old Palapye as it is uated roughly 20 krn WNW of Phalatswe, which more commonly known, is situated towards the rapidly became an important staging post for north-western end of the Tswapong Hills in the wagon-traders on their way north-west to Central District of Botswana, at approximately Ngamiland. 22"37'30" S latitude and 27"17'20 E longitude (Figure 1). Most of the visible remains at the site Even before the expansion of this trade, sev- consist of traces of former rectangular and sub- eral European missionaries had already estab- rectangular stone- or brick-built structures. A lished a presence among Batswana. Livingstone, number of less-substantial stone rings, piles and for example, spent almost four years among lines which probably represent the foundations of BaKwena in the south of the country, at various clay or wattle and daub houses, granaries Kolobeng, before embarking on his travels into and compound walls also survive. Interspersed Central Africa. The German Lutheran church was between these features are a series of artifact scat- also active in the region, and in 1859 established a ters of varying densities. mission at Shoshong. Khama, the first Tswana kgosi (ruler) to adopt Christianity, was baptised The site is known from documentary and oral the following year, while his father Sekgoma I sources to have served as the capital of the was still in power. In 1862, however, the Lutheran Ngwato state from 1889, following their move mission closed and was replaced by the London from Shoshong, to 1902 when the capital was Missionary Society (LMS) under John relocated again to Serowe (Parsons 1982). Mackenzie. Mackenzie and his wife were suc- Throughout this period BaNgwato were ruled by ceeded by James Hepburn in 1870, who moved Kgosi Khama 111 (c. 1835-1923). Khama was a with BaNgwato to Phalatswe in 1889. Following man of considerable political vision, and is gener- a falling out with Khama 111, Hepburn left the ally held to have been the driving force behind the NYAME AKUMA No. 45 June 1996

Figure 1 : Location of Phalatswe and other settlements mentioned in the text. NYAME AKUMA No. 45 June 1996

Figure 2: Location of areas planned in 1995-based on an orthomap prepared by the Dept. of Physical Geography, Stockholm University using aerial photographs taken in 1988. NYAME AKUMA No. 45 June 1996

Figure 3: Detail of part of Area 2, showing the arrangement of houses, courtyard walls and granary bases and typical ward layout.

- 4 \' C \\ 0 \\ 0 " (3:::. 0 \ \ a \ \

0 fJ \ \ D \\ m n am \\ 4 \ \\ a \ 0 '\ \\ 0 Q 8 \ \ 0 0 \\ V) Q \\ \ \ @ 0' \\ \ -0 0 . - a _-' . :.-

0 u V) E z+ 0 C - u 0 0 V) D .CI- - 1 .r 0 E 2 .so- L =3 aJ zgs%r 2 0 0 .z pwP' h ,-'U+$e1 , . , . I a -' , oemC3+gv, . , 11 :: 0 8 0 80 11 :: I1 :: 0 NYAME AKUMA No. 45 June 1996

town a broken man towards the end of 1891. use of an Electronic Distance Measurer (EDM); There followed a brief period when no missionary and (b) foot survey of cultivated and abandoned was present. Eventually, the Reverend W.C. fields that surround the main concentrations of Willoughby arrived in 1893 and remained with structural remains, with limited surface collection BaNgwato until after their move to Serowe. of artifacts. Aside from acting as important catalysts for However, as a University project, training stu- change among BaNgwato, the various European dents in various techniques of survey, surface col- missionaries and traders who visited their temtory lection, map reading, drawing and on-site record- left a legacy of written descriptions, sketches and ing was also a major component of the work. In photographs of late 19th century BaNgwato hous- addition, two Year 3 Archaeology students carried es, settlements and material culture (e.g. out parallel studies on, respectively, the use of Willoughby 1900; Hepburn 1970; Johnston 1970; space and resources within the settlement, and Blackbeard 1985). These documentary materials conservation and management issues affecting the provide a valuable record of the layout and physi- site. These will form the basis of their Fourth Year cal structure of the town against which to assess dissertations (Mathibidi in prep.; Motheletsi, in the archaeological remains. prep.). Other related studies undertaken during this season included the collection of interview data and oral histories concerning former and cur- Archaeological Fieldwork rent land-use and vegetation types (Widgren and Dahlberg, work in progress), and a preliminary Despite the historical importance of foot-survey of the agricultural lands in the area for Phalatswe and its status as a gazetted National traces of subsidary settlements coeval with the Monument, there has been very little investigation main occupation of Phalatswe (Segadika, in of the archaeological and architectural remains at the site. A brief description of the more obvious prep.). The latter work is expected to form the basis of an MA study of BaTswapong settlement ruins has been published (Vorting 1990). This in the area, both prior to and following the estab- concentrates exclusively on the remains of the lishment of the Ngwato capital at Phalatswe. European ward in the settlement and the African church which, since both the gable ends remain standing, is the most visible of all the structures at The Buildings Survey the site. The more modest remains of the areas occupied by the bulk of the African population, on Three main surface concentrations of building the other hand, are not discussed. This neglect remains were surveyed over a six week period. has created the mistaken belief that nothing actu- Detailed maps of these areas were prepared at a ally remains of these; at least not on the surface. scale of 1:1000. The locations of these areas are In turn, this has encouraged a rather negative atti- shown in Figure 2. In view of the neglect of the tude towards the site by many local people, and a African elements of the town and their greater perception that it is only the European elements fragility as archaeological resources, it was decid- that are protected under the terms of Botswana's ed to concentrate the inital efforts in the two areas antiquities legislation. As a direct consequence of (Area 1 and 2) with the highest concentrations of this, large parts of the site have been brought typical Tswana compound and house remains. under cultivation since the late 1940s, and espe- After the completion of the planning of these cially since 1960. areas, work began on recording the surviving fea- In June and July 1995, as part of a broader tures in the European ward (Area 3). Although investigation of the landscape history of thls area most of this area was also planned successfully, a of Botswana (see Dahlberg et al. 1995), a start few features.stil1 have to be mapped. In addition, was made on mapping the site and at trying to there is a further concentration of mostly define its original extent. This work entailed two European elements around the former remains of separate elements that ran more or less concur- the Mission School and 'African' Church. It is rently. These were: (a) the planning of visible hoped that these areas, along with the smaller surface features in various parts of the site, includ- clusters of house and compound remains located ing both African and European quarters, with the NYAME AKUMA No. 45 June 1996

during the foot-survey will be planned during the known for certain, although preliminary ethnoar- 1996 season. chaeological investigation suggests that they As Figure 2 demonstrates, the settlement formed the base for another type of granary remains at Phalatswe are found almost exclusively known as matlolo (Mathibidi, in prep.) on uncultivated land. However, the survey clearly Area 3: The dominating settlement features in shows that the adjacent fields once held the same the European ward are rectangular house founda- type of archaeological features, and a scatter of tions, ruins and dry stone walls. Most of the more stones, sometimes clustered, is a common sight on prominent compounds are enclosed by rectangular many fields. The main elements in each of the stone walls. All but one (which has two founda- planned areas are described below: tions) contain only one ruin or house foundation. Areas 1 and 2: The planned areas of former The terms 'house foundation' and 'ruin' are used African (Tswana) settlement are dominated by here to differentiate house remains with parts of two types of extant features: round house founda- the original stone or brick walls still standing tions and granary bases (see Figure 3). A Tswana (ruin) and those where these have either entirely house foundation of this period consists of a ring collapsed or are otherwise lacking (foundation). of partly buried stones, onto which a wall of pud- Apart from these features, two rectangular dled mud was moulded. Most of those found at stone kraals, a European styled garden and the Phalatswe are disturbed, and only a few examples location of the former blacksmith's yard and asso- form a complete circular or sub-circular ring. ciated forge (indicated by, among other things, a Where encountered, such features were interpret- scatter of charcoal, iron scaling and iron axle- ed as house foundations. Traditional Tswana hubs), have been recorded in Area 3. Tko other housing also includes a lolwapa (courtyard), recorded features which have been provisionally attatched to the house and constructed in much the interpreted are stretches of former streets, marked same manner as house walls. A few lolwapa by parallel stone lines, and square shaped stone foundations survive in both Areas. In addition, constructions, likley to have been cooking hearths there are a number of stone lines which could or bread ovens. A few round house foundations, have formed other compound boundary walls. other round structures and indeterminate stone The number of compounds recorded are approxi- clusters were also recorded within the European mately 32 in Area 1 and 21 in Area 2, if all house quarter. foundations and stone lines likely to represent a former lolwapa are included. The Foot-Survey Two different types of features have been recorded as the foundations of former granaries The aims of the foot-survey were to try to (difala). In Area 1, the granary bases are made of establish the former extent of Phalatswe, to collect stones forming a low (0.1-0.2 m), tight cluster information about variations in artifact densities with a round or oval shape. The granary bases in across the site, and to record any isolated features Area 2, with one exception, are of a different type or structural remains that might be encountered consisting of a centre stone flanked by a broken during the survey. As most of the former area of circle of satellite stones. In contrast to Area 1, rel- settlement has been brought into cultivation over atively fewer granary bases occur in Area 2 in pro- the last eighty or so years, the majority of the portion to the number of extant compounds, i.e. areas investigated were either fields or abandoned house and lolwapa foundations. The compounds fields. In a few places patches of uncultivated in Area 2 are generally more complete, but also scrub or woodland were also searched. Several of seem to be of a slightly different design. In both these latter areas revealed traces of relatively well- areas a variety of stone clusters without any spe- preserved features that included house founda- cific form occur. These were recorded as 'indeter- tions, compound walls and graves. More detailed minate stone clusters', and many may represent investigation of these will be carried out in 1996. disturbed house or granary foundations. In addi- A composite 1:20,000 orthophoto map of the tion, in Area 1 three stone features comprising 4-5 area (compiled by the Department of Physical stones arranged in a square were plotted (marked Geography at Stockholm University), was used to 'Sq' in Figure 3). The function of these is not NYAME AKUMA No. 45 June 1996

locate each field or similarly defined unit. These establish whether there is any distinctive pattern- were then line-walked by teams of between 6 to 9 ing to the distribution of different vessel types people, spaced roughly 30 metres apart. Larger and/or decorative motifs, since this may aid the fields were walked in a number of sweeps in alter- identification of the location of the wards of the nate directions, whereas smaller fields were different ethnic groups (such as BaKalanga and walked in only one direction. A total of 102 BaTswapong) known to have resided at 'fields' were searched in this manner over two Phalatswe. It may also provide an indication of week period in July, amounting to an area of differences in the economic or social status of dif- approximately 10 km2. ferent household clusters. For each field the type and density of vegeta- Secondly, despite the predominance of local tion cover, crops, degree of weathering and num- material, the survey results suggest that all sec- ber and species of trees present were recorded. tions of the community had access to imported The presence of any structural remains, their goods of European origin. Glassware, porcelain, approximate location in relationship to existing factory-made iron objects and beads were field boundaries and condition were also noted. observed in considerable quantities in many areas. Particular attention was paid to the density and In all but a few cases, where surface artifacts were distribution of surface artifacts. Densities were present these included at least some trade items. graded into four rough categories - None; Low (c. Identification and analysis of this material will c 1 per 10 m2; Moderate (c. 5-10 artifacts per 10 provide the type of detailed information regarding m2) and High (c. > 10 artifacts per 10 m2). Each the specific classes of goods being imported by 'field' formed a distinct collection unit. As a gen- European traders and local patterns of consump- eral rule, only a sample of diagnostic artifacts tion that are lacking from the available documen- were collected. These included rim and decorated tary sources. body sherds of local and imported ceramics; Some interesting points have already bases, necks and decorated segments of glass bot- emerged. For instance, one of Khama 111's more tles and jars; various types of glass beads; miscel- controversial edicts was the banning of the trade laneous iron objects including axe heads, hoes, in liquor within his capitals. The discovery of knives and plough shares; and a range of cartridge several fragments of glass beer bottles and others cases. Querns and rubbing stones were noted but which may be from wine and or spirit bottles, sub- not collected. stantiates the limited doucmentary evidence that In addition to the general fieldwalking survey, this ban was not entirely successful. Certain pat- two fields were gridded into 50 m squares for con- terns in the supply of weapons may also be evi- trolled, 'total' collection. The primary objectives dent from the surface material. Thus, out of a of this exercise was to investigate the degree of total of fourteen cartridge cases collected, at least variation in the densities of surface artifact scat- six different types and four different manufactur- ters at the site, and their potential value as indica- ers are represented. It still needs to be demon- tors of the spatial organisation of the settlement. strated, however, that all of these surface finds are Accordingly, the two fields selected both con- coeval with the occupation of Phalatswe (pace tained traces of a number of ploughed out domes- Dryer 1994). tic structures. These features were also planned at As well as providing an indication of the 1:100 scale and added to the general plot of arti- range and density of material culture to be found fact densities. in different parts of the site, the field-survey ele- ment helped to locate a number of previously Results and Discussion unrecorded structural remains, including house foundations, lolwapa (courtyard) walls, granary Analysis of the artifacts recovered during the bases and graves. Many of these are isolated foot-survey program is still at a very preliminary examples, surrounded by areas of cultivation and stage. A number of observations, however, can be the disturbed traces of similar features. made. First, the bulk of the surface material Knowledge of the location and condition of these observed can be considered to be of local origin, features will assist the preparation of a compre- mostly pottery. This material will be examined to NYAME AKUMA No. 45 June 1996

hensive plan for the management of the entire set- ters have now been planned at Phalatswe indicat- tlement. ing that the former capital had a similar social Even more unexpected was the discovery of structure. Although the layout of these was basi- the remains of two different lines of metal tele- cally the same, there appear to have been differ- graph poles. These poles appear to have been ences in granary type and compound design deliberately broken off at about 1.5 m above between wards. At present, however, it is not pos- ground-level, so that only the stubs remain in situ. sible to say if these differences were correlated So far, these lines have only been followed for a with differences in the ethnic affiliation of the res- couple of km away from the site. One runs rough- idents, with some other socio-economic variable ly east-west and therefore probably extended orig- or reflect chronological developments. In order to inally as far as Palapye Road Railway Siding, resolve this, a series of test excavations in differ- where a telegraph office is known to have existed. ent wards are planned for 1996. The other line runs roughly north-south and may The surface survey will also be continued in mark the line of the former coach route between 1996 so as to try to define the limits of the origi- Shoshong to the south-west and Tati to the north. nal settlement. This will enable an assessment of An objective in 1996 will be to try to trace both the various estimates made of the size and popula- lines along their full extent. Accurate plotting of tion of the town by different European visitors and their alignment may also provide a clue as to residents. According to one account, first pub- which of the ruined structures in the 'European' lished in October 1890 in the Cape Argus newspa- quarter represents the remains of the former per, Phalatswe covered roughly twenty square Telegraph Office depicted in Willoughby's miles and had a population of 30,000 inhabitants account (1900). (cited in Parsons 1982:121). James Johnston, who By using the photographs in Willoughby visited the town in March 1892, described it as (1900), two of the recorded compounds can easily covering only sixteen square miles and with a be identified as the 'Court House' and the population of 15,000 (1969:229). Bryce says that 'English Church', as Willoughby terms them. the settlement was "two square miles" in extent in Further, the foundation of what most likely was 1895 (1898:218). The trader Samuel Blackbeard, the 'General Store' with adjacent buildings and who took part in the move from Shoshong, put the the location of the 'main street' have been identi- number of people involved as 20,000 (1985: 190). fied, even though the evidence from Willoughby is The provisional interpretation of the 1995 results not as clear as in the case of the 'Court House' or suggests that the area of settlement was signifi- 'English Church'. Another picture in Willoughby, cantly smaller than all of these estimates other of the 'native elementary school, bears some simi- than that by Bryce, who may have been referring larity to two of the other planned house founda- to just the central portion. However, the eastern tions. A few more features in this general area, and northern limits of the settlement still have to however, still have to be mapped and only then be defined. More work is also needed before it will a comprehensive comparison with the avail- will be possible to reconstruct the precise layout able documentary and photographic records be of the town and the composition of the different possible. wards. Once this is completed it may be possible to determine the likely size of the resident popula- In terms of the African elements of the settle- tion. ment, the main achievement in 1995 was the map- ping of the two most extensive extant portions. The spatial layout of these resembles that of more Acknowledgements recent settlements, including the new capital Serowe, which are typically organised into a num- Permission to carry out the fieldwork was ber of separate wards. Members of each ward granted by the Office of the President, erect their household compounds close to each Government of Botswana under Research Permit other in a large circle or horseshoe shape facing No. OP 4611 XXXiX (81). Thanks are due to both an open space, usually containing a common cat- SAREC and the University of Botswana for grants tle kraal (lesaka) and a kgotla (court and meeting towards the costs of fieldwork. We would also place). On this basis, at least five clear ward clus- like to thank Alinah Segobye, Mats Widgren and NYAME A KUMA No. 45 June 1996

Annika Dahlberg for their comments on earlier Mothelesi, G. drafts of this report, and for inviting us to join the in prep. A District-scale perspective on Landscape History Project. Finally, we owe a lot Archaeological Heritage Management to all the UB students who participated in the issues in Botswana. BA dissertation, fieldwork and whose efforts contributed greatly to University of Botswana, Gaborone. such a successful season. Neil Parsons kindly commented on an earlier draft. Parsons, Q.N. 1973 Khama 111, the Bamangwato and the British - with special reference to 1895- References 1923. Unpublished Ph.D. dissertation, Blackbeard, M. University of London. 1985 Memoirs of Samuel Blackbeard (1930). 1982 Settlement in east-central Botswana, c. Botswana Notes and Records 1800-1920. In R.R. Hitchcock 17:187-191. and M. R. Smith, editors, Settlement in Botswana. Gaborone: Heinemann Bryce, J. Educational Books, pp. 115- 128. 1898 Impressions of South Afnca. New York: The Century Co. Segadika, P. in prep. Politics and strategies in the Tswapong- Dahlberg, A., A.K. Segobye and M. Widgren Ngwato landscape. MA Thesis, 1995 The utilisation of land and resources in University of Wales, Lampeter. the past: A study of Old Palapye settle- ment. Paper presented at the 10th Vorting, H.C. Congress of the Pan African Association 1990 Old Palapye the deserted capital. for Prehistory and Related Studies. Lekghapo - Khama III Memorial Harare, June 1995. Museum Review 1988-89 1: 34-39. Dryer, J. Willoughby, W.C. 1994 Cartridges and chronology: An exercise 1900 Native life on the Transvaal border. in relative dating. Southern London: Simpkin, Marshall, Afncan Field Archaeology 3: 44-48. Hamilton, Kent and Co. Hepburn, Rev. J.D. 1970 [I8951 Twenty years in Khama's country. London: Frank Cass and Co. Johnston, J. 1969 [I8931 Reality versus romance in South Central Africa. London: Frank Cass and Co. Mathibidi, N. in prep. Landscape history in Botswana: A study of the layout of settlement and spatial organization of Bangwato society at Old Palapye. BA dissertation, University of Botswana. Gaborone.