Original Research

A historical archaeological investigation of Sardelli’s shop, Sabiepoort, Kruger National Park

Anton C. van Vollenhoven

Received: 29 Sept. 2023; Accepted: 10 Jan. 2024; Published: 06 Mar. 2024

Copyright: © 2024. The Author(s). Licensee: AOSIS.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Historical information indicates that Sardelli’s shop at Sabiepoort, built before the South African War because of its proximity to an existing trade route, was utilised by Steinaecker’s Horse during the war (1899–1902). This shop was likely a satellite of another site, the Gaza Gray outpost. Steinaecker’s Horse was a British volunteer unit operating in the Lowveld and eSwatini. Research on the Sabiepoort site forms part of a larger project, the Steinaecker’s Horse research project, aimed at undertaking archaeological and historical research of sites associated with this unit. The site was excavated with the aim of contributing to the contextualisation of information and to show the extent of the connection between historical information and archaeological evidence with regard to Steinaecker’s Horse at Sabiepoort. Dateable artefacts from the site represent the period of the late 19th to the early 20th century, coinciding with the dates of the war. Archaeological evidence indicates the presence of indigenous people present during the same period, most likely Tsonga. Indigenous pottery at the site is similar to those discovered from the Gaza Gray site, confirming concurrent occupation. Research findings were compared to outcomes from the Makhadzi outpost, indicating social differentiation between the different units stationed at this outpost close to the Letaba Rest Camp and gave evidence of the role of black soldiers and local inhabitants during the war. There is a distinction between the main aim of the Steinaecker’s Horse project and the aim at this site. The latter is described here with the former aimed at writing the history of this unit.

Keywords: Sardelli’s shop; Sabiepoort; Steinaecker’s Horse; Bill Sanderson; Gaza Gray; corrugated iron; Tsonga; Kruger National Park.

Introduction

The South African War occurred between 1899 and 1902 and was fought between Great Britain and the two Boer republics, Transvaal (Zuid-Afrikaansche Republiek) and the Orange Free State (Pretorius 1999:247). It was initially referred to as the Anglo-Boer War, failing to recognise the many other population groups that played their roles in the conflict, supporting either of the two sides whose tension led to the war. Steinaecker’s Horse was a British volunteer military unit that was active during this war. They operated mainly in the Lowveld and eSwatini (Pienaar ed. 1990:343; see also Woolmore 2006, Van Vollenhoven 2010, Van Vollenhoven & Pelser 2004).

The unit served as a border control, preventing the Boers from making contact with pro-Boer people in the Portuguese East Africa (Mozambique). For this purpose, they established several outposts along the eastern border of the country and eSwatini. One of these outpost sites was located at the shop of Sardelli at Sabiepoort. Sardelli was a Greek businessman and later served as a member of Steinaecker’s Horse (Woolmore 2006:282). Little other information is available on him and even a search in the National Archives of South Africa revealed no documents. He used the site between 1892 and 1899 and had two other shops in the vicinity, one at Gomondwane and the other at the Crocodile River (likely at Crocodile Bridge). The site of Sardelli’s shop at Sabiepoort is situated approximately 7 km south-east of the Lower Sabie Rest Camp in the Kruger National Park (Figure 1). This is about 170 m from the southern bank of the Sabie River, and it is right next to the western bank of a small tributary thereof, the Nhlanganzwani Spruit. The site is believed to be a satellite of the nearby Gaza Gray outpost of Steinaecker’s Horse, situated approximately 2 km south-east thereof and consisting of four separate geographical sections, separated by natural bush. The distance between the main site and the shop is minimal, and it is therefore believed to only be an extension (or fifth section) of the Gaza Gray outpost.

FIGURE 1: Map indicating the location of the Sabiepoort Site and the nearby Gaza Gray Site in the Kruger National Park.

The remains of the shop consist of a concrete floor, with no walls. Two deflated middens were identified, both towards the south on slightly elevated areas, containing a concentration of cultural artefacts. The first is about 75 m from the shop and the second approximately 150 m (Figure 2). In between these features, cultural materials are found scattered over the surface in a diameter of about 220 m, stretching further towards the south. Although spread out over a reasonably large area, the cultural objects are not found in abundance. Imported European and indigenous artefacts are intermingled, but the former are concentrated around the shop, with the latter being concentrated at the midden. This indicates a geographical differentiation of people on site with Sardelli and the soldiers around the shop and local people further away.

FIGURE 2: Site map of the Steinaecker’s Horse Sardelli’s Shop site at Sabiepoort (the site where Bill Sanderson was interned).

Problem formulation

The Steinaecker’s Horse research project is aimed at investigating, in terms of archaeology and historical research, various sites associated with this British force. The main aim of the Steinaecker’s Horse project, however, needs to be distinguished from the specific aim at this site as each site has specific aims within the larger project because of the uniqueness of the archaeological evidence at each. There were seven sites excavated as part of the Steinaecker’s Horse project, namely the site of the Sardelli’s Sabiepoort Shop, his shop at Gomondwane, Makhadzi, Sabi Bridge, Gaza Gray, KaNwumariwane and Ngotso Mouth. The historical and archaeological research at the Makhadzi site indicated that there was social differentiation between the different units. Archaeological excavations gave an indication of the role of black soldiers and local inhabitants during the South African War (Van Vollenhoven 2010:181). These local inhabitants comprise different people residing in the area before and during the war. After completion of the excavations at Makhadzi, it was deemed important to contextualise this information by comparing and contrasting other such sites in this regard. Also, too little was known about the everyday life and circumstances of the unit, especially relating to comparative samples. This necessitated the excavations of the other sites. The sites at Gomondwane, KaNwamuriwane, Sabi Bridge and Ngotso Mouth seem to not indicate such a differentiation although it is possible that the Sabie River might have provided such a division, with white and black people staying on opposite sides of the river at the Sabi Bridge site (Van Vollenhoven 2010:235). As Ngotso Mouth was only used as a temporary camp at a place where local people already stayed, it explains the lack of differentiation (Van Vollenhoven 2014:7). This social differentiation also has existed at the Gaza Gray site (Van Vollenhoven 2016:86). The research at the Sabiepoort shop is regarded as the further investigation of this aspect, aimed at determining whether such a differentiation also existed here but also to obtain information to assist in reconstructing the history of the Steinaecker’s Horse unit.

Historical context

The history of the site discussed in this article is closely associated with one Demetrius Sardelli, who was also called ‘Sardelli the Greek’. He was a member of the Steinaecker’s Horse unit. However, he and his business associates, Tom Paulin and Charlie Woodlands, operated three shops in the Lowveld before the war. Two of these shops were utilised by Steinaecker’s Horse. The third one was close to Komatipoort along the Crocodile River. At the Sabiepoort site, a farmer by the name of Bill Sanderson was interned by Steinaecker’s Horse during the war as he was on commando with the Boers for a short time. As he was older than 60 years, the Boers allowed him to return home (Pienaar 2012a:321, 415; Woolmore 2006:282).

Sardelli was born in Greece on 25 October 1864. It is not known when he came to South Africa; but during the 1890s, he was a resident of the Lowveld. With his business partner, Tom Paulin, he stayed on the Lebombo Mountain range near Zenga Mountain (Woolmore 2006:282). Sardelli had eucalyptus trees planted at the shop at Gomondwane, perhaps an indication that he resided there. It was believed in those days that these trees would prevent malaria (Minnaar & Pienaar 2012:355; Pienaar 2012b:347). Sardelli’s wood and iron shops were apparently mainly erected to do business with black miners returning from the Witwatersrand by selling them wares of inferior quality and homemade alcoholic concoctions at high prices (Pienaar 2012a:321).

When the South African War broke out, he abandoned his shop at Gomondwane (and most likely the other two as well) and went to Delagoa Bay. With him was Woodlands, who is said to have operated the store at Gomondwane. It is said that only Sardelli and Woodlands’ money reached the bay, and the latter was never seen again (Woolmore 2006:282). Sardelli later returned as he enlisted into Steinaecker’s Horse at Komatipoort on 01 December 1900 and was listed as Trooper no. 1180. He had no former service, indicating that he took this opportunity during turbulent times, perhaps to receive the pay and generous rations provided by Steinaecker’s Horse. It is even possible that he joined to safeguard his business interest after the war. His occupation was recorded as being a fitter, and he was described as being 5’7” tall, with grey eyes and black hair (BNA, WO 100/365; BNA, WO 126/141; BNA, WO 127/23; Woolmore 2006:282). This height characteristic earned him the nickname, ‘Mfishane’, meaning the short man under the local population’s language (Pienaar 2012a:321).

Sardelli went missing at the Battle of Fort Mpisane, but later re-joined. After the war, he was discharged at Komatipoort on 20 August 1902. He qualified for the Queen’s South African medal with a Transvaal clasp as well as the King’s South African medal with two date clasps (Woolmore 2006:103, 282). It is strange that he received these, as according to Pienaar (2012c:422), he absconded during the mentioned battle and sold rifles, which he had stolen from the unit to local people. While the sources are not very clear, those who were part of the war seemingly received arms. He continued with his tricks of selling rifles to local people even after the war, seemingly doing anything to make money. Apparently, he also worked as a cattle herder near Mica, but nothing more of this venture is known. His date of death is unknown, but he died in a mental hospital of which the name is not indicated by any of the sources (Pienaar 2012c:422; Woolmore 2006:282).

Besides Sardelli, the site under consideration in this article is also directly linked with the story of Bill Sanderson, who was interned. Indications of his relevance in the war are that he was the reason for the use of the Sabiepoort site during the conflict. Only two sources provide a narrative of Sanderson’s capture and the resulting events, namely Pienaar (ed. 2012) and Woolmore (2006). Bill Sanderson lived near Legogote on the farm Peebles with his brothers Bob and Tom. He was a farmer of Scottish descent, and the Sanderson brothers were citizens of the Transvaal Republic (Woolmore 2006:64). Sanderson had many friends and acquaintances among the Boer population (Pienaar 2012c:416; Woolmore 2006:64), but the brothers were not prepared to fight against their British countrymen either and thus decided to stay neutral during the war. Pienaar (2012c:416), however, indicated that Sanderson had a short spell on commando with the Boers but was allowed to return to his farm as he was older than 60 and thus not eligible for military service.

Soon after Steinaecker’s Horse had settled in Komatipoort, they heard that an Englishman (Sanderson) was stranded among the Boers and was unable to get away. It was also reported that he possessed cattle and several good horses, which the Boers were likely to make use of. A message was sent to him that a force would be sent to rescue him, to which he replied that it was unnecessary as he was not unsafe (Pienaar 2012c:416; Woolmore 2006:282). This led Steinaecker’s Horse to believe that he was a traitor and in contact with the enemy. After establishing that he had been on commando with the Boers, it was decided to remove him by force. Therefore, sometime during 1901, Steinaecker’s Horse captured him, and he left with as much of his household goods as he could carry onto two wagons, together with his horses, cattle and other livestock (Pienaar 2012c:416).

As they were leaving his farm, shots were fired from the hill above. It was a Boer commando who pounced upon the remaining cattle that they regarded as spoils of war. The Boers concluded that Sanderson, whom they looked upon as one of them, had gone over to the British. They therefore cleared up everything that had been left behind, occupied his house and used the corn mill (Pienaar 2012c:416; Woolmore 2006:65).

Sanderson was having a difficult time being held captive in a building, which was very warm during the day. He was interrogated regarding his activities in the war and kept under observation (Pienaar 2012c:416). Woolmore (2006:65) specifically indicates that he was held in a wooden and iron shed on the Lebombo Mountain near Sabiepoort, where the local camp was commanded by Lieutenant (later Captain) Gaza Gray. This shed was excavated as part of the Steinaecker’s Horse project, and the findings thereof are discussed below. Sanderson’s best horses and rifles were commandeered, under promise of compensation, which he never received. After the members of Steinaecker’s Horse learned that the Boers were using his mill, they blew it up (Pienaar 2012c:416).

It is said that after his internment by the British, he was forced to become a member of Steinaecker’s Horse (Pienaar 2012c:416), but his name is not to be found on the known lists of the unit. These nominal roles from the British National Archives provide official information on all soldiers from a specific unit and as the name of Sanderson does not appear on that of Steinaecker’s Horse, it means he never became an official member. When eventually released, he returned to his farm where he had to start afresh because of the damage done and losing most of his livestock (Woolmore 2006:65).

The Steinaecker’s Horse unit, for which Sardelli fought, was formed by Francis Christiaan Ludwig von Steinaecker, a former Prussian-German soldier (Forsyth 1972:20–23). He came to South Africa in 1886, working as a cartographer in German South-West Africa (Namibia). He later settled in Natal (now KwaZulu-Natal) in 1890. Von Steinaecker became a British subject and joined the Colonial Scouts when the war broke out in 1899 (Van Vollenhoven 2010:22).

Von Steinaecker came to the attention of General R Buller, commander of the British Forces, during the early stages of the war. After participating in a series of successful campaigns against the Boers, he was given permission to raise his own cavalry unit called Steinaecker’s Horse. The unit eventually comprised about 450 white and 300 black soldiers, consisting of Swazi, Pedi and Shangane. Although most of the members of the unit were local inhabitants of the Lowveld, some officers and men were seconded to the unit from the British Army and the Cape Mounted Rifles (Diespecker 1996:98).

The unit did not encounter much military action, but they were involved in a few skirmishes with the Boers such as the Battle of Fort Mpisane on 07 August 1901 (Pelser 1999:50–59). The unit’s main function, however, was to act as border guard to prevent the Boers from contacting their supporters in Portuguese East Africa (Mozambique). For this purpose, several outposts were established along the 200 km stretch of the Lebombo Mountains, specifically in the area today known as the Kruger National Park (Van Vollenhoven 2010:28–29). One of these was called the Gaza Gray Outpost, which is closely associated with the site of Sardelli’s shop at Sabiepoort, believed to be an annex to the first mentioned outpost.

Sites associated with Steinaecker’s Horse

According to Stirling (1907:273), Steinaecker’s Horse had more than a dozen permanent outposts, but this is an under-estimation as research revealed more sites. Very little is known about the location of these outposts and one of the aims of the Steinaecker’s Horse Project is to find them. This is even though there has been extensive archival research conducted in South Africa and the United Kingdom but not many records were found. The lack of archival material emphasises the significance of archaeological investigations.

Harry Wolhuter, a game ranger in the later Sabie Reserve (which became part of what is today known as the Kruger National Park), was a member of the unit. In his book, Memories of a game ranger, he mentioned that they had pickets along the Lebombo Mountains (Wolhuter 2010:174). It was his task to patrol the country up to the Olifants River and decide on suitable places where pickets could be established (Wolhuter 2010:55–56). Twelve of these are shown on a map (Woolmore 2006:79), but the scale of the map is too large to give any assistance in physically locating them. Of the 12 pickets, 7 are located to the south of the Kruger National Park in eSwatini (two of these seven are known to be Nomahasha and Signal Hill on the Kalishan Mountain, while another is said to have been 96 km south of Komatipoort) (Diespecker 1996:99). The remaining sites include four located within the boundaries of the park and one is the headquarters of the unit at Komatipoort (Cattrick 1959:180; Diespecker 1994:4). The four sites in the park probably include (from south to north) the site of another shop of Sardelli at Gomondwane, the Gaza Gray outpost (including the shop of Sardelli at Sabiepoort), Muntshe and the site at Sabi Bridge (Van Vollenhoven 2010:59). Sardelli’s Shop at Crocodile Bridge is not indicated on the map, but it was apparently also commandeered for the use of Steinaecker’s Horse (Pienaar 2012d:475; Woolmore 2006:198).

From other sources (Pienaar ed. 1990; Wolhuter 2010; Woolmore 2006), more sites in the Kruger National Park were identified. These are KaNwamuriwane at Metsi-Metsi, Bottelkop (observation post on the south-western slope of a hill at Komatipoort), Outspan, Ngotso Mouth and the Northernmost (Letaba or Makhadzi) outpost (Pienaar 2002: personal communication, 10 August, 12 September; Piernaar 2003: personal communication, 17 January; Pienaar 2012c:414). Another known site is that of Fort Mpisane at Bushbuckridge just west of the park (Pienaar ed. 1990:345). Woolmore (2006:74) and Wolhuter (2010:56, 58, 78) also mention a post named Kilo 104 and one at Nwanedzi. This brings the total of Steinaecker’s Horse sites to 21, of which 11 are located within the boundaries of the Kruger National Park.

The site close to the Makhadzi Spruit was the first one to be physically located and excavated between 1997 and 2002 (Van Vollenhoven 2010:75, 131–180). The second site that was researched is that of the unit’s headquarters at Komatipoort. It became the headquarters of Steinaecker’s Horse after the British occupied the Lowveld in September 1900 (Pienaar ed. 1990:343). Before that, the headquarters were at Nomahasha in eSwatini (Diespecker 1996:99).

Muntshe is described as an outpost of Steinaecker’s Horse during 1900–1902. Apparently, the site was used by various hunters and explorers as a camping spot even before the war (De Vaal 2012a:254; Pienaar 2012f:374). KaNwamuriwane (also called Metsi-Metsi) was used in 1901 to guard the Nwaswitsontso Poort and consisted of trenches dug against the eastern slope of the KaNwamuriwane Mountain close to Tshokwane (Pienaar 2002, 2003: personal communication, 10 August, 12 September, 17 January; Pienaar 2012e:708). Pienaar (2012a:319) also makes mention of another site on the western bank of the Mbiyamithi Spruit, west of the Môrester windmill. He indicates that many relics from the South African War, such as bullets and military buttons, were found here.

Outspan is a known campsite used by travellers along the East Coast trade route since the 1840s. It was used by the first white pioneers in the area, for instance, the hunter G.R von Wiellich and salesman Percy Fitzpatrick, and it was also used by Steinaecker’s Horse during 1901 (Pienaar 1990:345, 618, 62). Pienaar (2012c:422) also makes mention of an outpost close to the Olifants River, Ngotso Mouth camp (Wolhuter 1948:45), believed to be a temporary camp used by members of the unit on their way to take supplies to the Northern outpost.

The Salitje outpost at Gabeni is also indicated on a map (placed loose inside of Pienaar [ed. 2007] and Pienaar [ed. 1990:617]). There also was a picket close to Nwanedzi, the first outpost established by Wolhuter, and where he also got sick of malaria and black water fever (Wolhuter 2010:58). It seems that the Kilo 104 post is situated outside of boundaries of the Kruger National Park, likely within the Sabi Sand Game Reserve but has not been located yet. The post was only manned by 6 soldiers. They made use of two railway carriages as shelter and storage (Wolhuter 2010:54–55; Woolmore 2006:74). The site used to be the railhead and was situated between Sabi Bridge and Fort Mpisane. Following the washing away of the bridge at the Sabie River in 1901, Sabi Bridge became the railhead (Wolhuter 2010:55; Woolmore 2006:18).

Steinaecker’s Horse had a large post at Sabi Bridge where they had a camp and erected a blockhouse at the eastern side of the southern end of the temporary bridge (Cartwright n.d.:1; Joubert 2012:562). They occupied the site from 1900 until 1902 and operated the railway train to move soldiers and supplies between Komatipoort and Kilo 104, approximately 34 km north of the Sabie River (Pienaar 2012c:422; Woolmore 2006:18).

Harry Wolhuter was placed in charge of a picket at Ngomandwane with 20 men under his command (Wolhuter 2010:72). It is not certain whether this is the same picket as the one at Gomondwane. The latter site is officially known as Sardelli’s shop and was commandeered during the war for this purpose (Pienaar 2002, 2003: personal communication, 10 August, 12 September, 17 January; Pienaar 2012c:415; Stevenson-Hamilton 1952:28). The second shop was also commandeered by Steinaecker’s Horse and is called the shop at Sabiepoort (or Sabiepoort site) to distinguish between these two. There is no historical information suggesting that the site was used for any other military purpose than the jailing of Sanderson. It therefore seems as if the utilisation of the site by the Steinaecker’s Horse was brief, which likely would not have resulted in many cultural objects being left here during the time (see discussion below).

The Gaza Gray outpost is closely associated with the Sabiepoort site (Pienaar 2012e:717; Woolmore 2006:282). Even before the war, Gaza Gray used this site to herd his cattle. During the war, between 1900 and 1902, it was used as an outpost by Steinaecker’s Horse. The site is named after the commanding officer, Captain Edward George (Gaza) Gray (Woolmore 2006:198). He was in command of three outposts of Steinaecker’s Horse: Sabiepoort site, the nearby one at Gomondwane and the one at Crocodile Bridge (Pienaar 2012d:475; Woolmore 2006:198).

Methods and initial analyses

Five excavations were conducted on-site. Three of these were on the remains of the shop and included the shop as well as the immediate surrounding area. Two excavations were done at different refuse middens, both towards the west of the shop. The floor of the shop was still cemented at the time. However, it was partially covered with soil, which had to be removed as part of the excavation. The area around the shop was included to see whether there were any other features.

The Middens

Two trenches were opened at the first midden: the first trench covered an area of 7 m × 5 m while the second was about 7 m × 6 m. The aim was to determine the depth thereof and to obtain cultural materials. The first midden was very shallow, with sterile soil being exposed at a depth of between 2 and 4 cm (depending on the natural slope towards the river, falling from south to north). The second one was deeper, between 6 and 11 cm, but certain sections in the excavation only contained surface artefacts. It would seem, therefore, that these were deflated middens probably as a result of soil erosion over time.

A reasonably low number of cultural materials was recovered from the first midden, including one ceramic shard, glass beads and glass fragments. The conclusion is that the area represents a deflated midden with a diameter of between 10 and 15 m. Most of the cultural materials collected were unearthed at the second midden. They included extremely fragmented faunal materials, ceramics, glass beads, metal and charcoal. It is concluded that this is a semi-deflated midden. In both cases, artefacts were seen downslope towards the river (north), confirming deflation by means of erosion.

The shop

The aim of the excavation was to unearth and investigate the remains of the shop and to collect cultural materials, which could be used to shed light on possible activity areas or the functions of rooms. The three excavations joined one another and together formed a rectangle of 12 m × 8.25 m. In the end, only one layer was excavated, which ended on a concrete floor. The latter is the original floor of the structure. The depth of excavated soil varied between 5 and 8 cm, with the concrete floor being an additional 5 cm thick. An exception was the south-eastern corner, where a hole, inserted in the concrete during construction of the shop, was investigated. Here, the excavation was 1 m deep. The hole seems to have been part of the original construction of the floor and was thus investigated to determine whether it had a specific function and if so, what this was. The shop excavations did not contain much cultural material that can be expected of a structure with a concrete floor as artefacts could easily be washed away from such a site. A few pieces of wood and a piece of bakelite, as well as glass, metal, faunal materials and plastic were unearthed. The bakelite indicates the re-use of the site after the war (see the discussion section below).

Discussion

In order to understand the site, a description of the remains is necessary. What remains of the shop mostly consist of an almost completely intact concrete floor (meaning very few areas where the concrete was completely destroyed were found) with linear marks indicating different rooms or features (Figure 3 and Figure 4). Considering that the research was principally about the Steinaecker’s Horse, there was no need to excavate below the concrete level as doing so would have meant going beyond the levels occupied by Steinaecker’s Horse without a reason. The centre of the structure consists of a rectangular room with sides measuring 3.15 m (north-south) × 2.95 m (east-west). This central room had a concrete floor and originally had corrugated iron walls indicated by typical impressions left in the cement. The impressions are vertical, meaning that the corrugated iron sheets were placed vertically in length. Corrugated iron was patented in Great Britain in 1829 and revolutionised architecture as it became possible to build large buildings (originally only temporary), which were self-supporting (Nicholson 2013:1). Because of the material being portable, it was used in abundance in the colonies, such as South Africa. Sizes varied between 1 m to 3 m long and 46 cm wide (Nicholson 2013:1–3). This means that the iron sheets had to be cut as it is unlikely that the room would be 3 m high or more. Looking at the size of the room, Sardelli must have used between 9 and 11 sheets per wall, depending on the overlap between sheets, which had to be provided to keep water out and strengthen the walls. Such overlaps usually are approximately between 5 and 15 cm (Nicholson 2013:2). Based on this information, it is thus suggested that he used sheets of 3 m in length.

FIGURE 3: The excavated building remains, photographed from the north.

Apart from the marks left by corrugated iron on the floor, at least one loose piece of cement found (several pieces were lying around) had clear marks left by corrugated iron. The historical information indicated that the building was made of wood and corrugated iron, but it is also known that this was common during late Victorian times (round about 1900), especially within a military context (Nicholson 2013:1–2).

Outside of the western wall, clear indications of reed marks were found along the entire wall, indicating that the building here had a double wall. The outer reed wall was likely constructed to keep the temperature down as it can be very hot in the afternoon.

No post holes were found, suggesting that the room either had no framework or the framework was placed on top of the floor. As no indication of poles were found at this room, it is suggested that the corrugated iron walls were set inside of the wet cement floor. It is likely that the sheets used for the eastern and western walls were bent around the corners of the northern and southern walls to strengthen it. This overlaps likely measured 10 cm more at each corner, which is why these two walls are 20 cm shorter than the other two.

FIGURE 4: Excavation map of shop area and reconstruction thereof.

It is clear that the corrugated iron was removed with caution (probably on demolition) as the floor here does not show much deterioration. This was likely done to re-use the material. Since no indication of typical screws and washers, which are usually used to fit corrugated iron sheets together, were found, it seems that these were removed with the metal sheets when the site was abandoned.

The room probably had a window in the western wall, indicated by window glass shards found here. Unfortunately, the fragments could not be pieced together, but a rough calculation indicates that the window was at least a 20 cm square. Based on the position of most of the glass shards, it seems to have been off-centre towards the south. A possible doorway measuring 95 cm was found off-centre towards the east in the northern wall. Nothing indicating the material from which the door was made remained. The way in which the floor was smoothed out here indicates that the door had a plank or branch at the bottom. The door probably was cut out of the corrugated iron wall and a plank fixed to the bottom for stabilisation. Whereas all walls were cemented from the outside as well as the inside to strengthen it, it was not the case with the door as it had to be possible to open it. It seems to have opened to the outside as the smooth edge was on the outside with the corrugated iron markings visible on the cement at the inside (Figure 5).

FIGURE 5: The central room, photographed from the north.

No indication of roofing was found, but as it seems the building was not built sturdy, one would rather expect that it had a roof consisting of branches or reeds. However, this is contradicted by a 15 cm – 20 cm wide section of floor along the southern side, which has marks indicating that it was a drop zone for water from the roof. This suggests a roof tilted towards the south in order to control the flow of rainwater. It also suggests that this room likely had a solid roof, perhaps also made of corrugated iron. Sardelli thus most likely used 3 m lengths for the roof, which must have then been laid from east to west as it would not cover the entire roof area if placed from north to south. The same number (between 9 and 11) sheets of 3 m long were likely used. Because of the runoff of water along the ridges in the roof, a similar drop zone section on the cement is found on the northern, eastern and western side of the building, but this is clearly less accentuated as the one on the south. Having a corrugated iron roof, perhaps confirms that a wooden framework had been used for the construction of the building, but that it was merely placed on top of the concrete floor.

The concrete floor seems to have been laid at different times. The floor at the central room was likely constructed first, but only as it was the starting point, meaning that other floors were constructed soon afterwards. It seems as if the second section of the floor that was constructed is towards the south as it is clear that this floor was laid over the mentioned water drop zone. This floor, which forms the central part along the southern end of the building, forms a second room.

This southern room measured 3.45 m (north-south) × 2.95 m (east-west). From marks in the cement, it is clear that this room had reed walls on its eastern, southern and western sides (Figure 6). The reeds were fixed together with stronger (likely leadwood) poles approximately 80 cm apart indicated by larger markings in the concrete. No clues were found indicating the wall on its northern side, but here it probably utilised the corrugated iron wall from the central room. Indications of a threshold in the north-western corner suggest that this was the entrance.

FIGURE 6: Note the crack indicating the different floor levels. Also note the marks left by reed walls. Some reeds were cut close to the site and placed in the reed marks as illustration. Also note the difference in texture of the two floors.

The floor of this southern room is smooth, whereas all other floors are rough. It is believed that all floors were originally smooth, but that the reeds from the walls in this instance collapsed on top of the floor, protecting it from deteriorating. Spots on the floor of what looks like pitch were observed. This may have been used as waterproofing, which may indicate that the smooth cement section had a roof, but it is doubted that the walls of the second room were strong enough to hold a roof. It may thus have been used to waterproof the bottom of the walls. No indications of a roof or window were found. It is believed that this room may not have had a roof and only served as open air space where things may have been stored, warranting at least walls.

The remainder of the floor of the structure was the last to be laid. It has a U-shape running along the western and eastern sides of the two rooms and along the northern side of the central room. It was clearly laid on top of the others as the cement overlays the last mentioned in certain areas. This likely was a veranda seemingly without a roof as the structure would not have been able to carry the weight. Various post holes were unearthed, of which seven were found in between the different floors of which one still had the remains of the wooden post inside. The remains of two posts found in situ on the northern side as well as that of two poles suggest that a row of poles was planted approximately 25 cm from the outer edge of the floor. This may have been a fence. The remains of one pole were found intact on the southern side of the concrete floor although it had fallen. No indication of such fence posts was found on other sides, but a large number of leadwood poles were identified on site, suggesting a fence or enclosure, perhaps further away from the building. Some of these fitted in the mentioned holes and a few were still bound together with wire, an indication that it was used as fencing. It is unlikely that such a feature would only have been erected on one side as wild animals would have been a major threat on site. The poles found were lying all over the site close to the building and may therefore represent an enclosure around it.

The wooden poles had two different sizes. Some are round about 10 cm in diameter and others approximately 25 cm, with the latter being outnumbered by far. The thicker ones therefore are likely corner posts with the thinner ones used in between. Reeds, found in abundance along the river, may also have been used, but none of these survived (also see Figure 6). The length of the poles varies, but as it may have broken it is not a good indication of length. The longest of the thinner ones is 1.9 m long and of the thick ones, 2.18 m. A few short poles, approximately 45 cm in length also were clearly cut with a 45° angle. These were likely used to fit poles together and strengthen the fence.

The size of the entire floor varies as the edge was not made straight. At most, it is 9.10 m (north-south) × 7.10 m (east-west). The front side is north facing towards the Sabie River, and it is likely that trade routes were also on this side as it is known that these routes sometimes followed rivers (De Vaal 2007:68). The functioning of the building is explained in the conclusion.

At the north-western corner, just outside of the floor, a layer of cement much thinner than elsewhere was found (1 cm vs 5 cm). This either was the area where the cement was mixed or the extra cement after completion of the floor was just discarded. Excavation on the outside of the concrete floor proved that there was no foundation underneath. However, the soil on the south-western side seemed very hard, indicating that it was possibly compacted to ensure a good base for the laying of the concrete floor.

Before the section of the floor in the south-eastern corner was laid, a hole was dug and laid out with stones. This hole, which clearly is a deliberate feature, is just east and outside of the southern room. When constructing this floor, the cement was edged around the hole with bits falling in, explaining pieces of cement and lime (probably used to mix the cement) found inside of the hole up to a depth of 60 cm.

The hole was filled with stones up to 84 cm deep, which gives the impression that it was some kind of a drainage system for water. The hole is square with sides of 70 cm, and the top of the large stones is 4 cm below the edge of the cement. While the cement was still wet, six poles were planted around the hole. These post holes are still visible and vary between 6 and 8 cm in diameter and they are 8 cm deep.

Conclusion

In summary, it can be indicated that the structure excavated consisted of a concrete floor, with two rooms and a veranda on the northern, eastern and western side, facing north towards the Sabie River (Figure 7). The few artefacts found give clues to the functions of the two rooms. It is believed that the main room was the shop, which may also have served as sleeping quarters. Expensive goods may have been kept here under lock. The second room was likely a storage area. The hole at the south-eastern corner is believed to be a sump, interpreted as a washing area as it is reasonably deep. The six post holes were either for poles used to fix a pump system for pumping water from the river or it could have been for shelves.

FIGURE 7: Drawing indicating a reconstruction of shop facades.

The artefactual evidence points to the presence of Europeans on the site. This is corroborated by the historical information on the site being one of the shops of Sardelli and being utilised by Steinaecker’s Horse. Artefacts with a European origin date from the late 19th and early 20th century and therefore fall in the period of the South African War (1899–1902) when the Steinaecker’s Horse unit was operational. This included metal artefacts and glass from bottles (alcoholic and non-alcoholic beverages) and window glass (Lastovica & Lastovica 1990:27–7; Weiland 2009:29–31). A food tin and bottle cap found are from items sold and used at the site. As only a few remains of glass bottles were located, it suggests that Sardelli did not sell his home-brewed alcohol in such containers, but in ceramic containers, produced by himself or local people.

The limited remains of tins from tinned rations and glass from bottles indicate that the site did not have had a long occupation by the British forces. Apart from these rations, their diet was likely supplemented by eating the meat of domesticated and non-domesticated animals. It is known that the members of Steinaecker’s Horse herded the livestock from local people in order to use them as food (Stevenson-Hamilton 1952:14–15). Remains of mussel, land snail shells as well as ostrich eggshell beads indicate that these formed part of the diet. Maize was also eaten on site, at least by the indigenous people as shown by the typical maize grinding stones on site (Huffman 2007:454).

European manufactured artefacts found at the two middens imply that these are contemporary with the shop. Parts of a pocket watch found, represent an expensive personal item belonging to someone present on site. A few objects found date from a slightly younger period (ca. 1900–1920), likely indicating re-use by game rangers. It is known that other sites associated with Steinaecker’s Horse were re-used (Van Vollenhoven, Pelser & Van Den Bos 1998:119).

As the Sabiepoort Shop is located close to the Gaza Gray site, the shop site was likely under the command of Captain Gaza Gray. As an extension of the Gaza Gray site, it makes sense that similar cultural material has been found. Artefacts with a similar date and European origin were excavated at both sites, and many of these are similar to ceramics excavated at other sites associated with Steinaecker’s Horse, suggesting the presence of this unit here. Finding artefacts with a military origin, such as ammunition, military buttons and insignia, serves as archaeological proof of the presence of a military unit, but no such cultural material was found at Sabiepoort. The only artefact, which may have a military origin, is a buckle from the bridle of a horse, mule or donkey. It may, however, also have originated from before the war. As the shop was only used for a short time during the war when Sanderson was held captive there, this is not surprising. However, military objects were found at the associated Gaza Gray site, including Eley shotgun, Martini-Henry and Lee-Metford/Lee-Enfield rifle cartridges, commonly used by the British Army (Simpson 1982:26). Although these types of ammunition were also used before and after the war, the headstamps of these are associated with the South African War (Langbride 1975:n.p.).

Locally produced ceramics were also excavated at both sites. The ceramics from one excavation suggest that the local people were closely associated with the members of Steinaecker’s Horse. Having a workforce at hand may have contributed to the location, but even before the war, Sardelli likely made use of local people as servants. It is known that Steinaecker’s Horse mainly picked their outposts in areas where there were already people present (Van Vollenhoven 2010:247). These local people served the soldiers for different tasks including being servants, cooks and doing military tasks such as patrolling. This indeed was the case at the Gaza Gray Outpost where they were mainly used for cattle herding (Van Vollenhoven 2016:82).

From the cultural objects excavated at the shop site, a few activities of the people on site can be deduced, including brewing alcohol, hunting and fishing. They probably also assisted with the herding of livestock and have planted maize. Many of the tasks indicated above may not have been performed by the soldiers who would have kept watch and went on routine patrols (Pelser & Van Vollenhoven 1998:45). During the internment of Sanderson, they would also have interrogated him.

The locally produced ceramics found at Sabiepoort are similar to those found on the Gaza Gray site, dating to 1850–1900 AD, indicating a 19th-century Nguni/Tsonga signature for the site (Biemond 2012). A bright yellow glass bead found at the Gaza Gray site was typically manufactured from the 1860s and indicating utilisation of the site during the latter part of the 19th century (Biemond 2010: personal communication, 01 December; Biemond 2016:personal communication, 05 January). This date is corroborated by historical information. Stevenson-Hamilton described the Sabi Reserve (the southern part of the Kruger National Park) as a desolate country in 1860 because of raids by the Swazi in the area. During the 1870s–1890s, various tribal wars were fought around the Sabie River Valley and in Mozambique, and Tsonga groups were also at war with the Portuguese, resulting in Tsonga groups fleeing to this area in the former Transvaal. Tsonga people still inhabited the southern parts of the park in the years just after the South African War (Eloff & De Vaal 2007:58–64). This information therefore indicates that both sites were only occupied by the late 19th century.

The Tsonga were well-known middlemen on the early trade routes (Changuion 1999:104; Ferreira 2002:34). Maps in Bergh (ed. 1999:9) and Pienaar (ed. 2007) do indicate that one of the early trade routes and an old wagon route passed not far to the north of the Sabiepoort site and another one of the known trade routes did go through the Sabiepoort (De Vaal 2012b:67). Trade items such as glass beads have been found at both sites. Sardelli probably erected his shop here because of it being close to the trade route.

Except for the indigenous people and members of Steinaecker’s Horse on the site, members of military units who associated with them, such as the Native Police, could also have been present. Wolhuter (2010:62, 85) indicated that the Native Police were stationed at Steinaecker’s Horse pickets.

It is concluded that the site was occupied before the South African War, during the latter part of the 19th century by Tsonga-speaking people. During the 1890s, Dimitri Sardelli and his associate, Charlie Woodlands, erected a shop here. Sardelli joined the Steinaecker’s Horse unit during the war. Apart from using the shop for the internment of Sanderson, there is no other historical evidence that Steinaecker’s Horse occupied the Sabiepoort site. There is also no artefactual evidence indicating that this was done for any other purpose. Their association with the site therefore seems very brief. Nevertheless, the research at the Sabiepoort site contributes to obtaining a better understanding of Steinaecker’s Horses’ activities in the area. In this case, it indicates their use of this site and its association with the nearby Gaza Gray outpost, as well as activities indicated above.

Acknowledgements

Staff of the South African National Parks, especially those in the Kruger National Park for their assistance in the planning and organisation of the project and for keeping us safe on the site, are thanked. Special thanks are because of Louise Swemmer of Scientific Services who served as our liaison with the Park as well as Marius Snyders, the regional ranger at Lower Sabie for his assistance. The volunteer members of the research team are all thanked for their support and hard work.

The author has declared that no competing interest exists.

The author declares that they are the sole author of this research article.

Ethical approval for the study was received on 20 November 2023 from North-West University Basic and Social Sciences Research Ethics Committee (BaSSREC) with the ethics number NWU01056-23-A7.

The author received no financial support for the research, authorship and/or publication of this article.

The author confirms that the data supporting the findings of this study are available within the article and its supplementary materials.

The views expressed in the submitted article are the author’s own and not an official position of the North-West University.

References