Neither of the settlements has been analyzed or published in full to date. For this reason a summary description of the most important features and the context of the pottery is presented here.
3.1 Uitgeest-Groot Dorregeest
3.1.1 SITE DESCRIPTION
The settlement was discovered in 1952, when part of the terrain, a Late Medieval ‘terp’ was levelled. The large scale excavations at Groot Dorregeest took place from autumn 1980 to spring 1983 and were carried out by the State Archaeological Service (Rijksdienst voor het Oudheidkundig Bodemonderzoek), Amersfoort. The excavated area is circa 3 ha (fig. 3.1); the habitation remains date from the Late Iron Age, the Roman period and the Medieval periods (4-5th and 9th-11th century AD) and possibly represent a continu-ous occupation from the Late Iron Age onwards (Besteman 1990a,b; De Koning 1998; Woltering 1982/83).
Fig. 3.1 shows the location of the site, fig. 3.2 the excava-tion trenches and numbers. As the excavaexcava-tion results have been analyzed only minimally to this date, the description of the features and the discussion of the context of the pottery are of a preliminary nature. Some work has been done by the author on the reconstruction and phasing of house plans and on the structure and dating of the wells, as well as on the geomorphologic processes, which formed the site (Abbink & ter Schegget unpublished). The small part of the site plan with all features (fig. 3.4) shows the extreme com-plexity and density of the archaeological features in the western area along the creek, the result of a continuous digging and re-digging of ditches, postholes, wells, etc. Most of these features date from the Roman period occupation. In the eastern part of the excavation the density of features was much lower, partly due to levelling of the land in that area prior to excavation.
The settlement plan in fig. 3.5 contains a selection of fea-tures, mainly those from which pottery is included in the sample for analysis and some other associated important features. They cover most of the area with traces of build-ings and house-sites dating from the Roman period occupa-tion. Pottery from the earliest fill of the creek can be dated to the Late Iron Age. Obviously these features are only a
minor part of all that were present within the excavated area. The selected pottery is likewise only a fraction of the total amount recovered1.
3.1.2 GEOLOGY
In an earlier publication (Woltering 1983/4), the main geo-logical and geomorphogeo-logical features at the site were pub-lished. The remains of a coastal barrier are the oldest sub-stratum within the excavated area. This sandy deposit was partly overgrown by peat, which was still remaining in the southeast and southwest corners of the excavated area. The base of the peat layer was dated to circa 3900 BP (RGD2 report, unpublished). The barrier and the peat were partly eroded by a large system of creeks formed during the Dunkirk I transgression phase. The course of the creeks changed continuously, resulting in the partial erosion of their oldest deposits by later courses. The radiocarbon dates of shells from the earliest phase are all circa 2400 BP (RGD report, unpublished). Fig. 3.1 shows the creek bed at the beginning of the settlement in the Late Iron Age/early Roman period and in its final stage, when erosion and re-deposition no longer took place. When occupation started at the site the creek had changed into a fresh water gully. The earliest fill contained concentrations of pottery in the north-ern part of the site, in trenches 10, 34, 37 and dating from the Late Iron Age (250 BC-0) (see Van Heeringen 1992). Usually this pottery was found spatially and stratigraphically separated from the earliest pottery of the Roman period. The radiocarbon dates (table 3.1) suggest that the residual creek filled up rather quickly during the Early Roman period. The deposits from the Iron Age and Roman period consisted of three layers of humic sand and clay, separated by thin sterile layers of windblown sand. All three layers contained quite a lot of occupation debris from these periods, with some concentrations of pottery, for example on the banks of the creek in trench 20 (fig. 3.7a+b). Roman imports were present in all three layers as well. In the windblown sand separating the fills human footprints and hoof prints of cattle and pigs were found in several places, indicating that the creek period-ically carried little or no water. During the post-Roman period, renewed erosion and deposition took place within the same channel. The14C dates from the three different layers
63
3
The settlements Uitgeest-Groot Dorregeest and
from the Roman period do not match the excavation data very well. The calibrated dates all overlap; it is possible that the channel filled up very rapidly, perhaps within a period of 50-100 years. The Roman imports in the first three layers, however, date from the 2nd and 3rd century AD (Zoetbrood 1985), while at the end of the Roman period a much thicker layer of windblown sand was deposited, separating this period from the beginning of the Early Medieval gully and occupa-tion phase3. The archaeological data therefore indicate that the creek actually functioned during a much longer period than is shown by the14C dates. This discrepancy is not easy to explain. Possibly, but not likely, material from older phases was redeposited in younger layers, or finds from a later date have moved down into underlying layers.
3.2 Features from the Roman period occupation
From the late Iron Age occupation no structural remains were recovered. The location of the pottery from this period in the creek suggests that the habitation was located to the North or East of the excavated area. The Roman period occupation took place mainly on the Dunkirk I creek deposits. The features consisted mostly of a large number of ditches, postholes and wells. Series of palisades along the creek, made of stakes, date from both the Roman and Medieval period. In only a few locations some of the old land surface remained.
The habitation area seems to be divided into an eastern and a western part. The ditches from the Roman period cover virtually all of the excavated area. The dwellings and the wells are concentrated in the western part of the site on the
high sand deposits alongside and parallel to the gully over an area of circa 30 ≈ 40 m. Especially in the western part the extreme density of settlement traces obscures the stratig-raphy. The number of house-plans published by Woltering (1983) and here in a slightly revised form (fig. 3.5) is almost certainly only a small percentage of all structures built dur-ing the Roman period.
Except for the broad distinction between features from the Roman and Medieval period, the features cannot as yet be distinguished stratigraphically and/or chronologically. Only for some locations within the excavated areas a relative stratigraphy could be established on the basis of re-cuts and finds.14C dates are available for most of the wells (fig. 3.3 and 3.6, table 3.1). A preliminary interpretation of the type of features was made for those features from which pottery was selected for this study. Three types of ditches were present.
1 Ditches associated with dwellings:
a ditches surrounding the walls of a three-aisled farm-stead
b ditches, which served as a wall-foundation trench 2 Ditches delineating a house site (a compound)
3 Ditches which are associated with a field system; some were dug into the banks of and parallel to the gully. Considering the overall plan in fig. 3.4, it will come as no surprise that the interpretations of the features, as presented in the next section, are only tentative and by no means definite.
64
65
3.2.1 HOUSE-PLANS
A series of three-aisled house-plans, minimally ten thus far, could be identified on the ridge along the creek in the most western part of the occupation area (see also Woltering 1982/83). Most of them are indicated on the simplified plan of fig. 3.5; in only one case, house 7, there is some evidence for the length of the building. All dwellings show an approx-imately SW-NE orientation. Their relative chronology can be established only in part, based on the depth at which features first appeared and on their overlap. House 8 appears to be younger than house 7, whereas house 7 is later than house 10. The ditch 21 very likely is a house ditch around house 7. In this area there is some evidence for a second building, house 9. Ditches 16 and 34 can also be part of the structures 7 and 9. The features 16, 19, 21, 22, 25, 27, 33 and 34 were probably small ditches surrounding buildings, perhaps drip gullies. It is less likely that they represent wall trenches, because quite a lot of pottery was found in these features. Further North, in trench 20, the remains of several three-aisled houses (1, 2, 3, 4 and possibly 6) occurred, but all of these are badly preserved due to intense rebuilding in the same area or by later ditches.
In trenches 19, 35, 14 and East of the line of dwellings discussed above, a second occupation area with many traces of buildings can be identified. Feature 33 is interpreted as a house ditch for house 11. Feature 28 is also interpreted as a dwelling (see below). This building is certainly the latest in this area. In its vicinity, feature 23, consisting of many short stretches of ditches, often re-cut, was excavated. The nature of these ditches is not clear. They may represent dwellings or dwelling areas, but also field ditches considering their width and orientation. Features 20 and 23E are of an older phase, like feature 22 but both are probably younger than feature 19 and feature 32. The relationship between these features and feature 33 is uncertain. Ditch 22 could represent a wall-ditch, with feature 24 as a house-ditch, although the former contained many sherds. The short stretches surviving of ditch 23 (D) and 27 are older than all other features in this area. Both contained large quantities of pottery and a relatively large number of partial vessels. Also a dog-burial was found in one of the ditches of feature 23. Considering the evidence from other sites, such as Schagen and Assendelft, such a burial represents a ritual context and the large amounts of sherds recovered from this area would also fit in well with ritual depositions.
The features 1, 14 and 28 are interpreted as wall trenches. They suggest the presence of so-called wall-ditch structures as defined in the Assendelver Polders Project (Therkorn & Abbink 1987) and interpreted as dwellings without a stalling area. Within the selected features of the Roman Iron Age these houses appear to be of a relatively late date. They contain Roman imports which unfortunately cannot be dated precisely.
3.2.2 WELLS
A substantial number of round wells from the Roman and Medieval period were excavated within the habitation areas (fig. 3.3 and 3.6). It seems likely that wells were associated with farmsteads, but as yet it is not possible to ascribe spe-cific wells to spespe-cific dwellings.
For most of the wells a radiocarbon date is available (table 3.1). The undated wells were ascribed to an occupation period on the basis of the pottery and/or the specific construction. 16 wells date from the Roman period. Only a few of them contained pottery in any quantity, but if one did it usually consisted of complete or partially complete vessels. By far the largest amount of pottery was present in well 18-1, which also con-tained the highest number of (almost) complete vessels (n = 13). Most of this pottery shows signs of secondary burning, with some vessels being very badly affected. The total set could represent a household inventory, dumped in the well after a fire. Unfortunately, the remains of the settlement from the Roman period in the area of the well were destroyed to a large extent by the removal in the 1950's of the church mound that was built over this area in the Medieval period. Three main types of construction, each with variations, were distinguished: 1 Wells with walls constructed from sods.
There are several variations in this type of construction: a Sometimes a wooden foundation is present, in the form of
a cartwheel, a square wooden frame or matting of twigs. The wheels all date from the Roman period, as do most of the frames.
WELL number: 7-3, 7-4, 14-3, 19-1, 31-1
b In some cases the sods have a humic layer, which proba-bly represents the original vegetation. Wells of this type seem to be mostly of the Roman period.
WELL number: 12-1, 19-2, 35-2, 8-1, 8-2 (ME), 14-1, 7-3, 7-5, 18-1, possibly 18-2, 19-2, 22-5, 33-2, 35-1 (also with basket), 35-2.
c In other cases the sods do not show any structure and these wells often have vertically placed wooden posts or planks lining the inside of the sod wall. This construction seems to be typical for the medieval period.
WELL number: 9-1, 22-4
2 Wells lined with wattle, sometimes in the form of a bas-ket, usually without foundation. This type of construction dates to the Later Roman and/or Early Medieval period. WELL number: 7-1, 7-2, 14-2, 14-4, 17-1, 19-3, 20-1, 22-2 (with a wooden foundation)
medieval period occupation.
WELL number: 2-1,12-2, possibly 22-6, 22-7 3.2.3 FIELD SYSTEMS
Features 8, 11 and 44 are interpreted as field ditches. These ditches were found mainly in the eastern part of the exca-vated area, which is lower than the western part. The field systems consist of a series of long parallel ditches, roughly with a NE-SW orientation. At least three systems can be identified because of a slight change in the orientation of the ditches. A second system of orientation is NW-SE. It is striking that both the houses and the field systems show similar changes in orientations. This correlation might prove to be of importance for establishing chronology of the site. Field ditches are also present within the densely habitated area, but obviously in a much-disturbed form. The amount of pottery recovered from these ditches is generally low compared to those associated with dwellings. Considering that levelling took place in the eastern part of the excavated area, these field systems seem to be of a relatively early date.
3.3 Pottery selected for technological and formal analyses
The selection criteria for the pottery were of course insti-gated by the purpose of the analysis, i.e. to gain information about the complete process of pottery production and about the intended and actual use of vessels. The selection criteria were:
– The size of the remaining part of a vessel in relation to the size of the complete vessel, which determimes: – The possibility for reconstructing the diameters of rim and
base, the maximum circumference and the height. – A well-preserved surface and indications for use on these
surfaces.
In first instance only complete and well-preserved profiles, from which all size measurements could be obtained, were considered. For the pottery of Uitgeest, however, the criteria rarely could be met, even after much trial and tribulations; very few complete profiles or vessels could be reconstructed. Less than 20 complete vessels were excavated and most of these are small cup-like vessels which have a higher chance to survive unbroken than pots of a larger size. It was there-fore necessary to add incomplete profiles, which met as much as possible the criteria stated above, to the sample. Exceptions were made for some of the wells, like 18-1, where smaller fragments were included to complete the information about the total composition of the contents of the well. This also applies to some of the habitation features containing exceptionally large concentrations of pottery (for instance features 16, 22, 23, 27). Most of these ‘partial pots’
are rim fragments, which extend to at least the greatest diameter, and include parts of the lower wall. Also some bases with a large part of the lower wall present were added to the sample for analysis. The total number of vessels used for the complete analysis of technological and functional features is 147. This sample of 147 vessels will be referred to as pottery sample 1 (see legend fig 3.5).
Additional samples of sherds were used for different pur-poses. Because of this, the actual number of individual pots will vary slightly for the analyses of the fabrics, the mor-phology, and the use. For the technological analysis, sixty sherds were added to sample 1 to provide additional data on the quantity of organic temper and quartz. The sherds either came from the same features as sample I (n=30) or were selected from the earliest creek-fill (fig. 3.5 and 3.7a,b; n=30). Sample 2, consisting of 625 sherds, was used in the analysis of the size, shape and use of pottery in chapter 8. This sample was selected from the total population of sherds that have been drawn.
The actual sample size and composition will be presented with each of the analyses; the databases can be found in the appendices to the chapters.
3.3.1 CONTEXT OF THE POTTERY IN SAMPLE 1
The distribution of the pottery for the features discussed above is listed in the legends of fig. 3.5 and 3.6. A consider-able number of the vessels in sample 1 were recovered from the wells. This pottery is by far the most useful group as far as context information, dating, and survival are concerned. Only very little pottery was selected from the northern occu-pation area (trench 20; fig. 3.2), not because less pottery was found there—on the contrary, masses of sherds were exca-vated from the banks of the creek, as well as from the creek fill itself—but because not a single complete profile could be reconstructed out of this massive deposition.
A disproportionate large number of vessels in the analysis was derived from the middle of the occupied area (trench 19 and, especially, 35). This is not due to a higher amount or better preservation of pottery in these trenches, but to a change in excavation techniques. The usual method of exca-vation and collecting finds was to dig through the first 10 cm of features with a shovel on each level of excavation. Finds were collected over the entire length and width of a feature. A mechanical digger then removed circa 10 cm to the next level of planning and finds collection etc. This method caused pottery, especially complete vessels, to be broken up and scattered through different levels of collect-ing. Consequently, parts of such vessels will be missing or very difficult to retrieve. The large number of recent frac-tures and the impossibility of restoring complete profiles or vessels, even in cases where sherds from several levels clearly belonged to the same vessel, confirm this. The much
68 Ge es te rw eg 0 50m 505.7 505.7 505.9 505.8 505.7 109.6 109.7 109.5
Fig. 3.1 Uitgeest-Gr.D.. Site location and excavated area, with the main geological features.
Legend
1 Modern ditches
2 Boundaries of excavated area
70 20-1 31-1 19-2 19-3 35-2 19-1 33-1 33-2 8-2 8-1 14-3 14-2 14-1 14-4 2-3 3-1 12-1 7-3 7-4 7-5 7-2 7-1 2-1 2-2 18-2 18-1 22-2 22-6 22-4 22-1 22-3 22-5 22-7 22-1a 1 2 3 4 8-1a 0 50m
Fig. 3.3 Uitgeest-Gr.D. Part of the excavated terrain, with the wells dating from the Roman period to the late Middle Ages.
Legend:
1 = Wells dating from the Roman period
71
Scale 1:250
74 FEATURE VESSEL NR: NR: 1 20-06 3 20-08; 31-01; 31-18 8 20-07 11 31-13; 31-15 14 31-14 16 19-20; 19-21; 19-25 16?/16.1 19-26; 19-13 17 19-02 19 19-14 19.2 35-01; 35-14 19.3 19-15 20 19-17 21 35-19; 35-16 22 35-17; 35-18; 35-24; 35-28; 35-39; 23 35-20; 35-21; 35-27; 35-31; 35-36; 35-38 27 35-03; 35-04; 35-05; 35-06; 35-07; 35-11 27.1 35-33 28 35-22; 35-23; 35-25; 35-29; 35-30; 30 35-40 32 19-06; 19-07 33 19-10 33.1 35-37 34 19-22; 19-23; 19-29 36 19-24; 19-27; 19-28 41 14-08 42 14-11 43 14-09 44 14-12 45 14-05; 14-07 46 14-06; 14-13 47 14-15 48 20-02 49 14-10 53 14-14 54 20-10 55 35-15 56 33-05 57 21-01 58 33-07; 33-09 60 31-02; 31-03 61 31-12 63 19-19 64 20-01 65 20-12 FEATURE VESSEL NR: NR:
THE CREEK FILL
50 13-01 20-03 20-04 20-05 34-01 34-02 34-03 34-04 34-05 34-07 34-10 34-12 51 16-01 28-01 32-01 33-01 33-02 33-03 33-04 33-08 52 30-02 UNKNOWN 8-06 75-01 75-02 75-03 Fig. 3.5: Legend 3
78
Fig. 3.7a Uitgeest-Gr.D. The horizontal stratigraphy of the fill in the remaining creek, north of the settlement area. Legend
1 Layer 1: oldest stratum in final Dunkirk-I creekbed, humic layer with some clay
2 Layer 2: clayey humic layer 3 Layer 3: sandy layer seperating layers 1 and 2 from the later fills 4 and 5
4 Layer 4: humic clay
5 Layer 5: sandy deposit of the (early) Medieval creek. 6 Layer 2 + 3 (no distinction
pos-sible)
7 Layer 3 + 4 (no distinction pos-sible)
8 Layer 4 + 5 (final stratum in creekbed, dated to the end of the Roman or the beginning of the Medieval period)
9 Layer 9: Dunkirk I deposits 10 Post-Roman deposits
79
higher amount of partial or nearly complete pots recovered from trenches 19 and 35 by changed excavation techniques also proves the point. When a concentration of sherds was seen in a feature, these were collected as one find-number, and, if necessary, the feature was then excavated to a lower level than the standard 10 cm. By digging this way consider-ably more reconstructable vessels were recovered than on the remainder of the site. Altogether, this seems to confirm the impression of the author, that many more partial and complete vessels have been present in other parts of the site, but were broken up and scattered by the standard excavation procedures. It must be concluded that the sample composition is far from ideal for the specific purpose of this study. The small num-ber of complete vessels, the uneven distribution over the occupation area, the poor control over stratigraphy and the specific context of the vessels within these features all hin-der control over the composition of the sample. The control over the temporal aspects of the potters technology as well as possible changes in the composition of the household inventory is poor. This study being, however, to a large extent an exploration of the possibilities to examine pottery production and use in a prehistoric context, the sample is adequate for this purpose. The pottery is in general in a good condition; surfaces are well preserved and contain many indications for the surface finishing techniques, the firing methods as well as the subsequent use of the vessel.
3.4 Schagen-Muggenburg 1
3.4.1 SITE DESCRIPTION
The site of Schagen-Muggenburg-1, referred to as ‘Schagen-M1’ below, was excavated during the summers of 1983 and 1984 by staff and students of the A.E. van Giffen Institute for Pre-and Protohistory, Amsterdam, under supervision of drs. L.L. Therkorn (1984, 1985). The settlement, located on peat, dates to around 300 AD. Most of the settlement area was excavated (fig. 3.8, 3.9). The features consist of the remains of many hearths and pits, some ditches and a small dwelling. Also present were the remains of human burials and of human cremation remains, covered by turned-over vessels. All habitation features belong to one period of occupation, estimated to be circa 20 years (Therkorn forth-coming). A second and comparable site in the near vicinity, Schagen-Muggenburg-III, dating from the mid-fourth cen-tury AD, was excavated in part in 1990 (Therkorn 1991). Both sites were analyzed in detail by Therkorn (forthcom-ing) and interpreted as a combination of the usual mixed farming settlement and a focus for seasonal ritual activities. This analysis is the basis for the following site description. Schagen-M1 was located on peat covering a Calais IVB clay or sandy clay deposit. The site was covered by a thin layer of Dunkirk III deposits. The peat layer was at least 70-100 cm thick at the time of the habitation, but only a very thin layer
of humic clay remained at the time of excavation. Due to the oxidation, features dug into the peat cover had disappeared and only those dug into the clay sub-soil survived. Fortu-nately most of the features on or in the peat could be recon-structed by their remaining contents such as pottery and bones; these materials had remained in situ (fig. 3.9). The hearths, for example, had subsided intact with the decaying peat and so had the contents of the shallower pits.
The site consists of two main areas, the northern and south-ern area, with a central area in between (fig. 3.8). The fea-tures in the northern area consist of clusters of pits and open air hearths. The pits differ in size and shape but most are rectangular or elongated. A ditch was still present (feature 107), running along the West-side of this area. The four burials and the pit with the cremation remains were located North-North-West of these clusters. Seven complete or nearly complete vessels, placed upside down and very closely together covered the cremation remains (fig. 3.9). It is not clear how many persons are represented in these remains. As graves and grave-sites are still very rare in the later Roman period in the western Netherlands, their pres-ence indicates the special importance of the site. In the central area a few pits were found.
The southern part of the excavated area consisted of a house site. The small dwelling had a clay flooring and contained a virtually intact hearth (feature 194; fig. 3.9) as well as two ritual pits (features 212 and 352). Pit 212 was constructed before both the clay floor and the hearth were renewed. South of the dwelling, a part of a house-ditch (feature 240; fig. 3.9) was marked by a fill of mostly pottery sherds; an interruption in the finds defines the entrance to the house, which was marked by three complete ox mandibles, a block of pink granite and two ash dumps. Around the house, especially to the East, again a large number of rectangular pits was present. It is quite likely that pits dug into the clay subsoil were used to extract the clay itself. Samples of this clay were used in the experiments for fabric composition and properties (chapters 4 and 6). The layout of the pits in all three areas is remarkable. Those in the northern area are in line with the northern wall of the dwelling and this line is also continued eastward by a series of pits. The pits occur in clusters of three and often show specific orientations within these clusters. All of the pits were involved in ritual activities with depositions of artefacts. The spatial arrangement plays an important role in the inter-pretation. At the end of the habitation, a low mound of soil mixed with large quantities of sherds covered both areas4. 3.4.2 STRUCTURE AND CONTENT OF SETTLEMENT
FEATURES
Hearths
covered by a layer of clay (fig. 3.9). This is the usual con-struction in the indigenous sites of the western Netherlands (Therkorn 1987). Several activities were carried out in and around the open-air hearths, as suggested by the associated finds, most importantly metalworking of bronze, iron and lead. Other activities were bone- and woodworking and very likely also the firing of pottery. It is assumed in this study that pottery was fired in open fires, not in kilns (see chapter 2), despite the fact that the remains of a possible oven wall were found in a trial trench excavated by local amateurs (Diederik 1981). These remains consist of very thick clay coils from a large circular construction, with a diameter of circa 100 cm. The wall was supported by twined rope within or at the outside of the coil. Similar lumps of clay, with remnants of rope impressions, were found among the pottery in the cremation pit; these probably are the remains of an oven wall as well.
A recent study shows that similar remains were found in several sites in the western and northern Netherlands (Boulonois 1997). They are definitely to be interpreted as parts of ovens although their specific function is less easy to determine. The use as pottery kilns seems supported by the fact that burnt pottery and pottery slag is often found in association with these remains. There are several arguments against such an interpretation, however. It is equally possi-ble that the pottery sherds were used in the ovens, for example as heat shields. Moreover, the number of sites with oven remains and the reconstructed number of ovens is so low, that their use as kilns seems very unlikely. Also, the data on firing methods of the pottery from this period from all known sites point against the use of kilns. Alto-gether, the remains are too scant to allow a decisive recon-struction of the complete form and the function of these ovens.
Pits
The pits were dug to different depths, some are shallow, and some are deep shafts reaching well into the clay subsoil. All of these pits contained depositions of artefacts and they were always backfilled immediately after the deposition of these goods; sometimes different layers of fill were distinguished associated with different depositions. These depositions and their associations are the basis for the interpretation of the site by Therkorn, see below. The following major find cate-gories were present alone or in combination (table 3.4): Selected bones or skeletons from animals are the most impor-tant category of deposited material. Three pits contained complete animal burials, of a horse and two dogs (features 62, 238, 250, fig. 3.8)5. Many others contained a specific selection of bones from these animals as well as from sheep and cattle, notably elements of the skull such as jawbones and teeth, and of the lower limbs such as phalanxes and
astragali. Thus all livestock animals are represented in the depositions. Already mentioned are the exceptional burials and cremation remains of humans at the site. Another impor-tant category of finds in pits is wood, both worked and unworked. Wood and bones were regularly found together in a pit. Other associated finds are querns, spindle whorls, dice, metal objects, etc.
Of special importance for this study are the pits in which pottery was deposited, often as nearly complete vessels. These features will be discussed in more detail below and in chapter 8.
Other features
The ditch along the western side of the northern area, feature 107, contained several categories of material, including many animal bones, especially butchering remains, and some pottery, mostly sherds. As mentioned, indigenous vessels were used to cover cremation remains in the northern area. The remains of a habitation layer, feature 30, associated with the open air hearths and pits in the north included many pottery sherds, some of which fitted with sherds from fea-tures 143 and 148. The house-ditch South of the dwelling also contained large quantities of sherds. Feature 258 in the southern area is a large elongated and irregularly shaped feature, which in the subsequent analyses is interpreted as a pit. Feature 259 is a habitation layer, which has subsided into the top of pit 258.
3.5 Site interpretation
The bones or skeletons of livestock animals are the most important category in this ordering (table 3.4, based on Therkorn). These data were combined with the age of the animals where available. In the ditch (feature 107) mainly butchering remains from cattle were found; these animals were slaughtered at another age than those found in the ritual pits; the latter contained bones of animals from vary-ing age categories, but especially from—very—young ani-mals. There is also a specific selection of the type of bones in these deposits, namely from the skull or lower limbs. These bones show a specific distribution, both horizontal and vertical, over the features. On the basis of historical sources from the Roman and Medieval period, the bones
were interpreted as ‘head’ and foot’ symbols, the latter being associated with divination practices. By combining informa-tion from several historical sources, three major seasons and associated ritual ‘dates’ were established and in turn were combined with the archaeological data from the site(s) (Therkorn 1991/2, forthcoming). These rituals were possibly based on the cosmological notions of the inhabitants in which the seasons were connected with the most important ‘economic’ cycles concerning livestock and agriculture. Three seasons were distinguished by the West-Frisians: spring, summer/fall and winter. The ritual dates connected with these seasons were established by Therkorn as: late spring (May-June), late summer or fall (October/November) 82 352 0.45 - 2.0 76 - + + + - - -1 212 0.63 1 2.8 136 1 + + - - - -118 0.38 - 2.5 63 - + + + + - + 120 0.62 2 5.4 127 - + + - - - + 2 115 0.32 3 1.2 28 1 + + - - - -121 0.47 - - 3 - - - -3 42 0.22 - 0.9 25 - + + - - - -21 0.15 2 3.2 99 - + - + - - -223 0.27 10 0.9 44 10 + + + + - -4 221 0.46 - 0.8 40 - + + + - - -222 0.16 2 3.2 99 - + + + - + -143 0.78 8 11.5 226 1 + + - - - -5 141 1.01 - 2.6 89 - + + - + + -142 0.28 - 2.2 193 - + - - + - -156 0.89 - 4.4 172 - + + - - - -6 154 1.00 3 4.6 135 3 + + - - + + 155 0.12 6 2.0 74 1 + - - - + -148 1.01 1 7.0 317 - + + + + + -7 79 0.53 9 13.1 175 2 + + + - + + 78 0.26 2 3.2 77 - - - + 22 0.26 1 1.8 46 - + + - - + -8 23 1.63 - 2.5 74 - + + - + + -24 0.56 - 1.0 61 - + + - - - -31 0.56 4 3.4 52 - + + - + - -11 187 0.79 - 0.3 12 - + + - - + -18 0.99 - 0.1 9 - + - - - - -linear 147 0.43 3 4.3 79 1 + + - - + -features 27 4.50 (3) 24.6 735 3 + + + + + -185 3.95 1 5.5 195 1 + + + - + +
Table 3.3 Schagen-M1. The context of specific categories of finds, associated with ritual deposition.
Cluster
of Pits Feature Size m
3 N vessels in sample kg N sherds Pottery compl. vessel Bones of domestic animals
Wood Stone Metal Roman import
and midwinter (December/January)7. In the ritual offerings specific goods were used in each of these seasons and deposited in specific pits. Through these deposits the north-ern, central and southern area were interconnected. The alignment of pits in the northern and southern area is part of this connecting and structuring of the activities. The site thus expresses and binds the cosmological ideas and practices of its inhabitants, in several ways: a. by a specific spatial order-ing of the features, b. by a specific selection of material categories and their placement in the features and c. by the time/timing of the ritual depositions as seasonal and cyclical. In chapter 8, the pottery groups resulting from the indepen-dent analyses of morphological variables are analyzed as to these positions within the ritual practices.
3.6 Selection of pottery
The selection of pottery for this study was guided by the preference for complete or nearly complete profiles and the site interpretation presented above. Schagen-M1 is in many respects an ideal site for the purpose of this study because the number of complete profiles, including complete vessels, is quite large, whereas the excavation methods provide detailed context information and ensured the possibility of reconstructing a large number of broken vessels8. The pot-tery from all features was sorted, described and where possible restored by ms. T. Spruyt, who also carried out most of the analyses of construction techniques for the sample. For many pits, few sherds remained after restora-tion; where this was not the case, for instance in feature 27, the sherds were found mainly in the top fill and could thus be part of the subsided covering layer. The analysis of the fabrics and surface treatment was carried out by the author. A total of 108 partial or complete profiles were selected for analysis (table 3.3 as legend to fig. 3.8), composed of pottery from the following categories of features. The number of vessels from pits in the sample is 67, of which 35 are com-plete profiles and seven are bases. Seven profiles, including one complete profile and two bases were recovered from ditches. From the hearths 26 vessels, of which seven com-plete profiles and three bases, were used in the sample. Of the seven vessels covering the cremation remains, four were complete profiles. Mixed with these pots were the remains of another vessel. This sample was used for the analysis of form and function and with some changes also for the technologi-cal analysis. Most of the pottery was in rather bad condition, probably due to the acidity of the soil.
3.6.1 POTTERY FROM FEATURES IN THE NORTHERN AND CENTRAL AREA
Hearths and occupation layer
The restoration work by Spruyt on the pottery from the hearths made clear that most of these features incorporated
at least one more or less complete but broken vessel. Large fragments and some complete profiles could also be recon-structed from the double hearth 157/177, including three complete profiles, as well as from hearths 35, 76, 78, 127 and 159 in the northern and 194 in the southern area (appendix 1c and table 4). Two hearths (features 135 and 157/177) even contained unbroken vessels, very small cup-like pots.
Pits and ditches
Many of the pits in the northern area contained pottery, including at least 11 complete vessels. Pottery was selected from the following features: 22, 27, 79, 143, 143, 154-155, and all vessels from the cremation pit (feature 345). All of these pits contained at least one complete profile or com-plete vessel.
One of the most important features in Therkorn's analysis is pit 27: a rather shallow and irregular but very large pit with large amounts of un-worked wood; this pit also contained many sherds, but relatively few complete vessel profiles. Unfortunately these few complete profiles were lost during transfer, before they were measured or drawn, but they are still included in the sample.
Feature 79, a deep round pit with a stratified fill, contained several complete indigenous vessels, as well as sherds. Feature 148 is another deep feature with many layers of fill, mainly consisting of organic material like twigs, charcoal, and bones, containing in the uppermost fill a nearly com-plete Roman import vessel, of which the rim is missing. No complete indigenous pots were present, but a few sherds fitted with those in the habitation layer, feature 30, and pit 143; one of these vessels is included with the pottery from 143, another is ascribed to the occupation layer, based on the context of the majority of the sherds.
Feature 142, 143 and 147 are less deep; pit 143 also had different layers of fill, containing many sherds, especially in the upper fills, from which 5 complete profiles were restored (fig. 3.9). No complete profiles could be restored from fea-ture 142 and 147, but the latter contained one complete roman vessel. This jar is dated to the third century AD. The features 154 and 156 are rather deep pits with many finds in the layered fills. Feature 154 (Fig. 3.9) contained much organic material—worked wood, dung, ash, etc.—as well as large amounts of pottery, including three complete vessels and the base of a Roman import vessel. Feature 155 is a very small and shallow pit, containing one complete profiles and one complete pot, as well as many sherds. Only one complete profile could be restored from the sherds in ditch 107. Feature 22 in cluster 8 also contained one complete profile.
The pottery from the pit containing the cremation remains could be restored to four vessels with a complete profile or
86
Fig. 3.9 Schagen-M1. Photographs of features 79, 143, 148, 155, 194, 223.
3.9A+B Feature 223. A: view to Northwest
B: view to Southeast, at lower level 3.9C Feature 194
3.9D Feature 155 3.9E Feature 240 3.9F Feature 148 3.9G Feature 143
3.9H Cremation pit, with pottery F 314, 325, 340-2
3.9I Detail of cremation pit and pottery
3.9A
3.9B
87
3.9D
88
3.9H 3.9I
even nearly complete; of three others the bases were miss-ing. As it is unlikely that this happened during excavation, the base may have been removed on purpose before deposi-tion, althought it is possible that bad conservation prevented the complete restoration. The eighth vessel is an incomplete profile restored from sherds in the same pit.
In the central area, between the concentration of pits and hearths in the northern part and the dwelling in the southern part, two clusters of pits (10, 11) were found in the vicinity of the horse burial (fig. 3.8). In cluster 10 no pottery was found. In cluster 11, feature 31 was filled with bones and wood as well as three complete vessels. In feature 185, a very large pit, a tiny vessel containing seeds of chickweed and orach was found.
3.6.2 POTTERY FROM FEATURES IN THE SOUTHERN AREA
Hearths and layers
One complete profile could be reconstructed from the hearth, feature 194, in the dwelling, together with large fragments of three other vessels. Incorporated in the hearth was a part of a ‘pedestal’ bowl, a specific type of indigenous vessel. Two incomplete vessels were restored from the sherds in layer 259, which formed the uppermost fill of feature 258. Pits and ditches
Pits 115, 120, 212, 222, 223 and 258 all contained at least one complete profile, mostly a complete vessel and often large fragments of other vessels. Feature 223 is of special importance. This rather small pit was completely packed with sherds from ten vessels and these were obviously deposited as whole, unbroken vessels. The restoration to complete profiles failed, however, in three cases because of the bad preservation of the sherds, which literally were falling apart when handled. According to excavation details and the postdepositional changes visible in the pottery, some of these vessels had been placed on their sides.
Most of the other pits contained much less pottery, sometimes only one vessel, like feature 212, the pit within the dwelling and feature 258, with a small vessel and a sherd used as a lid. In the house ditch 240 some large fragments but no complete pots were present within the large amount of sherds.
notes
1 The author has catalogued virtually all pottery; the number of sherds recovered exceeds 65,000, not including the smaller finds from the layers of the creek fill in trench 16. Four of these layers were entirely sieved on a 4 and 9 mm sieve, resulting in ca 388 kg of pottery.
2 Rijksgeologische Dienst.
3 Based on pottery from this layer, mostly Carolingian, recently confirmed by de Koning (1998)
4 Such mounds were also found in some of the levee sites in the Assendelver Polders, see Therkorn & Abbink (1987).
5 Animal burials are also found in many other sites in the Nether-lands from the Roman and Early Medieval period (Prummel 1993), including Uitgeest, where several horse burials have been exca-vated, but these have yet to be dated.
6 See also chapter 8.11 and 8.14; for more details the reader is referred to the original text by Therkorn, 1991 and in prep).
7 According to Therkorn (pers. comm.) there is some debate about whether or not the Germans recognized 'fall' as a separate season. In this study the Oct/Nov. rituals will be labeled as 'fall' in the modern sense, i.e. according to our own seasonal cate-gories.
8 All pits were dug stratigraphically, the sherds were collected separately from each layer and those obviously being from one pot were also collected separately. For each pit, the depth at which vessels were placed, and their position, that is, whether placed upright or on their sides etc. was also recorded.
