Porcelain waste and porcelain production in Worcester: the landscape evidence from fieldwalking

Summary Field survey can provide a longitudinal sample of ceramic production from the porcelain factories which operated in the city of Worcester from 1751 to 2009. Until the twentieth century domestic and industrial waste from the city was used as manure/hard core on fields throughout the county. We have found kiln waste and biscuit wasters from the earliest phases of production to wares made in the late 19th/twentieth century, along with contemporaneous domestic discard. This assemblage provides a series of snapshots of porcelain and semi-porcelain production allowing us to identify trends in types of fabrics, forms, manufacturing and consumption.


INTRODUCTION
This paper discusses the archaeology of industrial porcelain production in Worcester from 1751 to approximately the end of the nineteenth century.Traditionally, industrial archaeology was focused on monumentality, e.g., the conservation and recording of factories. 1To this has been added the study of production and consumption, in glass, pottery, and metal. 2 Johnson has characterized the archaeology of the period between 1750 and the present as the archaeology of capitalism, a perspective that Belford has also advanced for the West Midlands via the West Midlands Research Strategy. 3Belford highlights one of the biggest challenges of this era as the complexity and interconnectivity of factories, their outputs and the landscape, as the scale of production increases throughout the 19 th and 20 th centuries. 4n this paper we want to address the scale of industrially produced ceramics and its mundanity.This material had social relevance and meaning.Robb cites Gell's 'technology of enchantment' as a way of understanding how everyday technologies and their design are seen by people to have both social and economic relevance. 5Often in the study of porcelain it is the pieces of most importance to art that are studied.Pieces designed for elite consumption and produced by the elite artists at the factory.Robb suggests that there is also meaning to be found in the 'invisible' everyday items created by these factory systems which ultimately were discarded in huge numbers in the landscape of Worcestershire.This perspective gives scope for the study of the archaeology of both factory and domestic waste disposal, allowing for the sheer quantities involved to have meaning in and of themselves.
The archaeology of porcelain and refined earthenware production has focussed on the factory site, 6 with in situ analysis of wasters, e.g.Owen for Worcester Tonquin, 7 Edwards for Nantgwy, 8 or more commonly, through chemical analysis of items from museum or private collections thought to derive from specific factories. 9As Palmer, Nevell & Sissons state, industrial archaeology must also take a landscape approach to understand the impact of factory life on the surrounding environs. 10he project formed part of the University of Worcester's former undergraduate degree in Archaeology and Heritage, Introduction to Fieldwork module.The pedagogical aims were to train students in field walking and finds processing techniques.We used a total pickup strategy, which recovered everything from Middle Palaeolithic flint to modern plastic. 11In addition to the pedagogical aims we were aiming to understand the landscape history and archaeology of the environs of St. Johns a suburb of Worcester.This paper is concerned with our findings relating to the record of porcelain and refined earthenware production.Worcester from the 18 th to the twentieth century was home to a number of ceramic factories. 12The impact of these expanding enterprises includes more than the footprint of the individual manufactories, it includes transportation networks, the economic impact of the employees, and extensive transportation of factory and domestic waste to the rural hinterlands.
Specifically, we wanted to use material collected through field survey to contribute to the historical and technological narrative of porcelain, bone china and semi-porcelain production in Worcester; to assess the proportion of porcelain and bone china consumption in the local market; to identify periods of deposition of both domestic and industrial waste in the landscape.
Fieldwalking has several advantages for this type of study.Firstly, excavation of the sites of production is rarely possible so new information is only available as and when development takes place.Secondly, it is important to understand what the extent and composition of these artefactual assemblages is.Finally, fieldwalking is relatively cheap and can be learned relatively quickly and is an ideal technique to use with students, volunteers, and community groups.As none of the factory and domestic discard material was in situ we used a simplified fieldwalking technique that we had previously used in Mediterranean field surveys where large amounts of ceramic material is often found. 13By using these methods we were able in two mornings work to recover a sample of domestic and industrial discard from the city of Worcester covering the main periods of ceramic production.
The use of plough zone fieldwalking and collection has a well-established pedigree in archaeological methodology.Artifact collection from ploughed fields has previously been disregarded as lacking stratigraphic context and therefore useless in providing meaningful knowledge apart from site reconnaissance.From the 1970s onwards prehistoric archaeologists have studied and analysed plough zone artifact assemblages, producing reports which emphasize larger scale artifact distributions and interpretations. 14Historical archaeology has some different challenges in accessing plough zone data.Firstly, as Brooks et al. state  15 for the case of Australia, there may not that many historical sites which have been subjected to ploughing, and so the technique has tended to be under-utilized.Secondly, there is a perception that historical archaeology is more complicated, sites more detailed, and hence anything other than clear stratigraphic relationships are too coarse grained for study. 16ithin the United Kingdom, there is even less use of plough zone data in Historical Archaeology, despite the long pedigree in prehistoric research.There are a few field survey projects published with 18 th -twentieth century materials recorded or analysed, such as the Tamar Valley, Cornwall field training event, 17 and the Bingham Heritage Trails Association community project. 18In these cases, only limited analysis of the ceramic assemblage could be completed, due to the large quantities of materials recovered as well as a lack in expertise with refined earthenwares and porcelains.Historical Archaeology has traditionally focussed on in situ strata, whether domestic 19 or factory 20 .However, until the late 19 th and early twentieth century, waste dumping on agricultural land for both domestic and industrial purposes, comprised a significant cultural activity and should therefore be considered archaeologically relevant.
Rathje's long running Garbology project in Arizona demonstrated the cultural value of secondary deposition, albeit in a landfill context. 21Outside of landfill, the dumping of midden and factory waste for agricultural purposes means the archaeology will have undergone ploughing deformation.With that caveat, we can confidently analyze the data provided and answer a number of relevant questions.First, what was being deposited in the field and when was it deposited.Second, what is the make-up of the factory waste material being deposited, and what can it tell us about the industrial activities of the porcelain factories at that time.Third and related, what is the make-up of the domestic midden was being dumped, and what can it tell us about the consumption patterns of the residents of the city of Worcester and the St John's environs.
Fieldwalking has a vital role to play in contributing to the historical archaeological narrative of Worcester, both industrial and domestic.At the time of press, there are few published excavations relating to porcelain production in Worcester, though one is forthcoming, and so the data presented here represents a significant contribution in our knowledge of both manufacturing techniques and waste disposal habits.Fieldwalking is also a relatively cheap and accessible form of archaeological activity, can be learned relatively quickly and is an ideal technique to use with students, volunteers, and community groups, as demonstrated by the Tamar Valley project, amongst others.We used the techniques of collection pioneered in Mediterranean field surveys where large amounts of ceramic material are often found. 22By using these methods we were able in two mornings work to recover a sample of domestic and industrial

PORCELAIN WASTE AND PORCELAIN PRODUCTION IN WORCESTER
discard from the city of Worcester covering the main periods of ceramic production.

CONTEXT AND PREVIOUS WORK
Fieldwalking was carried out at Oldbury Farm (SO827 554) in 2016 (Fig. 1). 23CgMs had produced two desk-based assessments and a magnetometry survey of the area fieldwalked. 24The results of the desk-based survey showed that there are no known settlements within the area walked.The nearest habitations are Ambrose Mill/Farm to the east, and Oldbury Farm to the west, dating from at least the eighteenth century.The wider area began to be infilled in the middle of the nineteenth century, and remained dominated by orchards and farms well into the 1960s.
The site at Oldbury Farm consists of long standing arable which had been allowed to go fallow for a season before fieldwalking.Historic and Ordnance Survey maps indicate that this area has been under cultivation since the eighteenth century and was part of the Temple Laugherne estate. 25Fieldwalking was conducted by two supervised teams of students.Collection was based on the British School of Rome's survey technique, with students organized into transects.All cultural material was collected regardless of age, with the transect context retained throughout the analysis.Finds collected spanned from the Middle Palaeolithic through Roman, Medieval and Post-Medieval periods. 26he post-medieval and modern material recovered from Oldbury Farm could come from a number of activities, though the likeliest is through manuring activities from the newly established Ambrose Mill/Farm.This supposition is supported by the large quantity of buff slipwares and redwares which date locally from the late 16 th through to the 19 th centuries, and the lack of high-status materials such as tinenamelled earthenware (Delft).The appearance of 18 th and nineteenth century porcelains and refined FIG. 1 Location map showing oldbury farm and worcester porcelain factories, worcester, including holywell. 137arthenwares in the form of both domestic discard and factory waste coincides with the expansion of the St John's suburbs, as well as of factory production in the city itself.In 1764 a manure depot was established at Pitchcroft to receive the city 'miskins' or local middens, before redistribution on local fields. 27rregular dumping of city waste occurred in the environs surrounding Worcester until the establishment of the first official landfill site in 1930. 28

HISTORY OF PORCELAIN AND THE WORCESTER PORCELAIN INDUSTRIES
The history of the invention and development of porcelain dates from the Tang dynasty in China in the tenth century where the combination of kaolin clay, feldspathic rock and high fire kilns produced a glassy, translucent, almost vitreous fabric. 29In the thirteenth century during the Yuan dynasty, the Chinese potters developed what became known as 'eggshell' porcelain, introducing blue underglazes as decoration. 30Between the 16 th and early 18 th centuries, Italian, French and German potters began to develop successful versions of porcelain, aided by the availability of kaolin-type clays as well as royal patronage. 31ritish experiments in porcelain production were hampered by two key factors: the lack of appropriate clays 32 and the absence of royal or parliamentary patronage. 33The solution was the combination of access to Cornish white clays and the results of close to half a century of experimentation in key factories such as Limehouse in London, 34 Derby, 35 and Lund's in Bristol, 36 funded by early investors including the East India Company. 37However, with the lack of access to appropriate 'true hard-paste' porcelain clays, the end products were essentially synthetic porcelain substitutes. 38he 18 th and 19 th centuries are periods of industrial expansion and globalization of trade in the West Midlands.The city of Worcester expanded into light and heavy industry from the eighteenth century onwards.There were three main ceramic factories within the city limits which produced and traded in porcelain and refined earthenwares between the 18 th and 20 th centuries (Fig. 1).Dr John Walls opened the Worcester Tonquin (later Royal Worcester Porcelain) at Warmstry House in 1751, after a period of experimentation on a site on Broad Street. 39The first kilns were sited along the river at the end of Copenhagen Street. 40By the end of the eighteenth century a second porcelain factory had opened, Chamberlain's on Severn Street, and by the beginning of the nineteenth century, a third, Grainger and Wood joined them, producing semi-porcelain as well as porcelain at a site at Lowesmoor. 41In addition to these factories, there is evidence of early porcelain experimentation and a kiln site at Holywell House, St John's sometime in the mid eighteenth century (Fig. 1). 42

THE ARCHAEOLOGY OF WORCESTER PORCELAIN PRODUCTION
Sandon and Barker conducted a rescue excavation of an eighteenth century Worcester Tonquin kiln in 1968. 43Large quantities of the earliest soft-paste porcelain production were recovered including wasters of the prototype sauce boat which was Worcester Tonquin's earliest commercial success. 44This material, much of which is still in storage in the Worcester Museums archives, has not been fully published, with the bulk of material unavailable for access. 45A number of chemical analyses have been produced on a handful of waster, confirming the close relationship between the Bristol factory soft-paste recipe and Worcester Tonquin. 46ubsequently, two developer-led projects of porcelain factory remains have taken place, a Grainger's kiln site in Lowesmoor, and at the Chamberlain/Royal Worcester Porcelain factory at Severn Street.The Severn Street factory has produced a comprehensive historic building record, as well as revealing a quantity of modern material in fills and during watching briefs across the site. 47The industrial period materials were only analysed if found in sealed deposits or contexts. 48Unfortunately, this has inadvertently resulted in a lack of reporting of modern pottery in the finds reports.For example, in a report on a watching brief held for service trenches and ground reduction at the Severn Street factory, the pottery report excluded refined earthenwares, 49 though within the body of the report a number of such sherds are mentioned. 50The as yet unpublished excavation in advance of development of a Grainger's Kiln in Lowesmoor, revealed the floor and subsurface structures of a kiln and other factory elements, along with a substantial quantity of materials. 51There are a number of grey literature reports of excavations of pits and quarry pits in the city in which large quantities of factory waste material are noted, but it is not clear from the report summaries if the ceramic assemblages were kept or analysed.
Elsewhere in the UK there are a number of publications which have dealt exclusively with 18 th and nineteenth century porcelain materials from domestic excavations such as Pearce 52 and Pearce et al. 53 for London.These suggest that Worcester was a major exporter of porcelain in these early periods.There are also an increasing number of reports on excavations and finds analysis of other porcelain factories, for instance the Limehouse porcelain manufactory, London, 54 and Musselburgh Pottery, Scotland. 55n summary, despite the potential for the archaeological investigation of the three factories which

PORCELAIN WASTE AND PORCELAIN PRODUCTION IN WORCESTER
ultimately made up the entity of Royal Worcester Porcelain, there has been only limited archaeological analysis to date, with no large-scale exploratory excavations yet published.Scientific analysis has been restricted to small scale studies of a few sherds, and the largest potential collection of excavated sherds from the Warmstry House kiln has never been fully analyzed or published. 56The scale of the volume of material created through excavation has led to problems in which significant assemblages from the city languish in the stores for many decades unanalysed.Subsequent chemical analysis of both archaeological and museum materials demonstrates a complex history of porcelain innovation and competition in the city.The advantage of fieldwalking to locate the products of the factories is that the resource is readily available in the fields for collection.

RESULTS
The aim of this paper is to provide a chronological record of factory deposition on the site of Oldbury farm, as part of understanding the technological changes in manufacturing which occurred amongst the Worcester porcelain factories in the 18 th and nineteenth century.We will use factory waste as a proxy for the change and continuity taking place in Worcester, as the companies compete and merge over time.The finds also include evidence of domestic consumption and discard of both porcelain and refined earthenwares (e.g.creamware, pearlware and whiteware or 'CC' ware).The results presented here are those attributes which reveal these processes, including: (1) change from one fabric ware to another, (2) fabric variability within fabrics, (3) variations in sherd dimension, particularly wall thickness and sherd weight and decoration.For the domestic discard material, we only examined the porcelain for its origin and decoration.Aside from whiteware wasters, we will not be reporting in depth on the creamware and pearlware domestic material.
The sample (n ¼ 547) consists of porcelains, refined earthenwares and kiln furniture which have a production span of between c. mid eighteenth century to potentially the early twentieth century.The porcelain material is divided by ware types -soft-paste and hard-paste porcelain and bone china, and the refined earthenware is divided into the ware types creamware, pearlware and whiteware (Table 1).Each ware type is further divided into domestic discard and factory waste, with the porcelain categories dominated by factory waste, as will be discussed below.
Determining the presence of porcelain and bone china in large assemblages of modern material is problematic. 57We used a three-step process in identification, transmitted light to detect translucency, a scratch test to separate soft-paste from hard-paste and bone china and finally shortwave UV light to determine origin of the porcelain (European, British, Asian). 58In the absence of sophisticated instruments to analyse the paste we used the transmitted light colour as a proxy for variation in fabric composition. 59asters were identified by the following criteria: white bodied unglazed biscuit fired sherds, white bodied crazed or burned glazed sherds, warped sherds either biscuit or glazed. 60Domestic sherds were identified by sound glazing and clear decoration.These criteria are based on a conservative definition of waster which ignores the presence of factory seconds.This will inevitably result in the over representation of porcelain wasters in the domestic category.However, the results show that porcelain is a low percentage of the domestic assemblage overall, suggesting that the criteria as set is adequate for the analysis.
We used the Worcestershire Ceramic Database codes for our database. 61Though an excellent resource, it is, however, limited for post-medieval and modern ceramics.There are single codes for creamware (84), 'Worcester Porcelain' (83.1), and 'modern china' (85).There are no separate codes for pearlware or whiteware, or for soft-paste, hard-paste and bone china.

SOFT-PASTE PORCELAIN
Soft-paste porcelain production began in 1751 at Wall and Davis's Worcester Tonquin factory, Warmstry House, and continued in production until 1792, when it is said to have been replaced by a hard-paste fabric. 62Soft-paste porcelain is a synthetic porcelain comprising soaprock and ball china clay which when fired at high temperature produces a white bodied, thin walled, translucent vessel. 63irst period soft-paste fabrics undergo a number of modifications, some of which are better documented than others. 64There have been several attempts to investigate Worcester first period fabrics, usually using samples from museum or private collections. 65One of the few chemical analysis of sherds from an archaeological context is by Owen of wasters from the Warmstry House excavation. 66This analysis verified that prior to 1751, Wall and Davis were experimenting with a phosphatic-magnesium paste. 67 similar study by Jay of a group of sherds from the same excavation using Raman-spectroscopy and electron microscopy confirmed the identification of an early fabric utilizing flint glass. 68This fabric is soon replaced by the Bristol soaprock recipe, undergoing further modifications by 1760, 69 until the Flight and Barr period at the Warmstry House factory 1792, and soft-paste recipes are replaced by hard-paste (see below). 70ecoration of first period porcelain was usually blue hand painted underglaze, particularly pre-1760, followed by a clear over glaze and less commonly, polychrome. 71Further colours, including yellows and purples in addition to gilts occur more frequently after 1770. 72FT-PASTE WASTERS There were 42 waster sherds, representing 57% of the soft-paste sample.Within the waster sample, there were 36 biscuit fired sherds and six glazed sherds.The main type of transmitted light was a red colour, with 29 sherds and thirteen sherds in other shades (Table 2).The single waster with a green transmission was a glazed waster and may represent a pre-1760 firing. 73Glazed wasters had either a clear or a creamy glaze, with no transfer or painted decoration.
The waster sample was dominated by bases and body sherds, with a significant minority of teapot rims (Table 3).The reconstructed vessel classes included two bowls, two plate/platters, eight teapots, and fifteen teacups, including fourteen bases with reconstructed EVE diameters of between 50 mm and 100 mm.Two sherds had moulded decoration: a platter with raised features and a fluted sherd from a cup or teabowl.

SOFT-PASTE DOMESTIC DISCARD
The remainder of the sample (n ¼ 33) is domestic discard.Of that sample, 20 sherds were undecorated apart from a clear glaze, with the remaining thirteen sherds decorated in a variety of ways.Three sherds had blue and white hand painting, two sherds had blue transfer printing and another two had brown and white and grey and white transfer printing (Fig. 2).
The domestic discard is dominated by a clear transmitted light (n ¼ 20), with smaller numbers in the other colours.Three sherds had the characteristic green transmitted light thought to be associated with the earliest phases of Worcester.Both the wasters and the domestic discard are similar, in that they are dominated by a particular transmitted colour.In the case of the wasters, it is a red transmitted colour, and in the case of the domestic discard, it is clear.It shows that there were a variety of pastes in production and circulation at this time.There were proportionately fewer bases (n ¼ 4) and teapot rims (n ¼ 1) in the domestic sample which was dominated by body sherds (n ¼ 19) (Table 3).There were more rims (n ¼ 8) in the domestic discard possibly representing tea bowls or similar.As with the wasters, the assemblage was dominated by vessels associated with liquid consumption with teapots and teacups/bowls forming the majority with few identifiable fragments of plates or platters.

HARD-PASTE PORCELAIN
Using a scratch test, it was possible to identify a harder paste porcelain material.This was a surprising result as early British hard-paste porcelains are rarely reported from the archaeological and fine art literature. 74It was identified as hard-paste using two methods firstly a steel file leaves a grey mark and secondly is distinguished from bone china, European and Asian hard-pastes by UV analysis.It fluoresces a pale pink producing a similar colour to the soft-paste porcelain wasters (Fig. 3).There are several documented attempts at producing a synthetic hard-paste porcelain prior to 1780 in the UK, including Bristol (1749), Bovey Tracey (1750, 1766) and Plymouth (1768). 75Documentation of the transition from soft-paste to hard-paste in Worcester is unclear.The period in which hard-paste fabric recipes began to be widely adapted coincides with the end of the first period at Worcester, and the beginning of new ownership under John Flight and Martin Barr. 76However, it also coincides with the foundation of Chamberlain's porcelain works in Worcester, in direct competition with Flight and Barr. 77According to Godden, a chemical analysis of an early Chamberlain's cup yielded a recipe very similar to samples from both Bristol and New Hall, 78 sharing a high firing temperature.
When a similar analysis of firing temperature was carried out on a sherd of Flight and Barr 'hard-paste' from ca 1795, the sherd melted thus highlighting two very different fabrics. 79It is worth noting therefore that so called 'hard-paste' porcelain refers to the harder quality of the fabric against scratching and possibly not a higher kiln temperature.This is the first time this fabric has been recognised archaeologically as far as the authors are aware.As such it represents a significant addition to our knowledge of the production at the factory sites in Worcester and elsewhere.

HARD-PASTE WASTERS
The hard-paste waster sample is less than 50% of the total hard-paste collection (n ¼ 17).Ten of the wasters were biscuit fired and eight were glazed.The decorations for the wasters ranged from embossing, incising, and moulding, through to hand painted concentric lines (Fig. 4).There is a single waster sherd with blue hand painted chinoiserie.The transmitted colour for wasters was highly variable with red (n ¼ 5) and yellow (n ¼ 5) the most common, followed by clear (n ¼ 3), green (n ¼ 3) and orange (n ¼ 2) (Table 2).UV fluorescence was a consistent dull pink (Fig. 3).Sherd type was evenly distributed across base (n ¼ 5), body (n ¼ 5), rim (n ¼ 4) and teapot rim (n ¼ 3) (Table 3).Finally, there is a single thick slab of hard-paste porcelain, very crudely formed and unfinished, which may have served as a test tile for firing.

HARD-PASTE DOMESTIC DISCARD
The collection of 38 pieces of hard-paste domestic discard contains European and Asian, as well as presumably English porcelain examples.The transmitted colours were dominated by yellow (n ¼ 12) and clear (n ¼ 11), followed by green (n ¼ 8) and orange (n ¼ 6), with a single example of red (Table 2).The UV fluorescence revealed three different fabrics, with six examples of magenta and a single example of lilac, in addition to the 31 pieces with a pale pink fluorescence (Fig. 3).Asian hard-paste porcelain fluoresces magenta, and European hard-paste porcelain fluoresces lilac (Fig 3).There are six examples of Asian hard-paste porcelain, all body sherds, a single sherd of which is clobbered, and a single fragment of unglazed European hard-paste possibly a branch or vegetative fragment from a figurine (Fig. 4).
Of the 38 pieces of hard-paste domestic discard, eight were base sherds, 18 were body sherds, eight were rim sherds, with single examples of a handle, a spout and a teapot rim (Table 3).The decorations of the domestic discard ranged from simple hand painted blue lines, transfer patterns, including blue and white basket weave, moss/vermicelli, and geometric transfer prints, embossing and underglazing (Fig. 4).Though some sherds were decorated in chinoiserie, there were no examples of Willow pattern.

BONE CHINA
The first commercially successful bone china is produced by Josiah Spode around 1790. 80 According to Edwards, 81 'porcelain is defined as a vitreous ceramic material which has been subjected to high temperatures during which key elemental oxides and compounds used in its formulation have been converted into a translucent body … Bone chine is defined as a porcelain compound containing bone ash, a feldspathic material and kaolin, producing a ware with a translucent body which contains a minimum 30% bone ash derived from calcined animal bones.' Bone china has the distinction of being the strongest of the porcelains, as well as one of the easiest to pot and to fire.Edwards also notes that though bone china technically is in common use by 1815, variants of bone ash are a constituent component of many of the earlier soft-paste experiments, including

PORCELAIN WASTE AND PORCELAIN PRODUCTION IN WORCESTER
Worcester. 82Bone china or English china production continues today and is one of the most common fineware fabrics. 83

BONE CHINA WASTERS
There were 41 wasters identified in the bone china collection, of which 36 were undecorated biscuit fired and five were glazed.
The waster sample was dominated by body sherds (n ¼ 16) along with rims (n ¼ 10) and teapot rims (n ¼ 10), and finally, five base sherds (Table 3).All of the teapot rims recovered were wasters as well as the majority of the base and rim sherds.The reconstructed base diameters ranged from 50 mm to 180 mm, though the most common measurements fell between 40 mm and 80 mm, indicating cup or small bowl vessel classes.The reconstructed rim diameters fell between 70 mm and 200 mm, with the most common measurements between 70 mm and 100 mm, also confirming cup or small bowl vessel classes.Bone china sherds were noticeably thinner, lighter, and smaller than their earlier predecessors (Table 4).Simple comparison of means and modes of breadth and length comparisons, due to the clear differences.These fine vessels perhaps being at more risk of breakage into smaller fragments.They are also easy to spot due to their brilliant white appearance.
Of the decorated sherds, one had abstract embossing, one had a black and white hand painted floral design, one had a black and white floral transfer pattern, one had hand painted blue concentric circles, and one sherd had scallop patterns pencilled in, presumably as a guide (Fig. 5).In addition, the sherd with the hand painted floral underpaint design had evidence of grinding or smoothing along one edge (Fig. 5).

BONE CHINA DOMESTIC DISCARD
The sample of bone china domestic discard (n ¼ 21) consisted of fourteen body sherds, six rim sherds and a single base (Table 3).The domestic discard was dominated by clear (n ¼ 8) and green (n ¼ 9) transmitted light, with a single example of transmitted red and two examples of orange.The UV light fluorescence was bright white (Fig. 3).Overall, bone china, both waster and domestic discard, shows far less variability in transmitted light colour than either soft or hard-paste porcelain.Sherd types for the domestic discard included fourteen body sherds, six rim sherds and a single base (Table 3).There were no teapot rims, handles or other shapes.
Taking together both wasters and domestic discard is a similar pattern of discard to the hard-paste collection as a whole.The identifiable vessel types were almost exclusively teapots (n ¼ 10) along with two teacups/teabowls and a single larger bowl.The decoration on the bone china was far and away the most interesting and varied of the porcelain collection.There were fewer glazed undecorated sherds, and again, though some transfer patterns, no Willow pattern.The most common transfer pattern was three sherds of Royal Lily, a very popular pattern used by Worcester Porcelain, Chamberlain's and Caughley (Fig. 5). 84The Royal Lily pattern appears during the Flight period at the Worcester Porcelain and is found on soft and hard-paste porcelains and bone china from 1788 until well into the twentieth century. 85The most common pattern for both handpainted and transfer sherds are floral designs.A single sherd had embossing.WHITEWARE Among the wasters were a number with no translucency and which could not be scratched with a file.These could potentially be what Grainger's called 'semi-porcelain' though there is little archaeological literature around this material.Increasingly white bodied refined earthenwares, otherwise known as CC ware (e.g.Miller) appear widely after 1820. 86rainger & Co., began trading ca.1800, concentrating primarily on decorating porcelain blanks purchased from companies such as Caughley. 87They were also notable for producing a form of 'hybrid hard-paste', but also by the early to mid 1800s had developed 'chemical porcelain' and 'semi-porcelain', for use in industry as well as household items such as doorknobs and knife handles. 88They were bought out by Royal Worcester in 1889, though the factory continued production until 1902. 89Excavation in advance of development of a Grainger & Wood kiln in Lowesmoor, Worcester, has revealed substantial whiteware production. 90

WHITEWARE WASTERS
There were 23 sherds, two base sherds, fifteen body sherds, four rim sherds, and a single handle, lid, and miscellaneous piece (Table 3).The measurable diameters of both bases and the lid were 70 mm.The single vessel rim sherd had a measurable diameter of 140 mm.The only unidentifiable class was the piece of tile which was 400 mm x 360 mm in dimension, with a thickness of 60 mm and a weight of 14 g (Table 4).The whiteware wasters were undecorated biscuit fired sherds, with no evidence of underglazing or embossing.There are three whiteware biscuit fired wasters showing evidence of bloating, which Edwards attributes to the makers struggling with the chemical balance of their ingredients, particularly the presence of calcium. 91

WHITEWARE DOMESTIC DISCARD
There were 81 sherds identified as domestic discard.There were thirteen bases, 50 body sherds, fifteen rim sherds, a single handle, and a miscellaneous piece (Table 3).Decoration ranges from blue and white transfer ware, dominated by Willow pattern through to coloured transfer prints in chinoiserie, floral, vermicelli/moss and landscape patterns. 92n the overall assemblage we see a very different collection to the earlier porcelains and contemporary bone china with the emphasis on thicker heavier robust vessels and less emphasis on tea and other drinks consumption.The identifiable vessel types were plates, open bowls and a tile.We have not investigated the fabrics further.

KILN FURNITURE
Kiln furniture is made up of pieces of clay used to support and protect pottery during biscuit firing. 93The clays used can range from earthenwares, such as redwares, to whitewares. 94The forms vary greatly, and to date there is no published typology for kiln props, saggars, or other ceramic aids of manufacturing. 95here were 51 pieces of kiln furniture collected from Oldbury farm.Twenty-five pieces of kiln furniture transmitted light and so are porcelain or bone china, with the remaining 26 opaque.Ten pieces transmitted a clear light, ten pieces transmitted orange, and the remaining five pieces transmitted red (Table 2).No UV analysis was conducted on the kiln furniture.
The kiln furniture came in four forms: ring, ribbon, strap and triangular cone/prop (Fig. 6).There were 36 rings, ten ribbons, four straps, and a single triangular prop.All of the pieces which had transmitted light were rings.Ribbons had dimensions on average of 197 mm x 110 mm, with a mean thickness of 9 mm and weight of 2 g.Rings had dimensions on average of 310 mm x 120 mm, with a mean thickness of 11 mm and weight of 6.2 g.The four straps ranged between 220 mm x 130 mm to 320 mm x 170 mm in dimension, and between 3 g and 9 g in weight (Table 4).It was possible to calculate EVE's diameters for the thrown rings.Reconstructed diameters ranged from 50 mm to 310 mm, with a median and mode of 90 mm.If we understand that rings are used to support bases of pots during firing, we can estimate that the majority of the rings were used in the firing of small to medium sized cups, bowls and other vessels.Saggars were treated as brick and tile, counted in the field and discarded.Approximately 57 pieces of course kiln furniture, e.g.saggar fragments, were counted and discarded.

MISCELLANEOUS STONEWARE AND CREAMWARE WASTERS
In addition to the porcelain/bone china and whiteware wasters identified, there were low numbers of stoneware wasters, a single creamware waster, and sherds which either demonstrated inconsistencies in FIG.6 1. 16W57 creamware waster; 2. 16840 lilac waster rim; 3. Kiln furniture.

fabric and firing, or which demonstrate ongoing experimentations.
There is a single glazed creamware base, with visible inclusions in the fabric (Fig. 6).There is no documented evidence that any of the factories produced a creamware during the eighteenth century, though this suggests experimentation must have occurred.There are two stoneware wasters, despite the lack of a known stoneware producer in Worcester, and there is a single sherd of very thin, biscuit fired lilac coloured refined earthenware (Fig. 6).

CHRONOLOGY
Our results demonstrate it is possible to provide approximate dates for the episodes of factory waste disposal, using known criteria.Our results also demonstrate that a more refine chronology can be achieved by the use of both translucency and UV luminescence.In particular we have shown that the hard-paste porcelain recovered from Oldbury Farm is likely to have been the earlier product of the Worcester Company and dating to the last part of the eighteenth century.Most importantly, we have demonstrated the importance of using UV to distinguish between hard-paste and bone china sherds, which are otherwise very similar in translucency and texture.This demonstrates clear episodes of waste disposal, relating to clear changes in fabric composition.
Fabric variability as measured by mean and modal thickness of sherds, decreased from soft-paste through hard-paste, with bone china both the thinnest and most consistent of vessel walls.Visual inspection also demonstrated a great degree of visible inclusion matter in both soft-paste and hard-paste sherds, with little visible inclusion matter in bone china.

DEVELOPMENT OF PRODUCTION AND TRENDS IN CONSUMPTION OF WORCESTER PORCELAIN
It is possible to identify several trends in production from the material located from the field survey.As no makers marks were recovered and due to the relatively small sizes of sherds it was not possible to identify the individual factories producing either the finished products or the wasters.The exception is Grainger's which is the only factory thought to be producing whitewares (semi porcelain).Consequently, we are reliant on cross dating the fabric of our sherds with the historical record of porcelain production (Table 5).
Of the range of fabrics produced in Worcester we only have a sample of four types at Oldbury Farm, including the relatively rare hard-paste porcelain.It is documented that Worcester had a short phase of hard-paste production towards the end of the eighteenth century and beginning of the nineteenth century at several factory sites. 96y using transmitted light as a proxy for paste recipes we have been able to assess variations within fabric composition.While a novel approach this is a quick and simple way of determining broad patterns in a sample.We have identified five different colours of transmitted light in all the translucency tests.This probably represents many more variations in fabric at any given time, than simply five recipes.Further work is needed to see how these colours correspond to the chemistry of the fabrics.For example, Owen 97 associates a green transmitted light with the earliest Flight and Barr production of soft-paste porcelain, while Ramsay and Ramsey associate green with early Bow hard-paste. 98Given the variability, it is unlikely that transmitted colour is indicative of either period or place of production.Studies of soft-paste composition by Owen and Jay and for hard-paste porcelain by Godden and Ramsay & Ramsay all indicate variation in paste recipes and these results would tend to agree that paste composition could be highly variable within a particular fabric. 99he narrative history of porcelain manufacturing in Worcester is broadly unilinear, emphasising the development of consistent fabrics, 100 that are hard, 101 high fired, 102 white, 103 and translucent. 104And this is of course a primary concern in manufacturing development, producing a reliable, consistent and affordable product. 105However, such unilinear narratives tend to mask the complexity and richness of the production process, 106 which Barrett hints at in John Flight's diary entries. 107We suggest our results, along with earlier chemical analyses, indicate a much less straightforward trajectory of experimentation and production.These variations in fabric may relate to a number of manufacturing issues, including composition and consistency of raw materials, consistency of equipment, including kilns, as well as human factors such as skill and preference of the relevant factory.
The waster sample covers a period from the mideighteenth century to the end of the nineteenth century.However, for all types of fabrics and both wasters and domestic discard, other than whitewares, the forms are dominated by vessels associated with drinking hot beverages such as tea, coffee and chocolate.Larger forms such as plates and platters are rare.There are also few examples of moulding or embossing amongst the sample, for instance, a single example of fluting in the soft-paste.The emphasis is on simple moulded or thrown forms for painting.This is in contrast to the historical narrative.Godden, for example, suggests that Chamberlain's produces few tea sets in the period 1820 to 1852, about the time of the merger with Barr and Barr, apparently preferring figurines, vases, and other intricate decorative items. 108Grainger & Co. begin trading in 1805, using in the beginning a 'hybrid hard-paste' and 'derivative design', 109 moving into bone china in 1815.So, whilst there is an emphasis on decorative pieces in the antique literature, we see very little evidence for this in the survey sample.
The whiteware sample is larger and heavier than any of the porcelains.This is consistent with the observation that the attraction of porcelain as a fabric is that it can produce both lighter and thinner products, as opposed to the more durable and cheaper to produce whiteware fabric.There is also a preponderance of larger forms in the whiteware collection with more platters and bowls.The appearance of a creamware waster, a lilac biscuit waster and a piece of hard-paste tester/tile hints at the broad range of competitive product experimentation amongst the porcelain factories of Worcester.

FACTORY AND DOMESTIC WASTE DISPOSAL, WORCESTER IN THE 18 TH AND 19 TH CENTURIES
The results of the analysis demonstrate the preponderance of wasters, particularly porcelain wasters in this sample.This supports anecdotal evidence that the porcelain industries in Worcester disposed of significant proportions of their kiln waste on fields within Worcestershire.City wide collection of domestic waste did not begin until the 1930s when the first landfill sites were opened. 110Previously, domestic waste (miskins) was collected and emptied onto Pitchcroft racetrack and then carted away for manure. 111e results of the excavation of an early Worcester Tonquin kiln demonstrated that at least for the earlier period waste appears to have been deposited on site, or perhaps into the river. 112There is an anecdotal account that the earliest kiln was sited next to the Severn for convenience of waste disposal, but this was halted after the city complained that the river channel was being filled in.The Warmstry House factory site was sold to Chamberlain's in 1840, and production moved entirely to Severn Street in Diglis. 113Our results of the identification of softpaste suggests that certainly by 1760-1780, kiln waste was being carted outside the city for disposal.Subsequently, factory waste appears in quantity around the city, usually in fields, including allotments.The large quantity of wasters domestic discard at Oldbury Farm suggests that there were three processes in operation; firstly, the farm's own manuring, secondly manure derived from the city and finally the disposal of factory waste.
The dates of deposition of the waste are best achieved by looking at the fabric of the wasters.Softpaste porcelain is produced from 1751-1792.Bone china is introduced no earlier than 1815 and is continuously produced until 1997, leaving us a potentially long period in which that material could have been disposed of.Finally, hard-paste porcelain is thought to have been produced between 1780-1889, and Harry Frost suggests that the Worcester porcelain factory experimented with it on and off until WWII. 114We can at a minimum conclude that Oldbury Farm was repeatedly used for waste disposal from the city and factories between the mid 18 th centuries through to potentially the mid twentieth century, alongside domestic manuring (Table 5).
The dating of the domestic decorated porcelains is not much more precise than that of the wasters.Softpaste and locally produced hard-paste porcelains have the smallest potential production periods.Though some work on identifying forms and decoration was conducted, the small sherd size and preponderance of decorative styles, particularly printed ones, does not so far improve the accuracy of dating based on fabric type alone.The provenance of the domestic discarded porcelains is largely British, if not of Worcester origins.Though a local origin for particularly soft-paste domestic discard might be argued, it is safer to say that these sherds could have come from the broader area, specifically Bristol, Caughley's in Ironbridge, or Liverpool, all of whom produced significant material in the 18 th and 19 th centuries. 115Only a small number of Asian sherds were recovered, and only a single European sherd, suggesting that for the late 18 th and early 19 th centuries, Worcester was receiving limited materials from abroad.
There are few directly comparable sets of fieldwalking in the literature, but preliminary analysis of two other collections in the Worcester region show a similar low percentage of Asian and European material (Loney and Hoaen nd).We can make a cautious comparison with the analysis of the Duke Street excavations to demonstrate the availability of locally produced vs imported wares in Worcester as opposed to the larger city of London.Pearce found there were five times the amount of Asian porcelain as either continental or English porcelain, 116 despite the presence of several early porcelain factories in Limehouse and Chelsea. 117If you remove the Asian porcelain from Pearce's findings, the proportion of continental and English porcelain compared to the rest of the domestic discard is somewhat different to our results. 118For example, at Worcester soft-paste porcelain is present in high numbers in the domestic discard and assuming it to be broadly contemporary with creamware it represents a third of the late eighteenth century fineware assemblage.It may be suggested that over time that the proportion of porcelain and bone china consumed declines in Worcester as consumer taste switches to pearlwares and whitewares, though more research needs to be conducted to confirm this.We have only eighteen sherds of bone china while broadly contemporary types such as pearlware and whiteware are 129 sherds and 81 sherds respectively.
We have reported on an unusually high quantity of porcelain material at Oldbury Farm compared to elsewhere in the archaeological literature.Excavations at the Royal Worcester Porcelain factory on Severn Street, for example, have produced remarkably little reported whiteware or porcelain materials, 119 and no further analysis as to period or chemistry.We have already reported on the findings of the excavation of the Warmstry House kiln site, 120 which has cumulatively produced a sample analysis of the materials, the rest awaiting funding to complete.Projects such as at the Limehouse porcelain manufactury excavations show that it is possible to produce a thorough porcelain analysis despite recovery of over 1400 sherds. 121lsewhere in the literature, 18 th and especially nineteenth century whitewares are often lumped together, with little porcelain identified or described. 122Projects focussing on nineteenth century working class dwellings often find very little porcelain to report on. 123Porcelain is a rare commodity even within the context of high-status sites such as Manor House in Sheffield, and the Fulbourne Manor House, Cambridgeshire. 124The exception appears to be Pearce's reporting for eighteenth century London, in which Asian, European and English porcelains appear regularly, though representing around 5% of the assemblages overall, 125 and Ratkai's analysis of the Bullring assemblages in Birmingham. 126n summary, porcelain appears to be unevenly analyzed in the archaeological record in the UK, away from factory excavations.The high quantities and wide variety of fabrics with the porcelain material at Oldbury Farm points to a complex history of both factory production and local consumption at a key time in the development of the British global market.The recovery from fieldwalking indicates the importance of looking at the landscape of manufacturing as a strategy in Industrial Archaeology and justifies the intensive quantification and analysis of the resulting materials.

SURVEY DATA AND INDUSTRIAL ARCHAEOLOGY
Field walking is not a traditional avenue of data collection in industrial archaeology, particularly pottery manufacturing. 127Most published projects are based on excavation of in situ buildings and fills, with survey relegated to the identification of sites for future exploration. 128Though there is a perception that survey data lacks context, particularly in understanding factory practice, we have demonstrated the contribution of survey collected data to a greater nuanced understanding of the technical process of the porcelain factories during the 18 th and 19 th centuries.Further, we can record some key differences in waste behaviour between pottery complexes, for instance, the practice of broadly spreading factory waste in Worcester as compared to the construction of dense waster mounds in Staffordshire. 129inally, field survey is an accessible form of archaeological methodology, suitable for community projects as well as University fieldwork projects.It remains amongst the least expensive forms of data collection and working in modern material has the benefits of ease of initial recognition.It also allows a closer personal association, particularly in places such as Worcester which was an iconic employer as well as manufacturer, or Staffordshire, where the industry continues.Projects such as this one, the Tamar Valley and the Fulbourn Manor Estate though not explicitly about pottery waste nonetheless can provide important evidence about the surrounding industries. 130

CONCLUSION
Whilst still in its infancy compared to the art historical and technological study of the porcelain industry in Worcester, archaeology provides a complementary source of information to the record from the literature and from museums.Our analysis of wasters suggests that there was significant experimentation with the recipes for the types of pottery produced in Worcester over time, and that at no point does there appear to have been a standard paste.Later period production in Worcester is best known for its architectural and sculptural pieces, however the assemblages we have examined are dominated by pieces PORCELAIN WASTE AND PORCELAIN PRODUCTION IN WORCESTER associated with drinking.It is only in the whitewares that we see a shift towards more domestic eating ceramics such as plates and platters.Robb refers to Technologies of Enchantment as a way of understanding how certain elements of technology become powerful and influential. 131Using Gell, 132 he notes that the design of key things influences or recall feelings of power.In a world of the rationalist industrial revolution, he provides a more humane and socially centred explanation of how expensively produced and ultimately fragile items such as porcelain and bone china can remain at the top of the prestige hierarchy at the expense of equally lovely and more durable whitewares.That said, it also provides an explanation of how the history and recording of the factory outputs continue to emphasize the unique and highly prized items only available to the very few, whilst the actual outputs included the full range of domestic products, including transfer printed whitewares.This is an important concept, that of the everyday, the mundane, the 95% of stuff that makes up daily life, and the 95% of material found by archaeologists on a regular basis.
Material culture studies in industrial and historical archaeology emphasize the social context of the artefact, the use it is put to and the meaning we ascribe to its ownership, 133 community life, 134 and social and class structure. 135Material Culture studies may also put emphasises on the rare, the identifiable and the powerful, at the expense of the banal mass, 136 for instance, the unidentifiable waster and the discarded teacup.This is particularly so with historical pottery studies, which may still question the appropriateness of field midden for recovering and understanding the manufacturing process.Further, historical periods are not as well prioritised in the planning process in the uk, and consequently, much material is simply discarded.For the 18 th and nineteenth century materials, the inclusion of all elements of the factory product creates a balanced narrative of a highly prestigious industry that none the less created a very large amount of commercially successful products whilst maintaining the enchantment of the process.In this way, we can take the evidence of the less glamorous elements of Worcester porcelain production to provide balance to the art historical narrative of fine art and global influence, without detracting from the latter.scarti domestici e quelli di produzione provenienti dalla citt a furono impiegati come concime/riempimento nei campi di tutta la regione.Sono stati trovati scarti di fornace di prodotti finiti e di biscotti riconducibili alle prime fasi di produzione, cos ı come prodotti realizzati nel tardo XIX e nel XX secolo, unitamente a scarti domestici coevi.Questo nucleo fornisce una serie di spaccati sulla produzione di porcellana e semi-porcellana, permettendoci di identificare le tendenze relativamente ai tipi di impasto, alle forme, al processo produttivo e all'impiego degli oggetti finiti.RESUMEN: La prospecci on arqueol ogica puede producir una muestra longitudinal de la producci on de cer amica de las f abricas de porcelana que operaron en la ciudad de Worcester desde 1751 hasta 2009, ya que los desechos dom esticos e industriales de la ciudad se utilizaron hasta el siglo XX como esti ercol en los campos de la zona.Durante las prospecciones hemos encontrado restos de hornos y desechos de las primeras fases de producci on hasta finales del siglo XIX y XX, junto con desechos dom esticos contempor aneos.El conjunto nos ofrece datos sobre la producci on de porcelana y semiporcelana que nos permiten identificar tendencias en los tipos de pastas, formas, fabricaci on y consumo de la epoca.

TABLE 1 .
Sherd counts by fabric for soft paste, hard paste, bone China, whiteware and kiln furniture.

TABLE 2 .
Colour transmission of translucent sherds, soft paste, hard paste, bone china and kiln furniture.

TABLE 5 .
Approximate date ranges of production of british porcelains and refined earthenwares.