Determining the origin of Iron Age copper ingots from Zambia and Zimbabwe.

Southern and South Central Africa are home to numerous rich copper deposits, including the sediment-hosted deposits in the Central African Copperbelt, which runs along the border between the Democratic Republic of Congo and Zambia, the similar deposits of the Magondi Belt of northern Zimbabwe, the early Precambrian metavolcanic 'greenstone belts' of southern Zimbabwe, eastern Botswana and northeastern South Africa, the Limpopo Mobile Belt (mostly gneisses) parallel to the border between Zimbabwe and South Africa, and the Phalaborwa Igneous Complex (a carbonatite) in northeastern South Africa. All of these deposits were well known to local populations, and extensively worked for their ores, long before the arrival of Europeans in the area. These pre-colonial mines were the subject of a number of studies between 1920 and 1975, but subsequently have been largely overlooked, and many are now in danger of being lost due to large scale modern mining in the same areas.

The earliest known evidence of copper mining in the region is at Kansanshi in northwestern Zambia, which dates from the fourth century AD, about the same time as the first Iron Age farmers arrived in the region, and the ore appears to have been extracted using iron tools. Ore appears to have been processed at Kansanshi, and transported overland to Kipushi on the Kafue River for smelting; a distance of over 100 km. A variety of smelting techniques have been recorded from ancient sites in the Copperbelt, including tapping the metal directly into molds, allowing it to gather in the bottom of furnaces, and collecting it from prills (globules) trapped within the slag of melted ore. It was then typically worked with iron tools, including hammers and wire-drawing plates. The most common usage of copper in Southern Africa appears to have been to make jewelry, but stores of copper ingots have been found in a variety of locations, and it has been suggested that these may have been used as currency. These copper ingots have been the subject of a number of studies, which have focused on their age, distribution (there are several different forms, found in different areas), and possible uses, although the precise origin of the different forms of ingot have yet to be determined.

In a paper published in the journal PLoS One on 22 March 2023, Jay Stephens of the School of Anthropology at the University of Arizona, and the ArchaeometryLaboratory at the University of Missouri, David Killick, also of the School of Anthropology at the University of Arizona, and Shadreck Chirikure of the School of Archaeology at the University of Oxford, and the Department of Archaeology at the University of Cape Town, present the results of a study of a selection of Southern African copper ingots, using isotopic and trace element data to try to connect them to specific mining areas.

Locations of prehistorically exploited mines in southern Africa. Stephanie Martin and Jay Stephens in Stephens et al. (2023).

The exact age of the ores of the Copperbelt is uncertain. The copper is thought to have accumulated in a sedimentary basin between the Zimbabwean and Congo cratons (proto-continents), which came together to form part of the super continent of Gondwana between 650 and 500 million years ago. As this basin closed, the sediments were compressed and folded, before undergoing a series of faulting events, forming the Damaran-Lufilian Arc, which runs along the coast from western South Africa, through Namibia, and up into Zambia and the southern Democratic Republic of Congo. The formation of the copper minerals may have happened during, or prior to, the Lufilian (or Pan-African) Orogeny, though they were certainly remobilised and redistributed during this orogeny. The ore bodies appear to have received two influxes of uranium, one about 650 million years ago, and one about 530 million years ago, This has resulted in the presence of the isotopes lead²⁰⁶ and lead²⁰⁷, the decay products of uranium²³⁸ and uranium²³⁵, within the Copperbelt ore bodies.

The Kipushi Mine lies within the Copperbelt, but access deposits geologically distinct from other Copperbelt ore bodies, being younger, at about 450 million years old, higher within the stratigraphic column, and geochemically distinct. Kipushi is the largest of three zinc-lead-copper deposits within the Copperbelt area, with the others being Kabwe, also in Zambia, and Tsumeb, in Namibia. The isotopic signal of these deposits is quite different from that of the other Copperbelt ores, lacking in the radiogenic lead component. It is also possible to distinguish Kipushi ores from Kabwe and Tsumeb ores chemically, with the Kipushi ores highly enriched in copper, zinc, arsenic, silver, antimony, and lead, both within the sulphide ores themselves, and within the supergene zone (area at the top of or above the ore that becomes enriched in metal elements as water peculates through the ore body, dissolving metal ions then redepositing them), which is typically of a malachite colour and enriched in copper arsenate, carbonate, oxide, phosphate, sulfate, vanadate, and chloride minerals.

The Magondi Belt originally formed as a back-arc basin between 2.2 and 2.0 billion years ago, and was deformed and metamorphosed by the Magondi Orogeny between 2.0 and 1.9 billion years ago. These deposits were further deformed and re-deposited during the Pan-African Orogeny, about 550 million years ago. The ores of the Magondi Belt contain radiogenic lead isotopes, with a lead²⁰⁷/lead²⁰⁴ to lead²⁰⁶/lead²⁰⁴ ratio distinct from that of the Copperbelt.

Within the Magondi Belt, ores are generally stratiform, concentrated within the Deweras and Lomagundi group rocks on the eastern margin of the belt. Most of these deposits host only copper, or copper and silver with smaller amounts of gold, lead, platinum, and uranium, although the Copper Queen and Copper King deposits on the western margin of the belt contain zinc-lead-copper-iron-silver ores. 

Copper ingots from Southern Africa have been known since the 1960s, when the first such ingots were discovered in a burial at Sanga in the Upemba Depression of the southern Democratic Republic of Congo, and at Ingombe Ilede in the Zambezi Valley in Zambia. The oldest known ingots are of a type known as 'Ia'. These are small, rectangular ingots, dating from the 5th to 7th centuries AD, known from sites in the Copperbelt and at Kumadzulo. Two 'fishtail' ingots from Kamusongolwa Kopje, in Northwestern Province, Zambia, and Luano Main Site, in Central Province, also Zambia, dated to between the 9th and 12th centuries AD, are also considered to belong to the 'Ia' type. 

A second type of ingot, named HIH, appears at sites dating from the ninth century onwards, and is considered to be linked to a significant increase in the production and use of copper. These ingots tend to be 7-20 cm in length, and are 'H' shaped, with two short arms at each end of a longer central bar. These ingots are known from 9th to 14th century sites from the Upemba Depression in the Southern Democratic Republic of Congo to Great Zimbabwe, with a number of molds being known as well as the ingots themselves. HIH ingots appear to have been mainly trade items until the 13th century, but in the 14th century they became prestige goods in their own right, and are often found within Kambabian-period burials. All the known HIH ingots from Great Zimbabwe date from the 14th century.

In the 14th century two, apparently separate, spheres of ingot circulation developed. In the western Copperbelt and Upemba Depression, small HX and HH type croisette (cross-shaped) ingots, between 0.5 and 7 cm across, are found, with molds for these ingots found across this range. At the same time, in the eastern Copperbelt, at Ingombe Ilede in the Zambezi Valley, in northern Zimbabwe, and in the Dedza area of Malawi, much larger HXR type ingots appear. These are X-shaped, with a flange around their outer edge, 20-30 cm in length and weigh 3.0-5.5 kg. Molds for these ingots are only known from the eastern Copperbelt, although this includes Kipushi. 

'Ia' (rectangular and fishtail) and croisette ingot typology. Stephens et al. (2023).

These HXR ingots have been a subject of considerable archaeological interest, being known from the Ingombe Ilede burial site (eight ingots), eight from a hoard in the Dedza area of Malawi, and two from the Chedzurgwe site in Zimbabwe, and thirty other sites within northern Zimbabwe having produced a further sixty examples, often accompanied by pottery of a similar style to that of Ingombe Ilede. Oral histories from the area record that copper crosses were obtained by trade from a people called the Va-Mbara from the Urungwe area, who were renowned for their metal-working skills, and the sixteenth century Portuguese explorer Antonio Fernandes recording that such metal crosses were produced by a people called the Mobara from the land of Ambar, which maybe a reference to the same group.

Highlighted archaeological sites, geological districts, and geological mines. Stephens et al. (2023).

The copper crosses mentioned by Fernandes and the HXR ingots have been thought to be the same thing for some time, potentially suggesting that Fernandes' records might provide an insight into the origin of these ingots. Fernandes himself believed that the ingots originated from close to the 'Copper rivers of Manyconguo'. which may be a reference to the Niari Basin ore deposits, which lie near the mouth of the Congo River, and which were exploited by the Kongo Kingdom at this time. This is about 2100 km from the area where these ingots are found today, making it highly unlikely that Fernandes was correct on this issue, but several researchers have suggested that this might indicate that the ingots originate from the Copperbelt, which is in the same direction as Niari Basin, albeit considerably closer. 

Few of these large ingots have been directly dated, although, based upon their archaeological context, it is assumed that they were manufactured between the 14th and 18th centuries. The majority HXR ingots have been found at sites which have also produced Ingombe Ilede style ceramics, although examples are also known from Musengezi and Mutapa culture sites. As with earlier ingots, these items seem to have been used both for trade and as prestige items, something recorded by the Portuguese and which matches with the contexts in which they were discovered. 

At some time in the 18th century, production of these HXR ingots ceased, with their replacement by X-shaped un-flanged handa ingots, and I-shaped Ib and Ic ingots, which could weigh up to 30 kg. These were frequently recorded by European visitors to what are now Malawi, Zambia, and the Democratic Republic of Congo, although they have never been recorded from the Zimbabwean Plateau. 

Additionally, copper ingots of the lerale (rod-shaped) and musuku (cuboid) types were produced in South Africa, while nail head ingots, and more informal bun and bar ingots were produced in both South Africa and Zimbabwe, during the 12th to 20th centuries, although these are outside the scope of Stephens et al.'s current study.

Copper ore invariably contains a small amount of lead, with an isotopic composition dependent on the age, type, and other geological characteristics of the rock. This lead is incorporated into the smelted copper with its isotopic signature unaltered, something which has previously been used to trace the origin of copper artifacts from archaeological sites in Europe and the Mediterranean region since the 1980s. More recently, the technique has been applied to copper, bronze, and tin archaeological material from Namibia, Botswana, and South Africa with some effect.

Stephens used an Inductively Coupled Plasma Mass Spectrometry system at the University of Arizona to analyze 34 copper samples taken from ingots, 33 of which came from the collections of the Museum of Human Sciences in Harare, Zimbabwe and the Livingstone Museum in Livingstone, Zambia, for their lead isotope composition, and proportions of chromium, iron, cobalt, nickel, copper, zinc, arsenic selenium, molybdenum, silver, cadmium, tin, antimony, and lead.

The type Ia rectangular and fishtail ingots produced similar lead isotope ranges, all of which were consistent with an origin somewhere within the Copperbelt, as well as a radio-isotope age of 589 ± 15.4 million years, also consistent with an origin within the Copperbelt. It was not possible to assign any of the ingots to a specific mine within this region.

The HIH and HXR croisette ingots from Ingombe Ilede and northern Zimbabwe can be split into two groups, both of which contain both types of ingot. The first group, comprising eight HIH ingots and four HXR ingots, match the type Ia ingots in their lead isotope ratios, and a radio-isotope age of 627.25 ± 3.57 million years; again consistent with a Copperbelt origin, but not sufficient to pin that origin down to a specific mine.

The second group, comprising three HIH and 13 HXR ingots, as well as three large HH ingots from the Upemba Depression in the Democratic Republic of Congo, have isotopic signatures consistent with having originated from the Kipushi Mine deposit within the Copperbelt.

Another HXR ingot, recovered from the Kent Estate in Zimbabwe, has a quite different lead isotope ratio. This appears to match with ores from the Magondi Belt, although Stephens et al. are unable to confirm this with any confidence as relatively few ore samples from Magondi have been analysed.

Chemically, the Ia type bar and fishtail ingots are remarkably pure, containing less than 5 μg per gram of chromium, selenium, molybdenum, silver, cadmium, tin, antimony, and lead, and less than 50 μg per gram of nickel, zinc, and arsenic. The levels of cobalt and iron are slightly higher (if quite variable), with 5-204 μg per gram of cobalt and 27-71 μg per gram of iron. This high level of purity, and pattern of elements has previously been observed in copper artifacts from the Democratic Republic of Congo and Botswana. Once again, this elemental composition is consistent with the metal having originated within the Copperbelt, but is not sufficient to pin it down to a specific mine, with about 150 known mines potentially capable of having made these ingots.

The HIH and HXR croisette ingots were split into two groups on the basis of their isotopic composition, and their chemical content was found to match this. The  first group, comprising eight HIH ingots and four HXR ingots, which were had similar isotope ratios to the type Ia ingots, also had similar chemical compositions, with chromium, zinc, selenium, molybdenum, cadmium, tin, antimony, and lead all under 5 μg per gram, silver under 20 μg per gram, nickel concentrations of 5-42 μg per gram, iron concentrations under 50 μg per gram, and cobalt values of 2-144 μg per gram. This is again sufficient to suggest that these ingots originated from the Copperbelt, but not where, although the variation in nickel, cobalt, and silver seen across these and the type Ia ingots suggests that more than one mine was involved in production.

The second group, comprising three HIH and 13 HXR ingots, have much higher levels of zinc, at 13–146 μg per gram, with a mean of 61 μg per gram, arsenic, at 240–2515 μg per gram, with a mean of 869 μg per gram, silver, at 88–1966 μg per gram, with a mean of 1254 μg per gram, antimony, at 2-111 μg per gram, with a mean of 29 μg per gram, and lead, at 14-1465 μg per gram, with a mean of 378 μg per gram. Concentrations of cobalt, nickel, selenium, molybdenum, cadmium, and tin mostly remain below 5 μg per gram. The lead isotope ratio of these ingots matches the ore from the Kipushi Mine. Three large HH ingots which were analysed in an earlier study led by Frederik Rademakers were found to have similar chemical compositions and lead isotope ratios to this group. The enrichment in chalcophile elements (zinc, arsenic, silver, antimony, and lead) may have happened within the malachite ore, where ions of all of these elements could substitute for copper, or come from the supergene deposits (deposits which lie above the ore body) which are a similar colour to the malachite ore, and could be introduced accidentally. 

A single HXR ingot from Kent estates in Zimbabwe does not fit into either of these groups. This ingot has chromium, cobalt, zinc, arsenic, molybdenum, cadmium, tin, and antimony, levels below 10 μg per gram, lead and nickel below 30 μg per gram, an iron level of 490 μg per gram, selenium at 145 μg per gram, and silver at 1572 μg per gram. This enrichment in selenium and silver, while most other elements are depleted, is typical of ores from the Deweras Group of the Magondi Belt, such as those accessed by modern mines at Mhangura (formerly Mangula) and Norah. These ores also contain uranium, which could account for the high proportion of radiogenic lead in this ingot. 

Stephens et al.'s study suggests that croisette ingots were manufactures in at least three separate locations: the Central African Copperbelt, the Kipushi deposit, and the Magondi Belt. All the ingots appear to be made from ore from a single source, without any evidence of ingots being made from mixed ores or recycled copper from multiple sources, something which has also been observed in previous studies.

Inferred provenance conclusion for the rectangular, fishtail, and croisette ingot in the study. Provenance results indicate that objects travelled significant distances to reach certain destinations and that interactions between the Copperbelt and areas further south can be traced back to the 6th-7th century. Stephens et al. (2023).

Sixteen of the ingots have elemental and isotopic signatures which match the ores of the Central African Copperbelt, something which has previously been observed in copper objects from the Upemba Depression in the southern Democratic Republic of Congo and the Tsodilo Hills in northwest Botswana. These ingots are depleted in chalcophile elements and enriched in siderophile elements such as cobalt and nickel, as are the ores of the Copperbelt, although neither the elemental nor isotopic composition of the ingots allows any more precise diagnosis.

Molds for rectangular or fishtail ingots have never been found within the Copperbelt, but numerous croisette ingot molds have been found, along with considerable evidence for precolonial mining. More than 100 precolonial mines had been recorded in the Copperbelt by 1906, the majority in an arc from Kolwezi to Kipushi in the Katangan Copperbelt of the Democratic Republic of Congo, but with examples in the Kafue Hook and Domes Region of the Zambian Copperbelt. 

The ingots which Stephens et al. attribute to this source include e all three recorded rectangular and fishtail ingots from Zambia, dated to before the 12th century, an HIH ingot from the Harare tradition site of Graniteside, an HXR ingot from burial 8 at Ingombe Ilede, dated to the 15th-17th century, and , 10 HIH and HXR ingots from farms and towns in northern Zimbabwe. These ingots appear to have frequently been carried long distances from their point of origin; about 600 km to the 6th-7th settlement at Kumadzulo near Victoria Falls in southern Zambia, and as far south as Harare in Zimbabwe. This distribution also demonstrates that cultures such as Ingombe Ilede in southern Zambia, Harare in Zimbabwe and Musengezi in northern Zambia were in contact with one another and part of a common trade network.

Another 16 ingots could be linked directly to the Kipushi deposit on the southern border of the Democratic Republic of Congo. The isotope signature of these ingots form a particularly tight cluster even within the range observed at Kipushi, which Stephens et al. assume relates to the oxidized surface deposits, which are known to have been accessed by a precolonial mine. A fragment of casting spill from a furnace by the Kafue River in Zambia close to Kipushi, and a chunk of malachite ore found nearby, also match this signature. These ingots are also enriched in chalcophile elements, but depleted in siderophile elements, something observed at Kipushi in both the malachite ores, and the overlying carbonate deposits. This has also been seen in three HH ingots previously analysed, and presumed to have come from the same source.

There is a great deal of archaeological evidence for precolonial mining at Kipushi, albeit all on the Zambian side of the border. This includes at least 57 discrete smelting sites with slag heaps, two large habitation sites, one campsite, and 71 individual croisette molds (for both HIH and HXR ingots). This activity appears to have been going on since at least the ninth century, with a significant expansion in the fourteenth, which coincides with the first appearance of HIH ingots in cemeteries of the Upemba Depression and in the archaeological record of Zimbabwe.

The ingots considered to have come from Kipushi include undated HIH ingots from northern Zimbabwe, two HXR ingots from Ingombe Ilede (one from burial 2 and one from burial 8), dated to the 15th-17th centuries, two HXR ingots from Chedzurgwe in Zimbabwe dated to the 16th century, and nine HXR ingots discovered at undated surface sites in northern Zimbabwe. These ingots would have been transported 500 km to Ingombe Ilede and up to 780 km for finds near Harare, Zimbabwe.

Stephens et al. also note that there appears to have been a distinct shift in where the HXR ingots were being made, with 15 of 18 HXR ingots analysed coming from Kipushi, and HXR mold fragments associated with the increase in activity at that location, suggesting that Kipushi became the major producer of these ingots at this time, although quite who was responsible for this activity is unclear.

A single ingot, from Kent Estates in Zimbabwe, had an isotopic signature consistent with an origin in the Magondi Belt. There is considerable evidence for precolonial mining within this area, in the areas where the modern Alaska, Angwa, Mhangura, Norah, and Silverside mines are sited, although this has not been studied much by archaeologists. A mold for an HXR ingot was found at the Golden Mile Mine, which is about 20 km away from the Mhangura Mine in Zimbabwe. This appears to have been copied from the form being used in the Copperbelt, but to have been used to make ingots from local copper.

Ninety four HIH and HXR copper ingots have been recovered from archaeological sites in the Zambezi Valley and across the Zambezi Plateau. It was originally assumed that these were made locally, using copper obtained from the copper deposits of the Magondi Belt of Northern Zimbabwe. However, the subsequent discovery of molds for these types of ingots at Kipushi and other sites within the Copperbelt of northern Zambia and the southern Democratic Republic of Congo presented an alternative source, with the ingots potentially being transported to southern Zambia and Zimbabwe.

Stephens et al. analysed 29 of these 94 ingots geochemically, and conclude from this that 28 of them originated within the Copperbelt, and 16 at Kipushi. This suggests that a trade network carrying copper ingots (and presumably other goods) from the southwest of Zambia was in place by the sixth or seventh century, and that this had reached as far as Botswana by the eighth century. By the fourteenth century copper from the Copperbelt was reaching east to the middle Zambezi Valley, and the Urungwe District of northwest Zimbabwe, where pottery in the style of Ingombe Ilede in the middle Zambezi Valley has also been found. HIH and HXR ingots have not been reported from Botswana, but copper goods with similar chemical signals to the Copperbelt ingots has been found there. In addition, some HXR ingots do appear to have been manufactured in Zimbabwe.

See also...

Follow Sciency Thoughts on Facebook.

Follow Sciency Thoughts on Twitter. 

Mining giant Rio Tinto forced to apologise after loosing radioactive caesium¹³⁷ capsule in the Western Australian outback.

The mining giant Rio Tinto has been forced to issue an apology to the people of Western Australia, after loosing a capsule containing a small amount of the radioactive isotope caesium¹³⁷ while it was being shipped across the state earlier this month. The capsule served as a radiation source for a device which was able to measure the density of iron within ore samples, which was being transported between the Gudai-Darti Mine, to the north of Newman and the state capital, Perth, a distance of 1400 km, by a specialist contractor brought in by Rio Tinto to undertake the job. The device was inspected at Gudai-Darti and seen to be safe on 12 January 2023, but when inspected in Perth on 16 January is was found that a gauge, one of its four mounting bolts, several screws, and the caesium¹³⁷ capsule had gone missing, apparently shaken loose by the vibration of the truck.

A member of an Incident Management Team in Western Australia which is coordinating the search for the missing radioactive capsule. Western Australia Department of Fire and Emergency Services/Reuters.

The missing capsule is described as silver in colour, 6 mm in diameter and 8 mm long, small enough to raise concerns that it could be caught in the tread of a car tire. The radiation it emits has been described as equivalent to 10 medical X-rays per hour, not threatening to anyone simply driving past, but potentially harmful to anyone remaining within 5 m of the capsule for any length of time, who could potentially suffer burns or even radiation sickness, with the potential for harm to anyone directly handling the capsule being quite high. Authorities in Western Australia are searching for the missing capsule using portable radiation survey meters, which can be mounted on vehicles, and which should be able to detect the capsule from a distance of about 20 m. This method should be able to detect the capsule if it has remained on or close to the road along which it was transported, although if it has moved from this route, for example by being caught on another vehicle, then the task will become much more difficult, and the danger of members of the public having been exposed to harmful levels of radiation much higher.

See also...

Follow Sciency Thoughts on Facebook.

Follow Sciency Thoughts on Twitter.

Using strontium isotope ratios to determine the origin of textiles from archaeological sites in Nubia.

The manufacture of textiles by pre-industrial societies was typically a demanding, multi-stage process, which involved obtaining fibres from an Animal or Plant, spinning it into a thread, and weaving it into a cloth, with dyes potentially being added at different stages to achieve colours or patterns. As such, textiles from archaeological sites can be important sources of information about the societies that produced them, giving evidence about trade routes and manufacturing techniques, as well as the social status, gender and age of the wearer. 

Textiles from archaeological sites in Sudan have been studied since the early twentieth century, with material recovered from many sites in the mid Nile Valley. Wool and cotton are the most common fibres in these archaeological assemblages. Linen is also present, but in much lower quantities. Silk has never been produced in Sudan, so any silk discovered there can automatically be assumed to have been imported. Spinning of threads in Sudan was almost exclusively done in the S-direction (anticlockwise), while woven cloths were usually plain tabbies (a plain weave, with warp and weft threads crossing at right angles to form a criss-cross pattern) or weft-facing tabbies (cloths in which the warp is covered up by multiple, complimentary weft layers, enabling the building up of a pattern, which is typically angular in nature), often having bands or stripes in different shades.

However, determining the origin of fabrics from archaeological sites is complex, as plain and weft-facing tabbies are widely manufactured, and the textiles recovered from archaeological sites are typically very fragmentary. Even spinning direction is of limited use, as even in areas where the majority of spinning is done in one direction, the presence of local communities or even individual spinners doing the opposite cannot be excluded.

In a paper published in the Journal of African Archaeology on 11 August 2022, Magdalena Wozniak of the Department of African Studies at the University of Warsaw, and Zdzislaw Belka of the Isotope Research Unit at Adam Mickiewicz University, present the results of a study which used strontium isotope ratios to determine the origin of cotton and wool fragments from Late Antiquity and the Early Medieval Period recovered from archaeological sites in the Middle Nile Valley.

Strontium isotope ratios in water are strongly linked to local geology, and can be preserved in the tissues of Animals and Plants, making them a powerful tool in archaeology. Strontium isotopes have been used to track the movements of ancient Human and Animal groups, as well as tracing ancient trade routes by determining the origin of stone tools, glass, and textiles. This technique requires an understanding of the ratio of the isotopes strontium⁸⁷ and strontium⁸⁶ in the geology of the area, both locally and regionally. Armed with this it is possible to determine whether or not a material (in this case fabric) was manufactured locally, and in some cases where a non-local item originated from. Both Animals and Plants (the producers of wool and cotton) take up strontium from the environment in the ratios at which they are present, without any fractionation, with the strontium isotope ratios found in their tissues reflecting a mixture of that found in the bedrock and overlying soil as well as that from ground and surface water. However, the different ecologies of Animals and Plants mean that even when they live in the same area, they will have slightly different isotopic ratios.

Although obtaining an accurate strontium isotope ratio only requires about 100-150 mg of material, the desiccated and friable nature of most archaeological fabrics means that the process typically leads to the destruction of one or two square centimetres of fabric, so the choice of materials sacrificed must be made very carefully.

Wozniak and Belka used four textile samples from recent excavations near the fourth cataract, made available for the study by the Polish Academy of Sciences, as well as three samples provided by the Sudan National Museum and the National Corporation for Antiquities and Museums in Khartoum. All bar one of these was from a burial site, and all are of either wool or cotton. The sample was chosen to include samples which were presumed to be of both local manufacture and imported. None of the samples showed signs of decomposition or degradation, making it unlikely that they had had their isotope signatures overwritten by strontium from groundwater after being buried, something which can be a severe problem in wetter climates, but is less so in arid Nubia.

The first sample chosen, NT1, was a section of decorated warp-faced woollen fabric (woollen fabric in which the threads are packed together closely, hiding the weft) with roughly 14-16 warp threads and 8 weft threads per cm². This fragment was recovered from the bottom of a funerary chamber beneath Tumulus 24 in the El-Ar 1 cemetery, near the El-Ar village in Shamkhiya District. The El-Ar 1 cemetery has been dated to the 2-3rd centuries AD, with decorated fabrics being very rare there.

Fragment of warp-faced wool tabby, sample NT1. Magdalena Wozniak in Wozniak & Belka (2022).

The second fragment used, NT2, is a piece of dark brown plain tabby woollen fabric, woven from S-spun threads, recovered from the El-Ar 4 Christian cemetery, and which could indicate a date anywhere between the sixth and fifteenth centuries AD. The fragment has 8-9 warp and 7 weft threads per cm², and comes from a section of shroud, which is in turn likely to have been a cloak or blanket used by the buried person and re-used as a shroud upon their death.

Fragment of the wool shroud, sample NT2. Magdalena Wozniak in Wozniak & Belka (2022).

Sample NT6 is a course wool plain tabby, woven with 8 warp and 8 weft threads per cm², taken from the shroud of a naturally mummified body (i.e. preserved by burial in a dry environment, rather than Human intervention). The fragment is brown in colour, and very desiccated. The mummy was buried in a supine position, with hands resting on hips, which is indicative of a Christian burial, again implying the 6th-15th centuries AD.

Close-up of mummy’s shroud, sample NT6. Magdalena Wozniak in Wozniak & Belka (2022).

Sample NT7 is a section of woollen kilim (woven tapestry rug). This has a dense weave with 7 warp threads and 32 weft threads per cm², and is woven in the slit-tapestry style, which creates the same pattern on both sides. The warp is a 2-ply z-spun (clockwise) cream wool, while the weft is single-ply z-spun wool in a variety of colours, including red, green, blue, orange, yellow and pink, giving a pattern of geometric and floral designs against a red background. The kilim was discovered in the town of Meinarti near the 2nd cataract in 1963, and has been dated to the 14th century AD.

Fragment of wool kilim from Meinarti, sample NT7, before conservation. Magdalena Wozniak in Wozniak & Belka (2022).

Sample NT3 is a piece of weft-faced tabby cotton cloth from the El-Ar 4 Christian cemetery. The threads are S-spun and of a golden colour, with the weft threads thicker than the warp threads, and a thread-density of 11 warp threads and 19 weft threads per cm². This is a common, medium-quality cloth type at archaeological sites on the Middle Nile.

Fragment of cotton tabby, sample NT3. Magdalena Wozniak in Wozniak & Belka (2022).

Sample NT4 is a fragment of cotton plain tabby from a Late Antique (2nd to 5th century AD) burial site at El-Ar. It is comprised of Z-spun white threads with traces of blue, green, and red pigments, probably a sign of resist dying (dying a finished fabric, while using wax or a similar substance to control which parts of the fabric the dye reaches), and a thread density of 22-24 threads per cm² for both warp and weft.

Bi-coloured threads from the cotton tabby, sample NT4. Magdalena Wozniak in Wozniak & Belka (2022).

Sample NT5 comes from the shroud of a naturally mummified body of unknown providence in the collection of the National Corporation for Antiquities and Museums. The textile is desiccated, low density, cotton tabby with a thread count of 8 warps and 8-9 wefts per cm². The threads are Z-spun, and have a golden colour. This mummy was also preserved in a supine position, with hands on hips, indicating a Christian burial, between the 6th and 15th centuries AD.

Fragment of cotton tabby woven from Z-spun threads, sample NT5. Magdalena Wozniak in Wozniak & Belka (2022).

Precambrian rocks of the Arabian-Nubian shield are widely exposed between the 2nd and 5th cataracts of the Nile. These exposures show a large variety of magmatic and metamorphic rocks, separated by suture zones with small occurrences of ophiolites and deep-water sedimentary rocks. To the west of the Nile Valley these are bordered by younger sedimentary rocks, predominantly Silurian sandstones. These rocks are overlain unconformably by the rocks of the Nubian Sandstone, which spreads across much of northern Africa, and includes a range of continental, esturine and marine sediments (predominantly sandstones), which in Sudan are predominantly of Cretaceous age. These Nubian Sandstone deposits have been mostly eroded away in the northeast of the country, exposing the underlying Precambrian basement, but extensive exposures are still present in the Bayuda Desert in the south, and the area to the west of the Nile. Tertiary Basalt rocks, which have intruded into these overlying strata, are exposed in the south of the country, around Abu Hamad. The floodplains of Nile Valley is also home to an extensive succession of Tertiary Sediments. 

Simplified geological map of northern Sudan. Red asterisks  indicate places where the investigated wool and cotton textiles were found. Numbers refer to the samples numbers. Inset shows the location of the study area. Wozniak & Belka (2022).

The volcanic rocks of the Precambrian basement of northern Sudan have strontium⁸⁷/strontium⁸⁶ ratios in the range 0.7024 to 0.7071, while the plutonic and metamorphic rocks of the basement have ratios in the range 0.7158 to 1.0039. The Cretaceous sediments of the Nubian Sandstone have strontium isotope ratios reflective of those of the basement rocks, with two clusters of values, around 0.7070 and 0.7160.

The alluvial sediments of the Desert Nile Valley are dominated by material derived from the Blue Nile and Atbara rivers, which drain from the Ethiopian Highlands, an area dominated by Cainozoic Volcanic rocks. These sediments have isotopic ratios in the range 0.7047 to 0.7076, a range which includes that of modern Nile water, at 0.7062. Nile muds dating from the end of the African Humid Period, about 45 000 years ago, show a strontium isotope ratio of 0.7052–0.7057, while those dating from between 1000 BC and 500 AD (i.e. between 3000 and 1500 years ago) show an isotope range of 0.7058–0.7076.

Samples NT1, NT2, and NT6 are all low-density woollen tabbies made from S-spun threads, technologically consistent with local manufacture. Sample NT1 contains cream and dark wool, used to make a pattern, while NT1 and NT6 are made entirely of undyed, dark wool. Lighter, cream-coloured wools are rare in the El-Ar assemblage, which may reflect the genetic structure of the local flocks (i.e. many brown Sheep and few cream Sheep). Alternatively, the local Sheep of the period may have been entirely brown, with cream wool being an imported commodity.

Sample NT7 is quite different from these, made up of coloured, Z-spun threads woven together using the split-tapestry technique to produce a decorative pattern, which is likely to indicate that this was an imported item. Furthermore, furthermore, the red dye used on some of the wool appears to have been treated with a lac dye, derived from the Scale Insect, Laccifer lacca, which is found in South and Southeast Asia and South China, but quite alien to the Nubia, where red dyes were traditionally derived from the Madder Plant. This apparently non-local item dates from the 14th century, a time when Meinarti was occupied by the Beni Ikrima, a nomadic group from the Maghreb region.

All four wool samples yielded strontium isotopic ratios within the range 0.7075 to 0.7084. In order to better compare these to the local environment, Wozniak and Belka also obtained strontium isotope ratios from Sheep and/or Goat remains (the two are hard to tell apart) from several archaeological sites between the 2nd and 4th cataracts. All of these remains are known to be older than the wool samples, dating to between 2500 and 500 BC, and gave isotopic ratios in the range 0.7068 to 0.7082 (one set of remains gave a much higher reading, of 0.7109, though this is likely to indicate that the Animal was of non-local origin).

Diagram showing the strontium isotope signatures of the  investigated wool textiles (yellow spots; numbers refer to sample numbers) in comparison to Strontium isotope  composition of Ovis/Capra remains found in the region  between the 2nd and 4th cataracts. Wozniak & Belka (2022).

Three of the investigated wool samples, NT1, NT6, and NT7, had isotopic ratios entirely consistent with archaeologically derived Sheep/Goat remains from the region, and the fourth, NT2, fell only slightly outside this range, and still within the range of Cattle and Human remains from the Desert Nile Valley.

Of the three cotton samples, only one, NT3, was made of a yarn spun in an anticlockwise, S-direction, typical of local manufacture in Nubia or Egypt. The other two, NT4 and NT5, were spun in a clockwise, Z-direction, which may indicate non-local manufacture. One of these, NT4, also shows signs of a resist-dying technique, not known in Nubia in the 2nd-5th centuries AD, but common in textiles from India during this period.

These three samples show a narrow isotope ratio range, from 0.7084 to 0.7086, which makes it likely that they were derived from crops grown within a limited geographical area. As a crop, cotton requires a rather specific set of conditions, with a hot dry climate, and large amounts of available water for several months of the year. This can be provided in the Nile Valley between the confluence of the White and Blue Niles and the 2nd cataract, although water and plants derived from this region show isotope ratios below 0.7075, ruling this region out as an origin point for the cotton in the study. Cotton is also grown around the Nile Delta, and ancient cotton fibres and seeds have been recovered from other sites along the Lower Nile, suggesting that cotton may have once been grown more widely. Strontium isotopic signatures are not available for sediments from all of these areas, but the isotopic signature of the Nile water, at about 0.7069, makes it unlikely that any of these floodplains have an isotopic ratio above about 0.7075. Similarly, the sediments of the Atabara and Blue Nile valleys have average isotope ratios between 0.7041 and 0.7060. All of these locations therefore seem unlikely as a point of origin for the cotton used to make the textiles included in the study. 

However, cotton grown in the Western Desert of lower Nubia typically has an isotopic value of about 0.7085, and limestones from the Western Desert of Egypt can have strontium isotope signatures in the range 0.7077 to 0.7078. Cotton is known to have been grown in this region in the 2nd-3rd centuries AD, using irrigation systems which drew upon shallow groundwater reservoirs, supplied by rainwater that had filtered through Eocene carbonates, which could conceivably produce a cotton with a strontium isotopic signature of about 0.7085.

Cotton is also known to have been produced in Meroe, on the lower White Nile to the south of Khartoum, during the Roman period, although both the water and plants from this region tend to have a higher isotopic ratio than the cotton from the study, excluding this region as a point of origin. 

Diagram showing the strontium isotope composition of the investigated cotton textiles (green spots; numbers refer to samples) in comparison to Sr isotope composition of selected elements of the natural environment in the Nile and White valleys. Wozniak & Belka (2022).

There is strong evidence for the production of cotton in Egypt and Sudan in Roman and later times, but the possibility of material also being imported from further afield should not be ruled out. Cotton was widely cultivated in India and on the Arabian Peninsula during antiquity, and both raw cotton and finished textiles could easily have been imported to Nubia via the Red Sea. Cotton seeds and textiles excavated at Mleiha in the modern United Arab Emirates have been shown to have originated from western India by their strontium isotope ratios. Unfortunately, strontium isotope information is only available for a few locations on the Arabian Peninsula or from India or Pakistan. However, several regions would be compatible with the isotope signatures obtained from the cotton samples, including the lower Indus River Basin, in India and Pakistan, the Kathijawar Peninsula on the west coast of India, the area covered by modern Kuwait, and the Oman Peninsula.

The isotope signatures obtained from the wool samples from archaeological sites in the Nubian Nile Valley all support a local origin. This is also supported by the technological features of three of the samples, NT1, NT2, and NT6, but not the fourth, NT7. This sample, a kilim rug from 14th century Meinarti, appears to have been made with a technology entirely alien to this part of the Nile Valley. However, it also dates from a time when the city of Meinarti was occupied by an alien group, the Beni Ikrima, who originated from the Maghreb region. Since the Maghreb has a geology with similar strontium isotopic ratios to that of northern Sudan, it is possible that the kilim was brought to the area by the Beni Ikrima from their homeland. However, it is also possible that the kilim was made locally by Beni Ikrima craftspeople from wool obtained from local, Nubian, flocks. Wozniak and Belka conclude that, given the available isotopic evidence, the more likely scenario is that the kilim was made locally, contrary to previous expectations.

In contrast, the isotopic signature of the cotton samples examined fits poorly with an origin in the Nile Valley. The samples all have very similar isotopic signatures, which seems to imply a common origin, although this is at odds with the different spinning techniques used. One of the samples, NT3, has an isotopic signature that fits well with the Dakhla/Kharga oasis in the Egyptian Western Desert, which has previously been identified as a possible site of ancient cotton cultivation. This piece also shows a technology consistent with production in the Nile Valley, notably an S-spun yarn. Importing cotton textiles from an oasis in the Western Desert would only make sense if local, Nubian, production was failing to meet demand. The frequency of cotton fabrics from archaeological sites in Nubia has been observed to have dropped during the Late Antique and Early Medieval periods, so this is not implausible, although an alternative explanation could be that the deceased person for whom this item was used as a shroud had travelled during their lifetime, either moving from the Western Desert to Nubia and bringing the cloth with them, or at some point visiting the Western Desert and obtaining the item there.

Sample NT4, on the other hand, appears much more likely to be of Indian origin, where a block printing technique using resist dying was common at this time. The isotopic signature of this sample matches that of the Kathijawar Peninsula on the west coast of India, leading Wozniak and Belka to conclude that this item was most likely imported from India. This adds evidence to the inclusion of the area around the 4th cataract into long distance trade networks, something previously indicated by the discovery of glass beads of foreign manufacture in the region.

Sample NT5 also appears to be non-local in origin, with a Z-spinning technique having been used, and a non-local isotopic signature. However, there is insufficient evidence to give a precise origin for this fragment at this time, with possible points of origin including  Indus River basin in Pakistan, the west coast of India, Kuwait, and the Oman Peninsula.

Although Wozniak and Belka were not able to determine the point of origin of all the fabrics in the study, their study shows the potential for determining the origin of ancient fabrics using strontium isotope ratios. The development of a wider data set of strontium isotope values for textiles could add to the information obtained by examining the technological aspects of textile-making, to provide a better understanding of the manufacture of and trade in cloth in the ancient world.

See also...

Follow Sciency Thoughts on Facebook.

Follow Sciency Thoughts on Twitter.

The Cabeço da Amoreira burial: An Early Modern Era West African buried in a Mesolithic shell midden in Portugal.

The Tagus and Sado valleys of central Portugal contain numerous shell middens, dating back to the Late Mesolithic, roughly 6500 to 5000 years ago. As well as depositories for waste shells, these sites were used as burial grounds by the people who made them. A number of these sites were excavated by archaeologists in the 1930s, producing a series of sets of Human remains, buried within the middens without grave goods. One site, Cabeço da Amoreira at Muge in the Tagus Valley yielded an individual noted at the time as being both notably taller and better preserved than others recovered from such sites. More recently, scientists working on a database of Mesolithic European genomes have begun to sequence individuals from these Mesolithic Portuguese sites, including the Cabeço da Amoreira individual, in the process of which they found that this individual was not closely related to other individuals from Mesolithic burials in Portugal, or elsewhere in Europe, but rather appeared to be of African descent.

In a paper published in the Journal of Archaeological Science: Reports on 21 February 2022, Rita Peyroteo-Stjerna of Human Evolution at Uppsala University and the Centro de Arqueologia da Universidade de Lisboa, Luciana Simões, also of Human Evolution at Uppsala University, Ricardo Fernandes of the Department of Archaeology at the Max Planck Institute for the Science of Human History, the School of Archaeology at the University of Oxford, and the Faculty of Arts at Masaryk University, independent researcher Gonçalo Lopes, and Torsten Günther and Mattias Jakobsson, again of Human Evolution at Uppsala University, present the results of the follow up study which used multiple lines of enquiry to determine the origin of the Cabeço da Amoreira individual.

 

Location of Cabeço da Amoreira shell midden (indicated by the star), Muge, Tagus valley, Portugal. Peyroteo-Stjerna et al. (2022).

Radiocarbon dating of material from the Cabeço da Amoreira site, including bone, charcoal and shells, have produced dates of between 6500 and 5000 BC, consistent with a Mesolithic origin for the site, however, radiocarbon dating of collagen from the individual buried at the site yielded dates between 1529 and 1763 AD, and probably between 1631 and 1793, consistent with an Early Modern origin. 

Relationships between Early Modern Europe and Africa were dominated by the trans-Atlantic slave trade, which saw millions of people taken from Africa and shipped to European colonies in the New World, and to a lesser extent Europe itself. Portugal is estimated to have directly imported 2-3000 African slaves per year between the fifteenth and nineteenth centuries. Most of these remained enslaved their whole lives, though some were freed and able to live relatively independent lives, albeit very much at the bottom of the social scale. 

Because mitochondrial DNA is found in the mitochondria, organelles outside the cell nucleus, it is passed directly from mother to child without being sexually recombined each generation, enabling precise estimations of when individuals shared common ancestors, at least through the female line; this is known as the female haplogroup. It is also possible to trace direct ancestry through the male line, using DNA from the Y chromosome, which is passed directly from father to son without sexual recombination; this is known as the male haplogroup. Since everyone has mitochondria, it is possible to determine the female haplogroup of all Humans, but generally only males have a Y chromosome and can be assigned to a male haplogroup.

Genetic analysis of the Cabeço da Amoreira individual established that he had a Y chromosome, indicating that he was male. It was also possible to determine both his male haplogroup. He was found to belong to the E1b1a male haplogroup, which is the most widespread in sub-Saharan Africa, being commonly found in Nigeria, Congo, Cameroon, Gabon, Guinea-Bissau, and among Bantu-speakers in Southern Africa.

A principle component analysis based upon his entire recoverable genome revealed that Cabeço da Amoreira man showed a greater genomic similarity to West Africans than to other populations, and in particular, to people of Gambian or Mandinka origin. 

 

(A) Principal component analysis. Worldwide modern populations (circles coloured according to continent) and Cabeço da Amoreira man projected as a yellow, red outlined diamond. (B) Geographic distribution of the genetic affinity of the studied individual with modern African populations, measured by outgroup-f₃. The two highest f₃ scores are depicted with diamonds. Peyroteo-Stjerna et al. (2022).

Peyroteo-Stjerna et al. next looked for alleles (gene variants) associated with sub-Saharan populations, finding that Cabeço da Amoreira man had a number of alleles which would further support an African origin, notably the FY*B allele, which is associated with resilience to Malaria, and a number of skin pigmentation alleles, namely MFSD12 rs10424065; DDB1 rs11230664; OCA2 rs1800404; SLC45A2 rs16891982; and HERC2 rs6497271, which are more commonly associated with sub-Saharan African populations than with Europeans (skin pigmentation is complicated, genetically speaking, and it is not possible to directly determine someone's exact skin tone from their genome at the current time, but it is possible to associate allele abundances with specific populations). Cabeço da Amoreira man also lacked the alleles for lactase persistence (i.e. retaining the ability to digest milk into adult life), sugesting that he was lactose intolerant, something more common in Africans than Europeans.

A stable isotope analysis for carbon and oxygen isotopes, based upon bone collagen from Cabeço da Amoreira man, suggested that when he was growing up his diet comprised largely C₄ Plants, supplemented with seafood. A diet of C₄ Plants is not at all typical for Portugal (or elsewhere in Europe) in the Early Modern period, although it would have been common in parts of West Africa, notably the Sahel Region (which reaches the coast in the Senegambia region and southern Mauritania), where the principal crops for the time would have been Sorghum and Millet, both of which are C₄ Plants. Further south, in the West African forest zone, the principal crops were Rice (a C₄ Plant) in the west and a more mixed vegecultural diet (also based around C₄ Plants) in the east. Therefore, the C₄ Plant component of Cabeço da Amoreira man's diet makes it likely that he came from the Sahel region, and the seafood component further ties him to the Senegambia and Mauritania region.

 

Estimated area of origin of Cabeço da Amoreira man (mug019) in West Africa and place of burial in Portugal. Traditional plant food-producing systems in West Africa. Peyroteo-Stjerna et al. (2022).

Around 35 000 slaves were brought to Portugal from Africa between 1514 and 1866. Records of these movements are fairly complete after 1750, but older records are somewhat patchy, making the origin of Cabeço da Amoreira man difficult to reconstruct in this way. However, it is known that slaves were brought to Portugal from predominantly from Guinea-Bissau and The Gambia, with smaller numbers arriving from the Cape Verde islands, Princes Island and São Tomé, Bance Island (Sierra Leone), the Gold Coast (Ghana), Senegal and Whydah (on the coast of modern Benin). 

Most slaves in Portugal during this period would have been baptised as Christians, and buried in Christian burial grounds. However, there are records of slaves being buried in other ways, including by roadsides, in wastelands or in Olive groves. The Church generally kept good records of births, deaths, marriages, and baptisms during this period, for all social classes including slaves, which offered some hope of discovering the identity of Cabeço da Amoreira man. Peyroteo-Stjerna et al. were able to identify two deaths of interest in the Cabeço da Amoreira area in the seventeenth century, the first of an unnamed slave on 5 May 1633, for whom no burial location is listed, and the second of the murder of a man named João at Arneiro da Amoreira on 1 November 1676; João is described as being brown skinned, which may indicate that he was of mixed origins, but he was buried in a churchyard, so presumably was not Cabeço da Amoreira man.

One notable feature of the Cabeço da Amoreira burial is that the body does not appear to have been buried hastily, but rather laying upon a bed of sand which had been used to line the grave, something not seen in Mesolithic shell midden burials (the difference was noted at the time of excavation, but the significance of this, understandably, was not realised).This implies that the burial at this location was planned and carefully executed, rather than being the hurried disposal of the body of a slave or murder victim.

Shell midden burials, both ancient and modern, are known from the Senegambia region, and are still sometimes practiced among Serer fishermen in the Saloum Delta. Here, some families maintain temporary settlements on islands deep within the delta, which are used for four-to-five months each year, when shellfish are harvested. Since these sites are essentially located on shifting sandbanks, the shell middens that build up their form stable hardgrounds, which can be used for purposed such as supporting structures and burying anyone who dies while the temporary villages are in use.

 

Modern cemetery on a shell midden, at Fadiouth in the Saloum Delta, Senegal. Hardy et al. (2015).

This does not unequivocally tie the Cabeço da Amoreira burial to the Senegambia region, but does create a plausible scenario in which members of a community transplanted to Portugal, who had practiced shell midden burials in their homeland, might have chosen to recreate the practice in their new environment.

See also...

Follow Sciency Thoughts on Facebook.

Follow Sciency Thoughts on Twitter.