4 The pottery
4.1 The Early Medieval pottery at De Geer and Roomburgnext section
This is the first publication of the Early Medieval pottery at De Geer and Roomburg (see tables 3a, 3b and 4). Two phases have been distinguished. The first covers the Merovingian period from the fifth century up to about the mid-seventh century. This is followed by the Dorestad period which ends between 850 and 900 with late Badorf and early Pingsdorf ware.
Table 3a. The rim sherds of Early Medieval pottery from De Geer (typology: see
Van Es & Verwers 1980
). W = wheel thrown; H = handmade.
Table 3b. The rim sherds of Early Medieval pottery from De Geer (typology: see
Van Es & Verwers 1980
). W = wheel thrown; H = handmade.
The sherds from the latter period were determined in detail as to (sub-)type and fabric using the typology presented in Van Es & Verwers (1980; 2009). For the Merovingian period we still lack a detailed pottery typology that can be applied more generally to the settlement assemblages in these parts. We shall distinguish just three principal forms which comprise several (sub-)types and fabrics. For a more detailed classification of the pottery from De Geer the reader is referred to Bakker 1997.
4.2 Comparison of the Early Medieval pottery from settlements along the Rhine
In table 5 the wheel-thrown pottery from the Dorestad period is divided into three groups. This distinction is in part chronological. The earliest group, comprising types W V-VII, IX, X and XIV, dates from between c. 650/675 and 750/775 and the two others are dated between ca.750/775 and 875/900. There may be a slight chronological difference between these two, in the sense that W IIIA and B occurred widely even in the second half of the eighth century while the Badorf ware of types W I, I/II, II and IV is mainly ninth century (Van Es & Verwers 2009, 295, table 36). However, the difference is a mainly functional one: types W IIIA and B represent cooking pots, which are always a major constituent of crockery, while the Badorf ware, made up of amphoras, jugs and drinking cups (type W IV), will on the whole have served for the storage and serving of wine and other liquids.
Table 4. The rim sherds of Early Medieval pottery from Roomburg/Matilo (typology: see
Van Es & Verwers 1980
). W = wheel-thrown; H = handmade.
The assemblages presented in the tables are very different in size. In fact, they all pale into insignificance compared with the roughly 30,000 rim sherds that we recovered at Dorestad. The finds from the harbour zone along the Hoogstraat (HS) and those from the settlement area outside it are presented separately. In addition, the finds from the harbour are listed by excavation area (HS 0 to IV).
The pottery assemblage at De Geer is of such a size that it may well be regarded as representative. In the case of Roomburg this probably is no longer so and the finds from Rijswijk are just a sample. The data relating to Katwijk-zanderij Westerbaan are taken from Dijkstra’s (2008) description of the pottery.
Most of the other columns present data only for the Dorestad period. These are data that we gathered from different contexts (Van Es & Verwers 1985; Van Es & Verwers 1994). The absence of data relating to the Merovingian Period in these columns does not mean that Merovingian pottery was lacking in these assemblages. Apart from assemblages from individual sites there also are provincial pottery collections, comprising finds from various sites. The number of Dutch sites yielding wheel-thrown pottery from the Dorestad period is considerable (Van Es 1990, fig. 8).
4.2.1 Wheel-thrown pottery from the Merovingian period
The Merovingian wheel-thrown pottery was distributed over large parts of the Netherlands. Apart from the central Netherlands, it is found along the coast right up to the province of Friesland, where fairly large amounts were recovered from Wijnaldum (Gerrets & De Koning 1999, 96-98). It is thought to constitute almost 64% of the pottery dated between 550 and 650. Recently it was also encountered at Leeuwarden-Oldehoofsterkerkhof (Dijkstra & Nicolay 2008, 124-129) although the quantity here is far smaller. Moreover, in Merovingian and Carolingian times there is an overwhelming preponderance of handmade pottery. The wheel-thrown wares did also occur in the sandy regions to the north and south of the Rhine delta, but in Drenthe they are almost entirely absent. A comprehensive study of Merovingian pottery in the Netherlands is an urgent desideratum. In the present context, suffice it to say that wheel-thrown pottery fulfilled an important part of pottery requirements especially in the Rhine delta and certainly from the fifth century onwards. It is unclear to what extent handmade pottery was also in use here at these times. At any rate it does not seem to have played a significant role.
De Geer, Koudekerk, Katwijk-zanderij Westerbaan and Roomburg show similar pottery assemblages in Merovingian times. The rough-walled cooking pot (the so-called Wölbwandtopf) in its various manifestations accounts for about three-quarters or more of the total, and the smooth-walled biconical pots are extremely sparsely represented. The latter are overrepresented at Katwijk because there the wall sherds were also counted.
The differences between the four assemblages are fairly minor. We will refrain from broad conclusions until the material has been analysed in greater detail. The assemblages from De Geer, Koudekerk and Katwijk-zanderij Westerbaan appear to be somewhat older than that from Roomburg. For De Geer we assume continuity from the Late Roman period onwards. The category ‘various’ of Koudekerk, which is remarkably voluminous, contains some comparatively early sherds of Alzei types and pitchers that probably date from the fifth century. At Katwijk, the rim sherds of Wölbwandtöpfe include some of types Alzey 27 and 32/33. Such early sherds are unknown from Roomburg. Yet this assemblage is too small and from too limited an excavation to permit any definite conclusions. The foundation date of Roomburg remains uncertain, while for Koudekerk the fifth century is plausible. The Merovingian potsherds from Rijswijk are of course too few to present a proper pottery assemblage, let alone to reveal a foundation date.
Against the background of the widespread occurrence of Merovingian wheel-thrown pottery in the region, its absence in (the excavated part of) Dorestad is significant. Apart from perhaps a few potsherds, there is no wheel-thrown ware older than roughly the mid-seventh century. On the banks of the Kromme Rijn east of De Geer (and west of Leut?), the earliest habitation appears to have coincided with the start of Dorestad.
Table 5. Comparison of some Early Medieval pottery assemblages from sites near the Rhine and from the northern Netherlands (typology: see
Van Es & Verwers 1980
). W = wheel thrown; H = handmade; HS 0-IV = Dorestad, excavation areas Hoogstraat 0 to IV.
4.2.2 Wheel-thrown and handmade pottery from the days of Dorestad
In the overall assemblage of wheel-thrown pottery from Dorestad, the early types are the least well represented (c. 23%), while the cooking pots W IIIA and B and the Badorf ware are roughly balanced and together overwhelmingly predominate. When the pottery from the various excavation areas is considered separately, the proportions of the three categories of wheel-thrown ware are found to fluctuate. We shall not seek to explain these differences in this paper since they do not fundamentally deviate from the overall spectrum. The cooking pots W IIIA and B and the Badorf ware jointly account for three-quarters and locally even more of the wheel-thrown pottery. The fact that these two groups are roughly contemporary and cover about half of Dorestad’s chronology prompts the conclusion that pottery imports significantly increased from the mid-eighth century onwards. This is connected to the transition from the small-scale and scattered production of Merovingian times, of which our ‘early’ group represents the final stage, to production on a more ‘industrial’ scale in the German Vorgebirge.
As might be expected, the pottery assemblage from De Geer, which was part of the Dorestad agglomeration, parallels that of Dorestad. The difference compared to Roomburg and Koudekerk is obvious. There, the early types make up 40% or more of the Dorestad-period assemblage and very clearly outnumber the Badorf ware. The increase of pottery imports in Carolingian times is far less evident at Roomburg and Koudekerk. The difference is mainly due to the Badorf ware, the ‘drinking ware’. For a farming community such as Koudekerk this is only to be expected. At Roomburg, a former vicus beside the Fossa Corbulonis, this is more surprising. However, this pottery assemblage is not very voluminous and from a restricted area and is thus unlikely to be representative of the entire settlement beside (and within?) the castellum. This goes a fortiori for Rijswijk, which shares its high proportion of ‘early’ wares with Roomburg and Koudekerk. At both Roomburg and Koudekerk this ‘early’, i.e. Late Merovingian group is the tail end of a flow of Merovingian imports starting at an earlier date. The contacts pre-existed and did not need to be newly established in the second half of the seventh century. This must also apply to De Geer, but here the ‘early’ wheel-thrown wares have, in manner of speaking, been wholly overtaken by the subsequent spate of imports. The explanation may in part lie in changing relations between different types of settlement. During the heyday of Dorestad an ordinary villa may have had less direct access to the flow of imports monopolised by Dorestad.
Koudekerk and Roomburg have a far greater proportion of handmade pottery than Dorestad and De Geer: almost 50% against less than 20%. This is a Frisian feature. Whereas in the vicinity of Dorestad and upstream the handmade pottery seems to have become virtually extinct in Merovingian times, along the North Sea coast the Kugeltopf vessels continued to compete against the wheel-thrown cooking pots from the German Rhineland throughout the eighth century. The Kugeltopf vessels in Dorestad are presumably Frisian imports. They occur comparatively often in the harbour zone (Hoogstraat 0 to IV). Their high values around the mouth of the Rhine match those in other Frisian regions.
The right-hand half of table 5 (from the ‘Utrecht city’ column) should be regarded as an extra, relating only to the wheel-thrown pottery of the Dorestad period. The city of Utrecht and the coastal provinces from Zeeland right up to the isle of Texel show assemblages that are roughly comparable to that of Dorestad but the proportions of the ‘early’ category are rather low. This could mean that here too the flow of imported, wheel-thrown pottery accelerated during the heyday of Dorestad, i.e. the mid-eighth to the mid-ninth centuries. Yet as long as the Merovingian pottery from the region in question is not better understood, we should refrain from drawing conclusions. For the northern coastal zone (of Friesland and Groningen) and the sandy regions in the provinces of Gelderland and Overijssel, such an increment in pottery imports in Carolingian times is certainly plausible. Merovingian wheel-thrown pottery is rare there. Yet it appears that the western margin of Friesland (Westergo) is an exception, where at any rate locally this pottery does occur in some quantity. Westergo lay closest to the Rhine delta but here too the beginning and volume of Merovingian imports are yet to be determined. Dorestad must have played a major role in the exportation of Carolingian wheel-thrown pottery to the downstream Rhine delta and along the Dutch coast. For the interior, particularly the province of Overijssel, Deventer is a more likely transit port. This trading settlement on the IJssel emerged in the late eighth century, flourished in the ninth century and, in contrast to Dorestad, did survive the Viking raids (Bartels 2006). The wheel-thrown pottery in the province of Overijssel anyhow comes from excavations at Deventer. The assemblage of wheel thrown pottery at Deventer is somewhat ‘younger’ than that of Dorestad since it contains a larger proportion of Badorf ware (Van Es & Verwers 1985).