Bu Njem Ostraca

Gholaia: Roman fort, part of the Limes Tripolitanus, modern Bu Njem.

The nomads have arrived, bringing four asses and two Egyptians carrying letters to you and to Gtasazeihemus Opter, and a runaway slave.

It is just a notice, scribbled on a potsherd from Bu Njem, an oasis in Libya once occupied by Roman legionaries, who called it Gholaia. Archaeologists have found almost 150 of these reports. Although the texts have not profoundly changed our knowledge of ancient warfare, they can still offer quite interesting information about ordinary life in a not so ordinary fort.

The Limes Tripolitanus

Sabratha, Oea, and Lepcis Magna were three large towns in the northwest of modern Libya, which was called Tripolitana, “land of the three cities”. The cities were famous for producing olives. When Julius Caesar had to fine Lepcis, he demanded no less than three million pounds of oil. One of the country’s assets was the presence of cheap laborers: the Garamantes, nomads who had their winter pastures in the oases in the deep south, came to the north during the harvest season. Leaving their goats on the nearby fields, the Garamantes helped to gather the olives and used their dromedaries to plow. They also bartered meat, dairy products, and products from sub-Saharan Africa with ceramics and objects made of metal, until they returned home with their flocks when their jobs were done.

The people of the cities and the nomads needed each other, but tensions were inevitable. As the cities grew, there were more peasants, who needed more land, and this meant that it became increasingly difficult for the Garamantes to come to the north with their cattle – there was less pasture. If the nomads became restless and attacked Roman settlers, the army intervened, and after a brief retaliatory campaign, a new balance was found.

It was hard to prevent these small wars, which appear to have taken place frequently. Of course the Romans could build walls to keep the nomads out, but that would also prevent seasonal migration. Therefore, the Roman emperor Septimius Severus (r.193-211), a native of Lepcis Magna, adopted another policy: he built three forts at the main oases, which meant that access to water could be refused to aggressive nomads. They could only reach Roman territories by carrying bags of water on the dromedaries, which would slow their progress and make them vulnerable to counterattacks.

However, Severus’ solution created a new problem. In Bu Njem, one cohort of the Third Legion Augusta (480 men) and one unit of mounted auxiliaries were stationed (about 120 men with horses/dromedaries), and there was a large village close to the fort. If we add 500 villagers, we need about 900,000 kilograms of grain (or the equivalent) per year, which is more than the peasants near the oasis can have produced. Although it was possible to import food from the cities, the Romans opted for another solution: they decided to develop the semi-arid zone between Bu Njem and the next fort, Gheriat el‑Garbia. Along the wadis, fortified farms were built, which were probably handed out to army veterans. Direct evidence is lacking, but the settlers must have received government support: much capital had to be invested to build dams, dikes, cisterns, and fortified farms, and without some starting capital, it would not have been profitable to develop the zone.

The forts of Tripolitana were different from military frontier settlements elsewhere. Hadrian’s Wall, the Rhine, the palisade in the Black Forest, and the Danube were clear, recognizable boundaries “to separate the barbarians from the Romans,” as the author of the Historia Augusta describes the purpose of the limes. In Tripolitana, this was impossible: the Roman farmers and the nomads needed each other. The Limes Tripolitanus was more open than Rome’s northern frontier, and the army was especially occupied with checking the seasonal movements of the Libyan tribes and offering protection to nearby peasants. Gheriat el‑Garbia and Bu Njem were, essentially, police stations.

The Bu Njem Ostraca

Papyrus and parchment were expensive. For simple messages, the ancients used wooden tablets or potsherds (ostraca). Now ceramics can break but never perish, so there is a fair chance that we can find messages, and in fact, there is almost no part of the Mediterranean world from which written sherds are absent. The material on which the memos from Bu Njem have been written, is not exceptional, and neither is the fact that they were written by soldiers: there are similar collections from Dura Europos on the Euphrates and Vindolanda, one of the forts of Hadrian’s Wall.

Still, a collection of 146 documents, belonging to one and the same very brief period (the years 253-259), is sufficiently rare to be called an important find. Perhaps the most surprising conclusion was that, during the confused “crisis of the third century”, this was an orderly frontier zone. There are no indications of unrest, although during these years (the reign of Valerian), the empire was under attack from almost every side.

Nearly all ostraca were found in three buildings. 136 documents were excavated inside the headquarters (principia) and in a depot close to its southern entrance. Four pieces were discovered in the commander’s office (praetorium), two sherds are from the bathhouse, and the remaining four pieces were found elsewhere. The absence of written texts from the barracks confirms a statement by the Roman military writer Vegetius: only a couple of recruits had to be literate.note And indeed, all documents composed in Bu Njem were written by just four men. In the official publication, they are divided into six categories.

  1. 62 daily reports (#1-#62)
  2. 4 lists (#63-#66)
  3. 7 accounts (#67-#73)
  4. 44 letters received (#74-#117)
  5. 28 other documents (#118-#145)
  6. 1 document in Punic (#146)

The texts are not written in classical Latin. For example, the m at the end of a word is often dropped, and words that ought to be written with ae are often spelled with e, (e.g., Emilius for Aemilius). This is a normal development, also attested in other parts of the Roman Empire. On the other hand, the shift of f to u (Uabius instead of Fabius) is less common and may betray the Punic background of the soldiers.

The People of Bu Njem

The fort offered accommodation to one cohort from the Third legion Augusta, to which mounted auxiliaries were added. The name of their unit is not recorded: it is simply called numerus collatus, “an added unit”. From an inscription, we know of the presence of cavalry from the Cohors VIII Fida Equitata, “the eighth mounted, loyal cohort”, but they are not mentioned in the ostraca and appear to have left before 263. Another mystery is the exact rank and task of the men. Many of them are called quinanari, but we have no idea what that means: it may refer to soldiers living “in the fifth block” or to those who do their exercise “on the fifth hour”.

We have more certainty about the soldiers’ self-image: they were proudly Roman, as we can deduce from their personal names: Aurelius Felix, Julius Optatus, Paccius Maximus. A small minority combines Punic and Latin names, like a centurion named Marcus Porcius Iasuchan and the commander mentioned in the introduction to this article, Gtasazeihemus Opter. Outside the fort, these double names were the large majority: among the inhabitants of nearby Ghirza (one of the new towns along the wadis) is a family with the Roman name Marchius, which the bearers combined with native names like Nasif, Chullum, and Maccurasan. Mixed names were the rule, and the fully Roman names of the soldiers betray their attempt to present themselves as real Romans.

This can also be deduced from their religious beliefs. The only cult that is mentioned on the ostraca is that of the signa, the standards. Every day, one soldier guarded the little shrine in which they were kept. (The same practice is mentioned in the papyri from Dura Europos, where an officer and eight or nine soldiers guarded the standards.) This religious exclusivity comes as a surprise, because we know from inscriptions that the soldiers built a temple for Ammon in the oasis, and also venerated Cannaphar, a Libyan war god. However, within the walls of the fort, they presented themselves as soldiers of Rome.

We already noticed that the settlers along the wadis produced grain and oil for the soldiers in the forts. The army had not organized the transport of these products. On the ostraca, we encounter several dromedary-drivers, none of them belonging to the armed forces. Their names are interesting: although they must have had mixed names, like the Marchii from Ghirza, the scribes of Bu Njem recorded only the native part (e.g., Iddibal, Iassuchtan, Iaremaban), as if to stress that these merchants were less Roman than the soldiers.


The ostraca record many tasks and details. For example, there were at least four types of job to be done to keep the bathhouse functioning: someone was occupied with cleaning the building (ad balneu), several men were occupied with bringing five cubic meters of water to the two basins (ad aqua balnei), one soldier worked at the furnace (ad furnum), and one man was sent out to obtain wood (missus ad lignum) – a rare commodity that had to be brought from the coast. Seen from the outside, the inhabitants of the fort lived in incredible luxury.

Every day, the commander learned how many people were sick: no less than 41 of the 62 daily reports make reference to the health of the soldiers. Leaving out two sherds that refer to an epidemic, we can establish that on average, about 5% of the soldiers were ill, which is a normal figure for soldiers in the desert. Two legionaries, Caecilius Rogatus and Aurelius Celestinus, appear to have been chronically ill, but the army continued to take care of them.

Other references show that soldiers were sent to outposts like wells and cisterns, or guarded the station of the dromedary caravans. Horsemen and dromedary-riders were patrolling the area, and there are several references to captives: probably seasonal laborers who had not obtained a letter (or sign) of passage. These undesirable aliens may have been expelled, or detained until the man who wanted to hire them arrived to confirm that he needed laborers.

But perhaps the most intriguing activity of the soldiers at Bu Njem was to receive runaway slaves, like the fugitive mentioned at the opening of this article. Apparently, there was some sort of understanding with the Garamantes that they would send back people from the Roman Empire who had decided to improve their fortunes elsewhere. Rome’s southern frontier was not a clearly drawn line, but an open zone that did not impede people’s movements. However, on both sides, people knew who was a Roman and who was a Libyan tribesman. Essentially, the Limes Tripolitanus was not a territorial, but a social border.

Document 5 (another report):

XVI KAL(endas) [I]VL(ias) N(vmervs)


TIRONES N(vmervs)


IN HIS EQ(vites)










AD ST(ationem) CAMELLAR(iorvm)
















Translation of Ostracon 5:

16 June, number


Number of recruits


of which cavalry




To the bathhouse


From the shows


To the commander


To the dromedaries' station




Id., Rufus


On sick leave




Other details


The bathhouse received





This article was originally published in Ancient Warfare 1/2 (2007). The official publication is Robert Marichal, Les Ostraca de Bu Njem (1992 Tripoli = Libya Antiqua, suppl.9).