Comparing and Connecting: Comacchio and the Early Medieval Trading Towns

Early medieval archaeologists and historians investigating the same phenomena in different regions need to compare their results and methods. Unknown trading worlds are emerging from Africa to Rus, via the North Sea and the Adriatic. At Comacchio specialists discovered each other’s discoveries and created intellectual contacts and communications across geographic and disciplinary boundaries that still are too rarely broken. Reflecting on connections and comparisons was inevitable among so many scholars working on so many early medieval trading settlements. The remarkable information starting to come from the Adriatic excavations gains from new understanding developing out of the more established investigations in the north

Political boundaries raise the important question of royal power in creating emporia. I have argued elsewhere that, for some emporia, evidence for royal initiative is weak at best. This is very different from political authorities who sought to capitalize on or control and tax trading settlements that had already come into existence. New or old evidence undermines many classic cases of purported royal foundation: the Frankish royal annalist's assertion that the king of the Danes created Haithabu by transporting merchants there by force in 808 ignores the archaeological fact that activity had started on the site a full sixty years earlier. The princely authority once credited with the even deposit of the sand layer on which Ribe was built turns out, literally, to have been wind 3 .
At the Comacchio conference, Joachim Henning showed that Baltic peasants were capable of laying out even plots of the sort frequently attributed to royal power 4 . What about Comacchio?
The new settlement was born just as nearby Ravenna's main port at Classe was fading away 5 . Comacchio sprang up some five kilometers east of the main Roman road leading from Ravenna to Spina 6 . By that road, the exarchate's capital stood over half a day's distance to the south, about thirty kilometers, if anyone traveled overland rather than by boat. Today the open salt ponds of Le Valli lie between Comacchio and Ravenna.
If they existed in our period, at some ten km across they may well have complicated getting to Comacchio by land, as indeed the curve in the Roman road suggests. In other words, Comacchio looks paradoxically close to but separated from Ravenna.
But why should one have done business in tiny Comacchio rather than Ravenna?
It cannot have been in order to get closer to a bigger population and therefore to a bigger 3 For both cases, details in McCormick, "Where do trading towns come from?", 44-46; R. Fleming also has mounted a vigorous assault on the idea of princely foundation of the Anglo-Saxon emporia: R. Fleming, "Elites, boats, and foreigners: rethinking the birth of English towns", in Città e campagna nei secoli altomedievali, 1, Spoleto, 2009 (Settimane de studio del centro italiano di studi sull'alto medioevo 56), p. 393-425. 4 Joachim Henning discussed archaeological evidence from Rügen for this in the paper he delivered at the 360: "Et hoc item de illis hominibus, qui negotium fecerint sine uoluntate regis cum romano homine…". It is probably no coincidence that the opening paragraph of the 715 Lombard pactum with the Comacchian boatmen refers to the present restoration of times of peace in terms that could evoke recent political as well as judicial conflict: "Judicia homines uestri pararunt; nunc quidem deo auxiliante remota sunt, ut pacis temporibus pars parti perfruamur": Pactum, ed. L. M. Hartmann, Zur Wirtschaftsgeschichte Italiens im frühen Mittelalter. Analekten, Gotha, 1904, p. 123-124. Code remained the law of the land in Byzantine Italy, including Ravenna and, if any officials set foot there, Comacchio 9 .
If "offshore" and unregulated status was the attraction, Comacchio fit a broader pattern. It goes back at least to the fifth century, when emperor Valentinian III tried to outlaw what he called "stealthy business" (furtiva negotiatio): merchants were avoiding official business places and doing deals "in obscure and hidden places" where there were no tax collectors 10 . Down to the Lombard conquest, the apparatus of the Romano-Byzantine state with its well-established administrative routines, institutions and taxes ruled Ravenna and its markets. That may well be why Comacchio is not at Ravenna. Of course nature, geology and the silting of the Po at Ravenna might have had a word to say as well.
A second systemic feature follows from the emporia's marginal location and leads to the first of the material features that characterize many of them. As Stéphane Lebecque observed, the emporia tended to be relatively isolated from densely populated areas 11 . This seems paradoxical in the light of their commercial vocation. That necessitated markets which required significant populations. But it may well be connected with the fact that at the outset the emporia mostly lacked formal defensive structures. Building walls added to the costs of having an emporium; isolated sites in ecologically marginal zones needed less defending. Comacchio, Rialto, Torcello, Dorestad, Birka, Walcheren or Ladoga for instance, were all in some way completely or partially separated from their surrounding territories. This must have discouraged at least casual attempts at robbery or raiding. It also made it easier to police comings and goings to the site, enhancing the security of those active there, even as it may have lent the site a kind of politically neutral character such that, in some cases, they seem almost to lie in a kind of no man's land.
The absence of ramparts until late in their existence may provide another clue: for long the emporia enjoyed and likely required a broader sense of security for their sites to function well economically. The Vikings found the Frankish emporia easy pickings. The 9 For the laws against exporting to the barbarians wine, oil, garum and weapons or gold: Arabs had a harder time with Venice, which had faced multiple attacks earlier in the ninth century; Comacchio on the other hand did fall to them in 875 when Venice seems not to have been directly attacked 12 . It is perhaps no coincidence that when walls did go up, that tended to occur toward the end of an emporium's life-span, when the wealth long generated by a place seemed to justify the investment. Yet the expenditure apparently would not suffice to save some market settlements. The question of the relative defendability of a dying emporium's site versus that of its successor town may shed some light on the shifts from, say, Haithabu to Schleswig or from Dorestad to Tiel.
The ramparts that still mark for us the late stages of northern emporia's existences demarcate their size at that time. We need always to take into account the differing sizes of the emporia and how they may have grown or shrunk over time. As anyone who has actually visited both sites will recognize, Haithabu is physically bigger than Birka. What does that mean in economic and historical terms? Does it reflect a more thinly populated hinterland for Birka than in Haithabu's? How will Haithabu's hectares compare to those of Quentovic, Dorestad or Hamwic in the same decades? Obviously the task does not get any easier with emporia that began as polyfocal settlements such as the early Venetian nuclei or, apparently, Comacchio. But measuring the relative spatial extent and differing topographical structures of different emporia and how they changed are indispensable to gauging their growth, contraction and comparative significance. 13 This will supply some signals about how individual trading settlements may have functioned within the broader system.
Trash in substantial qualities also marks the denser human population and economic activities that characterized emporia. Their waste management practices require careful comparative study; they illuminate both the taphonomic processes that have shaped the archaeological record and the organization and governance --or lack thereof--of each town 14   We would expect refuse testifying to exceptional levels and varieties of craft-production whether it takes the form of the apparent glass wasters in Comacchio or the paws from fur pelts at Birka 16 . The concentration of wealth that occurs in a place given over to making money inclines one to expect among the household garbage signs that at least some residents enjoyed a distinctive diet that may have signaled luxury status 17 . Perhaps the trash might also shed some light on the differing social structures of the emporia. A Christian woman (matrona) who lived at Birka could have herself been a merchant or head of a merchant household, for Frideburg was very wealthy ("in saeculi quoque rebus dives") and had manifestly been to Dorestad, for she tells her daughter that, unlike Birka, it was filled with churches and clergy. Frideburg decided that her daughter should travel 1300 some km to Dorestad to distribute her bequest of alms to the many poor people who lived there. "Because," she said, "there are fewer poor people here" 18 . That could mean simply that Birka was smaller than Dorestad, as indeed it probably was. It could also, however, suggest that the distribution of wealth looked different in the Scandinavian trading town compared to the Frankish one, a proposition which differing patterns of food consumption might well clarify.  16 For later seventh-century wasters from glass blowing: M. Ferri, "Glass: production and material traces", in L'isola del vescovo. Gli  Money or at least concentrated moveable wealth such as hacksilver, gold or silver ingots or rings should be present in quantities that distinguish an early trading settlement from its surroundings. Hoards may not be surprising, but single stray coins suggest more reliably how wealth circulated in the emporium. Both give a revealing image of a place's nearer and further trading partners. This is certainly so at Dorestad. Even if Frankish authorities successfully limited the circulation of foreign issues north of the Alps, coins converged there from all over the Carolingian empire, including mints as far afield as In fact, the demand for coins in the emporia was so strong that established mints sometimes had difficulty meeting it. At least that is what the proliferation of imitation coins suggests. These are coins whose legends and design show they do not come from an official mint, but whose other features have not --thus far--suggested that their metal content was so low that they aimed at outright fraud. A number of coins that might fall into this category bear the Dorestad mint name, and imitation coins seem in general common in Carolingian Frisia 24 . An even more interesting insight comes from a new scientific analysis of Carolingian coins. It has always seemed strange that Venice struck Frankish silver pennies. As the dating formulas of Venetian documents tend to confirm, ninth-century Venice was technically a part of the Byzantine empire 25 . Indubitably authentic, the coins are nevertheless systematically lighter than the coins struck in mints that the Carolingian kings really controled 26 . Recent laser ablation inductively coupled plasma mass spectroscopy (LA-ICP-MS) of the specimens preserved in the Cabinet des Médailles of the Bibliothèque nationale de France has revealed yet another surprise: the eighteen Venetian coins contained significantly less silver than contemporary standard Frankish issues 27 . Their wide circulation in the Carolingian empire and, so far, their total absence from Venice itself -although, as mentioned, not from Comacchio, where one has turned one up in the Piazza XX Settembre excavation--seem to mean that Venice paid for whatever it bought from the mainland with Frankish coins that it minted itself 28 . The further implication is that Venice profited twice over, making the standard benefit that came from minting coins, and enjoying a discount by paying for its purchases with a coin that was not only lighter, but that contained ten to twenty percent less silver than normal Frankish coins. It seems undeniable that we have here a conscious policy aimed at maximizing profit. In terms of the typology of economic behavior developed by the economist Peter Temin, this can only signify an "instrumental" mindset, that is, one that Gelder, "Coins from Dorestad, Hoogstraat 1", at p. 218, concerning his p. 219, nos. 9-14; S. Coupland, "Dorestad in the ninth century: the numismatic evidence", Jaarboek voor munt-en penningkunde, 75 (1988), p. 5-26, at p. 18-19, challenges some of these interpretations, but there are other cases that he is inclined to admit: S.  Table 2. Eighteen of the nineteen coins analyzed produced usable results. Of these, one had 92% silver, two, ca. 90%, ten, 81.9-88.6%, five, 72.6-78%. The mode (six coins) ran between 85.2 and 86.9%. By contrast, the six contemporary coins of Carolingian Italy run between 97.6 and 93.6% fine, with one outlier from Pavia at 84.1%. 28 Illustrated in Gelichi, L'isola del vescovo, p. 19. is geared to markets that operate more or less as we know them 29 . So materials science illuminates the mental disposition of the authority who issued coins in Venice.
Intensive money use brought with it another significant feature of coin circulation in the nascent trading towns: outright counterfeiting was such a part of the monetary spectrum as to rival the discrete cheating of Venetian minters. Thus nine Islamic dirhems recovered in association with a shipwreck in the port of Haithabu: on first examination, they seemed to represent a remarkable set of dirhams struck at the same time in Baghdad and which had traveled together almost to the bank of the Schlei in Jutland. But they have turned out to be counterfeits 30 . Dorestad too was no stranger to counterfeit coins 31 .
The emporia's intensive use of money thus spawned fraud. Small wonder that touchstones appear at Dorestad. They should show up on other emporium sites as well: merchants were well-advised to test the fineness of the money that changed hands 32 .
Maybe this is another sign also that the hand of the king lay more lightly on the emporia zones, and that the currency that circulated there was less subject to royal controls. A related feature is the frequency and often sophistication of the scales and weights that occur on such sites 33 . The weight systems themselves can be tricky to reconstitute given the oxydation that often affected the objects, but they are certainly worth investigating.
Did Comacchio use a Byzantine, a Roman, a Frankish, or some other weight system or combination of systems? What does the answer to that question tell us about the economic horizons of the place, and how they may have changed over time?
Transport was essential to the trading networks that coalesced in the emporia. So the zooarchaeology of emporia should feature donkeys, mules, horses, and oxen whose relative proportions and spectrum may have changed over time. For instance, the donkeys which were scarcer than hen's teeth in Roman northern Gaul became more frequent By the same token, remains of transport containers should litter the site. There is small hope of course for perishable sacks or leather wrapping, but once in a while favorable conditions might preserve some scraps as well as baskets or boxes 41 . In a Mediterranean context, amphoras, possibly including some of local manufacture, feature prominently at Comacchio and this must also be true of its Venetian rivals 42 . Barrels probably played a greater role in late Roman trade than we are accustomed to think. The Adriatic rim seems to have been among the early epicenters of the Roman spread of the new container, as we know from texts, sculpture and even the occasional ancient or medieval barrel 43 . Classe Warehouse no. 17, which was destroyed by fire in the late fifth century, featured large numbers of amphoras, but also at least one barrel 44  Shabby temples to greed and acquisitiveness, the early trading towns were filled with valuable goods as well as money, real or counterfeit. And so it is no surprise that their residents made ample use of keys and locks to insure that their valuables were not "acquired" by other residents 52 . The human wares traded at the emporia needed to be secured as well, hence the appearance of iron shackles, which probably once figured alongside more ephemeral ropes and wooden yokes to control slaves 53 . Weapons too we may expect in some quantities 54 . Historical reports make clear that the wealth that traders transported needed defending under way. Whether they went about armed in the emporia we do not know, but we must imagine that weapons were part of their professional gear: the merchants with whom St. Anskar traveled to Birka successfully fought off a first pirate attack before losing their wares and wealth to a second one 55 .
Specialized craft production of course characterizes emporia. What was the range of types of such production and how did it change over time? On what scale did individual workshops produce and how did it develop? How many similar workshops operated at any one time? Were crafts organized spatially in specialized zones? If so was that due to special needs, for example, the access to much water and clay required for ceramic manufacture, or efforts to avoid unpleasant emissions? Could it be explained by other reasons, for instance to facilitate delivering shared raw materials or because the size of the settlement encouraged like industries to congregate so clients or suppliers could more easily find them or indeed, because of some sort of implicit "emporium tradition", some set of expectations that emporium dwellers and visitors carried with them? Did trading towns put out some or much of their craft production to sites in their hinterland?
How big was that productive hinterland? With respect to Comacchio, how much of the traders' salt was produced in and around Comacchio, and how much may have been acquired elsewhere along the coast? It will be illuminating to compare the emporia on this score. A southern raw material whose origins segregate promisingly between the western and central Alps is soapstone. The thermal qualities and workability of vessels made from it encouraged wide trading soapstone cookware aroaund early medieval Italy.
Advanced scientific methods distinguish the production zones. Soapstone extracted from the western Alps displays a distribution pattern geared to western and central Italy that may well reflect Tyrrhenian trading networks. 61 In our period, stoneware whose chemical Birka. Considering the annual rhythms of supply, it would be most interesting to try to discover the seasonal rhythms of commercial travel to, from and among the northern emporia. The next question would be to see whether, when or how they might correlate with those already established for some of the Mediterranean emporia by virtue of the seasonal patterns of southern ship movements.
As the archaeologists intensify their investigation of the Adriatic emporia, seasonality will therefore bear watching there too, through whatever window the taphonomic circumstances and resources and skill of the investigators will allow. Pollen from cores and waterlogged archaeobotanical macrorests retrieved from wet-sieving may be productive. Evidence drawn from other areas can also contribute. The timing of the arrival and departure of shipments from elsewhere in the Mediterranean will have paced the economic life of the Adriatic emporia. Written sources show that the overwhelming majority of ship movements in the early medieval Mediterranean occurred between April and October. Arrivals and departures -that is ports' most active periods-peaked in August. April and June came in second place 69 . The Adriatic rim's distance from the main basins of the great inland sea may have altered the timing somewhat, but the basic pattern probably held. 70 A second very important parameter will have been laid down by the seasonality of salt production, Comacchio's first great export. Judging from Venice's sixteenth-century records, salt production stretched from May to October, but clearly spiked in August, followed by July and then September 71 . If, as we may suspect, highprotein and easily preserved eels figured prominently in Comacchio's marketable resources in the ninth as in the nineteenth century, then the season of their harvest should have been a third factor that set the port's economic rhythms. In the nineteenth century, the peak catch of eels coincided with their annual autumn migration when they swam out of the lagoon and into the sea at the end of August and the beginning of September 72 .
So the production of two of its core wares as well as the schedule of arrivals of ships with wares from elsewhere in the Mediterranean were summer affairs. Indeed, all peaked about the same time, in August or September. But the really crucial element for making money out of Comacchio must have been the rhythms and time-scales of shipping on the Po and its tributaries, where the Lombard Pactum and subsequent records seem to locate Comacchio's main customers 73 . The wares that arrived or were produced in Comacchio in the summer and early autumn could be sold for a profit at the small towns scattered up and down the Po and its tributaries. A river transport system that covered, very roughly, some 500 km and served a good number of small markets is nothing to sneeze at 74 . The modern seasonal regime of the Po suggests that optimal travel conditions -limited flooding-would have occurred in the summer, perhaps from May to September, that is precisely when other activities were reaching a crescendo at Comacchio or indeed even before the eel harvest 75 .
We can gauge more precisely the temporal constraints imposed by the river conditions and distances if we remember that in August 949 it took Liudprand of Cremona only three days to sail more than 400 km down the Po to Venice from Pavia.
He traveled downstream at about five km an hour or over 130 km per 24-hour travel day; upstream travel probably took around twice as long, such that towing a boat loaded with goods from Comacchio to Piacenza when the current was not too strong might have taken five or six days without allowing for stops. In sum, a first rough estimate suggests that a round trip up and down the Po to one of places listed in the 715 Pactum implies a minimum of some eight to ten days. It would have grown lengthier with the number and duration of stops along the way. Although these rough and ready estimates should be refined by more detailed research, they offer a rough bench mark of what was possible.   From these considerations and convergent seasonal rhythms flow two important corollaries. The first is that because salt-producing, Po river transport, and Mediterranean shipping seasons mostly overlapped, it is unlikely that the same people made the salt, shipped and sold it upriver, and also sailed overseas to get long-distance goods elsewhere in Italy or further afield. In other words, the convergent seasonal ecology of these three activities imposed a specialization of labor, unlike what we see, for instance, with some ninth-century Rhenish potters who sold their own products up the Rhine. 78 Since Comacchiesi towed their boats up the Po and presumably made the salt, these social and economic actors must have been different and coexisted simultaneously at Comacchio.
We have, on the other hand, no unambiguous references to Comacchio ships in the Adriatic or anywhere else in the Mediterranean in the eighth and ninth centuries. It is conceivable that some, perhaps much of the wares that reached the wharves of Comacchio got there in ships whose homeport lay elsewhere. In this context, a ninthcentury report about a monk of Comacchio bears repeating. Returning from pilgrimage to the Holy Land in the 820s, the monk Dominic took a Venetian ship home from Alexandria. It is of course possible that the Venetian convoy was the first transit available. But we may also suspect that there were no ships from Comacchio in Alexandria's harbor those weeks. This might have a more general bearing on who was carrying Mediterranean wares to Comacchio at that date 79 .
The second corollary is that at least some of the wares that arrived or were produced in Comacchio in the later part of the summer only became available too late in the year to be carried immediately upstream toward Pavia and Milan. If it is true that Po shipping diminished considerably late in the season -again, a point which requires closer study--, that would imply that such goods had to be stored in Comacchio over the winter.
The immobilization of capital that this entailed would imply lesser efficiency in deploying the wealth that financed imports and production at Comacchio. At the same time, these seasonal rhythms would suggest that the local boatmen and merchants wanted to get underway as early in the spring as conditions would have warranted, to begin to turn their stockpiled wares -and immobilized capital-into profit as quickly as possible.
It may therefore be no accident that the Comacchio traders who concluded the agreement with the Lombards in 715 were in Pavia by 15 May that year.
Exotic wares have from the beginning figured as distinctive features of the emporia, so there is little need to insist on them. Nevertheless it is worth emphasizing that the emporia were economic phenomena, and that economics is not only about structures.
It is also about quantities. Hence the importance of even rough estimates of transport another, as new networks rose and old ones declined. To be able to see and track these changes, we need to quantify the visible proportions of imported and exported wares.
Most importantly, we need to quantify to be able to assess their economic significance.
Only by so doing was Janssens able to show that the Rhenish ware discovered at Haithabu, and which attracted such attention, was in fact only a very small percentage of all similar ceramic. The Rhenish vessels also showed a peculiar spatial distribution. That they concentrated in certain houses led Janssens to conclude that in all probability these finewares were not traded goods, but items that traders brought with them from the Rhineland for personal use. They thus point indirectly to the geographical horizons of the traders' activity, but are not themselves direct evidence for trade, and certainly tell us nothing about the volume of wares traded 80 . Thoughtful quantification should lead to a comparative list of how much of what was imported from how far, and how the proportions changed over time. That will begin to shed some light on the sensitivities of growing, declining, changing markets and trade networks: patterns of supply, but also of demand.
In the course of these much-to-be-desired studies, careful scholars will consider what we cannot see. Many -most?-wares likely or certainly traded through certain sites have left few obvious material traces when they were not produced on site: textiles, slaves, speciality woods, spices, furs, honey, wax, salt, perhaps even some precious ancient books if we were to suspect that a famous fifth-century manuscript of Livy was purchased at Dorestad. As unlikely a place as it might seem for a crucial link in the chain of transmission of ancient literature, the fact remains that a great ancient codex, the sole surviving witness to Livy's Fifth Decade, belonged to someone at Dorestad, whether because it had traveled there from Italy in some merchant or pilgrim's sack or because it got there in some other way. Around 800, the book was owned by a man who identifies himself, rather mysteriously, as "Theatbertus episcopus de Dorostat." 81 Dorestad was not a bishopric, but Theatbert's title raises the question of the ecclesiastical status of the new emporiaa question that was quickly settled in the Adriatic, but not in the north. 82 This important difference of ecclesiastical organization between the Adriatic trading sites and the northern ones probably helps explain why the Adriatic settlements enjoyed greater permanence than their northern counterparts. Among material features, the initial absence of defensive structures reflects the broader sense of security that commerce needs to flourish; the relative sizes and differing structures -polyfocal versus mononucleic-distinguish emporia, or groups of emporia, at least at some points in their existence. Trash in substantial quantities signals a denser population. How it was managed (or not), casts light on the emporia's organizational structures, while the presence of waster products from craft activity and refuse from luxury diets point to some of the residents' non-agrarian activities and unusual wealth.
Emporia feature an exceptional density of money or the precious metals which served as standards of value. Coins of distant origin appear alongside local mint production fired by concentrated demand for coins, a demand that encouraged the circulation of imitation and also counterfeit coins in these hotspots of commercial activity. This explains the need for touchstones in emporia, as well as the fine scales and weights suited to measuring coins and other precious substances that turn up there.
value that Carolingian scholars attached to ancient manuscripts, one cannot but wonder whether there was not some trade in them as well, including in the markets of the emporia. 82 The way Theatbertus describes himself suggests that he may have been of chorepiscopus (i.e. auxiliary rural bishop) assigned to Dorestad from the Utrecht see to which the port belonged.
Transport animals and gear, and the traces on human skeletons of heaving, hauling and carrying should appear regularly on trading sites because of the need to move wares. Elements of carts, boats and landing gear can be expected whether as derelict materials or recycled into other uses. So too some of the stouter containers such as barrels and amphoras as well as the occasional basket when preservation conditions allow.
Where barrels competed with amphoras, we might get some insight into the changing scale of trade, and dendrochronological analysis eventually will allow precise dating and localization of the barrels. How the containers were grouped and stored holds further insights into how early medieval people imagined and organized their commerce, as well as into its scale. The need to insure the security of the valuables that each trader brought or sought in the emporium implies locks and keys; the need to keep valuable human wares from running away explains the presence of iron shackles for the slave trade.
Weapons too were part of early medieval traders' kits.
Specialized craft production has long been a classic signal of an emporium. We would gain from trying to see the broad and changing constellation of workshops and their production activities in a particular emporium and its hinterland. The craftsmen's experience of a new, unfettered but also unsupported social context of production may well have helped elicit a new mentality and new approaches to production. The networks that supplied raw materials for the workshops ranged near and far, and are an obvious but still under-studied part of the picture.
Every emporium's activities surely followed seasonal rhythms. They need more investigation in each place, and we need to compare them among emporia and consider how they may or may not have been coordinated from trading node to trading node. changes, then answers will be forthcoming. Those answers will lead to the comparative understanding, and even the historical connections among these remarkable new entities that played such an extraordinary role in the origins of medieval Europe's economic networks. However short their life span, the early medieval trading towns connected and fostered, in the long run, the economic networks that they would transform and develop into the systems that spread from Europe outward in one long and tumultuous --but continuous--process of development. This is the early face of a globalization which in some meaningful sense continues even today, relayed and reinforced by new trading worlds that emerged out of the tiny, hesitant beginnings we perceive in the emporia of early medieval Europe.