African Gold-weights in the British Museum

Fiona Sheales

See all objects in this catalogue 

Share this catalogue

The manufacture and use of gold-weights


All weights can be divided into two broad categories: those of natural origin such as seeds, stones, shells, bones and flints which are not cast in metal; and those manufactured by man using a variety of methods.

Customarily, the basic unit in the Akan system of weights was the tiny, hard, red and black spotted seed of the Abrus precatorius, a tree that is found throughout the tropics. Seeds such as these (damma) were freely available to all and were probably in use long before cast metal weights became widely available. Every gold-weight, irrespective of its design, size and weight, is called abrammuo (pl. mrammuo) in Twi, the language spoken by the Asante.

The introduction of metal-working techniques from Western Sudan and the spread of the trans-Saharan gold trade into the tropical forests of the Gold Coast in the 14th century stimulated the production of locally made weights. Timothy Garrard argues that the earliest weights produced by Akan goldsmiths were based on those used by Islamic traders, but these were supplemented within a relatively short period by weights cast in the form of animals, insects, birds, tools and weapons (Garrard 1980). Garrard based his phased gold-weight chronology on the stylistic evolution of forms and informed guesswork. However, as no securely datable gold-weights have been recovered from archaeological contexts, the introduction and long-term development of the gold-weight corpus remains highly speculative. It is hoped that improved dating techniques, new technology and documented excavations will provide a more accurate and detailed chronological sequence in the future.

Historically, goldsmiths were also responsible for making other items of equipment that were used in the weighing of gold, such as the cast or riveted brass boxes used for storing gold-dust (adaka), shallow brass scoops used to remove impurities (famfa) and highly decorative brass spoons (nsawa) for putting the dust onto the scales (nsania).

A small balance scale, two gold dust boxes and a scoop, 19th century, Gold Coast

Besides these items casters also made two different sizes of brass balance scales; the larger, known as akontuma, were used to weigh big amounts and the smaller scales, known as mframa nsenia (literally ‘wind scales’), were so light and delicate that a breath of wind could cause them to fluctuate. Whereas the majority of gold-weighing equipment was cast in brass, Bowdich claimed that Asantehene Osei Bonsu’s (r. 1800–24) scales, blow pan, boxes and weights were made of the purest gold (Bowdich 1819, 312). Unfortunately, these items have not survived, but given the level of power and wealth enjoyed by the Asante at this time there is no reason to doubt Bowdich’s assertion.

The majority of weights were made using two related techniques: in the case of a direct casting, a clay and charcoal mould was formed around an actual object such as a beetle, crab claw or seed pod, which was incinerated when heated leaving a void that could be filled with molten metal. The second technique, known as lost wax casting or cire perdue, required a wax model or mould to be created. Evidence suggests that the wax (akaa) was sourced locally from the hives of wild bees or was traded from the north in the form of flat round cakes (akaa tatere) measuring about 1 ft (30cm) in diameter (Garrard 1980, 119). The mould (foa) was formed by rolling and shaping the wax on a wooden block (adwini pono), using a wood or bone spatula (adwini dua, adwini nua). Details were added by cutting, incising and carving into the wax using heated blades (sekan), wood or bone spatulas, a pointed bamboo stick and an iron needle (mpane). When the mould was completed the smith attached one or more thin, hollow wax rods in places where they would be unobtrusive in the final casting. These rods would eventually protrude through the clay mould and melt away during the casting process, leaving channels through which the molten metal could flow into the mould. The mould was then allowed to harden off completely.

Once the wax mould was hard the smith either dipped or painted a layer of clay slip, which had been mixed with finely ground charcoal, onto the exterior of the mould taking care to ensure that all parts were covered. After the first layer had dried further layers were added until the clay casing measured around half a centimetre in depth. An outer casing of clay was created by carefully pressing wet clay around the inner clay slip casing. The completed mould was then allowed to dry. When ready, the mould was heated in the furnace (ebura) and taken out and inverted. This procedure ensured that any molten wax that had not soaked into the mould walls was drained away. A small clay cup (semoa) containing pieces of brass was grafted onto the mould around the hole/s left by the wax rods. Clay was then plastered heavily around the joint and the mould was allowed to dry thoroughly (see McLeod 1981, 80–2 for a full description).

On the day of the casting, the mould and cup were placed in the furnace, the cup downwards, until the metal was judged to have become completely molten. This process required the charcoal fire to be kept at a constant high temperature by having air pumped into it by one or more bellows (afa). At the right moment the mould and cup were seized with a pair of tongs (odabaw) and swiftly inverted, which caused the molten metal to rush down the channels and into the cavity. The mould was held in this position, or placed on the ground while the molten metal cooled and hardened within the cast. The mould was then smashed open and the casting picked out. When it had completely cooled the casting was cleaned by picking off the black moulding material that had baked on in the heat. It was then scrubbed in water mixed with the juice of limes and trimmed.

Although the great majority of gold-weights were made using either the direct casting or the lost wax casting method, a small number were created using a combination of these techniques. Gold-weights cast in the form of seeds, pods and small fruits are direct castings, but their stems and stalks in some cases have been moulded into the shape of handles using wax in the way described above.

This gold-weight of a male oil-palm inflorescence surmounted by a handle combines direct and lost wax casting techniques

A very small number of weights have also been shaped from scraps of brass by hammering and filing; such examples possibly represent amateur work.

A gold-weight in the form of a European-style chair with curved backrest, made from machine-cut brass, 19th–20th century

There is also a final category of weight that will be referred to in this publication as appropriated weights (Garrard calls them pseudo-weights). These are objects made of brass and other base metals that were not intended to function as gold-weights but came to be used as such. Examples of this weight type include objects as diverse as glass beads, pebbles, cowrie shells, small animal and fish bones and carefully reshaped sherds of European china as well as a plethora of brass oddments of European, African, Indian and Asian manufacture.

A cast brass figure of a cow and calf, 19th century, probably made in India

By definition, all weights had to weigh an exact amount in order to function within the context of trade and exchange. Manufacturing an accurate weight was therefore dependent upon the caster’s ability to judge the exact size of the mould and how much molten metal would be required to fill it. Evidence suggests that in the event of a weight being very much under the required amount the caster would abandon his intention to make the particular weight he had first aimed for and diminish it to the next nearest lower standard. If the weight was very much over the required amount, he increased it up to the weight above; and if the weight were only slightly under or slightly over the required amount he would make it up or diminish it to the weight he originally intended. In each case the diminishing process was achieved by drilling holes, filing or cutting pieces off the weight, and the incrementation either by adding a ring or a link of metal or by pouring lead into the cells and cavities that form the relief decoration on the top of a weight.

The common practice of adding base metal to weights suggests that recessed motifs and patterns were not purely decorative but, as in cloisonné, provided convenient cells and compartments that could be filled if required. Interestingly, some additions of base metal have been sensitively done so as not to deface the decorative schema of a weight, whereas others are crudely applied with no reference to the underlying motifs.

A lost wax cast gold-weight in the form of an oware game-board, 19th–20th century. A number of the cups have been infilled with lead in order to increase the mass

A rectangular geometric gold-weight that has had lead added to it in order to increase the mass, 19th-20th century

The differences observed in the treatment of weights may represent interventions by users rather than by professional craftsmen. Anticipated weight adjustment also took other, less intrusive and more aesthetically pleasing forms as some weights were cast with a circular or ovoid hole in the base.

Evidence suggests that weights were also subjected to these types of alterations throughout their existence, not just at the point of manufacture. The Museum’s collection contains a small number of weights that show clear evidence of multiple interventions, specifically that they have pieces cut off and metal links or additions added which suggests that at different times their mass has been adjusted both up and down.

A gold-weight of an antelope which has had been modified in two different ways in order to decrease and increase its mass, 19th–20th century

These actions may reflect successive economic fluctuations (inflation and deflation) that occurred naturally over time. Therefore it is possible that gold-weights embody in their physical forms evidence of past local and global economic cycles.

The remains of the casting sprues (known as gyinae), which are frequently seen on weights, present casters with another method for altering the mass of a weight. These sprues are all that remain of the hollow wax pouring channels through which the molten metal entered the mould. In many cases, the caster has chosen to leave one or more sprues untrimmed in order to make up a deficiency in a weight’s mass. Commonly, the remains of sprues prevent a weight from standing upright or squarely, either because they are sited on the base or because they cause the weight to over-balance. However, the fact that a weight cannot stand upright does not impair its economic function.

An untrimmed casting sprue protrudes from the chest of this antelope gold-weight, 18th–20th century

In the past, most adult men possessed their own set of weights or had access to those of a relative or a friend. Some weights may have been commissioned, while others may have been inherited from maternal forefathers. This suggests that some surviving individual weight sets therefore represent the collective accumulations of generations of male family members. This practice may also partially explain the presence of many duplicates especially of the more commonly used weight units that are found in individual assemblages. For the Akan, there seems to have been no such thing as a ‘set’ of weights comprising one specimen of each size in a regular series. Another noteworthy aspect of weights is that the Akan believed that, like other intensely personal objects such as stools and cast metal ornaments, weights could become repositories and vehicles for their owner’s soul over time. This enduring and intimate relationship was believed to enable the spirits of the ancestral dead to project their authority beyond temporal and spatial boundaries. This suggests that ancestral spirits may also have been present at trading transactions, in the same way as they participated in other ritual contexts.

Customarily, when a boy approached adulthood he would practice weighing gold-dust using a small set of weights, miniature scales, spoons and scoops.

Goldweighers, Côte d’Ivoire, 1892 (from Monnier 1894)

Learning the correct way to handle gold-dust was an essential skill as it is sticky to touch and can easily escape from gaps in containers or be blown away on the wind. Opanin Kwaku Badu (b. 1880) remembered that it was important not to breathe or shake when weighing gold (quoted in Garrard 1980, 175). Rattray, who witnessed first-hand the weighing of gold-dust in 1923, described how the balance scales were held between the second finger and the thumb of the left hand, with the three free fingers ready to rest on or under the arm of the beam in order to prevent a too sudden drop of one of the trays, which would cause the dust to spill (Rattray 1923, 305). Several accounts record that any gold-dust that fell to the ground in the palace or the marketplace in Kumase could not be picked up again on pain of death as it was regarded as Asante state revenue (Bowdich 1819, 76, 257; Garrard 1980, 175–6). This gold was periodically reclaimed by panning the surface during times of national emergency.

The alluvial and terrestrial origins of gold-dust may have led to its being perceived by Akan communities as inherently ambiguous and liminal, unlike some cast gold items which functioned as important emblems of royal authority and reflected the wealth and status of the king. This difference in perception suggests that casting gold into a solid form fixed its potential meaning and value for a time, whereas gold-dust retained a certain amorphous, fluid and marginal quality that enabled it to flow, circulate and be divided easily. It is suggested that the versatile qualities associated with gold-dust in particular must have played a part in its becoming the exchange medium for the majority of all transactions on the Gold Coast by the mid-18th century.

Perceptions concerning the qualitative properties of gold-dust have their roots in the gold trade, and technical transformation processes such as casting may have provided the Akan with metaphorical and/or analogous models that informed social beliefs, values and social ethics. In effect the smelting and casting processes informed or reinforced notions regarding the properties and qualities associated with cold and hot, solid and liquid, of purifying and cleansing and the creation and dissolution of form and meaning in other contexts. In this way, Akan culture and technology can be understood to form a dialectical relationship, whereby society was partly structured by technology and technology was partly structured by society.

The act of weighing gold-dust may have also given rise to a conceptual model for structuring ethics and social values. In particular it appears to have informed the cultural importance that is attached to accuracy, balance, control, equipoise and even-handedness. All of these qualities are also associated with negotiation skills and an ideal inner state of coolness (dwa) which enabled individuals to assess, evaluate and judge effectively. These qualitative attributes, which were reinforced through the physical act of weighing gold-dust, are therefore implicated in every decision and action. This impacts on how gold-weights were perceived by their creators and users. Far from being merely utilitarian objects used for transacting business, they were also devices for the manipulation and regulation of internal emotional states. As a result, it seems likely that weights were also valued for their psychological and physiological associations as well as for their aesthetic, artistic and economic functions.

At the start of every transaction the buyer and seller first had to agree a price for the goods and then agree on the choice of weights to be used to weigh the gold-dust. Each weight user would check his own weights with those, supposedly of the same unit, of the person with whom he was dealing to ensure that they were equivalent. Individuals knew the value of each weight in their own weight set, which in some cases could number upwards of 100 weights or more. The form and size of their weights primarily depended on the wealth and status of the owner. We can therefore assume that the most attractive and large weights were the property of prosperous traders, rich officials or a chief or king. Some sources claim that the weights of senior chiefs and those used for some royal functions were heavier per unit than those of ordinary people (McLeod 1981, 124). The difference in weight constituted the commission of the official acting on behalf of the chief.

A heavy geometric gold-weight which was probably used within a judicial or royal context, 18th–20th century

It is also probable that in some cases the gold-dust was weighed twice, once by the buyer and then again by the seller, each using their own weights in order to check that the amount was correct. It has been widely assumed up until now that only one weight at a time was put in the balance scale when weighing gold-dust. On the face of it, this assumption seems reasonable given the wide range of weight units created and the size of the scale balances used; however, this does not preclude the possibility that smaller sized weight units were used in combination if a larger single weight was unavailable.

The amount of time taken to negotiate each stage of the transaction process in all probability correlated to some extent to the amount of gold-dust that was changing hands. Hence high-value purchases were more likely to be the result of protracted negotiations whereas small everyday transactions were probably concluded relatively quickly as these were straightforward. Eyewitness accounts provide circumstantial evidence that the majority of transactions took place in the public marketplace or in the semi-private threshold of a trader’s house. Akan thresholds, in the pre-colonial period, were architecturally defined spaces that took the form of open verandas known as adampan that ran the length of the front of a dwelling.


Dampans in Adum Street, Kumase 1817

This area interfaced between the private domestic sphere of the residence and the public arena of the street and therefore was the ideal site for the transacting of business, both economic and social. In this way, economic transactions were concentrated around public or semi-public contexts and were governed by conventions and norms that promoted social interaction.

The primary purpose of gold-weights was for use in trade, but some of the weights in the British Museum’s collection were cast with suspension loops or rings that enabled them to be worn, possibly as amulets and charms in order to bring good fortune or to preserve the wearer from harm.


A geometric gold-weight cast in the shape of an amulet known as a safi, 19th–20th century

Evidence suggests that they were occasionally worn by sickly children to restore them to good health (Garrard 1980, 201). It also appears that gold-weights were occasionally employed as a means of communication. Rattray, for example, noted that when a man owed another £1, the creditor might send the weight representing a suru (£1) to the debtor as a reminder to pay up (Rattray 1923, 306; Garrard 1980, 202). On the whole, however, the evidence for this type of weight use is very limited.

When not in use gold-weights were normally kept in special containers such as a leather bag (futuo) or a chest (adaka). Gold-weights owned by chiefs and kings formed part of their treasury (dampon kese) and as stool property (inalienable possessions of the state) were in the care of the head treasurer, known as the Gyasehene, who carried a gold key (sika safoa) or a large bunch of keys as a badge of office. The treasury collection might contain upwards of several hundred weights and were far larger than the ordinary man’s futuo. There is no evidence to suggest that particular weight forms or designs were reserved exclusively for chiefs, queen-mothers or other members of the elite. In southern Ghana and Asante a chief was not allowed to use weights or to operate the scales, nor could he open the leather bag or great chest (adaka kese) in which gold-dust was kept. Instead he employed trained weighers known as afotosanfo, who, as part of their duties, were responsible for keeping the chief’s household accounts. Interestingly, the afotosanfo were also servants of the bedchamber and functioned as royal barbers and manicurists. In this role they oversaw the safe dispersal of items of bodily waste, such as hair clippings and nail parings taken from the chief. At first glance, it looks as if these two roles are unconnected, but in fact they are both intimately concerned with the same action: that of distributing inherently liminal and transgressive substances.

Weights belonging to a chief were stored in a special container known as a sanaawith a store of gold-dust. The sanaa typically contained larger weights, with weighing values in excess of 80lbs, which were used only for the most important transactions and judicial fines. They were kept wrapped in cotton cloths, which were enclosed in a piece of leather made from an elephant’s ear that in turn was securely tied up with cords. The sanaa was kept together with the state regalia and ancestral stools in a particular room in the palace. The whole bundle would sometimes be placed in a brass vessel (kuduo) or an imported brass bowl (yawa) so that it did not touch the ground. Gold-dust, especially if there were great quantities of it, was also kept in cast brass or bronze kuduo along with precious items of jewellery. These beautifully shaped and richly decorated vessels were also believed to be connected to a man’s ntoro group and also with his soul (kra).


Cast brass kuduo vessel decorated with bands of linear geometric designs, mid-19th century, Asante

The gold-dust stored in the sanaa and kuduo was used to pay for public festivals and rituals, including marriages and the funerals of members of the chief’s family. Every time the sanaa was opened a libation was poured, accompanied by prayers to the ancestors, after which the treasurer struck the bag with his hand before it was unlocked by one of the afotosanfo. The popular belief was that, if this ritual was not observed, god would punish the transgressor with blindness. This conviction, shared by the whole population, safeguarded the treasure from theft. The sanaa was taken outside the palace grounds only on special occasions, for example during the major soul-washing rite known as Odwira, when it was publicly displayed. In times of emergency it would be hidden by trusted members of the court.

Another bundle, known as a futuo, contained the many smaller weights of diverse design that were used for daily domestic transactions. These weights, along with other weighing equipment including touchstones (twaboo) for testing the purity of nuggets or cast gold, were also wrapped in layers of cotton cloth. The bundle was further wrapped in an outer covering made from the skin of a civet cat (bata), antelope (adowa) or a monkey (akyeneboa). Small talismanic objects such as fragments of Neolithic axe-heads, glass beads and Islamic amulets (suman) were also added to the contents of futuo possibly as a means of protecting against theft, warding off harmful influences and bringing good fortune in trading (Garrard 1980, 182, McLeod 1981, 133). The futuo accompanied the chief whenever he travelled and was even taken on military campaigns. The futuo was carried in a finely embroidered shoulder-bag or leather pouch (kotokuo, pl. nkotokwaa) which was one of a pair, the other being used to hold the chief’s gold-dust. The bags had protective charms of gold and silver attached to them and were carried by the royal treasury bag bearer, known in Asante as okotoku kurani. Timothy Garrard was told that leather workers from northern Ghana once specialized in making these pouches for sale to the Akan. In 1817 Thomas Bowdich was presented with such a bag by Asantehene Osei Bonsu as he departed Kumase having successfully concluded the first Anglo–Asante trade treaty. This beautiful bag was donated by Bowdich to the British Museum on his return to London in 1819.

A fringed leather shoulder bag with appliqué decoration (kotokua) given to Thomas Bowdich in 1817

The care that was taken to safeguard gold-weights indicates the high value placed upon them as esoteric and functional objects.

1. The origins and history of gold-weights 
3. Gold-weights and proverbs 
4. Gold-weights at the British Museum 
5. A guide to weight types