Chapter 3

Conquerors

Emboldened by the victory at Civitate, the Normans of Apulia and Aversa began to expand their dominions. In 1054 Count Humphrey of Hauteville captured Conversano, a town 27km south-east of Bari. Troia, which Paul Oldfield notes had become ‘one of the key urban centres in inland Apulia’, agreed to pay tribute. Bari, Trani, Venosa, Otranto, and Acerenza acknowledged Hauteville authority, although Loud asserts that ‘only Venosa, and perhaps, Acerenza, were actually in Norman hands’. Benevento, a papal possession, was also besieged, but without success. In reference to otherwise unattested events after the battle at Montepeloso in September 1041, Malaterra insisted that the Normans possessed ‘most skilled’ engineers, capable of building ‘siege engines’. The reverse, however, seems to be true: the Normans’ inability to take the larger fortified towns can be explained by their lack of access to siege engines. An important aspect of siege warfare was psychological. Indubitably, fortified towns could be forced to capitulate by surrounding them with walls and ditches (circumvallation), in addition to the seizure of livestock and destruction of crops. But in addition to providing the ability to storm a fortified town’s walls and gates by clearing its battlements of defenders, stonethrowing machines instilled great fear in the inhabitants, whose homes and loved ones could be destroyed by artillery shot. Hence, siege weaponry added an all-important psychological element that, when used in conjunction with circumvallation, could ensure the desired outcome more swiftly. The defiant resistance of Reggio’s citizens metamorphosed into speedy surrender when engines were prepared before their walls (1060) – as did that of the people of Noto (1088) and Capua (1098) – and the inhabitants of Gerace surrendered (1062) when Roger Hauteville threatened to destroy their crops and deploy siege engines against their walls. However, while this strategy could be effective when it came to inland strongholds, towns and cities on the coast were a much greater challenge. Firstly, most (e.g. Bari, Trani, Siponto, Salerno and Naples) could not be completely surrounded given their access to the sea via fortified ports. As long as the walls held and the inhabitants could be provisioned by sea, they could potentially hold out indefinitely. Indeed, as will be noted throughout this chapter, the greatest coastal strongholds of the mainland and Sicily could not be taken by the Normans until they had both siege engines and ships at their disposal (although Reggio was an exception to this rule, for Guiscard did not possess a fleet until after its capture). In order to deny succour by sea, a sizeable fleet was necessary to blockade coastal cities. Once the Hauteville scions had a navy, the success rate of their sieges improved exponentially. That being said, naval blockades did not always guarantee success: Naples, jointly besieged by Guiscard and Richard of Capua, proved to be immune to a full-scale, lengthy investment by land and sea in the late 1070s. Siege engines were not deployed at Naples – nor were they used at Salerno (captured by means of treachery from within after eight months) – and perhaps their absence could be viewed as the ingredient missing from this unsuccessful investment.¹

Guiscard had already experienced the same problem in Calabria. ‘Not able to seize any fortified town or city’, he had resorted to stratagems such as seizing Peter of Bisignano in order to ransom him. On most other occasions scorched-earth tactics were employed. Given that the total number of troops available in Calabria to Guiscard c. 1046–57 could be numbered in the hundreds rather than thousands, even the tactic of circumvallation was presumably beyond his means when it came to the larger fortified towns. Conversely, Count Richard Quarrel of Capua evidently possessed an army of significant size, for he circumvallated the city of Capua twice in the 1050s. On the first occasion, Amatus related that Richard’s men erected three siege forts – presumably linked together by earthen ramparts – before the walls of Capua in c. 1052. With access denied to their crops, the Capuans were forced to launch desperate sallies. Although casualties were recorded on both sides, the Lombards were unable to break the deadlock. Having tried the military solution, the Capuans turned to the monetary alternative: upon receipt of 7,000 Byzantine gold coins, Richard abandoned the siege. He returned in 1057, probably in September/October as Amatus related that circumvallation prevented the citizens from harvesting their grapes. The Capuans again offered stout resistance until their food stores were depleted, forcing them to surrender and recognize Richard as their prince. As the Capuans surrendered no earlier than May the following year, the siege may have lasted for up to eight months.²

While the Capuans submitted the city to their new Norman prince, they did not surrender its walls and gates. Richard seems to have been happy enough with this arrangement – the princely title and the revenues that came with it were presumably agreeable for the time being – but his request for the city’s gates and towers to be handed over to him in 1062 was defiantly denied. The siege that followed is interesting for a number of reasons: Amatus’ account of it contains the first explicit reference to Norman use of archers in Italy. Another first is that that some of these archers were equipped with crossbows – that is, four years before the first recorded use of them by the Normans of France. Given that Amatus’ text survives only in an early fourteenth-century translation, an objection could be raised that the anonymous translator used the term arbaleste anachronistically. However, writing in the 1070s, William of Poitiers mentioned infantry ‘armed with bows and crossbows’ at Hastings in 1066. Hence, there is nothing anachronistic about the similarly phrased ‘bow and crossbow’ present in the Old French translation. Lastly, the siege is also notable for the first explicit, or believable, mention of the use of stone-throwing engines by the Normans. That is, if Malaterra’s unconvincing reference to the use of siege engines at Montepeloso (1041) is discounted, in addition to his subsequent mention, also unrecorded by Amatus and William, of them being deployed against the walls of Civitate in 1053 (although the context suggests that, unlike the ones ‘used’ at Montepeloso, they were rams or ladders rather than catapults).³

Capua’s formidable walls were protected by archers and slingers. Logically, Richard began the siege by deploying his archers and crossbowmen, undoubtedly ordering them to clear the battlements of missile troops. Amatus recounted that practically the entire city assisted in the city’s defence: wives supplied their husbands with slingshot, and fathers armed their sons. A twelve-year old archer named Auxencie is said to have inflicted mortal wounds on various Normans before succumbing to one himself. As Richard’s archers were unable to clear the battlements, he deployed ‘various instruments and engines for hurling stones’ (tension- or torsion-powered mangonels, or perhaps more powerful Byzantine-style traction [rope-pulled] or hybrid [traction and gravity] engines, given the destruction reported). The parapets were targeted and destroyed, the walls pummelled, ‘and many buildings damaged’. Still the Capuans resisted: damaged masonry was repaired and, despite the circumvallation, supplies were clandestinely smuggled into the city by river at night. But the ingenious method of securing much-needed nourishment was soon discovered, and Richard brought it to an end by having his own concealed vessels capture the boats. Thwarted at every turn, the Capuans sent their archbishop to appeal to the western emperor. Yet the prelate was unable to secure military assistance, and so the Capuans had little choice but to give Richard the keys to their city on 21 May 1058. ⁴

Calabria

Guiscard’s chances of penetrating deeper into the heart of Calabria improved immeasurably when Humphrey of Hauteville died in 1057. Succeeding his halfbrother as count, Robert now had access to many more troops. Providing him with vitally important assistance was Roger, his brother from the same mother (Frédesende), who arrived from Normandy around the time of Guiscard’s accession. Gian Piero Givigliano notes that the strategy appears to have been focused on venturing further south along the via Popilia, the imperial road that, like the modern A3 which generally corresponds to it, extended all the way from Capua to Reggio. Situated either on or close to this road were the tribute-paying towns of Bisignano, Cosenza and Martirano. With these funds in addition to those now available to him as count, Guiscard was in a strong position. The results were immediate; a daring raid was launched into southern Calabria. Exhausted from the trek through the mountainous region of northern Calabria, Guiscard treated his army to two days of rest in the vicinity of Cosenza and Martirano. Robert’s weary soldiers took the opportunity to bathe at some hot springs enjoyed by visitors since antiquity, located at modern-day Lamezia Terme in the province of Catanzaro. Malaterra related that the downtime was also used to ‘reconnoitre’ the land, presumably to the south of Martirano given that this town marked the limit of Guiscard’s personal experience and knowledge of Calabrian topography. Suitably rested, the Normans abandoned the via Popilia, marching to Squillace (40km southeast of Martirano), a fortified, eastern coastal town established in the late sixth century. What happened at Squillace is not reported by Malaterra, who simply noted that from there the army circled the coastline until it arrived before the walls of Reggio, the Byzantine capital of Calabria. Yet ‘seeing that neither threats nor blandishments could sway the citizens of Reggio’, after three days Guiscard ordered a retreat. Marching northwards up the via Popilia, Malaterra observed that the fortified towns of Maida (104km north-east of Reggio), Nicastro (13km north-west of Maida), and Canalea (location unknown) surrendered. While these towns are simply said to have ‘made peace’, it seems that, like Bisignano, Cosenza, and Martirano before them, their citizens also agreed to pay tribute.⁵

This renewed phase of campaigning in Calabria seems to have been more concerned with determining what sort of resistance could be expected in the region, in addition to ascertaining how well fortified it was. If conquest is to be defined as the direct seizure and subsequent rule of towns and cities, then Robert, outside San Marco, had conquered coffers rather than fortified settlements. Like Richard when it came to Capua, Guiscard no doubt intended to follow up the tribute-receiving strategy with direct control. Largely, however, this next logical step of consolidation would be taken by his younger brother Roger, for Guiscard was recalled to Apulia to attend to a significant threat to his rule. Peter I, son of Ami I, having earlier disputed the comital succession in 1046, chose to challenge the authority of the new count. In defiance of Guiscard’s suzerainty, Peter seized Melfi. The people clearly knew what tactics Guiscard would employ on his return to Apulia; they accordingly pleaded with Peter to protect their yet-to-be harvested fields. He agreed, brokering a truce with the count. Nonetheless, a dispute between Peter and Robert soon broke out, probably in the autumn of 1057, and the former was forced to flee Melfi when its citizens made it clear that he was no longer welcome. Peter fled to Andria, the town he had established in the hinterland of Barletta in the 1040s. Amatus recorded that a battle was subsequently fought there, in which both sides suffered substantial casualties. Robert seems to have been the victor, however, judging by the fact Peter was said to have begged for forgiveness afterwards, which the pragmatic Guiscard granted.⁶

Reported to have been left in command of sixty cavalrymen, Roger of Hauteville meanwhile secured a vantage point in the mountains of Vibona (Vibo Valentia; 74km north-east of Reggio). Givigliano suggests Mesiano (10km south-west of Vibo Valentia), but a more likely location is either the elevated part of the modern city, or the nearby town of Bivona, both of which formed part of the ancient city of Hipponion. In addition to being closer to the via Popilia than Mesiano, both sites overlooked the Tyrrhenian Sea and the plain to its east, and their strategic value is attested by the fact that great fortresses were later built at both locations: respectively, the so-called castello normanno-svevo (thirteenth century), and the castello di Bivona (fourteenth century). Surveying the coast and hinterland of the Italian ‘toe’, Roger used his precipitous camp as a base to terrorize the region. The raids launched from this base were evidently effective, for Malaterra noted that ‘all cities and fortified settlements of this province and the entire Saline Valley’ surrendered.⁷ None of the places captured were named, but one of them must have been nearby Mileto [8km south-west of Vibo Valentia], for it was later granted to Roger by his brother in 1058. Oaths were sworn and Roger was showered with gifts. He duly sent some of the tribute exacted to his brother in Apulia, and then did something yet to be done in Calabria: in addition to adding towers and ramparts to the fortress Malaterra called ‘Nicefola’, he garrisoned it with cavalrymen. Nicefola’s precise location remained a mystery until Vera von Faulkenhausen identified it in 2000 as the Rocca Niceforo (Angitola) in Maierato (province of Vibo Valentia). Referred to in a twelfth-century papal source as ‘the stronghold of Calabria called Nicephorus’, the fortress was named after the tenth-century Byzantine general and emperor Nikephoros II Phokas (Niceforo in Italian). Whether Roger captured the Rocca Niceforo or found it abandoned is unknown, but it was more than likely intended to serve as an advance base from which Reggio could be captured. Receiving the surrender of ‘the entire Saline Valley’ (the plain of Gioia Tauro) was a critical step towards achieving this goal, for its principal towns, San Martino (di Taurianova) and Oppido (Mamertina), lay in relatively close proximity to the Calabrian capital (respectively, 40 and 36km to the north-east).⁸

As future events demonstrated that its people were more than a little warlike, the rapid capitulation of these areas of Calabria is somewhat surprising. However, as was the case in Longobardia/Apulia, imperial garrisons were small and no doubt denuded significantly by the empire’s current defence problems. The empire at the time of the incursions of the Hauteville brothers in Calabria was hard-pressed on its all-important northern and eastern frontiers. The former border, fixed at the Danube, was overrun in the late 1040s by Turkic horsemen originally from the central Asian steppe, known to the Byzantines as Patzinakitai (Tur. Peçenekler [‘Pechenegs’]). While subsequently defeated and settled as mercenaries near Serdica (Sofia) in Bulgaria, the Pechenegs later captured Preslav and raided the Thracian hinterland of Constantinople. After various pitched engagements, the Pechenegs were finally subdued in c. 1053. Ironically, while Normans were raiding deep into Byzantine Calabria, imperial forces were assisted by ‘Hervé, son of a Frank, at that time commander of the same people’ against the Pechenegs.⁹

In the east, another people of Turkic origin, the Selçuk Turks, launched debilitating raids deep into the easternmost province of Anatolia (Armenian theme), both prior to and after the disbandment of Byzantine troops in the region – estimated by Treadgold to have numbered up to 50,000 – in c. 1053. No doubt many continued to fight, but the various districts of Armenia were now required to provide money rather than soldiers. Given that Anatolia was the region where a great many of the empire’s troops were recruited, undoubtedly many from other provinces of the empire were recalled to defend it. Guillou observed that there was a division (tourma) of cavalry based at Oppido in the Saline Valley, which was subdivided into three subdivisions (vanda/droungoi), each under the command of counts. When reconciled with the assessment of varying Byzantine numerical strengths in the previous chapter, this contingent might have consisted of up to 1,000 men. However, McGeer observed that in the late tenth century a vandon, ‘the smallest tactical unity in the cavalry’, consisted of fifty men, and hence it is possible that the commander of Oppido had little more than 150 cavalry at his disposal. Significantly, McGeer’s figure is based on the military manual attributed to the man after whom the Rocca Niceforo was named, and a Calabrian chronicle recorded that various hilltop locations were fortified at the same emperor’s behest. Moreover, since Phokas reorganized the structure of the Italian provinces in c. 969, it is likely that his revised system of military arrangement was not only adopted by the katepano in Apulia, but also by his subordinate in Calabria. Whether the Saline cavalry were recalled to the imperial heartland cannot be determined, but later events suggest that some semi-professional, thematic troops were still present in the region.¹⁰

Roger returned to Apulia to meet with his brother and discuss plans for the capture of Reggio. In the autumn of 1057 Guiscard and Roger, with a ‘large army of cavalry and infantry’, marched on the Calabrian capital. When traversing the Saline Valley, word reached them that the city’s citizens had gathered anything consumable from Reggio’s hinterland in order to deny the Normans the ability to forage. Roger was ordered to plunder the area around Gerace – 53km north-east of Reggio – in order to procure the supplies necessary for what was anticipated to be a lengthy siege. The younger brother’s expedition was successful, and he soon afterwards arrived before the walls of Reggio, then under siege by Guiscard. But autumn had ended by this stage, and the army abandoned the investment and wintered at Maida (22km north-east of Rocca Niceforo [Maierato]).¹¹

The walls of Reggio had proven to be an insurmountable obstacle, and yet another problem arose: sibling rivalry. Roger seems to have felt that his elder brother was not as appreciative of his efforts as he should have been. Malaterra noted that without sufficient funds to pay them, Roger’s cavalrymen were becoming increasingly restive. Guiscard’s refusal to accept his brother’s request for additional funds with which to pay his troops led to a rebellion commanded by Roger and his brother William of Hauteville: count of the Principate – created (1056) within the principality of Salerno at Prince Gisulf II’s expense – and grandfather of Roger of Salerno, who would become the third ruler of the principality of Antioch (r. 1113–19). From William’s Calabrian fortress at Scalea (40km north-west of San Marco), Roger’s biographer Malaterra noted that his protagonist ‘made many attacks against Guiscard, harassing on every side’. Count Robert then marched on Scalea, engaging in the customary destruction of vineyards and olive groves. William’s response was to sally forth from the fortress of Scalea, inflicting substantial casualties on the besiegers. Guiscard was accordingly forced to abandon the siege, and the acrimonious dispute was resolved, for the time being, when he formally received Roger into his service. But the reforged brotherly bond was broken a mere two months later when Roger returned to Scalea and, with 100 cavalrymen, again raided Guiscard’s territories in both Calabria and Apulia.¹²

The spring of 1058, Malaterra recorded, was not a happy time for the Calabrians. A three-pronged scourge afflicted them, the first of which should not come as a surprise: ‘the Norman sword was raging, from which barely anything was spared’. The second was famine; the third, disease. Noting the chaotic events of the first of several Hauteville civil wars, the Calabrians ‘began to shake off the yoke of the Normans’. Tribute payments ceased, and military service (servitium) was refused. Passive acts of defiance were followed by violence: the sixty-strong Norman garrison at Nicastro was massacred. Moreover, unbeknown to (or unreported by) Malaterra, all was not well among the Byzantines in Calabria. For reasons unknown, the Calabrian commander Leon Thrymvos ‘caused the Scrivones to be slain in the city of Crotone’. In earlier centuries, Scrivones were men who formed a unit of imperial guards attached to, but separate from, the tagma of the ‘the Sentinels’ (Exkouvitoi). Whether this was still true in the eleventh century cannot be ascertained but, much as in Longobardia, they were probably officers rather than members of a regiment. Whatever the case, when these troops/officers were either killed in battle or executed, Leon was forced to return, or was recalled, to Constantinople owing to the unpopularity of the act; indeed, a revolt erupted. As a Byzantine source observed, the rebellion played into the hands of Guiscard, allowing him to advance unimpeded to Reggio.¹³

While the revolt in Crotone certainly assisted the Normans, the rebellions led by Byzantine garrisons and by Roger were a considerable challenge to the duke’s authority. Hence, ‘seeing Calabria consumed and Apulia thrown into complete confusion’, Guiscard decided to grant his brother territory in Calabria. Roger, Malaterra related, received all land south ‘of Mount “Nichifola” [Rocca Niceforo, Maierato] and Mount Squillace’. Evidently this land was to be held in fief (not owned and from which service was due), since the only hereditary possession mentioned was the fortified town of Mileto. Roger now controlled around half of Calabria in theory: much of it remained in Byzantine hands, and later events demonstrated that even when they were overcome, Guiscard failed to honour his side of the agreement. With the land grant easing tensions immeasurably, the Normans were now free to suppress the Calabrian revolt. It was vital that the rebellious towns of the Saline Valley be seized if the Normans were to stand any chance of besieging Reggio without the danger of being attacked in the rear. Oppido, which may have been defended by its commander and his cavalry detachments, was besieged by Roger in 1059. While the siege was underway, an army led by the unnamed commander of Gerace and the bishop of Cassano lay siege to San Martino (presumably garrisoned some time after its capture in c. 1057). Roger promptly abandoned the investment of Oppido, overcoming the besieging Byzantines at San Martino from the rear.¹⁴

This success seems to have chastened the inhabitants of Calabria for a time, but they remained in possession of strategically important towns such as Gerace, Squillace, Oppido and Reggio. However, now that the locals had switched from the offensive to the defensive, the Hauteville brothers decided that the time was ripe for the capture of the capital. In the summer of 1059, Guiscard ‘went plundering all the way to Reggio’ until the onset of winter.¹⁵ That is all Malaterra had to say about the expedition, but when compared with William of Apulia’s contemporary poem, the events of this expedition can be discerned. Guiscard was besieging Cariati on the east coast of Calabria prior to being invested with the ducal title by Pope Nicholas II at Melfi (August 1059). After receiving the title, the duke returned to Calabria ‘with many cavalrymen’, and shortly afterwards Cariati’s people surrendered.

From here he travelled to other places:

Powerful Rossano, then to Cosenza, mighty in arms.

The well-known resources of wealthy Gerace were also secured,

And nearly all of Calabria was made subject to him.¹⁶

Malaterra’s raid, then, was actually more of a campaign to secure the all-important supplies necessary for the impending siege of Reggio. Cariati and Rossano (28km to the north-west) were not mentioned by Malaterra; they were new acquisitions. As Cosenza was already paying tribute, the duke likely secured supplies, cash and perhaps some troops (as noted below, the province of Cosenza was well known for its soldiers, hence William’s ‘mighty in arms’). The citizens of Gerace, the hinterland of which Roger had seized supplies from for the failed siege of Reggio in 1057, agreed to pay tribute but were not burdened with a Norman garrison (as Malaterra later made clear). Gerace’s commander and presumably his troops had participated in the recent attack on San Martino, and hence needed to be brought in line. Moreover, as Roger had discovered two years earlier, the region’s great wealth (agricultural and fiscal) coupled with its close proximity to Reggio (53km south-west), meant that the fidelity of its people was essential if the Normans were to have any chance of sustaining a protracted siege of the capital.

Guiscard wintered in Apulia and Roger in Calabria before the brothers marched on Reggio. Malaterra’s chronology is often vague and contradictory, and given that he mentions the impending siege occurring after the winter, his assertion that it began in 1059 cannot be correct. Also, when his chronology is reconciled with William of Apulia’s, the preceding plundering expedition occurred during the late summer or early autumn of 1059, and hence the subsequent siege of Reggio feasibly commenced in 1060. While normally being specific about the crops destroyed or harvested, on this occasion he simply mentioned that the siege began during the harvest. Accordingly, if the people of Reggio were busy harvesting wheat, then the investment started in July/August of 1060 and, if they were picking grapes, in September/October of the same year. The citizens of Reggio resisted the siege with great determination, but chose to capitulate when they witnessed ‘siege engines’ being readied for use. Unfortunately Malaterra did not relate whether these engines were towers, ladders, rams, or catapults, but in any case they must have been formidable as the well-fortified city was said to have surrendered before the machines were deployed. The first credible mention of siege equipment being used by the Normans in Italy is something of a turning point, and while indubitably such technology was known in France, it was more than likely made available to them by subjected Byzantine engineers (perhaps from Cosenza, later noted to have been protected by catapults).¹⁷

Malaterra recorded vaguely that two men ‘who seemed to rule the others’ were allowed to leave Reggio with their men. One of these leading men could have been the strategos of Calabria, although the title of duke first granted to Argyros had either decommissioned or considerably lessened the power of this office. The commanders and their men withdrew to the fortress at Squillace on the east coast, where they prepared for the inevitable siege. While Guiscard rested at Reggio, Roger turned his attention to capturing the twelve remaining fortified towns and fortresses in Byzantine hands. Only one of them was named by Malaterra (Squillace), but included among them must have been Oppido (siege abandoned in 1058), and Mesiano (unmentioned until c. 1062). The allegiance of the unnamed eleven towns was gained variously by the use of threats and promises of future rewards. The heavily fortified town of Squillace, effectively now the Byzantine capital of Calabria, proved to be a harder nut to crack. Initial efforts to take it by force failed, so Roger erected a siege fort in front of its gate. Disbanding the rest of the army, the fort was manned with cavalry, whose presence prevented access to and from Squillace. With the rest of Calabria in Norman hands, the Byzantine commanders had little hope of receiving supplies and reinforcements. Accordingly, after an unspecified amount of time had elapsed, they chose to leave for Constantinople by ship. Now leaderless, the citizens of Squillace decided to come to terms with Roger. By the end of the year 1060, Calabria was entirely subject to Hauteville dominion, although fewer than half of its towns seem to have possessed garrisons. Malaterra concluded his account of the siege of Squillace with the comment that all of Calabria was now at peace. This would, however, be shortlived: both Normans and Calabrians ensured that warfare would continue to be a feature of the region for lengthy periods in both this decade and the one that followed it.¹⁸

Despite the fact that Guiscard had earlier granted Roger authority over the yet-to-be-conquered south of Calabria, for reasons unspecified he reneged on the agreement once the region had been captured. Without land that could bind his men to him through ties of service, up to this point Roger had little choice but to maintain his troops through cash payments and booty secured from raids and campaigns, a system not only employed widely by the Hauteville scions in Italy, but also in Syria (Tancred, 1104). While he may have established a small cadre of household troops at Mileto – the only fortified town in his undisputed possession – the majority (or all) of Roger’s soldiers required regular remuneration, and hence he was faced with a significant financial problem. Other than desiring to resolve this fiscal crisis, Roger also wanted to provide his bride Judith d’Évreux – a relative of William the Conqueror’s – with a dower equal to her distinguished birth. Undeterred by his brother’s refusal to honour the land grant, Roger appealed for the arbitration of the prominent lords in Apulia. When they chose not to rule in his favour, he hired cavalrymen and strengthened the fortifications of Mileto. Guiscard raised an army in 1062 and marched on Mileto, but discovered upon arrival that Roger was recovering from a serious illness at Gerace (40km south-east). Roger quickly gathered his men and set out for Mileto, in the hinterland of which he encountered his brother’s forces converging at the base of a hill Malaterra called Sant’Angelo. A battle ensued, during which Judith d’Évreux’s brother was killed. Roger’s forces did, however, manage to prevent Guiscard from establishing a base on the summit of Sant’Angelo or on the neighbouring hill called Monte Verde (although, since the abbey of Sant’Angelo [SS. Trinità] was later founded by Roger on Monte Verde, it is possible that Malaterra’s two hills were one and the same). The younger brother’s cavalry continued to assault the elder’s, forcing Guiscard to build two siege forts before the walls of Mileto, one in front and the other to the rear. Roger mounted daily assaults on the forts, choosing to attack the one not manned by his brother, and then passing through the middle of the town in order to attack the second one when Guiscard came to relieve the first.¹⁹

One evening, Roger rode to Gerace with a force of 1,000 cavalry and received the submission of its people. In response, Guiscard garrisoned the siege forts and marched with the rest of his army towards Gerace, encamping outside its walls. A prominent citizen called Basil invited Guiscard to lunch, and he entered the town hooded in order to prevent recognition. Nevertheless, word soon got out that the Norman was inside the city, and the citizens began arming themselves, fearing that a coup was in progress. Basil’s house was soon surrounded by an angry mob: both he and his wife Melita were killed, and the mob, now pacified by Gerace’s leading officials, allowed Guiscard to be taken into custody. Roger, who was not present during these events, promptly returned to Gerace to secure his brother’s release. Despite thanking the emissaries for the loyalty their people had shown, Malaterra stated that Roger suffixed his compliments with threats:

If you attempt to delay for a long time, you will witness the uprooting of your vineyards and olive groves. You will see that your city, when besieged by us, will offer no protection against our prepared siege engines.²⁰

Those who spoke with Roger were most concerned that Guiscard honour their earlier agreement: that is, no garrison would be installed. Roger assented to this request, which was subsequently sealed by Guiscard’s oath to that effect.

The sibling rivals were reconciled and Guiscard agreed to honour his previous agreement with Roger. However, evidently unhappy that Roger’s troops had seized the two siege fortresses outside Mileto, demolishing one and garrisoning the other, Guiscard reneged again. His anger did not abate when his younger brother subsequently surrendered the fortress and freed the garrison. Roger now resumed hostilities by seizing the fortress at Mesiano, situated 5km north-west of his headquarters at Mileto. The seizure was enough to force Guiscard to reconsider his position, which he changed in his brother’s favour soon afterwards. Roger, now a count, was in formal control of southern Calabria, although Malaterra later made it clear that the younger brother was only granted a half share in the various towns and fortresses entrusted to him. The count, whose troops were lacking in arms and horses on account of the recent conflict, decided to exact tribute from the various towns now under his control. Since the people of Gerace had been involved in the siege of San Martino (1059) in addition to imprisoning his brother, Roger, to keep a watchful eye on them, began construction of a fort close to the section of the town belonging to him. The town’s officials objected, stating that the count was violating the oath sworn recently by the duke. Roger begged to differ, reminding them that he was not bound by an oath he did not swear, and that his brother’s agreement remained in force in the half of Gerace subject to him. The citizens’ desire to retain a degree of autonomy, however, did not waver: the count agreed to stop construction of the fortress in return for a cash payment.²¹

Calabria was relatively peaceful for the next two years. Since the beginning of the Hauteville forays into Sicily (1061), the region was exploited for its cash, supplies, horses and men. In 1064 Guiscard ‘assembled a large army from Apulia and Calabria’ for use in another Sicilian campaign. Prior to the departure for Sicily, the brothers captured a fortress called ‘Rogel’ near Cosenza. The reason for the siege was not explained by Malaterra, but if his earlier contention that all of Calabria had submitted to the Normans by the end of 1060 was not disingenuous, then perhaps the inhabitants of ‘Rogel’ started what was to become a large-scale rebellion occurring ‘in the province of Cosenza’. Various locations mentioned correspond to the modern-day province, but that the rebellion was not restricted to this region is confirmed given that the next disturbance occurred in the modern province of Crotone (south-east). Having disbanded the army used for the Sicilian campaign, in early 1065 Guiscard razed the fortress of (Petilia) Policastro to the ground, repatriating its people to Nicotera (12km south-west of Mileto). He then marched on Aiello (22km south-west of Cosenza), whose residents were made to endure a four-month siege. The ‘warlike Aiellans’ sallied forth, launching devastating attacks on Robert’s forces with bows, slings, and javelins (see Appendix). Frustrated by the unrelenting hail of missiles, Guiscard’s cavalrymen ‘rushed forth, launching a charge into the midst of the enemy, pressing where they were most packed’, but were repulsed with some notable casualties. Among those killed were two men identified as members of Guiscard’s inner circle. One of them, Roger, was ‘pierced by a javelin’, the impetus of which dislodged him from his horse. The other was Roger’s distressed nephew Gilbert, who was killed in the process of trying to assist his wounded uncle. Greatly upset by the loss of these two men, Guiscard arranged for them to be buried at Santa Eufemia, the monastery he founded in 1062 with Judith d’Évreux’s half-brother, Robert of Grandmaisnil.²²

Based on Malaterra’s description, the force which successfully assailed the Normans at Aiello was a skilled conglomerate of what had been known in Greek since antiquity as slingers (psiloi/sphendonetai), javelineers (peltastai/akontistai), and archers (toxotai). While at first glance javelineers might seem to be ineffective against heavy cavalry, their lighter armour gave them greater mobility, and the ability to hurl javelins with deadly force, as well as parry incoming blows with their shields, made them invaluable. As advised in the tenth-century military manual attributed to Nikephoros II Phokas, javelinarmed infantrymen were meant to ‘engage with opposing cavalry and handle them’ and, as will be discussed in the next chapter, they were later deployed by Alexios I Komnenos in the same manner at Dyrrachion in 1081. Since the archers and slingers of Aiello were unsuited to mêlée combat, troops armed with javelins, while tactically different, were nevertheless able to achieve the goal of heavy infantrymen (whose tactical aim, as Raoul de Caen put it, was ‘to repulse a charge’). The duke clearly respected the martial adroitness of Aiello’s troops, for despite wishing to punish them for the killing of his friends, he instead chose to make peace with them.²³

Following the submission of Aiello, the region around Cosenza, ‘mighty in arms’, remained quiet for a decade. There was, however, one further revolt further south at Stilo (83km north-east of Reggio). On Christmas Day (c. 1071), Constantine Kondomikitas, one of Guiscard’s soldiers stationed at Isola (di Capo Rizzuto) near Crotone, removed his kinsman Constantine Pelogas – a fellow Byzantine appointed by Guiscard earlier as governor of Stilo – from power. The consequences of this coup, wrote Malaterra, were dire:

As a result, for almost six years insurgents throughout various regions of Calabria created much disturbance for the duke.²⁴

Unfortunately Malaterra provided little detail on the various events of the Calabrian rebellions, choosing instead to devote most of his narrative to the ongoing, and increasingly successful, campaigns in Sicily. Accordingly, a coherent account of the events in Calabria cannot be provided. However, based on what little information was given, it seems that the revolt at Stilo inspired another one at Rossano. Emphasis should be placed on seems, for Malaterra recorded that, much to the great consternation of its inhabitants, Guiscard in 1072 established a keep or donjon in the town. No reason was given for the construction of the donjon within Rossano’s walls, but it is more than likely that it was built following the suppression of a revolt. In addition to the rebellions led by Greek-speakers were the ones actioned by those who spoke French. Abelard, son of Guiscard’s predecessor Count Humphrey, having already participated in a rebellion against the duke in 1067–68, began raiding the duke’s Calabrian possessions in 1072. Another participant in the earlier rebellion, Robert Arenga, who had similarly been granted land in Calabria when coming to terms with the duke in 1068, followed suit from his base at Castrovillari (58km north of Cosenza). Abelard launched his attacks from Santa Severina, an impregnable hilltop town 20km north-west of Crotone. Count Roger returned from Sicily to lay siege to Abelard’s fortress, and was later joined by Guiscard. At around the same time, Guiscard’s son Roger Borsa was tasked with the investment of Castrovillari.²⁵

Dating the events of the siege is difficult, primarily because Malaterra stated that the duke arrived after the capture of Salerno in 1076, whereas Amatus, who does not mention the count’s participation, indicated c. 1073–75. Whatever the case, the siege lasted for over two years. A major problem afflicting the inhabitants of any besieged fortress, however impregnable, was access to provisions. Abelard was able to circumvent this challenge by waiting for the enemy troops to finish foraging, then seizing the fruits of their endeavours. This tactic seems to have been successful for quite some time: that is, until Guiscard established three siege forts at strategic points around Santa Severina. According to Amatus, these forts were linked by ditches and palisades. Despite now being deprived of the ability to secure supplies from enemy foragers, Abelard refused to capitulate. The frustrated Guiscard then abandoned the siege, leaving garrisons in the siege forts. He later returned, and was eventually able to bring about Abelard’s submission in c. 1075, although two different reasons were given for it: either the town was surrendered in return for the release of Abelard’s brother Herman, whom Guiscard had recently imprisoned (Malaterra); or, moved by the pleading of Santa Severina’s undernourished inhabitants, the disgruntled nephew chose to come to terms with his uncle. Guiscard then joined his son at Castrovillari. Evidently Robert Arenga had managed to withstand Roger Borsa’s siege by utilising the same tactic as Abelard, but Guiscard put a stop to it with the same solution. In addition to the fact that Castrovillari’s people were starving, upon hearing of Santa Severina’s capitulation Robert Arenga could do little but surrender.²⁶

With the Calabrians up in arms, and the Normans at each others’ throats, the time was ripe for the Muslims to strike. In 1074, the west-coast town of Nicotera was not only captured, but razed to the ground by forces commanded by relatives of the Zirid amīr of Ifrīqiya (Tunisia), Tamīm bin al-Mu‘izz bin Bādīs (r. 1062–1108). The Muslims seem to have been consistently well informed about political developments in Calabria, for both this sack of Nicotera and the subsequent one (1085) occurred when members of the Hauteville family were waging war against each other.²⁷

Malaterra’s rebellion of ‘almost six years’ evidently ended with Abelard’s surrender of Santa Severina, although it may have well continued until 1079, the year in which William of Apulia recorded yet another uprising in the region of Cosenza. Guiscard suppressed the revolt himself, returning to Apulia with a force of ‘excellent infantry’ gathered from the region. Some of these high-quality infantrymen, like those from Aiello, were likely to have been javelineers since they were valued by Guiscard for the customary agility the Byzantines associated with this troop type: ‘The duke chose some men from Cosenza, whom he recognized took the lead in speed’. The warlike troops from the province of Cosenza, despite being recognized as a great military asset, nonetheless continued to pose a threat to Norman control of central Calabria. Another revolt, this time at the expense of Guiscard’s son and successor, was launched. When the insurrection began is difficult to pinpoint, for Malaterra commented that by the time Duke Roger Borsa organized an army to suppress it in May 1092 (not 1091 as stated), the rebellion had been raging ‘for a long time’. Given that the count had spent three days scorching Cosenza’s olive groves and vineyards in the May/June of the previous year (before sailing for Malta in July), the insurrection must have existed for at least twelve months. The seriousness of the revolt was indicated by the forces gathered to suppress it. Not only did the duke summon the military ‘aid’ of his brother Marc, better known as Bohemond, but also of his uncle Count Roger, who dutifully arrived from Sicily with ‘many thousands of Muslims … and plenty of cavalry’. Preparing for a lengthy siege, the inhabitants of Cosenza duly gathered supplies, strengthened the battlements with palisades, and prepared their arms and armour. Once again Malaterra paid little attention to geography, merely stating that Duke Roger Borsa occupied the plain, while Count Roger attempted to invest the walls near the foothills. Based on this description, the duke probably approached Cosenza from the south, his uncle from the east. The count’s forces, however, were driven back by archers and projectiles hurled from catapults. He accordingly decided to enclose his chosen side of the city by establishing a network of siege forts linked together by palisades. While the circumvallation strategy was successfully deployed, missile troops still prevented the besiegers from advancing on the walls. However, despite the determined resistance of the people of Cosenza, they were forced to seek terms with the Rogers when it became clear to them that, without access to fresh supplies, their days were numbered. To ensure that a rebellion would never again foment in Cosenza, the duke ordered masons to build a fortress ‘on the city’s highest summit’. Thanking the count for his ‘service’, Duke Roger granted half the city to his uncle, who subsequently erected a donjon within it prior to leaving for Sicily.²⁸

Sicily

When Robert Guiscard received the ducal title from Pope Nicholas at Melfi in 1059, not only was he invested with Apulia and Calabria, but also Muslim Sicily. Like the Byzantines before them in 1038, the Normans realized that the capture of Messina in the island’s north-east (Val Dèmone) was crucial. Not only was the city in close proximity to Reggio, but also as the Byzantines had calculated before launching their earlier campaign, the Normans could hope for the assistance of those who formed the ethnic majority in Sicily’s north-east: Greek-speaking Christians. The count and duke were no doubt kept abreast of the politically fractious situation in Sicily by the Muslim community of Reggio, who quite understandably went to great lengths to prove to their new overlords that they should not be regarded as a potential fifth column. Suitably furnished with Muslim intelligence, soon after the surrender of Reggio in the summer/autumn of 1060, Count Roger led a reconnaissance expedition to Sicily. Crossing the Strait of Messina in what must have been Byzantine ships commandeered after the fall of Reggio, Roger reportedly disembarked with sixty cavalrymen. When arriving in the vicinity of the populous city of Messina, Roger’s force was confronted by a concerted attack by the city’s garrison. According to Malaterra, Count Roger ‘at first simulated fear, leading them far from the city’; he then turned and routed them with an irresistible charge. While there is no reason to doubt Roger’s feigned retreat – an extremely difficult tactic requiring great cohesion and discipline – it seems likely that the count’s victory was more than a little exaggerated for, as Stanton observes, ‘the fact is they were forced to withdraw’. Having ascertained what sort of resistance could be expected in the Val Dèmone, Roger returned to Reggio.²⁹

The increasing political disintegration in Sicily, which the Byzantines had taken advantage of in 1038, was related in the previous chapter. By 1052/3 the Kalbid dynasty had come to an end with the death of amīr Ḥasan, and the capital Palermo seems to have been subsequently governed by a council of elders. The rest of the island appears to have been divided among three rulers: Ibn al-Ḥawwās (Agrigento and Castrogiovanni), Ibn Mankūd (Marsala and Tràpani), and Ibn al-Maklātī (Catania). On these emirates, which more or less correspond to the island’s three regions – Val di Noto (south), Val di Mazara (west), and Val Dèmone (north-east) – unfortunately little is known. The first and last power blocs, however, appear to have been on friendly terms, for Al-Maklātī of Catania was married to al-Ḥawwās’ sister Maymūna. The balance of power in the island’s east was shattered when the ruler of Syracuse, Ibn al-Ṯumnah, not only defeated al-Maklātī in battle, but also seized his city and married his widow. Maymūna managed to escape to her brother at Castrogiovanni (hereafter called by its modern name, Enna), which her new husband subsequently besieged without success, incurring substantial losses in the process. Ibn al-Ṯumnah’s failure to depose Ḥawwās considerably weakened his position; he accordingly looked for allies. In February/March 1061 he met with Roger in Calabria and secured an alliance. While Ibn al-Aṯīr’s later claim that al-Ṯumnah promised to make Roger ‘lord of the island’ was highly exaggerated, the Normans must surely have been promised some land, which Robert ‘the Cunning’ presumably planned to use as a base from which the rest of the island could be conquered.³⁰

In March 1061 Roger launched his second Sicilian campaign, this time with 160 cavalry under the command of one of Guiscard’s best Norman lieutenants, Geoffrey Ridel. Guiscard ordered Geoffrey to pay heed to al-Ṯumnah’s counsel, suggesting that the latter may well have possessed overall command of the expedition (as Amatus later implied). The ensuing campaign is notable for the first specific reference to the presence of ‘arms-bearers’ or ‘squires’ in a field force. Chalandon, followed more recently by Stanton, held Ibn Khaldūn’s figure of 600 to be a more probable figure, with the suggestion that ‘the rest [440] consisted of squires and valets’. However, Ibn Khaldūn’s 600 is mentioned with specific reference to Enna (Fig. 8), a precipitous town which the Normans did not approach until their third campaign in May/June of the same year (which featured a similar figure of ‘seven hundred’). While he did not engage in a discussion of the presence of squires, Chalandon was nonetheless prudent to mention them. Admittedly, Malaterra only uses the singular form (armiger) in the pertinent passage, but that cavalrymen other than Roger were attended by arms-bearers is evident in a later passage, where mention was made of ‘the arms-bearing party who were carrying the plunder’. That a Norman cavalryman was commonly attended by a squire is confirmed in one of the Apulian annals: when besieging Brindisi in 1070, forty Normans were captured as well as ‘forty-three attendants’. Throughout the twelfth-century French-speaking world, squires were generally ‘knights in waiting’, and while normally tasked with the maintenance of equipment, horses, supplies and plunder, they were nonetheless blooded in battle as a sort of rite of passage to knighthood. A general parallel can perhaps be drawn here, but there is another explanation that better suits the south Italian context. That is, that these men may well have been, in nascent form, the servientes listed in the mid-twelfth century Catalogue of the Barons: landless men, as Matthew observes, ‘acting as attendants on knights and ready to serve with them’. Interestingly, the first use of this term is to be found in an eyewitness, anonymous chronicle of the First Crusade: ‘a Lombard serviens’ in the Hauteville contingent was mentioned in reference to the siege of Antioch (1098). But to return to the ‘arms-bearers’ indentified by Malaterra, two things can therefore be posited tentatively: firstly, his figure of 160 men on horseback can be doubled (and potentially all subsequent figures provided by him). Secondly, whether the theoretical 440 squires participated in battle cannot be determined with any certainty – they are only described as riding in relatively close proximity to their masters in battle – but it remains possible that they fought when needed.³¹

The fleet disembarked at the Punta del Faro (also known as Capo Peloro), the island’s north-eastern promontory named after the lighthouse (pharos) built by the Greeks in the classical period. The location had been of great strategic importance since antiquity: the Syracusans occupied it with a fleet and land forces to oppose the Athenian crossing from Reggio (425 BC), as did Sextus Pompeius against Octavian’s fleet commanded by Marcus Agrippa (36 BC). Unfortunately, the precise sequence of events after the landing is difficult to determine. While Malaterra stated that the army marched on Milazzo with the intention of plundering it, Amatus believed that Messina was instead subjected to a surprise attack. Confusingly, Malaterra suggested that it was actually the Normans who were the victims of a surprise attack: while advancing on Milazzo (27km north-west of Messina), they were intercepted by troops from Messina led by one of Ibn al-Ḥawwās’ relatives. As Roger was not wearing a hauberc at the time of the attack, the army was clearly on the march. Both sources agree that hostilities commenced during the evening, and while Malaterra recounted a skirmish, Amatus wrote of a lengthy pitched engagement which the Normans won decisively. Returning to Faro with the spoils gained from raids in the vicinity of Milazzo, Messina and Reggio, the Normans now attempted to return to Calabria. However, an immediate crossing was prevented by tempestuous weather. According to Malaterra, who may well have been referring to the same clash related by Amatus, a Muslim force of cavalry and infantry now chose to launch an attack. Roger sent his nephew Serlo ahead with a detachment, confidently ordering him to prevent the retreat of the Muslim force. He then charged the enemy, who in their subsequent flight were caught in a vice between his and Serlo’s troops. Having inflicted heavy casualties on Messina’s garrison, the Normans now boldly attempted to capture the city. The greatly depleted garrison was supplemented by non-combatants including women, whose determined resistance subsequently forced the invaders to abandon the siege. The strait was still too stormy to cross, and Roger became particularly worried about the likelihood of a counterattack launched by neighbouring Muslim forces. As William the Conqueror would later do at Saint-Valery in 1066, Roger decided to appeal to a saintly power for the calming of the sea, pledging to rebuild the Byzantine church of Agios Andronikos at Reggio with the proceeds of the campaign. Unsurprisingly for a monk, Malaterra believed that Roger’s spiritual devotion to the saint paid off: the sea calmed ‘allowing him to cross with impunity’.³²

The third campaign was launched in May 1061 and, while its predecessors were concerned with plunder and reconnaissance, this one was clearly concerned with securing a permanent foothold. A sizable fleet was gathered from Apulia and Calabria (mostly from the latter) for the invasion. While Byzantine naval terms were used with precision at times, they were often used interchangeably. For example, a dual-masted galley featuring two banks of oars on either side was known as a chelandion or dromon, the former originally denoting a cavalry transport, the latter a faster ship with a single bank of oars (dromon means ‘runner’). Dromon was a term rarely used by Latin writers in the tenth and eleventh centuries; they instead consistently used both variations and transliterations of chelandion to denote ships variously employed for the transport of troops and sea-based combat (featuring marines and equipped with flamethrowers). There is, however, one notable exception to this general rule: Malaterra specified dromundi and germundi, which he stated were present respectively in the Muslim and Norman fleets (as observed by Malaterra’s editor Pontieri, the latter could be a corruption of the former term). Given that the sources subsequently mentioned the use of ships to transport cavalry across the Strait of Messina, Malaterra’s germundi can be understood to be dromones/chelandia (pl.) modified for the purpose. It was noted in Chapter 1 that vessels of this type were built in Calabria and that a sizable fleet was maintained in the region; it can therefore be assumed that the Normans used them to transport their cavalry. When the Byzantine leaders abandoned Squillace for Constantinople in 1060 they only took one ship. It follows that various vessels remained in the region, and since the Normans had only recently gained control of the coastal areas where the Byzantine fleet was based, they are unlikely to have organized the construction of additional ships at such an early date. Lending credence to the idea that the Normans used chelandia/dromones as cavalry transports is the later, but well-informed, testimony of Anna Komnene: Guiscard’s ‘horses and armed cavalrymen embarked on the dromones’ (1081), a necessity since opposition by mounted forces was anticipated when landing. This passage is interesting for another reason, as the presence of battle-ready cavalrymen on ships suggests the assimilation of technology famously used by Nikephoros Phokas at Crete in 960 – that is, by means of ramps (klimakes), fully equipped cavalry were able to ride onto the beach. Indeed, while Malaterra’s description is ambiguous, it is likely that Count Roger’s cavalry did the same at Malta in 1091.³³

The other type of ship present in the Norman fleet mentioned by Malaterra was the galea, a transliteration of the Greek term for a fast galley which, as Hélène Ahrweiler noted in her classic work on the Byzantine navy, was variously employed for reconnaissance and espionage missions, in addition to the transport of guards and officials bearing messages and intelligence. That some of the Norman ships conformed to the Byzantine model is confirmed by Amatus, who mentioned two ‘most slender and very fast’ galéez. However, it would be problematic to accept that all of Malaterra’s galleys were of this type: firstly, because Amatus only mentioned two of them, and secondly because the combat readiness of Calabrian ships – manned by Greek- and Arabic-speakers from Reggio – is suggested by Amatus’ earlier mention of a sea battle in the Strait of Messina prior to the launch of the third campaign. While Stanton has referred to this fleet as ‘makeshift’, given its ability to engage the Muslims with the loss of a single ship and eleven casualties, it was evidently up to the task.³⁴

The size of the fleet was unfortunately not given, but Malaterra wrote that it outnumbered the Muslim flotilla of twenty-four ships (according to Amatus) sent to prevent the crossing to Punta del Faro. Malaterra insisted that the Muslim fleet dispatched from Palermo consisted of catti, golfari and dromundi, whereas Amatus stressed that all were catti. Stanton observes that catti were almost certainly larger than dromones/chelandia, which explains Malaterra’s comment that direct confrontation with the enemy fleet was deemed unwise. Accordingly, two of the fleet’s fastest ships were sent to reconnoitre the coast during the evening, with the aim of finding the safest place to disembark. Having chosen Tremestieri (6km south of Messina), Roger landed with 270 (Amatus) or 300 (Malaterra) men, half of whom were cavalrymen; and Amatus noted that thirteen ships ferried the troops. That these ships were chelandia seems certain as in the ninth century they were capable of ferrying twelve horses each. Using this evidence as a benchmark, thirteen chelandia could in theory transport 156 cavalrymen, a figure that accords with the 150 specified by Malaterra. However, after the landing, the same ships were used by Guiscard to ferry 170 mounted troops, suggesting that they were in fact capable of transporting thirteen to fourteen mounts (as was later the case at Malta in 1091 [14]). When it comes to the first landing, the transport of the non-mounted troops needs to be considered. Depending on which authority is favoured, there were 120 or 150 infantry. Using thirteen as the average capacity for cavalry per ship, each of the same number of vessels needed to ferry nine or twelve infantry, an achievable figure given that Thietmar of Merseburg noted that a chelandion was large enough to be manned by a crew of 150, a total that correlates with the numbers specified in Byzantine sources.³⁵

With few troops to defend Messina on account of Roger’s previous campaign, the coastal stronghold was stormed with little resistance. It was then subjected to a particularly brutal sack; those unable to escape to nearby ships were put to the sword. Malaterra, who at times liked to insert what we would call ‘human interest’ stories, recounted that, in order to spare her from the indignity of rape, a noble Muslim woman was killed by her brother. The Muslim navy, realizing that all was lost and driven by the need to flee the increasingly rough conditions in the strait, returned to Palermo. Garrisoning Messina with ships and cavalry, the Hauteville brothers set their sights on nearby Rometta (12km south-west). However, before the army arrived before Rometta’s walls, having heard of the horrors of Messina’s sack, its citizens chose to capitulate, duly swearing their allegiance to Guiscard on copies of the Qur’an. Soon afterwards, Greek-speaking inhabitants in the hinterland of Frazzanò (72km south-west of Messina) pledged their loyalty, welcoming the Normans and showering them with gifts. This event seems to have triggered a general trend, for other fortified towns in the Val Dèmone were later said to have followed suit. That being said, other towns, presumably with sizable Muslim garrisons, continued to resist: the citizens of the hilltop town Centuripe (97km south-west of Messina), for example, whose battlements were protected by catapults and archers, forced the Normans to abandon their siege. Accompanying the Normans as a commander was Ibn al-Ṯumnah, whose scouts informed them at Paternò (15km south-east of Centuripe) that Ibn al-Ḥawwās’ army was advancing on their position.³⁶

The Muslim force, consisting of both local troops and ones from Zirid Ifrīqiya, was said by Malaterra to have numbered 15,000. His total was almost certainly an exaggeration, but not as much as Amatus’ figure of 15,000 cavalry and 100,000 infantry! According to Amatus, the Norman army consisted of 1,000 cavalry and as many infantry, although Malaterra, who neglected to mention the latter, stated 700 of the former. These numerical discrepancies can be explained by the fact that since al-Ṯumnah was said to be one of the commanders of the campaign, he presumably led his own corps of 300 cavalry. The Normans, now encamped near Enna (55km west of Paternò), drew up two battle lines. As Malaterra neglected to mention the infantry, he seems to have meant that two batailles of cavalry were formed. The Muslims arrayed themselves into three lines, and one of them began the battle. The assault was repulsed, and the Norman horsemen soon turned the ensuing retreat into a rout, driving the enemy cavalry into ditches guarding the approach to the base of the hill on which Enna was situated (Fig. 8). So numerous were the horses captured, observed Malaterra, that those who had lost their mounts in the battle could be compensated with ten. Despite this decisive victory, the invaders were unable to bring about Enna’s capitulation. The ever-restless Roger decided to lead 300 juvenes (youths), presumably his household troops, south-west into the island’s western province (Val di Mazara), raiding all the way to Agrigento (68km south-west of Enna) and returning with much plunder. With winter on the way, the brothers chose to abandon the siege and return to the mainland. Meanwhile, al-Ṯumnah conducted raids from his base at Catania (72km east of Enna).³⁷

Just before Christmas in the same year (1061), Roger embarked on his fourth Sicilian campaign. With a force of 250 cavalry, he again plundered all the way to Agrigento. The Greek-speakers welcomed him throughout the Val Dèmone, and the citizens of Troina allowed him to spend Christmas with them in their town. By the beginning of 1062 much of the Val Dèmone was loyal to the Normans, although important hilltop strongholds such as Centuripe and Enna remained in Muslim hands. Capturing the latter was particularly important because, as Loud observes, it ‘barred any permanent advance into the south of the island’. Roger returned to Sicily for the fifth time in early 1062 and, with al-Ṯumnah’s assistance, captured the hilltop town Petralia Soprana (30km north-west of Enna), which the count garrisoned with ‘cavalry and stipendiaries’. The latter term can be translated as ‘mercenaries’, but probably meant fideles without land rather than ‘guns for hire’ or ‘soldiers of fortune’. Having garrisoned Troina (43km east of Petralia) similarly, Roger returned to Calabria. Al-Ṯumnah continued to conduct raids and forge alliances in the region – that is, until he was assassinated at Entella soon after Roger’s departure.³⁸

Without the critically important support of al-Ṯumnah and his troops, the garrisons at Petralia and Troina were forced to abandon their posts and flee to Messina. Al-Ṯumnah’s death, coupled with the abovementioned series of civil wars in Calabria raging throughout the 1060s–70s, meant that the flames fuelling Roger’s Sicilian obsession were reduced to embers. The same applied to Guiscard, who in addition to revolts in Calabria and Apulia was still preoccupied (until 1071) with subduing the rest of the Byzantine strongholds in the latter province. Accordingly, the conquest of Sicily effectively stalled. While campaigns continued to be launched during this period, at times with success, it would be imprudent to assume that the conquest of Sicily took thirty years to complete solely on account of prolonged political instability on the mainland. Although much of the Val Dèmone capitulated with relative ease, the region was predominantly Christian, whereas the Val di Noto (southeast) and Val di Mazara (west) were overwhelmingly Islamic; accordingly, the invaders had to overcome resistance on an unprecedented scale (although they were nonetheless greatly aided by disunity among the Muslims).

Having finally secured the territory promised by his brother in Calabria, Roger, accompanied by his wife Judith, returned to Sicily in the summer of 1062 with 300 (probably household) troops. The primary aim of the campaign seems to have been to reverse some of the losses of 1061 – namely, to reacquire the towns whose garrisons fled when news reached them of al-Ṯumnah’s assassination. If the Normans were to have any chance of penetrating into the Val di Mazara, possession of towns near to its frontier was crucial. Roger advanced on Troina, whose citizens again gave him access to their hilltop town. According to Malaterra, the people of Troina were not as welcoming as they had been when Roger spent Christmas with them the previous year. The change of heart can be easily explained: like the towns of Calabria, Troina’s people were happy enough to pledge their allegiance and provide assistance, but did not appreciate being garrisoned. Of course, the garrison had long since fled by the time of Roger and Judith’s arrival, but the people had no doubt enjoyed their short-lived period of autonomy. And it was short-lived indeed: not only did Roger strengthen Troina’s fortifications – presumably adding a wooden donjon – but he also garrisoned it again, in addition to partitioning the town into two sections (taking half for himself). Particularly unpopular was the count’s choice to billet his troops, for the affected households were said to have been greatly concerned about the honour of their daughters and wives (which could well imply acts of sexual indecorum by the billeted soldiers).³⁹

The mounting tension was soon replaced by violence: while Roger was busy raiding in the vicinity of Nicosia (19km south-west), Troina’s citizens revolted. The count returned with the greatest haste when news reached him of the revolt, for the seriousness of the matter was greatly exacerbated by the fact that Judith was among the besieged. The Byzantines of Troina, meanwhile, had constructed a barricade to separate their half of the town from that held by the count. All attempts to breach the stoutly defended palisade failed, and the increasingly desperate situation was worsened by the arrival of Muslim troops from neighbouring garrisons, who promptly surrounded the town. Prevented from leaving Troina to procure food, the Normans began to starve. Roger attempted to disperse the Muslim circumvallation, but the cavalry were driven back and the count himself was almost killed when his horse was felled by javelins. While he was able to extract himself from this dangerous situation, the count’s plight was clearly hopeless. However, his run of luck was restored when the four-month siege was brought to an end by the onset of a particularly harsh winter. While this seasonal change could well be the sole explanation for abandonment of the siege, Malaterra’s insistence that it was abetted by Roger’s capture of the Muslim camp during a particularly cold evening may well be true. The citizens of Troina now chose to surrender and the leaders of the revolt were hanged. After restoring the town’s fortifications, Roger departed for Calabria to gather replacements for the horses killed in battle (and perhaps eaten in desperation?), leaving Judith in charge of Troina’s garrison.⁴⁰

Roger returned to Troina in 1063 with an unspecified number of troops. Included among them seems to have been his youthful cadre of household troops – who, judging by earlier mentions of them, appear to have been maintained at a strength of 300 – since he was subsequently mentioned as wanting to test their mettle in battle. When word reached him that Zirid reinforcements had arrived from Ifrīqiya, 500 of whom had been sent to garrison the strategically important, seemingly impregnable Enna, Roger decided to march south-east. Sending his nephew Serlo ahead in charge of three conrois (thirty men), the count followed at a distance. Roger’s nephew was instructed to ride up to the base of the precipitous hill on which Enna was built (Fig. 8), with the idea of provoking the garrison to attack what seemed to be a pitifully small force. The ruse worked, but only two of Serlo’s men managed to arrive at the prearranged ambush point unharmed. As usual, Malaterra drew attention to the heroics of his protagonist, but the fierce fighting that ensued seems to have been evenly balanced for the most part. Indeed, while he wrote of great spoils being secured by the Normans, all that could have been gained was, at the very best, some horses, arms and armour. Roger now returned to the Val Dèmone, raided the hinterland of Caltavuturo (62km west of Troina), after which he returned to Enna, hoping in vain to entice its cavalrymen to mount another sally. A raid was now launched southwards into the Val di Noto, reaching as far as Butera (42km south-west of Enna). While a significant amount of plunder, including heads of cattle, was seized in the region, this leg of the raid seems to have been a little impetuous, even excessive, for Malaterra reported that many of the horses died of thirst.⁴¹

Although Zirid troops in 1063 were clearly assisting their co-religionists against the invaders, their presence in the region was the result of the Tunisian dynasty’s direct intervention in the island’s increasingly fragmented political affairs. Tamīm’s sons Ayyūb and Ἁli, in the same year, respectively gained control of Palermo and Agrigento, the latter being captured at Ibn al-Ḥawwās’ expense (his garrison was surely disadvantaged given Roger’s previous two plundering expeditions in the region). The numbers of Zirid troops committed to the island, however, cannot have been large: since renouncing the Shia belief of their nominal Fatimid overlords in favour of Sunni Islam (1052), Ifrīqiya had been a warzone for ten years. Those responsible for the ensuing chaos were Egyptian troops hailing from the semi-nomadic Banū Hilāl clan, who consequently refused to acknowledge the leadership of their Zirid commanders. The dynasty’s control of Tunisia and its environs became increasingly tenuous: in 1052 the Zirids were decisively defeated by the Hilālīs, and a sizeable fleet was wrecked near Pantelleria, an island in the Strait of Sicily (between Ifrīqiya and Sicily). Although the Zirids had re-established political accord with Fatimid Egypt by 1055, the Hilālīs continued to diminish their power in the region, most notably by sacking the capital Qayrawān (1057): an unfortunate outcome, as Metcalfe relates, for the sack ‘marked the decline of one of the great cities of the early medieval Islamic world’.⁴²

Having returned to Troina, Roger was informed in June 1063 that a large Muslim army consisting of both Sicilian and Zirid troops was in the vicinity of Cerami, a fortified town 13km north-west of Roger’s base. He gathered his troops and scouted the enemy army, which upon arrival had bivouacked on the opposite bank of the river bearing the same name. With thirty-six cavalry, Serlo was sent ahead to garrison the town. Having done what was asked of him, the count’s nephew chose to launch a sortie against a detachment of Muslims, whom Malaterra alleged numbered 3,000 ‘except the infantry, of whom there was a boundless multitude’. The monk was clearly exaggerating, for it is unlikely that such a small force was capable of defeating 3,000 cavalry and countless infantrymen. That being said, the outcome of the skirmish need not be questioned: the Muslim detachment sent to capture Cerami was driven off. Roger subsequently arrived with 100 horsemen, and now considered whether to follow up on his nephew’s victory. Roussel of Bailleul, who would soon gain both great fame and infamy in Byzantine service, said that he would no longer provide ‘aid’ (auxilium) to the count if he chose not to pursue the enemy. Roger relented, ‘creating two wedges/battalions [cunei] out of his men’.⁴³

The term used by Malaterra is particularly interesting: it may well indicate Norman adoption of, to quote from an imperial military manual, the ‘triangular formation of the kataphraktoi’, an array in which they may have charged with imperial cavalry at Troina in 1040 (see Chapter 2). However, it is important to stress that while from antiquity the term often denoted a triangular formation, some medieval writers used it less precisely to describe a battle line of infantry or cavalry. Malaterra, however, did not use it in this sense, contrasting the battle lines (acies) of the Muslims and the cunei of the Normans. Significantly, up to this point in the narrative Malaterra had used acies to denote a Norman cavalry formation: Count Humphrey ‘arranged his acies’ at Civitate (1053), as did Guiscard who ‘set in order two acies’ near Enna (1061). The wedge (and rhomboid) formation had been favoured by Sarmatian, Thracian, Greek, Macedonian and Roman/Byzantine cavalry since antiquity owing to its greater piercing power and, to use Phil Sidnell’s apt explanation, ‘the narrow frontage meant that wheeling to change direction could be done relatively quickly’. Another advantage of the cuneus (sing.) was that, unlike the flat, serried formation favoured by northern French cavalry, its flanks were protected. Malaterra only mentioned the use of cunei on one other occasion in the same year, perhaps suggesting that Roger was either experimenting with different formations, or adapting tactically to the task at hand.⁴⁴

Serlo, Roussel and a certain Arisgot from Pucheil (Normandy) were assigned command of the first cuneus, which formed the vanguard, while Roger took command of the rearguard. They rode in battle formation towards the enemy camp, and both the van and rear were shadowed by enemy cavalry, whose intent was no doubt to entice the Normans to break formation. At this stage in the narrative Malaterra, the only source for the battle, chose to render the ensuing clash in a religious manner. He had already prepared the reader for his theological set piece by utilising the Old Testament motif frequently employed by medieval writers – ‘How could one chase a thousand, and two put ten thousand to flight’? (Deuteronomy, 32:30) – in reference to Serlo’s earlier triumph. The outnumbered Normans, riding into certain death, clearly required divine assistance. Accordingly the Byzantine military saint Agios Georgios appeared on horseback, and the formerly frightened Normans, yelling ‘God and St George’, followed his banner into battle against the figurative Saracen dragon. Malaterra’s choice of saint begs the following question: why George instead of the bellicose Norman patron saint, Saint-Michel, whose victory over Lucifer and the fallen angels had an obvious temporal parallel with Christian victories over the Muslims? Miraculous manifestations aside, thankfully the monk provided enough secular data to gain a general idea of the clash.

As had been the case in various battles since the classical period, what is likely to have won the engagement was a full-scale rout caused by the death of the enemy commander. William of Poitiers, in reference to the chaos caused by the rumour of William the Conqueror’s death at Hastings, summarized this feature of ancient and medieval warfare succinctly:

The majestic Roman army … accustomed to conquering by land and by sea, fled on occasion after its commander was either known or supposed to have been killed. The Normans, believing their duke and lord to have been slain, should therefore not be shamed too much for withdrawing by flight.⁴⁵

When Roger’s rearguard was attacked, he engaged with the ‘qā’id of Palermo’, whom Malaterra, in a rare description of armour, described as near-invincible as his coat of mail was covered by a ‘glittering cuirass’ featuring ‘iron plates which were fastened with bonds’ – that is, a lamellar corselet. Naturally, the always clever Roger was able to pierce both layers of protection by deploying a powerful lance thrust towards the gap between two of the iron plates on the commander’s lamellar coat. The count may well have intended to do this on the spur of the moment, but the intervals between rectangular plates on lamellar coats were miniscule – let alone the layer of mail beneath the qā’id’s corslet – so Roger’s success can be attributed more to sheer luck than to Malaterra’s ‘ingenuity’. While the monk insisted the battle continued to rage fiercely after the qā’id’s death, it is more likely that the Muslims now began to flee. Malaterra had earlier noted that the warlike commander was perceived to be invincible by his men. Accordingly his death cannot have failed to lower morale, create general confusion and change the course of the battle as a result. Moreover, it is interesting to note that the so-called Sicilian History, an anonymous thirteenth-century Latin chronicle that largely drew on Malaterra’s work, stated that after the qā’id ‘was killed by the lance of the most glorious count’, his troops fled. What should have been a massacre, therefore, turned into a great victory. How costly this victory was to the Muslims, however, is difficult to determine. Malaterra’s insistence on the slaughter of the rearguard numbering 15,000 by fewer than 150 cavalrymen – although that number should be multiplied by a minimum of three given the abovementioned use of squires or sergeants – is clearly hyperbolic. Regardless, the Normans were undoubtedly outnumbered, and accordingly their victory was nothing short of impressive.⁴⁶

News of the great victory spread throughout Italy: Pope Alexander II, whom Roger had sent a gift of four captured camels, demonstrated his approval by sending a papal banner for the Normans to rally behind in Sicily. The fallout throughout the island was immediate: Christian merchants from Pisa ‘who were in the habit of profiting frequently at the port of Palermo’, were subjected to ‘certain injustices/injuries’, presumably meaning not only that their trading privileges had been revoked, but also that some were victims of violent behaviour. The Pisan government accordingly ‘assembled a fleet of ships from everywhere’, which in the same year (1063) set sail for an unspecified port in the Val Dèmone. Upon the fleet’s arrival in Sicily, its leaders sent a delegation to Roger at Troina requesting the assistance of his cavalry in an assault on Palermo. Probably because Robert and Roger had already planned the attack to be launched against Palermo in the following year, the count declined to provide aid. Undeterred, the Pisan fleet set sail for Palermo, but other than severing the chain which, like the famous one at Constantinople, prevented unauthorised access to the harbour, they were forced to abandon the attack.⁴⁷

Before Roger returned to Apulia to confer with his brother in the summer of 1063, ‘he led a cavalry division to plunder Golisano on the first day, Brucato, the second, and Céfalu on the third’. ‘Golisano’ is the modern Collesano (14km south-west of Céfalu), and Brucato (Ar. Būrqād) was a hilltop town located in the north-eastern foothills of Mount Calogero. Given Brucato’s close proximity to the via Valeria (2km), and Céfalu’s location on it, Malaterra’s plundering expedition was rather more than that. In order for the Normans to advance unimpeded in an westerly direction to Palermo, it was vital to secure control of the old imperial road – which generally corresponds to the modern A19 – skirting the coastline from Messina on the east coast to the island’s westernmost point (Marsala). It therefore follows that Roger chose to raid all the way to the via Valeria in order to determine what sort of resistance could expected on the old imperial road. As no hostilities were recorded during the expedition, two things can be posited: Roger reported this discovery to Guiscard in Apulia, and the brothers accordingly resolved to advance on Palermo in the following year (1064) by means of the via Valeria.⁴⁸

Returning from Apulia in the autumn, the count launched another raid into the south-easternmost extremity of the Val di Mazara. Since Guiscard had granted him an additional 100 cavalry from Apulia, Roger’s forces had now doubled. For the third time, the count plundered the hinterland of Agrigento, securing considerable spoils in the process. A sizeable number of his troops were ordered to advance ahead with the plunder, while Roger followed with two squadrons arrayed in the cuneus formation used previously at Cerami. On the return home, the vanguard entrusted with the spoils was ambushed by an unspecified number of troops, probably hailing from Enna. The cavalrymen tasked with protecting the squires who carried the plunder fled, and some of the latter were killed. The cavalry fled to a nearby hill, evidently abandoning their mounts as Malaterra stated that they climbed it. By this stage, the two squadrons of cavalry arrived on the scene, and the count exhorted his men to descend from the hill and avenge the fallen. Eventually they were convinced to come down and fight, and together they were able to drive off the assault, suggesting that the ambush force was a small one. Having recovered their spoils, which were initially seized and later abandoned by the ambushers, the Normans returned to Troina.⁴⁹

In the following year, the brothers launched what they must have been planning for some time: a campaign with the sole aim of capturing Palermo. They crossed the strait of Messina with a ‘great army assembled from Apulia and Calabria’, although characteristically Malaterra again neglected to provide its total number, choosing only to mention the total number of cavalry (500). From Punta del Faro the army marched unimpeded to the hinterland of Palermo, presumably by means of the quickest route – the via Valeria. Evidently the intelligence available to the brothers was deficient, for they chose to camp on a pernicious hill overlooking the city of Palermo. Afterwards known as Tarantino, the hill was densely populated by an enemy the Normans had not yet encountered: tarantulas, which launched sorties from their silken burrows. Malaterra noted that while the venom caused great pain and distress, it was not lethal. Instead, the duke’s men were afflicted with a particularly gaseous malady. The incessant, painful flatulence suffered by those unfortunate enough to have been bitten could only be partially relieved by the application of a heated poultice to the affected area. If Count Stolberg’s itinerant account is anything to go by, the eleventh-century cataplasmic method was more advanced than that used in the nineteenth: the German humanist recorded that the ‘Grecian’ Tarentines of Apulia believed the pain ‘could only be cured by violent dancing’! Quite unsurprisingly given humanity’s age-old penchant for toilet humour, Malaterra related that the victims became the source of much amusement for those lucky enough to have escaped the fangs of Tarantino’s theraphosid residents.⁵⁰

Defeated by Tarantino’s spiders, the Normans were forced to encamp elsewhere. They were able to plunder some of the districts outside of the great city’s walls, but after three months of slim pickings the siege was abandoned. While the brothers’ audacity and ambition can be underscored, the campaign should nonetheless be regarded as poorly planned, even impetuous. Palermo had been for some time an important entrepôt of the Mediterranean. Like Constantinople, the city was defended by stout walls and a protected harbour. As the Arab geographer al-Muqaddasī observed in the late tenth century, Palermo’s northern fortifications extended so close to the sea that ‘the water batters the walls’. Not only was the entire city surrounded by walls featuring four large gates, but its centre was also fortified. Accordingly, as the Pisan emissaries had no doubt earlier made clear to Roger, the city could not be taken without a coordinated assault by land and sea. Hence, a land-based investment without possession of a fleet, let alone a siege train, was an exercise in futility. Naturally, it can be expected that the Normans realized the importance of naval support, but chose not to deploy their fleet on account of their enemy’s access to one greater in size. The twenty-four vessels sent to prevent the crossing in 1061 were sent from Palermo, and this fleet undoubtedly formed just a small portion of the total number of galleys available to the city’s commanders.⁵¹

The army now took a south-westerly route, later arriving before Bugamo in the vicinity of Agrigento. The town was stormed, then subsequently sacked; its unfortunate occupants were taken into captivity and later repatriated to Scribla, Calabria. Agrigento was besieged next, and a determined sally launched by its garrison was driven back. Despite this initial success, without siege engines the Normans stood little chance of storming the well-defended town, and Guiscard accordingly abandoned the investment, thereafter returning to Calabria in order to suppress revolts in the region. For the next two years, little progress was made in Sicily, other than adjustments in 1066 to the fortifications of Petralia Soprana, which had earlier been abandoned by its Norman garrison (1062). Perched in the precipitous southern slopes of the Madonie mountains, the town was already well fortified, for al-Muqaddasī recorded that ‘Petralia … is walled; inside it stands a fortress, with a church’. As Malaterra observed in regard to the count’s renovations, ‘he meticulously strengthened it with towers and ramparts outside of the gate’. With more than a little exaggeration, the monk asserted that his protagonist now dominated much of Sicily from Petralia, but regardless it became an important advance base for subsequent westerly incursions into the Val di Mazara.⁵²

Over the next two years Roger continued to conduct raids from Petralia, and Malaterra related that his protagonist managed to secure tribute from some unspecified towns in the vicinity. By 1068 the Palermitans had decided to rid themselves of the Norman gadfly, mustering an army that intercepted the count’s raiding force soon afterwards near Misilmeri. Unfortunately, Malaterra devoted only a few sentences to the ensuing battle: all that can be gleaned is that Roger ‘wisely arranged his cavalry division’, and his forceful charge routed the Muslim ‘multitude’. The monk then related that among the spoils secured were cages containing carrier pigeons. Evidently Roger and his entourage were not in the habit of carrying ink with them, for ‘small notes written in blood’ were attached to the pigeons; when freed, the feathered messengers flew back to Palermo to bring news of the defeat to the horrified Palermitans.⁵³

Disunity among the Muslims continued to assist the Hauteville cause greatly in Sicily. Ibn al-Ḥawwās, still in possession of Enna and Castronuovo, launched a campaign to recapture Agrigento from its Zirid ruler Ἁli, but was subsequently killed in a pitched engagement. Ἁli’s brother Ayyūb, probably after the victory at Misilmeri, was later confronted by a revolt in Palermo. Perhaps on account of continuing problems in Ifrīqiya along with those in Sicily, he and Ἁli decided to abandon the island in 1068/9. From this point onwards, the Sicilians would only receive occasional assistance from the Zirid fleet. While it is the duty of the historian to assess what occurred, rather than what could have happened, it is nonetheless tempting to pose the following question: had the Muslims been able to unite against a common enemy, would the Normans have been successful? After all, given the often smallish size of their forces, a single decisive victory could have driven them from the island.⁵⁴

Nothing else of importance seems to have occurred in Sicily for three years – that is, until July 1071 when Guiscard began preparing a second, much better-organized campaign against Palermo. The brothers later met at Catania on the east coast, which Roger had just captured and garrisoned with forty men. Malaterra asserted that the count and duke chose to assemble at Catania in order to fool the enemy into thinking that their sights were set on Malta, not Palermo. The brothers may have been worried about Zirid naval patrols, and therefore wanted to divert their attention. But this ruse, if it was intended to be one, failed: William of Apulia noted the presence of Zirid ships at the ensuing naval battle in the bay of Palermo. As Bari had recently capitulated after an epic three-year siege (1068–71), Guiscard’s fleet was now substantially larger. Indeed, one of the Bariot annals mentions fifty-eight ships departing for Sicily in July, a figure which more or less agrees with the one given by Amatus for the total size of Guiscard’s fleet (fifty-one, including ten large war galleys [catti]). Adding to the effectiveness of the fleet was a contingent manned by experienced Greek-speaking sailors and marines from Bari. Given the linguistic divide, it is possible that a Byzantine commanded these ships: William of Apulia later mentioned that Stephan Pateranos, who had arrived with an imperial fleet to relieve Bari in 1069, was given control of the Bariot fleet on its return to Apulia from Palermo in 1072. It is therefore possible that he commanded it beforehand. Naturally Guiscard took overall command of the navy, which ferried ‘an abundance of infantry’ from Catania to Palermo, while Roger rode towards the city with an unspecified, but ‘large’, number of cavalry.⁵⁵

The fleet approached the bay of Palermo (Fig. 9), whereupon a fierce naval battle was fought. William of Apulia related that the fleet drew up in formation, and in order to protect the sailors and marines from missiles hurled by javelineers and slingers, the ships were covered by tarpaulins. The Palermitans, aided by Zirid ships from Ifrīqiya, engaged the enemy fleet with great determination. Nonetheless, they were driven back towards the lengthy chain – around 200m in length given the estimated width of harbour’s mouth – which guarded the entrance to the fortified inner harbour (Le Cala). While the chain was being lifted, Guiscard’s fleet burst through, capturing some of the ships and burning the others. Cut off from the sea, the Palermitans would now have to withstand a lengthy siege, sustaining themselves only on supplies already contained within the city. Accordingly, it was this stage of the siege that was most crucial to its later success.⁵⁶

After ‘blockading the city from the ocean by ships’, Malaterra noted imprecisely that Guiscard approached the city’s walls from one side with his Calabrians and Apulians, while his brother was tasked with another section of the defences (Amatus similarly stated that Roger attacked from ‘the other side of the city’). Since he later clarified that the duke attacked the walls ‘lying near to the navy’, Guiscard evidently assailed the city’s north-eastern walls, while Roger concentrated on those to the south-east. In order to storm the city’s outer walls, it was necessary to clear the missile troops manning the battlements. Accordingly, without catapults to achieve this objective – although Malaterra alone suggested some were later assembled – the duke deployed a contingent of slingers and archers. Since Malaterra recorded that the duke’s infantry were Calabrians and Apulians, among these missile forces must have been Byzantines hailing from the former province, a region noted above for the quality of these troop types (i.e. Aiello and Cosenza). Evidently worried that these lightly armoured troops would be decimated by Palermitan sallies, Guiscard made sure that a contingent of ‘armoured cavalry’ monitored their progress. William’s ‘armoured’ qualification, used only once in his entire work, is particularly interesting. Throughout his poem, equites/milites generally denoted heavy cavalrymen, and hence these ones were already wearing helmets and haubercs. Two possibilities therefore come to mind: either, like the heaviest Byzantine kataphraktoi, the horses were armoured as well (perhaps to lessen casualties from missile troops); or ‘armoured’ was used to indicate that, in addition to the presence of the usual heavy cavalry at the siege, there were also less heavily armoured mounted troops (conterati?). Guiscard’s fear was soon realized when the infantry contingent, having bombarded the battlements with missiles, was routed by a determined sally. The armoured cavalry rode to their assistance, driving back the assault. The missile troops regrouped, as while some of the casualties reported by William were inflicted by lance and sword, most of the Muslims were killed by slingers and archers (particularly the latter).⁵⁷

What finally brought the five-month siege to an end in January 1072 was the scaling of the seaward walls, which by this stage, Malaterra observed, were unprotected. The city’s gates were opened soon after the walls had been climbed, and subsequently the cavalry and infantry forces poured through them. The city’s capture, however, was not yet complete. While Guiscard had overcome Palermo’s outer fortifications, he still needed to overcome the inner ones. It is here that the crucial importance of the naval blockade is underscored: a potentially protracted siege of the inner city was prevented by the fact that the Palermitans had been starving for months. Exhausted by malnourishment, they had little choice but to capitulate and agree to pay what could be called a ‘reverse ǧizya’ – that is, the poll tax previously paid by ḏimmī (Jews and Christians) for the freedom to practise their religion. Having settled matters in his new city, Guiscard chose to keep Palermo for himself, thereafter granting the Val Dèmone and any other future territory seized to his brother (to be held as vassal).

If the Normans had believed that the capture of the Sicily’s greatest city would cause a ‘domino effect’, they had hoped so in vain: the conquest would not be completed for another twenty years. One reason for this was that Guiscard never again returned to the island. Accordingly, Roger had to complete the conquest with fewer men, although it must be stressed that the number he could call on was larger than those available to him prior to the capitulation of Palermo: firstly, Guiscard left behind a ‘small part’ of the army before returning to the mainland, and secondly Roger increasingly recruited Muslim troops. Despite the fact that, after subdividing the Val Dèmone among Arisgot and Serlo, the count now had relatively cheaper access to cavalrymen – i.e. fideles/vassals may still have been entitled to a share in any spoils seized, even when they did not participate in a campaign (e.g. after Malta [1091]) – those without lands still had to be paid with cash or promises of future reward (land, plunder, or both).⁵⁸

Soon after the capture of Palermo, Roger’s nephew and vassal Serlo was ambushed and killed by a force hailing from Enna. This unfortunate event no doubt reminded the grieving count that the capture of Sicily’s most impregnable fortress was of the greatest importance. He first decided to seize a base to the west of Palermo, from where he could later conduct campaigns against the west coast and hinterland of the Val di Mazara. The base chosen was Mazara (del Vallo), a coastal town situated 85km south-west of Palermo. Having captured it in 1072, Roger established a donjon within its walls, something which he subsequently did at Paternò (17km north-west of Catania) in the Val Dèmone. By 1074, the count decided to test Enna’s defences, shrewdly constructing a fortress at Mount Calascibetta (3km north), from where he conducted raids around the hinterland of the lofty fortress. In the following year Mazara was recaptured by a Zirid fleet from Ifrīqiya, although Roger’s garrison remained in possession of the donjon. Upon hearing the news, the count rode to Mazara with great haste, routing the enemy within the town’s walls, and thereafter managing to drive the besiegers back to their ships.⁵⁹

Roger’s family suffered another blow when, in 1075, Hugh of Gercé, a noble hailing from another recent Norman acquisition (county of Maine, 1064), was killed in an ambush. Not only was Hugh the count’s son-in-law, but he had been entrusted with overall command of Sicily when Roger was recalled by his brother to deal with the revolt of their nephew Abelard in Calabria. Hugh, whom the count had earlier granted authority over Catania, was expressly told by Roger not to venture too far from it, lest he be attacked by forces commanded by the amīr of Syracuse, Ibn al-Ward. Hugh, thirsting for glory, disregarded the count’s advice; he collected Jordan and ‘the count’s military household’ from Troina and returned to Catania. Using the same tactics Roger had used against the garrison of Enna in 1063, Ibn al-Ward’s ‘great force of picked heavy cavalry’ concealed themselves at a considerable distance from Catania’s walls. Subsequently a detachment of thirty men rode within sight of Catania’s garrison, with orders to flee once they had managed to lure out the enemy’s cavalry. The ruse worked, as Hugh, and Roger’s illegitimate son Jordan, stormed out of the city’s gates. While they wisely sent a detachment of thirty ahead to scout for signs of an impending ambush, the Syracusan cavalry cleverly waited for the vanguard to pass, after which they charged headlong into the unsuspecting rearguard. While those sent ahead to scout managed to flee unscathed to Paternò, the others were massacred; only Jordan managed to ‘slip away with a few to Catania’. The deeply distressed Roger returned to Sicily in the summer of 1076, waging a campaign of vengeance in the Val di Noto. He first razed the fortified town of Judica near Caltagirone (59km south-west of Catania) to the ground; the men were killed and the women captured. Unable to seize anything else, Roger chose to burn all fields encountered throughout the region. This act was particularly disastrous for the region, as the harvest season (June) had just begun; the result should by now be a familiar one: famine.⁶⁰

In May 1077, Roger launched his most ambitious campaign to date: an assault by land and sea on the western coastal town of Tràpani. While Guiscard seems to have taken much of the fleet with him back to Apulia in 1072, Roger nonetheless appears to have had access to a considerable number of ships. Launching into verse, Malaterra exclaimed with some exaggeration: ‘Alexander the Great’s fleet was not as illustrious as this’. In addition to the ships stationed at Reggio and Messina, there were those captured during the siege of Palermo, and while the city belonged to Robert, it seems likely that the services of its substantial arsenal would not have been denied to the count. With the fleet presumably ferrying the infantry, Roger rode to the destination with his cavalry. Eager to win glory for himself, Jordan ‘held counsel with members of his household’, choosing 100 of them to conduct a night-time operation. According to Malaterra, the destination chosen was the verdant spit of land situated around ‘ten miles’ from the town’s walls, which was used for the grazing of animals. Jordan landed his force there in the evening, thereafter finding a suitable place from which to launch an ambush. When the citizens came to graze their livestock in the morning, Jordan’s men burst out from their concealed position and attacked them. Soon afterwards, a sizable force stormed out of the town’s gates, and a fierce struggle ensued, although Jordan’s men were eventually able to gain the upper hand. Not long after the Muslims had been driven back into their town they chose to capitulate, accepting terms similar to those offered by Guiscard to the Palermitans in 1072. Roger subsequently garrisoned the town with some cavalry, in addition to strengthening the fortifications with towers and ramparts.⁶¹

Using Tràpani as a base, the count now concentrated on capturing towns in the surrounding districts. The task was not a simple one, for Malaterra noted that these towns were fortified and well defended. Unfortunately, the count’s biographer vaguely related the subsequent surrender, by means of sieges, of ‘twelve most famous fortified towns’; suspiciously the same figure he had earlier used in reference to Calabria (which could well mean ‘I don’t know how many, but more than a few’). A positive fiscal outcome for Roger was that he now had more land to parcel out among his cavalrymen; accordingly he awarded an unspecified number of them land to be held from him as vassals. Thanking his men for their military service, Roger disbanded the army and rode to Vicari, a hilltop town south of Palermo, where he enjoyed a break from his constant campaigning. The weary count, however, soon received an invitation to capture the lofty fortress at Castronuovo (17km south-east of Vicari). The town’s commander was despised, even by his ‘vassals’ (fideles) – perhaps those who held land (qaṭā’i) from him – and the tense situation had reached boiling point. Having been subjected to a flogging, an unnamed miller gathered together a force and captured the town’s highest point, from where an invitation to take control was sent to Roger at Vicari. After arriving at the base of the hill that the fortress crowned, the Normans were able to access the citadel by means of a rope lowered by the miller. Witnessing the arrival of enemy troops, the commander gathered everything he could and fled. The count garrisoned the town and made a point of generously rewarding the miller, hoping that the act would encourage other towns to follow suit.⁶²

In February/March 1079, Roger set his sights on the capture of the impregnable town of Taormina. In the late tenth century, al-Muqaddasī noted that the precipitous town was crowned by a stone fortress (castello Saraceno) on Mount Tauro, in addition to having a port on the Ionian sea (Figs. 10 and 11). Taormina’s impregnable nature is attested by the fact that it was one of the last of the Byzantine strongholds to fall to the Muslims (902) – that is, long after the period (827–59) during which much of Sicily (excluding Taormina, Catania, Syracuse and Rometta) was captured. The town soon expelled its Muslim garrison, resisting attempts to retake it until 962, when its inhabitants could no longer withstand the detrimental effects of a thirty-week siege.⁶³

Given Taormina’s near-invincible reputation, the count no doubt prepared for a lengthy investment. As he had earlier done at Tràpani, Roger surrounded Taormina by land and sea. His land forces took the precipitous paths up to the town in order to prevent its inhabitants from securing supplies, while the fleet performed the same task in the harbour below. Not only was Taormina hemmed in by men on the landward side, but it was also surrounded by a network of twenty-two siege forts, interlinked by palisades. Presumably contained within this network of fortifications was the lofty Byzantino-Arab fortress Castelmola (see Fig. 10) which overlooks castello Saraceno; indeed, Malaterra’s indication that the people of Taormina were denied access to the ‘great mountain’ suggests that both fortresses were captured. The blockade was maintained for six months and Taormina’s starving inhabitants eventually surrendered in July/August 1079. The successful siege might have been prevented had a Zirid squadron of fourteen ships chosen to engage the count’s fleet, but according to Malaterra its commander chose to secure peace instead, claiming he had been ordered by amīr Tamīm to patrol the region for seaborne freebooters. Since from the Zirid perspective the Normans were pirates, Malaterra’s account could well be defective in this respect. Indeed, the timing of the fleet’s arrival is more suggestive of an attempt to lift the naval blockade than a serendipitous appearance. Assuredly the count possessed more ships, but one unresolvable mystery remains: as Malaterra noted that the ‘rigging/tackle’ of the count’s ships had been ‘dismantled’, they posed little immediate threat in this state and, to use the appropriate colloquial expression, were therefore ‘sitting ducks’.⁶⁴

Towards the end of the siege of Taormina, Roger was displeased to learn that the Muslim occupants of (San Guiseppe) Jato (23km south-west of Palermo), a fortified town perched on top of a mountain (Mount Jato) so large it could not be surrounded, had refused to ‘to pay service and tribute [i.e. ǧizya]’. The count began the siege in June 1079, but the site’s inaccessibility only led to frustration. Further exacerbating the situation was the receipt of news that the citizens of Cinisi (22km north-west of Jato) had chosen to follow the precedent set by their co-religionists at Jato. No doubt uttering an expletive or two, the count summoned some Muslim cavalrymen to whom he had earlier granted land in Partinico (13km south of Cinisi) and Corleone (20km south-east of Jato). After ordering them to take over the siege, Roger advanced on Cinisi. While his Muslim cavalry devastated the crops at Jato, the count attempted to do the same at Cinisi, only to be prevented by its citizens, who chose surrender over famine. Malaterra neglected to relate Jato’s fate, suggesting that it remained independent for the time being.⁶⁵

Roger’s fame in Italy and Normandy had spread to the south of France, as in 1080 emissaries from Count Raymond IV of Toulouse (r. c.1041–1105) – later a leader of the First Crusade who became a bitter rival of the count’s nephew Marc Bohemond – requested the hand of Roger’s daughter Matilda in marriage. The count consented to the union and the wedding was celebrated in Sicily. He was later summoned to the mainland to assist his young nephew Roger Borsa, whom Guiscard had placed in charge of Apulia while he invaded the Byzantine empire. But Roger was soon forced to return to Sicily to attend to another troubling development. The amīr of Syracuse and Noto, Ibn al-Ward, said by Malaterra to have been ‘experienced and devoted to military training’, had been able to capture Catania by means of treachery from within. A force said to have consisted of only 160 cavalry was hastily organised by Jordan, Robert of Sourdeval (Normandy) and Elias of Cartomi – the latter a Sicilian Muslim who had converted to Christianity and served Roger as a commander during the siege of Taormina. Unsurprisingly, Malaterra provided a stupendously large figure for the Muslim force engaged near Catania: 20,000 infantry and an unspecified number of cavalry. Al-Ward’s infantry formed the right wing, the cavalry the left. Evidently the infantry column was protected by heavily armoured infantry and missile troops, as three separate charges were repulsed. The Normans then assailed the formation of cavalry; after some fierce fighting, the enemy horsemen were driven back to the city’s gates. Now faced with a potentially lengthy siege, al-Ward chose to slip away to Syracuse, presumably by sea.⁶⁶

Having returned to the island, Roger was presented with a challenge, this time by one of his own vassals. Ingelmar, a ‘low-born cavalryman’, had earlier been granted the hand of Serlo’s widow in marriage, whose dowry included one quarter of the town of Geraci (not to be confused with Gerace, Calabria). Ingelmar, having greatly improved the height and strength of the donjon established by Roger, was ordered by the suspicious count to return the keep to its former state. Emboldened by the fealty of the town’s Byzantine population, Ingelmar chose to defy his lord. Roger, ‘ready to maintain his legal rights’, disavowed the oath of protection to ‘his man’, thereby giving him the right to undertake a siege of Geraci. The count invested the town for an unspecified amount of time, and when it became clear to Ingelmar that Roger was not going to give up, he fled. Roger subsequently rode to Messina, where he supervised a team of specialist masons tasked with improvement of the town’s fortifications. Not only were extensive battlements added, but so too was a donjon. Roger, according to Malaterra, held Messina to be ‘the key of Sicily’. Accordingly, the count’s attention to the town’s fortifications indicates that he was deeply concerned about the continuing threat al-Ward posed to the Val Dèmone from his bases in the south. Moreover, while Malaterra made no specific mention of it, Roger may well have improved Messina’s arsenal. Not only was a substantial fleet for the subsequent campaign against al-Ward’s Syracuse fitted there (1085–86), but after the foundation of the Kingdom of Sicily (1130), as Matthew notes, ‘Messina became the chief royal dockyard’.⁶⁷

While the count was right to be worried about enemy incursions from the south during his prolonged absence, ironically the threat to the Val Dèmone came first from within. Although Malaterra hinted that the count’s son may have been contemplating collusion with the Muslims, Loud shrewdly observes that the rebellion had more to do with Jordan’s desire to secure land endowments for himself and his followers, and hence the dispute can be likened to the earlier ones between Guiscard and Roger. Further exacerbating an already tense situation was the fact that Jordan, despite being the only one capable of succeeding his father, was illegitimate. With the difficulties faced by William the Conqueror being a case in point, the accession of a bastard was a guaranteed recipe for civil war in Norman society. In 1083/4 Jordan first captured San Marco d’Alunzio, the fortress established west of Messina in 1061 by Guiscard. He then seized control of Mistretta (109km south-west of Messina), a town to the northwest of Troina. Jordan’s subsequent attempt to capture the latter town, however, failed. While Roger later came to terms with his son, his forgiveness did not extend to Jordan’s circle, twelve of whom were blinded.⁶⁸

Following Guiscard’s death in July 1085, a dispute broke out between his sons Marc Bohemond and Roger Borsa. While Count Roger was busy assisting the latter against the former, his fears of an attack by al-Ward were realized, although it came by sea and was focused on Calabria instead of Sicily. Nicotera, previously sacked by a Zirid fleet in 1074, suffered the same fate. Sailing south to Reggio, the fleet managed to despoil two churches in the vicinity, thereafter moving up the east coast and ransacking Rocca Asini near Squillace. Al-Ward, having demonstrated how easily the count’s heartland could be attacked, returned in triumph to Syracuse. In response, Roger ordered the preparation of a fleet at Messina in October, which departed for Syracuse in May 1086 (not 1085 as Malaterra believed). Three stops were made on the east coast, and having disembarked to the south-west of the target, Roger rendezvoused with Jordan and the cavalry near Rosolini (41km south-west of Syracuse). Here they decided to send a fast galley commanded by Philip, the son of Gregory, a Byzantine of patrician rank, to reconnoitre Syracuse. The decision to scout the large, horseshoe-shaped ‘Great Harbour’, which to this day retains its ancient Greek name (megas limen = Ital. Porto Grande) was particularly wise: for 1,500 years it had been the submarine burial ground for untold numbers of men and ships. After preventing escape from the Porto Grande’s mouth by sealing it with their ships in 413 BC, the Syracusans were able to triumph over Athens, the greatest naval power in the Mediterranean. The same tactic was successfully deployed in the ninth century, when a large Muslim army besieged Byzantine Syracuse by both land and sea (827–29). Increasingly addled by both disease and extensive casualties suffered at the hands of recently arrived reinforcements, the Muslim ships tried to cut their losses and flee ‘but the Romans prevented exit, positioning the fleet at the mouth of the harbour’ (Fig. 12). Unable to exit the Porto Grande, the Muslims had little choice but to burn their ships and retreat by land. Subsequent sieges were undertaken in 866 and 868, but the Byzantines continued to resist. The city did finally succumb, but only after a nine-month siege (877–88).⁶⁹

Manned by men who could speak both Greek and Arabic, Philip’s ship was able to enter the Porto Grande, maintain its cover, and successfully ascertain the disposition of al-Ward’s fleet. As it was reported to be prepared for battle, it is possible that the overarching purpose of the amīr’s Calabrian raid was to entice a naval attack on the Porto Grande, where his fleet could hope to encircle or entrap the enemy ships in a harbour clearly suited to such a strategy. The plan might well have worked, had it not been for the fact that al-Ward’s sea-based forces suffered a crippling blow to their morale, just as his land-based co-religionists did when their commander was killed at Cerami in 1063. Soon after entering the mouth of the Porto Grande, the two fleets engaged. Al-Ward directed his ship towards the one commanded by Roger, but was wounded by a javelin flung by a Lombard marine named Lupino (a later source claimed that ‘the bold count’ did the deed, but the amended attribution was intended to glorify the Hauteville family). The count then boarded the enemy ship and, while attempting to leap to a nearby vessel, the wounded amīr fell into the sea. Since he was fully armoured, al-Ward drowned quickly. Witnessing the death of their warlike commander, the Syracusan ships disengaged, and many of them were destroyed during the retreat. As al-Muqaddasī recorded in the late tenth century, Syracuse – blessed with ‘a marvellous harbour’ – consisted of ‘two cities joined to one another [i.e. by a causeway]’. Unfortunately, based on his description of the naval battle and subsequent siege, Malaterra knew almost nothing about Syracuse and its Great Harbour. Nevertheless, it can be safely assumed that while the clash between the two fleets raged, Jordan established control of the coastal area where the modern city is located, and the ensuing siege was concentrated on Ortigia – that is, the small island connected to the mainland with a mole. Heavily fortified and only allowing direct access via its causeway, Ortigia was virtually impregnable. It therefore would come as no surprise that, despite being assailed by both land and sea, the defenders of the island fortress were able to hold out for five months. Indeed, its garrison could have resisted even longer had it been able to secure additional supplies, and it seems likely that al-Ward’s wife chose to evacuate only when the existing food stores ran out. Interestingly, given that she and her entourage were able to escape at night with two ships, evidently Roger’s ships had been unable to capture Ortigia’s north-eastern dock. Called the ‘smaller harbour’ by Thucydides, in classical antiquity it featured a neorion (arsenal/dockyard), and an Italian source confirms that it was still there in the twelfth century.⁷⁰

While al-Ward’s wife fled to her deceased husband’s stronghold at Noto, the Syracusans came to terms with the count. Al-Ḥawwās, killed when attempting to recapture Agrigento from the Zirids, had been replaced by a certain Ḥammūd (perhaps his son). He took possession of the town after Ἁli abandoned it, in addition to inheriting Sicily’s most impenetrable stronghold: Enna (Fig. 8). Turning his covetous eyes in the direction of Ḥammūd’s strongholds, Roger resolved to save the most impregnable town for later. Arriving before the walls of Agrigento in April 1087, Roger began a lengthy siege. While Malaterra had little to say about the events of the siege, it seems that, like the citizens of Reggio in 1060, the Agrigentans chose to surrender on 25 July when they saw the deployment of siege engines. Among the captives were Ḥammūd’s wife and children, and Roger ordered his men to treat her with the utmost respect, planning to use them as a bargaining chip for subsequent negotiations with her husband at Enna. These negotiations were so successful that Ḥammūd agreed to surrender his stronghold and convert to Christianity in return for his family and some land in Calabria. He planned to hand over the town by stratagem, and even though Enna’s people chose to resist when the plot was uncovered, after a brief period of skirmishing they nonetheless surrendered to Roger. Such was the bloodless capitulation of one of the greatest fortified towns ever built by the Byzantines: a kastron so well fortified and situated, that despite many attempts to capture it from 829 onwards, it did not fall to the Muslims until January 859, and only then after its garrison had resisted a protracted siege.⁷¹

In the interim between the capture of Agrigento and Enna, eleven towns between these two strongholds were subjected to debilitating raids. After a short time, all of them came to terms with the count. Only Noto and Butera (respectively, 28km south-west and 99km north-west of Syracuse) remained to be captured in Sicily, and Roger accordingly turned his attention to them. He arrived before the walls of Butera in April 1088, surrounding the town and plundering its hinterland. While siege engines were being prepared, a legate arrived to inform the count of Pope Urban II’s arrival in Sicily. Roger went to meet him at Troina, after which he returned to Butera to complete the siege. The town soon surrendered, and the news allowed Roger to receive the submission of the last Muslim stronghold in the Val Di Noto without a single blow. Holding resistance ‘to be the edge of insanity’, emissaries tasked with securing favourable terms were sent from Noto to the count in 1090/1. They were successful, securing a two-year exemption from the ǧizya/poll tax (census). The count granted control of the town to his son Jordan, who strengthened the donjon soon after his arrival. Evidently displeased at the outcome of the negotiations, Ḥammūd’s widow set sail for Ifrīqiya.⁷²

By 1091, all of Sicily was subject to the authority of this most successful of Norman dynasties. But Roger was not yet content. Having spent much of his life campaigning, the count remained restless, pondering which of the ‘overseas kingdoms’ he should conquer first. The target he settled on was an interesting choice for more than one reason: not only would possession of the archipelago of Malta guard against potential seaborne attacks emanating from the central Mediterranean, but it would be a valuable base if the Normans, at some later stage, decided to annex weakened Ifrīqiya. The count at this time was presumably still intending to honour the treaty of non-aggression earlier secured with Tamīm: an entente which had led him to refuse to take part in the Pisan attack on the capital Mahdiyya in 1087. This treaty was of the utmost importance when many of the island’s coastal strongholds remained in Muslim hands, but now that Sicily was firmly under Roger’s control, it is possible that the count may have considered the treaty to have outlived its purpose. Whether Malta was intended to ‘future proof’ Sicily’s defences, or to be used as a stepping stone to future conquests, remains an open question. At any rate, as will be noted in Chapter 6, Count Roger’s son of the same name would capitalize on his father’s conquests in the central Mediterranean.⁷³

Much to the chagrin of his son Jordan, Roger, now aged sixty, decided to command the Maltese expedition. A sizeable army was ferried from Capo Scalambri (Scaramia) on the south-eastern coast (53km south-west of Noto). The fleet consisted entirely of dromones or chelandia and Malaterra noted that Roger’s flagship, despite being the fastest and hence the first to disembark, was nonetheless able to transport fourteen cavalrymen. Where the fleet landed is difficult to ascertain given Malaterra’s vague description. It definitely disembarked on the southernmost, and largest, island (Malta) of the Maltese archipelago given that the northern island of Gozo was attacked on the return to Sicily. Since the rocky west coast of Malta possessed no harbours, the fleet must have landed in either of the two large ones on the eastern side of the island: Marsamuscetto or the Grand Harbour to its south, both of which the modern-day capital Valletta – not established until 1566 by Grand Master of the Hospitallers, Jean de Vallette – overlooks. Notable, also, is the fact that these two harbours are located due south of Capo Scalambri, no doubt indicating Roger’s choice to launch the fleet from a Sicilian port hitherto unmentioned by Malaterra. Having landed, Roger’s men mounted their horses and attacked the troops sent to oppose the landing, driving them back with substantial losses. When all soldiers had disembarked from the other ships, they encamped on the shore. The following morning, while the bulk of Roger’s troops invested the walls of what must have been Mdina in the centre of the island, the cavalry plundered the hinterland. Seeing that further resistance was futile, Mdina’s qā’id negotiated a peace, surrendering weapons and horses, as well as releasing all Christian prisoners in addition to agreeing to pay annual tribute. Soon after the fleet left Malta for Sicily, the count spotted Gozo to his west. After landing and promptly laying waste to the region and pillaging in the usual manner, the island’s leaders agreed to submit to Roger’s authority.⁷⁴

Other than the various occasions on which he was recalled to the mainland to suppress rebellions, Roger spent the last ten years of his life consolidating the administration of Sicily. In addition to distributing more land among his cavalrymen, the count rewarded those fideles already in possession of tenancies with a share in the Maltese spoils. The armies he took with him to deal with revolts on the mainland consisted primarily of Muslim troops, strongly suggesting a concerted policy of improving the island’s defences by maintaining a permanent French-speaking presence in the most important towns and fortresses. In addition to relying on Muslim troops, Roger, referred to as ‘sovereign’ in Arabic coins minted at Agrigento, also made use of Arabicspeaking officials (qādī-s) to administer the affairs of the island’s Muslim majority, and such men would later form the nucleus of the royal chancery (Ar. dīwān; Lat. duana; Gk. sekreton). The same applied to the Byzantines of the Val Dèmone, whose officials (strategoi and kritai) attended to the administrative tasks of those who spoke the same language. Greek-speakers, already prominent in both Roger’s civil administration and the navy, would soon provide the first amiratus (1107), a Latin transliteration of the Arabic word amīr, which via French entered English as ‘admiral’. As Hubert Houben notes, the first amiratus Christodoulos, who had replaced a fellow Greek-speaker Nicholas (Roger’s leading official), was ‘a sort of prime minister’. Military responsibilities came with the office, however, and over time the army’s commander-in-chief was known as ‘the admiral of admirals’, the most famous of whom was Christodoulos’ Greek-speaking protégé, George of Antioch (Fig. 15), whose naval expertise was of inestimable value to the count’s son and later founder of the Kingdom of Sicily, Roger II. Taking all of this into consideration, while they would continue to evolve and be more precisely honed, the foundations for the militarily successful kingdom were very much laid by the remarkable Count Roger of Hauteville.⁷⁵

Apulia

Robert Guiscard’s marriage in 1058 to Sichelgaita, the sister of Prince Gisulf II of Salerno (and hence daughter of the great Norman patron, Guaimar IV) greatly improved the legitimacy of the Hauteville family on the mainland. As Loud has summarized, the marriage was equally beneficial to the Salernitan prince: not only could Gisulf expect protection from the depredations of both William Hauteville (of the Principate) and Count Richard Quarrel of Capua, but the alliance with the powerful Norman could keep the covetous designs of his uncle Guido in check. The benefit of hindsight allows the historian to see that Gisulf’s decision to accept the marriage proposal, while understandable, was ultimately a poor one. The Lombards continued to be influenced by the practices of the imperial court – garbed in a silken mantle and wearing a beard Gisulf looked like a Byzantine on his visit to Constantinople, and he also had himself attired in the regalia of an emperor on coinage – and the ‘marriage strategy’ up to this point seems to have imitated the one employed at Constantinople. That is to say, marriages with foreigners were best avoided: the danger was that ‘barbarians’ (varvaroi) or ‘foreigners’ (xenoi) marrying legitimate Byzantine princesses could seize the state from within. Gisulf certainly seems to have been worried about this potential outcome, prevaricating for some time ‘because the Gauls were seen to be a savage, barbarous, cruel and boorishly minded people’, as William of Apulia put it. As already observed, Normans had only received the hands of (almost certainly) illegitimate daughters (Drogo), and nieces (William ‘Iron Arm’ and Rainulf Drengot). If Drogo’s wife was illegitimate, then, as Loud posits, Guiscard’s marriage was ‘qualitatively different’. Should a son be borne of the union, he could not only hope to succeed as prince if the circumstances later favoured it, but he could also expect the loyalty of his Lombard subjects.⁷⁶

As mentioned above, at Melfi in 1059 Guiscard received papal sanction for his acquisitions to date, and for those still to be acquired in Apulia, Calabria, and Sicily. Now a duke married to the most important Lombard woman in southern Italy, Guiscard’s future prospects looked very bright indeed. Buoyed by his legitimacy in southern Italy, the duke turned his attention to the conquest of the considerable portion of Longobardia still in Byzantine hands. His first (re)acquisition, in 1060/1, was the largest and most important town in the Capitinata, Troia (near Foggia), whose leaders had refused to honour their payment of tribute. As attempts to storm Troia were thwarted by slingers and archers mounted on the battlements, the starvation tactic was employed. Unable to harvest their crops, in addition to running low on food and water, the town’s citizens sued for peace. In order to prevent further attempts at rebellion, the duke erected a donjon on Troia’s highest point, flanking it with towers. Among the town’s inhabitants were (Byzantine?) engineers, whom the duke ordered to construct ‘a stone-thrower and other engines’. Having installed a keep and garrisoned it with troops, in 1061 he captured Acerenza (82km south-east of Troia), although it is not clear whom it was taken from. By 1062 two strategic fortified settlements on imperial roads had capitulated: Oria (via Appia) and Brindisi (via Appia Traiana).⁷⁷

After these gains, Guiscard’s progress stalled, primarily on account of the abovementioned revolts in Calabria (1062, 1064–65), in addition to large-scale rebellions (1064, 1067–68) led by his nephews Abelard and Geoffrey of Conversano, as well as members of the rival Ami family, Ami II (later of Giovinazzo) and Joscelin of Molfetta (Ami’s father-in-law). That being said, further acquisitions were seized at Byzantine expense in Apulia, but Guiscard was neither a participant nor a direct beneficiary. Peter I of Andria, from the Ami family, gained Taranto (May 1063), his son Geoffrey seized Otranto (1064) and his nephew Ami II ‘entered Giovinazzo’ in 1068. All other towns were captured by members of the duke’s family: his nephew Robert of Montescaglioso (Matera, April 1063), his sister’s son Geoffrey of Conversano (Montepeloso?) and his brother Geoffrey of Hauteville who, together with his son, Robert of Loritello, gained a foothold outside Byzantine territory in the Abruzzi (see Epilogue). Given the reliance on a few sparse entries in two Apulian annals, little can be said about the manner in which the fortified towns after Troia were acquired. That is, other than Montepeloso, which Guiscard besieged in 1068 when Geoffrey of Conversano refused to provide the ‘service’ due from the town, arguing that no ‘aid’ was owing since ‘it had been gained from the enemy’ by himself, not the duke.⁷⁸

The Byzantines, facing increasing incursions on their north-western and eastern frontiers, were unable to commit substantial forces to stem the Norman tide. As opined in the continuation of the history by Byzantine chronicler Skylitzes – copied verbatim into George Kedrenos’ early twelfth-century historical compilation – those forces still present in Italy were greatly disadvantaged by Guiscard’s relatively unlimited access to troops and money (after succeeding Humphrey as count in 1057). Appointed katepano in 1064, Avoulchares devoted himself to organizing garrisons in the southern coastal towns of the Salento Peninsula. Evidently Constantine X (r. 1059–67) was displeased with Avoulchares’ performance in Italy – probably owing to the fall of Otranto – for in the same year Perenos, despite being based at Dyrrachion (Illyria/Albania), was made ‘duke of Italy’ (and subsequently of Dyrrachion). It will be recalled that Argyros was the first to be granted this title (1051), and evidently in 1064 the katepano was less commander-in-chief, and more a military commander invested with plenipotentiary powers by the duke based in Illyria.⁷⁹

Perenos now opted for the age-old Roman tactic of ‘divide and rule’. He treated successfully with Joscelin, Ami II and Abelard. In three particularly cryptic and inelegant Latin sentences, one of the Bariot annals indicates that the Byzantine-backed rebellion began in 1064, when ‘many nobles went to Dyrrachion at the request of Perenos to receive rewards’. Afterwards, a fleet led by the duke and Joscelin engaged a Calabrian squadron of unknown size, managing ‘to set fire to’ a chelandion. In the same annals, Guiscard, some time before this event, went to Bari and secured an ‘oath’ from its nobles, which should probably be read as a pact of non-aggression to ensure his Apulian possessions would not be attacked while he was campaigning in Calabria and Sicily. The date of 1064, accepted by Chalandon, has since been questioned by Wolfgang Jahn, who argues that the revolt began in 1067. However, the problem remains that Amatus noted Guiscard’s presence in Calabria when the revolt occurred, but unfortunately did not provide a date. Since the duke was not recorded by any of the annalists or chroniclers to have been in Calabria in 1067, it remains possible that the revolt – that is, the initial part played in it by Joscelin and Perenos – did in fact occur in 1064–65, at a time when the duke was absent from Apulia for up to a year: Calabria (Cosenza), then Sicily (Palermo, Bugamo and Agrigento), and again in Calabria (Policastro and Aiello [four months]). It was therefore the perfect time to strike, and another piece of information lends some credence to this hypothesis: Malaterra observed that the citizens of Aiello, despite their lengthy, bellicose resistance, obtained a favourable peace settlement because the duke was required elsewhere. Admittedly Malaterra did not explain why Guiscard needed to leave Calabria, but it is important to stress that unless Roger played a part in them, the monk often had little to say about Guiscard’s actions in Apulia. The most obvious example of Malaterra’s general indifference to the duke’s actions when Roger was absent is that, unlike Amatus and the anonymous Bariot annals, he made no mention of Joscelin’s pact with Perenos, nor of the efforts made by Guiscard to suppress the related (but subsequent and separate) revolts of Abelard and Ami II (1066, 1067–68).⁸⁰

Having sown the seeds of a dangerous revolt in Apulia, the ‘brawn’ method, surely intended to complement the preceding divide-and-rule strategy, was launched in 1066. A fleet of chelandia led by Michael Mavrikas (or Maurix) sailed from Dyrrachion to Bari, and it contained a most important cargo: a contingent of the Varangian Guard. Owing to the general silence of the chroniclers, little can be discerned about Mavrikas’ subsequent military endeavours. All that can be said is that Brindisi was retaken – captured by Guiscard in 1062 but not regained until 1071 – and almost certainly Vieste in the Gargano Peninsula (northern Apulia). Lastly, another set of Apulian annals observed: ‘Count Geoffrey, son of Peter [I of Andria], wanted to advance into Romania with many troops, but was opposed by the commander of the Greeks called Mavrikas’. Since Romania in contemporary Latin usage meant the empire’s heartland across the Adriatic (the Balkans and Anatolia), this passage literally means a transmarine assault on Illyria, which Mavrikas’ fleet intercepted and defeated. Geoffrey, not to be confused with his Hauteville namesake at Conversano, attempted to raid the Illyrian coast: a fact confirmed by William of Apulia, who added that assistance was provided by Geoffrey’s younger brother Peter II of Andria, whom Guiscard loathed because ‘he had tried to attack Dalmatia without his permission’. The classicizing poet did not mean Dalmatia (Croatia) literally; rather, he used the correct term for the southern half of the Roman province of Illyricum (modern-day Croatia and Albania). While Chalandon suggested that Dyrrachion was the target, it seems more likely that the port in closest proximity to Otranto (Geoffrey’s base since 1064), Avlona (95km south of Dyrrachion; 92km north-east of Otranto), was the intended destination. As Guiscard did not launch an attack on Avlona until 1081, Geoffrey’s expedition set a precedent and, despite its failure, it was an audacious enterprise.⁸¹

Luckily for posterity the annalists, unlike the chroniclers, chose to record the events, albeit in a Spartan manner. Making up for their general silence in the years 1061–68, Amatus, William and Malaterra thankfully awoke from their chronological slumber in order to record the ‘main attraction’: the siege of Bari, which the duke ardently began in 1068. The city was situated on a triangular peninsula, only the base of which could be approached by land. As the other two sides and tip of the triangular spit were flanked by the ocean, a full-scale investment by both land and sea was necessary. According to William of Apulia, Guiscard attempted to take the city by threats. Declaring that he would be satisfied with possession of Argyros’ palatial home located on the city’s highest hill – next to which he no doubt planned to build a donjon – the duke demanded that the Bariots surrender it to him. As Argyros had died earlier that year, in making this request Guiscard undoubtedly wanted to remind the citizens that they stood little chance without their most famous citizen to protect them. The demand was met with a torrent of abuse from the battlements or, as William put it, ‘the Baresi gave a sharp response’. Perhaps using the same engineers that had built siege engines for him at Troia, the duke had additional ones prepared before the walls of Bari in August/September. Collating the descriptions provided by William, Amatus and Malaterra (the first, as usual, providing the most details), it appears the siege equipment built included ‘siege forts’, ‘battering rams’, ‘various catapults’, ‘shelters’ or ‘mantlets’ (see Appendix), in addition to a wooden tower equal to the height of the battlements that was protected by two rock-throwing engines. To break the cycle of bombardment and prevent the tower from reaching the walls, the garrison burst out and drove the besiegers back to the safety of their forts. Meanwhile, given the need to prevent supplies and reinforcements reaching Bari from the sea, the Calabrian fleet blockaded the city’s harbour by means of linking ships together with chains.⁸²

Access to the fleet, arrayed in a crescent-shaped formation around Bari’s seaward walls, was made possible by the construction of one (William, ‘Lupus’) or two (Malaterra) jetties/bridges. While William suggested that a tower was constructed to prevent sallies on the single jetty, Malaterra maintained that the wharves were protected by the duke’s cavalrymen. However, despite such protective measures, the ‘citizens … cast frequent stones as well as javelins’, afterwards managing to destroy the tower and dismantle much of the bridge. Either the ranks of the naval bridge of boats had diminished, or they were scattered by tempestuous weather, for soon afterwards the commander Vyzantios managed to sail out of the harbour, subsequently defeating a force of four Calabrian galleys – two were sunk – that attempted to prevent his escape to Constantinople. On his return to Bari some time in 1069, Vyzantios again defeated a detachment of ships, capturing two in the process. However, this small victory might have been preceded by an earlier defeat. One collection of Bariot annals noted that Stephan Pateranos, whom Amatus believed was sent with Vyzantios from the capital, engaged a Norman fleet at Monopoli (40km south-east of Bari). Stephan was said to have possessed twelve ships, and the result of the engagement was expressed in the following words: ‘many men died, and others were seized by the Franks and mutilated’. Despite returning from Constantinople with extra ships, reinforcements and money, Vyzantios was assassinated in July 1070 by the cronies of ‘Argirizzo’, whose name seems to be an Italianate rendering of a Greek compound (Argyros + the diminutive itzes = Argyritzes [‘little Argyros’]). As Amatus noted that Bari was divided into two factions, it seems the parties earlier led by Argyros (no relation) and the pro-Norman Adralisto – who died in the same year as his rival (1068) – continued to be a feature of Bariot politics.⁸³

Argyritzes clearly lacked the numbers to mount a coup, for the Bariots continued to defend their city with much gusto. The defender’s spirits were once again raised when another Byzantine fleet appeared on the horizon in February 1071. It was led by Duke Perenos’ ally, Joscelin of Molfetta, who, if Malaterra’s ‘of Corinth’ is extrapolated, might have been assigned a command in the theme of Ellas (Greece); indeed a later Italian source claimed that he was a ‘duke’. Joscelin had unsurprisingly managed to become a favourite of the warlike emperor, Romanos IV Diogenes (r. 1068–71). As Malaterra put it, Romanos valued Guiscard’s nephew ‘because he was resolute in arms and versed in strategy’, and his close relationship with the emperor was also noted by William. Joscelin was entrusted with twenty chelandia laden with an unknown member of mercenaries, not to mention various riches which he had convinced Romanos would lure some of the covetous Normans into imperial service. This would not have been a ‘hard sell’, for Roussel of Bailleul, who had fought alongside Roger at Cerami in 1063, was currently serving Romanos as commander of the battalion of ‘Frankoi’. Unfortunately for Joscelin, his ship was captured, another was sunk and the remaining eighteen managed to find refuge in Bari’s harbour (Amatus specified nine captured, but all other sources mention one or two). The captured Joscelin, ‘wonderfully attired in the Greek style’ as Malaterra observed, was triumphantly paraded before the duke’s ebullient troops.⁸⁴

The commander responsible for this victory of a kind was Roger, who had recently arrived from Sicily with ‘a large number of oarsmen’. The success, however, came at a high price: when rushing to one side of the ship in order to board Joscelin’s chelandion, 150 cavalrymen repurposed as heavy infantry drowned when the galley flung them overboard due to the sudden weight imbalance. Meanwhile, Guiscard sailed south to Brindisi, either taking the town by siege or receiving the submission of its people (it was simply related that he ‘entered’). While the duke’s naval and land forces still showed no signs of being able to capture Bari, the worn out and increasingly impoverished inhabitants nonetheless chose to capitulate on 15 April 1071. The Bariots would prove to be restive under Hauteville rule, and eight years after their surrender ‘Bari rebelled, thereafter casting out the duke’s garrison’ (recaptured in 1080). Another rebellion was launched in 1082–83, in favour of the Byzantine emperor Alexios I Komnenos.⁸⁵

Campania

The surrender of Bari effectively concluded the Hauteville conquest of Byzantine Longobardia. While Guiscard’s attention was diverted by a campaign in Sicily and the suppression of revolts in both Apulia and Calabria for the next five years, by 1076 he was ready to focus on the conquest of coastal Campania, a goal all the more achievable given the duke’s access to a sizable fleet. The decision to extend control to the coast certainly made sense: the south-eastern section of the Salernitan principality conquered primarily by Guiscard’s younger brother William in the 1050s – which came to be known as the Principate – was landlocked. Moreover, Salerno was on the via Popilia, which extended all the way to the Strait of Messina. While Prince Gisulf II of Salerno had no doubt expected Norman expansion within his principality to cease when granting Sichelgaita’s hand in marriage to Guiscard, attacks continued well into the 1060s, culminating in Pope Alexander II’s excommunication of William of Hauteville and Guimond of Moulins in 1067. Guiscard’s coastal ambitions in Campania first became evident in September/October of 1073, when he agreed in return for tribute to protect the duchy of Amalfi against his brother-in-law. In the early 1070s, Gisulf launched a series of successful attacks against the duchies of Naples, Amalfi, Gaeta, and Sorrento (the latter two held respectively by Richard of Capua and Gisulf’s estranged uncle Guido). Many of these assaults were naval based, but Amatus also noted that the prince led a contingent of infantry and cavalry, with which Gisulf captured various coastal fortresses, in addition to establishing new ones.⁸⁶

Having reached agreement with the Amalfitans, Guiscard, then preoccupied with the siege of Santa Severina in Calabria, sent ships and troops to assist them, but they were defeated and captured by Gisulf en route. While Gisulf’s attacks on Naples, Gaeta, Sorrento and Amalfi attest to a continuance of Guaimar IV’s expansionist policy, his designs on the latter maritime city were partly driven by his desire to punish its people for their complicity in his father’s murder. Moreover, it will be recalled that Guaimar had extended his authority over the duchy and, as far as the papacy was concerned, the ducal title was inherited by his son (endorsed as late as 1088, when Pope Urban II referred to Gisulf as ‘our dearest son, the prince of Salerno and duke of Amalfi’). Guiscard’s intervention in Amalfitan affairs, then, was an act of aggression from the perspective of both Gisulf and Pope Gregory VII. The duke’s ambitions in Campania became even more apparent when Richard Quarrel of Capua, having supported a revolt against Guiscard in 1072–73, subsequently came to terms with him. An alliance between the rival Norman dynasties was brokered: Richard agreed to help Guiscard capture Salerno in return for assistance in besieging Naples. Both Amatus, evidently for personal reasons, and Malaterra, in order to vindicate Guiscard’s unlawful attack on Salerno, denigrated Gisulf. William of Apulia took the opposite approach, presumably because Gisulf was both ally and confidant of popes Gregory VII and Urban II (whom William claimed was a patron of his poem). Gisulf’s response to Guiscard’s treaty with Amalfi was to petition for what has been called a ‘proto-crusade’ against his brother-in-law. Already angered by Norman raids in the Abruzzi (as far as Pescara, 155km north-east of Rome), Gregory’s decision-making was also informed by a letter received from Byzantine emperor Michael VII Doukas (r. 1071–78), which requested military assistance against the increasingly serious Selçuk incursions into the empire’s Anatolian heartland. Accordingly, in the winter of 1073–74 the Pope wrote to various European princes asking for troops, stating that ‘when the Normans are pacified, we may cross over to Constantinople’, later proposing that once the Byzantines had been assisted, the army would advance on Jerusalem.⁸⁷

Guiscard was excommunicated in March 1074, and soldiers from Salerno, Rome, Tuscany and Pisa began mustering, later assembling before Gregory near Viterbo in June. By August, the duke had managed to secure an alliance with the very same ally whom Gregory was already planning to assist militarily after the Normans had been ‘pacified’. In addition to the receipt of 1,200 pounds of gold, Guiscard arranged for the marriage of his daughter Olympias to Michael VII’s son Constantine. As it turned out, the duke’s negotiations with the Byzantine emperor were unnecessary: at around the same time, plagued by disputes among its leaders (including Gisulf), Gregory’s army disbanded. Yet the ties with empire increased Guiscard’s prestige and wealth, and the fact that he relished his entrance into the imperial nobility is evidenced on his Greek seal: ‘Robert, the most noble duke of Italy, Calabria, and Sicily’. From the Byzantine point of view, Guiscard was now the latest holder of the office and titles first granted to Argyros of Bari in 1051, and hence southern Italy had theoretically been reincorporated into the empire (although Guiscard unquestionably viewed the arrangement rather differently). Michael VII’s policy of turning enemies into allies was commensurate with the attempt by his predecessor Romanos IV to attract Normans into imperial service during the siege of Bari (c. 1070). Had it not been for the deposition of Michael VII in 1078, the new relationship between the Doukas and Hauteville dynasties might well have borne mutually beneficial fruit.⁸⁸

When considering Guiscard’s next action, the words of Stalin come to mind: ‘The Pope! How many divisions has he got?’. Clearly unfazed by his excommunication and Gregory’s continuing threats, in May 1076 the duke chose to besiege the city of the Pope’s greatest ally in southern Italy. As William of Apulia put it in verse: ‘The fiery duke advances on Salerno with an innumerable army, and prepares a siege by land and by sea’. Thankfully, Amatus provided more details about Guiscard’s large army, noting that it was divided into three columns based on ethnicity: Byzantines, Muslims, and Latins (i.e. Normans, Franks and Lombards). While Salerno’s harbour was blockaded by the ducal fleet, troops invested the city’s southern walls with siege forts. Richard of Capua’s soldiers also erected forts on the other side of Salerno’s walls, linking them together with ditches and palisades. Despite the efforts of archers, crossbowmen and slingers to clear the battlements of defenders, the Salernitans continued to resist. Interestingly, none of the chroniclers mentioned the use of siege engines, so Guiscard’s strategy seems to have been entirely focused on starvation. It worked, but only after eight months and only through treachery. In December some disgruntled Salernitans opened a gate for the besiegers. Their prince promptly took refuge in the Torre Maggiore (castello di Arechi; Fig. 13), a fortress perched on the mountain overlooking the city: ‘no stronghold was considered to be this impregnable in any other regions of Italy’, declared William of Apulia. Gisulf now bombarded the besiegers with stone-throwing catapults, a shot from which struck Guiscard in the chest, inflicting a serious but non-fatal injury. But the prince’s provisions were all but finished, forcing him soon afterwards to surrender himself and the Torre Maggiore to the duke. ‘Robbed of the honour of Salerno’, as William of Apulia saw it, Gisulf went to Rome, where he became an important member of the papal court. Evidently impressed with the well-fortified, wealthy trading city, Guiscard decided to make Salerno his capital, arranging soon afterwards for the rebuilding of the cathedral which, along with the castello di Arechi, is still one of Salerno’s greatest tourist attractions.⁸⁹

According to Amatus, towards the end of the investment of Salerno Richard of Capua advanced on Naples (although Salernitan annals stated that he in fact began the siege in May 1077). True to his word, Guiscard sent Calabrian and Amalfitan ships to blockade the city by sea. As Alexander of Telese observed in the late 1130s, Naples had since time immemorial been a sizable city surrounded by precipitous walls, and the approach to its south was not only guarded by this formable defensive perimeter, but by the Tyrrenhian sea. Accordingly, the city was considered to be ‘impregnable’; it could only be captured by ‘famine’. Despite Richard’s great military ability, his attempts to capture the city were thwarted at every turn. Not only were the Neapolitan infantry and cavalry able to inflict defeats on the prince’s forces, but they also managed to destroy his siege forts. Having received Gisulf’s surrender at Salerno, the duke went to Naples with extra ships and troops. Despite spending thirty days there, Guiscard was unable to help the Capuan prince capture the city. Leaving his fleet with Richard and some soldiers, in December 1077 the duke advanced with the rest of his forces on the city of Benevento, whose citizens had acknowledged papal overlordship since 1073. The siege lasted until April the following year, and Richard of Capua provided assistance to the duke. The city was circumvallated by siege forts, and Pope Gregory’s response in March 1078 was to excommunicate ‘all Normans taking pains to attack the land of Saint Peter’. While incursions into papal territory were cited as the reason for the sentences of anathema (renewed, in Guiscard’s case), Campania was also mentioned. It is evident that the capture of Salerno informed the Pope’s ruling in regard to the duke: Gregory later reminded Robert of the possessions ‘which you hold wrongfully’ – Salerno and Amalfi – stating in 1080 that he would wait patiently for Guiscard to return them to their rightful owner (Gisulf, whom in 1082 the Pope referred to as ‘our most dear Prince of Salerno’). Richard had little time to ponder how many divisions the Pope had, as he died one month after his excommunication. He left to his son a powerful and extensive principality, and Jordan, having entered Gregory’s good graces, abandoned the siege of Benevento. Without the military support of the new prince of Capua, Guiscard chose to follow suit. Jordan also discontinued the siege of Naples, although he retained direct control over much of the Neapolitan hinterland – e.g. Pozzuoli in the west, Acerra to the north – and seems to have secured tribute from the duchy in return for abandoning the lengthy siege.⁹⁰

Presumably after the capture of Salerno, the duchy of Sorrento was absorbed into Guiscard’s Campanian possessions, for in addition to other peoples subject to his rule, the duke mentioned Sorrentines in a diploma of the late 1070s. By c. 1078, then, the duke possessed Salerno, Sorrento, and Amalfi, while Jordan of Capua held Gaeta and Naples (not directly, but as a protectorate). These acquisitions marked the end of eleventh-century Hauteville expansion in Campania, although Naples, Gaeta, and the principality of Capua had all been annexed to Roger II’s kingdom by the middle of the following century.

1. Basilica San Nicola di Bari Frieze (c. late 1080s–90s), Puglia (Paul Brown). Late 11th-century Norman frieze; note various spear grips, different armour types, and banner.

2. Bayeux Tapestry (c. 1070s), Scene 51 (Wikimedia Commons). Observe various spear positions, saddles, banners, and archers.

3. Theodore Psalter, F.145r, 1066 (British Library). Byzantine cavalrymen in battle; note throwing of lance (left) and underhand thrust (right).

4. Theodore Psalter, F.63v, 1066 (British Library). Byzantine cavalry formation featuring lances, maces, axes, and bows; armour is scale or lamellar.

5. Monreale Cloister Capitals, c. 1174–82, Palermo, Sicily (Neville Finlay). The couched-lance position; Sicilian knights engage enemy cavalry.

6. Theodore Psalter, F.182, 1066 (British Library). Slingers were a common troop type in both southern Italy and the Byzantine empire.

7. Kantara kastron, Cyprus (Paul Brown). Almost certainly founded during the reign of Nikephoros II Phokas (960s).

8. The lofty town of Enna, Sicily (Wikimedia Commons); known as Castrogiovanni in the 11th century.

9. Palermo, Sicily (Paul Brown). The bay in which Guiscard’s ships defeated the Palermitan fleet (1071).

10. Taormina, Sicily (Ian Brown). Castello saraceno, perched high above the fortified town; taken from an even loftier fort (Castelmola).

11. Taormina, Sicily (Ian Brown). The vast harbour was blockaded by Count Roger’s fleet in 1079.

12. Ortigia, Syracuse, Sicily (Paul Brown). Castello Maniace (after Byz. general George Maniakes). Ortigia’s southernmost tip guarding entrance into the Porto Grande (left).

13. Castello di Arechi, Salerno (Wikimedia Commons). The Torre Maggiore above Salerno. Gisulf II retreated to it in 1076, as did Roger II’s garrison in 1137.

14. Castel Sant’Angelo, Rome (Wikimedia Commons). Pope Gregory VII sought refuge in this formidable fortress when Henry IV captured Rome (1084).

15. Mosaic of George of Antioch, la Martorana (1140s), Palermo, Sicily (Paul Brown). Roger II’s formidable ‘admiral of admirals’.

16. Mosaic of Roger II, la Martorana (1140s), Palermo, Sicily (Paul Brown). Commissioned by George of Antioch, the Greek inscription reads: ‘King Roger’.

17. Akrokorinthos, Corinth, Greece (Wikimedia Commons). The lengthy walls of Akrokorinthos, a fortress with commanding views of the Saronic and Corinthian gulfs.

18. Monemvasia, Peloponnese, Greece (Wikimedia Commons). The aptly named Monemvasia (‘one entrance’); a large Sicilian fleet was unable to seize it in 1147.

19. Castello Svevo, Bari, Puglia (Paul Brown). Built in 1132, demolished in 1155 and 1156; it was reassembled during the reign of Frederick II.

20. Basilica San Nicola di Bari, Puglia (Wikimedia Commons). Construction began in 1087; it was later used as a temporary keep (1155).