Byzantine Anatolia quickly became the target of Muslim expeditions after the Byzantine evacuation of Syria and northern Mesopotamia in the midseventh century. Muslim historical traditions disagree on who led the earliest Arab raids through the mountain passes into Anatolia and the ‘land of the Romans’. Ibn ‘Abd al-Hakam’s Futuh misr reports the earliest Muslim expedition against Amorion in 644 (AH 23), when Constans II was still too young to be capable of developing his polity’s defences.11 These early raids penetrated deep into Byzantine territory. Mu‘awiya had already commanded an incursion into Asia Minor in 643, and he probably led another expedition against Amorion in 646. His expeditions disrupted the Anatolian interior, forcing the Byzantines into defensive countermeasures. The antecedents of later Muslim warfare and diplomacy can be traced back toMu‘awiya’s governorship of Syria, and to the period after he became caliph in 661.12 Summer raids (saifa) began from around 640. The raids of the 640s were often launched from Mesopotamian and Syrian towns such as Homs and Antioch, and the raiders entered Anatolia by way of passes such as the one at Hadath (between Germanikeia and Melitene) and the Cilician Gates (using bases such as Mopsuestia and Tarsus once these were in Arab hands). Whether or not Mu‘awiya himself went on the important early expedition against Amorion in 644,13 he led a number of other campaigns into Anatolia at a time when Byzantine resistance was beginning to harden. As Constans II tried to fortify Byzantine cities and strongholds and to develop a coherent resistance, Mu‘awiya gained experience in how to fight and to negotiate with the Byzantines, becoming familiar with the problems and challenges of their Anatolian terrain, climate and logistics. Probably no other caliph had as much personal military experience against the Byzantines as Mu‘awiya did. Despite this, Mu‘awiya’s offensives against Byzantium did not result in any lasting Muslim conquests in Anatolia between 643 and his death in 680.Muslim raids became an almost annual event, penetrating up to 1,000 kilometres into the Anatolian plateau. They were not restricted to summer, and a winter raid would sometimes follow hard upon a summer one.14 The raids contributed toMu‘awiya’s prestige, helping to enrich theMuslims and attracting ever more tribesmen to take part, while the Muslim casualties probably remained relatively modest. However, this persistent raiding seriously damaged the empire’s infrastructure: the Byzantines’ territories were devastated; they lost property and human lives; many were taken captive; and their commerce and agriculture were destroyed. The raids also kept Byzantium off-balance, forcing them onto the defensive and preventing them from launching major offensives of their own against Muslim Syria. Under Mu‘awiya’s able command, the Muslims were innovative and capable of taking their opponents by surprise. Despite a lack of any Arab seafaring tradition, they embarked on combined land and naval operations, highlighting their readiness to adopt new strategies and techniques of warfare. Although the literary evidence about naval expeditions against Byzantine-controlled islands such as Arwad and Rhodes is contradictory and impossible to verify, epigraphic evidence confirms an intensification of Muslim military and naval activities, including devastating Muslim raids on Byzantine Cyprus in 649 and 650 at Soloi.15 Some 120,000 Cypriots are said to have been deported, marking a serious change in the island’s fortunes. Although the Byzantines did receive forewarning of the Arabs’ preparations for some of these naval expeditions, the raids further reduced Byzantine resources and naval capabilities in the Mediterranean, jeopardising yet more of the empire’s coastline and islands. Muslim power was proving capable of extracting financial concessions from regions hitherto beyond its reach. In 654 or 655, Mu‘awiya’s naval forces decisively defeated Constans II at the battle of Phoenix off the south-western Anatolian coast, also known as the ‘battle of the masts’. The late seventh-century ascetic Anastasius of Sinai testified to the shock of this Byzantine naval defeat.16 Ibn Abi Sarh, governor of Egypt, commanded the Arab fleet, and its crew members may have included many Christian Egyptians. The Muslims then mounted a threatening but brief and abortive combined land-and-sea operation, reaching almost to Constantinople itself.17 The seventh-century Muslim naval offensives culminated in their costly and disastrous assault and naval blockade of Constantinople from 674 to 678. The Arabs failed to plan adequately and they also encountered a new Byzantine weapon, Greek fire, which devastated their warships and inflicted heavy casualties (see above, p. 233). Mu‘awiya’s governorship of Syria and his caliphate extended Arab territorial control, with Cyprus and most of Armenia falling underMuslim influence. The period of Mu‘awiya’s ascendancy also saw larger-scale expansion inNorth Africa, asMuslim military pressure on Anatolia reduced theByzantine government’s ability to reinforce and defend vulnerable positions in the empire’s western approaches.Mu‘awiya’s prestige derived from his military victories; from the fact that he received recognition from the Byzantine emperor; from his control of the holy places of Christianity and Islam; from the line of successors from the ProphetMuhammad; and also from the messianic attributes of his leadership. His aggressive, risky and unpredictable strategies challenged a number of arrogant assumptions of the Byzantines about the Arabs: as with their earlier stereotypes about the Persians, they assumed that the Arabs could not fight in cold weather and instead became lethargic.18 Muslim winter expeditions into Anatolia brought home to the Byzantines just how wrong their notions about Muslim warfare could be. Winter campaigns were costly to both sides; it was gambling with the lives of Muslim soldiers to keep them in such a totally hostile environment for long periods, but they unquestionably disrupted the Byzantines’ way of life in Asia Minor and kept them on the defensive. So why did rapid Muslim expansion not continue northwards under Mu‘awiya, in the wake of the extraordinary early successes? Although no explicit treatise outlining Mu‘awiya’s strategy exists, the failure does not seem to lie in flawed tactics, nor can it be put down to Byzantine policy or military victories. The harsh Anatolian climate and terrain played their part, as did the sheer logistical complexity of mounting lengthy, long-distance raids into the interior. The Muslims encountered tougher resistance, the closer they penetrated to the rather more ethnically and religiously homogeneous core areas of the empire; there could be no realistic expectation of winning over many converts to Islam there. Resources for potential Arab expansion were also squandered on the ill-fated naval siege of Constantinople itself (see above, pp. 232–3) and, above all, on the first and secondMuslim civil wars.Muslim leaders started to see the sense in exploring temporary arrangements with the Byzantines, rather than engaging in perpetual warfare. Another complication worked to the Byzantines’ advantage and helps to explain the caliphate’s reluctance and inability to provide whole-hearted commitment to invading and fully subjugating Anatolia: the Arab incursions were frequently undermined by rivalry and envy among their leaders. One of the most daring Muslim commanders, Khalid bin al-Walid, was much admired ‘because of his usefulness to the Muslims in Byzantine territory, as well as his bravery’.19 However, his fame and success appeared to threaten other military leaders, and the caliph himself allegedly contrived al-Walid’s poisoning in 666/7, on his return to Homs after a raid into Anatolia: Mu‘awiya feared his growing prestige among other Syrian Arabs. Although al-Walid’s death may perhaps signal other problems, including tensions betweenMuslims and Christians atHoms, the reports of his death there highlight the rivalries and tensions among Muslim commanders. As Byzantine intelligence on Muslim strategy and tactics improved, so did their response to Muslim aggression. Byzantine resistance began to take shape, notably during the reign of Constans II, who inherited sole rule at the age of eleven in 641.20 Constans faced various hurdles, including factional and dynastic infighting, the need to justify his authority and policies, and internal military strife.21 Yet his military and diplomatic activities in Anatolia between 641 and 663, and in the central and western Mediterranean between 663 and his assassination in Syracuse in 668, present us with something of a riddle. The last memories of military victories – especially in the east – were those of Constans II’s grandfather,Heraclius, who had personally risked his life and reputation in campaigning, even if his efforts against the Muslims had failed catastrophically. Constanswent out on campaign in Armenia and in Anatolia, and this pleased his troops. His Armenian campaign of 652/3 was an unsuccessful attempt to restore his claim to authority, faced with the prospect of the Armenians becoming clients of the caliph, although virtually no Armenian conversions to Islam took place at this time (see above, p. 342). The emperor’s presence in person was needed to make the military system work, as would still be the case many hundreds of years later.22 This obviated the risk of disobedience or incompetence on the part of his commanders, but it was not always practicably possible. Echoes of earlierHeraclian accusations of betrayal resound in the accusations made by Constans II’s courtiers’ against Maximus the Confessor and PopeMartin I (649–55).23 It was difficult for his officials to explainByzantine disasters at Arab hands. Heraclius and Constans both resorted to public accusations, ridicule and the denunciation of those whom they charged with harming the empire and the dynasty. When Constans received a letter from Caliph ‘Uthman (644–56), summoning him to Islam and proposing that he become the caliph’s subject, his reaction was to have it deposited on the altar of St Sophia and to invoke a passage from Isaiah.24 Here Constans acted as both head of state and mediator to the deity. Constans II ruled at a time when the balance of military power between empire and caliphate was fundamentally unfavourable to the former and when a Byzantine collapse was not out of the question. This obliged him to enter into diplomatic relations with theMuslims in the form of embassies.25 In 650 theMuslim commander Busr bin Abi Artat led a raid into Isauria and netted 5,000 prisoners. Constans requested and received a two- or threeyear truce in return for his payment of tribute. However, with the impact of the Arab conquests of Syria, Palestine, Egypt and upper Mesopotamia still reverberating in imperial circles, Constantinople remained highly suspicious of anyone who made – or who was in a position to make – unauthorised, local contacts with Muslims, whether commander or churchman.26 The imperial administration was equally suspicious of anyone who dissented from imperial policy, whether civil or religious, and Constans II’s attempts to censure PopeMartin I for unauthorised contacts with the Arabs are comparable with those of his grandfatherHeraclius: he, too, had tried to prevent unapproved negotiations between local leaders and Muslim commanders. By the ninth century, however, it would be impractical to enforce such rigid policies along the border. Although not providing the only explanation, the fitnas were as important a factor in theMuslims’ inability to crush the Byzantine empire in the Umayyad period as was Byzantine institutional restructuring.27 The first Muslim fitna, fought betweenMu‘awiya and ‘Ali, son-in-law of the prophet Muhammad, from 656 to 661, forced Mu‘awiya to purchase an expensive temporary peace with Byzantium in 657,28 and he had to keep this until his decisive victory over ‘Ali in 661 or 662.Only then wasMu‘awiya free to turn his and his armies’ attention to the situation on the northern approaches of Syria, although even then the Kharijite rebellion remained a formidable problem for them. According to Ibn Sa‘d’s Kitab al-tabaqat al-kabir, a Muslim army first established its winter quarters in Anatolia, ‘in the land of the Romans’ (’ard al-R¯um) in 662/3,29 but Ibn Sa‘d does not identify the expedition’s leaders, the number of raiders or their provenance, nor exactly where they wintered. Arab winterings in Byzantine Anatolia were more perilous for local life and disruptive to agriculture than were their summer raids. But they were also risky for the Arabs,30 for they prompted the Byzantines to strengthen their defences in Asia Minor.31 It is noteworthy that the earliest references to some form of thematic units in Byzantine Anatolia occur only a few years after the initial Muslim winterings there, whether or not these units as yet had any of the social or economic ties with particular areas that they would eventually form (see above, pp. 239–41, 266–7). The campaign theatre of Anatolia does not seem to have been a priority for the early Muslim historians. Those records which do survive come from Iraq, an area from which relatively few raids into Anatolia originated, because of the formidable logistical hurdles such as distance, heat and supplies. 32 We have no source-material comparable to the extensive narratives on other regions; this may simply not have survived, or details of the Anatolian conquests were either unavailable, or deemed unworthy of historical attention by al-Tabari and other later historians.33 The brevity of allusions in the extant Muslim histories to seventh-century raids into Anatolia may well owe at least something to the following considerations. Firstly, many raids started out from Homs or points further north, where there were few Muslim scholars in the mid- and late-seventh century. The surviving raiders were probably often some distance away from historical writers or their copyists who might have been in a position to record information about them and pass it on to later generations. Secondly, in Syria, Egypt, Iraq and Africa there eventually arose issues of tax and property rights which, although they might contaminate the source-materials, at least gave reason to put on record details about relations with the local inhabitants. There was no such incentive in the case of Anatolia, for it had not been conquered by the Muslims. A third possible reason for the lack of sourcematerial on the early Anatolian raids is that the motive for recording such expeditions was the pious commemoration of the names of participants, including those who perished, partly so as to add fame and distinction to their families, groups and clans back in Syria, Iraq and even in Egypt. But all that was needed for this purpose was the lists of their names and the dates – whether accurate or not – for those events. A fourth and final possible reason for the dearth of Muslim source-material about the early raids into Anatolia may be that it concentrates so heavily on the house of Mu‘awiya, the Umayyad caliph, celebrating its feats.34 The cessation of the first fitna was not the only factor behindMu‘awiya’s adoption of a more active approach towards Byzantine Anatolia.35 Another likely catalyst was Constans II’s departure for Italy and Sicily in 662/3 in an attempt to strengthen military defences in the west: this coincided with the ending of the fitna and the release of extensiveMuslim resources – both human and material – for offensives against the empire.36 The date for the first Muslim wintering was neither accidental nor random.37 Although the military situation in Anatolia worsened for the Byzantines after 663, the Arabs failed to establish any permanent base north of the Taurus mountains. Indeed, the series of Muslim raids from that time onwards could even be seen as indirectly attesting the overall effectiveness of the Byzantine defensive system. However awkward the Muslim winter campaigns made the situation for the Byzantines in Anatolia, the raids were, from the Byzantines’ point of view, preferable to irreparable Muslim conquest. If Constans’ move westwards offered the Arabs an opportunity, the fates of Asia Minor and the more distant Mediterranean were now more closely intertwined. Exchanges betweenDamascus and Constantinople intensified during the mid- to late seventh century, and Muslim officials and military commanders were not infrequently switched between Anatolia and North Africa. To take just one example, Fadhala bin ‘Ubayd was transferred from campaigning in the east to join Ruwayfi bin Thabit al-’Ansari in the major raid on the North African island of Jerba; this raid probably occurred in 677/8.38 Constans II lacked the skills that Heraclius had shown in exploiting his domestic and Persian enemies’ internal strife, and it was internal discord that ultimately overwhelmed Constans and led to his murder. He also lacked his grandfather’s skills in identifying external enemies’ weak points and then applying pressure to them, and he did not have his sense of timing in battle: Constans was able neither to divide theMuslims, nor to decapitate or neutralise their leadership.