early m e d i eva l trading worlds was Haithabu (Figure 20.1), across the Schlei from Schleswig, its later medieval successor. Godofrid's action was a decisive step in the rise of this emporium which linked Scandinavia, the Baltic, and the North Sea economies. That the king did this, and that his risky application of violence seems to have worked, to judge both by the fortune of Haithabu and the utter disappearance of the old Baltic emporium, implies that the Danish king had at least a passing understanding of commerce and how to funnel it for his own profit. That the royal annalist tells the story in this detail, and that Charlemagne was angry about it, implies that both Franks also grasped the economic implications.23 The point is only driven home two years later. When a Danish fleet attacked Frisia and collected one hundred pounds' tribute from the Frankish empire's trading people par excellence, the outrage finally drove an aging and ill Charles to the drastic decision of attacking Denmark directly.24 Charlemagne's personal interest in commerce had already been made clear when he used it as an economic weapon against an Anglo-Saxon king ten years earlier. And it is hard for me not to see a practical grasp of the potential economic gain behind the complex and dangerous effort to subjugate Venice, even as the conflict with Denmark was escalating.25 If one accepts that importantly, if not exclusively, Carolingian coins were meant for commerce, then the remarkable restoration of royal control over the coinage effected by Pippin III, Charlemagne, and Louis the Pious points in the same direction. Even if one doubts this position (on which more below, Ch. 23.1), there is no denying a probing analysis of Carolingian charters that touch on markets. So far as we can tell, the Merovingian kings and the first Carolingians were indifferent to markets. That changed with Pippin III and his successors, who got very interested in places of commerce, precisely because of the new revenues they supplied.26 2. Trading worlds beyond the Carolingian empire The places with which we have detected communications are what, compared with Frankish Europe, we might call the "developed" economies of the early Middle Ages. The Islamic lands and to a lesser degree, the more deurbanized Byzantine empire were societies with a full panoply of well-developed currencies, and juridical and administrative institutions, and which boasted permanent markets and real beyond the carolingian empire 23 As well, of course, as they grasped the political issue of a satellite people who had been pillaged in the Danish expedition; Ann. re^ni Franc, a. 808, p. 126 ("quod [emporium] ... magnam regno illius com-moditatem vectigalium persolutione prae-stabat"), and a. 8og, p. 128. 24 Ibid., a. 810, p. 131. 25 Embargo on Anglo-Saxon merchants: Gesta Font, 12,2, p. 87; Venice: see e.g. R.295-Contra: Latouche 1966,166. 26 Endemann 1964, 38-42. Figure 20.1. Haithabu, aerial view from the south. This gives a clear idea of the nature of the site at the narrow base of the Jutland peninsula, on the long tongue of the Baltic known as the Schlei. The half circle of trees marks the wall put up in the tenth century; it enclosed an area of c. 24 hectares. The second founding of this trading town lit a fuse which exploded in war between Charlemagne and the Danes; the Frankish ruler's famous elephant died south of here during that campaign. In this place communications from an envoy and member of the Byzantine imperial family (Figure 8.1) arrived alongside freshly minted dirhams from Baghdad (A20 was discovered in a shipwreck off the beach), and Carolingian missionaries and merchants rubbed shoulders with troops of enslaved Christians. Courtesy of the Archäologisches Landesmuseum der Christian-Albrechts-Universität, Schloss Gottorf, Schleswig. early medieval trading worlds cities. They command our attention. But they were not the only societies with which the Carolingian empire was geographically disposed to trade, and not all roads to the developed economies led through sunny Italy or Spain. Some roads ran to the mistier trading worlds to the north and the east of the Frankish empire The northern arc in particular, as we have seen, stretched toward Central Asia. To the southeast, the Danube and Balkan corridors cut through forests and mountains toward the Bulgarian empire, the Black Sea, and Byzantium's doorstep. The south: Mediterranean trading worlds If much remains unclear about the directions, rhythms, and flow of trade in the Islamic lands in the eighth and ninth centuries, the situation is a little better for Byzantium. Our first task will be to characterize the main places and features of trade in both economies. We will then turn to the new evidence supplied by the sea bottom. Shipwrecks indicate some of the cargoes afloat on the early medieval Mediterranean, and supply new evidence that tends to confirm, in several instances, the changing patterns of shipping routes that issued from the entirely independent data of communications. They will show that goods traveled along at least some of the same routes as our travelers, strengthening the suspicion that the changing patterns of communications are connected with commerce. By way of a very approximate order of magnitude, one estimate guesses that, a century or two after our period, in terms of inhabitated area and population, the Muslim economy was roughly double the size of that of Byzantium.27 One suspects that Byzantium, then at its medieval zenith, had gained ground on the house of Islam since the ninth century. Recent historical work is developing a new vision of the Arab economic expansion which attaches less importance to the role of trade. In its place, scholars are emphasizing the Arab "green revolution" which boosted agrarian productivity. Even so, market-oriented industry, particularly rural and urban textile production in early medieval Egypt, is reckoned to have played an important role, against the backdrop of what is thought to have been a slow demographic increase in the Muslim world, from the seventh century forward.28 And, whether or not we assign it the primary role, the signs of trade are unmistakable. 27 For an interesting discussion and rough estimates of the relative sizes and populations of the Arab and the Byzantine empires, see Issawi 1981, 392: though the gross areas of the ioth-C. Arab and nth-C. Byzantine empires were very different (9 or 10millionsq km vs 1.2 millionsq km), the inhabited areas were much closer (2 million sq km vs. 1 million sq km). He reckons the Arab population as 35-4° million (30-35 million? c. 750: p. 387) vs. 15-20 million for the Byzantines. 28 For a succinct description of this new vision, Shatzmilleri994,43-50, emphasizing the contributions of L. Bolens, A. Watson and G. Frantz-Murphy. See also below. 582 EARLY MEDIEVAL TRAU1 in u vvui^jjo It is easy to enumerate the main places of the Islamic world liable to link with the west.29 The cities which lay on that world's northern fringe concentrated the wealth of their own regions and conveyed that of more distant ones. Large internal migrations from the Arab and Iranian heartlands helped fuse the far-flung edges of the new empire. The rise of locally specialized craft centers and general easeofcommunications.includingthe transport of raw materials "greatly stimulated long-distance trade in the prize products of the Muslim towns, such as textiles, metalwork, pottery and leather. Long distance trade in foodstuffs also developed, to service the large populations of the great cities."30 The ease of communications suggested in the eighth and ninth centuries by Spanish, Maghrebin, and Egyptian rabbinical consultations with the Gaonic academies in Iraq foreshadows the intensive flow of information and letters mirrored in the later records of the Cairo Genizah.31 Al Andalus, particularly its northern city of Saragossa and its capital Cordova, would have been worthwhile destinations for European merchants interested in the products of the new economic and cultural empire. Saragossa lay less than 300 km from Barcelona, a Carolingian bulwark since 801, and so was especially inviting.32 Across the Mediterranean, Tunis and Alexandria were important centers at the juncture of land and sea routes, not to mention other places springing up further inland, like the burgeoning garrison and government towns of Kairouan and Old Cairo.33 What we can see in North Africa and Egypt suggests a 29 See in general Abulafia 1987; Constable 1996b, 1—51. 30 Abulafia 1987,412. 31 See Mann 1917,477-88. 32 In the period following ours, Almeria, a ioth-C. foundation, and Seville seem to have led Andalusian trading centers, followed notably by Cordova: Constable 1996b, 17-23. The Carolingian sources (below) scarcely allow us to see where merchants went beyond Saragossa, on which see M. J. Viguera, EI 9 (1995): 36-8; on Cordova and Seville: C. F. Seybold and M. Ocana Jimenez, Ei 5 (1986): 509-12, and J. Bosch Vila and H.Terrasse, EI 4 (1978): 114-18. See also Levi-Provenqal 1950-3,3: 335-9 and 353-95. There may have been some kind of fair in Saragossa in mid-August: this would explain why Prudentius - himself of Spanish extraction - specifies that Bodo-Eleazar, who had already sold his companions as slaves, traveled to Saragossa by that date, "miser-rima cupiditate deuinctus." Ann. Bert., a. 838, pp. 27-8. The allusion to cupiditas seems otherwise inexplicable: cf. Blumenkranz 1963,183^. For a local Christian merchant at Cordova in 850, see Paul Alvarus, Indiculusluminosus, 5, 277.1-22; Eulogius of Cordova, Memorial* sanctorum, 1,9,2.377.4-29. 33 On Tunis, e.g. G. Jehel, LMA8 (1996): 1092-5 and M. Talbi, "al-Kayrawan," EI4 (1978): 825-32; on Alexandria in general, see S. Labib, EI 4 (1978): 132-7- A survey of one excavation there concludes that in the 8th C, imported pottery disappeared and was replaced by a flood of local Egyptian table ware, suggesting the stimulating effects on local production of changes in trade patterns: Rodziewicz 1982,44-5. On Old Cairo, see the next note. 584 b b 1 u in u 1 u k CAROLINGIAN EMPIRE prosperous economy.34 Toward the middle of the eighth century, for instance, the merchants of the coastal towns of Alexandria, Tinnis, and Damietta presented an appetizing target for a governor looking for cash.35 In the second half of the eighth century, Iraq and the Persian Gulf experienced rapid economic growth. They stood at the center of an empire and economy which sprawled from the Adantic eastward toward India and beyond. Persian ships and Muslim merchants sailed past Ceylon to China itself, while a Christian envoy traveled to Alexandria from India seeking a bishop from the Coptic patriarch.36 Chinese ceramics begin to crop up in Old Cairo in the ninth and tenth centuries.37 In the later eighth and early ninth century, the Persian Gulf port of Siraf took off. The combined presence of Chinese and Iraqi ceramics in its ruins testifies to the ambit of its commerce.38 In the ninth century, Iraq became the empire's economic, as well as its political center of gravity.39 Toward ancient Mesopotamia traveled a continuous flow of the merchandise of east and west, "from India, Sind, China, Tibet, the lands of the Turks, the Dailamites, the Khazars, and the Abyssinians," according to a ninth-century resident.40 The list of goods imported there in the mid-ninth century is a long one. It includes Yemen incense, silk, cinnamon, paper, ink, and ceramic from China, sandalwood, ebony and coconut from India, fine textiles, papyrus, and Judaean balsam from Egypt, felt from the edges of North Africa and Armenia, prepared fruits and nuts from around the Muslim world, paper from Samarkand, slaves, chain mail and 54 There are gleanings in Talbi 1966, e.g., 41-4, on cultural links with theeast. OnEgyptand Old Cairo, see Kubiak 1987, whose topographical approach illuminates how this administrative city mushroomed (76-94), but only alludes to its economic circumstances, its harbors (whose toponyms hint of different arriving goods: 117-18), trading streets (e.g. the Saddlers', attested c. 785, 112) and the short-lived canal linking the Nile to the Red Sea (118-20); cf. Scanlon '994- Brighter flashes of commercial life c. 810-20 come from the papyrus archive of textile merchants of the Fayum and Old Cairo: e.g. Marchands d'eroffes, 5,1. The Arabic History 0/ the Patriarchs of the CopticChurch ofAlexandria, 5.97, between 744 and 768. The often valuable Greek and Coptic materials for this history were translated into Arabic in the 10th C. and assembled in their present form in the nth: Graf 1947,300-6. 36 Ashtor 1976,107-8; HistoryqfthePatriarchs, 5.36-7; on the commercial development of the Persian Gulf, e.g., Szymariski 1991. On the impact of the Arab economy and trading links to India and beyond, Chaudhuri 1985, 34-52. 37 T'ang splash ware (from China), gth-C. deposit, Fustat-B: Scanlon 1989,4on49; two pieces from this "proletarian quarter," dated 10th C. and pre-1050: ibid., 47-8; see also Gyllensvard 1973 and Scanlon 1971. 38 Hodges and Whitehouse 1983,I34~4i;cf. also, idem, 1996,174-5, and esP- Rougeulle 1991. 39 See in general, Abulafia 1987,402-23. On the rise of Iraq, e.g., Mantran 1991,161—6; Ashtor 1976, 86-90, on its demographic growth; Hodges and Whitehouse 1983, 126-31 and 151—6 on the archaeological evidence. 40 Ya'kubi, Lespays, tr. Wiet, p. 4. 585 helmets from the Khazars.41 We know from a Byzantine raid that merchants still congregated in Antioch in the first half of the ninth century, which hints at the continuing economic importance of Syria.42 The general economic tendency in this vast economy, or series of interlocking economies, probably swerved more than once between the seventh and the tenth centuries. The initial decades of conquest and reorganization must have seen important shifts, as large amounts ofwealth suddenly changed hands. When the enormous Persian treasury and countless private and public treasures of the Roman empire fell to the victors, a very important dethesaurization probably triggered a sharp spike in demand.43 Substantial numbers of Arab immigrants migrated to new lands and founded new cities, attracting to them also a service population comprising large numbers of natives, if one can judge from the case of Old Cairo.44 Pilgrimage will have helped to sustain the immense patterns of movements set in train by the conquest and migrations. In the eighth century, settlement and industrial growth - metal and stone production - jumped sharply at the head of the Red Sea, according to the most recent archaeology.45 The ancient fiscal structures of the Roman empire, not to mention the Sassanian inheritance of Persia, underwent redirection toward the top, if not at their base and in their human staffing: in the new empire's western lands, until the eighth century, accounts were kept in Greek by people who look to be the descendants of late Roman bureaucrats in language, culture, and religion.46 Initial generations of improvization finally yielded a remarkable new and flexible monetary system that fused the Roman legacy of gold with the Iranian one of silver.47 No later than the Abbasid dynasty, a general economic upswing was underway, a "boom," as it has been called.48 When the economic expansion ended has been variously dated to the second or third quarter of the ninth century.49 The level of exchange sustained permanent markets in the main cities. Merchants are well attested inside the economy of the Caliphate, and its very size 41 The anonymous treatise attributed to Al Djahiz, A Clear Look at Trade, tr. Pellat. On this text, see Lewis 1977,13, andMiquel 1973-88,1: iio-ii. 42 R459; on the city's final decline, in the 9th and 10th C, Kennedy 1992,194,196-7; cf. too Ashtor 1977,254-6. 43 For the traces left by the dethesaurization on 7th-C. monetary history, see Grierson i960, 259-64. 44 Kubiak 1987,79-82. 45 Avner and Magness 1998. 46 Theophanes, a.m. 6199,1.375.31-376.7 and a.m. 6251,1.430.32-431.3. 47 Grierson i960.9th-C. papyrus records from Egypt prove that transactions there involved silver coins as well as gold: Marchands d'e'tojfes, e.g., 5.1.10.4; 29.14-16. 48 Ashtor 1976,70-114; "boom": 115; cf. Hodges and Whitehouse 1983,130, about sea trade. 49 2nd quarter: Hodges and Whitehouse 1983,149-51 and 160; 3rd quarter: Ashtor 1976,115-21. ind dynamism probably contribute as much as cultural factors to explain why so few Muslim traders were tempted to wander off to the far less developed puropean markets. Periodic markets supplemented the permanent markets, including one at a time and place which made it very attactive to westerners.50 This was the fair oflerusalem, the holy site toward which Willibald, Bernard, the future doge of Venice, and so many other western Christians traveled in our period. A European eyewitness of the 680s reported that the fair attracted an "almost countless multitude of different peoples from everywhere." It was still aoing strong in the ninth century.51 That the fair was connected with the feast of the Exaltation of the Cross (14 September) meant that its conclusion coincided with the most favorable season for sailing to Europe against the prevailing winds.52 These markets featured some goods that were not otherwise available to the Prankish empire. None was really indispensable to life, so demand for them in the west had to be culturally conditioned. To that extent, demand was elastic. They were typically very high value items: silks and other textiles, perfumes, spices, incense, and the like. Exotic foods perhaps might also be mentioned, as new edible plants spread through the Islamic world on a schedule which is obscure but which might extend back into our period.53 Then there were the new drugs that announced the birth of Arab medicine (Ch. 24.3). The southern Mediterranean also continued to produce and transport the ancient staples of oil and grain, which were regionally available also inside the Frankish empire (Tables 20.1-20.7). Egyptwas still manufacturing and exporting papyrus on a substantial scale.S4 Linen was also a major export.55 Across the frontier, Byzantine economic expansion may have gotten under way in the ninth century. Things were certainly looking up by the tenth.56 Trade 50 For an informative overview of cross-cultural and economic research into the history of periodic markets, see De Ligt '993.1-32- 5' Gil 1997,241,citinglbnMasawayh. 52 The Frankish pilgrim Arculf said it began on 12 September. Typically, his reason for telling us about the fair was less its economic significance than a miraculous rainstorm. This flushed away the dung heaps left by the merchants' camels, horses, donkeys, and oxen and which, he sagely observes, posed no small annoyance to residents' noses and feet. Adomnan, De Iocis sanrtis, 1, 7-13, CCL175.185.25-186.59; quote: "Diuersarum gentium undique prope innumera multitudo." Sailing season: above, p. 451. 53 Watson 1983; Al Djahiz, A Clear Look, 113, tr. Pellat p. 159, also mentions two plants which were reported to have been imported into the Caliphate from Byzantium. 54 Lewis 1974, 90-4. Grohmann 1954,1: 63-71. 55 Mackieig89, 82. 56 Opinions differ on this score. Kaplan 1992, 529-40, argues for demographic stability or growth, and extension of the arable down to the middleofthe 9th C; particularly in light of his arguments for technological stagnation, this implies that the 586 587 early medieval trading worlds played a part, particularly in the capital, toward which ships converged from the north, the east and the west, as one observer put it in 8oi.57 The livelihoods that benefited from Irene's tax breaks sketch the main sectors of trade in the capital-poor hunters and fishermen, swineherds, sheep merchants, wine dealers, butchers, weavers, coppersmiths, leatherworkers, dyemakers, spice dealers and architects, woodworkers and goldsmiths all rejoiced at the new arrangements.58 Then there were the large shipowners from whom Nicephorus I extracted forced loans.59 Although the volume is at present indistinct, scattered texts show that ships and commerce flowed along two main axes pivoting on the capital. One reached toward the north and the Black Sea, and seems to have developed considerably in our period.60 By the later ninth century this stream branched off in three directions. Toward the northwest was the Bulgarian empire and the Rus traders beyond it. Northward, across the Black Sea, on the doorstep of the Don River and the Khazar realm, stood the fortified outpost of Cherson over which Theophilus (829-42) tightened Constantinople's grip.61 We may well have heard an echo of the businessmen established there in a homily by St. Constantine-Cyril (Ch. 7.2). Overland trading links connected this region and its merchants to central Asia. We have already seen that, whether or not they were interested in trade, Jews immigrated to the Khazar kingdom from Baghdad, Constantinople, and Khorasan, and extended the network of connections.62 Finally, toward the east, ships made along the southern shore of the Black Sea toward the Caucasus and its mountain passes leading into Iran. In the 790s, merchants from Amastris had been active further east atTrebizond, the meeting place of shipping and caravan Footnote 56 (cont.) general economic trend was broadly similar. But the conclusion has not found widespread acceptance: see Kazhdan 1994, esp. 87-8; Harvey iggs, 250-55, opposes to this recent archaeological surveys which point to demographic decline in 7th-C. Greece, and further evidence for renewed growth by the 10th C. See also Harvey 1989, 207-19, for the recovery of some towns and growing interaction with the countryside from the 8th and gth C, and Haldon 1994, 79-80. 57 Theodore Studite, Ep. 7,25.53-4. 58 Ibid., 25.57-26.68. 59 On the conditions which funneled wealth to Constantinople: Hendy 1985,554-69; on the possibility ofa differing commercial regime in the provinces, see Oikonomides 1993,658-9; commercial installations of the medieval city: Magdalino 1996,19-25, and 58, on the beginning of Constantinople's recovery in the 8th C. Forced loans: Theophanes, a.m. 6302, 1.487.11-13 and 17-19; as an attempt to stimulate trade: A. Laiou, ODB 1:489. Patlagean 1993 argues that Constantinople was a "port of trade" in the technical sense ofPolanyi. 60 Lilie 1976, 276-g. 61 ConstantineVII, Deadm. imp., 42, 1.182.27-184.54; Sicilian coins at Cherson: above, p. 507. 62 Zuckerman 1995a, 251; cf. Ch. 7.2. For a Jewish trading diaspora reaching toward China, see Haussig 1989. beyond the carolingian empire routes from the interior of Anatolia and Iran.63 A generation or two later, a local man hailed Amastris as the market place where Byzantines and northern barbarians ("Scyths") met. Probably around 800, a hagiographer described a trip that reflects this shipping route along the southern coast of the Black Sea (cf. Map 20.3)- His purpose was to investigate relics and St. Andrew's travels, and to escape the Iconoclasts. In so doing, he sailed to Nicomedia, Daphnousia, Herakleia, Amastris, "Dorape," "Karousia," and Sinope.64 By the tenth century, customs duties on trade at Trebizond, particularly in textiles, were producing around a thousand pounds of gold for the imperial treasury.65 In 911-12, the imperial authorities at Constantinople assumed that the bulk of precious substances used in the Byzantine fragrance trade entered the empire atTrebizond.66 The history of textile production in central Asia appears also to reflect this Black Sea axis of trade. In the ninth or tenth centuries, the silks produced there underwent a marked transformation in decoration and technique which has been ascribed to the imitation of Byzantine textiles, presumably exported over one of these routes.67 The second main trading current reached south into the empire's maritime heart, the Aegean Sea and, probably, beyond.68 It may have been gaining strength in the eighth century. By then the important fair at Ephesus had sprung up (R229). Slave sales in the Aegean were active enough around 800 to elicit a surcharge from an emperor looking for money (R303). From the Aegean, one suspects trading links along the southern shore of Asia Minor, toward Attaleia and beyond. Cyprus, which paid taxes jointly to the caliphs and the Byzantine emperors, must have attracted merchants from both empires to its neutral ports.69 Early in the tenth century, taxes on imports and sales of Muslim booty and slaves in the Byzantine port at Attaleia were producing some 300 to 400 lb. of gold.70 In the ninth century, a regional shipping hub linked Cyprus to the southeast coast of Asia Minor (Ch. 18.3). These two currents of trade met at Constantinople. Together, they defined the main axis of maritime trade within the Byzantine empire of the early Middle Ages. As an Iraqi observer noted in the later ninth century, ships from the Black Sea and 63 Merchants: e.g. R225; R415; market: Nicetas David, Laudario S. Hyacinthi (BHG 757), 4. PG, 105.421C-D. 64 Epiphanius, Vita Andreae apastali (BHG 102), PG, i20.22iB-224A;cf. Beck 1959, 513. 55 Ibn Hawqal, tr. M. Canard, in Vasilievand Canard 1935-68,2.2.416-17, with Vryonis "-)?', is-i6andHendyig85,174-5; Oikonomides 1993, 652-4. The sum is •'Pparently an annual one. 6 Leo VI, Liber Praejecti, 10,1-6, no.470-7. 67 Muthesius 1997, 94-100. 68 Malamutig88,2:538-43. 69 McCormick 1998b, 4ing; and, in general on its status, Browning 1989, III, and Kyrris 1984. Cf. the anecdote from the later 7th C. preserved in al Mas'udl, discussed by Canard 1964, 49-50. 70 Ibn Hawqal, tr. M. Canard, Vasilievand Canard 1935-68, 2.2.414-15. Ibn Hawqal's first edition gives the date a.d. 913, the second a.d. 932. 588 589 early medieval trading worlds the "Syrian" sea plied the central segment of these two routes, between the Bosporus and the lower Aegean.71 Archaeology does not yet yield so coherent a picture, but some ofwhat can be discerned fits this pattern. Amphoras continued to be produced and shipped. But the general impression they make contrasts with the late Roman situation. In the eighth century, a multitude of small centers of production, many of which do not obviously continue the late antique ceramic series, appears to prevail.72 Some amphoras sustain the two main commercial currents that can be deduced from the texts. A late seventh-century Cypriot imitation of Syrian amphoras (type Bii) has been found both on Cyprus and at Constantinople.73 A series of amphora types that runs from the ninth or tenth to the twelfth or thirteenth centuries documents the northern current. What appears to be the oldest type (Giinsenin I) occurs all around the Black Sea and at Constantinople, and then trails out along southern Asia Minor as far as modern Alanya, about halfway between Attaleia and Seleukeia.74 A ship that sank off Crete (Table 20.5, ship 31) also carried this type. The production site has just been located at Mount Ganos (mod. Gazikoy, Turkey) on the Sea of Marmara, where seven new shipwrecks loaded with tens of thousands of amphoras have come to light.75 The prominence of Cyprus and the links between Byzantine and central Asian silk weaving provide a first suggestion. This the customs' take at the Byzantine ports of entry of Attaleia and Trebizond confirms: trade linked the economies of the Muslim world and the Byzantine empire, at least in periods of peace.76 Byzantine merchants seem to have been active on the Nile sometime between 715 and 718.77 According to an eighth-century Syrian hagiographer, a Byzantine victory over the Arabs resulted in a seven-year truce, during which merchants from both states flocked to each other without interference and Byzantine pilgrims headed untrammeled for Jerusalem (R146). If this information is reliable, the merchants may have been traveling overland, for the fact that the last bubonic plague reached Constantinople from Sicily and Africa shows that direct shipping links with the Muslim Levant were not operative in 743-4 (Ch. 17.1). ' Abbasid legal treatises in any case imply that, after that date, foreign merchants 71 Ibn Khurradadhbih, tr. De Goeje, 76. 72 Although it must be conceded that a large proportion of the amphoras are still unidentified and typologies are incipient at best. 73 Arthur 1986, 659. 74 Giinsenin 1989, 269-71. 75 Giinsenin 1998, where she mentions dates of the 10th and early nth C. 76 See in general the discussion of Canard 1964,48-56. 77 History o/thePatriarchs, 5.68-9, describes how the governor of Egypt hunted down "strangers," people lacking a passport, their boats, and Romans on the Nile, whom Canard 1964,49, plausibly identifies as merchants. Canard's dating of the episode also explains it, for it occurred in the midst of the mobilization for the great assault on Constantinople, when there was good reason to curtail contacts with the enemy. beyond the carolingian empire were active in the Caliphate, and some enunciated measures to facilitate and protect their operations.78 By around 850, the Byzantine empire had become the supplier of a series of highly desirable goods to the much larger economy centered on the caliphal capital. A list of the most select imports was drawn up in Iraq ;1t that time: prom the Byzantines' country [we import]: gold- and silverware, dinars of pure gold, medicinal plants, gold-woven textiles, abrun [?], silk brocade, spirited horses, female slaves, rare copperware, unpickable locks, lyres, hydraulic engineers, agrarian experts, marble workers and eunuchs.79 The detailed discussion of how to discern different qualities of silk textiles - sheen, designs, colors - includes the imperial workshops' silk and shows the importance and value of the silk exports. It also suggests that Arab merchants encountered competing imitations, presumably from the same Byzantine source.80 We can detect the return flow of this commerce in the first decade of the tenth century, when the Byzantine silk trade with Syria and Iraq was closely regulated. Syrian merchants brought their textiles, spices, incense, and dyestuffs to the Byzantine capital. At the Byzantine end, other Syrian immigres who had resided in Constantinople at least ten years were involved in the transaction. The merchants from the Caliphate were allowed to stay in the capital for up to three months, selling their wares and acquiring the kinds of goods we have just seen. The arrangement might have been new in the 88os.81 The caravans of Trebizond show that some part of the trade between Byzantium and the Muslim world traveled overland. Tenth-century Byzantine generals used as spies merchants who traded across the land frontier.82 But some of these external and internal trading currents traveled in ships. The vessels themselves have recently begun to emerge from the sea and tell their own tale, beginning with their cargoes. Although the numbers are still relatively small, and the dates, origins, and many identifications uncertain, we cannot neglect these extraordinary new data. Since ceramic typology of the early Middle Ages - and therefore dating and provenance - is still developing, it is wiser to cast a wide net, without losing sight of the probability that some of the ships went down before or after our period. The recent corpus of shipwrecks lists forty-six vessels which Probably date from between the seventh and the tenth century. 590 78 Discussion in Canard 1964,50-1, confirmed for North Africa by Ibn Sahnun, R.5I5. 79 Al Djahiz, A Clear Look, 14, my translation from Pellat's French, p. 159. ;So ^d., 12, p. 158. 1 Leo VI, LiberPraefecti, 5, 94.263-96.295, and Reinert 1998,130-5, who stresses that the development was recent. For the range ofgoods ofthe myrepsoi, seealso A. Kazhdan, ODB 3:1627-8. 82 De uellitatione bellica, 7,162.14. Seealso Canard 1964, 50, on Armenia; 52. 591 early medieval TRAJJIXNVj wwiv^uo The geographic distribution of the ships makes an important point. Notwithstanding the far superior archaeological recording that prevails in the western Mediterranean, Tables 20.1-7 show that most (59 percent: 27 of46) early medieval wrecks occur in the eastern Mediterranean basin. This contrasts sharply with antiquity, when the western Mediterranean dominates overwhelmingly.83 This sample thus reinforces the general opinion that the eastern Mediterranean was economically more developed than the western basin in the early Middle Ages. The ships also cluster along the northern rim. This reflects both ancient shipping patterns and those of modern prospecting.84 Most wrecks present some evidence of cargo, chiefly in the form of amphoras: indeed, ampho-ras are sometimes all that remain. Ships carrying materials other than amphoras are archaeologically less visible, so that the wrecks do not give a true picture of the range of cargoes actually transported across the sea. Just as in antiquity, the textiles, lumber, grain, spices, or slaves will almost never show up on the sea bottom. Even so, this first summation strikingly conforms to some of the patterns of communication uncovered in Part IV. Tables 20.1-20.4 summarize the evidence from ships which sank in the western Mediterranean. Tables 20.5-20.7 record the wrecks found in the eastern basin. Each table, save the last, corresponds to a shipping route or zone, as they emerged from patterns of communications; the ships are given in chronological order. When the cargo or other evidence indicates where the ship was last loaded before she sank, or when the crew's possessions or other evidence points to a home port or region, it has been possible to deduce all or part of the ship's route.85 The numbered segments on Map 20.2 identify the route; the location of the individual numbered wrecks appears on Map 20.3. Here the ancient centrality of the main trunk route running toward the eastern Mediterranean is obvious. The clustering of the wrecks in the first half of the seventh century corresponds to the penultimate stage in the contraction of the late antique shipping world. Significantly, four out of the five seem to have had home ports in the Byzantine east. Just as revealing, eastern ship 3 appears to have been carrying a west Mediterranean cargo and working what looks like a western regional circuit. One could not hope for a better illustration of our earlier conclusion that, as antiquity ended, eastern ships predominated even in western waters. So far no wrecks unambiguously document the persistence of this route in the 83 Cf. Parker 1992, 6-7. 84 Only three wrecks have been located along the southern rim: Parker 1992: Dor, no. 367, 600-700; Israel, no. 525, 500-600 and Marsa Lucch, no. 660, 500-650? On the superiority of northern courses, Pryor 1992, 20-4. 85 The nth-C. vessel excavated at Serge Limani, Turkey, shows how much well-preserved wrecks can reveal. Its broken glass, orpiment, dried fruit, recycled amphoras, pork, and weights allow Van Doorninck 1991 to deduce it was carrying Greek-speaking merchants from Syria. 592 beyond the carolingian empire n c 00 0j o Oh 3 O J* a g E -s be c o "c3 o -a c H .M u 15 cd E tu a a ts uj j < H rg o a. c c 0£ wi c o '■P fO u O — II Z O tu to a U - O O o m O o CL O Cl QJ E £ ro U 3 S To V tq o o O ,_ ro o CO CO o_ •Jf C c ro ro rt3 QJ \ f pi &» in cm \ % 1 % ■ M s r a\ 3. 1—, t/ u V 1_ - ^ sL 0 _g -5 c o a p. is a U iO Qj < QJ g » u c a. -5 -a 11 5 - a n aj , a u 3 9. £J -a ■3 ° S -a 5 .5 FJ P o TO O a. c c TO — o t-f a. 1 l 2 O in a o Cn o Cn OO 1— + + rsi j* c 3 qj a °: O--' -C oj il i 3 3 .5 r= I * ■n _ c « ? o o c cn Oj < X} 0} TO ^ QJ E ° £ y o > Cl OJ c ij ^ (13 o " „ 1 "s o cc: _aj < 3 X) Xj O X> O O E l/l 4_f tu 5 JS >■ CL ljT o o o o o CTi iv. ' 1 i— ro 'u O O ^. TO >■ >. r^. 1/1 z> Q. Q. > f>- c c c C 33 10 TO - y, Oj ° ™ CL C E =S Q o o o ^ ;.- 3= < o o X) o- c t -o 11 5 a: o o beyond the carolingian empire IB O a. c e ks 0 Cj Is 2 w a o u a? £ 1 5 2 1, O 3 u 3 £ o + + ed C clj a. "c QJ OJ CJ Ol (v o c C C E c c o CZ ra ra o o >- im >~ o >- m CD CD lo CD a? o o o o o a- E o o o o o u o o o o 3 £ a. > H Cr, TABLE 20.6 pastern Mediterranean: early medieval shipwrecks in the Cyprus shipping zone Date Name, location of wreck Cargo Cargo Parker provenance Route no. 400-650 Cape Andreas A, east tip 35 450-650? Cape Andreas C 36 450-650? Cape Andreas E 37 500-650 Arwad Reef B, Syria 38 500-700 Thalassinies Spilies, Cyprus 39 600-700? Cape Andreas B 40 600-800? Cape Kiti A, near Larnaka 41 800-900? Syria roof tiles 17 amphoras Riley LR1 and LR1A Aegean ?Cyprus 17 ?Syria ?Cilicia "Byzantine baluster" amphoras, terracotta sarcophagi fragments rilled, pear-shaped Byzantine amphoras Byzantine globular amphoras Aegean? (LR2?) Byzantine amphoras (like Riley LR13), tiles, ?glass vessels Byzantine amphora fragments 17 18 17 17 17 jars 1.7x1.2 m; Arab 3-handled Caliphate 18 amphoras Source: Parker 1992 any cargo from the opposite, western basin. One ship (wreck 22) probably was from the Aegean Sea or Marmara region itself, judging from the cooking ware, while the possible evidence of a kosher pot might indicate a Jewish captain in a ship transporting Palestinian cargo (wreck 21). Some of the amphoras ofwreck 22 were reused; it has recently been suggested that this ship was not a freelance merchant, but an ecclesiastical shipper and very late case of the state transport system.92 Texts and shipwrecks thus begin to delineate some main currents of trading in the two great economies which lay to the south of the Frankish empire. But there were other economies, whose stage of development more closely approximated that of Frankland, and to which the geographic obstacles to communication were Perhaps less oppressive. 9a V,ln Alfen 1996, 208-13. 202 204 206 59 1,145 203 212 1,125 EARLY MEDIEVAL TRAUinu LE 20.7 ern Mediterranean: various early medieval shipwrecks :k Date Name, location of wreck Cargo Cargo provenance Route Parker no. 400- -650? Porto Longo, Sapiendza "Byzantine potsherds" Byzantine empire Trunk (11) + ? 889 500- 625? Neseber B, Bulgaria amphora fragments, including 1 globular Greek graffito 16 + ? 738 500- -650? Marsa Lucch, Libya amphoras, Riley LR8a, over 6x2 m; pottery funnel N.Africa? 20? 660 500- •600 Israel pottery, 6th C; ballast stones; 2 m high 18 525 600- 700 Dor, 40 km S. of Akko local storage jars, on rope, straw Palestine 18 367 z: Parker 1992 West and east: new trading worlds Across the Channel from Carolingian Frankland, the burgeoning Anglo-Saxon kingdoms had long welcomed Frankish exports. Hamwic was an important emporium in Wessex when Willibald and his family began their epic pilgrimage. There they boarded the ship that took them to a market at Rouen, the first stage in a long trip to the Levant (Ch. 5.1). Excavations have uncovered Hamwic's trading links to Frankland, and we shall see more of them shortly (Chapter 23). The Slavic lands to the east mostly lack written records. But, as Frankish power expanded eastward, they surely offered, at the least, the usual wealth of "forest economies," notably slaves. So far, this eastern trading world is visible chiefly through Frankish eyes, and we shall return to it as well in the next chapters. Further to the southeast was the growing power of the Bulgarian empire, toward and through which the new overland corridor connected the Frankish empire to the shores of the Black Sea. More than just ambassadors and armies marched along those routes. For the central European corridor, the Raffelstetten inquest documents mostly local commerce on the upper Danube. It mentions salt, foodstuffs, wax, slaves, and horses as the goods that merchants and estate workers transported through this area around 900.93 At the other end of the corridor, exchange between the 93 Rafjelstetten Plea, i, 4-7, g, pp. 250-2; see Mitterauer 1964 on local trade and transport. For a possible allusion to Jewish traders in central Europe in the 840s, see above, ch. ^75. protobulgarian empire and Byzantium was sufficiently important that in 812 the khan added a clause regulating it when he proposed renewingan earlier treaty (c. 716). He suggested that traders dealing with both countries be recognized by means of documents and seals; any who lacked them would have their goods confiscated by the two countries' fiscal authorities. The treaty was rejected, but not because ofits commercial provisions.94The khan's newproposal shows that Byzantine trade was important to the Bui gars and that money was to be made by placing it under stricter state control. A very similar treaty was finally signed three or four years later.95 Early in the tenth century, Bulgarian merchants were importing linen and honey into Constantinople. There they purchased purple and other silk textiles, including, apparently, ones which had been imported into the Byzantine capital.96 Bulgarian merchants seem also to have exported slaves to the Byzantines.97 In the light of the disastrous wars triggered by the exclusion of Bulgarian traders from the capital, the trade in honey, linen, and textiles shows that c. 900 at least, Constantinople was one pole of Bulgarian-Byzantine commerce, feeding the north-south current we have observed above. Debeltos (mod. Stari-Debelt, Bulgaria), just inland from the Black Sea on the southern bank of the Sredecka river, was an important frontier post staffed with Byzantine trade officials (Jcotn-merkiflrioi) no later than 832/3.98 A second pole was Thessalonica. Seals of kommer-kiarioi established there begin in 712/13. By around 800, a new official appears. He was apparently modeled on the toll collector at Abydos for he was called the kom-merkiarios and abtjdikos ofThessalonica, and his office implies that trade was voluminous enough to warrant taxing. The quantity of kommerkiorioi seals has been taken to reflect the town's economic activities. If this is so, they surge in the ninth century; the ubydikoi and related seals end in the tenth.99 Some of this trade may 94 Theophanes, a.m. 6305,1.497.16-498.4; Besevliev 1981, 249-50; cf. 198-9. 95 Although its text is only partially preserved, the surviving segments have been taken to show that it essentially followed Krum's proposals, including those on trade: Nadpisi, no. 41 and Besevliev 1981,276-9. Oikonomides 1988 has proposed an ingenious explanation of the 812 and 716 clauses, according to which controlled trade (not tribute) was the intent of the earlier one, and the proposals of 812 aimed to restore earlier commerce interrupted by war. Whatever the interpretation, the 812 Proposals emphasize the importance of commerceatthatdate. g6 Liber Praefccti, 9,6,108.439-49; cf-t00 Koder's translation, ibid., p. 109. Cf. ibid., 5,1, p. 94.264-6, which suggests that for some textiles, the Constantinopolitan merchants were acting as middlemen, retailing cloths produced in Syria or elsewhere in the Caliphate. On the Byzantine silk industry in general and the technical terms used in the Greek text, Muthesius 1995, 255-314' 97 Besevliev 1981,414. g8 CBSDO i: 172, with no. 76.2; Oikonomides 1988,31; TIB 6:234-5; kommerkiorioi: Ch. 18.2. 99 See CBSDO 1:50; cf. Oikonomides 1992b, 247- 604 1^ j| 605 have traveled southwestward via the Gulf of Corinth route, for an imperial fiscal arrangementaround goo seems to connectThessalonica to Kephalenia (Ch. 18.2) Geography suggests that some trade also came from Bulgaria. In the tenth century, Constantine Porphyrogenitus reckoned Thessalonica a natural point of departure for the Danube and the Bulgarian town of Belgrade.100 So too we have already seen that the papal envoys of 869 who entered the empire at Thessalonica had traveled overland via Bulgaria (R592). The clincher comes in the observation that two of the kommerlriarioi of Thessalonica known from seals of the late ninth or early tenth century, Cosmas and Stauracius, show exactly the same names as the two greedy merchants from Greece who schemed c. 889-93 t0 Dar all Bulgarian commerce from Constantinople and shift it to Thessalonica so that they could unjustly tax it.101 A brief florescence of kommerlriarioi seals from the main rivers of Thrace connecting with Bulgaria has suggested further commercial activity.102 The trade was important enough that the Bulgarian tsar Symeon protested against the changes introduced by Cosmas and Stauracius. When Leo VI turned a deaf ear, Symeon began the invasions in which he would pile up victory upon victory for over a quarter of a century.103 In sum, by the late ninth century, at the eastern end of the Danubian route a trading world had sprung up whose commerce was valuable enough to start wars. The northern arc Scandinavia enjoys a unique archaeological tradition, reinforced by the circumstances of ancient pre-Christian burial practices and modern civic attitudes towards the law of treasure finds, not to mention the systematic exploration of the rich sites of Haithabu and Birka over several generations. All this makes the northern lands' long-distance communications with Asia, and also western Europe, the best studied of the early Middle Ages. The abundant archaeological and numismatic finds - more than 80,000 Near Eastern coins from the seventh through the early eleventh century (with significant deposits beginning c. 850) - 100 Deadm. imp., 42,1.182.15-17; cf. Runciman, ibid., 2.153-4, and Tapkova-Zaimova 1979, no. XXVI, 169-70. 101 CBSDO1:66, no. 18.38, and 68,18.44, with Oikonomides 1992b, 247 and Theoph. Cont., 6, g, 357.22-33, KCtKcos kou. UEpKEiiovTEs, as punctuated by Oikonomides 1992b, 246. On the date and circumstances, Tsankova [=Cankova]-Petkovaig68, 88-97. 102 Oikonomides 1992b, 247, referring to CBSDO, 1:125 a°d 126, nos. 44.5, a.d. 822/3? (could the unusual imperial portrait be that of the usurpers Thomas and his adopted son? Adrianople was in their hands that year) and 44.6, a.d. 838/9?; I: r33, no. 50.1 (s. ix1), and 1:108, no. 39-5> with the references on 1:107. About travel on the Thracian rivers, see Todorova 1984-47- 103 Theoph. Cont., 6,9, 357.30-4; cf. e.g., Browning 1975, 58-67. |iave been plausibly construed into a vast "northern arc" of pre- and proto-his-toric contacts and trade. Paradoxically, some contemporary experts seem more and more cautious about deducing trade from the increasingly abundant numismatic ancj archaeological remains. Reacting perhaps against simplistic assumptions of earlier generations, they see instead plunder, tribute, or simply more neutral "imports" in the foreign objects discovered in northern soil.104 Others continue to see trade as the primary vector of goods.105 Another important qualifier to earlier enthusiasms is the growing evidence that some Islamic-looking material presumed to come from the caliphal heartlands actually was produced on their marches, closer to the northern homelands.106 It is growing clearer that the exchange networks (to use a neutral term) with the east were complex and, in many cases, involved uncertain numbers of intermediaries, as well as varying degrees of violence. Nonetheless, some of the abundant material remains of eastern origin surely came byway of trade. For this we need not rely on the later and controvertible evidence of the sagas. We can turn to the unimpeachable testimony of contemporary and well-informed Arab observers, as well as the provisions for Rus merchants active at Constantinople codified in the Russo-Byzantine treaties of 911 and 944, not to mention the growing number of excavated settlements whose features seem hard to explain in other terms.107 The same is true for the relations with the west, since the Frisians seem to have linked Scandinavia to the Frankish empire. King Alfred's milieu and the Old English Orosius supply moreover the contemporary testimony of Ohthere and Wulfstan, two men who indubitably were northern traders.108 The webs of exchange began around the North Sea. A first network linked the southern reaches of England and the Continent to the Jutland peninsula. At the Frankish end, it reached into the lower Rhine and Meuse rivers, and particularly the Frisian trading settlements of Dorestad and of Domburg, whose Arab coins have already caught our attention (Ch. 12.2). In Jutland, it connected to the emporium of Ribe, on the west coast, and Haithabu, on the east. A very large market or craft zone (at least 200x50 m) was laid out at Ribe in 720/1; activities continued there into the ninth century.109 The trading settlement at Haithabu was apparently peopled by travelers from the west c. 750-800, who settled there on J°4 Oriental coins from plunder or tribute, not trade: Sawyer 1982,124-6, and 123 on improving hoard evidence after c. 850; "imports," not trade goods: Jansson 1988, 566; ibid., 569, on the number and general chronology of coins; see also Ch.12.2. I05 Most insistently Hodges and Whitehouse ■983,101-68, and esp. the systemic overview in Hodges 1989. But see also, e.g., Steuer 1987a, 113-17. 106 Imitation Kufic coins, possibly from Volga Bulgars: Jansson 1988, 572. 107 For the Arabs, see below; Rus treaties: R780 andRSn. 108 Frisians: Lebecq 1983,1:249-71; Old English Orosius. 109 Feveilei994. BEYOND THE CAROLINGIAN EMPIRE the Baltic shore of the neck of the Jutland peninsula. As the presence of Ohthere inrJ Wulfstan at Alfred's court makes abundantly clear - to cite only that example - even the southern reaches of Anglo-Saxon England were involved in this North Sea economy.110 The distribution patterns of traded or transported goods are especially distinct for raw materials, but, generally, they hold also for finished products. They reveal that the Dorestad-Haithabu route reflected a North Sea zone of exchange, even as Haithabu's Baltic connections are unmistakable and, in some respects, predominant. Western goods generally diminish as one moves east into the Baltic and north toward Sweden.111 To judge from its cemeteries and overall archaeology, Haithabu grew steadily from the eighth century until its decline in the tenth century; at its largest it may have comprised 800 to 1,000 permanent residents, alongside a substantial number of transients.112 Recognized Carolingian exports to Haithabu include wine and Rhenish ceramics.113 The pottery, however, was more for the personal use of rather numerous visitors than a commercial commodity.114 A later ninth-century Frankish visitor heard psalmody arising from a crowd of chained slaves who were being dragged past him at I laithabu. He recognized and redeemed a Christian nun.115 These were probably not the only slaves for sale in the history of Haithabu. The Christian slaves on view farther east, at Birka, decades before, must have transited through the Judand peninsula.116 Some textiles might also have traveled toward Haithabu, but the natural conditions do not preserve them well; most of what has been found appears to be of local manufacture, except for grave goods and some possible imports from the west.117 Querns from the Eifel region ofGermany, glass and the no Lebecq 1983,1:139-163 and 225-47; Jankuhn 1986,115-17; 119-83; 204-5. in Steuer 1987a. 112 Jankuhn 1986,204-5. "3 Ibid., 150-2. U4 Janssen 1987 stresses that imported 115 Prankish ceramic is significant, but less voluminous than previously believed. Its use is confined mostly to what appear to be higher status zones of the site. Overall, Frankish import ceramic constituted 7 percent by weight (90 kg) of the total ceramic finds uncovered as of that date at Haithabu; the other 93 percent was made locally or comes from the Baltic, lanssen's careful evaluation leads him to conclude that these finds represent not an 116 export ware per se but the residue of per- 117 sonal-use materials of a limited group of people present at Haithabu from the final decades of the 8th C. to the end of the gth (and beyond). These people had strong connections to the Rhineland around Cologne. VitflRimberti (BHL7258), 18, pp. 95-6. Cf. Jankuhn 1986,142-3. The fact that she resorted to psalmody might mean that she spoke a different language from the Frankish prelate-Protoromance, Anglo-Saxon or Irish. A more sinister interpretation might suggest that the lord archbishop had turned a deaf ear to the entreaties of his enslaved countrymen until the woman signaled her religious (and social) status. See below, nizg. HSggiggi; Jankuhn 1986,148-9. 609 early medieval trading worlds famed swords were also imported from the Frankish empire.118 Conclusive evidence of grain imports is lacking, although some southern foodstuffs have turned up (e.g., walnuts, also used for dyeing). But it is for now impossible to go beyond the observation that Frisians purchased cereals along the Rhine, that a ninth- or tenth-century dock at Hamburg was covered with a thick layer of grain and that during a famine, Charlemagne forbade the sale of food outside the empire.119 Toward Haithabu flowed also raw materials and products from Norway, Sweden, the Arctic, and even the east.120 The picture bears comparison with the west Slavic settlement of Starigard/Oldenburg, some 80 km to the east.121 The flow of eastern goods into the Baltic delimits a second zone. Here the main sites are Birka, Gotland, and various places along the Baltic shore, such as a recently published craft and trading settlement of the ninth century, near Gdansk. Among the Polish finds were ten complete dirhams, two Sassanian coins and fifty-four fragments. The dirhams were issued between 767 and 815.122 Such settlements communicated with the Slavic hinterland and the chain of trading towns associated with the Rus which, inland from the Gulf of Finland, ran south and east. In places like Staraya Ladoga and Beloozero, Scandinavians appear in the ninth century. Eastward these travelers continued down the Volga toward the Khazars and beyond them toward the Caspian Sea. Southward, the Rus boated down the Dnieper and took control of Kiev, en route to the Black Sea and Constantinople, which they firstraided in 860.123 Baltic amber seems to have been reaching Iraq toward 850.124 Before 885, the Rus were already purveying beaver furs, black fox, and swords as far as Baghdad.125 In the early tenth century, other Arab observers would add slaves, especially young women, to these merchants'wares.126 If texts are so far the main source for exports to the Middle East, the earth has yielded a broad array of durable goods which traveled in the 118 Ibid., 154-5; 158-9 and 161-3. ng Lebecq 1983,1:26-8; Jankuhn 1986,140; MGHCapit. no. 44,4,1,123.2-3. 120 Jankuhn 1986,153-4; 156-8; 163-9. 121 Gabriel 1988 and 1991. 122 Jagodziriski and Kasprzycka 1991. 123 See in general, Franklin and Shepard 1996,3-138- 124 It is mentioned by al Kindi (d. 861), as quoted by al Biruni, as well as in a 9th-C. dictionary, and information on it increases considerably in the early 10th C, according to Lewicki 1963,3-5, knowledge of which I owe to the thought-fulness of Jonathan Conant. 125 Ibn Khurradadhbih, in Pritsakigzi, 256 (trans.); 253 (text).Thissegmentofthe text is in a jumbled position, within the famous account of the Radhanites, and was already so in his source: e.g. Pritsak 1971,243-4. For the date of the treatise, see below, ch. 23n7o. 126 E.g., c. 903-4, Ibn Rusteh, trans. Wiet, pp. 163-4 mentions the Rus' slave-hunting raids among the Slavs; cf. Ibn Fadlan, trans. Canard p. 115. The most comprehensive collection of Arabic material remains Jacob 1891, here 6-17. 610 beyond the carolingian empire Opposite direction. They include the tokens of trade: coins, folding scales, and weights. Oriental ornaments, rock crystal and carnelian (a reddish chalcedony) beads, silk and metal-worked textiles and probably even clothing, including, certainly, the Islamic innovation of buttons, figure next to double axes (which Scandinavians adopted with such relish that they served as the distinctive weapon of those in Byzantium's imperial service) and domestic utensils. Find sites of eastern wares diminish as one advances westward across the Baltic.127 They diminish, but do not stop. We know that ninth-century travelers, including Carolingians like the missionary St. Anskar and his followers, sailed to Birka. They presumably traveled via Haithabu, and so crossed from the western-dominated zone to the eastern one.128 The Christian slaves they found at Birka must have come from the west; whether any of them were sold further east we do not know, but there is no evidence against it.129 Brooches, buckles, and the like reached Haithabu from the eastern edge of the Baltic and beyond. A silver belt fitting found there seems to come from the land of the Khazars or the Volga Bulgars.130 Folding scales of a type associated with the Caliphate spread generally throughout Scandinavia in the last decades of the ninth century, apparently in tandem with the large-scale penetration of dirhams. The connection between scales and Arab coins makes sense in terms of Bernard's eyewitness report that cash payments were always weighed in the Caliphate (Ch. 5.1). The scales have been found at Haithabu, where some seem even to have been produced.131 Distinctive fine weights triumphed across Scandinavia at the same time. Over 100 such small polygonal objects have been discovered at Haithabu, some bearing pseudo-Arabic inscriptions.132 Crystal and carnelian beads probably imported from the eastare also well attested there.133 Silk, eastern textile trim, and a button 127 See the systematic catalogue and discussion in Jansson 1988. 128 Rimbert, Vita Anskarii (BHL 544) 10-11,14, 19-20, MGH SRG, pp. 31-3,36,39-46. 129 Ibid., 11, p. 32:"Multietiamapudeos captivi habebantur christiani, qui gaude-bant iam tandem se mysteriis divinis posse participari." 130 Müller-Wille 1988,741-62 and 778. 131 Steuer 1987b, 462-7 (his Type 3), with list 6. nos. 33-44 (p. 524). 132 Cf. Steuer 1973,13 and 20, which states that they first appear in the late 9th C., at the earliest, but does not explicitly identify the grounds for this date; and Steuer 1987b, 470, Abb. 7. On the regional metrical considerations that explain the frequency of glass weights for coins in the Caliphate, Grierson i960,248-57; Bates 1991,53-60. 133 In addition to Jansson 1988,584-91, cf. von Müller 1970, who dates the Haithabu finds to the first halfof the 10th C, because similar beads show up in two graves so dated in Sweden, and Arrhenius 1978, who gives stronger archaeological evidence for a date in the later 9th C. for Birka and argues, from technical grounds, for the presence of eastern gem cutters, noting also, however, the possibility of a Scandinavian source for the raw material. According to Al Djahiz, A Clear Look, 15, trans. Pellatp. 159, Iraq's bestcarnelian came from Yemen. 6ll imji jvl r. u i t v /\ l i iv^ ' tiEKDiNiu int. UAROLINGIAN EMPIRE have also turned up, though they are no more precisely dated than "Middle Viking" (c. 850-1000).134 And we have already observed the Islamic coins deposited in Viking contexts of ninth-century Frankland and England (Ch. 12.2). So goods flowed northward from western Europe and from the southeastern lands. To some degree, they overlapped at Haithabu. At least one distinctive product of western Europe made it across the entire length of the northern arc, to users in Baghdad: the famous Frankish swords (Ch. 25.1). With every decade, the picture of communications and exchange has grown more detailed. Excepting swords and Arab coins, few objects can be traced along the entire complex of routes, between the Frankish empire and Baghdad. But there is strong ninth-century evidence for a northern trading world comprising a series of interlocking exchange zones. ^ 1 On every side, the Frankish empire was flanked by trading worlds which were coming to life. Whether we look to the great economies of the south, or the nascent ones of the North Sea and the northern arc, or Bulgaria, the pace quickens in the eighth and ninth centuries. It is true that, at least in the Byzantine world and along the northern route, the regional component of those trading worlds was pronounced, and perhaps predominant. One suspects something similar in the vast Muslim economy, or economies. Yet the economic regions overlapped. In the course of the ninth century, these different trading worlds began to intersect in new ways and in new places. The clearest example is the Black Sea and its hinterlands. Rus traders and raiders from the northern arc reached it and Constantinople itself, sailing past the shores of Bulgaria. Others trekked overland toward the Caspian Sea, beyond which lay Iran, and Iraq, the magnet for the wares of the world. Byzantines occupied Cherson on the northern shore of the Black Sea, while northerners crossed that sea to trade with Byzantines on its southern coast. But this thickening web of connections, of commerce - of communications -was not confined to the Black Sea or Iraq. They merely happen to be the best-documented places. Syrian traders were immigrating to Constantinople by 900, and ships loaded with Cordovan and other Spanish goods were sailing along the Riviera at most a few decades later. The proverbial "big picture" of trade across the Muslim world, from Spain to the Persian Gulf, implies similar webs of trade, whether wide or narrow, linking the regions along the southern rim of the Mediterranean Sea. And, as the route by which the last bubonic plague reached Constantinople attests, these webs had taken shape since 746. The southern and eastern places of prosperity in North Africa, Egypr> Syria-Palestine, Bulgaria, and Byzantium are the same ones to, from, or through 134 Jansson 1988, 639 and 640, with 606. 612 which our travelers moved. In general terms, the cargo-laden ships that sailed ilong these routes and sank confirm what the aggregate evidence of communications depicts. This brings us a step closer to recognizing a link between communications and commerce. The Carolingian empire was surrounded by a series of distinct trading worlds which were now beginning to intersect and interweave. Are we to believe that it alone went unwashed by currents of trade? No one contests that the Frankish world traded to some degree toward the northern arc: people using pottery manufactured in the Rhineland sojourned at Haithabu; we know that the places that supplied relics to Frankish hoards changed between the seventh century and 800, and came to feature particularly the Arab Levant, Constantinople, and Ephesus; we know that suddenly, around 775, Arab coins began to flow into Italy, at the same time that communications in general surged with the distant shores of the Mediterranean. It is time to return to the Carolingian merchants. 613