The Great Arab Conquests

Although the Persians had been decisively defeated, the conquest of Syria and Palestine had a very damaging effect on Byzantine power in the Levant. Apart from the bloodshed caused by the warfare, it seems that many of the Greek-speaking elite emigrated to the security of North Africa or Rome.¹ The fighting had been very destructive, especially in the towns, but perhaps more important was the loss of the tradition of imperial rule and administration. For most of the period of Muhammad’s mission, Syria and Palestine were ruled by the Persians, not the Byzantines, and it was not until 630, a couple of years before the Prophet’s death, that Byzantine control was re-established. Nonetheless, this control must have been very patchy, and there were probably many areas where Byzantine government hardly existed. Most younger-generation Syrians would have had no experience or memory of imperial rule, and no cause to be loyal to Constantinople. Even as Byzantine government was being slowly re-established, the religious differences that had divided Syria in the sixth century came to the fore again. The emperor Heraclius was determined to enforce religious conformity on a Christian population that in large measure rejected his doctrinal position.

Byzantine control over Syria had been established for more than half a millennium. If Islam had been born fifty years earlier, and the early Muslims had attempted to raid Syria and Palestine in the 580s not the 630s, there can be little doubt that they would have been seen off very quickly, as the provinces were firmly controlled by the government and the defences well organized. The coincidence that the first Muslim armies appeared in the area immediately after the traumatic events of the great war between Byzantium and Iran was the essential prerequisite for the success of Muslim arms.

The country was familiar territory to the leaders of the early Islamic community and it was natural that it would be the first objective of the new Muslim armies. The tradition that the Prophet himself visited Syria before the start of his mission is ancient and well attested. A Syrian city, Jerusalem had been the first focus of prayer for the earliest Muslims, before the adoption of Mecca. Abū Sufyān, leader of the Meccan opposition to Muhammad, owned property in Jordan, including the village of Qubbash, in the fertile Balqā district south of Amman, which he used as a base for his trading activities.² The towns of Syria were the entrepôts along the desert margin and many members of the new Islamic elite had visited the country and knew it well. When Muhammad, at the end of his life, was looking for areas to provide new resources for the Muslims, it was natural that he looked north. Syria was very different in this respect from Iraq, which few of the new Islamic elite had visited before the conquests began and which was essentially unfamiliar territory.

The Muslim attacks on Syria had begun on a small and not very successful scale in the last two years of the Prophet’s life. Visitors to Jordan, travelling south along the ‘King’s Highway’, the ancient route that runs along the fertile ridges to the east of the Dead Sea, from Karak to Petra, are shown the tombs of the early Muslim heroes just south of the village of Mu’ta. The tombs, with their neat domes and groves of trees, are quite modern, but their position seems to be a genuine relic of the first encounter between Muslims and Byzantine forces. In 629 Muhammad had sent a raiding party in the direction of Syria, probably just looking for booty in the turmoil that followed the withdrawal of the Persian army. As the small band of Muslims rode north up the King’s Highway, they were met by a detachment of Byzantine soldiers, mostly local Arab tribesmen, marching south down the road to re-establish Byzantine rule in the area. In a short clash at Mu’ta, the Muslims were defeated and forced to flee, several of the leaders being killed and buried in the tombs we still see today. Among the Muslims who fled to fight another day was Khālid b. al-Walīd, the ‘Sword of God’ who was later to play such an important role in the conquest of Syria.

The defeat at Mu’ta was a humiliation for the nascent Muslim state but Muhammad seems to have been undeterred and was still determined to pursue the project of raiding Syria. In 630 he sent a carefully planned expedition against Tabūk in the northern Hijaz which may have been a trial run for attacks on Syria. Among the commanders who gained useful military experience there was Amr b. al-Ās, the man who was to conquer Egypt for the Muslims a decade later. There can be no doubt that when the early Muslim high command embarked on the conquest of Syria, they were pursuing a policy already begun by their Prophet.

Immediately after Muhammad’s death, the caliph Abū Bakr sent another expedition to Syria, an expedition that marked the beginning of the real conquest of the country. The sequence of events becomes extremely confused at this point. We have a vast mass of traditions about major battles and minor engagements and about the capture of cities. But the truth is that there is no way of reconciling the different chronological schemes that were elaborated by different Muslim editors, and there are very few external sources to give us any sort of guidance. As the great Muslim historian Tabarī complained when he was collecting the conquest narratives, ‘in fact, one of the most annoying things about this study is the occurrence of such differences as the one I have noted above about the date of this battle. Such differences arose because some of these battles were so close together in time’.³ In the end, we can only be certain that campaigning began in earnest from 632 and that eight years later, in 640, all of Syria was under some sort of Muslim rule with the exception of the coastal city of Caesarea. The account that follows is based on the most generally accepted chronology, but it should be treated with considerable caution.

The objective of these early expeditions was to assert the control of Medina over the Arab tribes on the fringes of the settled land. On the western borders of the fertile land of Iraq and along the edges of the Nile valley in Egypt, the border between the desert and the sown is a comparatively firm line between one ecological zone and another. In Syria the distinction is much less clear cut. Moving east from the well-watered Mediterranean coast, the landscape becomes gradually more arid. At the line of the 200mm isohyet (the line beyond which there is less than 200mm annual average rainfall) settled agriculture is impossible without oasis irrigation. West of the line is a zone that can be used as pasture by the Bedouin or by dry farming. Many Bedouin have also been part-time farmers, cultivating small fields of grain as well as pasturing their animals. The policy of securing the allegiance of the Syrian Bedouin to Islam led the Muslims inexorably into conflict with the Byzantine imperial authorities and their Arab allies. It was a very conscious and deliberate policy move by the caliph Abū Bakr and the rest of the Muslim leadership: all nomad Arabs were to pledge their allegiance to the Muslim state and those who did not do it voluntarily were to be coerced.

Abū Bakr is said to have dispatched four small armies to operate independently in the frontier zones to the east of the Dead Sea and the Jordan valley, attaching banners to the spears of the leaders as a sign of authority. His choice of commanders was to be very important in the history of the early Islamic state. One of them was Yazīd, the son of Abū Sufyān, who took with him his brother, Mu^cāwiya. As we have seen, the family already had properties in Syria and knew the area well. Yazīd was to be one of the leading Muslim commanders in the conquest, and this enabled him and his brother to establish the power of their family in Syria. Yazīd died of the plague before the conquests were finally complete, but his brother Mu^cāwiya inherited his role. The power base he built up in Syria during and immediately after the conquests enabled him to establish himself as the first Umayyad caliph in 661 and rule the entire Muslim world from Damascus.

Another appointment with long-term consequences was that of Amr b. al-Ās, shrewd and cunning rather than a great warrior, the wily Odysseus of the early Islamic armies. His background as a merchant trading in Gaza had recommended him to the Prophet, who had chosen him to collect taxes from the tribes on the road from Medina to Syria. He chose to lead his men, said to have been about three thousand in number, many from Mecca and Medina,⁴ to the area with which he was already familiar. He travelled along the Red Sea coast as far as the head of the Gulf of Aqaba then turned west, camping with his men in the great sandy depression between Jordan and Israel known as the Wadi Araba. From there they climbed up the escarpment to the plateau of the Negev before heading for the sea at Gaza. Here Amr began negotiations with the local military commander, probably demanding money, and there is a tradition that the Byzantine governor attempted to capture or murder him as they were parleying. Finally, on 4 February 634,⁵ there was a battle in which Amr and his men defeated the small Byzantine army at a village called Dāthin, near Gaza, and killed its commander. The Arab victory made an immediate impression. News travelled fast, and we are told that a Jewish community near Caesarea openly rejoiced at the death of a Byzantine official and the humiliation of the imperial authority.⁶

The Muslim victory at Dāthin may have been on a fairly small scale but it alerted the Byzantine authorities to the new threat from the south. Overall command lay with the emperor Heraclius. He was around 60 years old at this time and was certainly no pampered denizen of the vast and luxurious palaces of Constantinople; rather, he was a man with a vast amount of military experience, well used to the hardships of campaign. He was also at the height of his powers and, even as the earliest Muslim raids on Syria began, had just celebrated a major triumph with the return of the True Cross to Jerusalem. Heraclius never led his armies against the Muslims in person (but neither did the Muslim caliphs lead the armies of Islam) but he remained behind the lines in Syria, in Homs or Antioch, directing operations, appointing generals and issuing instructions. The portrayal of Heraclius in the Arabic sources is very interesting.⁷ He is renowned for his shrewdness and wisdom and his ability to foresee the future. In one story, Abū Sufyān, the Meccan aristocrat, tells how he saw Heraclius when he was visiting Syria with a group of merchants. ‘We arrived there when Heraclius had just defeated the Persians and driven them out of his territory recapturing from them the great cross, which the Persians had stolen . . . Heraclius then left Homs, which was his headquarters and walked on foot . . . in order to pray in the Holy City. Carpets were spread for him and aromatic herbs were thrown on the carpets. When he reached Jerusalem, Heraclius prayed together with the Byzantine nobles.’⁸ He is shown here victorious but modest and pious.

In a number of anecdotes, Heraclius is said to have recognized the greatness of Muhammad and would have become a Muslim if the Byzantine nobles had not been so hostile to the idea. To the Arabs, he was the key, symbolic leader of the Byzantine resistance to the armies of Islam, the ancient enemy. He is shown to be proud and autocratic but he also goes through moments when he alone of his advisers and courtiers can see how strong the Muslims are and recognizes that they are bound to prevail. The image the Arab sources give of Heraclius is not entirely unsympathetic: he is a tragic figure whose failure to embrace Islam meant that his career ended in humiliation and failure.

Up to this point, the Muslim attacks on Syria had amounted to little more than pinpricks along the frontier. The next phase of the conquest began with the arrival of Khālid b. al-Walīd and his men after the forced march across the desert from Iraq, where he had been raiding along the desert frontier. Khālid’s march across the Syrian desert, with perhaps five hundred of his troops, has been enshrined in history and legend:⁹ Arab sources marvelled at his endurance; modern scholars have seen him as a master of strategy.¹⁰ The story is often told of how he crossed six waterless days of desert by making some of his camels drink more than their fill, binding their jaws so that they could not chew their cud, then slaughtering them one by one so his men could drink the water from their stomachs. At another stage, when Khālid and his men were stumbling along, suffering from extreme thirst, he asked one of his men, Rāfi, who had been in the area before, whether he had any idea about water. Rāfi said that there was water near at hand: ‘Go on and look for two hummocks which look like two women’s breasts and then go to them.’ When they arrived he told them to search for a thorn bush like a man’s buttocks. They scrabbled around and found a root but no sign of a tree, but Rāfi told them that this was the place and they should dig there. Soon they uncovered damp ground and small quantities of sweet water. Rāfi, greatly relieved by the discovery, said to Khālid, ‘O Commander, by God, I have not come to this waterhole for thirty years. I have only been once before when I was a boy with my father.’¹¹ So, the account goes on, they prepared themselves and attacked the enemy, who could not believe that any army could cross the desert to them.

The trouble is that the accounts of this expedition, though vivid, are very confused. We can be certain that Khālid did cross the desert from Iraq to Syria some time in the spring or early summer of 634, that it was a memorable feat of military endurance and that his arrival in Syria was an important ingredient of the success of Muslim arms there. The problem is that some sources suggest that he went on the long southern route by Dūmat al-Jandal, while others are equally certain that he made the journey via Palmyra to the north. There are good arguments on each side and simply no knowing which version is correct.

The Arabic narratives give pride of place to Khālid as the commander who provided the most effective leadership, even after Umar had dismissed him from supreme command and replaced him with Abū Ubayda. It was Khālid who united the different Muslim armies on his arrival, it was Khālid who began the conquest of Damascus by opening the East Gate, and it was Khālid who devised the tactics that won the battle of Yarmūk. He then went on to take a leading role in the conquest of Homs and Chalkis (Ar. Qinnasrīn). His reputation as a great general has lasted through the generations and streets are named after him all over the Arab world. Despite his undoubted achievements, however, his reputation in the sources is mixed. He came from one of the most aristocratic families in Mecca and like many of people of his class he had been deeply suspicious of Muhammad with his preaching of social justice and simple monotheism. He had not been one of the early converts to Islam; indeed, he had been among the enemies of the Prophet, actually fighting against him at the battle of Uhud, but he converted to Islam soon after. Once converted he become staunchly Muslim and began to devote all his considerable military talents to the support of the new Muslim state. On Muhammad’s orders, he destroyed one of the most famous of the old idols, the image of the goddess al-Uzza at Nakhla near Mecca. He enjoyed the confidence of the first caliph Abū Bakr and was entrusted with commanding the armies against the rebel Arab tribes in the ridda wars. He won great victories but also gained a reputation for ruthless and sometimes over-hasty reactions: on one occasion he massacred a whole group of Muslims by mistake and compounded the offence by immediately marrying the widow of one of his victims.¹² His later fame seems to have rankled with some early Muslims, notably the caliph Umar, who strongly believed that early commitment to Islam was essential for anyone who wished to be a leader, that late conversion did not suffice, and that a little humility would not go amiss. A story told of Khālid attempts to explain his life and rehabilitate him. In a dialogue with the Armenian general Jurjah immediately before the battle of Yarmūk, Khālid is made to justify his career and explain why he was popularly called the ‘Sword of God’.

God sent us his Prophet, who summoned us, but we avoided him and kept well away from him. Then some of us believed him and followed him, whereas others distanced themselves from him and called him a liar. I was among those who called him a liar, shunned him and fought him. Then God gripped our hearts and our forelocks, guiding us to him so that he followed him. The Prophet said to me ‘You are a sword among the swords of God which God has drawn against the polytheists’, and he prayed for victory for me. Thus I was named the Sword of God because I am now the most hostile of Muslims to the polytheists.

Khālid had been instructed by Abū Bakr to march as quickly as possible to aid the conquest of Syria, which had now reached a critical state. On Easter Day, 634 (24 April), he and his forces suddenly appeared and fell upon the Ghassānid Christian allies of the Byzantines, who were celebrating the festival amid the lush grass and spring flowers of the Meadow of Rāhit just north of Damascus.¹³ He then turned south to join up with the other Muslim commanders already operating in Syria, who now seem to have been united under his command to face the challenge posed by the Byzantine imperial armies. They began with an attack on the city of Bostra.¹⁴

Bostra lies just north of the modern Syrian-Jordanian border in a flat but fertile landscape strewn with the black basalt boulders characteristic of much of the area. To the north of the city, and clearly visible from its walls, rise the volcanic hills of the Hawrān. Though the mountains are rugged, if not especially high, they contain, like many volcanic areas, patches of extremely fertile soil. The hinterland of Bostra was the closest area to Arabia which could supply the wheat, oil and wine the Bedouin desired. The city had become rich as a trading entrepôt, and it was widely believed that the Prophet himself had visited it in his youth and had been instructed in the mysteries of the Christian faith there by the monk Bahira. Bostra was also a political centre. When the Roman emperor Trajan had annexed the Nabataean kingdom in 106 and turned it into the Roman province of Arabia, he had moved the capital from distant Petra in the south to the more accessible (accessible from Rome, that is) city of Bostra. Built of tough, unyielding black basalt, the ruins of the ancient city of Bostra are among the most impressive in the Near East. The huge Roman theatre there still survives almost intact, forming the centre of a later medieval fortress. Columns and paving stones indicate the routes of ancient streets, and there are the remains of baths and a number of important Christian churches, including a magnificent round cathedral.

After the surrender of Bostra, the Muslim force marched west to meet up with Amr b. al-Ās. Amr, after his first victory at Dāthin, was now confronted by a large Byzantine force which had gathered south-west of Jerusalem on the road to Gaza. Khālid and the others crossed the Jordan valley without apparently encountering any resistance and met up with Amr and his men. The combined Muslim army is said in one source to have been about twenty thousand strong and was under the command of Amr, who is the only Arab general named in the sources, where his image is consistently one of shrewdness and intelligence. He is described spying out the enemy camp in person or sending agents to do so, while the Byzantine general writes to him as someone equal to himself in cunning.¹⁵ The armies met at a place the Muslim authors called Ajnādayn, and a major battle developed. We have no detailed information about the nature of the conflict but it is clear that the Byzantines were defeated and that the remnants of their army withdrew to Jerusalem and other fortified sites. News of the victory of the Muslims spread far and wide, and it seems to be the battle referred to in the Frankish chronicle of Fredegar composed some twenty years later in France. He includes the interesting, and possibly true, detail that ‘the Saracens’ (the Muslims) offered to sell back to Heraclius the booty they had just taken from his defeated men but that the emperor refused to pay for any of these stolen goods.¹⁶

The contemporary Armenian chronicler Sebeos tells how the Byzantine forces were ordered by the emperor to remain on the defensive.¹⁷ Instead, they left their camps by the river and took refuge in the city of Pella, on the east bank of the river. Pella was a prosperous city in the fertile lands of the Jordan valley and an easily defended acropolis rose above the classical streets and porticoes on the valley floor. Here they were attacked again. As usual, the course of battle is not entirely clear but some features seem to have been remembered. The Byzantine troops had crossed the Jordan valley from Scythopolis on the west bank and, in order to delay the pursuing Muslims, had cut some of the irrigation ditches, allowing the water to spill out and the flat lands of the valley bottom to become an ocean of mud.¹⁸ The Muslims charged on, not knowing what the Byzantines had done, and many of their horses became stuck in the mire, ‘but then God delivered them’. In the end, it was the Byzantines who were trapped in the mire and many were massacred.

The remnants of the Byzantine forces now withdrew to Damascus. The Muslims pursued them. The siege of Damascus became one of the set pieces of the conquest of Syria. To a remarkable extent we can retrace the progress of the siege because of the detailed descriptions of the sources and the preservation of the fabric of the city. The walls of old Damascus, Roman or earlier in origin and continuously restored since, are still largely intact. Only at the western end where the city expanded in Ottoman times is the old circuit breached. All except one of the ancient gates survive and they bear the same names today as they do in the early Arabic sources: it is an astonishing example of the continuity of urban geography and architecture through almost fourteen centuries. We are told that Khālid b. al-Walīd was stationed at the East Gate (Bāb Sharqī), Amr b. al-Ās at St Thomas’s Gate (Bāb Tūma), Abū Ubayda at the now demolished Jābiya Gate on the west side and Yazīd b. Abī Sufyān at the Little Gate and Kaysān Gate on the south side.

The Muslims also took the precaution of stationing a force on the road north of Damascus. This proved a wise move because Heraclius, who is said to have been in Homs at this time, sent a force of cavalry to try to relieve the siege but they were intercepted and never made it.¹⁹ Just how long the siege lasted is not clear. Disconcertingly, the Arabic sources give widely differing estimates, anything from four to fourteen months. The Muslims do not seem to have had any siege engines, or any equipment more sophisticated than ropes and ladders, and even the ladders had to be borrowed from a neighbouring monastery. ²⁰ It seems that all the attackers could do against the substantial Roman walls of the city was to mount a blockade and hope that famine, boredom or internal disputes would cause the defenders to give up. When it became clear that no relieving force was going to appear, the defenders of the city began to despair. According to one account, the end came when a child was born to the patrikios (Byzantine commander) in charge and he allowed his men to relax and eat and drink to celebrate. Khālid b. al-Walīd, who was always on the lookout for opportunities and knew exactly what was going on in the city, decided to take advantage. He had ropes and ladders with him. Some of his men approached the gate using inflated animal skins to cross the moat. They threw their ropes around the battlements and hauled themselves up, bringing the ropes up after them so that they would not be seen. Then, at a given signal, with the cry of ‘Allhu akbar’ (God is great) they rushed the gate, killing the gatekeepers and anyone else who resisted.

Meanwhile, at the other end of the town, the Damascenes had begun opening negotiations for a peaceful surrender and Muslim troops began to enter the city from the west. The two groups, Khālid’s men from the east and the others from the west, met in the city centre in the old markets and began to negotiate. Terms were made, leaving the inhabitants in peace in exchange for tribute. Properties belonging to the imperial fisc were confiscated for the benefit of all Muslims, becoming part of the fay (the communal wealth of the Muslim community).²¹ As usual there was booty to be divided up and the commanders were careful to keep a share for those who had been stationed on the road north, for though they had played no direct part in the siege their presence had contributed to the victory and they had earned their booty. The complicated stories which evolved of the taking of Damascus, from two different ends in two different ways, may be an attempt to solve the thorny issue of whether the city was taken by force or by treaty (see above, pp. 18-20). In this case the authorities seem to have tried to reach a compromise that allowed it to be neither one thing nor the other.

The accounts of the fall of Damascus also reflect divided loyalties among the population. The city was a centre of imperial power with a military governor appointed by the emperor himself, but many if not most of the inhabitants were Christian Arabs. It is evident that many of them had split allegiances and that they felt closer to the Arabs outside the walls than they did to the Greeks and Armenians who composed a large part of the garrison. Whatever the explanation, it is clear that Damascus was spared the horrors of bombardment and sack. In the century that followed, the city became the capital of the whole Muslim world and entered what came to be its golden age.

Around the time of the fall of Damascus, and as usual the chronology is very uncertain here, the elderly Abū Bakr, the successor of Muhammad and first caliph of Islam, died in Medina. We know that his death occurred in July 634. What is less clear is what stage this was in the story of the conquest, but there are a number of reports that news reached the Muslim armies in Syria during the course of the siege. The new caliph was the austere and formidable Umar, who is portrayed in many accounts as the mastermind behind the conquests. There was no opposition to his succession among the forces in Syria but the new caliph had clear ideas about command. As we have seen, Umar disliked Khālid b. al-Walīd intensely. The fact that Khālid had fought so brilliantly for the Muslim cause against the ridda in eastern Arabia and again in Iraq and Syria did little to improve his standing with the new caliph. He now abruptly ordered that Khālid be removed from command and return to Medina. In one account, Abū Ubayda, now appointed as supreme commander in Khālid’s place, was ordered to demand that Khālid should confess to being a liar. If he refused, as he was bound to do, his turban should be pulled off his head and half his property confiscated. Faced with this ultimatum the great general asked for time to consult, not as might be imagined with his friends or subordinates, but with his sister. She was clear that Umar hated her brother and if he admitted to being a liar he would be removed all the same. There was no point in trying to placate the caliph by admitting to crimes he did not believe he had committed.

In an interesting reflection on the power of the caliph and the unity of the Muslims, Khālid felt that he had no choice but to go to Medina. A Byzantine general in that position might well have raised a rebellion and called on his troops to support him in a bid for the throne. By contrast, the greatest general of the Muslim army meekly accepted dismissal and humiliation. When he arrived at Medina, Umar pursued his vendetta. Whenever he met Khālid he would taunt him: ‘Khālid, take the property of the Muslims out from under your arse!’, to which Khālid would meekly reply that he did not have any of the ‘Muslims’ property’. In the end a settlement was reached, with Khālid paying over most of his fortune so that he was left only with military equipment (^cuddat) and slaves (raqīq). He was soon back in Syria, playing a major role in the battle of the Yarmūk and the subsequent conquests of Homs and Chalkis, where he finally settled. In the end Umar is said to have recognized that he had maligned the ‘Sword of God’ and that Abū Bakr, who supported Khālid, had been a better judge of men.²² The great general died peacefully in 642, a brilliant, ruthless military commander, but one with whom the more pious Muslims could never feel entirely comfortable.

Meanwhile, the emperor Heraclius was preparing one more major effort to drive the Muslim invaders out of Syria. After the fall of Damascus he had retreated to Antioch in the north of Syria, the traditional capital of the entire area. Here he set about directing what was to prove his last campaign. The Byzantines assembled all the troops they could recruit. Arab sources give very large numbers, over 100,000,²³ but comparisons with other Byzantine armies of the period make it clear that this is a huge exaggeration, with numbers between 15,000 and 20,000 being more probable. The armies comprised a very diverse collection of men. There were Byzantine Greeks under the command of Theodore Trithurios, a large contingent of Armenians under Jurjah and the local Christian Arabs led by the king of the Ghassānids, traditional allies of the Byzantines, Jabala b. Ayham. The overall commander was an Armenian called Vahān. The different contingents would have spoken different languages - Greek, Armenian and Arabic - and they may have found it difficult to communicate with each other. There were also profound religious and cultural differences. The Greeks and Armenians would have come from settled, probably rural village backgrounds and were used to living and fighting in upland, mountain terrain. The Arabs, on the other hand, were nomads, used to the mobile traditions of desert warfare. All the troops came from Christian backgrounds but both Armenians and Christian Arabs were regarded as heretics by the orthodox Byzantines. How far these divisions really affected the performance of the Byzantine army is not clear, but the sources are awash with rumours of disaffection, of Jurjah converting to Islam at the hand of Khālid b. al-Walīd on the eve of battle and of the Christian Arabs going over to the Muslim side in the course of battle. The Arab sources also talk of the Byzantine soldiers being chained together so that they could not flee, but this is a story found in many accounts of the conquests, used to contrast the free and motivated Muslims with the serf-like soldiers of their enemies: there is no real evidence for such an impractical idea being put into effect, though it may be a distant reflection of the practice of infantry locking shields together to make a protective wall.²⁴

The Byzantine forces probably assembled in Homs and marched south through the Biqa valley, past Ba^calbak with its great pagan temples - now almost empty of worshippers but still magnificent in their decay - and so to Damascus. In anticipation of the arrival of this force, the Arabs seem to have withdrawn from the city, allowing Byzantine forces to reoccupy it unopposed. We have no information on how they found the town but there are reports of tension between the Byzantine generals demanding supplies for their men, as was the usual Byzantine practice, and the local financial administrator, the Arab Mansūr, who maintained that the city did not have sufficient resources to feed them. Certainly the army did not use Damascus as a base but moved on south.

The Byzantine army assembled at Jābiya in the Golan Heights. This was the traditional summer pasture of the Ghassānids. According to the most probable reconstruction, it was now August 636 and the Golan would have provided much-needed food, water and pasture for the army. Meanwhile the Muslim forces prepared to oppose the Byzantines and hold on to their newly won gains. Their army also assembled in the Golan area, to the south-east of the Byzantines. The different Muslim armies had now come together under the command of Abū Ubayda, or possibly Khālid b. al-Walīd. Yazīd b. Abī Sufyān and Amr b. al-Ās both led contingents. According to Muslim sources, the Arab army numbered about 24,000. In view of the downward revision of the numbers on the Byzantine side, it is possible that the two armies were not very different in size.

The battle that ensued between the Christian and Muslim armies is generally known as the battle of Yarmūk and conventionally dated to the summer of 636.²⁵ The battle of Yarmūk is, along with the battle of Qādisiya in Iraq, one of the major conflicts that has come to symbolize the Muslim victories in the Fertile Crescent. As with Qādisiya, the Arab accounts are extensive and confused and it is difficult to be clear about exactly what happened. There is no contemporary or reliable account from the Byzantine point of view. Both sides are said by the Muslim sources to have been inspired by religious zeal. As the Byzantines remained in their fortified camp, preparing for battle, ‘the priests, deacons and monks urged them on lamenting the fate of Christianity’.²⁶ On the other side, Khālid b. al-Walīd addressed his men: ‘This is one of God’s battles. There should be neither pride nor wrongdoing in it. Strive sincerely, seeking God in your work, for this day also has what lies beyond it [i.e. the afterlife]’, and he went on to urge them to stick together and fight in unison.²⁷

The River Yarmūk, a perennial watercourse, flows down from the plateau of the Hawrān to the Jordan valley, just south of the Sea of Galilee. In the course of its descent into the rift valley, it has gouged out a steep gorge, with high cliffs on each side. On the north side, it is joined by a number of smaller valleys, notably the Wadi al-Ruqqād. These steep ravines were to define the course of the battle and may have proved disastrous to the defeated when they attempted to flee from the scene. The actual site of the battle, between the Yarmūk gorge in the south and the Golan in the north, is a land of rolling, rocky hills, dotted with villages and farms. It was, in fact, good open country for cavalry manoeuvres, but it also provided some cover from rocks or trees for men to hide or set up an ambush. Since 1948 this site has been politically very sensitive, lying as it does on the border between Syria (north of the river), Jordan (south of the river) and the Israeli-occupied Golan. This has made access to the battlefield very difficult for historians. It was not always thus, however. Before the First World War, when the entire area was part of the Ottoman Empire, the battlefield was visited by the great Italian orientalist Leone Caetani, Prince of Sermoneta. He used his first-hand observations and knowledge of the Arabic sources to produce a geographical setting for the battle, which has formed the basis of the most plausible modern accounts.²⁸

The battle of Yarmūk was a series of conflicts that probably lasted more than a month and culminated in a major battle towards the end of August.²⁹ The first encounters took place in the Jābiya region, after which the Muslims retreated east towards Dar^ca. There followed a period of waiting and skirmishing as the Byzantines prepared their army and tried to sow divisions in the Muslim ranks. It seems that the real fighting began when the Muslims feigned a retreat from their positions and lured elements of the Byzantine army into rough terrain, where they were ambushed. During the Muslim counter-attack, the Byzantine cavalry became separated from the infantry, enabling the Muslim cavalry to inflict great slaughter on the foot soldiers while the cavalry were making their way through the Muslim ranks.³⁰ Khālid b. al-Walīd is said to have organized the Muslim cavalry in a ‘battle order which the Arabs had not used before’. He divided the cavalry into small squadrons (kards), between thirty-six and forty in number, apparently so that they would appear more numerous in the eyes of the enemy.³¹ The Byzantines may also have been unsettled by a dust storm. The main Byzantine force was now driven west and hemmed in between the rugged valleys of the Wadi’l-Ruqqād and Wadi’l- ^cAllān, with the cliffs of the Yarmūk gorge behind them. Any prospect of retreat to the west was destroyed when Khālid b. al-Walīd took the old Roman bridge across the Wadi al-Ruqqād, and Muslim forces went on to storm the Byzantine camp at Yāqūsa on the road to the Sea of Galilee. As the enemy pressed home their advantage, the Byzantine forces were further demoralized by rumours that the Christian Arabs had defected to the Muslims. Morale broke and the Byzantine forces lost all cohesion. There are reports of exhausted and dejected soldiers sitting down, wrapped in their mantles, lamenting the fact that they had not been able to defend Christianity and waiting for death.³² Others were driven down the cliffs into the wadis. The Muslims took very few prisoners.

The defeat on the Yarmūk was catastrophic for the Byzantines and news spread far and wide. In distant France, the author of the chronicle of Fredegar recorded it twenty years later as a terrible defeat. He has the Muslim army at 200,000 strong. According to him, the night before the battle ‘the army of Heraclius was smitten by the sword of the Lord: 52,000 of his men died where they slept’. Not surprisingly, the survivors were seriously disheartened. ‘When on the following day, at the moment of joining battle, his men saw that so large a part of their force had fallen to divine judgment, they no longer dared advance on the Saracens but all retired whence they came.’³³ Towards the end of the seventh century, the ascetic St Anastasius of Sinai in his remote monastery remembered it as ‘the first and fearful and incurable fall of the Roman army’.³⁴

In the aftermath of the victory, the Muslims continued to reduce the cities of Syria to obedience. One force, led by Abū Ubayda and Khālid b. al-Walīd, went north from Damascus to Homs, an important city in late Roman times.³⁵ They besieged the city through the winter (probably 636-7) despite the bitter cold and sorties by the Byzantine garrison. The defenders were convinced that the cold would force the Arabs, shod only in sandals, to give up the siege, though when spring came and they were still there, voices were raised in the city urging peace negotiations. According to another account, the Muslims were aided when the walls were badly damaged by an earthquake, a sure sign of God’s favour to them. In the end the two sides made peace. As usual, the inhabitants were obliged to pay taxes to the Muslims, some apparently at a fixed rate, others at a variable rate according to their prosperity at the time. All their lives, property, city walls, churches and water mills were to be guaranteed to them except for a quarter of the church of St John, which was to be turned into a mosque.³⁶ At the same time we are also told that half the houses should be made available to the conquerors. The general leading the Muslim conquest of the city is said to have ‘divided it up among the Muslims in lots so that they might occupy them [the houses]. He also settled them in every place whose occupants had evacuated it and every abandoned garden’.³⁷ Homs was an important centre on the fringes of the Syrian desert and it may have been thought that this was a suitable place for the Bedouin to settle. The city was probably the first in Syria to have a substantial Muslim population.

The clause about giving up a quarter of the church for use as a mosque may seem curious and perhaps improbable: after all, how could these two religions, whose followers had just been engaged in violent warfare, end up by sharing the main religious building in the town? We are told, however, that it also happened at Damascus, where the Muslims used half of the cathedral as the first mosque. Only at the beginning of the eighth century, sixty years after the conquest, were the Christians expelled and a purpose-built mosque constructed. Even then, compensation was paid and the Christians made a new cathedral in the church of St Mary, about half a kilometre east of the mosque, and this remains the cathedral of the Melkite (Greek Orthodox) community of Damascus to the present day. Interestingly, we find archaeological confirmation of this practice from a small town in the Negev, Subeita. Here there are two large, finely built Byzantine churches. In the narthex or porch of one are the foundations of a small mosque. We can tell it is a mosque because of the mihrab, the niche showing the direction of Mecca, which is clearly visible. All this evidence suggests that, after the political defeat of the Christian forces, the two religious communities could and did coexist, if not in harmony, at least in a measure of mutual tolerance.

The next city up the road to the north was Chalkis, which the Arabs called Qinnasrīn.³⁸ Whereas Homs is still one of the most important cities in Syria, Chalkis has virtually disappeared from the maps. Only recently have archaeological surveys and excavations in a little village just east of the Damascus-Aleppo road revealed the ancient site. Chalkis stood in the middle of a fertile, grain-growing plain; although an important administrative centre, it was never a very big city. The ancient acropolis can be distinguished, as can the early Islamic town, which lay outside the confines of the classical city: the Arabs settled outside the walls in what was effectively a new suburb, not in the city itself. After the fall of the city, Khālid b. al-Walīd decided to make it his home and he was joined there by his wife.

Heraclius seems to have moved from Antioch after the battle of Yarmūk and settled in Edessa, where he tried to organize the defence of northern Mesopotamia and south-eastern Anatolia. He then moved on along the upper Euphrates before turning west, heading for Constantinople, the capital he had not visited for the last ten years. There is no evidence, as some have suggested, that he was disabled by senility or depression, but he must have been weary and painfully aware of the scale of the Byzantine defeat. Arab authors put a number of sad and resigned speeches into his mouth and Heraclius’s farewell to Syria was widely reported. In one he is made to say, ‘Peace be upon you, O Syria. This is the farewell after which there will be no reunion. No Byzantine man will ever return to you except in fear [as a prisoner] until the coming of the Anti-Christ. How sweet will be his deeds [because he will fight the Muslims] and how bitter their outcome for the Byzantines [because he will be defeated].’⁴⁰ In another version he just says as he goes through the passes in the Taurus mountains and looks behind him, ‘Peace be upon you, O Syria [Sriya]! What a rich country this is for the enemy!’⁴¹ As he withdrew, he took with him all the garrisons from the districts along the new frontier, creating a sort of no man’s land between Byzantine and Muslim territory at the north-eastern corner of the Mediterranean.⁴² A later Syriac source, deeply hostile to everything Byzantine, says that Heraclius ‘gave order to his troops to pillage and devastate the villages and towns, as if the land already belonged to the enemy. The Byzantines stole and pillaged all they found, and devastated the country more than the Arabs’.⁴³

With the departure of the emperor, the remaining Byzantine cities were left to their own devices. Antioch, the ancient capital of Syria, put up little resistance and the remaining inhabitants seem to have made no effort to use the mighty walls the emperor Justinian had built around their city less than a hundred years before to keep the attackers out: probably there were just too few of them to defend the huge circuit. They are said to have rebelled against Muslim rule later, but this may mean only that they refused or were unable to pay taxes and had to be coerced into doing so. In other small towns, the surrender to the Muslim armies had an almost carnival atmosphere. At the little town of Shayzar on a bend of the Orontes river in central Syria the inhabitants came out to meet the Muslims with players of drums and cymbals, as was customary when greeting important visitors.⁴⁴ The same happened at Ma^carrat al-Nu^cmān and Apamea, once the proud capital of the Roman province of Syria II but now in deep decay after being sacked ferociously by the Persians sixty years before, in 573. It was not always as easy as that: when the people of Dar^ca⁴⁵ in southern Syria came out to greet the caliph Umar with drumming and singing, carrying swords and bunches of myrtle, the puritanical monarch ordered that they be stopped. His general, Abū Ubayda, by now used to the customs of Syrian small towns, explained that it was their custom and if he stopped them doing it, they would think he was breaking the agreement the Muslims had made with them. Reluctantly, the grumpy caliph allowed them to continue.

The most vigorous resistance the Muslims encountered was in the cities of the Syrian and Palestinian coasts. These had always been the areas in which Greek civilization was most firmly established and most deeply entrenched. It was also because the Byzantines were able to resupply and reinforce these towns from the sea. The Byzantine forces in Palestine had largely withdrawn to Egypt, but Gaza and Caesarea still held out. Gaza had been the scene of the first encounter between Amr b. al-Ās and the Byzantines at the very beginning of the conquest, and it seems that he now returned to the city and succeeded in taking it. It was natural from that position that his thoughts should turn to Egypt, with which Gaza had such close connections.

Further up the coast, the strongest resistance was in the city of Caesarea. While Gaza has been continuously inhabited and built over so that very few traces of its classical past have survived, Caesarea is largely deserted and the outlines of the ancient city, founded by Herod the Great (73-4 BC) as a window on the Mediterranean world, can still be seen. The city remained prosperous into the sixth century, with new residential quarters laid out between the great monuments of the classical period. Down by the harbour, an elegant octagonal church overlooked the quays and docks. It seems that the city held out for some years, possibly as late as 641, five years after the defeat of the Byzantine forces at the battle of Yarmūk, and we are told that it fell only when one of the Jewish inhabitants showed the Muslims how to enter through a concealed water channel. It is said that the man who led the conquering army was Mu^cāwiya b. Abī Sufyān. If so, it was a first military triumph for the man who was to become, twenty years later, the first Umayyad caliph and who was to rule the entire Muslim world from his base in Damascus. Because they had resisted so long and the city had been taken by storm, many of the inhabitants were enslaved and taken to the Hijaz where, we are told, they worked as secretaries and labourers for the Muslims (fi’l-kuttb wa’l-a^cml li’l-Muslimīn).⁴⁶ Perhaps we see here the beginnings of the Muslim appropriation of Greek culture, so characteristic of the early Islamic period.

In Latakia, modern Syria’s largest port, the inhabitants closed the great gate of their city walls against the invaders. The Arabs are said to have made a great effort and dug ditches deep enough to conceal a man and a horse. Then they pretended to be retiring to Homs. When night fell, they returned to their hiding places. In the morning the inhabitants opened the gate to drive their cattle out to the pastures; this was obviously a very agricultural town. The Arabs emerged suddenly from their hiding places and forced the gate, taking possession of the city. Here the inhabitants were allowed to keep the whole of their church and the Muslims built a new mosque for themselves.⁴⁷ The cities of Lebanon, Beirut, Tyre and Sidon put up no resistance. Only at Tripoli did the Byzantines hold out for a long time and, supplied from the sea, the city was defended until the beginning of the reign of the caliph Uthmān in 644. The Muslims built a small fortress outside the walls to keep watch on the inhabitants and finally they woke up one day to find that the defenders had all been evacuated overnight in Byzantine ships.⁴⁸ Its fall meant the final end of Byzantine control of any part of the eastern seaboard of the Mediterranean.

There was one city whose conquest was more of symbolic than military importance, and that was Jerusalem. The city had great significance for the early Muslims, as the first focus for their prayers and later as the site from which Muhammad is said to have begun the famous night journey on which the secrets of the heavens were revealed to him. Jerusalem at the end of the sixth century was a thriving centre of pilgrimage and ecclesiastical administration. The walls enclosed roughly the same area as the Old City today. We have an unusual insight into the appearance of the cityscape because of a document known as the Madaba map.⁴⁹ This is a mosaic map of the Holy Land laid on the floor of a church in the little Jordanian town of Madaba, probably at the end of the sixth century. The city of Jerusalem figures prominently in it. We can see the classical colonnaded streets, which follow the same route as the main streets of the Old City do today. We can see the walls and tower and the great church of the Holy Sepulchre, marking the site where Christ was crucified, buried and rose again. We can also see the great New Church, the Nea, built by the emperor Justinian as part of his campaign to beautify the city. Excavation since 1967 has recovered the foundations of the church and the new street that led to it, confirming the accuracy of the map. There is one area of the city which the map does not tell us about, the Temple Mount. This is the vast platform where Herod’s temple had stood and which had probably been empty since the Romans destroyed the temple in AD 70. Sixty years after the Muslim conquest, the Umayyad caliph Abd al-Malik was to build the Dome of the Rock on the spot, generally regarded as the third most holy place in Sunni Islam after Mecca and Medina. It would be fascinating to know what, if anything, Umar found on the site but, tantalizingly, the mosaic has been destroyed just at the point where the temple platform should be: if the accident of survival had spared another few centimetres of the ancient tesserae, we might have an answer to this question.

The man in charge of Jerusalem was the newly appointed patriarch, Sophronius. He was a Greek churchman, educated and sophisticated, with a lively contempt for the rude Bedouin. For Sophronius, the appearance of the Arabs was a sign of God’s anger at the sins of the Christian people. In a fiery sermon he berated them: ‘Whence occur wars against us? Whence multiply barbarian invasions? Whence rise up the ranks of the Saracens against us? Whence increases so much destruction and plundering? Whence comes the unceasing shedding of human blood? Whence the birds of the heavens devour human bodies? Whence is the cross mocked? Whence is Christ Himself, the giver of all good things and our provider of light, blasphemed by barbarian mouths?’ ‘The Saracens’, he went on, ‘have risen up unexpectedly against us because of our sins and ravaged everything with violent and beastly impulse and with impious and ungodly boldness.’⁵⁰ This is the authentic voice of high Greek culture, appalled and dismayed about the Muslim conquest of Syria.

The opportunity came when Umar visited Syria. As usual with the Arabic sources there is considerable confusion about when he did this and, indeed, whether there was one visit or several.⁵¹ The most likely scenario is that the caliph came to Jābiya in 637 or 638 and while he was staying there, dealing with a wide range of administrative matters, a delegation from the city arrived to make terms. They came on horseback, wielding swords, and some in the Muslim camp assumed they were hostile raiders. But the caliph, preternaturally wise as always, was able to reassure them that they were only coming to negotiate. The purported text of the agreement that was reached has come down to us:

In the name of God, the Merciful, the Compassionate. This is the assurance of safety [amn] which the servant of God Umar, the Com mander of the Faithful, has given to the people of Jerusalem. He has given them an assurance of safety for themselves, for their property, their churches, their crosses, the sick and healthy of the city and for all the rituals which belong to their religion. Their churches will not be inhabited by Muslims and will not be destroyed. Neither they, nor the land on which they stand, nor their cross, nor their property will be damaged. They will not be forcibly converted. No Jew will live with them in Jerusalem.

The people of Jerusalem must pay the taxes (jizya) like the people of other cities and must expel the Byzantines and the robbers. Those of the people of Jerusalem who want to leave with the Byzantines, take their property and abandon their churches and crosses will be safe until the reach their place of refuge. The villagers [ahl al-ard, who had taken refuge in the city at the time of the conquest] may remain in the city if they wish but must pay taxes like the citizens. Those who wish may go with the Byzantines and those who wish may return to their families. Nothing is to be taken from them before their harvest is reaped.

If they pay their taxes according to their obligations, then the conditions laid out in this letter are under the covenant of God, are the responsibility of His Prophet, of the caliphs and of the faithful.⁵²

Whether the text really is that agreed by Umar, or an ancient fabrication, we cannot be sure, but it gives a clear impression of how the Muslims should respond to their newly conquered Christian subjects. The fact that it bore the name of Umar undoubtedly gave it added weight and authority. The emphasis on the security of religion is not surprising, given the particular status of Jerusalem. Rather more unexpected is the provision that Jews are not to be allowed to settle in the city. This prohibition had been a feature of Roman law and the fact that a Muslim source records it suggests that the Christian negotiators had played a strong hand. Some of the clauses throw interesting light on the circumstances of the city. The provision made for Greek officials to leave points to an emigration of the upper and official classes, and the clauses about the country people who have come to the city are a clear reflection of contemporary circumstances.

Umar then visited the city in person. The fullest account of his visit is given in the Christian Arabic chronicle of Sa^cīd b. Batrīq, also known by his Christian name of Eutychius.⁵³ Writing in the eleventh century, he preserved traditions intended to show how Umar had safeguarded the position of the Christians in the Holy City. According to his account, Sophronius welcomed Umar to the city and the people were given guarantees about their property and freedom of religious observance. When it was the time for prayer, the patriarch suggested that the caliph should pray in the Church of the Holy Sepulchre itself, but Umar refused because he said that if he did, the Muslims would take it as a shrine and it would be lost to the Christians. He then issued a document in which the Muslims were forbidden to pray in the precincts of the church and, as a result, the church has remained in Christian hands ever since. Umar then requested a site to build a mosque and the patriarch took him by the hand to the rock on the platform where Herod’s temple had once stood. The narrative is clearly fashioned to make it clear that the status of the Christians in Jerualem was based on the unimpeachable authority of Umar himself.

In the Arabic tradition, Umar was guided by one Ka^cb b. Abhar, a Jew who was converted to Islam and is said to have introduced many stories and traditions about the Jews into the new religion. In response to the caliph’s question, Ka^cb suggested that the rock, which sticks up in the centre of the platform, should be the direction of their prayer on that day, but Umar rejected this, making it clear that God had reserved this role for the Ka^cba in Mecca. Umar was well aware that the site marked the position of the Jewish temple which had been destroyed by the Romans after the great Jewish rebellion in AD 70 and had been left as a rubbish tip in Byzantine times. He set about clearing the site himself and the people followed his example. He may have ordered the erection of a simple place of prayer. Certainly when the European Christian pilgrim Arculf visited Jerusalem after the Muslim conquest but before the beginning of the building of the Dome of the Rock in 685, he found a basic place of worship there. It is for this reason that the Dome is sometimes quite misleadingly referred to as the mosque of Umar (or Omar).

By 640 the whole of Syria, apart from one or two coastal towns, had come under Islamic rule. The northern boundary of Muslim rule was established at Antioch, the ancient city of Cyrrhus and Manbij. Garrisons were established and the local people enjoined to let the Muslims know of any approaching Byzantine forces. For the moment, however, the Byzantines were too devastated by defeat, the death of Heraclius in February 641 and the subsequent struggles for succession to the imperial title to be able to mount any sort of counter-offensive.

The completion of the conquest of Syria opened the way for Muslim armies to cross the Euphrates and begin the conquest of the Jazira. The Arabic word jazīra means ‘island’ but since the seventh century the term has been used to describe the ‘island’ between the Tigris and Euphrates rivers in the lands of modern Syria and Iraq. To the north, the Jazira was bordered by the mountains of the Anti-Taurus in south-east Anatolia, the border running more or less along the modern Turkish frontier. The landscape is mostly flat open plains and desert. A recent historian notes: ‘the Jazira is rather like the Mediterranean, an ocean of steppe punctuated by archipelagos of river valleys and hills and settled unevenly on its shores’.⁵⁴ There is a natural unity to this area and communications are quick and easy, but at the time of the Muslim conquest it was divided between Byzantine territory in the west and Sasanian lands in the east, with the frontier near the ancient city of Nisibis, more or less along the line of the modern Syria-Iraq border. This division determined the way in which it was conquered, Muslim forces from Syria taking the lands on the Byzantine side of the frontier, forces from Iraq taking the ex-Sasanian lands further to the east.

In the river valleys there were a number of ancient cities of which the most famous was Edessa. Edessa was one of the centres of early Christianity. In the first century AD its then king Abgar is said to have been the first monarch in the world to accept Christianity. Its great cathedral, of which nothing now survives, was one of the most magnificent buildings in eastern Christendom. It was also an important political centre, and Heraclius had made it his base in the final stages of his campaign in Syria.

The conquest of the Jazira was an important stage in the consolidation of Islamic rule in the Fertile Crescent. If it had remained in Byzantine hands it would have been a major threat to Syria and Iraq. Despite its strategic importance and the antiquity of its cities, the conquest of the Jazira is recounted in a very laconic fashion in the Arabic sources, and such accounts as there are are more concerned with the terms of capitulation than the course of the military campaign. ⁵⁵ Most agree that the conquest was led by Iyād b. Ghanam, who was ordered by the caliph Umar to lead a force of Syrian Arabs across the Euphrates. According to one account he had just 5,000 men with him,⁵⁶ but despite these small numbers he encountered little serious opposition. It seems that the withdrawal of the Byzantine imperial army left the local people with little choice but to capitulate and agree to the comparatively easy terms the Arabs were offering. Even at Amida (Diyarbakr), whose mighty city walls are one of the great glories of ancient and medieval military architecture, there seems to have been no resistance, and the same was true of the great castle the Byzantines had built in the sixth century at Dara to repel Persian attacks.⁵⁷ Edessa seems to have capitulated quickly on condition that the Christians could keep their cathedral but agreed not to build any new churches and not to aid the enemies of the Muslims. The city of Raqqa on the Euphrates also fell after a short resistance. The exact route of Iyād’s army as it toured the province accepting the surrender of smaller cities cannot be determined, but it seems that he may have finished by raiding along the ancient road that led to Armenia before halting at Bitlis. He then returned to Syria, where he died.

Syria had been conquered by Arab armies recruited in the Hijaz. This did not, however, result in a vast influx of new immigrants from Arabia. The Quraysh and their allies in the Muslim elite knew Syria well and they wanted to keep control over its resources. They did not want to share it with a mass of impoverished Bedouin. These were encouraged to move to Iraq instead. In British army parlance it could be said that Syria was for the officers, Iraq for the other ranks. They did not found new Muslim towns as later happened in Iraq and Egypt. All the towns that were important under Muslim rule had been important in Roman times (though some cities, such as Scythopolis, which had been important in Roman times, declined and virtually disappeared in the Islamic period). At one stage there seems to have been a project to establish a new town at Jābiya in the Golan, the old summer camping ground of the Ghassānids. It was here that the caliph Umar had come to meet the leaders of the victorious armies on his visit to Syria. But Jābiya remained just that, a summer camping ground: no mosque was built there, no government palace and no plots were allotted to different tribes. Instead the Muslims seem to have preferred to settle in established towns. We have seen how houses in Homs were made available for them. At Chalkis and Aleppo what were effectively Bedouin suburbs were established outside the walls of the old cities.

In part this was possible because sections of the Byzantine elite had fled to Constantinople or further west, leaving space in the towns. After the fall of Damascus, many people left the city to join Heraclius⁵⁸ and prominent Muslims were able to take up residence: Amr b. al-Ās owned several residences in Damascus and estates in Palestine. The treaty Umar made with the citizens of Jerusalem assumed that elements of the Byzantines would leave, either voluntarily or under coercion. It also looks as if many areas of Syria had been depopulated by plague and warfare and the Muslim conquerors had driven out many of the Byzantine inhabitants of the coastal cities.⁵⁹ It was hard to find men to garrison the port cities of the Mediterranean coast. Mu^cāwiya was obliged to settle Tripoli with Jews, in the absence of any Muslims who could be persuaded to take up residence there. Muslims were also settled in villages around Tiberias and were sometimes given deserted agricultural land on condition that they brought it under cultivation. There is no evidence of great tribal migrations of the sort that are well attested in Iraq.

Something of the day-to-day working of the relationship between the Arab tribes and the inhabitants of the towns and villages can be seen from a group of papyri found in the little town of Nessana in the Negev.⁶⁰ Some of these are bilingual, Greek and Arabic, demand notes ordering the Christian people of the town to provide the Bedouin in the area with supplies of wheat and olive oil and, in some cases, cash. The payments seem to have been made directly to the tribal chiefs and there was no complicated bureaucracy involved. How the local people collected the supplies and divided up the burden seems to have been up to them. The documents, which date from 674-5, a generation after the conquest, show how simple and, in a way, informal the Arab occupation could be.

The pattern of Arab settlement in Syria had another consequence. In Iraq and Egypt the Muslim settlers in the towns were directly dependent on the state for their pensions, often their sole means of livelihood. In Syria, by contrast, many of the new elite had urban or rural properties off which they could live. Within a generation members of the Muslim elite in Syria were building themselves luxurious and imposing residences in the countryside, something, again, which seems to have been virtually unknown in Iraq or Egypt.

So, if there was no massive influx of Arabs sweeping away Graeco-Roman civilization, what actually changed in Syria as a result of the Muslim conquests? At the most obvious level, the government and administration were controlled at the top by Arabic-speaking Muslims, but at second glance even this change was not as dramatic as it might initially appear. For the first half-century, the bureaucracy continued to use Greek and was staffed in large measure by local Christians. There was a new elite religion, but it seems to have made little impact on the built environment. In Iraq, in the new towns of Kūfa and Basra, the mosque lay at the heart of the Muslim city; in Damascus at the same time the Muslims had to make do with a half-share of the cathedral church in the city centre.

There is little evidence, either, for the bedouinization of the countryside. The impression that the Arab conquests resulted in hordes of nomads coming in and ravaging the settled lands seems to have been generally untrue, although there may have been incidents of violence and destruction in the course of the invasions. In those fragile, marginal areas such as the Syrian steppe east of Homs, Transjordan and the Negev in southern Israel, areas where the boundaries between cultivated lands and the pastures of the nomads shifted according to political and cultural changes, the evidence suggests that the first century of Muslim rule saw an expansion of settled agriculture. Not until after 750, when the Syria-based Umayyads were overthrown by the Iraq-based Abbasids, did the boundaries of settlement retreat and the Bedouin areas expand.

The Muslim conquest of Syria did, however, have profound effects on the long-term history of the area. It brought to an end almost a thousand years of rule by Greek speakers with contacts in the Mediterranean world. From this point on, the most important links were not with Rome or Constantinople but with Mecca and Medina, and later with Baghdad and Cairo. The emergence of Islam as the dominant religion and Arabic as the near-universal language could not have occurred without the conquest. These deep changes in language and culture may have taken some time but they could not have occurred without the military conquests of the 630s.