The Great Arab Conquests

Sind was very remote from Arabia and the heartlands of the early Muslim state.¹ The overland route led through the pitiless deserts of Makrān, where the track led from one parched oasis to another and where supplies were almost impossible to obtain. Alexander the Great had been one of the few men to try to lead an army through this land, and it proved one of the toughest struggles he had had to face. The alternative route was by sea, along the barren southern coast of Iran and Makrān to the ports around the mouth of the Indus river. In either case, the distances and the nature of the terrain made the journey very difficult.

Our knowledge of the Arab conquest of Sind in the early eighth century is very limited. The area is largely neglected by the classical Arab authorities. Only Balādhurī gives a systematic account and that occupies only about a dozen pages of the text.² There is no indication that he, or any of the other Iraq-based chroniclers, had ever visited this remote outpost of the Muslim empire, and the few details they give shed little light on the country or its conquest. Nor are local sources much more forthcoming. The only Sindi chronicle to deal with the conquests is the Chāchnāmah³ of 1216 by Alī b. Hāmid al-Kūfī, a translation of a lost Arabic original, said to have been collected by written by a qādī of Al-Rūr who claimed descent from the Thaqafīs, the tribe of the leader of the original conqueror, Muhammad b. Qāsim. The second half of this work is essentially an account of the first phase of the conquest.⁴ The Chāchnāmah has not been held in high regard by historians and it contains many legendary accretions, but much of the core narrative seems to be derived from early Arabic sources: the author names the historian Madā’inī, and the outline of the narrative, and some specific incidents, are closely based on Balādhurī’s texts. Two themes are stressed in the text. One is the powerful role played by Hajjāj in distant Iraq. He is described as having absolute day-today control of the campaigning. Muhammad b. Qāsim scarcely moved without writing to his master and waiting for the reply, which always came with improbable speed by return of post. On one occasion the text describes Hajjāj ordering Muhammad to draw a sketch map of the River Indus so that he can give advice on the proper place to cross it.⁵ What is meant to be conveyed, clearly, is the authority Hajjāj had over his commanders in the field. A second theme is the role of soothsayers and wise men, who are constantly telling the Sindi princes that the Arab conquests had been predicted and that there is nothing that can be done to prevent them. The Chāchnāmah contains some material said to have been conserved among the descendants of the original Arab conquerors which may be genuine, and some Arabic poetry that was not translated into Persian along with the rest of the book. This too may be of eighth-century origin.

Arabs had had contact with Sind, before the coming of Islam. In late Sasanian times there was a growing trade by sea between the Gulf and Sind and one group of Arabs was especially important in the development of this trade. The Azd tribe of Uman may have been remote from the centres of early Muslim power in the Hijaz but they were well placed to play a role in the maritime trade in the Indian Ocean. They converted to Islam and played an important part in the conquest of Fars and other areas in Iran. They formed a powerful lobby, wanting to invade Sind to further their commercial aspirations.

Sind at this stage has been described as ‘the wild frontier of Indian civilisation’,⁶ but for the early Muslims it was ‘the land of gold and of commerce, of medicaments and simples, of sweetmeats and resources, of rice, bananas and wondrous things’.⁷ It derived its name from the Sanskrit Sindhu, the name of the river known in the west as the Indus and to the Arabs as the Mihrān. Sind is created by the Indus river system in the same way as Egypt is created by the Nile. Arab geographers of the tenth century recognized the resemblance: ‘It is a very great river of sweet water,’ Ibn Hawqal wrote. ‘One finds crocodiles in it, like the Nile. It also resembles the Nile by its size and by the fact that its water level is determined by summer rains. Its floods spread over the land, then withdraw after having fertilized the soil, just like the river in Egypt.’⁸

At the time of the Muslim invasion, the settled parts of the country were ruled by a dynasty of kings of Brahmin origin. This had been founded by Chāch (c. 632-71) and was led in the early eighth century by Dāhir (c. 679-712), who led the resistance to the Muslims.⁹ The king seems to have lived in the city that the Arabs called al-Rūr, and the main port was the city of Daybul. The shifting course of the Indus delta has made the identification of this site very difficult, but it is probably to be identified with the ruins at Banbhore which now lies in desolate salt flats about 40 kilometres from the sea east of Karachi. The city first appears in the historical record in the fifth century, when it was a distant outpost of the Sasanian Empire. In the time of King Chāch and his son Dāhir it seems to have been a base for pirates, attacking trade between the Gulf and India, and suppression of this was one of the reasons for the Muslim attack.¹⁰

Much of the country was occupied by semi-nomadic tribes like the Mīds and the Jats, known to the Muslim sources as the Zutt. The Mīds supplemented the meagre livelihood they could scrape from their barren homelands with piracy against merchant shipping. The Zutt were agriculturalists who used water buffalo to cultivate the swampy lands by the Indus and grow sugar cane. According to Muslim sources, some Zutt were transported to southern Iraq by the Sasanian shāh Bahrām Gūr (420-38) to cheer his people up with their music-making! ¹¹

We are on firmer historical ground with the campaign of 710-12. According to Balādhurī, the immediate reason for the expedition was that the ‘king of the Island of Rubies’ (Sri Lanka) sent Hajjāj, governor of Iraq and all the east, some women who were daughters of Muslim merchants who had died in his country. The author adds the note that it was called the Island of Rubies ‘because of the beauty of the faces of its women’.¹² On the journey, the ship was attacked by some Mīd pirates sailing out of Daybul and was captured with all its passengers. One of the women, in her distress, was said to have called on the name of Hajjāj, and when he heard of the attack, he determined to take action.

He first wrote to Dāhir, the king, ordering him to set them free, but the king replied that he had no control over the pirates who had taken them and was unable to help. Hajjāj then sent two small expeditions, but in both cases they were defeated and the leaders killed. He then decided on a larger-scale campaign. He chose as its leader a young cousin of his called Muhammad b. al-Qāsim al-Thaqafī (from the tribe of Thaqīf, originally from Tā’if). Muhammad was something of a golden boy and was described as ‘the noblest Thaqafī of his time’.¹³ He is said to have been given high command at the age of seventeen but he proved an efficient commander and a wise and tolerant governor. His brief, meteoric career and tragic end left a lasting memory in both Sind and the central Islamic lands. Hajjāj ordered him to gather an army in the newly founded city of Shiraz in south-western Iran; 6,000 professional soldiers from Syria were dispatched to form the core of the army and he sent all the equipment he needed, ‘even including needles and threads’. When all was ready, they set out on the long land route through southern Iran and then into Makrān, taking the city of Fannazbūr en route. Meanwhile ships were sent with men, weapons and supplies.

The forces met up outside Daybul. Muhammad immediately began to invest the city, digging siege works. He ordered that lances be set up with tribal banners flying from them and that the troops encamp by their own flags. He also set up a catapult (manjanīq) served by 500 men, which was known as the Bride. This suggests a large, hand-pulled swing-beam siege engine, and is one of the very few examples we have of the Muslim forces using siege artillery during the conquests. One of the main features of the city was a temple described as a budd, like a great minaret in the middle of the city; this was probably a Buddhist stupa. On top of the temple there was a mast (daqal) from which a great red banner flapped and twisted in the wind. This mast now became the target of the siege engine and when it was brought down morale in the city collapsed. Muhammad ordered up ladders and his men soon began to scale the walls, so taking the city by force.¹⁴ Dāhir’s governor fled and there were three days of slaughter in which all the priests of the temple, among others, were killed. Muhammad then ordered the building of a mosque and laid out plots for the settlement of 4,000 Muslims.

Muhammad now made his way inland to the fortified city of Nīrūn near the banks of the Indus. Here he was met by two Buddhist (Samani) monks¹⁵ who began negotiations. They made peace and welcomed him into their city, giving him supplies.¹⁶ As he pressed on up the river, the pattern was repeated, with Buddhist monks frequently acting as peacemakers. According to the Chāchnāmah¹⁷ the city of Sīwīstān fell because of divisions among the local population. On one side was the Buddhist party, on the other the Hindu governor of the fort. The Buddhists told the commander of the fort that they would not fight: ‘Our religion is peace and our creed is good will to all. According to our faith, fighting and slaughtering are not allowed. We will never be in favour of shedding blood.’ They added that they were afraid that the Arabs would take them to be supporters of the governor and would attack them. They urged him to make a treaty with the Arabs because ‘they are said to be faithful to their word. What they say they do’. When the governor refused to listen to their advice, they sent a message to the Arabs saying that all the farmers, artisans and common people had deserted the governor and that he was now unable to put up a prolonged resistance. The fort held out for a week before the commander made his escape by night. The Muslims entered the city, which was looted in the customary way, except that the possessions of the Buddhist party were respected. As always with the Chāchnāmah, it is difficult to separate fact from fiction, but the narrative does suggest that Buddhist pacifism may have been a factor in the success of Arab arms and that division between the ordinary people and the Hindu military caste allowed the Muslims to take some towns with comparative ease.

Dāhir still remained as the leader of the resistance. Muhammad, on the west bank of the Indus, confronted him across the river.¹⁸ The Chāchnāmah gives a detailed account of how Muhammad crossed the river to attack Dāhir.¹⁹ He decided to build a bridge of boats and collected boats filled with ballast of sand and stones linked together with connecting planks. Meanwhile, Dāhir’s supporters gathered on the east bank of the river to oppose their landing. Muhammad ordered that all the boats should be brought together along the west bank until the row of boats was as long as the river was wide. Then brave armed soldiers gathered on the boats and the whole row was swung round in the stream until it reached the other bank. Immediately the Arabs drove the infidels back with volleys of arrows and the horsemen and foot soldiers landed.

The final confrontation between Muhammad b. Qāsim and Dāhir is given a few terse lines in Balādhurī but is described in dramatic terms in the Chāchnāmah. The Sindi army was composed of 5,000 veteran warriors (or 20,000 foot soldiers) and sixty elephants. Dāhir was mounted on a white elephant, armed with a tightly strung bow, with two female servants with him in the litter, one to hand him betel leaf to chew, the other to keep him supplied with arrows. There were speeches made on both sides and the names of numerous Arab warriors are given, a sure sign that this part of the Chāchnāmah at least is based on an Arab original. We are also told how Arabs who had previously joined Dāhir’s forces, for reasons that are not explained, now came to give Muhammad b. Qāsim vital information about his opponents’ movements. In the fierce fighting that followed, the Muslims used flaming arrows to set fire to the litter in which Dāhir was fighting and the elephant threw himself into the water. Dāhir was seized and decapitated, his body being identified by the two slave girls who were with him in the litter. The historian Madā’inī preserves a short poem of triumph said to have been uttered by the Arab who killed him:

The defeat and death of Dāhir meant the end of organized resistance. Many of Dāhir’s women committed suicide, burning themselves, their attendants and all their possessions, rather than be captured. The Chāchnāmah puts a little speech into the mouth of the dead king’s sister: ‘Our glory has gone and the term of our life has come to its close. As there is no hope of safety and liberty, let us collect firewood and cotton and oil. The best thing for us, I think, is to burn ourselves to ashes and so quickly meet our husbands in the other world.’²¹ They all entered a house, set fire to it and were burned alive. Despite this self-sacrifice, the chronicle says that many high-caste women of great beauty were sent to Hajjāj in Iraq. He in turn passed them on to the court of the caliph, where they were sold or given away to favoured relatives and supporters. The remnants of Dāhir’s forces were pursued to Brahmanābādh, near where the Muslim city of Mansūra was later founded, where they were again defeated. The Chāchnāmah preserves an account of the dealings of Muhammad b. Qāsim with the inhabitants of Brahmanābādh which probably reflects many of the issues raised by the Muslim conquest. His immediate response was to spare all the artisans, traders and common folk and execute all the military classes. ²² He soon became aware of the need to recruit local officials for the administration. His first move was to conduct a census of the merchants and artisans, who were obliged to convert to Islam or pay the poll tax. He next appointed village chiefs to collect the taxes. Meanwhile the Brahmins sought to secure their status under the new regime. They came to Muhammad b. Qāsim with their heads and beards shaved, as a sign of humility, and petitioned him. First they secured a safe conduct for all surviving members of Dāhir’s family, including his wife Lādī, who was brought out from her inner chamber. She is said to have been purchased by Muhammad b. Qāsim and to have become his wife.²³ It is interesting to compare this with the contemporary marriage of the son of Mūsā b. Nusayr, conqueror of al-Andalus, with the daughter of the Visigothic king Rodrigo. In both cases the Arab conquerors were seeking to ally themselves with the old ruling house, perhaps in the hope that their descendants would become hereditary rulers. In both cases, however, they were thwarted by the robust action of the government in Damascus.

More complaints were raised by the keepers of Buddhist temples.²⁴ They had previously survived on charitable donations but these had dried up because the people were afraid of the Muslim soldiers. ‘Now’, they lamented, ‘our temples are lying desolate and in ruins and we have no opportunity to worship our idols. We beseech that our just governor will allow us to repair and construct our Buddhist temples and carry on worship as before.’ Muhammad wrote to Hajjāj, who replied that, as long as they paid their tribute, the Muslims had no further rights over them, so they should be allowed to maintain their temples as before. In a meeting held just outside the town, Muhammad gathered all the chiefs, headmen and Brahmins and permitted them to build their temples and carry on commerce with the Muslims. He also told them to show kindness to the Brahmins and to celebrate their holy days as their fathers and grandfathers had before and, perhaps most importantly, to pay three out of every hundred dirhams collected in revenues to the Brahmins and send the rest to the treasury. It was also settled that the Brahmins (presumably those who did not benefit from the tax revenues) would be allowed to go around begging from door to door with copper bowls, collecting corn and using it as they wished.

A further problem was the status of the Jats.²⁵ Muhammad’s advisers described their low status and how they had been discriminated against in the reign of King Chāch: they had to wear rough clothing; if they rode horses they were not allowed to use saddles or reins but only blankets; they had to take dogs with them so that they could be distinguished; they were obliged to help as guides for travellers day and night; and if any of them committed theft, his children and other members of his family were to be thrown into the fire and burnt. In short, ‘they are all of the wild nature of brutes. They have always been refractory and disobedient to rulers and are always committing highway robberies’. Muhammad was easily convinced that these were ‘a villainous set of people’ and that they should be treated accordingly.

These discussions are very interesting, not because they are necessarily an accurate record of what happened but because of what they tell us of the Muslim settlement and how people viewed it. At the most obvious level, it shows how the Muslims took over the existing administrative personnel and left prevailing social structures largely intact. The accounts serve a twin purpose, of explaining to a Muslim audience how it was that Brahmins continued to be so influential under an ostensibly Muslim government, and why temples should be tolerated. They also showed how the Buddhists were to be tolerated and allowed to practise their religion. For all the non-Muslims, they showed how their status had been accepted by the founding father of Sindi Islam, Muhammad b. Qāsim, and his great adviser, Hajjāj himself. To the unfortunate Jats, they simply showed that the coming of Islam was to bring them no benefits at all.

Muhammad b. Qāsim’s march now became something of a triumphal progress, and at one point the Muslims were greeted by the people dancing to the music of pipes and drums. When he enquired, Muhammad was told that they always greeted their new rulers in this way.²⁶ The next main objective was al-Rūr, described as the biggest town in Sind. Dāhir’s son Fofi had fortified himself in the city and intended to resist. According to the Chāchnāmah,²⁷ Fofi and the people of al-Rūr believed that Dāhir was still alive and that he would soon come to rescue them. Even when Muhammad produced his widow Lādī and assured them that he was dead, the defenders accused her of conspiring with the ‘cow-eaters’ and reaffirmed their faith that he would come with a mighty army to save them. According to this imaginative account, they were convinced only by the testimony of a local sorceress. When consulted, she retired to her chambers and emerged after a few hours saying that she had travelled the whole world looking for Dāhir, producing a nutmeg from Ceylon as proof of her voyages, and had seen no sign of him. This intelligence persuaded many of the inhabitants that they should open negotiations with Muhammad, whose reputation for virtue and justice was well known. That night Fofi and his entourage slipped away, and when the Arabs began to assault the city the next day, the merchants and artisans opened negotiations, saying that they had given up their allegiance to the Brahmins and were convinced that the forces of Islam would triumph. Muhammad accepted their overtures, after receiving assurances that they would abandon all military operations. The population gathered at a shrine called the Nawbahār (the same name as the great Buddhist shrine at Balkh) and prostrated themselves before the marble and alabaster image. Muhammad asked the keeper of the temple whose image it was. He also took one of the bracelets from the arms of the statue. When the keeper noticed that it had disappeared, Muhammad teased him, asking how it could be that the god did not know who had taken the bracelet. Then, laughing, he produced it and it went back on the arm.

After the surrender, Muhammad ordered a number of executions of fighting men, but Lādī interceded, saying that the people of the town were ‘good builders and merchants, who cultivated their lands well and always kept the treasury full’, so Muhammad spared them. Once again the narrative points to the compromises and working relationships that accompanied the conquests: the temple was undisturbed and the livelihoods of most of the inhabitants continued uninterrupted. The Muslim conqueror was celebrated, not for his zeal in the rigid enforcement of Islamic norms, but for his tolerance and easygoing humour. This is also in marked contrast with the destruction of temples and religious figures during the Arab conquest of Transoxania at exactly the same time. It is difficult to know whether this was the result of the pacific approach of the Buddhists or simply because the Muslims were too few in number to challenge existing customs. When the town had submitted completely, Muhammad left two of his Arab followers in charge, exhorting them to deal kindly with the people and look after them.

The other main city, Multān, fell soon after. The conquest, called ‘the opening of the house of gold’ by the victorious Arabs, marked the furthest point of the Muslim advance in Sind at this stage. The town was rich and the temple (budd) was a major pilgrimage centre. The inhabitants put up a stout resistance and the besieging Muslims ran very short of supplies, being obliged to eat their own donkeys. The end came when they were shown how drinking water entered the city and were able to cut off the supply. The people surrendered unconditionally. The men of fighting age and the priests were all put to death and the women and children enslaved. The Muslims acquired vast amounts of gold.²⁸ Curiously, there is an old tradition that Khālid b. al-Walīd, known to history as the conqueror of Syria, is buried in Multān and his supposed tomb is the oldest Muslim building in the city.

The conquests in India posed a new sort of problem for the victors. In most of the lands that the early Muslims conquered, the majority of the population could be considered ‘people of the Book’, which meant that they could be allowed to keep their lives, property and religious practices as long as they accepted Muslim rule and their status as dhimmis. After the conquest of Iran, it had been gradually accepted that the Zoroastrians could be considered as ‘people of the Book’ too. The problem in Sind was that the population were mostly either Buddhists or Hindus. As far as most Muslims were concerned, Buddhists and Hindus, with their elaborate images and statues, were no more than idolators plain and simple, who could be exterminated at will if they did not convert to Islam. The Arab conquerors of Sind soon tempered their religious enthusiasm with pragmatism. After he had taken al-Rūr, Muhammad is said to have reasoned that ‘the budd are like the churches of the Christians, the synagogues of the Jews and the fire-temples of the Magians’ and that they should be respected in the same way. In practice this meant that both Buddhists and Hindus should, in effect, be accepted as dhimmis. In many cases local Brahmins and Buddhist monks continued to run the local administration for their new Muslim masters.

The initial conquests were brought to an abrupt end by events in the Muslim heartlands. In 715, when Muhammad had been in Sind for three and a half years, there was a major change of government. Hajjāj, his relative and patron, had died in 714 and the caliph Walīd I followed him the next year. The accession of Sulaymān to the Umayyad throne saw a violent reaction against Hajjāj and his officials. Muhammad was unceremoniously ordered back to Iraq, where he was imprisoned and tortured by the new governor and soon died in captivity. He deserved better. Like his contemporaries, Qutayba b. Muslim in Khurasan and Mūsā b. Nusayr in Spain, he found that his achievements in the service of Islam were no protection against vengeful political rivals.

Muhammad’s dismissal marked the virtual end of active campaigning. In the short period of his rule, Muhammad had laid the foundations for Muslim penetration of the subcontinent. He had established the legal framework and precedents that would allow Muslims to live at peace with Buddhists and Hindus. Compared with later Muslim invaders of the Indian subcontinent like Mahmūd of Ghazna in the early eleventh century, he left a reputation for mildness, humanity and tolerance, and the natives wept over his disgrace.²⁹ He had also made a vast amount of money. Hajjāj is reported to have made a simple balance sheet for the whole campaign. He reckoned he had spent 60 million dirhams on equipping and paying Muhammad’s forces, but his share of the booty had amounted to 120 million, a tidy profit by anyone’s standards.³⁰ As usual, the figures may well be exaggerated, but this is the only record we have of someone attempting such a clear calculation in the whole history of the early Muslim conquests. The sums make it clear that such expeditions could be a very useful way of generating revenue.

The Muslims were now in possession of most of the lower Indus valley. The area from Multān south to the mouth of the river was to be the limits of Muslim settlement on the Indian subcontinent. It was separated from the rest of India (Hind) by the deserts that now divide Pakistan from India to the east of the Indus. To the north of Multān, the Punjab was outside Muslim control until the early eleventh century, when the Ghaznevids from eastern Afghanistan extended Muslim rule further to the north and east.

There is an interesting footnote to the Arab conquest of Sind. As we have seen, some of the Zutt were already settled in Iraq before the coming of Islam. Many more seem to have arrived as a result of the Muslim involvement in their native lands in the Indus valley. Soon the Umayyad caliphs moved some of them to the hot plains around Antioch in northern Syria, along with their water buffalo. Some of those in northern Syria were later captured in a Byzantine raid on Ain Zarba and taken away with their women, children and their precious water buffalo. Gypsies under the Greek name of Atsinganoi appear in the neighbourhood of Constantinople in the eleventh century. The Zutt of Iraq remained as a restless element in the local population, but they disappear from history after the year 1000. In 1903 M. J. de Goeje, the great Dutch orientalist, published a monograph in which he suggested that these Zutt were the origin of at least some of the Gypsies of modern Europe.³¹ The Gypsy language clearly originates from north-west India, and they may have emigrated from Syria through the Byzantine Empire to the Balkans, where they first appear in the fifteenth century. There is, however, no direct evidence for this, and the theory remains no more than an intriguing speculation.

Tāriq may have been aware that there had recently been a major political upheaval in the Visigothic kingdom of Hispania. The Visigoths had conquered the Iberian peninsula in the fifth century. From their capital at Toledo they had ruled one of the most successful of the Germanic kingdoms, which had taken over the lands of the western Roman Empire. Though the kingdom had been in existence for almost three centuries, there is no indication that it was feeble or decadent. It is true that the cities were small and comparatively undeveloped and that much of the countryside seems to have been very sparsely populated, but the monarchy was strong and successful and there was no tradition of internal rebellion or separatist movements. The church was well established and a long series of councils held in Toledo testified to the vitality of its organization and activities.

On the face of it, the idea that a small group of Berbers with a few Arab officers could attack and destroy this formidable state was most implausible. The kingdom was undergoing a short-term crisis, however. In 710 King Witiza had died. He had left adult sons, but for reasons that we do not fully understand the throne had been seized by Rodrigo, a noble who may or may not have been related to the royal house. The sons of Witiza, and their friends and allies, were powerful and resentful. Rodrigo had had no time to establish his authority before the Muslims invaded. Tāriq also had more immediate reasons for planning his invasion. The men he commanded were mostly Berbers who had joined the Muslim army in the previous few years. It is most unlikely that any regular system of payment had been introduced to reward them for their allegiance to the new faith. If he was to retain their loyalty, he needed to find a source of revenue quickly. Spain was the obvious area where this could be done.

In the earliest Arabic work to describe the conquest, the history of Ibn Abd al-Hakam,³³ considerable prominence is given to the story of ‘Julian’. This mysterious figure is said to have been lord of Ceuta, a port city just to the east of Tangier which may still have been under Byzantine overlordship. According to the chronicler, ‘Tāriq wrote to Julian, paying him compliments and exchanging presents. Now Julian had sent his daughter to Rodrigo [the Visigothic king of Spain], for her education and instruction, and Rodrigo had made her pregnant. When this news reached Julian, he said, “I do not see how I can punish him or pay him back except by sending the Arabs against him”.’ He then goes on to describe how Julian transported some of the men one evening and sent his ships back to the African coast to bring more the next. The people on the Spanish side did not pay them any attention because they were just like the merchant ships that were often going to and fro. Tāriq came in the last boat and the fleet remained at Algeciras while the Muslim army marched north, just in case anything went wrong and they had to be rescued. It is impossible to know whether there is any truth in the story or indeed whether ‘Julian’ ever existed. It does not come from the usual repertoire of Arabic conquest narratives, however, and it may reflect the reality of a widespread discontent with Rodrigo’s kingship.

It was probably in April or May 711 that Tāriq embarked his small force in boats to take them across the straits. The force was unlikely to have been more than seven thousand men, of whom only a small minority were Arabs. The intention may have been simply to launch a large-scale plundering raid. Once across, the Muslims were able to take the ‘Green Island’, where the port of Algeciras stands today. This was to be a base but also allow them to retreat to the African coast if events turned out badly.

Rodrigo was campaigning against a Basque rebellion in the far north of his kingdom. When he heard of the Muslim raid, he hurried south, pausing at his residence at Cordova to gather more men. Like Harold of England and the Anglo-Saxons at the battle of Hastings in 1066, his army must have been exhausted by long marches to confront the invaders. Tāriq pursued a cautious policy. Rather than pushing on to attack Seville or the Guadalquivir valley, he kept close to his base and requested reinforcements from Africa; 5,000 more Berbers arrived, giving him a total of perhaps 12,000 men. He is also said to have been joined by some of the partisans of the sons of Witiza, opposed to the new king. The role of the Visigothic ‘opposition’ is controversial. From a modern Spanish point of view, it is easy to see that, if they did indeed aid the Muslim invasion, they were traitors. On the other hand, they, like most of their contemporaries, probably saw the Muslim invasion as no more than a raid, which would last a summer season at most. They could not have known that Muslims were to rule parts of the Iberian peninsula for the next 800 years.

The Muslim invaders may have enjoyed some support among the Jewish communities of the Iberian peninsula. This too is a very controversial issue with obvious contemporary resonances. The reality is that we have no hard evidence for this at all. We know that the Visigothic kings had introduced increasingly harsh anti-Jewish legislation, ending with the edict that they should all be converted to Christianity. It would be natural, therefore, for the Jews to welcome the Muslim invaders as potential liberators. There is no indication that this legislation was ever enforced, however, and there is absolutely no evidence that any Jews gave the Muslims active support.

The decisive battle was fought near the little town of Medina Sidonia. The exact site of the conflict is not known but it is generally believed to have been on the little river Guadalete.³⁴ Accounts of the battle are very sparse. The Latin Chronicle of 754 simply observes that ‘Roderick [Rodrigo] headed for the Transductine mountains [location unknown] to fight them and in that battle the entire army of the Goths, which had come to him fraudulently and in rivalry out of ambition for the kingship, fled and he was killed. Thus Roderick wretchedly lost not only his rule but his homeland, his rivals also being killed’.³⁵ The Arabic sources say that the battle took place on 19 July 711 and, like the Chronicle of 754, suggest that divisions within the ranks of the Visigothic army allowed the Muslims to triumph when the partisans of Witiza’s son Akhila turned and fled.³⁶

Seven hundred men, all of them mounted, were sent to Cordova under the command of the mawla ū Mughīth. The fall of Cordova, soon to be the capital of Andalus, is recorded with some circumstantial, and probably fictitious, details in the Arabic sources.³⁷ When Mughīth was approaching the city along the south bank of the Guadalquivir river, his men captured a shepherd who was looking after his flocks. They brought him to the camp and began to interrogate him. He said that the city had been deserted by all the leading citizens and only the governor (bitrīq) with 400 guards and some non-combatants (du^cafa) remained. On being questioned about the defences, he said that they were in good order except for a breach above the gate that led to the Roman bridge across the river. That night Mughīth led his men across the river and attempted to scale the walls with the aid of hooks, but it proved impossible. They returned to the shepherd, who guided them to the breach. One of the Muslims scaled the wall and Mughīth took off his turban and used it to pull others up. Soon there were a considerable number of Muslims on the wall. Then Mughīth came to the Gate of the Bridge, which was then in ruins, and ordered his men to surround the guards on the walls. Then they broke the locks and Mughīth and his men were soon inside.

When the governor (called al-malik in this account) heard that they had entered the city he fled with 400 of his men east to a church in which they fortified themselves. Mughīth laid siege to it. The resistance went on for three months until one day Mughīth was told that the governor had fled on his own, intending to establish a stronghold in the mountains behind the city. Mughīth set off in single-handed pursuit and caught up with him when his horse fell into a ditch and threw him. Mughīth found him sitting on his shield, waiting to be taken prisoner. ‘He was’, the chronicler goes on to explain, ‘the only one of the kings of al-Andalus to be taken prisoner. All the others either made terms for themselves or escaped to distant regions like Galicia.’ Mughīth then returned to the church. The defenders were all executed but the governor’s life was saved so that he could be sent to the caliph in Damascus.

Tāriq himself headed for the capital, Toledo. This seems to have been largely abandoned by its inhabitants: according to the Chronicle of 754 the archbishop, Sindered, ‘lost his nerve and like a hireling rather than a shepherd, and contrary to the precepts of the ancients, he deserted Christ’s flock and headed for his Roman homeland’.³⁸ Ibn Abd al-Hakam’s only contribution to the history of the taking of the Visigothic capital is the story of the sealed room which, like the story of Julian, has been passed down in history and legend. According to this there was a room (presumably in Toledo) with many locks. Every king added another lock on accession and none opened the room. Rodrigo, on becoming king, insisted on opening the room. On the wall they found pictures of Arabs and there was an inscription which said that, when the room was opened, these people would conquer the country.³⁹

Tāriq may have pushed on along the road that led to the Ebro valley, perhaps taking Guadalajara before returning to winter in Toledo. Meanwhile his superior, the governor of Ifrīqīya, Mūsā b. Nusayr, decided to join in what was looking like a very profitable venture. The next spring, of 712, he gathered an army of 18,000 on the coast opposite Gibraltar. This was a very different force from the one Tāriq had led the year before. The majority of them were Arabs. It included some tābi’ūn (followers, that is men who became Muslims in the generation after the Companions of the Prophet) and leaders of the main Arab tribes. In June 712 the army crossed to Algeciras. Rather than hasten to meet up with Tāriq in Toledo, Mūsā seems to have decided to consolidate the area of Muslim rule in the south. He began with some smaller towns, Medina Sidonia and Carmona, before turning his attention to Seville, one of the largest cities on the peninsula. Resistance does not seem to have been very prolonged and the Visigothic garrison evacuated the city and withdrew to the west.

Mūsā then went north along the Roman road to the city of Merida. Merida, now a medium-sized provincial town, had been one of the main capitals of Roman Spain and the impressive classical ruins still testify to its wealth and status. In early Christian times it had become the centre of the thriving cult of St Eulalia. Here the Muslims encountered much more serious resistance than they had in Seville or Toledo. It seems that Mūsā was obliged to lay siege to the town through the winter of 712-13 and that the city did not finally surrender until 30 June 713. Mūsā then set out to meet up with Tāriq, but before he did so he sent his son Abd al-Azīz back to Seville, where resistance had broken out. Mūsā advanced east along the Tagus to the Visigothic capital at Toledo, now held by Tāriq. Here he forced his subordinate to hand over the treasury and the riches he had confiscated from the churches. The Arab sources are, as often, very interested in the booty and its distribution. In this case, they report the rivalry between Tāriq and Mūsā. The focus of conflict was the ‘Table of Solomon’, kept in a castle outside Toledo. This was immensely valuable, made as it was of gold and jewels. It had been taken by Tāriq but Mūsā insisted that he should have it. Tāriq reluctantly agreed to hand it over but took off one of the legs and fixed an imitation in its place. Mūsā installed himself as a veritable sovereign in the ancient city while Tāriq retired to Cordova in high dudgeon.⁴⁰ As with the story of Julian, this clearly legendary material may point to wider political tensions, in this case the rivalry between Tāriq and his Berber followers and Mūsā and his largely Arab army.

The caliphs in Damascus were often very suspicious of successful conquerors, fearing, perhaps rightly, that they might escape from government control. The death of Walīd I in 715 meant that Mūsā b. Nusayr, like Muhammad b. Qāsim in Sind, was removed from office and brought back to Iraq to be punished. Both Mūsā and Tāriq were ordered to come to Damascus. Before they left, the two generals made an attempt to subdue the areas around the northern mountains. Tāriq took Leon and Astorga and then moved on over the Cantabrian mountains to Oviedo and Gijon. Many of the inhabitants abandoned the cities and fled to the mountains of the Picos de Europa.

Only then did the two conquerors decide to obey the caliph’s orders. Mūsā appointed his son Abd al-Azīz as governor of al-Andalus; other sons were appointed to Sūs and Qayrawān. This had the makings of a dynastic state and in other circumstances, in late Merovingian France, for example, the Muslim west might have developed as an independent lordship ruled by the family of Mūsā b. Nusayr. In the early Islamic empire, the ties that linked the most distant provinces to the centre were too strong. Muhammad b. al-Qāsim in Sind and Mūsā b. Nusayr in al-Andalus both accepted their fate, obeyed their orders and returned to the central Islamic lands. In both cases the conquering heroes were humiliated, dispossessed of their gains and imprisoned. Mūsā died in 716-17, probably still in confinement. Of the fate of Tāriq we know nothing at all, but he must have died in the Middle East in complete obscurity.

In the name of God, the Merciful, the Compassionate. This text was written by Abd al-Azīz b. Mūsā b. Nusayr for Tudmīr b. Ghabdush, establishing a treaty of peace and the promise and protection of God and His Prophet (may God bless him and grant him His peace). We [Abd al-Azīz] will not set any special conditions for him or for any among his men, nor harass him, nor remove him from power. His followers will not be killed or taken prisoner, nor will they be separated from their women and children. They will not be coerced in matters of religion, their churches will not be burned, nor will sacred objects be taken from the realm as long as Theodemir remains sincere and fulfils the following conditions we have set for him:

He has reached a settlement concerning seven towns: Orihuela, Valentilla, Alicante, Mula, Bigastro, Ello and Lorca.

He will not give shelter to fugitives, nor to our enemies, nor encourage any protected person to fear us, nor conceal news of our enemies.

He and each of his men shall also pay one dinar every year, together with four measures of wheat, four measures of barley, four liquid measures of concentrated fruit juice, four liquid measures of vinegar, four of honey and four of olive oil. Slaves much each pay half of this.

The treaty is a classic example of the sort of local agreements that were the reality of Arab ‘conquest’ in many areas of the caliphate. It is clear that rather than embark on a difficult and costly campaign, the Muslims preferred to make an agreement that would grant them security from hostile activities and some tribute. It is a pattern we can observe in many areas of Iran and Transoxania. It is interesting to note that much of this tribute was taken in kind (wheat, barley vinegar, oil, but of course no wine). In exchange for this, the local people were allowed almost complete autonomy. Theodemir was clearly expected to continue to rule his seven towns and the rural areas attached to them. There is no indication that any Muslim garrison was established, nor that any mosques were built. Theodemir and many of his followers may have imagined that the Muslim conquest would be fairly short lived and that it was worth paying up to preserve their possessions until such time as the Visigothic kingdom was restored. In fact it was to be five centuries before Christian powers re-established control over this area. We do not know how long the agreement was in force: Theodemir himself died, full of years and distinction, in 744. It is likely that it was never formally abolished but rather that as Muslim immigration and the conversion of local people to Islam increased in the late eighth and ninth centuries, its provisions became increasingly irrelevant.

The governorate of Abd al-Azīz was brought to an abrupt and unfortunate end. According to Ibn Abd al-Hakam,⁴² he had married the daughter of Rodrigo, the last Visigothic king, who brought him vast wealth and an exalted idea of royal prestige. She was dismayed by the modest state he kept and the informality with which his Arab followers approached him, not prostrating themselves before him. According to the story, she persuaded him to have a low door constructed in his audience hall so that they all had to bow before him as they came in. The Arabs resented this strongly and some even alleged that she had converted him to Christianity. A murder plot was hatched and the governor put to the sword. Clearly the story belongs to the genre that contrasts the simple, even democratic nature of Arab government with the hierarchy and pomp of the empires and kingdoms it replaced. It may also reflect a tension between those Arabs who had married rich heiresses from among the local people and the rank and file of the invading army.

The new rulers of Spain began to make their mark on the administration almost immediately. We can see this most clearly in the case of the coinage. The arrival of Mūsā b. Nusayr was marked by the minting of a new gold coinage, based not on Visigothic but on North African models. The earliest of these coins have the Latin legend ‘In Nomine Domini non Deus nisi Deus Solus’, a direct translation of the Muslim formula ‘There is no god but God’, an unusual mingling of Muslim and Latin traditions. This was probably produced in mobile mints that accompanied the army to recycle booty, perhaps valuables taken from churches, into cash money which could be more easily divided among and spent by the military.

The Muslim conquerors of Spain were not settled in military towns: there was no Iberian equivalent of Fustāt or Qayrawān. It seems rather that there was a much more dispersed pattern of settlement, in some ways more similar to the ways in which the Germanic invaders of the western Roman Empire in the fifth century settled in Gaul and Hispania. It looks as if the Arabs, who must mostly have come from urban backgrounds in Fustāt or Qayrawān, chose to settle in the cities and villages of the Guadalquivir and Ebro valleys, around Cordoba, Seville and Zaragoza, while the Berbers, who came from more pastoral backgrounds, established themselves on the high plains of the Meseta in the centre and the southern mountains.

The conquest had been astonishingly successful. Within five years of the initial invasion, almost the whole of the Iberian peninsula had been brought under the control of the Muslim armies. There was, however, an important and, as it turned out, fatal exception to this rule. In the north of Spain, as in some areas of the Middle East, the 1,000-metre contour line represented the limit of the territory held by the Muslims. This meant that in the high southern valleys of the Pyrenees and the Picos de Europa further west in the Asturias, small groups of refugees and indigenous inhabitants gathered to protect their independence from Arab rule. In the Picos de Europa, the movement is said to have been led by one Pelayo, who may have been a Visigothic noble and member of Rodrigo’s court. We know nothing about the history of this rebellion from the Arabic sources, but for the Christians of the Kingdom of the Asturias, the story of the rebellion was the foundation myth of their realm. As recounted in the Chronicle of Alfonso III,⁴³ probably composed soon after 900, Pelayo was about to be arrested by the Arabs but was warned by a friend and fled in the Picos de Europa. The landscape of the Picos is rugged, with steep gorges and rocky outcrops. Frequent rains mean that it is astonishingly green, with well-watered fields and forests and swift-flowing rivers. It was a very different landscape from the open plains of the Meseta to the south and a world away from the deserts of North Africa and Egypt. It had never been really part of Roman Spain, no big cities were established there and no Roman roads led through it.

Pelayo, according to the Chronicle, was able to escape when he came to the bank of a swiftly flowing river and swam across on his horse; his enemies were unable to follow. He fled into the mountains and established a headquarters at a cave which became the centre of resistance for people from all over the Asturias. The Arab governor was furious and sent an army of 187,000 men, a wholly fantastical figure, to put down the rebellion. They were led by an Arab commander, whom the source calls Alqama, and a mysterious bishop called Oppa, who is presented as a collaborator. The Muslims confronted Pelayo at a place called Covadonga, high in the mountains. The bishop addressed Pelayo and asked him how he thought he could withstand the Arabs (Ishmaelites) when they had defeated the entire Gothic army shortly before. Pelayo responded with a pious little homily, saying that ‘Christ is our hope and through this little mountain which you see, the well-being of Spain and the army of the Gothic people will be restored’.

After the breakdown of negotiations, the Muslim army attacked. Huge numbers of them were slain and the rest fled. The battle of Covadonga, usually dated to 717, has acquired mythical status as the beginning of Christian resistance. The failure of the Muslim forces to suppress the revolt led immediately to the loss of control of northern settlements like Gijon and the foundation of a small, independent Christian kingdom. It was this kingdom, and similar small entities in the Pyrenean valleys and the Basque country, which was the foundation of the later Christian reconquest.

There were other areas of the early Muslim world where independent principalities coexisted with the Muslim authorities, moderately peacefully - in the mountains of northern Iran, for example. The Christian principalities of mountainous Armenia were in a position not entirely dissimilar from that of the Christians of northern Spain. None of these, however, seriously threatened Muslim rule in the areas to the south. When Daylamite mountaineers from northern Iran conquered much of Iran and Iraq in the tenth century, they did so as Muslims, and they soon lost their identity among the wider Muslim populations. The Armenians maintained their independence but they never sought to make conquests beyond their traditional homelands. What distinguished the principalities of northern Spain was that they maintained, if only just, their Latin Christian high culture. At the same time they kept alive the memory of the Visigothic kingdom and the idea that the whole peninsula had once belonged to the Christians and should do so again. They also had access to and links with a much wider Christian polity to the north. These factors meant that, unlike the northern Iranian or Armenian principalities, the Christians of Spain came to be a serious long-term threat to Muslim control, until eventually, 800 years later, they finally drove them out.

The ambitions of the Arabs did not end with the Pyrenees. Muslim forces were soon raiding up the Rhône valley and through the fertile lands of Aquitaine. Unfortunately we have only the briefest accounts of these adventurous campaigns. The course of the raids is often quite unclear. The Arabic sources are frequently just one-line reports and we have brief notes in some Latin monastic chronicles. This first encounter between the peoples of north-west Europe and the Muslims is shrouded in obscurity. The first raids are said to have been directed by Tāriq b. Ziyād and to have reached Avignon and Lyon before being defeated by Charles Martel.⁴⁴ The Muslim raiding parties always went around the eastern end of the Pyrenees: Barcelona, Girona and Narbonne all came under their control, though Muslim rule in Narbonne was shortlived and ephemeral. Later Arabic sources allege that Mūsā b. Nusayr had conceived the massively daring and ambitious plan of marching his armies through the whole of Europe and the Byzantine Empire back to Syria.⁴⁵ Sometimes they must have felt that they were unstoppable.

They were not always successful. In the summer of 721 the governor ^p of Andalus led a raid into Aquitaine but the duke, Eudes, fortified himself in Toulouse. In a sharp conflict on 9 June, the Arabs were driven back and the governor himself killed. In 725 the Arabs launched the most ambitious raid so far. They began with the Roman and Visigothic fortress of Carcassonne, which they took by storm. They then moved east through the Midi. Nimes surrendered peacefully, giving up hostages who were sent behind the lines to Barcelona. The governor^q then led his men on a lightning raid up the Rhône valley, encountering little serious resistance. The army reached deep into the heart of Burgundy, taking Autun, which they pillaged thoroughly before returning to the south.

The climax of the Arab invasions of France came with the conflict generally known as the battle of Poitiers.⁴⁶ Since the late eighth century this battle has acquired a symbolic fame, marking the point when the Arab advance into western Europe was finally brought to an end by the Carolingian warlord Charles Martel. Within a couple of years Bede, in distant Northumbria, had heard of it and felt able to say with confidence that ‘the Saracens who had devastated Gaul were punished for their perfidy’. Gibbon, in one of his more elegant flights of fantasy, allowed himself to speculate about what might have happened if the fortunes of battle had been different.⁴⁷

A victorious line of march had been prolonged above a thousand miles from the rock of Gibraltar to the banks of the Loire; the repetition of an equal space would have carried the Saracens to the confines of Poland and the Highlands of Scotland: the Rhine is not more impassable than the Nile or the Euphrates, and the Arabian fleet might have sailed without a naval combat into the mouth of the Thames. Perhaps the interpretation of the Koran would now be taught in the schools of Oxford, and her pulpits might demonstrate to a circumcised people the sanctity and truth of the revelation of Mahomet.

In 1915 Edward Creasy, in an influential work of popular history, included it as one of his ‘Fifteen Decisive Battles of the World’. In reality it does mark something of a watershed. Until this point the Muslim armies had raided far and wide in France, even if they had not made any permanent conquests. As the people of Central Asia were finding out at exactly this time, Arab raids could be the prelude to more lasting conquest. After this point, Arab military activity was largely confined to the area around Narbonne, and al-Andalus began the transformation from a jihādist state to a more settled government.

For Western military historians the battle of Poitiers has acquired a further significance. It has been argued that Charles Martel was successful because, for the first time, he used the heavily armoured mounted warriors, the knights, in a coordinated charge that destroyed the enemy. According to these theories, this marked the beginning of the dominance of the battlefield by the heavily armoured horsemen which became characteristic of western Europe in the Middle Ages. With the rise of the knight there came the emergence of feudalism as the characteristic form of fiscal and social control.

It is all the more frustrating therefore that our information on what actually happened is short and confused, and even the date of the conflict is uncertain, though the traditional date of Saturday, 25 October 732 is as likely to be as right as any other.⁴⁸ The earliest important account is given in the Christian Chronicle of 754. Writing no more than twenty years after the events, the chronicler seems to have been fairly well informed, probably by Muslim survivors of the expedition who had returned to Cordova. He describes how the governor, Abd al-Rahmān al-Ghāfiqī, first defeated a Muslim rebel, Munnuza, in the mountains of the eastern Pyrenees. Munnuza had sought support from Duke Eudes of Aquitaine and Abd al-Rahmān now went in pursuit of him. He caught up with the duke and defeated him on the banks of the Garonne.

Abd al-Rahmān then determined to continue the pursuit. He sacked Bordeaux and burned the famous church of St Hilary at Poitiers. He then decided to go on north along the Roman road to despoil the great church of St Martin at Tours on the Loire. While he was on the road from Poitiers to Tours he was confronted by Charles Martel, ‘a man who had proved himself a warrior from his youth and an expert in things military, who had been summoned by Eudes’. The two armies probably met at a small town still known as Moussais la Bataille.

After each side had tormented the other for almost seven days with raids, they finally prepared their battle lines and fought fiercely. The northern peoples remained immobile like a wall, holding together like a glacier in the cold regions, and in the blink of an eye, annihilated the Arabs with the sword. The people of Austrasia [that is, the followers of Charles Martel], greater in number of soldiers and formidably armed, killed the king Abd al-Rahmān, when they found him, striking him on the chest. But suddenly, within sight of the countless tents of the Arabs, the Franks despicably put up their swords, saving themselves to fight the next day since night had fallen during the battle. Rising from their own camp at dawn, the Europeans saw the tents of the Arabs all arranged according to their canopies, just as the camp had been set up before. Not knowing that they were all empty and thinking that inside them were phalanxes of Saracens ready for battle, they sent scouts to reconnoitre and discovered that all the troops of the Ishmaelites had left. They had all fled silently by night in tight formation, returning to their own country. But the Europeans, worried lest the Saracens deceitfully attempt to ambush them in hidden paths were slow to react and searched in vain everywhere around. Having no intention of pursuing the Saracens, they took the spoils and the booty, which they divided up fairly, back to their country and were overjoyed.

The accounts are not nearly as detailed as we would like, but certain things do emerge clearly. The first is that this was no cavalry battle. The author of the Chronicle of 754, with his image of the glacier, suggests strongly that the Franks fought on foot as a sort of phalanx. He also makes it clear that they were very disciplined. The failure to follow up victory by pursuing the enemy that night is evidence not of cowardice but of the need for discipline and the dangers of chasing an enemy in the dark through unknown country. Most of the Arabs may have saved their lives, but they certainly abandoned their tents and much of their military equipment.

The defeat of the Muslims at Poitiers effectively marked the end of large-scale raiding in France. It became clear that they were not going to the able to conquer the country, or even to continue raiding with any degree of success. The military prowess of the Franks, like the ‘northern glacier’, was only one of the reasons for the end of expansion. The Muslims were probably short of manpower. The North African conquests had been made possible because large numbers of Berbers had joined the Muslim armies; these same Berbers had formed a major contingent in the armies that invaded al-Andalus. There are no reliable reports of Franks or other inhabitants of France joining the invading armies. Perhaps they were too alien to allow easy cooperation, perhaps their presence was always too transitory to inspire confidence, but whatever the reason, the lack of local support left the Muslim armies very isolated and vulnerable.

But perhaps the most important reason for the change was the great Berber rebellion in North Africa in 741. The brutalities of the slave trade had caused massive resentment throughout the Maghreb and the Berbers almost succeeded in driving the Arabs out altogether. Only the sending of a massive army from Syria restored Muslim control in the area. This great conflict meant that neither Berbers nor Arabs were able to spare manpower for extending the conquests further in the cold and unfriendly fields and forests of the north.

SPAIN AND PORTUGAL