The Value of Historic Window Glass

Transcription

1 The Historic Environment: Policy & Practice ISSN: (Print) (Online) Journal homepage: The Value of Historic Window Glass David Dungworth To cite this article: David Dungworth (2011) The Value of Historic Window Glass, The Historic Environment: Policy & Practice, 2:1, 21-48, DOI: / X To link to this article: W. S. Maney & Son Ltd 2011 Published online: 18 Jul Submit your article to this journal Article views: 1196 View related articles Citing articles: 8 View citing articles Full Terms & Conditions of access and use can be found at Download by: [ ] Date: 08 January 2018, At: 15:20

2 the historic environment, Vol. 2 No. 1, June, 2011, The Value of Historic Window Glass David Dungworth English Heritage, UK This article reviews the nature of window glass employed in historic buildings in England over the past fi ve centuries and presents new results on the nature of that glass. Chemical analysis of available samples has shown that window glassmaking technologies have changed over time. This provides information which can be used to identify the period of the manufacture of particular panes of glass. It is hoped that the use of chemical analysis to date historic window glass in this way will be used to enhance building conservation decisions. keywords conservation of glass, glass blowing, window glass Introduction Windows are a vital part of almost every building; they admit light and provide views of the outside world while keeping out the worst of any inclement weather. 1 Windows also give character to the external appearance of any building; the extent and distribution of the glazing, as well as the size of individual windows and panes, all contribute to a building s personality (Figures 1 and 2). 2 For historic buildings an important contribution is also made by the aesthetic quality of the historic glass. 3 Depending on when (and how) the glass was made it may have a distinct tint and surface texture (Figures 3 and 4). The character of such glass makes a significant contribution to the overall atmosphere of a historic house. 4 Unfortunately the value of historic window glass is not always appreciated, leading to unsympathetic replacements. Recent concerns to increase energy efficiency and reduce carbon emissions may have also contributed to a drive to replace older windows. 5 Managing and conserving historic window glass depends primarily on developing a sense that such glass does form an important part of the historic fabric of a building. 6 Any management of historic assets, however, needs to be made from a position of knowledge. 7 Knowing that a window is old is not really enough. We need to know the age of windows in a particular room; are they original or later replacements, and, if they are later replacements, when were they installed? What we need is a dating technique for window glass. W. S. Maney & Son Ltd 2011 DOI / X

3 22 DAVID DUNGWORTH figure 1 Grevels House, High Street, Chipping Camden, Gloucestershire. Photographed by Nathaniel Lloyd 1927 English Heritage, CC Reproduced by permission of English Heritage NMR A brief history of English window glass manufacture to 1960 Over the last 500 years a variety of different types of glass have commonly been produced in England. The selection of different types of glass has depended on a wide variety of factors, including technology, economics, and aesthetics. Thus, the

4 VALUE OF HISTORIC WINDOW GLASS 23 figure 2 231/233 Kennington Lane, Lambeth, London. Photographed by Eric de Mare English Heritage, AA98/ Reproduced by permission of English Heritage NMR figure 3 Quality of glass and transmitted image, Marygate, York (left modern float glass, right nineteenth-century crown glass). Photographs: author

5 24 DAVID DUNGWORTH figure 4 Quality of glass and transmitted image, King s Manor, York (left image focused on the glass, right image on the exterior). Photographs: author manufacture of colourless lead-based glass has been largely restricted to the production of tablewares (crystal) and only from the late seventeenth century onwards. 8 Although window glass manufacture in England has a long history, documentary sources rarely provide the level of detail required to characterize the nature of glass produced at different periods. Almost all window glass has been, and continues to be, produced from sand, which is composed almost entirely of silica. In order to melt silica, however, it is necessary to add a flux. A flux is any material which will lower the melting temperature of the silica. The most common fluxes used in window glass manufacture are the alkalis sodium and potassium. The more alkali that is added, the lower the melting temperature; however, an excess of alkali will leave the glass prone to weathering. In order to increase the chemical durability of glass a stabilizer is usually added. A wide variety of elements can act as glass stabilizers but the most important have been calcium and magnesium. Early glassmakers were not necessarily aware of the range and concentration of different elements in the raw materials they used, yet in spite of this they developed working practices that were successful. In many cases the presence of elements can be detected in historic window glass which were not deliberately added but which were fortuitously present in one or more raw materials. During the medieval period, glass was produced in a few forested areas, in particular Cannock Chase, Staffordshire 9 and the Weald of Surrey and Sussex. 10 We surmise that much of this glass was made using the ashes of woodland plants (with sand) 11 and that the flat glass was formed using the broad (also known as cylinder or muff) glass technique. 12 A bubble of glass would be blown and elongated to form a cylinder; the ends were removed and the cylinder cut along its length and flattened (Figure 5). The flattening process was undertaken with blocks of wood while the glass

6 VALUE OF HISTORIC WINDOW GLASS 25 figure 5 Broad window glass manufacture: a bubble of glass is elongated to form a cylinder which is then cut open and flattened to form a large sheet. Image Pilkington Group Limited. Reproduced with permission

7 26 DAVID DUNGWORTH was still sufficiently soft, but this left a more-or-less rough surface on the glass. At some stage a second forming method was developed which came to be known as crown glass. 13 The glass was blown as a bubble which was then opened and spun round until it formed a disc of glass (Figure 6). Crown glass did not have to be flattened and so had a superior surface finish compared to broad glass. The drawback with crown glass, however, was the central swelling or bull s eye which prevented the production of large panes of glass. Almost all glass of this period (whether broad or crown) was made using plant ashes and so has a natural tint or colour which comes primarily from the traces of iron present in the raw materials used. Relatively little plain (that is unpainted) medieval glass survives in buildings. It has long been suggested that the English glass industry went into decline in the sixteenth century and was only revived by the arrival of glassmakers from France from 1567 onwards. 14 This suggestion owes much to the oft-quoted Charnock s Breviary, which contains the lines: As for glassmakers, they be scant in this land. 15 Unfortunately, it is rarely acknowledged that Charnock s Breviary was, in poetical form, a series of instructions to the (al)chemist, including how to obtain the rather arcane glassware necessary for this profession. 16 While it may have been difficult to obtain alembics and other specialized glassware, archaeological evidence indicates that more utilitarian production (including window glass) took place in England in the mid sixteenth century. 17 The late-sixteenth-century influx of glassworkers (mostly from France) is well known from documentary sources; 18 however, it has not been clear what effects this had on the nature of the window glass produced. The glassmakers from Lorraine are traditionally associated with the production of broad glass, while those from Normandy are usually thought of as crown glass manufacturers. 19 Window glass continued to be manufactured in the Weald and Cannock Chase until the beginning of the seventeenth century. The early seventeenth century saw a major change in the glass industry as a whole when the use of wood as a fuel was banned; as a result the industry relocated to the coalfields. 20 The effects of the relocation of the glass industry on the nature of the window glass produced are uncertain. Documentary sources are equivocal on the types of window glass made during this period. This is not helped by the fact that the term broad glass (or its cognates, cylinder or muff glass ) was widely used to mean any sheet glass suitable for windows. 21 Although crown glass was described as first made in England in there is abundant archaeological evidence for its manufacture in the early sixteenth century. 23 The identification of broad or crown glass in extant seventeenth-century windows is made difficult by the fashion for small diamond panes or quarries of a rather small size. The quarries were held together with lead strips (came) and each window would often contain quarries. The small size of the panes, the lead came, and the greenish quality of the glass all contributed to a window which would provide much less light than a modern window and which could give only a rather distorted view of the outside world (Figure 4). The late seventeenth century saw the introduction of the sash window 24 which quickly made use of larger panes of glass (Figure 3). The sash window became increasingly popular during the eighteenth century and was used to allow more light into a room and a more useful view of the outside world. During the eighteenth century the

8 VALUE OF HISTORIC WINDOW GLASS 27 figure 6 Crown window glass manufacture: a bubble of glass is spun to form a circular sheet which is cut to form small panes. Image Pilkington Group Limited. Reproduced with permission

9 28 DAVID DUNGWORTH production of broad glass declined and crown glass came to dominate the market. While crown glass may have been an aesthetically superior glass, the system of taxation on window glass production worked in its favour. 25 Window glass manufac turers were taxed on the weight of materials they used to make glass but derived profit from the area of the glass they made and sold. As crown glass was routinely thinner than broad glass there was more profit to be made in it than in broad glass. It is not clear whether broad or crown glass manufacturers employed different types of raw materials and whether this may have also affected the colour and transparency of the glass. Writing in the early eighteenth century, Neve describes several different types of glass then made in England, including several grades of crown glass. 26 For the later eighteenth century there are a number of documentary sources which mention the use of kelp (seaweed ash) as a flux for crown window glass manufacture. 27 A new technique for producing sheet glass suitable for windows was developed in France in the late seventeenth century. The molten glass would be cast onto a metal table and rolled flat. Once cooled, the glass was laboriously ground and polished to produce two perfectly flat and polished surfaces. 28 The long polishing process made plate glass very expensive compared to crown or broad glass (Neve indicates that in the early eighteenth century plate glass was five to twenty times more expensive than ordinary sheet glass) 29 and so its use was largely restricted to mirrors. Plate glass was also used for the windows of coaches but was only used for domestic windows in the houses of the fabulously wealthy. Plate glass production was introduced to England in the late seventeenth century but production seems to have been short lived. 30 The technology was reintroduced in the late eighteenth century and continued into the twentieth century. 31 Very little is known about the sorts of ingredients used in plate glass manufacture but the apparent concern with transparency is likely to have encouraged the use of more select materials than those employed by crown or broad glass manufacturers. The nineteenth century saw many significant developments in the English glass industry, including the nature of the raw materials and forming techniques. For centuries all window glass had been made using the ashes of selected plants, but by the eighteenth century it was increasingly appreciated that the active ingredients in plant ashes were the alkalis sodium and potassium. 32 It was recognized that common salt was rich in sodium but that salt would not react with sand to form glass. Towards the end of the eighteenth century Nicolas Leblanc developed a technique for converting salt (sodium chloride) into sodium carbonate. 33 The technology was imported to England in the late 1820s and by the late 1830s Leblanc soda had largely displaced plant-based alkalis. 34 The 1830s saw Chance Brothers introduce what they called improved cylinder glass. 35 This glass was made using some variation on the traditional broad glass technique although it is not immediately apparent what the improvements were. Chances made use of skilled labour brought from continental Europe and appear to have been able to supply sheets which were larger than those made by traditional broad glass producers. 36 It is likely that, like so many industrial developments of this time, the improvement was actually embodied in the manual skill of the workers and so was not easily captured in a written text. Contemporary observers contrasted the quality of traditional, English broad glass, coarse in texture, of a wavy wrinkled surface, and very cheap, 37 with the new improved cylinder glass. The quality of improved cylinder

10 VALUE OF HISTORIC WINDOW GLASS 29 glass probably benefited from the recent availability of Leblanc soda. 38 Its production expanded rapidly and by 1844 accounted for almost a quarter of the flat glass made in England. 39 Crown glass production continued to decline 40 and by the early twentieth century the technique was described as nearly obsolete. 41 Further developments at the beginning of the twentieth century resulted in the application of steam power and compressed air which saw the near complete mechanization of the industry. 42 The Lubbers process (1910) used compressed air and hoists to draw cylinders up to 8 m in length. 43 The Fourcault process (developed just before the First World War but not widely used in England until the end of the 1920s) drew a sheet of glass by pushing a former into the molten glass. 44 Variations on such sheetdrawing techniques were developed during the twentieth century but all suffered to a greater or lesser extent from surface waviness that could distort images viewed through the glass. The search for a flat glass forming technique which would allow the cheap production of glass with a perfectly polished surface finally succeeded in the late 1950s when Pilkingtons developed the float technique. 45 Glass was poured directly from the tank onto a bath of molten tin. This allowed the glass to form plane and parallel surfaces and gave the glass a fire-polished finish which did not require abrasive polishing. Almost all glass is now produced using the float process. The history of window glass manufacture provides little information that can be used directly to determine the age of manufacture of a particular window. Leaving aside the difficulty of actually identifying a particular production process from a simple visual assessment of a pane of glass, both broad and crown glass appear to have been produced for over five centuries and the later developments are compressed into less than one century. The texture and tint of historic window glass give it its essential aesthetic quality but translating this into a date is beyond the skills of most people. The chemical composition of glass as an indicator of the date of manufacture The historical sources provide many indications that window glass manufacturing technologies have changed radically over the last five centuries. While few of these are obvious to the naked eye, it is now clear that many of these changes can be detected in the chemical composition of the glass. Hundreds of samples of historic window glass have been analysed in order to characterize chronological changes in composition. There are between five and eight periods and each can be characterized by the chemical composition of the glass; for each period the raw materials, recipes, and technologies used gave rise to chemically distinct glass. All of the analysed samples have been collected from contexts (archaeological or architectural) which have been able to provide some independent dating information. Information on the composition of recently produced glass is also available from twentieth-century literature. 46 Sampling problems and opportunities Any project which hopes to characterize chronological changes in window glass production through chemical analysis faces significant sampling problems. Window

11 30 DAVID DUNGWORTH glass is easily broken and replaced. When faced with a historic building, how certain can we be that any of the extant glass is original? Many of the samples analysed for this research project were recovered from archaeological rather than architectural contexts. This approach has several factors in its favour: the archaeological context provided some indication of date and the glass was already broken, making ethical considerations associated with sampling largely irrelevant. The window glass from Silkstone provides a good example of the use of samples from a well-dated archaeological context. 47 The late-seventeenthcentury glassworks at Silkstone did not produce any window glass 48 but glass was found in overlying contexts which were dated to the first three decades of the eighteenth century by association with clay pipes of known age. Naturally the dating of such a context only really applies to the date of deposition and it is possible that the window glass was produced before the eighteenth century. Glass from the excavation of historic buildings may also be assigned to particular periods depending on the history of the building. The excavations at Basing House, Hampshire yielded window glass which is unlikely to pre-date the construction of the house in 1530 and certainly cannot post-date the demolition of the house at the end of the Civil War (1645). 49 On occasion the relatively short life of a building provided even more precise dating. The samples of glass from the excavations on the site of Basing Grange were valuable as the Grange was built in 1677 but was demolished by Even closer dating has been provided in the case of the First World War army training camp at Brocton which was built in 1915 and demolished in The search for samples of historic window glass has often followed unusual paths. Samples of window glass from the Crystal Palace would have been extremely useful. The Crystal Palace was built for the Great Exhibition in 1851 and we know who supplied the glass (Chance Brothers). Unfortunately the Crystal Palace was moved to south London and destroyed in a fire in However, the architect (Joseph Paxton) had in 1837 designed a conservatory at Chatsworth for the Duke of Devonshire, again using Chance s glass. 52 This glasshouse was demolished (using dynamite!) in the 1920s and the site is now occupied by a maze. Enquiries to the gardens department at Chatsworth resulted in the donation of 184 samples of window glass recovered from the site of the conservatory. 53 Important evidence for the nature of window glass manufactured at different periods has also been obtained from production sites. Several medieval and early post-medieval glasshouses in both Staffordshire and the Weald have been excavated and analyses of the production waste undertaken. 54 The archaeological investigation of later glasshouses has provided samples which can be assigned to particular historical periods. Documentary evidence indicates that a glasshouse was in operation at St Thomas Street in Bristol between 1712 and 1774 and that at least one of its products was crown window glass. 55 The analysis of glassworking debris from this site has allowed the characterization of crown window glass of the mid eighteenth century. 56 The selection of window glass samples from a building faces many problems but also presents opportunities not available with material from archaeological contexts. The collection of samples of window glass from Shaw House, Newbury (undertaken during the restoration funded by the Heritage Lottery Fund) included samples from

12 VALUE OF HISTORIC WINDOW GLASS 31 a window which had gone out of use and been blocked at the beginning of the eighteenth century. 57 Therefore, any glass within the blocked window must have been produced before c Similarly, the architectural history of the house showed that the eastern side of the house was remodelled at the beginning of the eighteenth century and all of the windows were replaced. All of the window glass in these windows must therefore have been produced after c The search for samples has also taken advantage of opportunities on my own doorstep. I live in a street which was built in 1894 and The glass in these houses represents a palimpsest of the original glazing and a host of later piecemeal replacements. The analysis of samples of window glass from four separate properties (mostly undertaken when owners replaced older windows) has identified a glass which recurs in all four; this is interpreted as the original glass. 58 A full list of all of the sources, sites, and buildings is provided in Table 1 (see also Figure 7). Results The samples have all been analysed using laboratory-based techniques in order to obtain high-quality data on the chemical composition of the glass. Details on the methods used can be found in each of the individual site reports (Table 1) but it is worth stressing that high-quality data can be obtained from samples less than 0.1 g in weight. The results of the analysis of over 800 samples (in combination with data published by others) are summarized in Figures 8 15 and Table 2. Each figure shows the concentration of a single element in the glass; each point represents a single building or a single archaeological or architectural context. The vertical axis indicates the concentration of the element in question and the horizontal axis corresponds to the date. The uncertainty over date is represented by the horizontal error bars. A small number of samples with very poor dating information have been omitted completely. For each data point the vertical error bars (variation in chemical composition) were generally low and have been omitted for clarity. The chronological changes in the composition of the glass manufactured in England seen in the figures allow the identification of five major phases with three possible minor phases (Table 2). Within each phase, the chemical composition of manufactured glass displays limited variability; between each phase one or more elements in the glass changes significantly. In many cases the identification of the boundary between one phase and the next can be correlated with historical information. Phase 1: forest glass (before c. 1567) The first phase of historic window glass manufacture covers the medieval period up to the arrival of French glassmakers from The glass of this period is characterized by high levels of potassium and magnesium. The earliest known glass manufacturing sites within this tradition are Blunden s Wood 90 and Little Birches 91 which have yielded archaeomagnetic dates in the fourteenth century. Documentary sources hint that glass production may extend back into the thirteenth century. 92 The glass of this period was made using sand and ashes from woodland plants. While documentary sources appear to recommend the use of a wide range of plants, 93 the chemical composition of most English medieval glass is most consistent with the use

13 32 DAVID DUNGWORTH TABLE 1 SOURCES OF INFORMATION ON THE CHEMICAL COMPOSITION OF HISTORIC WINDOW GLASS Site County Date Analyses Comments Basing Grange 59 Hampshire Samples recovered during archaeological excavation Basing House 60 Hampshire Samples recovered during archaeological excavation Beverley Minster 61 Yorkshire Samples collected during conservation of extant glass Blunden s Wood 62 Surrey Samples of window glass and working waste from archaeological excavation of glasshouse Brocton 63 Staffordshire Samples of window glass recovered during archaeological excavation Chastleton House 64 Oxfordshire c Samples recovered during archaeological excavation Chatsworth Conservatory 65 Derbyshire Samples recovered during gardening Cheese Lane 66 Bristol Samples of working waste from archaeological excavation of glasshouse Flint Lodge 67 Portsmouth Samples recovered during building renovation Fort Cumberland 68 Portsmouth Samples recovered during building renovation Highland House 69 Portsmouth Samples recovered during building demolition Horsebridge 70 Sussex Samples of window glass and working waste from archaeological excavation of glasshouse Idehurst North 71 Sussex Samples of working waste from archaeological excavation of glasshouse Idehurst South 72 Sussex Samples of working waste from archaeological excavation of glasshouse June Hill 73 Surrey Samples of working waste from archaeological excavation of glasshouse Knightons 74 Sussex Samples of window glass and working waste from archaeological excavation of glasshouse Little Birches 75 Staffordshire Samples of window glass and working waste from archaeological excavation of glasshouse Margam 76 Glamorgan Samples collected during conservation of extant glass Nailsea 77 Avon Samples of working waste from archaeological excavation of glasshouse Newent 78 Gloucestershire Samples of window glass and working waste from archaeological excavation of glasshouse Palace House Mansion 79 Suffolk Samples collected during building renovation

14 VALUE OF HISTORIC WINDOW GLASS 33 TABLE 1 CONTINUED Site County Date Analyses Comments Portwall Lane 80 Bristol Samples of window glass and working waste from archaeological excavation of glasshouse Shaw House 81 Berkshire Samples collected during building renovation Sidney Wood 82 Sussex Samples of window glass and working waste from archaeological excavation of glasshouse Silkstone 83 Yorkshire Samples recovered during archaeological excavation St Thomas Street 84 Bristol Samples of window glass and working waste from archaeological excavation of glasshouse Tanland 85 Sussex Samples of working waste from archaeological excavation of glasshouse Tower of London 86 London Samples collected during building renovation Welch Road 87 Portsmouth Samples collected during building renovation Wentworth 88 Yorkshire Samples collected during building renovation York Minster 89 Yorkshire Samples collected during conservation of extant glass of bracken as a source of alkali. 94 There are some consistent differences in chemical composition between English medieval glass and that produced in either France or Germany. 95 The differences probably reflect different glassmaking practices, especially in the use of plant ashes. Unfortunately, the glass of this period is not particularly stable and it is rare for it to have survived in contemporary windows. This glass type was also used to manufacture a wide range of other goods including drinking vessels. Phase 2: high-lime, low-alkali glass (c to c. 1700) The second phase appears to start in 1567 with the arrival of French glassmakers and ends c The glass of this period is characterized by a high calcium content and relatively low levels of sodium and potassium. It has long been argued that the French glassmakers who came to England from 1567 onwards brought with them a number of technical developments. 96 It is now clear that the glass made by the French glassmakers was of a new type. 97 This new high-lime, low-alkali (HLLA) glass appears to have been produced using different types of plant ash as the source of the alkali. The high lime content of this glass strongly suggests the use of wood ash. Contemporary sources mention the use of a wide variety of plant ashes including trees (oak, birch, fir etc.) and a variety of other plants, such as thistle, hops, brambles, and

15 34 DAVID DUNGWORTH figure 7 Map of England and Wales showing the location of the sites and buildings investigated. Source: English Heritage, with permission even tobacco. 98 HLLA glass was used to manufacture a wide range of goods including tablewares and bottles. 99 The possible subdivision of HLLA glass into two phases is based on a rather subtle difference in manganese content. The sixteenth-century HLLA glass contains higher levels of manganese than the seventeenth-century glass. 100

16 VALUE OF HISTORIC WINDOW GLASS 35 figure 8 Sodium concentration in historic window glass. Source: English Heritage, with permission figure 9 Magnesium concentration in historic window glass. Source: English Heritage, with permission

17 36 DAVID DUNGWORTH figure 10 Phosphorus concentration in historic window glass. Source: English Heritage, with permission figure 11 Potassium concentration in historic window glass. Source: English Heritage, with permission

18 VALUE OF HISTORIC WINDOW GLASS 37 figure 12 Calcium concentration in historic window glass. Source: English Heritage, with permission figure 13 Manganese concentration in historic window glass. Source: English Heritage, with permission

19 38 DAVID DUNGWORTH figure 14 Arsenic concentration in historic window glass. Source: English Heritage, with permission figure 15 Strontium concentration in historic window glass. Source: English Heritage, with permission

20 VALUE OF HISTORIC WINDOW GLASS 39 TABLE 2 AVERAGE COMPOSITIONS OF HISTORIC WINDOW GLASS Phase 1 2a 2b 3 4a 4b 5a 5b Start c c c c c c c End c c c c c c c Na 2 O 2.5P P P P P P P P0.4 MgO 7.3P P P P P P P P0.2 Al2O 3 1.6P P P P P P P P0.2 SiO P P P P P P P P0.5 SO 3 0.3P P P P P P P P0.1 Cl 0.4P P P P P0.1 <0.1 <0.1 <0.1 P 2 O 5 3.2P P P P0.2 <0.2 <0.2 <0.2 <0.2 K 2 O 11.4P P P P P P P P0.1 CaO 15.3P P P P P P P P0.6 MnO 1.26P P P0.20 <0.10 <0.10 <0.10 <0.10 <0.10 Fe 2 O P P P P P P P P0.01 As 2 O 3 <0.20 <0.20 <0.20 < P0.16 <0.20 <0.20 <0.20 SrO 0.07P P P P P P P P0.01

21 40 DAVID DUNGWORTH Phase 3: kelp glass (c to c. 1835) Phase 3 sees wide-ranging changes in the manufacture and use of window glass. Phase 3 glass is a mixed alkali glass; it contains substantial proportions of both sodium and potassium. This glass still contains enough phosphorus to indicate that it was made using a plant ash. The glass contains a relatively high proportion of strontium suggesting the use of a marine plant. 101 The analysis of strontium isotopes in this glass has confirmed the use of seaweed. 102 Historical sources corroborate the use of seaweed (usually referred to as kelp) from at least the middle of the eighteenth century. 103 The chemical analysis of window glass shows that it was used in windows from at least the beginning of the eighteenth century. 104 There are some signs that the earliest use of kelp in glassmaking belongs to the later seventeenth century; however, its use appears to have been restricted initially to tablewares. 105 Phase 4: Leblanc soda glass (c to c. 1930) The beginning of phase 4 can be placed c with the introduction of the Leblanc process for the manufacture of sodium carbonate from common salt. Window glass made using Leblanc soda is a soda-lime-silica glass which contains very low levels of impurities compared with any of the earlier plant-ash glasses. In particular, glass made after c contains little or no phosphorus. 106 The glass of this period rarely has any significant colour, probably due to the use of better quality sand. The glass used at Margam in 1834 included some phase 3 kelp glass, 107 while the Chance s glass produced for the conservatory at Chatsworth between 1837 and 1840 was exclusively made using Leblanc soda. 108 Documentary sources indicate that Leblanc soda was first manufactured in England in the late 1820s 109 but it did not make any serious impact on the glass industry until the 1830s. 110 There is the possibility that phase 4 can be divided into two subphases (4a and 4b) based on the nature of the refining agent used in the glass. Refining agents are used to remove bubbles from glass; this is usually achieved paradoxically by using a material which produces glass bubbles. The introduction of arsenic, saltpetre, and even potatoes 111 will evolve gas bubbles which, being large, will rise quickly and tend to gather small bubbles as they rise. Glass made for the first few decades after the introduction of Leblanc soda usually contains significant concentrations of arsenic (phase 4a). The use of arsenic as a refining agent is mentioned in several contemporary documentary sources. 112 Around the 1870s, the arsenic disappears from window glass but there is a small increase in the potassium content of the glass (phase 4b). Documentary sources mention the use of saltpetre (potassium nitrate) as a refining agent. 113 Phase 5: mechanized glass (from c. 1930) The early twentieth century saw the mechanization of window glass production. By the late 1920s sheet drawing techniques, especially the Fourcault process, had been introduced to England. The scale of this process was such that the glass produced had to be kept hot for longer periods compared to the earlier manual techniques and this resulted in a tendency for the glass to devitrify. 114 The formation of small crystals in the glass obviously detracted from its transparency. Various attempts were made to alter the composition of the glass to avoid this problem but the lasting solution was found to be the replacement of a proportion of the calcium in the glass with

22 VALUE OF HISTORIC WINDOW GLASS 41 magnesium. 115 Window glass produced after c contains wt% MgO. 116 There is the possibility that phase 5 can be divided into two subphases based on the MgO concentration. The MgO concentration shows a strong bimodal distribution with one group containing wt% MgO and the other with wt% MgO. The limited dating evidence for glass of this period suggests that the former group is earlier than the latter. It is possible that the higher MgO group belongs to the period after the introduction of the float process. Discussion of the chemical analysis of historic window glass Chemical analysis offers the potential to date the manufacture of individual panes of glass. A consideration of the concentration of relatively few elements will allow a particular pane of glass to be quickly assigned to one of the five major phases. The phosphorus concentration will indicate whether or not the glass was made using plant ashes (phases 1 3) or synthetic soda (phases 4 and 5). Glass of phase 3 is easily distinguished by its strontium content. Glasses of phases 1 and 2 can be separated by examining the proportions of magnesium, potassium, and calcium. Glasses of phases 4 and 5 can be distinguished by their magnesium and calcium content. The concentration of other elements may allow further refinements of the estimated age of manufacture. It has been presumed that there has been little or no reuse of window glass; for example, if construction of a building began in 1587, it is assumed that none of the glass will have been made before this date. This assumption seems to be sound; however, there are a few exceptions which need to be considered. Before the late sixteenth century domestic window glass was usually regarded as a luxury item and often a portable one at that. Among the landed classes with several houses, it was not unknown for the window glass to be removed and brought with them when they moved from one house to another. 117 Windows could be specified in a will separately from the house until the end of the sixteenth century. 118 By the seventeenth century, plain window glass had become a relatively inexpensive commodity 119 and there is little evidence for the reuse of plain glass. The reuse of decorated window glass is known to have occurred on occasion. The glass in York Minster includes Romanesque glass from an earlier building on the site as well as fourteenth-century window glass from New College Oxford which was inserted into a fifteenth-century part of the Minster in the eighteenth century. 120 Many secular houses of the rich have incorporated reused fragments of ecclesiastical stained glass, in particular during the Gothic revival during the nineteenth century. It is known from late medieval and early post-medieval documents that window glass was obtained from English producers as well as those in mainland Europe. 121 A fifteenth-century contract for the glazing of the Countess of Warwick s chapel specified that the glazier use no glass of England, but glass from beyond seas. 122 Various medieval glazing contracts specify that the white glass (that is, glass which was not deliberately coloured) be obtained from England. 123 A recent examination of glassworking waste and data from the analysis of glass from English cathedrals suggests that 40 per cent of the glass was English while the remainder was obtained from continental Europe. 124 It is likely that importation became less significant from

23 42 DAVID DUNGWORTH the late sixteenth century due to the growth of the English glass industry. 125 Writing in the early eighteenth century, Neve described Newcastle glass as the most used in England. 126 He also mentions German, Dutch, and French glass, although most is described as not much used now in England. 127 By the middle of the nineteenth century English window glass manufacturers began to experience competition from producers in mainland Europe and North America, although initially this was almost entirely restricted to export markets. The second half of the nineteenth century saw significant importation of Belgian window glass; 128 however, at this relatively late date it is unlikely that there were significant differences in the recipes used by English and Belgian manufacturers. Although Figures 8 15 give the impression that the change from one glass type to another occurred almost instantly (and that this was universal throughout England) the situation was inevitably more complex. Many technological inventions can be turned to economic advantage and efforts are often made to keep the details of such innovations from competitors. The modes of production in the glass industry, however, have often conspired to prevent such industrial secrecy. Glassworkers have often been a relatively powerful workforce. The French glassmakers who came to England in the late sixteenth century were often referred to as gentleman glassmakers. They were engaged in manual labour but were the owners of their glasshouses and they maintained social relations with other immigrant glassmaking families in other parts of England. The movement of (especially young) glassmakers between one area and another promoted the rapid diffusion of technological knowledge. The change from forest glass to HLLA appears to have taken perhaps a generation to be adopted throughout the country. 129 This seems to have been to a large extent because the French glassmakers were reluctant to share their technological knowledge with the indigenous glassworkers. 130 The change from HLLA glass to seaweed glass appears to be complete within a decade or two at the most. 131 The change from seaweed glass to glass made using Leblanc soda appears to have been completed in less than a decade. 132 Subsequent technological developments in the glass industry were increasingly protected through patents but the economic advantages were such that competitors would either go out of business or take up the new technology under licence. 133 Future research The methodology employed for the chemical analysis of historic window glass for the research presented here is a destructive one. Samples of broken glass as small as 1 mm 3 have been analysed using laboratory techniques in order to obtained high-quality data. This approach works well in situations where historic window glass has been brought to a conservation studio for repairs or restoration. The size of the samples required is small enough for the remaining pane to be replaced after conservation treatment with no effective loss to its aesthetic qualities (the site of the sample being obscured by lead came or glazing bars). Sampling (and chemical analysis) under these circumstances can be used to inform conservation procedures; for example it can be used to identify earlier repairs, it can provide the data needed to specify glass for sympathetic replacement, and it can provide information about

24 VALUE OF HISTORIC WINDOW GLASS 43 glass deterioration. Information about glass deterioration can inform broader conservation decisions such as isothermal glazing. Isothermal glazing comprises the installation of new plain glazing in the position of the old (usually stained) glass and the remounting of the old glass with a gap between the two. This gap allows the circulation of air and so prevents condensation on the interior surface of the old glass. 134 The methodology employed to date in this project, however, cannot be used directly with in situ window glass which is neither broken nor the subject of any conservation intervention. Nevertheless, conservation officers, historic buildings inspectors, and the curators of historic buildings often want to know the age of extant glass precisely so they can make informed decisions about the conservation of such glazing. Recent developments with portable X-ray fluorescence (XRF) spectrometers, however, offer the potential to carry out non-destructive chemical analysis of window glass without removing the glass from the window. Portable instruments are now available which are light enough to easily carry out in situ chemical analysis in historic buildings. In addition these instruments are capable of providing information about the concentration of light elements (such as magnesium and even sodium). The application of a portable XRF instrument for the in situ chemical analysis and dating of historic window glass will be the subject of a future publication. Conclusions Window glass can be among the most vulnerable components of historic buildings. Window glass is easily broken and the need to keep buildings weatherproof may lead to replacement with readily available modern glass. The nature of the original glazing, however, makes a significant contribution to the overall character of a historic building. The survival and retention of original glazing can considerably enhance the heritage value of a historic building. There are two aspects of original glazing which define its aesthetic qualities: the surface finish of the glass and its colour or tint. The most stark difference in surface finish seen in most buildings is between almost any pre-1960 glass and modern float glass. Float glass usually has a perfectly plane surface which distorts neither reflected nor transmitted images. In addition, this glass is effectively free from any bubbles and is of uniform thickness. The production techniques used for the manufacture of almost all historic window glass leave the glass with a non-plane surface, variations in thickness, and bubbles and other defects within the glass itself. One type of glass manufactured before the development of the float process was capable of producing glass with aesthetic qualities which approach float glass: plate glass. The polishing process employed in the manufacture of plate glass left it with a perfectly flat surface which may be impossible to distinguish from float glass. Other early flat-glass forming techniques yielded glass which could have very similar aesthetic qualities, such as improved cylinder, Lubbers, and Fourcault glass. Most people experience window glass as a fully transparent material and, since the early nineteenth century, most window glass has been manufactured from materials which gave virtually no tint to the glass. Before the industrial production of

25 44 DAVID DUNGWORTH synthetic soda, however, almost all glass had at least a slight tint which gave a distinct character to transmitted light and images viewed through the glass. The exact tint of the glass could vary considerably depending on the composition of the glass and the way in which the original glass had been melted. The most significant element in glass which affects its colour is iron. Before the use of synthetic soda the sands employed by window glassmakers all appear to have contained enough iron to give the glass at least some colour. Glass made with kelp has only a slight blue-green tint which had little effect, while medieval forest glass had a slightly stronger colour but usually more green than blue-green. The HLLA glass of the late sixteenth and seventeenth centuries, however, had a much stronger blue-green colour which would significantly affect the quality of transmitted light and images. It is hoped that the research presented in this article will encourage a more sympathetic approach to historic windows and their value. The results obtained to date also provide some interesting ideas about how windows and their functions have changed over time. The iron content of glass manufactured prior to the eighteenth century is often rather high, giving rise to a distinctly green or blue-green glass. The techniques employed to form glass of this period (broad and crown) both left significant variations in thickness. While this early glass acted as a barrier to keep out rain and wind, and could be used to let light into a room, it could only provide rather obscured views of the outside world. The nature of early window glass suggests that people were not concerned with looking through a window. This perception of the nature and function of window glass accords with the medieval church tradition where it was used to keep the weather outside and let light in, but where a view of the exterior could be an unwelcome distraction. It can be argued that there was little or no expectation that windows were supposed to provide a view. The end of the seventeenth century saw a change to kelp glass with a lower iron content and greater transparency. In addition other aspects of glazing technology and fashion changed; traditional casement windows with lots of small panes of glass held together with lead came were increasingly replaced by sash windows with wooden glazing bars which held much larger panes of glass. Early-eighteenth-century windows are the first which allow those inside a building to really look at the world outside. Notes 1 Hentie Louw, The Development of the Window, in Windows: History, Repair and Conservation, ed. by Michael Tutton and Elizabeth Hirst (Donhead St Mary: Donhead, 2007), pp Jill Kerr, The Repair and Maintenance of Historic Glass, in Practical Building Conservation. English Heritage Technical Handbook. Volume 5. Wood, Glass and Resins, ed. by John Ashurst and Nicola Ashurst (Aldershot: Gower, 1988), pp Environment and Heritage Service (NI), Conservation of Historic Glass, Technical Note 14 (Belfast: Environment and Heritage Service (NI), 2006). 4 Historic Scotland, Managing Change in the Historic Environment: Windows, consultation draft (Edinburgh: Historic Scotland, 2009). 5 Chris Wood, Bill Bordass and Paul Baker, Research into the Thermal Performance of Traditional Windows: Timber Sash Windows (London: English Heritage, 2009). 6 English Heritage, Making the Past Part of Our Future (London: English Heritage, 2005). 7 Kate Clark, Informed Conservation: Understanding Historic Buildings and Their Landscapes for Conservation (London: English Heritage, 2001). 8 David Dungworth and Colin Brain, Late- Seventeenth-Century Crystal Glass: An Analytical Investigation, Journal of Glass Studies, 51 (2009), Christopher Welch, Glass-Making in Wolseley, Staffordshire, Post-Medieval Archaeology, 31 (1997), 1 60.

26 VALUE OF HISTORIC WINDOW GLASS George Hugh Kenyon, The Glass Industry of the Weald (Leicester: Leicester University Press, 1967). 11 James Smedley and Caroline Jackson, Medieval and Post-Medieval Glass Technology: A Review of Bracken in Glassmaking, Glass Technology, 43 (2002), David Martlew, History and Development of Glass, in Tutton and Hirst, pp Michael Forsyth, Window Glass, in Materials and Skills for Historic Building Conservation, ed. by Michael Forsyth (Chichester: Wiley-Blackwell, 2007), pp E. Wyndham Hulme, English Glass-Making in the Sixteenth and Seventeenth Centuries, Notes and Queries, Series 8 Volume 6 (1894), Ruth Hurst Vose, Glass (London: Collins, 1980); Hugh Willmott, A History of English Glassmaking AD (Stroud: Tempus, 2005). 16 Stephen Moorhouse, Medieval Distilling-Apparatus of Glass and Pottery, Medieval Archaeology, 16 (1972), Eric Wood, A Sixteenth Century Glasshouse at Knightons, Alfold, Surrey, Surrey Archaeological Collection, 73 (1982), Eleanor Godfrey, The Development of English Glassmaking (Chapel Hill, NC: University of North Carolina Press, 1975). 19 Ibid. 20 David Crossley, The Performance of the Glass Indu stry in Sixteenth-Century England, The Economic History Review, 25 (1972), ; David Crossley, The Archaeology of the Coal-Fuelled Glass Industry in Britain, The Archaeological Journal, 160 (2003), Godfrey, p. 5, n Richard Neve, The City and Country Purchaser and Builder s Dictionary (Newton Abbot: David and Charles, 1969 facsimile reprint, originally published in 1726), p Wood, Knightons. 24 Hentie Louw, The Origin of the Sash Window, Architectural History, 26 (1983), Theodore Cardwell Barker, The Glassmakers. Pilkington: The Rise of an International Company (London: Weidenfeld and Nicolson, 1977), p Neve, pp Richard Watson, Chemical Essays (London: Dodsley, Cadell, Evans, Merrill and Fletcher, 1782). 28 Ronald Walter Douglas and Susan Frank, A History of Glassmaking (Henley-on-Thames: Foulis, 1972), pp Neve, pp Douglas and Frank, p Raymond McGrath and Albert Frost, Architectural Glass (London: Architectural Press, 1937). 32 Watson, p Archibald Clow and Nan Clow, The Chemical Revolution (London: Batchworth, 1952). 34 Sheridan Muspratt, Chemistry: Theoretical, Practical and Analytical (Glasgow: Mackenzie, 1860), p. 322; M. Gray, The Kelp Industry in the Highlands and Islands, Economic History Review, 4 (1951), Barker, pp John Chance, A History of the Firm of Chance Brothers & Co. Glass and Alkali Manufacturers (privately printed, 1919), p. 7; McGrath and Frost, p Andrew Ure, A Dictionary of Arts, Manufactures and Mines, 3rd edn (New York: Appleton, 1844), p Chance, p Barker. 40 Harry Powell, Henry Chance and H. G. Harris, The Principles of Glass-Making (London: George Bell, 1883), p Walter Rosenhain, Glass Manufacture (London: Constable, 1919), p Michael Cable, The Development of Flat Glass Manufacturing Processes, Transactions of the Newcomen Society, 74 (2004), McGrath and Frost. 44 Cable. 45 Lionel Alexander Bethune Pilkington, Review Lecture: The Float Glass Process, Proceedings of the Royal Society of London, Series A, Mathematical and Physical Sciences, 314 (1969), Antonin Smrček, Evolution of the Compositions of Commercial Glasses 1830 to Part I: Flat Glass, Glass Science Technology, 78 (2005), David Dungworth, Composition of Early Eighteenth Century Window Glass from Silkstone, Yorkshire, Research Department Report 18/2006 (Portsmouth: English Heritage, 2006). 48 David Dungworth and Tom Cromwell, Glass and Pottery Manufacture at Silkstone, Yorkshire, Post-Medieval Archaeology, 40 (2006), David Dungworth, Basing House, Old Basing, Hampshire: Chemical Analysis of the Window Glass, Research Department Report 92/2009 (Portsmouth: English Heritage, 2009). 50 David Dungworth, Basing Grange, Old Basing, Hampshire: Chemical Analysis of the Window Glass, Research Department Report 91/2009 (Portsmouth: English Heritage, 2009). 51 Roger Wilkes, Brocton WWI Army Camp, Cannock Chase, Staffordshire: An Investigation of the Window Glass, Research Department Report 88/2010 (Portsmouth: English Heritage, 2010). 52 McGrath and Frost, p David Dungworth, Chatsworth House Greenhouse, Chatsworth, Derbyshire: An Investigation of the Flat Glass, Research Department Report 90/2009 (Portsmouth: English Heritage, 2010).