Holocene clays from deposits in different locations in North-Holland were used to study the chemical composition, the firing ‘behaviour’ and the resulting fabric properties of such clays under controlled conditions1. The test clays should
provide an independant reference set for the fabric analysis of the pottery from Uitgeest and Schagen, specifically for identification of the types of clay used for the pottery and for the firing methods. The test set was also used for chemi-cal analysis by Röntgen-fluorescence. The choice of the test clays was based on location, that is, their availability to prehistoric and Roman Iron Age potters in North-Holland, and on their quality as a potting clay. The latter was assessed by Ms Tineke Spruyt, a professional potter, and by the author. The hypothesis is that the potters chose clays available in the near vicinity of the sites and selected spe-cific layers within the deposits.
4.1 Clay deposits and samples
The western regions of the Netherlands partly consist of Holocene marine and fluviatile clay deposits, mainly illites and montmorillonites. In the Later Iron Age and Roman period clay deposits from several ‘transgression phases’2
were ubiquitous throughout most of North-Holland either on or near the surface. For the sites of Uitgeest-Groot Dor-regeest and the Assendelver Polders, Dunkirk-I clay deposits were lying directly behind the sandbars and levees on which settlement took place. In Schagen-Muggenburg, Calais-IV deposits formed the substratum below the occupied peat-layer.
Because of the ubiqitous presence it was argued, that all indigeneous pottery recovered from pre-and protohistoric sites in the Western Netherlands will have been made of locally available clays. There is thus no need for an exten-sive program on provenance studies for most pottery com-plexes in North-Holland, at least not in the usual sense, to determine specific clay sources for pottery which can not have been made of locally available clays, or clay sources which were used for pottery production on a regional or larger scale, like the Medieval pottery production at Schin-veld or Mayen. Needless to say that one should have an open eye for exceptions to this rule. The ubiquity of avail-able clay deposits in the Western Netherlands does not
mean, however, that the clays used by the potter can be easily traced. The composition of clays can and indeed does vary considerably from place to place even at a local level, as they are made up of materials from different sources and deposited in specific patterns of sedimentation. In Schagen-Muggenburg the clay compositions differed notably within 40 m in the excavated terrain. Moreover, not all clays will have been considered suitable by the prehistoric potters. As argued in chapter 2.4, variations in composition can affect the specific properties of the pottery fabrics. Characterisation studies of raw materials and fabrics can therefore be of great help in establishing the ways in which potters dealt with these local variations. The main types of variation to be expected in the composition of the Holocene marine clay deposits are, briefly, as follows.
Most soils contain iron in varying degrees and forms; both iron bound to the clay matrix and iron oxides in the form of concretions are quite common. The amount and size of sand particles in the clay deposits can vary considerably, depend-ing on the velocity of the water and the kind of deposits being eroded. The presence of calcium, as Ca-compounds within the clay-matrix, calcitic nodules, or shell fragments can be expected in marine deposits such as we are dealing with. Clay deposits will also vary in the content of organic material and humus and in the amount of natural salts and diatoms (Van der Leeuw et al. 1987). Diatom-analysis has proven valuable in differentiating between marine and fluvi-atile clays (Van den Broeke 1986). To look for the exact locations of clays procured by prehistoric potters within the near vicinity of the sites is, because of all possible varia-tions, looking for the famous needle in a haystack, even when the available techniques would be precise enough to detect these variations, which they are not. A more general problem is that the composition nowadays can be substan-tially different from that in prehistory through soil formation processes that took place since the Roman period3. Such
changes are relatively unimportant for the purpose of the experiments, to test and control the influence of nonplas-tics—quartz, iron, calcium and organic matter—on the resulting properties of fabrics after firing. Whether these nonplastics are of a primary or secondary origin is not rele-vant for these properties. However, as a reference set for
4
Chemical composition and fabric properties of Holocene clays
identification of the clays used in the pottery there should of course be a good match between the composition of both. Test clays were collected from the following sites, all in the province of North-Holland:
Site Deposit Sample nr.
Schagen-M1 Calais IV-B clay samples nrs. 62, 63, 64 Uitgeest-G.D. Dunkirk I clay samples nrs. 81, 82 Assendelft-F Dunkirk I clay sample nr. 65 Opperdoes Dunkirk 0/I clay sample nr. 60 Bovenkarspel Calais IV clay sample nr. 61
The choice of the clays was aided by archaeological evidence. In the sites Schagen-Muggenburg, Assendelft site F, and the Iron Age site of Opperdoes there were strong indications that certain pits were dug specifically for the extraction of clay (Therkorn & Abbink 1987, Therkorn 1984; Woltering, per-sonal comm.). The clay from Bovenkarspel was collected from within a Bronze Age occupation area (pers. comm. G. IJzereef). The three clays from Schagen-Muggenburg (nrs. 62-64) were all collected from the Calais IV B deposits within the excavated area (North of cluster 11, see fig. 3.8). Clay nr 64 was extracted adjacent to a series of settlement pits which were cut into the Calais IV-B layer and this clay was expected to be most similar with those of the pottery fabrics. Clay nrs 62 and 63 were extracted from a nearby trench without archaeological evidence for clay winning, but they were potentially good potting clays. Clay nr 65 from site F (Assendelver Polders) was taken from a deposit just along the outer limits of the levee system, where the sandy loam of the levees changed into clay deposits (Abbink in Therkorn & Abbink 1987). Here also numerous pits were dug and always into the same layer of grey clay. Wether this clay was used for pottery production and/or for other purposes, such as walls of buildings, is of course uncertain, but the evidence did provide a good starting point for the comparison of these clays with the fabric of the pottery. No suitable potting clay was present within the excavated boundaries of Uitgeest-Groot Dorregeest as the site is located on Dunkirk-I sandy deposits. Therefore clays were obtained from the nearest deposits, some 500 m West of the site4.
4.2 Methods
Of each of the clays test tablets were made with a thickness of 3, 6 and 12 mm and a length of circa 100 mm. The clay was thoroughly kneaded, but otherwise left untreated. No temper was added. The briquettes from clay nrs 61-65 were fired at the temperatures of 700, 800, 850, 900, 950, 1000, 1050, 1100 and 1150 °C. Those of clay nr 81 and 82 were fired at the temperatures 800, 850, 950, 1050, 1100, 1150, and of clay nr 60 at 700, 800, 950, 1050, 1100, 1150 °C (fig. 4.1a,b). All tablets were fired in an oxidizing atmosphere
in an electric oven with a temperature trajectory control. The maximum temperatures were reached after a period of 12 hours, maintained for 60 minutes, after which the oven was allowed to cool off. The total firing process took 24 hours. The production of the test tablets and all the tests were car-ried out by mr. E. Bolte, ROB. The tablets, mainly those with a thickness of 12 mm, were used for the following observa-tions:
– the size and quantity of sand;
– the percentage apparent porosity (%AP);
– colour and colour changes as an indication for iron and calcium-content of the clays; including colour identifica-tion by Munsell soil chart;
– the size and quantity of iron concretions and their changes; – for most temperatures and clays, one tablet was subjected
to a dilatometer, to check the known firing temperature with its dilatometric values5;
– thirty grams of dried pulverized clay from samples 62-65 was used for wetsieving on soilsieves with the standard meshes of 210, 150, 105, 75 and 52m. The residues on each sieve were compared with the data on the fired tablets.
For every clay sample, a piece of 3 ≈ 3 cm of the tablets was sawn lengthwise to create a core surface. This surface was used to count the inclusions and quartz-particles. Colours of the fired tablets were defined with the Munsell soil-chart, but were also arranged by the finer colour nuances together with textural information which could be seen within the rather coarse definitions of the Munsell colour-charts.
4.2.1 PERCENTAGE OF APPARENT POROSITY
The standard technique of the ROB laboratory was used for the measurement of the %AP. The samples are boiled in destilled water for exactly two hours, then allowed to cool to room-temperature in about 30 minutes. They are left in water while the sherds are being weighed in an immersed state: this is the volume weightV. The sherds are then left to dry for 15 minutes and weighed again, giving the saturated weightA. Subsequently, sherds are dried for circa 18 hours at a temper-ature of 130 °C, left to cool for one hour inside and one hour outside the oven to room temperature, after which the dry weightB is measured6. The %AP is calculated as the
differ-ence between saturated and dry weight according to a for-mula, developed by Brongers (Brongers 1983):
(saturated weight - dry weight) (A-B)
——— = %AP (saturated weight - volume weight) (A-V)
The procedure was followed for all %AP measurements carried out in this study7
4.2.2 CHEMICAL ANALYSIS OF SIX CLAYS
The chemical composition of sample nrs 62-65, 81 and 82 was analyzed by means of XRF (X-ray fluorescence) and ICP (Induced Coupled Plasma/Atomic Emission Spectrome-try) (van Haaren 1990/1, internal report)8. Both methods
measure the kind of elements in the sample and their quanti-ties in parts per million. XRF measures the main elements and some of the trace elements whereas ICP measures both. In the XRF results the amount of CaO is expressed as a percentage. Both the unfired clays and parts of the test tablets fired at 850 °C were tested. This temperature was choosen as the most likely maximum firing temperature of the pottery, based on dilatometric readings for some sherds from Uitgeest and on the %AP of the clays at this tempera-ture, in comparison with the pottery.
4.3 Composition of the test clays 4.3.1 NONPLASTICS:QUARTZ AND INCLUSIONS
The data on quartz particles, iron-rich concretions (hence-forth referred to as Fe-inclusions) and other components for most of the clays can be found in table 4.1. The tablets of the three clays from Schagen-Muggenburg (nrs. 62-64) share a characteristic fine and homogenous structure of the matrix and are all lacking in quartz particles coarser than 75-105m. In the sieve residues of the fractions 105 m and lower a very small amount of quartz was recovered, but most of the particles were smaller than 52m. The amount of quartz was very low as well, at most 1 teaspoon out of 30 grams of clay. These samples also contained rather a lot of colourless mica, mainly between 75-150m. In all sieve residues fine organic material was present. In clay nr. 62 and 63 some very small Fe-rich nodules were present both in the tablets and the sieve sample. Sample nr. 62 also contained hard calcitic nodules, probably very tiny shell fragments, which occurred in every residue and were more abundant than the iron concretions. These nodules were visible at every temperature upto 11000C, but in
diminish-ing quantities. The sieved residues of clay nr. 64 consisted mainly of very fine organic material with some iron and calcium nodules. The latter did not show up at all in any of the fired tablets. The tablets of nr. 63 also contained an occasional calcitic inclusion, probably shell, at lower firing temperatures, while at higher temperatures yellow streaks appeared within an overall pale red matrix. This change will be discussed below. Clay nr. 65, 81 and 82 (Uitgeest & Assendelver Polders) are rather different and contain vari-able amounts of quartz. In nr. 65, quartz particles to 150-210m (circa 10 particles per cm2) occurred both in the
tablets and sieve residues, as well as Fe-concretions and muscovite fragments. Clay nr. 81 and 82 contained quartz up to 210-300m in rather large quantities of 60 and 25 particles per cm2
. The Fe-concretions in nr. 82 occur
frequently (12 per cm2). These three clays do not contain Ca-concretions. Clay nr. 60 and 61 (Opperdoes and
Bovenkarspel) were not available for sieving, but the tablets showed only finer quartz particles up to 75m. Clay nr. 61 has a large number of Fe-concretions.
The test clays can be divided into two main groups on the basis of the nonplastic inclusions. The three clays from Schagen are rather similar, they all consist of a very fine, homogeneous clay with muscovite, organic particles, and with only a slight amount of the finest quartz fractions. Clay nr. 60 is very similar to them. They differ only slightly in the amount of iron and calcium inclusions. The clays nr. 65, 81 and 82 contain more and coarser quartz particles, as well as Fe-concretions, but lack the Ca-rich inclusions. Clay nr. 61 is similar to this group in texture, even though no coarse sand particles are present. In general, the texture of these clays is less homogeneous and coarser than of those from Schagen.
4.3.2 COLOUR
The Munsell colours were determined for all clays at all temperatures. In fig. 4.1, the tablets are arranged according to their Munsell code, but also by eye on minor colour variations which cannot be ‘measured’ by Munsell colours. In table 4.2, the colour codes for three temperatures, 850, 950 and 1100 °C are presented. As was expected, the colours of the tablets all look alike at lower temperatures. At 850 °C the chroma and HUE-values vary from 7.5 YR to 2.5 YR. Between 950 and 1050 °C, however, differences become apparent; some of the clays show in increase in the Red, others in the Yellow codes and at 1100 °C a clear distinction can be made between, on the one hand the prob-ably Ca-rich clays (nr. 82 and 60) with a HUE of 10YR and, on the other hand the probably Fe-rich clays (nr. 65, 61, 64, 62) with a HUE of 2.5 YR up to 10 R. The third group (nr. 81, 63) shows a remarkable mixture of both Yellow and Red chromas. Within the matrix of the tablets a lamination can be seen caused by trails with a light yellow colour. This separation of colours and chemical compounds is visible even at low firing temperatures in clay nr. 63 and change into ‘trails’ of alternately reddish-brown and yellow-ish-green at higher temperatures (see fig. 4.2 and discussion below).
haematite turns to dark red and purple and changes slowly into magnetite. Magnetite is a larger-sized crystal of Fe2O3
and is black (Keramiek 1973). The Fe-concretions follow the same pattern, but their colour is usually brighter, that is, their chroma-value is higher, than that of the clay matrix, due to the higher Fe-content. The change of haematite to magnetite seems to vary between clays, but clearly takes place from 1050 °C onwards.
The presence of calcium in clays in general makes the colours after firing lighter and weaker. If no Fe is present, the clays will fire white. Ca-compounds also change in composition and crystal forms. With increasing firing tem-peratures dissociation into the unstable CaO takes place between circa 750 and 850 °C, while with higher tempera-tures ghelenite, a stable compound, is formed which changes again into anorthite above circa 1100 °C. (Maggetti 1980, Heimann 1989). However, many clays and certainly the marine deposits in the Western Netherlands, contain both Ca and Fe. The colours of such mixed clays can alter drasti-cally with increasing temperature, for example clay nr. 63 (fig. 4.1b), but “the reason for these changes are still poorly understood” (Heimann 1989, 137).
4.3.3 CHEMICAL COMPOSITION
Samples of the unfired clays and tablets fired at 850 °C were examined. The results for a selection of the elements can be found in table 4.3. The samples have been ranked according to their amount of Ca(O), because of the impact of the Ca:Fe ratio on colour, as discussed above.
Ca and Fe content and ratios
Nrs. 81, 82 and 63 have a relatively high Ca-content both in unfired and fired form, whereas nrs. 61, 62, 64 show low percentages. As expected the Ca-content is largely responsi-ble for the colours of the taresponsi-blets for each clay: compare table 4.2-3 and fig. 4.1. The amount of Fe in these clays is more or less the same, although the Ca-rich clay nrs. 81 and 82 also have slightly higher amounts of Fe. Sample nr. 63 shows an interesting difference between its Ca-content in unfired and fired state, it being much higher in the first. This is in contrast with to all other clays in which the Ca-content increased slightly when fired at 850 °C. The increase is due to the dissociation of CaCO3into CaO which is taking place between ca 750 and 850 °C (Maggetti 1980). It is not clear why clay nr. 63 is an exception9
. Rather a surprise is the high content of Ca in clay nr. 65. The red colour must be due to the high amount of Fe in this clay, by far the highest of all. Thus, not only the absolute amounts of, but also the ratio of Ca : Fe are important determinants for colour (table 4.3). The change from red- to ‘white’ (= yellowish/green) firing seems to take place when the ratio is higher than 1.0-1.5 within this sample of Holocene clays.
Other elements
The element Sr (strontium) is often an indication for the presence of Ca as well; this is borne out by the present data. Van Haaren (1990/1) suggested that the iron occurred mainly in compounds with phosphor. Again this is supported by their relative amounts in the analyzed clays (table 4.3). The amount of manganese and sulphur in the clays are noted mainly for comparison with the composition of the sherds. The sulphur content of the sherds may have been affected by fertilizer.
4.4 Porosity and melting points
The %AP is rather high for all of the clays involved, cer-tainly when compared to tertiary clays (see Brongers 1983). There are small but consistent differences between the clays. The %AP of clay samples nr. 63 and 82 is relatively high, that of sample nr. 61 relatively low. The %AP of the clay nrs. 62, 64, 65, and 81 are rather similar (fig. 4.2-3). There is a clear increase in %AP at 800 °C and a slight decrease at 850 °C for all clays except clay nr. 81. The increase at 800 °C is caused by the final dehydration and the desintegration of the clay minerals. The drop in %AP at 850 °C is to be associated with the beginning vitrification. At higher temper-atures the %AP of the clays stay more or less the same or increases slightly again before the drastic drop which is associated with complete vitrification and melting. It is not clear why the %AP is increasing mainly in the Ca-rich clays above 850 °C. These data do not correspond with the general suggestions in the literature about increasing vitrification especially in Ca-rich clays (a.o. Maniatis & Tite 1981) nor by the data on tertiary clays tested by Brongers (1983). It is suggested here that the increase must be due to the higher amount of Ca together with the recristallisation of CaO. Moreover, the melting temperature of a clay co-varies with its Ca-content as well: clay nrs. 60 and 82 were completely, and clay nr. 63 was partially vitrified at 1150 °C, but the remainder did not show any sign of melting. The lower melting temperatures must be due to the Ca acting as a flux. Clay nr. 61 is an obvious exception, perhaps due to the very high number of Fe-inclusions.
why the %AP of clay nr. 63 is higher than that of clay nr. 81, notwithstanding the lower amount of Ca. Vice versa, the high %AP of clay nr. 65 is due to the high Ca content which partly negates the effect of the high Fe-content of this clay. There is no clear correlation between the amount and/or size of quartz particles in the clay and the %AP. The sandy clay nr. 82 with particles up to 400m, has the highest %AP of all clays, but also the highest Ca-content. Clay nr. 65 is moder-ately sandy but the %AP is similar to clay nrs. 62 and 64.
4.5 Discussion and summary
On the basis of colour, chemical composition and firing properties, the test clays can be divided into three main groups:
group 1: clay nrs. 61, 65 group 2: clay nrs. 62, 64, 81 group 3: clay nrs. 82, 60, 63
In group 1 and 2 the Fe-content is the main colourant, despite the rather high Ca-content of clay sample 65. Although the
chemical composition of the clays in these groups is quite similar, they have been distinguished on the basis of the non-plastic components and the slight differences in colour. Group 1 is formed by the Fe-rich clays with a low Ca:Fe ratio and a low %AP. These clays also contain considerable amounts of, mainly, Fe-inclusions. Group 2 contains no quartz and few or no inclusions. Clay nrs. 62 and 64 are sligthly less ‘red’, and more ‘yellow’ than nrs. 61 and 65. Group 3 is defined by a relatively high Ca : Fe ratio. Sample nr. 82 of group 3 has a high Ca-content as well as a high Ca : Fe ratio and is charac-teristized by a lack of inclusions. The Ca-content is so high, that it ‘overrules’ the Fe-content as colouring agent. Clay nr. 60 is very similar in colour and in %AP to clay 82.
Clay nr. 81 (group 1) and nr. 63 (group 3) are exceptional because of their specific firing properties and chemical composition. The Ca-contents and the Ca : Fe ratio's have values in between those of groups 2 and 3. The colour of clay nr. 81 is sligthly more red than that of clay nr. 63 and its %AP is lower, due to higher amount of iron.
900 950 1000 1050 1100 1150 800 850 700 5 15 10 20 25 30 35 40 45 0 750 1200 clay 61 clay 62 clay 63 clay 64 clay 65 clay 81 clay 82 % A. P. °C
The firing properties of the clays point to some interesting chemical changes with increasing firing temperatures. Firstly, there is some difference in the reactions of the Fe-and Ca-inclusions. Fe-concretions apparently remain as such and become more pronounced with increasing firing temperatures in clay nrs. 61, 65, 81 and 82. The Ca-inclu-sions in the form of shell fragments also remain present but seem to diminish in size and quantity at higher tempera-tures and to blend in with the matrix. Secondly, the melting temperature of a clay co-varies with its Ca-content as well. The lower melting temperatures must be due to the Ca acting as a flux. In the case of clay nr. 61 the melting is perhaps also due to the very high number of Fe-inclusions. The remainder did not show signs of melting. Thirdly, the Fe and Ca, as chemical elements in the clay matrix, usually determine the overall colour of the fabric as a homogeneous unit. However, in the clay nrs. 63 and 81 and possibly 82 (fig.4.1b) a dissociation seems to take place into seperate compounds in the matrix with increasing firing temperature. The colouring suggests that the Ca-compounds tend to form separate trails when the clay is fired at 950 °C or higher. Although these are less pronounced, trails of Fe can also be seen. This lamination shows up most clearly at 1100 and 1150 °C. In Clay nr. 81 the lamination is less pronounced
and visible only at 1100 and 1150 °C. Clay nr. 82 also shows some vague Ca-trails in the tablets fired at higher temperatures. Although there is no clearcut explanation for this separation of Ca-rich or Fe-rich trails from the matrix, it seems likely that it is somehow connected with the spe-cific amounts and ratios of Fe and Ca, or with the mineral forms. Possibly there is a relation with the differences in the temperatures at which decomposition and recristallisa-tion of Ca- and Fe-minerals take place. Ca, when dissoci-ated, can act as a flux in the vitrification process, although mainly at higher temperatures (Keramiek 1973, 308). Per-haps the fluxing effect results in a incomplete and local vitrification within the sherd, in which the Fe is reduced and at the same time perhaps ‘expelled’ from the vitrifica-tion area, while the Ca is recrystallizing. Iron does not have this fluxing effect, at least not when the clay is fired in an oxidizing atmosphere. Clay nr. 63 indeed still contained some Fe-concretions at 1000 °C, but the Ca-inclusions are practically gone. As even the experts (see Heimann 1988) have no definite answers to the question how Ca and Fe exactly react in all the processes taking place during firing, no further explanation can be offered here for the observed differences in firing behaviour10
. This problem is further discussed in the next chapter, but clearly more research
Ranking By Colour 800 °C ranking by % A.P. 1050 °C ranking by % A.P. 1100 °C ranking by % A.P. 850 °C ranking by % A.P. 950 °C ranking by % A.P. 1 Clay 60 2 Clay 82 3 Clay 63 4 Clay 81 5 Clay 62 6 Clay 64 7 Clay 65 8 Clay 61 60 82 63 65 64 81 62 61 82 82 82 82 63 63 63 81 81 60 60 63 65 81 81 60 64 65 65 62 62 64 62 64 60 62 64 65 61 61 61 61
using the appropriate laboratory techniques is needed to solve it.
Finally, colour proved to be a reasonably good indicator for the Fe- and Ca-content of clays. The absolute amount of Ca is an important determinant for the firing properties of the clays, but the effect is intervened by the amount of Fe. The ratio of Ca : Fe also suggests that at circa 1.0-1.5 there is a change from ‘whitefiring’ to ‘redfiring’ of the clays, while this ratio also determines the %AP11. Especially the latter is an
impor-tant result in relation to the properties of the pottery fabrics.
notes
1 The experiments were part of a more encompassing research program on the ceramological properties of Dutch fluviatile and marine clay deposits, carried out by the State Service for Archaeol-ogy (ROB) under supervision of dr. A.J. Brongers. The program was an extension of previous research on tertiairy clays from Lim-burg and the Eiffel (Germany), which were used in the Medieval pottery industry in both regions (Bardet 1995; Brongers 1986; see also Bruyn 1979).
2 ‘Transgressions’ refer to periods of increased activity of the sea, resulting in renewed erosion and deposition of material behind the coastline. It is now clear that not all transgression phases did take place simultaneously in different coastal arease. For the sake of clarity, however, the traditional names are used here. The Dunkirk I trangression around Uitgeest can be dated to circa 350 BC (Vos 1983).
3 It is indeed virtually impossible to determine what alterations occurred and in that respect the test clays cannot be used as a direct reference for the prehistoric fabrics.
4 As no excavations were being carried out at the time, the clay deposit was located by drilling and a small pit was dug to extract clay that was considered suitable and representative (assesment by the author). The clays from Opperdoes en Hoogkarspel were not tested from a potters point of view.
5 These measurements, as well as the %AP measurements were carried out by E.Bulte, ROB.
6 For most sherds, the procedure for measuring the dry weight was slightly changed: to prevent re-hydration, the sherds were weighed immediately after cooling to about room temperature in the oven.
7 The formula differs from those mentioned in most other studies on AP. Usually the more simple formula of A-B/V is used in archaeology.
8 In 1987 a X-ray diffraction analysis was carried out for test tablets of clay sample 61, fired at 800, 900 and 1050 °C, and clay sample 62, fired at 800 and 1050 °C by Wevers (1987; internal report). The tests were aimed at establishing the different phases in recristallization of clay compounds, in relation to the porosity and shrinkage. The results of this very limited analysis do not add significant information for the present study.
9 It is, however, not unlikely that the unfired sample contained a shell fragment.
10 According to Heimann (1989, 137) iron can in some clays be adhered to the crystal lattice and in others will be expelled during firing, which can cause a difference in the colour. Further testing of the clays by for example DTA is needed to establish all the different phases of change in Fe- and Ca- compounds for each temperature.
Fig. 5.1 Examples of various types of fabrics, inclusions and Fe-infiltration in original and refired sherds, see page 107-108
Fig. 5.1.3 Details of Fe-inclusions and iron infiltration. Scale 150%
5.1.3a,b,d Fe-inclusions in a reduced state in the reduced core and oxidized in the surface layer 5.1.3c,e Fe infiltration and oxidation just below the surface and all around the edge of a sherd.
Fig. 5.1.1 (Uitgeest) and 5.1.2 (Schagen)
Fig. 5.1.1a Uitgeest-Gr.D., sample 1. Examples of clay types of refired sherds (core or surface). Scale 1:1 Legend 1a, from top to bottom row
Vessel nr. 18-4 core 18-13 surface 18-6 core Clay type 2 3 + scum 2 Type inclusion A/Ca(/Fe) (Fe)
-Vessel nr 18-8 core 18-1A core 18-1B surface Clay type 2 1 originally burnt Type inclusion A/Ca A/Fe Fe
Vessel nr 31-15 surface 31-15 core Clay type 3+ scum?
Type inclusion All?
Vessel nr 20-4 surface 19-23 surface 19-23 core Clay type 1(,2)+ scum layer 3 + scum layer
Type inclusion All All
Fig. 5.1.1b Uitgeest-Gr.D. sample 1. Original and refired core surfaces. Scale 1:1 Legend 1b
Vessel nr 31-8 refired 31-8 original 19-7 original Clay 1(,2)
Type inclusion Fe - A/Ca Vessel nr 19-15 refired 19-15 original 19-7 refired
Clay 1 Clay 3
Type inclusion All A A/Ca(/Fe) Vessel nr 31-3 refired (sherd 34-6-94) refired 19-22 refired
Clay 1(,2) Clay 1 Clay 2 Type inclusion none Fe A/Ca(/Fe)
Fig. 5.1.2a-c Schagen-M1. Examples of clay types and of inclusions in the original and refired sherds. Scale 1:1
Legend 2a, top to bottom row
Vessel nr 240-6, original surface 212-4, original core
Type inclusion A/Fe A
Vessel nr 240-6, original core 212-4, refired core Type inclusion A/Ca Fe/A
Vessel nr 240-6, refired core 212-4, refired surface Clay type 3 1 (Fe infiltration) Type inclusion
Legend 2b, top to bottom row
Vessel nr 194-1 original core 194-2 original core Type inclusion Ca/A A/Ca
Vessel nr 194-1 refired core 194-2 original core (oxidized)
Clay type 2 1
Type inclusion A All
Vessel nr 143-7, original core 127-1, refired surface
Type inclusion A?
Legend 2c, top to bottom row;
Vessel nr 159-1 original core 223-3 original core
Type inclusion A A
Vessel nr 159-1 refired core 223-3 surface Clay type 1 (Fe infiltration) 3
