Over 1,200 thin-sections have been examined and a large amount of supporting binocular microscope analysis has been undertaken in connection with this research. It is appropriate to question the results of this vast expenditure of time and effort, not only to evaluate their contribution to archaeological knowledge but also to ensure that methodological conclusions reached as a result of this work are made available to future workers.
Petrological analysis can be used to answer several different problems. It is possible to take a sample in isolation and determine from geological literature where the constituents of this fabric originated. Many examples of this method could be cited and the results are extremely variable in their usefulness. At the worst this method can provide no indication at all of the source of the sample, or even be positively misleading by implying a local source through some formula such as "all the inclusions present in this same could have originated in the locality in which sample was found". At best, this method provides the most forceful results possible if it can be shown that the inclusions in the sample have a limited geological outcrop. The work of Peacock on the Gabbroic clays of the Lizard region of Cornwall used in the south-western peninsula in the Neolithic period is the classic example of the method (Peacock, 1969). The source of the Gabbroic clay can be shown by geological argument alone to a small region on the south Cornish coast.
One such clear-cut result has been obtained in this study, from the analysis of Malvern Chase wares. On petrological grounds alone the source of these wares can be localised to a region c.20 miles by 6 miles in south-west Worcestershire so that it was possible in 1977 to publish a type series and distribution of the products of the Hanley Castle pottery before any archaeological evidence for the kilns had been discovered. Confirmation of this characterisation has come from the field-work of P. Ewence, who has discovered several areas of medieval and later potting waste in the parish of Hanley Castle.
Petrology can also be used is to compare the inclusions in one pot sample with those in others. The comparative samples can be pot sherds; other ceramic material such as brick or tile; or briquettes made from clay samples of known origin (Howard, 1982). This is the standard method used in this study and has lead to the characterisation of Minety ware, Gloucester TF41b and other groups.
Those fabrics which contain inclusions other than quartz sand are normally variable in the relative quantities, identity, size and roundness of their inclusions. Using these three characteristics it is possible to compare thin-sections, or groups of thin- sections, to show whether or not there are any differences present. If no differences are found then the samples can be placed into a single fabric group. A combination of archaeological and geological judgement can then be used to interpret the grouping. Not all acceptable petrological groupings are the products of a single production centre. A grouping might be a fabric produced at several kilns forming a contiguous area, as in the case of Post-medieval Welsh Borderland wares, or it might be a ware used only at one centre, such as that used to make Ham Green ware jugs.
The likelihood of the ware being produced at several places is lessened considerably if the ware contains an unusual combination of rock-types or minerals, even if these types on their own have a wide occurrence. If, as at Ham Green, the fabric is formed by the tempering of a clay with material which does not occur naturally in the same area, then vessels made of this fabric were either produced at the same site or were made by potters with the same training.
The final use of petrology is to examine changes in the fabrics produced in a single industry. This is often much more difficult than the previous two applications, since the range of inclusions is likely to be the same. To investigate possible changes, accurate measurement of differences in size, relative frequency and shape are needed. In each case this information is available from a study of thin-sections but is laborious to collect and analyse.
An example of these methods can be found in the work of Darvill and Timby, who have compared the quartz grain size frequencies in samples of different forms made at a Romano- British kiln site in north Wiltshire (Darvill and Timby, 1982). A conclusion from this study was that there was a correlation between the form or intended function of the vessel and the fabric. Similar results have been found in this thesis, the most common differences are between 'kitchen ware' and 'fine ware' and between thrown forms and roof furniture. It is also possible to show changes in clay or temper sources with time, resulting from the exploitation of new resources. Examples of this change are the Malvern Chase 'pink' fabric, utilised only from the late 16th century, and a sandy variant of Minety ware, which may be limited to the last stages of production at Minety (Chapter 2).
In Bristol, Glamorgan, Worcester, Malvern Chase and Ashton Keynes wares there are differences in the fabric of contemporary ridge tiles and pots. The ridge tiles in all cases have more tempering which includes larger fragments than are found in pottery and in some cases are in a less well-mixed fabric.
There are two possible explanations for the difference between ridge tile and pottery fabrics. Firstly, it is possible that the ridge tiles were more heavily tempered because they are thicker wares, which needed more tempering to allow water vapour to escape during drying and firing. The other possibility is that there was less care taken with the preparation of ridge tile fabrics than with pottery fabrics.
It is possible to choose between these two hypotheses by examining wares in which no difference is found between ridge tile and pottery fabrics, for example Hereford A7b and Minety wares. In both these cases it is likely that the clay was used as dug without cleaning and that no tempering had to be added. Therefore, when a clay was already suitable for potting without further work no attempt was made to add more temper to the ridge tiles.
In the examples where differences are found they are probably due to the clay needing to be prepared by tempering or cleaning before use. In either case, less care may have been taken with the ridge tiles than with the pottery. Bristol ridge tiles often contain large quantities of clay pellets, which were probably present in the parent clay. These are less evident in the pottery fabric, which may therefore have been crushed and sieved to remove them.
A difference between the fabric of 'kitchen ware' and 'fineware' is less common but when present is often more clearly defined than that between pottery and ridge tiles. At Ham Green, the cooking pots and jugs were made from quite different clays. The cooking pots were made in a red- firing clay while the jugs were made in a light-firing clay although the same quartz and limestone sand temper was probably added to both clays. There are also differences between the fabrics used for Coarse Border ware lobed cups and other vessels although the same clay was probably used for both types. The lobed cups were not tempered but other vessels have a coarse quartz sand temper.
The difference between Malvern Chase cooking pots and tripod pitchers in the 12th century is very marked. There is however no proof that the two types were produced by the same potters, only that they were made in the same area. Malvern Chase cooking pots and jugs in the late 13th to 14th century were more similar in fabric. The same clay may have been used for both types but with more sand or gravel was added to the cooking pot fabric. Differences between the fabrics of these types become less and less evident with time and by the late 14th to early 15th centuries there is no difference at all. In south-east Wiltshire, fine sand-tempered Laverstock jugs were produced in the same kilns as coarse sand-tempered cooking pots.
Industries in which no difference is found between the 'kitchen wares' and 'finewares' include Minety, Hereford Fabrics A2 and A3, Worcester, and the later medieval and post-medieval Malvern Chase wares. Some, such as Minety and the later Malvern Chase industry probably utilised clays which could be thrown without much preparation. The Hereford fabrics would have needed tempering since the gravel found in them is not found naturally mixed with Devonian marl.
Thin-section analysis and related studies can therefore reveal a considerable amount about the source and method of manufacture of medieval pottery and tile fragments. It could, however, still be asked to what end this information is needed. In Chapters 10 and 11 data provided by the characterisation of pottery has been used to show changes in the sources of supply of pottery with time and it is suggested that the evidence implies that by the later 12th and 13th centuries pottery was being distributed by highly efficient means, so that every settlement was able to acquire pottery from a number of sources with little difference between the type of pottery available in a remote village and that present in a major town. This is not a new conclusion. It was suggested by Beresford and Hurst on the basis of a number of excavations on deserted medieval village sites producing mainly 13th to 14th century pottery (Beresford and Hurst, 1971). This study it has been shown that this means of supply was used from the 12th century, and probably earlier.
Having shown that this pattern of supply is similar over a wide area of southern and western England it is doubtful whether further studies in other parts of southern England are needed to show that the same pattern is general. There are, however, large areas of northern England, Scotland, Wales and Ireland where the early history of the pottery industry is demonstrably different from that in southern England. At York, for example, there ware no break in the tradition of wheelthrown pottery production in the 11th century and in Scotland, Ireland and Wales local pottery production began in the 13th century. Quantified comparison of these areas with southern England is likely to be rewarding. Even if the conclusion is that there is no difference in the type of distribution found in these areas from that found in southern England, which would show that previous history and social composition had no effect on the operation of the pottery industry.
Using the features outlined in Chapter 6, the majority of the late Saxon and medieval pottery in the study region can now be placed into a chronological framework. Usually this means that a type can be dated within a century and sometimes within a half-century. Assemblages can sometimes be dated even more closely, to periods of 20 or 30 years. Further precision has not been achieved from pottery evidence alone. This is probably because the working lifetime of an individual potter would also have been of this order of magnitude and the period of use of a vessel, once made, may have been of similar duration. This was certainly the case for the late 17th to 18th centuries.
It is not possible to date a pottery type closely if it is not present in any stratified assemblage and is devoid of datable typological features nor if the pottery type is known in stratified assemblages but has a long life-span with no typological development. For example, even a large, stratified, later medieval assemblage, can only be dated within very broad limits since pottery types at that time changed very slowly.
'Early Medieval' cooking pot sherds are also extremely difficult to date unless found in stratified contexts. The type was current from the beginning of the 11th century through to the mid -14th century or later. Some Newbury B pitchers, for example, were originally published by Jope as late Saxon, although they are now known to be definitely post-c.1150 and possibly even of early 15th century date (Jope, 1947). Using petrological analysis sherds from vessels of identical form and made by the same techniques can now be given accurate relative dates (for example Gloucester TF41a and Gloucester TF41b sherds can look identical but while the first is unlikely to be later than the mid-11th century the second could be of mid-13th century date.
For the purposes of this study most medieval pottery is satisfactorily dated so that unstratified assemblages can be used to show the presence of a ware at the site and even the approximate frequency. However, too much reliance should not be placed on differences in relative frequencies, either for dating purposes or to reconstruct marketing patterns. Many of the tables in chapter 6 and the catalogues in the gazetteer in volume III show clearly that later assemblages almost always contain pottery of earlier date. Only if it is certain that these wares were not still in use can they be discounted. Therefore the relative frequencies of contemporary wares in a group are usually 'diluted' by residual pottery.
Despite intrinsic limitations there are a few aspects of the pottery chronology which are capable of being refined by archaeology. For example, confirmation is needed that all of the Late Saxon wares in the study region are later than the early 10th century since this has the historical implication that commercially produced pottery was neither used in the late 9th and early 10th century towns of the region nor in their surrounding hinterlands. It is therefore important to show that this late starting date is as true, for example, in Gloucestershire and Wiltshire as it appears to be in Somerset and Herefordshire.
The nature of late Saxon pottery in the north Cotswolds, Berkshire and on rural settlements in the Severn Valley is still unknown, as is the dating of their replacement by Early Medieval wares.
Late medieval groups datable to the second half of the 14th century or later are also required to show whether there are differences between the pottery of this period and that of the previous century and to see whether the introduction of 'Tudor' forms, such as the cistern, the lobed cup and the conical bowl, was sudden, as it now appears to have been, or gradual, as in the south-east of England.
Ceramic evidence for trade routes is equivocal, since the presence of a ware at a site may have any number of explanations. Documented examples of the movement of people and households in the medieval period are common and any one of these, or a unique event may explain the discovery of a single sherd. There is nevertheless more chance of explaining a consistent pattern, however uncommon the pottery type. The most likely explanation for such pottery movements is that they are the by-product of regular movements in other goods or people.
For the 10th century, there is so little data that it is not possible to distinguish the normal from the unusual. Therefore no evidence for Late Saxon trade routes can yet be presented. From the 11th century, onwards there is evidence for several trade networks, in which pottery was incidentally carried, allowing the network to be examined.
The first such networks links Bristol, Gloucester and Dublin and is indisputable evidence for direct movement of people or goods from Bristol to Dublin and from Gloucester to Dublin, or between other sites close enough to these two towns to be receiving the same wares (chapter 11). The evidence for this route is continuous at Dublin from the 11th century to the mid-late 13th century, at which point the archaeological sequence becomes very fragmentary. Other English coarsewares are found, for example south-east Wiltshire cooking pots and tripod pitchers, but since these are all present in Bristol there is no evidence for direct contact with Wiltshire. After the 11th century is there no evidence for Gloucester area pottery in Dublin, nor for any of the other 12th to 13th century wares which supplied Gloucester.
From the late 12th century onwards the Bristol-Dublin trade became a more general coastal trade with large collections of Bristol region wares at various sites along the South Welsh coast, declining in frequency to the west but still forming a substantial proportion of the pottery, for example, from Loughor on the Gower peninsula. This is in marked contrast with the evidence from the southern coast of the Bristol Channel, which is also liberally scattered with medieval sites but where Bristol region imports, although still found, are much rarer.
The evidence for this trade carrying on into the later medieval or post-medieval periods is scarce. Bristol wares are found at various South Welsh sites, but predominantly at sites in Gwent, such as Caerleon and Chepstow, which could be interpreted just as easily as part of a cross- river trade. Evidence for trade with Bristol from south Wales Glamorgan westwards or from Ireland is very rare, probably because by that time both areas had flourishing pottery industries. The 16th century Malvern Chase distribution as being part of a Bristol-based coastal trade but by that time there is also evidence for Severn Valley boats themselves travelling around the Welsh coast, carrying mainly malted barley (Lewis, 1927).
Another important 11th century trade route is that to Droitwich (chapter 11). Bath fabric A cooking pots and spouted pitchers are not only found at Droitwich itself but also at Worcester, Pershore and sites along the Cotswolds and in the Severn Valley. As traded pottery, there is no reason for their presence since there were several pottery sources much closer. The most likely explanation is that this pottery reflects the trade in Droitwich salt and that pottery in the late 11th century was an acceptable commodity to take to Worcestershire for trade, since the county did not supply its own pottery. It is interesting that in the 12th century, after the start of the Worcester- type industry, the quantity of non-local pottery at Droitwich decreased almost to zero as did the incidence of Bath fabric A vessels in Gloucestershire and North Wiltshire. However, the extraction of Droitwich salt continued (Berry, 1957). This may mean that the local distribution system may have ceased or altered or simply that pottery was no longer an acceptable item of trade.
The network which operated between the East Midlands and the study region from the 10th century to the early 13th century was more diffuse. The main ware in this trade was Stamford Ware and within this ware predominantly glazed jugs. It is possible that these were distributed non- commercially as 'prestige goods' since they were by far the best pottery serving vessels obtainable at the time. However, in the 11th century the trade also included utilitarian vessels such as East Midlands and Stamford greyware cooking pots and St. Neots type cooking pots and bowls. There was a decline in the trade in the late 12th to early 13th century but Developed Stamford Ware and St. Neots type jugs are found in the study region. Lyveden ware, which replaced Stamford ware locally, is not found in the study region, except for a single vessel from Holm Castle, Tewkesbury. It would appear that this East Midlands trade was mainly directed to the larger towns of the region, such as Hereford, Worcester, Droitwich, Winchcombe and Gloucester. However, very few rural settlements of early 13th century or earlier date have been examined in the detail needed to find these relatively rare sherds.
Other evidence for non-regular pottery trade is almost non-existent. There is a sparse distribution of Oxford AM jugs throughout Gloucestershire to Hereford and Worcester which is not matched by their distribution to the south into Berkshire, nor to the east into the Lower Thames valley. This distribution probably reflects the main east- west routeway from London to Wales but if so it is remarkable that the pottery does not travel in the opposite direction, to London.
The fall-off in frequency of Nuneaton glazed wares is also probably non-regular and examples are found more often in east Gloucestershire and North Wiltshire than one would predict. This may also be due to travel along a major routeway, following the Cotswold ridge and the Fosse Way from Bristol to the East Midlands. Bristol wares are also occasionally found on this route, for example at Cirencester, while the Minety production area lies just to the south of it and sent a huge proportion of its pottery to Bristol, from whence it reached other sites in Avon as well as being shipped to South Wales, the Wye valley and Ireland.
These examples, however, are rare exceptions and it would be true to say that the material from a site such as 143-5 Bartholomew Street Newbury is more typical (Vince, forthcoming b). Here, out of a vast collection of pottery covering the late 11th to 16th centuries, only a handful of non-local sherds were found, mainly of Oxford AM, Minety and possibly Salisbury/Laverstock wares. Three imported sherds were recovered, one 16th century Andalusian Albarello and the others mottled green-glazed Saintonge ware. There is no evidence that the site was particularly impoverished and indeed the collection of non-ceramic artefacts was unusually broad (Vince, forthcoming b). There is also documentary evidence for the presence of foreign merchants connected with the trade in wool and cloth but from the pottery one could not infer that Newbury was a flourishing market town let alone one of the most important centres for the 15th century cloth trade in the country.
The distribution of pottery in the Late Saxon to post- medieval periods does not therefore appear to be very responsive to variations in the local or national economy and it is therefore likely that the relative frequencies of pottery from different sources do not reflect the degree of contact between such places and the site. "There is no better proof of the deceptive failure of pottery to travel the obvious trade routes than the apparent absence of Italian wares from any Southampton context earlier than the late 15th century" (Platt and Coleman-Smith, 1977, 29).
However, this lack of correspondence is unlikely to be random. Detailed comparison between the documentary and archaeological evidence for trade may indicate in which circumstances pottery was transported and in which circumstance trade took place but pottery was not transported.
Within the study region there is insufficient documentary evidence for local and interregional trade for such comparisons to be made but international trade is relatively well-documented for the later medieval and post- medieval periods and it would be possible to compare the continental imports from Bristol and, to a lesser extent, the south Welsh ports with historical evidence for international trade (Carus-Wilson, 1933).
.fi 12tile.txtThe initial aims of this thesis were: to produce evidence for production and distribution of medieval pottery; to extract from the data regularities in location of production and distribution of the finished products and to examine the processes which governed pottery production and distribution.
The first aim has been adequately achieved for the 13th century onwards but for the 10th to 12th centuries there still needs to be much initial fieldwork before even the basic facts can be accepted. For example, there is an almost complete lack of data from rural sites until the 12th century but it is unlikely that earlier settlements were aceramic. It is probable that it is a consequence of the smaller pre-12th century population.
From the 11th to the 15th century pottery was used predominantly for two purposes, cooking and serving. In both areas of use there were other materials which could fulfil the same function but all were more expensive. However, in the late 13th to 14th century there was a steep decline in the use of pottery cooking vessels. Wood and metal were also used at this time for the manufacture of drinking vessels. In the later 15th century pottery versions of such cups became more frequent. During the 16th century the production of tableware in pottery became much more common, leading in the early 17th century to industries whose production was dominated by bowls, dishes and plates.
No principles governing the changeover from one material to another have been found. It is likely that the underlying principle is such that a slight fluctuation in the cost of production of vessels in one material could tip the balance between the different materials. For example, the changeover from ceramic to metal cooking vessels in the late 13th to 14th century was due to the increased availability of metal vessels, which itself has several causes, one of which is that the economy was able to support bronze-foundries where economies of scale and the use of casting techniques could bring down the price of metal vessels.
The later changes from wood to pottery drinking vessels and tableware and may have been influenced by advances in the techniques of production of the pottery vessels. Both types require more control over the clay in throwing than is needed in the production of ceramic jugs and cooking pots. Advances in the use of glaze and decorative techniques might also have been a factor although fluctuations in the relative cost of the raw materials involved may also have been important.
Towards the end of the sixteenth century potters producing drinking vessels had to compete not only with wood workers but also with locally produced glass. A number of glasshouses are known in the study region, mainly founded by Huguenot immigrants (Vince, 1977d fig.1). Earlier glass is very rarely found in archaeological contexts in the study region and cannot have had a serious effect on the production of pottery.
Changes in the quantity of pottery used were as important as these qualitative differences. Ceramic cooking pots of 12th to 13th century type would have been rough and ready utensils whose period of use might be as little as a year. This would create a steady demand for replacements. A comparison of the relative sizes of cauldrons and ceramic cooking pots shows that the metal vessels had a greater capacity (although this difference could not be quantified using examples from the study region). This difference may indicate variations in function between the types but is just as likely to be due to the inability of the potter to produce stable vessels of the size of the metal cauldrons (it was within the capability of the potter to make vessels of that size, for example storage jars, but it is possible that they would not have withstood the sort of treatment given to a metal cauldron).
Conversely, pottery vessels have an advantage over wooden ones for tableware (if glazed, they are easier to clean) but again it is possible that they would have been more fragile and therefore would need more frequent replacement. Royal orders for pottery from Kingston-on- Thames show that large quantities of jugs were ordered each year (Hinton, M., 1980). Five orders survive between 3rd November 1264 and 26th December 1266. Between 500 and 1000 jugs were ordered at a time and one order for 600 pitchers was followed only a month later by an order for 500 pitchers. Thus in 25 months the royal household obtained 3800 pitchers from one source alone. We have no information on the consumption of jugs by lesser households but it is likely to have been on a similar scale for the seigneurial households. People at the bottom of the social scale may well have had to take more care over their jugs, hence the occasional use of lead plugs to repair jugs, for example at Loughor Castle and Gloucester (Site 53/69). Social differences in the treatment of pottery vessels could give rise to identical assemblages from sites of differing social status which result from quite different patterns of use.
This difference in attitude towards pottery reflects its varying value in the society. It is likely that when glazed jugs were first introduced they would have been more highly prized than in the late 13th to 14th centuries.
The rate at which vessels were broken and the value placed upon them by their users would affect how much they would be willing to pay to replace them and this in turn governs both the rate and location of the production of new vessels.
The state of the economy would have played an important part in the development of the industry. All things being equal, the quantity of pottery in use in a society and the rate of acquisition of pottery both depend on the amount of economic surplus. The 'richer' the society is the more non- agricultural craftsmen the peasantry can support and conversely the desire for manufactured goods is one reason for the production of an agricultural surplus. The social structure would also play an important part in this equation, since if the land-owning class purloined all of the economic surplus then there would be nothing left for the acquisition of manufactured goods.
The spread of manufacture and use of glazed serving vessels in the 12th century may therefore be seen as a measure of the available agricultural surplus. Other evidence for the presence of an increasing surplus during the 12th and 13th centuries comes from the granting of markets and fairs to numerous settlements in the region. These were given not only to established towns but also to settlements which, to judge by present day topography, were mere villages.
We do not know how successful these rural markets were; only the large ones are mentioned in subsequent documents, and these documents are predominantly of the Tudor period and later (Everitt, 1967). It seems reasonable to suppose that the majority of these markets were not merely speculations by the grantee and that they fulfilled a need for a network filtering agricultural produce in one direction and manufactured goods, services and different agricultural goods in the other. .cp5 If this correlation is correct then it is of interest that the present dating of the introduction of glazed wares would place it consistently earlier than the granting of markets or borough status in the same areas. It may also be relevant that the foundation of both markets and boroughs is earlier in central southern England than it was in the counties of Herefordshire and Worcestershire (Beresford & Finberg, 1973). This, together with the general increase in pottery distributions in the later 12th century, may be strong evidence to show that the foundation of towns and markets was built upon a pre-existing surge in the local economy rather than bringing such an upsurge into existance.
Although individual industries may have had unique histories it is possible to recognise the same stages in the development of the pottery industries of most regions of southern England, for example the London area, the South West (J. Allen, pers. comm.), East Anglia (Clarke and Carter, 1977; Jennings, 1981) and the South East (Streeten, 1980, 1982).
The early ceramic development is different in all areas, but from the late 11th century onwards the same pattern is found. Small-scale 'early Medieval' industries existed everywhere except in the western Celtic fringe. At some stage, thought to be in the mid-late 12th century, there is an improvement in technology, sometimes involving the introduction of the wheel and an increase in average distribution distance. This marks the beginning of the 'Medieval' industries, even though in some areas the change is less obvious than others.
In the next stage, industries of regional scale emerged out of the 'medieval' industries. The distinction between 'medieval' and 'regional' industries is an arbitrary one made solely on the basis of distribution. There is, however, a definite break between those industries supplying centres more than 40 miles away with more than 10% of their pottery and those that do not, most of which are substantially smaller (chapter 10). The earliest example is that of the Surrey-Hampshire border, which achieved 'regional' scale in the mid-14th century. Within the study region the only example is that of Malvern Chase in the late 15th to 16th centuries. In the south-west two regional scale industrys produced North Devon and South Somerset wares respectively. These industries too had their origins in the medieval period and underwent a period of growth in the late 15th and 16th centuries.
The late 16th to early 17th centuries saw the emergence of factories producing specialised finewares, such as light-bodied earthenwares, tin-glazed wares and, from the late 17th century onwards, stonewares. The final stage in all the sequences, is a network of small country potteries producing lead-glazed earthenware (Brears, 1971). These potteries can be recognised as early as the 16th century and still exist today.