How Medieval Caravans Crossed the Sahara: Camels, Guides and Salt

On February 13, 1352, a caravan left Sijilmasa, the oasis city on the northern edge of the Sahara in what is now southeastern Morocco.¹ Among its members were merchants of the town and the Moroccan traveler Ibn Battuta, then in his late forties, who had bought camels there and, in his words, “a four months’ supply of forage for them.”² Two months later, to the day, the caravan reached Walata, the northernmost outpost of the Mali empire, where the sultan kept a deputy.²

His account of that crossing, and of a return journey through the central Sahara in 1353, is one of the most detailed descriptions of a working caravan to survive from medieval West Africa. It is not the only witness. Nearly three centuries earlier, the Andalusian geographer al-Bakri had assembled reports from merchants about the towns, tolls and goods of the trade. Read together with archaeology from the caravan towns, the two writers show a system that ran less on endurance and romance than on agreements, hired specialists, careful water planning and taxes.

Two different kinds of witness

The sources need to be read for what they are. Al-Bakri (died 1094) lived in Córdoba, Seville and Almería and, as far as is known, never left Iberia. His book of routes and kingdoms, completed in 1068, was compiled from earlier writers and from people he met, sometimes copying older texts almost word for word.³⁴ The Northwestern University and Block Museum exhibition “Caravans of Gold” warns that medieval Arabic descriptions of the Sahara usually rest on “second- or third-hand reports,” but adds that they record things archaeology cannot, especially the movement of people.¹

Ibn Battuta did travel, but his book is not a diary. He dictated it after returning to Morocco, and Ibn Juzayy, who wrote it up, says plainly that he abridged it; the dictation was finished on December 9, 1355.⁵² The English used here is H. A. R. Gibb’s 1929 selection, itself shortened. Exact figures and anecdotes may have been shaped by memory and retelling, but the routine details (where water was taken on, what was paid and to whom) are less likely to have been invented.

It also matters which period is meant. “Trans-Saharan trade” covers more than a thousand years. The Caravans of Gold project dates the main desert routes to the fifth century CE and an escalation of the trade from the eighth century, which contributed to the rise of Ghana (seventh to thirteenth century), Mali (thirteenth to sixteenth) and Songhai (fourteenth to sixteenth).¹ Al-Bakri describes Ghana in the eleventh century; Ibn Battuta describes Mali under Mansa Sulayman, brother of the better-known Mansa Musa. They are not snapshots of one unchanging desert.⁵

Animals and loads

The dromedary made the long crossings practical, but it was a relative newcomer. A 2022 review of camel history by Bernard Faye places its domestication in southeastern Arabia roughly 3,000 to 4,000 years ago, its presence in Egypt’s western desert from the end of the fifth century BC, and its arrival in the central Sahara “slightly before the Roman Empire,” on the evidence of rock art in the Libyan desert.⁶ The precise chronology is still debated, and these are broad brackets rather than firm dates.

Ibn Battuta’s narrative shows how camels were used in practice. At Taghaza, the salt-mining station 25 days south of Sijilmasa, salt was cut in thick slabs, and “A camel will carry two of these slabs.”² On the return trip he set out with one camel to ride and a she-camel for his provisions; she gave out after the first stage, and the caravan’s guide had her load shared among his own people.⁵ At Takadda he paid 37 and a third mithqals of gold for two riding camels. In Mali, by contrast, he rode a camel because horses were expensive, a hundred mithqals each.⁵²

Camels did not carry everything everywhere. Al-Bakri’s informants describe salt entering Ghana by the donkey-load, which suggests that south of the desert, goods moved on to other animals and other routes, including, as Caravans of Gold notes, the Niger River, long a trade corridor in its own right.⁴¹ Al-Bakri also notes that at Zawila in the Fazzan, in what is now southwestern Libya, camels worked the irrigation of palm groves.³

Water, oases and the hired scout

Caravan routes ran between water sources. Caravans of Gold describes the main Saharan routes as threading between oases, fed by groundwater.¹ Ibn Battuta’s caravan took on water at Taghaza for the next leg, ten nights with almost no water on the way. That year they were lucky: rain had left pools in the desert. Even so, the desert was dangerous. He and his companions had been riding ahead of the caravan to find grazing until one of their party was lost; after that, he writes, he neither went ahead nor lagged behind. Later, when another caravan told them some of its people had strayed, they found one of those men dead under a shrub, about a mile from water.²

The next halt, which he calls Tasarahla, had underground water. Caravans rested there for three days to repair their waterskins, refill them and stitch on sackcloth covers against the wind.² From there the caravan sent ahead a takshif, a man of the Massufa, a Sanhaja Berber group (as Gibb identifies them) whose members also appear as residents of Walata and Timbuktu in the same account. He went ahead to Walata carrying letters from the travelers to friends or to reputable merchants, who rented lodgings for them and came four nights out into the desert with water to meet the caravan.²⁵

This was a paid and dangerous job. Ibn Battuta says the takshif sometimes died on the way, and when that happened the people of Walata did not know the caravan was coming, and many or all of its members could die. He gives the local explanation, demons that confuse a lone rider, and also a physical one: “there is no visible road or track in these parts, nothing but sand blown hither and thither by the wind.”² His own caravan paid its takshif a hundred gold mithqals. Its guide, he observed, was “blind in one eye, and diseased in the other, yet he had the best knowledge of the road of any man.” On the seventh night after Tasarahla they saw the fires of the party sent out from Walata.²

Caravans of Gold describes this division of labor more generally: Amazigh nomads who knew the desert routes managed the caravans, North African merchants worked under Islamic commercial law, and Wangara merchants south of the desert carried trade onward along the Niger and its tributaries.¹ Once inside Mali, the logistics changed. Ibn Battuta went on from Walata to the capital, 24 days away, with only a hired guide and three companions, because the road was safe enough not to need a caravan. Travelers there carried no food or coin, only salt, glass beads and some aromatic goods to trade for millet, milk and chickens in the villages.²

Salt south, gold north, and much else

Salt and gold are the famous pair, and the sources show why. At Taghaza, Ibn Battuta reports, the salt was dug by enslaved workers belonging to the Massufa, who lived on dates and millet brought in from elsewhere. A load of salt sold for eight to ten mithqals at Walata and for twenty to thirty, sometimes forty, at the Mali capital; salt was also cut into pieces and used as money.² In the tenth century, the geographer Ibn Hawqal had already described Ghana’s dependence on salt arriving from the north through Awdaghust, and said a load could fetch 200 to 300 dinars in the remoter interior.⁴

Al-Bakri’s informants explained how Ghana’s king profited. “On every donkey-load of salt when it is brought into the country their king levies one golden dinar and two dinars when it is sent out.” Copper paid five mithqals a load and other goods ten. The best gold came from Ghiyaru, eighteen days from the capital, and nuggets from the mines were reserved for the king, leaving gold dust to the public.⁴ Awdaghust, at the southern end of a desert route, appears in the same work as a town of wells and date gardens with a market so crowded and loud that a man could barely hear his neighbor, where “Their transactions are in gold.”⁴

Archaeology adds detail. At Tadmekka, at the desert’s southern edge in northeastern Mali, the ruins include caravanserais and livestock enclosures, and excavations there, first carried out in 2004 and 2005, have yielded glass from Egypt or western Asia, glazed ceramics and a fragment of Chinese porcelain.¹ The archaeologists Sam Nixon, Thilo Rehren and Maria Filomena Guerra identified ceramic moulds there used to cast gold coins in the ninth and tenth centuries; they argue the coins were probably blank pieces made for the caravan trade, to be melted down or made into inscribed coins once they reached the Mediterranean coast.⁷ At the northern end, Sijilmasa’s mint was striking gold coins by the tenth century.¹

Not all trade was salt and gold. Ibn Battuta describes copper at Takadda being cast into bars that served as currency and were sent south and east, though where Takadda was is still debated; even his 1929 translator noted observers who had found no copper at the site usually proposed.⁵ People were traded too. Al-Bakri reports enslaved people bought at Zawila for short lengths of red cloth and sent north to Ifriqiya, and Ibn Battuta’s northbound caravan from Takadda in September 1353 included six hundred enslaved women.³⁵ Caravans of Gold notes that slavery in this trade is visible almost only in written sources, and that enslaved West Africans, mostly women and children, crossed the desert on foot, many dying on the way.¹

Timing, tolls and protection

Ibn Battuta’s dates give one traveler’s calendar, not a rule for every caravan. He left Sijilmasa in mid-February 1352, reached the Mali capital on June 28, and left it on February 27, 1353. He went on to Timbuktu, sailed down the Niger to Gao, then joined a large caravan of merchants from Ghadamas for Takadda.² He set out north from Takadda on September 11, 1353, with seventy days’ food, because no grain could be bought between there and Tawat, only meat, milk and butter paid for in cloth.⁵ He reached Sijilmasa in December and left for Fez on December 29, through heavy snow.⁵

Passage depended on local powers. On the road to Takadda the caravan entered the lands of the Bardama, and Ibn Battuta says no caravan crossed there without a guarantee of their protection, a woman’s guarantee counting for more than a man’s. In the country of the Hoggar (Gibb’s Haggar), a Berber people of the central Saharan mountains, a chief stopped the caravan until it paid in cloth and other goods. Because the caravan arrived during Ramadan, he adds, it was not raided, as the Berbers of that route did not attack caravans during the fast.⁵ Such payments were part of the cost of the journey, much like the tolls al-Bakri recorded at Ghana.⁴

The caravan as these sources describe it was a temporary business partnership. People who had bought animals and fodder traveled on a schedule set by water, paid specialists for guidance and protection, sent letters ahead to arrange lodging, and were taxed when they arrived. Al-Bakri’s description of Zawila sums up the geography: “Caravans meet there from all directions and from there the ways of those setting out radiate.”³

Footnotes

1. Berzock et al., Caravans of Gold, Fragments in Time (2019), “Saharan Frontiers” and sections. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹

2. Ibn Battuta, trans. Gibb (1929), pp. 317-339, Fordham transcription. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³

3. Lane, introduction, and al-Bakri on Zawila, in Levtzion and Hopkins, Corpus, pp. 63-64, via medievalslavery.org. ↩ ↩² ↩³ ↩⁴

4. Al-Bakri and Ibn Hawqal in Levtzion and Hopkins, Corpus (1981), pp. 79-81 and related excerpts, via Boston University African Studies Center. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶

5. Ibn Battuta, trans. Gibb (1929), pp. 333-339 and notes, Internet Archive OCR. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰

6. Faye 2022, Animal Frontiers 12(4), pp. 8-10. ↩

7. Nixon, Rehren and Guerra 2011, Antiquity 85, abstract. ↩