The Alignment and Location of Medieval Rural Churches

ASPECTS OF THE ALIGNMENT AND LOCATION OF MEDIEVAL RURAL CHURCHES by Ian David Hinton being a Thesis submitted for the degree of Doctor of Philosophy in the School of History, University of East Anglia August 2010 © This copy of the thesis has been supplied on the condition that anyone who consults it is understood to recognize that its copyright rests with the author and that no quotation from the thesis, nor any information derived therefrom, may be published without the author’s prior, written consent St Mary’s and St Lawrence’s, South Walsham, Norfolk two churches in the same churchyard, but aligned 10° differently ABSTRACT This thesis explores the alignment of medieval rural churches and discusses whether their differing alignments have any specific meaning. It also examines the location of rural church sites and the chronology of church creation in relation to the process of settlement nucleation, the topography of church sites and their possible reuse. A survey of almost 2000 rural medieval churches provides the basis for this study. Part I provides a broad context for the detailed consideration of the results of the survey and their significance. It summarises earlier church alignment studies and the issues that they raise; the practice of alignment more generally; studies of the rural church and its place in the landscape; and earlier studies of medieval rural settlement. Part II describes the survey methodology and its basic results, applies the results to the theories advanced in earlier studies and evaluates them in the light of this new evidence. Part III discusses and analyses two significant variations which have been uncovered: the clear pattern of spatial variation in church alignment between the east and the west of the country, and the fact that between two and three times as many churches were built on east-facing slopes as on west-facing slopes. Possible reasons for these variations are evaluated and discussed. It suggests that harvest dates may have been a factor in the decision to build a church and that churches appear to be aligned with sunrise at early harvest completions. It also examines the chronology of the adoption of church sites and the development of local burial in rural areas. The possible relationship between earlier pagan sites and church sites, as part of the process of “Christian substitution”, is discussed, particularly in relation to the use of east-facing slopes. It proposes that local burial sites were adopted in villages early in the settlement nucleation process and that these graveyards provided the sites for the later building of churches, resulting in a bias of churches on east-facing slopes as the middle-Saxon burial sites seem to have sought them out previously. iii CONTENTS Appendices List of Figures Acknowledgements Glossary and Terminology Abbreviations v vi ix x xiii General Introduction 1 PART ONE – CONTEXT Chapter 1 Historiography of church alignment studies 7 Chapter 2 The historical use of alignment 51 Chapter 3 Historiography of church location, church origin and rural settlement 69 PART TWO – SURVEY DESIGN AND BASIC RESULTS Chapter 4 Survey method and basic results 107 Chapter 5 Survey results applied to earlier theories 127 PART THREE – ANALYSIS AND SYNTHESIS Chapter 6 Variation in church alignment across the country 171 Chapter 7 Relationship between slope and church alignment 215 Chapter 8 Location of churches; adoption of local church sites; development of local burial; Christian substitution 233 PART FOUR – CONCLUSIONS Chapter 9 Conclusions 277 BIBLIOGRAPHY 283 APPENDICES (listed overleaf) 304 iv APPENDICES Appendix 1 Survey results and analysis of Victorian churches 304 Appendix 2 Table of weekly sunrise positions by latitude 324 Appendix 3 Example church survey sheet 325 Appendix 4 Values of magnetic declination 1999-2008 by county 327 Appendix 5 Additional church alignment tables 329 Appendix 6 Sunrise position calculation formulae 330 Appendix 7 Details and additional analysis of sloping sites 331 Appendix 8 Easter calculation formula and east/west Easter dates 336 Appendix 9 Calculations to establish possible Norfolk minster-church sites 337 Appendix 10 Harvest festival details 346 Appendix 11 Alignment of Norfolk Monastic sites 351 Appendix 12 Detailed tables of church alignment and slope 352 Appendix 13 Technical details of digital Norfolk topography analysis 357 Appendix 14 Details of the sites of abandoned Norfolk churches 358 Appendix 15 Site-specific assessment of churches in Norfolk built on sloping land 362 Appendix 16 Comparison of minster-church and other church sites 366 Appendix 17 Churchyard finds in Norfolk 368 Appendix 18 Excel Spreadsheets of: Medieval church survey data (alignments, church and church site measurements, sunrise azimuths) Harvest festival data Victorian church survey data v CD inside rear cover LIST OF FIGURES St Mary’s and St Lawrence’s. South Walsham, Norfolk Frontispiece 1.1 1.2 1.3 1.4 1.5 1.6 The delaying effect of an elevated horizon on sunrise Meridiana at Santa Maria Novella, Florence and Palermo Cathedral St Mary’s, Rydal, Cumbria - built along the slope St Mary’s, Rydal, Cumbria – looking up the slope Sunrise positions at St Peter’s, Drayton, Oxfordshire Methods of correcting for church alignment errors for sundials 13 15 30 30 39 46 2.1 2.2 2.3 Major sun and moon setting positions The 18.6 year lunar cycle Remains of a ‘clava’ cairn at Balnuaran, Inverness 56 56 63 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 3.10 3.11 3.12 3.13 3.14 Alignments of York Minster and St Michael le Belfrey Tewkesbury Abbey during the floods in 2007 St Michael’s, Tirley, Gloucestershire, in the same floods in 2007 St Mary’s and hall at North Aston, Oxfordshire St Nicholas’, Harray, Orkney – next to an iron-age broch St Mary’s, Breedon on the Hill, Leicestershire – in iron-age hill-fort Spring below St Ishmael, Pembrokeshire Spring enclosure in St Lawrence’s, Gumfreston, Pembrokeshire Spring enclosure in St Lawrence’s (detail) View of the altar from spring enclosure, St Lawrence’s, Gumfreston Yew tree in the churchyard at All Saints’, Alton Priors, Wiltshire Knowlton church and henge All Saints’ Rudston and monolith Map of landscape classifications and survey areas 73 77 77 80 83 83 84 84 85 85 88 93 93 97 4.1 4.2 4.3 Survey areas Alignment graphs of urban and rural churches Church alignment by longitude and latitude 108 111 124 5.1 5.2 5.3 5.4 Sunrise azimuth and church alignment St Mary’s, Reepham and St Michael’s, Whitwell in the same yard St Andrew’s, Lammas, Norfolk – misaligned chancel St Hermes’, St Ervan, Cornwall – misaligned chancel 130 140 141 142 vi 5.5 5.6 5.7 5.8 5.9 5.10 5.11 Two-degree misalignment at St Lawrence’s, Castle Rising, Norfolk and at Holy Trinity, Goodramgate, York 144 Possible variations in nave/chancel alignment 145 Proportion of misaligned churches 147 Proportion of misaligned churches with post-medieval rebuilt chancels 149 Church tower at St Mary’s, Helmingham, Suffolk 161 Church tower at St George’s, Dunster, Somerset 161 ‘Frost days’ maps for Cornwall and Shropshire 163 6.1 6.2 6.3 6.4 Scattergram of alignment by longitude Scattergram of alignment by latitude Alignment of churches by longitude Alignment of churches in Norfolk by longitude 173 174 177 178 6.5 6.6 6.7 6.8 6.9 6.10 6.11 6.12 6.13 Archaeo-magnetic calibration for England Magnetic declination between 900 CE and 2000 CE Actual magnetic declination for England in 2002 Approximate magnetic declination for medieval England Church alignment by age of earliest fabric Areas investigated for Harvest Festivals Comparison of harvest festival dates 1870-1899 Harvest festivals in West and East Norfolk Illustration of calendar drift for harvest times 180 181 183 183 187 195 202 204 208 7.1 7.2 7.3 7.4 Possible effect of slopes on church alignment Alignment comparison by ‘slope-effect’ Churches on platformed sites by direction of downslope Churches on sloping sites by direction of downslope 217 219 223 223 8.1 8.2 8.3 8.4 8.5 8.6 8.7 8.8 8.9 8.10 8.11 Saxon settlement – Mileham, Norfolk Saxon settlement – Wellingham, Norfolk St Mary’s, Bawsey, Norfolk – hilltop site St Mary’s, Bawsey, Norfolk, view from the east Middle-Saxon churchyard finds in Norfolk Saxon settlement – Deben Valley, Suffolk All Saints’, Ramsholt location map All Saints’, Ramsholt, Suffolk from the northeast All Saints’, Ramsholt from the south Pre-parochial chapels in western Cornwall Early-Saxon and Middle-Saxon cemeteries in Norfolk 238 238 239 239 241 243 244 245 245 248 251 vii 8.12 8.13 8.14 8.15 8.16 8.17 All Saints’, Swanton Morley, Norfolk View of All Saints’, Swanton Morley from the east All Saints’, Ellough, Suffolk, on site of ‘pagan temple’ St Mary’s, Stody, Norfolk, from the east All Saints’, Fring, Norfolk St Peter’s, Mundham, Norfolk 261 261 267 270 270 271 In Appendices A1.1 A1.2 A1.3 A1.4 A3.1 St Barnabus’, Swanmore, Hampshire, aligned south east St Andrew’s, Kirkandrews, Cumbria, aligned at rightangles to the view from the house Alignment of post-medieval churches Alignment of post-medieval churches by date of building St John the Baptist’s, Barnby, Suffolk (with survey form) 310 310 313 320 326 (all photographs were taken by the author, unless otherwise acknowledged) viii ACKNOWLEDGEMENTS Thanks are particularly due to my wife, Maggy Chatterley; for her assistance in a good part of the fieldwork; for devoting many of her holidays over several years to this survey; for the enormous improvement in her map reading over the years and her ability to find remote and obscurely located churches, especially when occasionally having to use outdated maps; but mostly for her patience with, and interest in, my obsession. Thanks are also due to my supervisor Professor Tom Williamson at the University of East Anglia for his support and guidance throughout, despite fearing that both I and the whole concept were slightly mad at the outset. I must also thank several friends and colleagues who have had an interest in the subject and who have listened and made suggestions, particularly those who have helped with specific aspects of the research: Larry Newitt of the Canadian Geological Service for his attempts to calculate detailed values of magnetic declination from the past; John Davis of the British Sundial Society for his assistance with, and checking of, my sunrise calculations; the staff at the Norfolk HER; Bill Wilcox, fellow research student at UEA for obtaining and organizing the raw data for the calculations of the digital topography of Norfolk and Robin Forrest for allowing me to use his topographic basemap of Norfolk for some of the figures. As a life-long agnostic, I also need to thank my good friend Beryl Tooley, the local church organist, churchwarden and occasional lay preacher, for her interest in the project and her interpretations of religious matters, but who sadly did not live to read the final draft. Finally, thanks are due to the many keyholders, churchwardens and incumbents around the country who were prepared to give their time to unlock the many (sadly) locked churches and share their various opinions on this topic, as well as frequently bringing me back to earth by commenting “but churches face east, don’t they?” ix GLOSSARY & TERMINOLOGY Throughout this thesis when there are key numerical or statistical references in the text to important data which are contained in the relevant tables, these are highlighted in the same colour in both text and table to improve readability. There is no significance in the particular colours used in any instance. Alignment The term ‘alignment’, when used in relation to churches in this thesis, refers to the line parallel with the walls of the nave or chancel to the point on the horizon, measured in degrees from North, rather than using the more traditional term - the orientation of a church, which in its strictest sense refers to alignment towards east. Arc of sunrise The portion of the horizon where the sun appears to rise, between the Midsummer and Midwinter Solstices. Azimuth The actual position of sunrise, measured in degrees from North when horizon elevation and the position of the observer are taken into account. Azimuth values for the latitudes surveyed, over a level horizon, are shown in Appendix 2. The effect that horizon elevation has on the azimuth of sunrise is explained in Chapter One. Calendar (Julian) The Julian calendar, instigated by Julius Caesar, measured the year at 365.25 days (approximately eleven minutes too long). This resulted in a gradual drift between solar time and calendar time, such that by the late sixteenth century there was a ten day difference between the two (Cheney 2000, 17-19). Calendar (Gregorian) The Gregorian calendar, proposed by Pope Gregory, was designed to put the sun and calendar back into synchronization so that the date of Easter could be calculated correctly. In Europe it was adopted in 1582, but in Britain, with a doubt of all things Catholic, it was not adopted until 1752, when eleven days were removed from October that year (Cheney 2000, 17-19). Declination (magnetic) The amount by which North, as indicated on a magnetic compass, varies in an east-west direction from the position of True North. It was first measured in 1576 in London. Since this date, its value at Meriden (the geographic centre of England) has varied between 12° east in 1576 and 24° west in 1820 (Clark et al. 1988, 659). It currently (June 2010) stands at 2° 21′ west 1. The values for each area for the years surveyed are shown in Appendix 4. Declination, taken together with Inclination (variation in a north-south direction), makes up Secular Magnetic variation. 1 http://geomag.nrcan.gc.ca/apps/mdcal-eng.php (last accessed 18th June 2010) x Declination (star) Measures the position of a star in the sky in relation to the North Star and the celestial equator. Ecliptic Represents the plane of the Earth’s orbit around the sun. It is only when the moon crosses this plane that it can be eclipsed. Equinox (Autumn and Spring) The two days in the year when day and night are of equal length, each is halfway between the summer and winter solstices. Spring equinox is towards the end of March (around the 21st) and the autumn equinox towards the end of September (around the 23rd). On these two days the sun rises due east and sets due west wherever the observer is on the Earth (see Kaler 1996, 61-62; Heilbron 1999, 56). Moonrise standstill (major and minor) In a similar pattern to the movement of sunrise between the summer and winter solstices, the moon rises and sets between certain positions. Over a period of 18.6 years, these limits vary between the widest range (major standstill at each end) and the narrowest range – the minor standstill. This pattern is shown in diagram 2.2 in Chapter Two (Ruggles 1999, 36-37). Orientation The state of being orientated, in other words, facing the east (Chambers Twentieth Century Dictionary, revised edition 1970, London: Chambers). In modern times its use has been broadened to include leanings or affinity in many situations. Orbital Precession The result of the same phenomenon as Polar Motion. It causes the equinoxes to appear to move relative to the stars in the sky, but has no effect on the relationship between poles, equator and ecliptic for an observer on earth, which “will maintain their same configurations relative to the horizon year after year” (Kaler 1996, 151), so has no effect on sunrise for the purposes of this thesis. Polar Motion The rotation of the polar axis over a period of 26,000 years, so that North is indicated by different points in the sky (Kaler 1996, 149151), but has no effect on the relationship on compass directions and sunrises for the purpose of this thesis. “Remote” and “Isolated” churches Throughout this survey when discussing churches that are some distance from their settlement, Remote has been used where only one or two other buildings are close to the church, and Isolated has been used where the church is completely by itself. Saxon period The terminology and dates used during this thesis, are those adopted by both the Norfolk HER and the Suffolk SMR for Early Saxon (c.411–650 CE), Middle Saxon (c.651–850 CE) and Late Saxon (c.851–1100 CE). xi Secular Magnetic variation The amount by which the position of True North and Magnetic North differ; it is made up of figures for Declination, which measures the difference in an east-west direction, and Inclination, which represents the difference in a north-south direction. Solstice (Summer) The point at which the azimuth of sunrise changes from moving northwards to moving southwards, signifying the longest day, usually June 21st. Solstice (Winter) The opposite of the summer solstice, signifying the shortest day, usually around December 21st. Statistical tests Standard Deviation - Provides a measure of the amount by which a group of figures is spread about its mean (average) value; the larger the standard deviation, the wider the spread of figures. Confidence - An arithmetic mean is a point estimate of the average value of a sample group of figures, and degrees of confidence indicate how much reliance can be placed on the mean value for additional cases. Confidence coefficients or ‘degrees of confidence’ provide a range within which the mean value of the whole group is likely to fall – the greater the confidence required, the wider the range (Hayslett 1973, 150). For example, an arithmetic mean of a sample group of figures might be 55, but the actual value for the whole group will be in a range of 53 to 57 at 95% confidence and a range of 52-58 at 99% confidence – in other words, there is a 95% chance that the actual mean for the whole group lies within the first range and a 99% chance that it lies within the wider range. Generally, as the number of figures in the sample group grows, the more confident is the calculation of the mean value for the whole group and the narrower its range becomes (see Hayslett 1973, 150162 for a more detailed explanation of the methods of calculation). Throughout this thesis the automated methods of calculation of confidence levels within Microsoft Excel have been used (STDEV and CONFIDENCE). Where the whole group has been surveyed (such as the churches in western Cornwall or in Norfolk) the use of confidence figures is not applicable as the ‘sample’ and the ‘whole group’ are the same, as there are no other churches in the county to survey, so the calculated mean value is accurate and correct. True bearings The direction, measured in degrees from North after the value for the current Magnetic Declination has been subtracted, usually noted as “ °True”. xii ABBREVIATIONS USED IN THE TEXT (-)SMR (County or National) Sites and Monuments Record AOD Above Ordnance Datum (Mean Sea Level) BCE (CE) Before the Common Era (The Common Era) CBA Council for British Archaeology CCS Cambridge Camden Society DMV Deserted Medieval Village HER Historic Environment Record (the renamed Sites and Monuments Record in Norfolk) OE, ON Old English & Old Norse OS Ordnance Survey OSPSGA Oxford Society for the Promotion and Study of Gothic Architecture TASC Trevor Jones’ research - Saints Cults: Towards an Electronic Atlas. Found at http://le.ac.uk/elh/grj1/database/data.html August 2008) xiii (last accessed ASPECTS OF THE ALIGNMENT AND LOCATION OF MEDIEVAL RURAL CHURCHES GENERAL INTRODUCTION One end of every Church doth point to such a place where the sun did rise at the time the foundation thereof was laid … … and by the standing of these churches, it is known at what time of year the foundations of them were laid. (Sir Henry Chauncy 1700, 43) For at least three hundred years, the alignment of churches has been variously considered by antiquarians, ecclesiologists, folklorists, historians, archaeologists and has recently been investigated by geologists. Many studies of the alignment of churches have been published since the late nineteenth century investigating various ideas that churches were aligned towards specific points on the horizon; each of which will be examined in detail. Chauncy’s conclusion, quoted above, is a good example of the apparently definitive results that many of these studies have produced. Like Chauncy, the majority of these earlier researchers investigated alignments with sunrise, but many concentrated on sunrises that occurred on the feastday of the patron saint to which the individual churches were dedicated. Almost two thousand medieval rural churches have been surveyed for this thesis and their alignments vary by up to ninety degrees, which is exactly one quarter of the entire horizon. It is therefore easy to see why, over the years, reasons for this variation might have been sought when it was generally accepted that churches were aligned eastwards. The location of rural churches has also been studied for many years. Landscape historians have investigated the links between church location and possible previous ritual use of the site in question, whether through Christian substitution, readoption of a site or as an expression of Romanitas (for example: Bell 1998, 4; Eaton 2000, 14-17; Blair 1992; 2005, 377; Morris 1989; Rattue 1995; Stocker & Everson 2007), although recent work at Shapwick in Somerset (Gerrard and Aston 2007) has shown that care has to be taken when drawing conclusions based only on the current siting of the 1 church and settlement (because both are now known to have moved). Writers have also commented on the relationship between the site of the village church and that of the manor house (including Dymond 1968, 29; Morris 1989, 131; Blair 2005, 385; Williamson 1993; Scarfe 1987) and the location in relation to the village which the church currently serves (Morris 1989, 235-252; Ellison 1983; Wade-Martins 1980b; Williamson 1993). Medieval settlement patterns and settlement movement, particularly between mid-Saxon times and the twelfth century, have also been studied in detail and are intimately involved with the location of the churches which serve them. Other examples illustrate the complexity of the situation where the siting of churches can be explained by more than one possible reason, such as a combination of Christian substitution and the use of a locally prominent site. This thesis explores whether the differing alignments of medieval rural churches have any specific meaning. In the past it has been variously considered that churches faced east for liturgical reasons or reasons of Christian religious belief; that they faced Jerusalem; that they faced sunrise on the day that building started; or that they faced sunrise on their patronal-saint’s feast day, and, in the cases where nave and chancel were aligned differently, that this represented religious symbolism. In addition it has been suggested that churches were set out with a compass and therefore towards magnetic east when they were built, rather than true east. This thesis also explores the location of rural churches, especially the topography of their sites and the sites’ possible reuse. It also explores the timing of the adoption of the sites that now contain village churches, particularly in relation to the processes of settlements fixing their position and settlement nucleation, which, together with topographical elements, suggest that there are indications that some church sites may have determined the location of the settlement that they now serve rather than vice versa. 2 THESIS STRUCTURE Part I (Context) provides a broad context for the detailed consideration of the results of the survey and their significance. Chapter One briefly considers references to church alignment in church texts, and examines specific studies of church alignment in more detail. The methodology of these earlier studies is described along with their results and conclusions. The issues that are raised by the results of these studies, and in some cases the issues that their methodologies raise, are also discussed. Chapter One also considers the factors concerning the position of sunrise, such as the changing seasons, elevated horizons and calendar change, as well as considering the issues surrounding church dedication, all of which are elements that are central to the basis of most of the earlier church alignment research and its conclusions. Chapter Two discusses the background to the subject of alignment generally, including Palaeolithic and Bronze Age examples; and it covers aspects of alignment between two or more objects, and the alignment of objects towards distant features, both on the ground and in the sky. Chapter Three outlines the historiography of the origin of rural churches, especially the ‘minster model’; it also considers the location of churches, particularly in relation to the possible reuse of earlier ritual sites and the proximity of both the village and the lordly residence. It also outlines the historiography of rural settlement studies between the eighth and twelfth centuries, a time of much church building. Part II describes the author’s survey of almost two thousand rural medieval churches. Chapter Four outlines the sample selection as well as the procedures adopted for the survey itself, and provides an overall analysis of the basic survey results. In Chapter Five the results of the survey are applied to the various theories of alignment outlined in the earlier studies described in Chapter One. In particular, it uses the survey results to consider whether churches were aligned towards sunrise on their patronal-saints’ feast day; whether churches that have naves and chancels with different alignments represent religious symbolism; whether churches were aligned with sunrise on the day they were set out; whether churches were aligned towards sunrise at Easter and whether churches were aligned towards Jerusalem. It concludes that none of these earlier theories can be supported by the results of this survey, and that there is a more rational explanation for the observed alignment variations. 3 Part III (Analysis & Synthesis) enumerates and discusses two significant variations in alignment which were uncovered during the analysis of the results. In Chapter Six, the first of these variations – a clear pattern of spatial differences in church alignment between the east and the west of the country – is considered. This disparity has been revealed for the first time due to the size of the survey; three possible reasons for it are evaluated and discussed. Firstly, the possible influence that the variation in the position of magnetic north may have had in the setting out of church buildings is considered. Secondly, the possibility that there is a chronological element to the variation in alignment is examined, by investigating a possible chronology of church building and then comparing the alignments of the churches built in the different periods. Finally, the likelihood that climate played a part in the spatial pattern of the alignment of churches is investigated by examining whether harvest dates may have been affected by climatic differences across the country. The possibility that churches may have been first set out after a particularly early and successful harvest is then investigated, which might have been seen as an auspicious time to build a church. Chapter Seven expands on the second significant variation uncovered by the survey – that of churches sited on sloping land. The survey revealed that between two and three times as many churches were built on east-facing slopes as were built on west-facing slopes; possible explanations for this are put forward and discussed. The differences between this pattern, and the different distribution of churches built on artificially levelled platforms on sloping sites, is also considered. To test whether the bias of church sites towards east-facing slopes is a real one, or whether there is actually more land that slopes eastwards which may account for the inequality, a computer based analysis of the topography of the entire county of Norfolk is undertaken. In addition, the significant numbers of lost and ruined churches in Norfolk which were located in the same parish as an extant church are used to investigate whether the slope of their sites played any part in the selection of one church for retention over the one which fell into disuse. Chapter Eight develops the ideas raised in the previous chapters concerning the relationship between the siting of the church, the timing of the adoption of church sites, settlement nucleation, and east-facing slopes. The timing of the selection of the church site, and the development of local burial in rural areas, is discussed, particularly in relation to the apparent adoption of religious sites very early in the settlement 4 nucleation process. The proximity of the church and manor house and the sequence in which they were built is also examined, as it is usually considered that many of the country’s small rural churches were sited by the manorial lord on his own land, usually close to the manor house, which has been referred to as the church/hall focus. The issue of possible pre-Christian use of church sites is also discussed and is related to the reported attempts of the church to incorporate earlier ritual sites as part of Christian substitution. It also explores whether there is any indication that the origins, and hierarchy, of medieval rural churches influence specific church locations, by attempting to establish whether there were different factors affecting the location of different types of church, for example, were the influences on the decisions that were taken when siting a minster church, the same as those when deciding the site of a field church? Finally, in Part IV (Conclusions), Chapter Nine draws together the considerations of the previous chapters. In addition to indicating that the conclusions of many of the previous studies of church alignment can be shown to be in error, it proposes that in many situations there is a more practical, and simpler, set of influences on alignment than previously supposed. The overall aim appears to have been alignment eastwards, including aligning closer to east when the opportunity arose through rebuilding or extension, but it is suggested that the spatial variation in church alignment identified in the results of this survey can be explained by climatic variations reflected in harvest dates. An alternative process to the minster model in relation to the creation of some local church sites is also proposed. Instead of churches being built, and presumably their sites selected, later and later in the Saxon period as lower and lower levels of the lordly hierarchy are considered, it suggests that many religious sites were selected by local villagers in their settlements as a graveyard which later became the site for the building of what is now the village church. 5 PART ONE CONTEXT 6 CHAPTER ONE HISTORIOGRAPHY OF CHURCH ALIGNMENT STUDIES ‘There is no one but knows that every old church is built east and west.’ (John Mason Neale 1841b, 7) INTRODUCTION This chapter outlines the historiography of church alignment by examining earlier studies and surveys of church alignment and by summarizing references to alignment in broader church-related texts. It also considers other associated aspects which impact upon church alignment and its research; the position of sunrise and the factors which affect it, such as horizon elevation and calendar change, both of which alter the actual position on the horizon where the sun appears; and church dedication, which is at the centre of many of the previous theories of the reason for specific church alignments and the variation between them. Despite the apparent certainty of the quotation from Sir Henry Chauncy in 1700, noted in the General Introduction to this thesis, in which he concluded that churches were aligned with sunrise on the day that they were set out, Chauncy himself did not publish any corroborative details or explanations; his statement appears to have been made to try and explain the obvious differences that he had noticed between the alignments of individual churches. After this reference, the subject of church alignment appears to have been largely ignored for almost 200 years, apart from a poem by William Wordsworth about Rydal chapel written in 1823. Even the Ecclesiologists during the middle decades of the nineteenth century ignored it in the main, although Richard Morris has commented that “alignment intrigued” them (Morris 1989, 208). The Cambridge Camden Society (CCS) did produce The Orientator2 to measure church alignment as part of a church recording exercise suggested in an early publication by 2 The Orientator – a piece of equipment produced by the CCS, which was designed to assess the alignment of a church and compare it with the point of sunrise Chapter 1 7 John Neale (Neale 1841a, 10), but the survey seemed to fail completely (White 1962, 60), probably due to difficulties in data collection. However, apart from this exercise and very brief mentions of alignment in the pamphlets they aimed at church builders, the CCS not only had little to say about the subject of orientation but did not try and influence Victorian church builders either; in other words, they were only descriptive not prescriptive3. They were abrupt and forthright in their comments on many aspects of architecture and building form, particularly which aspect of pointed architecture was used, the size of the chancel, the chancel window style, the altar’s position and its height above the nave floor, but they did not appear to comment about the alignment of the church, or mention any requirement for new buildings to face east (or indeed any other direction, such as a saint’s-day sunrise). It has also been suggested that the Oxford Movement was interested in the orientation of churches. As Johnson wrote in 1912, quoting Victorian authors, “the practice of orientation had grown lax in the years preceding the founding of the Oxford Movement in 1833” (Johnson 1912, 206), although the Oxford Movement seems to have concentrated on liturgy and belief rather than on church buildings, apart from their decoration. In his book on the Oxford Movement, the Dean of St Paul’s, R. W. Church, wrote a 24-page chapter entitled The ideal of the Christian Church in which church buildings are not mentioned at all (Church 1892, 360-384) and it has been suggested that the Oxford Movement was on an altogether more spiritual and less material plane than the CCS (Brine 1990, 15). The fact that neither the Oxford nor Cambridge movements were really interested in alignment is illustrated by the results of the survey of Victorian churches described in Appendix 1, where it is shown that alignment towards east in Victorian times was rather less rigorously observed than by medieval church builders; indeed the only examples of churches in this survey that are aligned towards north or south are from the Victorian period. The lack of influence over alignment, particularly by the CCS, is also illustrated by the fact that there is no difference in alignment between the churches designed by architects who were members of the CCS and those who were not (see table A1.2 in Appendix 1 on page 309). Taken together, these clearly confirm the absence of any instruction from the 3 see Appendix 1 for a more detailed examination and analysis of the writings of the Victorian Ecclesiologists, including their few references to alignment and the details of the survey form they developed Chapter 1 8 Society in this matter and probably reflects the true Victorian interest in the issue, rather than that commented on by Morris (1989, 208) and Johnson (1912, 206). Interest in church alignment appears to have re-emerged during the late nineteenth century after which time a few local detailed surveys were published (Shore in 1886, Eeles in 1914, and comments on alignment in passing by Anderson (1898, 154) and Griffith (1908, 37)). There was a revival in interest shown by increasing numbers of published studies after the middle of the twentieth century which has continued until the present day (published studies by Cave 1950; Benson 1956; Searle 1974; Davies 1984; Abrahamsen 1992; Dymond 1999; Hoare & Sweet 2000; Ali & Cunich 2001, 2005; Hinton 20034; Muirden 2005; Wall 2006 and comments by Trubshaw in 1989). Some of these studies suggest alternative reasons why churches are aligned in different directions, frequently arriving at answers as definite as, but different from, Chauncy’s conclusion. The majority of these earlier surveys also investigate alignments towards sunrise, particularly sunrise that occurred on the feastday of the patron saint to which the individual churches were dedicated; others investigate sunrise on the day that building started; some have studied sunrise at Easter and others have suggested that churches faced Jerusalem. In addition, several have investigated the often-noticed difference in alignment between the nave and chancel in a single church and the proposal that this has religious symbolism (Cave 1950; Benson 1956; Hoare and Sweet 2000; Muirden 2005), with varying conclusions. This chapter will examine each of the studies of alignment, outlining the main issues that they raise and their main conclusions. Where it is felt that there are errors in the data, or that erroneous conclusions have been drawn from the data, these are identified and discussed briefly. Additional comments on the quality of the results of these studies will be raised in the following chapters when the results of this much larger survey are applied to the earlier theories. The summarised alignment results of all of these studies are shown in Table 1.3 on page 47, but prior to their detailed examination, mentions of church alignment in more general texts are outlined. Several general works concerning churches mention church alignment, sometimes only in passing; for example, John Blair in The Church in Anglo-Saxon Society only refers to alignment once, in a footnote, as “this difficult topic” (Blair 2005, 416). Earlier writers such as Baldwin Brown and Walter Johnson referred to orientation 4 An unpublished Masters Dissertation (Hinton 2003) Chapter 1 9 at greater length. Brown thought that “orientation was made much of in the later days of universal Christendom, although the church in the earliest times, and the Roman Church throughout, were indifferent to it” (Brown 1903, 22). Johnson, writing in 1912, devoted a whole chapter to the subject (1912, 204-242), in which he listed the findings of earlier surveys and the discussions of earlier theorists, especially the possibility of saint’s day sunrise alignments and the phenomenon of differently aligned naves and chancels, but ultimately sat on the fence by concluding that “it would be a bold man … who should affirm [a single] explanation [for alignment variation], and harder still … to dismiss every case as the result of chance or ignorance” (Johnson 1912, 242). John Harvey in 1974 referred to “the much disputed question of varying orientation” and that it had now been “settled in favour of patronal-saint sunrise” (Harvey 1974, 60), accepting the results produced by Rev. Benson in his 1956 study of Oxfordshire churches which is examined later in this chapter. Harvey went on to refer to two Cathedrals (York and Winchester) which were realigned to face east when rebuilt by the Normans, where the earlier Saxon cathedrals had followed the general alignment of the Roman street pattern (Harvey 1974, 74-75, 90-92). Richard Morris notes the same change in the adoption of alignment eastwards, particularly at York (Morris 1979, 116119) and he suggests that this was part of “a new concern with alignment during the eleventh and twelfth centuries … evident in cathedrals which were rebuilt and reorientated” (Morris 1989, 208). Nevertheless, this new concern with alignment, if it existed, was not universal even amongst larger churches, as not all cathedrals rebuilt at this time were realigned; for example at Canterbury, where “the walls of the final phase of the Anglo-Saxon cathedral lay parallel to the Norman cathedral … but five metres to the south” (Blockley 1993, 126; also in more general terms in Pounds 1994, 29), so the realignment appears to have been limited to cathedrals in towns with Roman street grids which were significantly different from east-west. Morris goes on to say that “there are signs that similar changes [in alignment] were made at village level” (1989, 208), this aspect is discussed further in Chapter Three. Stephen Friar lists most of the possible explanations for alignment that were discussed by earlier writers, but concludes that the reason for orientation has “remained a mystery” (Friar 1996, 326). According to Cruden’s Concordance, there are six references to East in the New Testament Gospels (Cruden 1769, 127-128) and a further three in Revelation. Most refer to the Second Coming on Judgement Day, for example: Chapter 1 10 Acts I, v.11, “[the second coming] shall so come in like manner as you beheld him going into heaven [to the east]” Matthew XXIV, v.27, “For as the lightning cometh out of the East …, so shall also the coming of the Son of Man be”. References to East in the Bible cannot have been references to buildings or their orientation, as purpose-built churches did not exist at the time the New Testament Gospels were written. The quotation below is taken from the Dictionary of Liturgy & Worship and is followed by an acknowledgement that Constantine buildings had their sanctuary at the west end, but that from the middle of the fourth century, the practice of locating the sanctuary at the east end was adopted and [alignment] became almost universal “but without complete accuracy in every case” (Davies 1972, 303). The siting of a building so that its sanctuary points to the east derives from the Christian practice of facing east for prayer. … the Christian eastern tradition could well have developed in contrast to the Jewish custom [praying towards the temple at Jerusalem], but would also have been influenced by the general pagan understanding of the time that the east is the direction in which the good divine powers are to be found, a view connected with sun worship (Davies 1972, 303). However, even though the Jewish religion required those praying to face the temple at Jerusalem, east was also obviously very important to the Hebrews; Cruden noted that they referenced all compass directions with relation to east; using words to signify ‘before’ for east, ‘left’ and ‘right’ for north and south and ‘behind’ for west (Cruden 1769, 127), and the Old Testament Books of The Bible contain 38 references to east (Cruden (1769, 127-128), reflecting its apparent importance; beginning with Genesis III, v.24, “God was placed at the east end of the Garden of Eden” (Cruden 1769, 127). Almost all of the church alignment studies discussed in this chapter focus on sunrise on specific days, so, prior to the summary and discussion of their conclusions, two specific issues which impact on the position of sunrise on the horizon need to be considered – horizon elevation and calendar change – as well as a general explanation of how the azimuth of sunrise moves throughout the annual seasons. In addition, as the authors of many of the studies of church alignment consider that churches faced sunrise on the feastday of their patronal saint, the issues of when churches were originally dedicated, and whether they were subsequently rededicated, become important. This section briefly discusses these issues, avoiding the complexities of the variations in the orbital motion of the Earth and of spherical trigonometry, and summarises the relevant aspects of church dedication. Chapter 1 11 THE POSITION OF SUNRISE Seasonal changes Sunrise at the Spring and Autumn Equinoxes over a level horizon occurs due east and sunset occurs due west everywhere on the planet, as the axis of the planet during the equinoxes is at right-angles to both the plane of the Earth’s orbit and to the sun itself (Heilbron 1999, 56). From the Spring Equinox, the position of sunrise moves north along the horizon, at a rate which is dependant on the latitude of the observer, until the Summer Solstice. At this point it reverses, moving south along the horizon, through east at the Autumn Equinox, until it reaches its most southerly point at the Winter Solstice, whereupon it turns north again. All over the Earth the sun is always due east at 6 a.m. (solar time), due south at midday, due west at 6 p.m. and due north at midnight, although at England’s latitude it is below the horizon, and therefore not visible, for some of this time. The actual length of day and night is again dependent on the latitude of the observer, as well as the season of the year (for a more detailed explanation see Kaler 1996, 61-80). The further north the observer’s position, the more quickly the position of sunrise moves northwards from east as the year progresses after the Spring Equinox, and the wider is the arc of sunrise between the two solstice dates. In extremis, at the Arctic Circle, the sun just sets on mid-summer’s eve at midnight and immediately rises again, therefore rising due North; and only just rises above the southern horizon at midday on mid-winter’s day, an arc between sunrises of 180º (90º either side of East). In England, the extreme positions of sunrise for the counties examined in this survey are at the most southerly point in Cornwall (50ºN latitude), where the sun rises between 54º at midsummer and 126º at midwinter (an arc of 72º, 36º either side of east), and at its most northerly point in Cumbria (55ºN latitude) where the range varies between 49º and 131º, an arc of 82º, 41º either side of east5 (full weekly sunrise details are shown in Appendix 2 on page 324). 5 Sunrise Positions based on local sunrise time taken from the US Navy website http://aa.usno.navy/mil/cgi-bin/aa.rstablew.pl last accessed Aug 2008, and converted using calculations from Davis 2004 (Appendices) – the formulae are shown in Appendix 6 of this thesis. Chapter 1 12 Horizon elevation In addition to the changes due to the seasons, the position of sunrise on the horizon is also affected by the height of the eastern horizon relative to the observer. An elevated horizon delays sunrise and makes the sun appear slightly later, therefore further south on the horizon, with a delay at England’s latitude of approximately 1.5º along the horizon for every degree of horizon elevation6. This effect is illustrated in Figure 1.1, showing a 17° delay in sunrise over a horizon elevated by 12° for a church that would have faced sunrise due east (90°) had the horizon been level with the church. The delay is caused by the fact that the Earth rotates completely every 24 hours – a rate of 15° every hour (360°/24) – so that at this latitude the sun moves in an arc, therefore during the delay of the sun’s appearance caused by the elevated horizon, the Earth has continued to rotate. This delay would apply to the many churches located in valley bottoms or on slopes that rise in an easterly direction. Therefore a church with an elevated eastern horizon aligned with sunrise on its patronal-saint’s day would be facing in a different direction from a church aligned with sunrise on the same day but with a level (0°) horizon. Figure 1.1 – The delaying effect on sunrise of an elevated horizon 6 The detailed calculations were based on formulae supplied by Dr John Davis (British Sundial Society) in a pers. comm., October 2002, and (Davis 2004, Appendices). The general spherical trigonometry is discussed in Kaler 1996, 198-201 and 467-472 (Kaler: Appendix 3) Chapter 1 13 The opposite situation, where the church is higher than its eastern horizon, would have the reverse effect, advancing sunrise relative to that over a level horizon and making it appear slightly more to the north. This has been observed in very few cases as part of the horizon measurements made in this survey, as a church on top of a hill is much more likely to be facing a similar hill the other side of the valley (effectively making its horizon level, or even elevated) than to be located on top of a hill facing out over a flat plain, or on the edge of a high cliff overlooking the sea. Calendar Drift The second issue concerning the position of sunrise is calendar drift, which progressively affected the relationship between the calendar date and the solar date before it was corrected in England in 1752 (Cheney 2000, 18). The drift impacts upon the exact position of sunrise on a specific date, such as a saint’s feastday, particularly when attempting to relate it to modern sunrise on the same date. The calendar date in medieval times was several days ahead of the same date today, after the calendar correction. In the medieval period there was concern about the calculation of the liturgically correct day for Easter which required a precise reckoning of the Spring Equinox (Heilbron 1999, 24-28; Cheney 2000, 4-6), the details of which are discussed in a later chapter. The error grew steadily after the introduction of the Julian calendar in 45 BCE and the adjustment to the calendar in most of the Roman world was made in 1582 CE by deducting ten days, but in the Protestant parts of Europe it was rejected at that time, not because it was inaccurate, but because it was popish (Heilbron 1999, 45). The change was made in 1752 CE in Britain, when the error was corrected by deducting eleven days from the calendar (Cheney 2000, 18). As part of their investigations to correct the calendar, Renaissance astronomers used churches as observatories to establish the exact day of the Equinox by constructing meridiana (Heilbron 1999, 62-68). These consisted of a line on the church floor which had to run due north-south, along which was traced the position of the sun at midday each day, by projecting the sun through a small hole in a south-facing wall. This enabled the solstices, the shortest and longest days, to be identified and by Chapter 1 14 inference the equinoxes. Incidentally, these constructions highlight the range of very different alignments of Italian churches, reflecting Baldwin Brown’s comment, noted earlier, about the Roman Church being indifferent to alignment. The requirement for a north-south direction for the meridian line meant that it would normally be expected to be roughly at right-angles to the axis of the church nave, which would be expected to be constructed close to east-west. However, the meridian line at Santa Maria Novella in Florence “stretched 58 metres up the nave, which was aligned near north-south, just reaching the choir” (Heilbron 1999, 69). At Palermo Cathedral, the meridian line runs diagonally across the nave and into the north trancept, showing the church to be aligned approximately southwest-northeast. Figure 1.2 – Meridiana at Santa Maris Novella, almost directly up the nave (left) and almost diagonally across Palermo Cathedral (right). Santa Maria Novella image sourced from http://math.nus.edu.sg/aslaksen/pictures/10143-s.jpg (accessed 24th July 2009) The calendar change has a particular impact when sunrise on a specific date, such as a saint’s feastday, is considered, as the sun appears at a different place on the horizon today from where it did on the same calendar date in the year that the church was set out. During the period between the middle of the tenth century and the middle of the fourteenth century, when most churches were being built, the error grew from six Chapter 1 15 days to nine days (Duncan 1999, 41-52). The difference between a specific date in the twelfth century, a period of much church building, and the same date today, is approximately seven days – sunrise, according to the calendar date, occurring effectively seven days earlier then. Seven days earlier translates to a difference in sunrise position of approximately 5° further north on the horizon around the autumn equinox, when the sunrise position is moving south, but 5° further south at the spring equinox, when sunrise is moving north, and virtually no difference (a tiny fraction of one degree) at the summer and winter solstices, when there is little day-to-day change in sunrise position (see Figure 5.1 on page 130 for an illustration of the movement of sunrise on the horizon and Appendix 2 for the actual figures). So, if churches were aligned with their patronal-saint’s feastday sunrise in medieval times, the calculations and adjustments required to establish where modern sunrise occurs on the day are fairly simple, depending only upon the year that the church was built. However, it is far more difficult to assess the adjustments required if churches were aligned with sunrise on the day that the building was set out, as the sun rises at the same point on the horizon twice in every year. Therefore, the changes in sunrise position brought about by calendar change relevant to each individual church depend not only on the year in which the church was built, but also on which season, as the adjustment would need to be made in the opposite directions for churches set out in Spring and those set out in Autumn, as sunrise is moving north in spring, day by day, and south in autumn. This means that any adjustment of sunrise position to correct for calendar change when attempting to prove building-start sunrise alignment is as accurate as tossing a coin. Chapter 1 16 THE HISTORIOGRAPHY OF CHURCH DEDICATION The final subject considered here as part of the background to church alignment studies is that of the dedication of churches. When, and to whom, churches were dedicated forms an important part of the topic of church alignment since much of the earlier work has made one of two assumptions; either that churches have always been dedicated to the same saint, or that churches that are aligned similarly must once have been dedicated to the same saint if they have different dedications now. The Reformation of the Church of England, the Puritan era and the Commonwealth period caused the knowledge of many church dedications to be lost or altered. John Ecton and Browne Willis worked in the eighteenth century to establish lists of dedications which had been lost. They had difficulties in some areas where dedications had been lost completely from folk memory and “made many assumptions” (Orme 1996, 47-50). Frances Arnold-Forster published a list of dedications in the late nineteenth century, but Francis Bond commented in the early twentieth century that Arnold-Forster’s volumes also contained a large percentage of dedications of doubtful authenticity, as well as many still unknown (Bond 1914, 14). More recently, dedications in specific counties have been investigated more thoroughly, for example by Alan Everitt in Kent (1986), Nicholas Hoggett in Hertfordshire (1988), Peter Northeast in Suffolk (1995) and Reverend Linnel in Norfolk, who also acknowledged that Arnold-Forster was less accurate than Bond (Linnel 1962, 4). In addition, the dedications of Saxon and Norman Monastic Houses have been studied by Alison Binns (1989) and Tim Pestell (2004). Much of this recent work has delved deeper into early documentary records, particularly the work of Richard Clark (1992) in Derbyshire, identifying high levels of rededication; Wilhelm Levison (1956) on ninth-century churches; Lawrence Butler (1985) on the Anglo-Saxon churches surveyed by Harold Taylor; and of particular saints, such as Oswald, by Alison Binns (1995). The West Country has been particularly well studied, with recent publication of work by Nicholas Orme (1996), Susan Pearce (1985, 2003), Catherine John (2001) and Sam Turner (2006). Most recently, Trevor Jones’ work, (Saints Cults: Towards an Electronic Atlas [TASC online database] and Jones 2007), has altered the thinking about rededications, and has, in particular, pointed out the possible error of applying results based on the proportions of rededications identified in the work of Clark and Orme to the rest of the country. This is discussed in more detail in later sections. Chapter 1 17 Definition of Dedication The reference to a church being dedicated to any saint is merely convenient shorthand, as early writers held that no early church was originally dedicated to a saint, but only to God. Both Bond and Muncey quoted Hooker (the Protestant writer) writing in the late sixteenth century, as saying that “Churches were consecrated to none but the Lord only” (Bond 1914, 1; Muncey 1930, 2). Bond quoted Saint Augustine of Hippo as saying “To the saints we appoint no churches, because they are not unto us as gods, but as memorials as unto dead men, whose spirits with God are still living.” (Bond 1914, 3), whereas Muncey expressed it slightly more generally: “It is not properly correct to speak of a church being dedicated to a particular saint or event – it is a convenient way of expressing that we mean that it is dedicated to God in memory of a particular saint or event” (Muncey 1930, 3). The practice of dedication to a saint may have originally been confined to the altar of the church, particularly if it contained relics. Certainly in later years, as altars and cults proliferated, the frequent side altars in a church were dedicated to a range of saints, different from the dedicatee of the high altar (Duffy 1992). This aspect is confirmed by Graham Jones’ definition of dedication as “the commemoration of a saint, angel or aspect of the Divine, by naming a place or object of devotion in their honour, part of a larger phenomenon of religious cult” (Jones 2007, 16). Richard Hooker devoted two chapters, albeit short ones (Keble 1888, 44-51), of his eight volume Ecclesiastical Polity to the issue of the dedication of churches, which included an explanation that it was now the church that was so named and not the altar, which would have been too idolatrous for the Protestants to accept, although he did say that as far as dedication was concerned “sometimes they [idolaters] may judge as rightly what is decent about such external affairs of God” (Keble 1888, 49). His writing certainly implied that originally it was only the altar that was the subject of the dedication. The incumbents, churchwardens and parishioners encountered during the survey for this thesis have universally referred to their church by the name of its dedicatee only, for example St Mary’s or St Peter’s, and they almost universally assume that the dedication of their church has remained the same since time immemorial. Chapter 1 18 Why dedicate a church? A church was supposed to be consecrated before Mass could be celebrated in it. This task had to be undertaken by a bishop and originally had three associated ceremonies: Dedication of the church to God; frequently [but not always, as is shown below] naming the church in honour of a saint; and consecration of the altar and enclosure of any relics within (Orme 1996, 4-5). The chosen saint would then act as an intercessor, or intermediary, between the prayers of the individual and God, with the aim of magnifying their power. In addition it was believed that the presence of the saint magnified the effectiveness of oaths and contracts, which was one of the reasons for the proximity of churches and market places and churchyard fairs (Jones 2007, 17), still seen today by the use of The Bible on which to swear oaths. The choice of saint in early days was most likely to be made by the patron of the church, whether thegn or bishop, but by the High Middle Ages, the views of the bishop “are likely to have been decisive in many cases, though paying careful attention to the wishes of influential individuals, particularly those providing funds” (Jones 2007, 20). The selection of a particular saint to act as the patron for a church was made for a variety of reasons. Dedicatees could be selected to promote Christianization, by using the apostles such as Peter, Andrew or Bartholomew; for reasons of practicality, by using saints whose feast days coincide with specific points in the farming year which reflect the local needs, such as Brigid and Anthony, whose feastdays occur during the birthing of different animals, or for supporting local trades, such as Peter for fishing. Graham Jones expresses this as “it is permissible to imagine that the feast saint was more likely to be chosen for their appropriateness for the secular seasonal cycle, especially the agrarian calendar” (Jones 2007, 51). In addition to the reasons noted above, the choice of saint might reflect the contemporary popularity of a particular saint or cult. Saints such as George and Mary waxed and waned in popularity both over time and geographically (Jones 2007, 13), and Giles and Leonard were popular in the twelfth century with monastic founders, and may also have been chosen as parish church patrons during the same period (Orme 1996, 31). Alison Binns recorded 514 dedications of monastic houses between 1086 and 1216, 235 of which were to Mary, 49 to Peter and Paul and fifteen to Andrew (Binns 1989, 18-19), but popularity changed during this period, Peter and Andrew were popular at the beginning (10661100) but almost no dedications are recorded to them towards the latter end, between Chapter 1 19 1150 and 1216 (Orme 1996, 27), emphasising changing popularities. Finally there might be political reasons for a choice of dedicatee, either nationally or locally. At a national level, there were dedications to Thomas Becket, whose death “evoked one of the most popular movements in Christian history” (John 2001, 109), whilst at a local level, a favourite saint of a major patron might be used, such as at the churches in manors controlled by the Beaufoy family of Norfolk, all of which were dedicated to St Andrew (Linnel 1962, 9). Another example would be the choice of the same patron saint for daughter churches as the one venerated at the mother church, such as at the cluster of ten adjacent dedications to St Mary in south Suffolk and north Essex, attached to the minster church of St Mary’s, Stoke by Nayland (Cooper 2000, 161-168, Webb 2006, 27-28). The popular explanation for the predominance of dedications to St Mary on the roads to the shrine at Walsingham in Norfolk is that they reflect the pilgrimage routes to the shrine, which is dedicated to The Assumption, but Reverend Linnel thought this was purely conjectural as only two of the churches dedicated to Mary can be definitely linked with the Assumption rather than one of her other festivals (Linnel 1962, 8). When were churches dedicated? Any one of the reasons noted above could have been the basis for the final selection of the dedicatee for a church, but when was this choice made? If the saint was not selected, and known by the builder, before the foundations of the church were set out, then the building could not have been laid out on the correct alignment for the patronalsaint’s sunrise, and any sunrise alignment could only have been correct by chance. Churches have apparently been dedicated to saintly figures since very early Christian times. By the end of the fourth century it was becoming usual to place the church under an additional saintly patron (Orme 1996, 4). Bede recalled a few churches dedicated to particular saints – Alban and Martin in St Albans and Canterbury; whilst in Cornwall, the first dedication recorded was to Docco, in the early sixth century (Orme 1996, 12). Certainly by the ninth century, a saintly dedication appears to have been the norm, as the Synod of Chelsea in 816 CE decreed that “when a church is built, it shall be consecrated by a proper diocesan, who shall take care that the saint to whom it is dedicated be pictured on the wall, or on a tablet, or on the altar” (Muncie 1930, 89). Chapter 1 20 Nevertheless, despite Anglo-Saxon law and Synods, many churches remained undedicated as late as the twelfth and thirteenth centuries, “probably because the ceremony of consecration became so long and elaborate” (Linnel 1962, 8). Butley Priory in Suffolk was still undedicated seventeen years after its foundation and Cirencester Abbey in Gloucestershire remained undedicated for an even longer period (Orme 1996, 5). It seems unlikely therefore, if large monastic and abbey churches could remain undedicated for decades, that parish churches would have been treated very differently. If anything, parish churches would be more likely to escape ‘under the radar’, perhaps with greater numbers not dedicated to a saint when originally built, although this is difficult to substantiate since few parish church records survive from before the thirteenth century. One example of a parish church which remained undedicated is at Leuchars in Fife; it was still unconsecrated 60 years after its completion in 1184 (Linnel 1962, 9). In 1237 the Papal legate, Cardinal Otho, required that all such churches should be dedicated within two years, as without it no masses would be allowed (Orme 1996, 5). This implies that the problem of undedicated churches was considered both important enough, and still widespread enough, to have required intervention at the highest level. Once dedicated, the annual patronal festival was obviously important to the church hierarchy, as indulgences were offered by the Bishop of Exeter to people for keeping them – 24 days in Exeter in 1231 and 30 days in St Buryan in Cornwall in 1238 (Orme 1996, 8). Were dedications changed? There are many ways in which the rededication of a church might occur. Many of these are true rededications, either at a change of owner or as part of the rise of a specific cult, but in other cases, where the modern and medieval dedications are different, the change may well have been caused by the methods used by eighteenthcentury antiquarians in ‘discovering’ lost dedications. These methods are discussed below. All of the factors affecting the possible choice of saint for the original dedication, noted earlier, could also apply if a church was rededicated; changes in farming patterns in the area, changes in the popularity of particular saints and cults, changes in politics and even at a change of owner, particularly in early times. The Reformation of the Church in England saw an increase in the number of changes to Chapter 1 21 dedications to apostolic saints, and to All Saints or All Hallows, that had started in the fourteenth and fifteenth centuries (Bond 1914, 191), which was seen as part of the reduction of idolatry and image worship, particularly under Edward VI, when “dedication was further marginalized” (Orme 1996, 42). These changes throughout the Reformation, and after, meant that the original dedications were gradually lost from folk memory. The reduction in the number of mentions of the name of the church in testator’s wills in Derbyshire illustrates this, with indications “that perhaps by the end of the sixteenth century, church dedications were no longer part of popular knowledge” (Clark 1992, 54). A similar pattern was noted in wills around Beccles in Suffolk during the later sixteenth century all of which refer to burials only in unnamed churchyards (Pers. Comm. David Lindley, Oct 2007). The loss from folk memory of dedications is also illustrated elsewhere in Suffolk, when the new vicar at Exning in 1823 found that “not one parishioner knew the saint-name of the church” (Northeast 1995, 201). In some churches, the political importance attached to relics, and the veneration of saints, meant that devotion to a particular figure could cause the dedicatee of the church to be altered. This “proactive use of dedications is witnessed at Hoxne in Suffolk”, where dedications changed from Ethelbert to Edmund to Peter and Paul over a period of about 300 years between 800 CE and 1100 CE, “and provides a potent example of the dynamics behind the choice of the patron saint of a church” (Pestell 2004, 94). Politics also influenced cases where the dedication of a wealthy and popular chantry overshadowed the dedication of the church in which it was founded, the chantry saint’s name ultimately taking the place of the original dedication for the whole church (Muncie 1930, 99). Muncie quoted examples where this was the case in the City of Cambridge, at Hitchin in Hertfordshire and at Walton and Griston in Norfolk7 (Muncie 1930, 99-100). Once consecrated, a church may not be consecrated again, with a few exceptions; if it was “polluted by blood”, had fallen into ruins or had been almost entirely destroyed by fire (Muncie 1930, 94-95). Whether the extension of a church through the addition of an aisle required re-consecration, or just consecration of the new, previously unconsecrated, floorspace, is unclear, but in any of these situations where a re-consecration was performed, it could have included a rededication. Alison Binns, writing about churches dedicated to Oswald, notes that precise evidence for the 7 The last two examples were cited by Muncie from Blomefield’s History of Norfolk, Vol II, 290 & 316 Chapter 1 22 year of dedication for churches dedicated to St Oswald survives only in very few cases – in 1241, 1349, 1447 and a rededication in 1287. She feels that there was “good reason” to suppose that each of these was a rededication or a new building on an old site (Binns 1995, 243) and assumed that this indicated a continuation of popularity of Oswald, but does not offer any evidence that supports an Oswald dedication before these dates. Feastdays were also moved, without changing the dedication, in parishes where the patronal feast happened at an inconvenient time in the farming year, such as at seed-planting or at harvest time. It became normal to transfer the feast to a more suitable time of year, often around Michaelmas at the end of September, when harvest gathering in was completed (Muncey 1930, 32-33). This period is also when the later annual labouring contracts were coming to an end and is considered to be the end of the farming year, before the winter activities started for the new farming year (Kussmaul 1981). Orme quotes an example at St Dominic in Cornwall, where the feastday was allowed to be altered in 1445, from 30th August, during harvest time, to the 9th May, a slacker time in the agricultural calendar (Orme 1996, 9). In a similar vein, at Hatfield in Hertfordshire in 1226, a fair was granted for four days on the feast of St John the Baptist (24th June). In 1318 this was altered to the vigil of St Etheldreda (October 16th) and the two days following (Doggett 1989, 10). Doggett went on to speculate that the dedication of the church was changed at the same time “to give impetus to the cult of St Etheldreda” (Doggett 1989, 10), but the change also shifts a long holiday from the time of the hay harvest to an agriculturally quieter time of the year. To complicate matters further, by the time of the Reformation, fairs were only held in Hatfield on the feasts of St George (23rd April) and St Luke, an apostolic saint, whose feast was conveniently on the 18th of October (Doggett 1989, 10), and presumably more acceptable as a celebration than for a female saint. The fact that each of these revised feastdays and fairs would have indicated to the antiquarians John Ecton and Browne Willis, when collating their lists of dedications, that the church dedication itself had been altered will be discussed below, however, the church at Hatfield is still dedicated to St Etheldreda today, thus avoiding misinterpretations by Ecton or Willis. Richard Morris quotes examples from all over the country where settlement names, with continuity since Domesday, are based on the presence of a church and its owner’s/founder’s name such as Baschurch, Shropshire (now dedicated to All Saints), Chapter 1 23 Colkirk, Norfolk (now dedicated to St Mary), and Offchurch, Warwickshire (now dedicated to St Gregory) (Morris 1989, 157 and the relevant Pevsner). The fact that these settlements were known by these names could mean that the churches themselves were not dedicated to a saint at the time. Morris also quotes two similar examples – Alvechurch in Warwickshire and Pucklechurch in Gloucestershire – but as names from pre-Conquest sources. He goes on to suggest that “it would be interesting to know who the dedicatee of the church at Pucklechurch was at this date” since it was rededicated to Thomas Becket in the twelfth century (Morris 1989, 157). This has now been truncated to just St Thomas (Verey 1970a). Whilst in some areas of the country up to two-fifths of the parish churches appear to have been rededicated at some point in their history, it is becoming apparent that this level of change was not necessarily universal. Richard Clark has identified that forty per cent of the churches in Derbyshire have changed dedication between the sixteenth century and the present day (Clark 1992, 49-61); Nicholas Orme has showed that almost thirty per cent of the modern dedications of Devon’s churches are different from pre-Reformation ones – 140 out of 482 – and even in Cornwall, where many patron saint names are preserved in the name of the village, 30 out of 218 churches (14%) are dedicated to different saints now from when the villages gained their name (Orme 1996, xii). In Kent, however, according to Alan Everitt, “very few [dedications] seem to have been altered at the time of the Reformation, in most cases they are recorded in early wills or other early documents; and in a significant number in the Domesday Monachorum or in other pre-Conquest sources” (Everitt 1986, 227). Other researchers do not quantify levels of rededication, for example Sam Turner in the West Country “does not rely heavily” on church dedications in his investigation of the medieval church in the west because there is “normally no way of discovering when they [dedications] were first used at any specific church” and so many of them have changed over the years (Turner 2006, 9). Despite some lower figures in other counties, it might still be assumed that the level of rededication discovered in Derbyshire and Devon was common in all areas, but Graham Jones states that “Assuming that this level of change was universal, is inaccurate” (Jones 2007, 48). He goes on to list several counties where rededication rates are far lower than in Devon and Derbyshire; particularly in Worcestershire, with “attrition rates” of less than ten per cent, rededications of just under a quarter of the churches in Leicestershire, and post- Chapter 1 24 Reformation changes in dedication in the West Midlands seem to have affected between ten and fifteen per cent of parochial churches (Jones 2007, 49 – based on TASC figures). Whether these lower rates will apply to other areas of the country remains to be seen. Even today, 26 of the 1,926 churches in the survey for this thesis have no current dedication recorded either in Pevsner’s Buildings of England or on the notice board, or in the porch, of the church itself. Two of these were described by Pevsner as “old church”, where the original church had been replaced by a later church in the parish, leaving 24 churches, most still consecrated and used, without a dedicatee. This either reinforces the suggestion that some dedications became lost from folk memory after the Reformation and in these cases were never recovered or reinstated, or that some churches were never dedicated to a saint in the first place and these escaped the pressure of the medieval church hierarchy to rectify the situation. Establishing lost dedications Reference was made earlier to the work of John Ecton and Browne Willis in attempting to identify dedications that had been lost. There was a revival of interest in the eighteenth century in church dedications and Ecton, and later Willis, began to compile gazetteers of church dedications and festivals. In doing so they used the eighteenthcentury parish feast dates to conjecture the date of the medieval patronal-saint’s feastday (Orme 1996, 48; Jones 2007). For many reasons, not least that in many cases the feast day had been altered, and that there was confusion as to exactly what the parish feast was celebrating, these lists were inaccurate. In Devon only 54 (31%) of the known medieval dedication feasts coincide with the eighteenth-century parish feast date (Orme 1996, 48/9). Willis assumed “without exception” that the parish wake (feast) occurred on the nearest Sunday to the patronal feast (Clark 1992, 52). Clark then quoted examples in Derbyshire where the wakes took place on dates which did not reflect the medieval dedications, known by evidence from wills. The fact that Ecton had had to use wakes as an indicator of the saints day, implies that the clergy and parishioners at this time did not know the true date. Recorded by post-Reformation wills, almost all of the wakes fell between June and November, with September the peak month, an influence of seasonal activities (Clark 1992, 53). This matches well Chapter 1 25 with the attempt by Henry VIII to move the annual feast of the dedication for all churches to the 1st October from “that called commonly the church holy day” (Orme 1996, 10). This refers to the celebration on the anniversary of the original date of dedication, rather the celebration of the feastday of the patronal saint. Since a church should not have been used for Mass before it was consecrated, it is unlikely that consecration would have been delayed until the saint’s feastday, so there are likely to have been two separate feast dates in most parishes. As Graham Jones put it “Patronal festivals need to be distinguished from feasts on the anniversary of [the churches’] consecration – known as the dedication festival”. This may have caused confusion in later times when the date of the parish wake was taken to represent the feast of the patron saint (Jones 2007, 18). These factors contribute to the considerable inaccuracies in the eighteenthcentury lists, carried through into Frances Arnold-Forster’s gazetteer in the late nineteenth century. Many writers since then have commented on the detailed local work required to establish documentary evidence of medieval dates in order to have any likelihood of establishing the original date and saint, Clark concluding “never assume continuity, unless it can be palpably proved” (Clark 1992, 54). CHURCH DEDICATION: CONCLUSIONS It appears that some, possibly many, churches may not have been dedicated to a saint when they were first built. Some churches have been rebuilt, and some of those may have been rededicated. Although not unusual, rededication is not apparently as common as had been thought, and Graham Jones’ work argues against the previous thinking – that between a third and a half of all medieval churches were once dedicated to a different saint from their current dedication – but it still appears that at least some rededication took place in all parts of the country. Perhaps it is not unreasonable to estimate that around 20% of all churches have been rededicated since they were first built, whether their original dedication was altered intentionally by the owner or the authorities, or altered unintentionally by antiquarians anxious to provide a complete list. Despite the possible problems that rededication might cause when assessing patronal-saint sunrise alignments, it is still important to use the large dataset provided Chapter 1 26 by the survey for this thesis to examine whether churches face specific sunrises. If the proposal that churches do face their patronal-saint’s sunrise is true, then whilst the churches may not face sunrise of their current dedicatee, the range of overall alignment should reflect a pattern of saints that were popular prior to any rededication. It will also allow the results of all the earlier alignment studies to be placed in context. Chapter 1 27 EARLIER CHURCH-ALIGNMENT STUDIES Each of the published studies of church alignment that was mentioned in the introduction to this chapter is summarised below and the main arguments and conclusions are considered and commented upon. Prior to this, a poem by William Wordsworth which mentions church alignment is analysed. Although not strictly speaking a study of church alignment, Wordsworth’s poem, written in 1823, has been quoted as a source in several of the later studies on the subject, and so is discussed here as the first in chronological order. William Wordsworth ‘On the same occasion’, The Literary Associations of Rydal Church, E. Jay, editor, Rydall: Armitt Trust, 1993, unpag This extract is from the second of two poems written to Lady Fleming of Rydal Hall entitled “On seeing the foundation preparing for the erection of Rydal Chapel, Westmorland” in 1823. This second poem was subtitled, “On the same occasion” (the second, third and fourth stanzas are quoted below). Then, to her Patron Saint a previous rite resounded with deep swell and solemn close, through unremitting vigils of the night, till from his couch the wished-for sun uprose. He rose, and straight – as by divine command, they, who had waited for that sign to trace their work’s foundation, gave with careful hand to the high altar its determined place; Mindful of Him who in the Orient born there lived, and on the cross his life resigned, and who, from out the regions of the morn, issuing in pomp, shall come to judge mankind. This poem has been quoted in earlier work on church alignment (Cave 1950, 47; Benson 1956, 206; Ali & Cunich 2001, 155; 2005, 56), but was it poetic licence? Did Wordsworth actually see this process – a vigil watching for sunrise on the feastday of the patron saint and the subsequent fixing of the position of the altar and alignment of the church – or was he converting Simon Domville’s seventeenth-century idea, which uses many of the same words, into poetry? Domville’s manuscript, which was sold at his death in 1678, is cited in Johnson (1912, 225), as follows: Chapter 1 28 In the days of yore, when a church was to be built, they watched and prayed on the vigil of the dedication, and took that point of the horizon where the sun arose from the East … . I have experimented (sic) some churches and found the line to point to that part of the horizon where the sun arises on the day of the saint to whom the church is dedicated. At Rydal, the theory of patronal-saint sunrise alignment actually appears to work. The church is dedicated to St Mary, and is aligned at 66º True, but with an eastern horizon elevated by 12º, thus delaying sunrise by approximately 17º, as shown in Figure 1.1 earlier, which means that to appear at the elevated horizon at 66º True, sunrise at a level (0º) horizon would have to have been at 49º True (66° minus 17°). At this latitude, this is close to the summer solstice at around the end of June. There are two feast days for St Mary at this time of year; Salutation on June 25th and Visitation on July 2nd. Unfortunately, this fairly well publicised nineteenth-century event has been seen as the proof of what had gone before and has given credence to the theory, despite the fact that the measurements here, which appear to confirm patronal-saint sunrise at this site, are the first time that such statistical ‘proof’ has been published for this key site. Previously it had only been taken on trust, probably based on the authority of the reporter (Wordsworth), which has been enough for some writers to be convinced that it must have been an ancient ceremony and applied at most, if not all, churches and in all periods. However, whilst Rydal chapel apparently faces close to sunrise on one of two midsummer Marian festivals, they are not the most important ones. After the Reformation, Mary’s Assumption (August 15th) was omitted from the list of festivals and her Nativity (September 8th) was then considered the most important (Friar 1996, 277), rather than her Salutation or Visitation, so why was the church not aligned with sunrise on her Nativity? The fact that it is built at rightangles to the axis of a 1 in 15 slope (6.6%), as shown in Figures 1.3 and 1.4, probably had a far greater influence in the final alignment and position of the church than any particular sunrise date. If the church had been aligned with sunrise on Mary’s Nativity, when sunrise would have appeared at 100º True (Sept 8th sunrise at 83º 8, +17º horizon “delay”), it would have been aligned much closer to east, but it rotates the church 34° to the south of its existing alignment; almost diagonally down the slope, rather than along it. Building at such an angle to the axis of the slope would make the building process more difficult in terms of buttressing for twisting loads, and, equally importantly, it makes the internal 8 See Appendix 2 for sunrise details at 55° North Chapter 1 29 liturgical layout difficult, with the chancel floor several feet lower down the slope than the floor of the nave. So even here at Rydal, where the church is actually aligned with one of the feastday sunrises of its patron saint, the saint’s day sunrise alignment is probably coincidental and the result of other much more practical factors. Figure 1.3 – St Mary’s, Rydal – built along the 1 in 15 slope Figure 1.4 – St Mary’s, Rydal – looking up the slope Chapter 1 30 There are also doubts about whether Wordsworth actually saw the process of the setting out of the foundations at this church in the way he describes. A sevenvolume anthology of the letters of the Wordsworth family enables another view to be taken. William Wordsworth, his wife Mary and his sister Dorothy lived in a house named Rydal Mount adjacent to the church site for 37 years, between 1813 and 1850. They were tenants of Lady Fleming of neighbouring Rydal Hall, with whom they fell out in May 1822 as a result of their complaints about the condition of the house (de Selincourt 1939, 71). The poem about the setting out of the church must have been written towards the end of 1822, certainly after August 6th, when Dorothy wrote to Edward Quillinan “my brother has not composed a single verse since you left us [in May 1822]” (de Selincourt 1939, 88); and before February 1823, when Mary wrote to Lady Beaumont, enclosing copies of the two poems, saying that she “hoped that they had the power of a peace offering [to Lady Fleming]” (de Selincourt 1939, 104). On November 19th 1822, in the middle of a long letter to Edward Quillinan, Dorothy wrote an apparently disconnected sentence “The church is to be built, in the orchard next to our field” (her underlining) (de Selincourt 1939, 98), implying that there had been some doubt about whether the church was to be built or not, and also where it was to be built. In February 1823, building work on the church had apparently still not been started as Mary, in a letter, asked Lady Beaumont “if you or Sir George could send us any hints, or sketch for a chapel that would look well in this situation, it is possible that it could be made useful through her [Lady Fleming is identified in a footnote] agents” (de Selincourt 1939, 104). In November 1823, Dorothy wrote to Catherine Clarkson “our church is near finished on the outside and is very pretty and you can have no idea how beautiful in connexion with the village, especially when seen from the other side of the Lake” (de Selincourt 1939, 126). These are the only references to the church in the published letters of this period of eighteen months. This raises several questions about the whole building process at this site. The church was built during 1823, and the setting out referred to in the poem, if it took place at all, must have happened in late June 1822 to coincide with one of the midsummer Marian feasts. Firstly, if the church was to be built in the orchard, were the trees grubbed out before the vigil was undertaken so that sunrise could be seen over the horizon, effectively marking the site, or afterwards, just before building started? Chapter 1 31 Secondly, if Wordsworth had actually witnessed the vigil at the location where the church was to be built (to set its alignment), why was Dorothy later expressing surprise that it was to be built in the orchard next to the house? Thirdly, why didn’t this unusual action (Wordsworth being out in an orchard around 3:30 a.m.) feature in the letters of one of these three prolific letter writers, when many other minor events, including several mentioning the church, were included? Although there is no reason why Wordsworth should have written the poem soon after he had witnessed the overnight vigil, it is odd that such an unusual action went unrecorded until the poem was written the following year. It seems eminently possible that this poem was just that – poetic licence – and that Wordsworth was putting in to words an idea that had certainly been around for at least a century and a half. The possibility that it is just a story is strengthened by the fact that Wordsworth himself admits to using such license in some of his poems by writing to Isabella Fenwick in 1843 that “I do not ask for pardon for what there is of untruth in such verses, considered strictly as matters of fact. It is enough if, being true and consistent in spirit, they move and teach in a manner not unworthy of a Poet’s calling” (de Selincourt & Darbishire 1947, 415; Nuttall 1974, 114). However, whether it was real or not, many subsequent church researchers have treated it as an actual observation, and more importantly, as proof of a centuries-old tradition. Shore T., 1886, ‘Orientation of Churches in Hampshire’, Walford’s Antiquarian Magazine and Bibliographer 10, 105-108 Shore measured the alignment of over ninety churches in Hampshire and Wiltshire. Fifty-seven named churches, plus “others”, were measured as being aligned between 20° and 22.5° north of east and were described as “Saxon”. Sixteen named churches, again plus “others”, were aligned between 10° and 15° north of east. Both these groups, according to Shore, were aligned “east-northeast”. Shore felt that this pointed either to a Celtic survival, in that this was the position of sunrise on May 1st 9 and these churches “were aligned with the sunrise on a continuation of the Celtic spring festival” [of Beltane]; or that they are part of “the Anglo-Saxon worship of the Virgin, the beginning of May being particularly dedicated to her” (1886, 108). He differentiated between these seventy-three churches, and those he described as “Norman”, thirteen of which he measured “to have an east-west alignment, which is their usual direction, or 9 In fact, the position of sunrise on May 1st at this latitude is 66° (24° north of east) see Appendix 2 p.324 Chapter 1 32 are built on a line south of east” and that “I have met with no church built entirely in Norman time, or begun in that age, which has an east-northeast orientation” (1886, 97). He felt that this therefore confirmed his assertion that the churches aligned eastnortheast were Saxon – and that east-northeast was the “usual line of orientation of a West Saxon church” (1886, 107). However, eighteen churches in Shore’s survey have also been surveyed as part of the survey of churches carried out for this thesis. Of these eighteen, only six alignment readings in the two surveys are within 3° of each other, six are between 4º and 10º apart, and the largest difference is 17º. The methodology for the measurements for this survey, outlined in Chapter Four, ensures consistency between readings at different churches, casting doubt on these particular measurements of Shore’s, and therefore perhaps, all of Shore’s readings. One of the ‘Saxon’ churches he described in the east-northeast group (20°-22.5º north of east) is in reality very close to due east – All Saints’ Minstead (85º) ̶ while another at Corhampton (no dedication), which he measured at “10°-15º north of east” (1886, 106), is actually aligned slightly south of east, at 92º. If the reason for the differences was that Shore did not allow for magnetic declination in his measurements, then all the readings would vary by the same amount and in the same direction, rather than by up to 9º to the south and up to 17º to the north, so it is most likely that his readings were affected either by iron in, or near, the walls, or were due to faulty equipment. Without publishing any details, he also concluded that “the usual explanation that the line of old churches is in the line of sunrise on the day of the saint to whom the church is dedicated, does not hold good in Hampshire” (1886, 107). Shore’s use of the term “usual explanation” here seems to point to a much wider knowledge of, and perhaps an on-going general discussion of, this issue. Since this study was the first published presentation of survey results in England, it strongly indicates that Wordsworth’s poem was not the only earlier reference to this topic, and that it was a subject that had been considered by a wider audience, but, without any earlier measurements with which to assess it and without any apparent published discussion, patronal-saint sunrise alignment appears to have been generally accepted as the norm. Chapter 1 33 Eeles F., 1913-14, ‘The orientation of Scottish Churches’, Proceedings of the Society of Antiquaries of Scotland 48, 169-183 Eeles published a survey of 62 churches in north-eastern Scotland (Aberdeenshire, Banffshire, and Morayshire), only 30 of which were medieval, or built on medieval foundations, as “so many of the ancient churches have been destroyed or rebuilt” (1913, 169). He referred to the orientation (here meaning east-facing) of churches as an “almost universally adopted practice”, and noted that there were exceptions “chiefly in Italy” (1913, 169) and “in the Scottish Episcopal Church in the eighteenth century [when] orientation was impossible under the Penal Laws” (1913, 174). He stated that his intention was “not to advance any theory … but merely to give exact orientation of a group of churches” (1913, 176), but he did mention the “theorising in England” about church alignment with the sunrise on the festival of the patron saint, and went on to say “This has been vehemently denied, and with good reason” (1913, 176). The sample of 62 churches varied in alignment between 55º and 110º, with a mean direction of approximately 83º. Almost two-thirds of them (61%) were aligned to the north of east, and while he measured eight churches as due east, he noted that three of these had “older stones” aligned to the north of east (1913, 182-183). The 30 medieval churches within the sample also ranged from 55-110º, with the same mean alignment (83º) while 63% of them were aligned north of east. He extracted twelve churches dedicated to three particular saints – Andrew, Moluoc and Drostan – and showed that churches dedicated to Andrew and Drostan varied by 45º (St. Andrew 65110º, St. Drostan 55-100º) and those dedicated to St. Moluoc ranged from 70º to 95º. In general terms he showed that there are many churches dedicated to saints with winter festivals that had alignments to the north of east – when the sunrise is far to the south of east – and “whilst it is not possible to prove a negative in all cases … we may safely say that we find nothing here to warrant the holding of the saint’s day sunrise theory” (1913, 180). Eeles briefly mentioned the variation in the position of magnetic north over time, but erroneously assumed that the variation was regular both in period and within a specific range; and he discussed the possibility of the use of a compass in the original building of the churches, noting that “this theory is not supported by the facts, even if we could believe that the compass was known and used here so long ago” (1913, 180). Chapter 1 34 Cave C., 1950, ‘The Orientation of Churches’, Antiquaries Journal 30, 47-51 The first of the modern surveys of church alignment was a compass survey of 642 churches conducted before and after the Second World War, which gave results for 633 churches. It appears to have covered much of the country to judge from the specific churches that were mentioned in the text. The alignments vary between 39° and 130°, with a mean of 86°. (The data presented in the article were amalgamated into five degree groups and the calculation of this mean is based on assumption that the original data were distributed evenly within each group). The survey included many urban churches, where Cave noted local influences on alignment, mentioning churches in Winchester and Chichester aligning with the Roman street pattern at 100-109° and 94100° respectively, although he said that this did not appear to apply in York where his measurements varied between 39 and 125° (similar to the author’s measurements noted in Chapter Five). He also noted that site restrictions had appeared to affect alignment, quoting effects on several larger churches, including Rievaulx Abbey which is orientated close to north-south because of its narrow site. This is confirmed by Anthony New, who states that “because of the steep slope, the church [Rievaulx] actually lies almost north-south”. (New 1985, 312) (Measured by the author at 163° True in June 2000) Cave also addressed the issue of alignment with sunrise on the patron-saint’s feastday. He used a subset of 151 churches dedicated to St Peter, St Andrew and All Saints taken from his survey and found that “not one church aligned with sunrise on its patronal-saint’s day” (1950, 48-49). For each of these saints, their patronal sunrise was outside the most extreme range of alignments. He also referred to suggestions that churches were aligned with sunset, rather than sunrise, on the feast day of their patronal saint. Again he found no evidence for this, “all the alignments for churches dedicated to St Peter fell south of sunset, whilst all the churches dedicated to St Andrew were aligned to the north of their feastday sunset” (1950, 50). He concluded that the “deviation from orientation due east” was “due to the direction of sunrise when the line of the foundations were first laid out” (1950, 50) – thus agreeing with Sir Henry Chauncy, writing two and a half centuries earlier. Cave identified 99 churches in his survey where the nave and chancel had different alignments, which he called “skewed”. The chancel was aligned to the north Chapter 1 35 of the nave in 56 cases, and 43 were aligned to the south. This aspect is considered in more detail in Chapter Five when the possibility that the different alignments represent religious symbolism is discussed. Cave felt that the “small percentage of churches with skew chancels and the way that the numbers fall off with increasing deviation shows … that the deviations were accidental, due to faulty laying-out” He went on to say that “there is nothing to show any symbolic meaning was in question” (1950, 55). Cave also stated that he “could not make his measurements agree with [those of Shore]” but it is not clear whether he meant measurements at specific churches or that his overall results were different. Cave assumed that this lack of agreement was “because he [Shore] was not aware of the magnetic field which is found in so many churches and which may cause the deviation of the compass by many degrees”. However, as already noted, Shore’s results are a little suspect, but more importantly, the results of this survey will show that there is considerable variation in mean church alignment across the country, so the fact that Cave’s measurements did not agree with those of Shore could easily have been because they were surveying in different areas. Benson H., 1956, ‘Church orientations and patronal festivals’, Antiquaries Journal 36, 205-213 Benson’s survey of 237 churches in Oxfordshire was undertaken on the premise that there was a definite link between church alignment and sunrise on the patronal-saint’s day. Specifically, he quoted Cave’s earlier (and three times larger) survey, as failing to find any churches that aligned with their patronal-saint’s sunrise, but suggested that this was because there are other feast days for the saints in question that Cave did not consider. In particular, Benson quoted two additional feast days for St Peter, 22nd February and 1st August; in addition to the near midsummer date of June 29th with “its extremely northerly sunrise” (1956, 206). He then found seven, out of 25, churches dedicated to St Peter that aligned with sunrise on these dates; so, even after including these additional feast days, eighteen of the 25 churches dedicated to St Peter in Oxfordshire still did not face a St Peter’s sunrise. He explained away the negative result by saying that an earlier church on the site would have “provided years of opportunities to sight the sun on the correct day” (1956, 206). This appears to ignore the fact that if there was an earlier church on the same site that was dedicated to the same saint, it would already be pointing in the right Chapter 1 36 direction, but it shifts the problem one step further back – how was the alignment of that earlier church established? If the new church was to be a rebuild and was to be dedicated to a different saint, perhaps at an early change of owner, how would the builders have ‘spent years’ looking in a different direction before knowing in which saint’s direction to look? The details of Benson’s results were used in such a way as to assume that churches aligned in the same direction, irrespective of their current dedication, were once dedicated to the same saint, which he termed “a cluster”; for example a St. Michael cluster for churches aligned at 96° (correct for sunrise at the end of September – St Michael’s day), with the assumption of, but no evidence for, a later change in dedication. To locate the exact sunrise point at each of the churches, Benson used details of horizon elevation and the shift in dates involved in the adoption of the Gregorian calendar in 1752 (outlined earlier) and came to the conclusion that 212 of his 237 churches faced sunrise on only eleven different dates, “nine of which happen to be festivals of the Holy Church – a very significant fact” (1956, 210). Many of the details of his argument do not stand up to close scrutiny. Apart from the 25 churches that do not face sunrise on a specific date, mentioned above (237 minus 212), which were “isolated cases, or difficult to measure” (1956, 210), and the 34 churches aligned with sunrise on the two dates that are not “festivals of the Holy Church”, the remaining dates of the festivals he identified are not necessarily principal ones; he uses February 22nd as a festival of St Peter (his chains), not mentioned at all in the calendar in The Oxford Dictionary of Saints, and as “Petrus in cathedra in Antochia” in Cheney (2000, 81); and he uses August 22nd as the “octave day of the Assumption” (one week later). Octave days were introduced after the seventh century for some saints’ days; “among the oldest being SS Peter and Paul, St Lawrence and St Agnes, … from the twelfth century the custom was extended to observing the days in between the first and eighth days. The number of feasts with Octaves was greatly increased in the Middle Ages; they were reduced, however, by the Breviary reforms of Pope Pius V [after 1566]” (Cross 1957, 974-975). At best then, February 22nd was only a minor festival and less likely to be celebrated in place of the main feastday of St Peter on June 29th. If the intention was to align the church with the saint’s feastday, why not align it with sunrise on the main feast in June? Similarly, the Octave of the Assumption, on August 22nd, Chapter 1 37 was not only unlikely to have been a celebrated event when the churches in question were first being built and their alignments fixed, but why was the church not aligned 4º more northerly, in other words one week earlier, to align with sunrise on the day of the Assumption itself? This example particularly, appears to be a case of stretching the argument to breaking point, when the real focus, if there actually was one, was so close by. Benson’s limited ‘eleven festival sunrise dates’ are between seven and fourteen days apart, therefore he was effectively adding together churches into 5º or 10° groups, as the sunrise point is moving along the horizon at almost 5º per week around the time of the equinoxes. This amalgamation of results runs counter to the whole tenet of his paper which extols the need for accuracy by taking detailed note of calendar changes and horizon elevations. Further examination of some of the other details in Benson’s study highlights other areas of concern; in particular, the data he quoted for churches dedicated to St Peter where Cave “failed to identify alignments” (noted above). Of the four churches that Benson stated faced sunrise on the minor festival of St Peter on 22nd February (1956, 206-207), three were measured as part of this thesis. Drayton St Peter, South Newington St James and Wilcote St Peter are all aligned at 103º True, which is close to the expected level-horizon position of sunrise on 22nd February (details of sunrises shown in Appendix 2). However, the church at Wilcote faces an eastern horizon elevated by 1º, South Newington faces a horizon elevated by 4º and Drayton’s is elevated by 8º. This has the effect of delaying the sunrise at each of these sites, as shown in table 1.1 and Figure 1.5 below. This means that the church at Wilcote actually faces sunrise on 27th February (rather than 22nd February), South Newington faces sunrise on March 8th and on March 16th at Drayton (three weeks after the St Peter’s feastday), effectively removing all three from Benson’s “St Peter cluster”. In order for these three churches to be aligned with the rising sun on February 22nd over their elevated horizons, they would need to be aligned at approximately 105º, 109º and 115º respectively (Drayton’s sunrise details are shown in Figure 1.5). Notice also that the church at South Newington (dedicated to St James) is only in Benson’s group because of its similar alignment to the others – Benson offered no evidence for any rededication from St Peter, other than its alignment – perhaps St Felix (March 8th) would have been a more applicable dedication here. Chapter 1 38 Table 1.1 Benson’s survey - sunrise and St Peter Drayton St Peter S. Newington St James Wilcote St Peter Church Alignment Sunrise date – level horizon Actual Horizon elevation Delay of sunrise (to nearest degree) Actual sunrise date (at 103°) Alignment required for 22nd Feb sunrise (to nearest degree) 103º 22nd Feb 8º 12º 16th Mar 115º 103º 22nd Feb 4º 6º 8th Mar 109º 103º 22nd Feb 1º 2º 27th Feb 105º Figure 1.5 – Sunrise movement and elevated horizon at St Peter’s, Drayton, Oxfordshire – aligned with sunrise three weeks later due to the elevated horizon Benson dealt in depth with churches that had naves and chancels with different alignments – which he called “crooked churches” – and he treated them as an extension to his patronal-saint alignment argument. He explained the difference in alignment between nave and chancel by the movement of sunrise due to calendar change and that a later rebuild, particularly when chancels were lengthened or rebuilt after the Fourth Lateran Council of 1215, was still aligned with the patronal-saint sunrise, which had moved slightly along the horizon. He suggested that feast days in the first half of the Chapter 1 39 year, when sunrise is moving northwards from the extreme south of midwinter sunrise, require a sunrise correction to the north and vice versa. Where this idea did not fit, as with the three churches where the nave and chancel are misaligned “by 4° or 5º ”, Benson acknowledged that for these churches this was “far too much for the requirements of the Julian calendar [shift]”, and concluded that “this must indicate a change of dedication” (1956, 212) – another example of fitting the data to the assumption. Searle S., ‘The Church points the way’, New Scientist, 3rd Jan 1974, 10-13 Searle was the first to consider seriously the possibility that churches had been set out by magnetic compass, and that the known movement of magnetic north over time explained the variation in church alignment. His sample was very small and included only nineteen churches, but he concluded that there was a direct correlation between date of building and the alignment of the church reflecting the changes in the position of magnetic north. Unfortunately, Searle dated four of these churches because they fitted in a particular place on his alignment curve – “author’s dating, within the century the church was said to have been built, as indicated by its orientation” (1974, 11, table 1 footnote) – thus using a circular argument to position them. The small sample of churches only had a range of alignments of 23° (with magnetic north between 5° west and 18° east of true north) rather than the range of 90° difference in alignments observed in the larger surveys. Magnetic variation is discussed in greater detail in later chapters, but the lack of scientific rigour in Searle’s study means that his conclusions can be safely set aside. Davies R., 1984, ‘Church Orientation in Rutland’, Rutland Record Vol. 4, 142-143 In 1984, Davies, with sixth form students at Oakham School, undertook a compass survey of the 46 churches in Rutland. Converting his magnetic survey results to true bearings, by deducting the local declination at the time (6°)10, the churches ranged between 57° and 108°, with a mean of 85°. His starting hypothesis was that the alignment of the chancel coincided with sunrise on the saint’s day to which the church is dedicated, but he found “no correlation” in the results. 10 http://geolab.nrcan.gc.ca/geomag/e_cgrf.html (accessed 15th April 2002) Chapter 1 40 Abrahamsen N., 1992, ‘Evidence for Church Orientation by Magnetic Compass in Twelfth Century Denmark’, Archaeometry 34 (2), 293-303 Abrahamsen’s survey included 572 twelfth-century churches in Denmark. His particular interest was magnetic changes, and initially the survey was conducted from early maps. A sample of 204 of these churches in two districts was measured on the ground using a compass. The results for the different areas are shown in table 1.2 below. Abrahamsen discounted the accuracy of the measurements from maps due to the magnetic bias of many of the eighteenth-century maps, which confused magnetic and true North and appeared to rotate the churches clockwise, in other words, align them even further to the south of east. He mentioned three studies in Germany and Denmark carried out in the nineteenth and early twentieth centuries which proposed that churches were aligned with patronal-saint sunrise, but found that “typical results are a normal distribution, symmetrical around 0º ” this would be “hard to explain, as saints days are scattered all over the year and no biased orientation would result” (1992, 300) Much of the article is concerned with trying to extricate statistical patterns from the alignment data. In the two areas he measured on the ground, Thisted and Aarhus, there appeared to be two different underlying patterns – in Thisted particularly – which he suggested could have been caused by different methods of setting out. The histogram for the alignment of churches in Aarhus appears to be far more regular – almost a bell-curve. He found that the simplest explanation for the clockwise rotation of the alignments “appears when comparison is made with the general magnetic declination in the Danish area, being systematically east of north between 1000 CE and 1600 CE” (1992, 301). He noted “that the mean direction of [English] churches of all periods was approximately 85º”, and wondered whether “[the] UK magnetic declination of a few degrees west during the fourteenth century provided a possible magnetic explanation for the difference in orientation from the Danish churches” (1992, 294). Magnetic changes and the effect that this might have had on church alignment are discussed later in Chapter Six. What Abrahamsen did not take into account, however, was that during the eleventh, twelfth and thirteenth centuries magnetic declination in England was to the east of north (Clark et al. 1988, 659; Merril et al. 1996, 3), the same as in Denmark, though slightly less extreme, so is unlikely to explain the difference. Chapter 1 41 He concluded that “some churches were probably laid out by sun or stars … … the remaining strongly rotated group may have been laid out using a magnetic compass” (1992, 302). He felt that his results “indicated a fairly common use of the magnetic compass soon after its appearance in Europe” (1992, 303). Table 1.2 – Summary of Abrahamsen’s Danish survey results Measured on the ground Area name Number Thisted 88 Aarhus 116 Mean 204 Measured from Maps Mean 368 OVERALL Range 78-120° 68-116° Mean 97.3° 91.8° 94.2°° 58-140° 98.8° 26.1 97.2° 28.8 572 % N of East 25.0 40.5 33.8 Dymond D., 1997, ‘Churches and Churchyards’, Historical Atlas of Suffolk, D. Dymond (ed.), Suffolk County Council, 54 &197 Dymond undertook a survey of 23 churches in the Thedwastre Deanery in central West Suffolk, measured from 1/2500 scale maps, using a “best-fit” line to represent the alignment. The alignments ranged between 69° and 103°, with a mean of 86°. Sixteen were north of east and seven south of east. Little variation was noted in alignment between churches of different sizes, or in those that had different floorplans. Similarly, there was little difference between the alignments of churches that were mentioned in Domesday (85% of the total) and those not mentioned: 85.9° and 86.5° respectively. Hoare P. & Sweet C., 2000, ‘The Orientation of Early Medieval Churches in England’, Journal of Historical Geography 26 (2), 162-173 Hoare and Sweet’s survey, consisting of 183 churches with substantial elements of visible Saxon or Saxo-Norman fabric identified in The Taylors’ Anglo-Saxon Architecture, produces a mean alignment of 88°, with individual churches ranging between 42º and 128º. The results were analysed by Saxon sub-periods and indicated similar results irrespective of period, from the earliest ‘A’ period (600-800 A.D.) at 88°, to 87° for the latest, Saxo-Norman, period (1066-1116). This result firmly shows that any variation in the alignment of churches is not related to the period in which they Chapter 1 42 were built. Their mean result was approximately 2° above the mean of 86° from Cave’s larger survey, but their sample excluded large areas of the west of the country which have no extant Saxon buildings (2000, 165), areas which, as the analysis of this survey will show, contain churches which exhibit a numerically lower mean alignment. They discuss the possible use of the compass in the setting out of the churches, but discount it on the basis that magnetic east was well to the south of true east during the Saxon period, and this would have resulted in churches aligned considerably further south than they are (2000, 167). They also conclude that the “time-honoured support for the various sunrise models is misplaced (although an occasional building may have been so aligned)” (2000, 168). There were few churches with different nave and chancel alignments measured in this survey – which they called “crooked” – as the survey was only concerned with elements of surviving Saxon fabric, and few churches had both a Saxon chancel and nave. The few that did have both were dealt with by taking an average reading for both parts if they fell into the “crooked” category. Ali J., & Cunich P., 2001, ‘The Orientation of Churches: Some New Evidence’, Antiquaries Journal 81, 155-193 Ali and Cunich’s survey investigates the alignment of 143 large parish and Monastic churches with known dedications that were built between the mid eleventh and late twelfth centuries. They conclude that 33 (almost a quarter) of these churches were aligned with sunrise on the day of their patronal saint. However, they also conclude that 28 others were aligned with their patronal-saint’s day sunset; 37 were aligned with Easter sunrise; 25 were orientated east-west; twelve were aligned with magnetic east; five were aligned for topographical reasons; 31 were aligned with Julian calendar sunrises or sunsets and only 20 of the sample of 143 were unclassifiable. This would seem almost to be a return to the 1960s, when alignments were ‘discovered’ in every direction (such as at Stonehenge – see Chapter Three), and importance was given to the fact that an alignment match existed, without investigating whether it had any significance, or had occurred through chance. No explanation was offered as to why different church builders might have employed so many different alignment formulae. Chapter 1 43 Muirden J., 2003-5, ‘Crooked Churches and Saintly Sunrises’, Church Archaeology 7-9, 33-43 (published 2005) No church alignments were included in the article for the 194 churches surveyed, the data had been converted to sunrise dates; but an unpublished interim version of the survey shows data from the first 49 churches surveyed in the range 46° – 123º, although there are no rural churches with alignments to the north of 52º, the two churches aligned at 46º being located in Exeter11, apparently aligned with the Roman street pattern. Muirden adopts Benson’s earlier methods [of deducing the dedication of the church by the sunrise direction it faces] “because his results were so impressive” (2005, 33). Since “fewer than one-fifth of Devon churches have a dedication record pre-dating 1300, the number that can be traced back to Norman times is tiny”, he states the subject has to be tackled from the “opposite direction” (2005, 35). This involved “determining their alignment sunrise data, to see if [this] suggests any favoured festivals” (2005, 35). According to the table of results, (2005, 35) none of the 194 churches surveyed faces sunrise between May 10th and August 4th, or between November 25th and January 18th, a total of 20 weeks, amounting to almost 40% of the year. During these 20 weeks, there are only six days when no principal saint’s feast is celebrated – July 12th, Nov 26th, December 2nd, 19th and 20th and January 11th (as recorded in the feastday calendar in Farmer 1987, 474-478). This period is when the sunrise is closest to north-east or south-east, rather than due east, and includes the major feastdays of several of the most popular current dedications – John the Baptist, Peter, Peter and Paul, Margaret, Andrew, Nicholas and John the Evangelist. If the theory of patronal festival sunrise alignment was true, then these major festivals should have at least some churches aligned with their sunrise. All of this also assumes that churches were originally dedicated to a specific saint when they were first built, as was discussed earlier in this chapter. Muirden deals with churches with different nave and chancel alignments – which he calls “crooked” – in detail. There are 53 such churches in Devon and he extends the proposal that churches face their patronal-saint’s sunrise in order to explain such churches by suggesting that differences in alignment between nave and chancel reflect calendar drift during the period between the building dates of the two parts of 11 Pers Comm. James Muirden - Interim version of the survey received July 2003 Chapter 1 44 the church; usually a later chancel built on a new alignment to maintain its alignment with the shifted sunrise of the original patronal festival. In nine cases this proposal was stretched to the extreme to include churches where the nave and chancel walls were built parallel to each other, but the east wall of the chancel was not built at rightangles to the north and south walls (2005, 36-37). It was assumed in these cases that a line at rightangles to the east wall of the chancel was aimed at the shifted sunrise. He identifies a concentration of churches “that share a popular sunrise window” (2005, 41), and he notes that Benson found a similar group “for which he presented a persuasive argument … that they faced sunrise on the Octave day of the Assumption, though there seems to be no discernible reason why this festival should attract more reverence than the Assumption itself” (2005, 41). His final conclusion is that “If churches were carefully aligned on their patronal sunrise, then analysis … could throw light on matters beyond the scope of present orthodox research, since their original patron saint and approximate date of foundation could be deduced” (2005, 41). Ali, J. & Cunich, P., 2005, ‘The Church East and West: Orienting the Queen Anne Churches 1711-34’, Journal of the Society of Architectural Historians 64, 56-73 This article starts by reiterating the statistical significance of the patronal-saint sunrise alignments, as well as the Easter Sunday and patronal-saint sunset alignments, discussed in their earlier article on large churches (Ali & Cunich 2001, 155-193). The basis for this study was eighteen eighteenth-century churches which “the Church Commissioners and [major architects] thought ought to be aligned east-west”, but whose actual alignments range between 57º and 115º. Their investigation of individual sites and church dedications shows that “sun control can be largely discarded as the principal control … for [the alignment of] Queen Anne Act churches” (2005, 66). Sir Edmond Halley was appointed to the Church Commissioners in 1712, and the authors thought it “inconceivable that he would not have offered the practical solution of declination-corrected compass bearings to determine the geographic axes of the church sites” (2005, 67). Their conclusion, for these eighteen churches, which vary in alignment by up to 58º, is that “in all cases an easterly alignment was achieved”. Further, that “there is a considerable gap in our knowledge of the underlying architectural precepts in common usage during the [medieval] period. That such knowledge had already been lost when Wren commenced his rebuilding … seems Chapter 1 45 almost certain” (2005, 70). As can be seen from the results of the survey of Victorian churches shown in Appendix 1 of this thesis, the overall alignments of medieval churches and those of the Victorian era are, apart from a small number of churches, very similar, so it is difficult to see what knowledge had been lost. One of their final conclusions is that “It has always been assumed that an eastwest alignment was important in church building, we now have quantitative information with which to amplify that assumption” (2005, 71). This does seem a little presumptuous from a sample of eighteen churches whose alignments vary by such a large amount. Wall J., 2006, ‘Church Orientation’, Bulletin of the British Sundial Association 18(i), 16-17 As might be expected from the name of the journal in which it was published, this was not a survey of churches, but the article refers to seventeenth- and eighteenth-century sundials, particularly those that are set at an angle to the church wall (canted-out), or had offset (or declining) gnomons, in order to compensate for the poor east-west alignment of the churches involved. Whether or not this was likely to be due to an improved ability to measure east accurately at the time of the creation of the sundial was not explored. Figure 1.6 – Sundials compensating for the poor east-west alignment of the church “canted-out” at St Mary’s, Gilcrux, Cumbria (church @ 64°) (left), “offset gnomon” at St John’s, Ickham, Kent (church @ 108°) (right) Chapter 1 46 Table 1.3 – Summary results for each of the surveys discussed in this chapter SURVEYS in U.K. Shore Eeles Cave Benson Searle Davies Dymond Hoare & Sweet (Saxon) Ali & Cunich (Monastic) Ali & Cunich (18th C) Muirden Location Hampshire Scotland England Oxfordshire Various Rutland Suffolk England England London Devon 62 63 72 85°° 86°° 88°° 90°° 87º 86º 65 69 60 51 63 58 39-130° 86.0°° 63 1,926 38-128º 86.1º 63 88 116 204 78-120° 68-116° 97° 92° 94° 25 40 34 See text 55-110º 39-130° 45-117° See text 57-108° 69-103° 42-128° 49-163° 57-115º 46-123º 1,386 This survey Thisted Aarhus Mean Chapter 1 83º (86°°) 83°° Number of churches 90+ 62 633 229 19 46 22 183 143 18 49(see text) TOTAL Surveys elsewhere Abrahamsen (Denmark) Mean % aligned North of East 47 Range CHURCH ALIGNMENT STUDIES: CONCLUSIONS In terms of the data they employ, some of the earlier alignment studies outlined here are better than others. Some leave accuracy to be desired and others seem to have allowed the stated aims of the work to influence their conclusions. As an overall body of work however, the results seem to indicate that the raw data have real value as the sample size grows, in that the bottom-line results of all the surveys added together exactly match those of the survey for this thesis, with an identical range of alignments, the same mean value and the same proportion of churches aligned to the north of east. What all these surveys do seem to confirm is that there was a general focus for church alignment which was close to east, as each of the studies exhibits a mean value a few degrees to the north of east (bar one, which was exactly east). The details will be more closely examined in later chapters. Wordsworth’s poem about sunrise on the saint’s day at Rydal church has been used as the proof that the method must have been used before, and has given credence to the theory, which seems to have convinced some writers that it must have applied at most, if not all, churches in all periods, despite the obvious evidence to the contrary and Wordsworth’s own admission about his use of poetic licence. The fact that some churches do face sunrise on the patronal-saint’s feastday does not mean that they were actually meant to. Some churches are bound to face sunrise on their patronal-saint’s day, purely by chance, since only eight (0.4%) of the 1,926 churches in this survey are aligned outside the sunrise arc during the year for their location. With a maximum of only 80º between midwinter and midsummer sunrises in much of England, and almost 2000 churches surveyed for this thesis, it would statistically remarkable if none at all faced the relevant sunrise. Benson’s work appears to be founded on minute accuracies involving Julian date alterations; discussions of whether the upper limb, the half orb or full orb of the sun should be used for assessing the exact moment of sunrise; and the elevation of the horizon. However, he then proceeded to generalize the results by amalgamating the data into 5º or 10º groups and concluded that almost all of the 229 churches that he surveyed face sunrise on only eleven specific days, most of which conveniently correspond to particular saints’ feastdays. As has been shown, some of these dates are at best doubtful as celebrated festivals, and others seem to ignore the impact of the Chapter 1 48 elevation of the horizon which had been so carefully considered earlier in the article. Muirden supports the majority of Benson’s methods and conclusions and criticises others for “wholesale methods of analysis [which] may have served to conceal, rather than reveal, the case for patronal saint sunrise” (2005, 41), but analysis to substantiate an argument of this sort, which both Benson and Muirden claim applies to most churches, requires the broader picture, rather than just the investigation of a small sample of individual churches that happen to fit the criteria. Identifying a group of churches that face the same direction, but are dedicated to different saints, and concluding that they must have originally been dedicated to the same saint because of their alignment, seems little better than the circular argument used by Searle, noted earlier, to date churches by their alignment, and is one way of ensuring that the proposal that churches faced their patronal-saint sunrise is true. In fact, based on the results of all the other surveys, it is the only way that it can be true; patronal-saint sunrise alignment will always be very difficult to prove when the results of the other surveys noted here have shown, and the results of this survey will show, that the great majority of churches are aligned close to east. It is just possible, as Muirden argues (2005, 39-41), that “crooked” churches represent realignments in order to continue to align with a specific sunrise as the calendar shifted in relation to the sun, and those misaligned to the right are dedicated to summer saints, and those to the left to winter saints. But since this idea is based on dedications that can often no longer be traced and that the whole idea of saint’s day sunrise alignment does not appear to be supportable, then the simpler idea that this pattern represents a desire to face closer to east seems far more attractive and will be shown by the results in this survey in Chapter Five to be far more likely. It also ignores the obvious fact that some of the churches that have chancels that are misaligned to the right (which according to Muirden must have been dedications to summer saints) are aligned well south of east and therefore aligned towards sunrise during the winter months, so the realignment is in the wrong direction taking it even further away from the sunrise of a summer saint. Similar cases exist of churches with chancels misaligned to the left, therefore dedicated to winter saints according to Muirden, but are aligned towards summer sunrises, so are also realigned the wrong way. Abrahamsen’s reference to the differences in magnetic declination between England and Denmark cannot explain the difference in the alignment of English and Chapter 1 49 Danish churches. He referred to the westward declination in England during the fourteenth century, but this was only for a very short period, and for at least 500 years before this magnetic declination in England was to the east of north, the same as in Denmark, though slightly less extreme. The fact that the results of the surveys noted here are very similar in the patterns of alignment that they exhibit has led the majority of the authors to similar conclusions – that the general focus for church alignment was just north of east and that some of the theories suggesting specific targets for church alignment could not be supported. The similarity of their results also lends weight to each of them in that none is glaringly different, although there are minor differences between them in different areas of the country which will be explained by the analysis of this much larger survey in later chapters. Two of the authors, Benson and Muirden, have used their data to arrive at a very different conclusion – that patronal-saint sunrise alignment is not only supportable, but probable – this will be examined more closely in the analysis of this survey’s results in later chapters. Chapter 1 50 CHAPTER TWO THE HISTORICAL USE OF ALIGNMENT Each generation has the Stonehenge it deserves – or desires (Jacquetta Hawkes 1967, 174) INTRODUCTION This chapter outlines the issues that will be considered in more detail later when applying the results of this survey to the alignment, and particularly the location, of churches. It considers the background to the use of alignment generally in history and prehistory by identifying the objects and features that have been assumed in the past to have had intentional alignments and explores the significance that has been placed on them. Alignment appears to have been part of most aspects of prehistoric life; from the houses that people lived in, the monuments associated with the ceremonial and ritual aspects of their lives, and even after death – from the tombs or memorials in which they were buried to the specific ways that their bodies were laid out in the ground. Not all of these aspects necessarily applied at the same time, but examples can be found from the Palaeolithic period, through the Neolithic, Bronze and Iron Ages, the Roman and post-Roman periods, and right up to modern times, especially in relation to death, burial and worship. Alignment has also been recently considered as part of some fringe mystical aspects of modern life, such as ley lines and orthotonies (lines along which UFOs were believed to appear). Jacquetta Hawkes’ comment, quoted above, could probably apply to many other monuments as well, but at Stonehenge, interpretations have ranged from a Pagan Temple – Aubrey (1693), a Druidic temple – Toland (1726), a solar observatory – Lockyer (1909) and Somerville (1927), a hub for ley-lines – Watkins (1925) to a lunar eclipse predictor or computer (amongst other things) – Hawkins (1966) and Thom (1967), or even the possibility that it was just built as an offering to the Gods and had no practical function – Johnson (2008, 254). Chapter 2 51 Research work is always open to reinterpretation, especially when new evidence is discovered; when research methods change and improve; when new methods of analysis are invented or developed; or the focus of investigations changes. It is also open to reinterpretation if faulty research methods or assumptions are made, or if relevant data are ignored because they do not fit the original presumption. Some of the Stonehenge interpretations noted above have been disproved or altered by later researchers by tightening research criteria and exploring all aspects, even those that do not apparently fit the required pattern. These last aspects were used particularly by Williamson and Bellamy (1983), Ruggles (1999) and Burl (2000), when examining aspects of previous work in their area of study which will be covered in more detail below. ALIGNMENTS In the following sections, each type of subject noted above – houses, monuments, tombs and burials – will be considered in turn, roughly chronologically within each. In some instances the alignments are towards objects in the sky, such as the sun, moon and stars; in others the alignments are towards other objects on the ground; and in other cases it is a combination of both, for example in the situation when the moment of sunset combines with a topographical feature on the horizon. Since in each of these cases it is the alignment itself which appears to have been important to the builder, they are all discussed together. 1) Houses In the Bronze Age and later, most houses were circular, the majority having their entrances facing eastward, mainly due east, perhaps evoking a cosmological metaphor for the house and the universe (Bewley 2003, 88; Parker Pearson 1993b), where east was associated with light, warmth, life, and good, while west was associated with the opposite in each case (Parker Pearson & Richards 1994, 50). From a practical point of view, however, a southeast-facing doorway, rather than an east-facing one, would allow in far more midwinter light at a time of year when light was short-lived. It has also been suggested that an east-facing door was used to face away from prevailing southwesterly winds (Bewley 2003, 88), also proposed by Pope, with a sample of 690 prehistoric houses from 253 sites in north and central Britain (Pope 2008, 19-20) where Chapter 2 52 the large majority have entrances facing eastwards, which was attributed to a combination of the maximization of light and winter shelter from the wind. Pope felt that houses that did not face this way in the survey were revealing the importance of local topography in those cases (Pope 2008, 20). 2) Monuments Prior to the appearance of cursuses there is little to suggest that Neolithic builders were concerned with precise layout at all – “houses, barrows and causewayed enclosures were all less than regular and displayed little sense of alignment” (Loveday 2006, 114). Good indications of this are the Neolithic ‘Long mortuary enclosures’, of which Loveday shows 42 examples in a diagram, distributed across much of the Midlands and East Anglia, which are aligned towards every point of the compass (Loveday 2006, 5558). However, recent work at the Ness of Brodgar on Mainland, Orkney has discovered elements of alignment involving a large stone building located between the Stones of Stenness and the Ring of Brodgar which “is aligned on the Maes Howe cairn” (Catling 2009, 6) although, as Maes Howe is almost due east of the building, it cannot be certain which was the target – Maes Howe or east – but there does appear to have been a purposeful alignment. Cursuses, by virtue of the nature of their linearity, possess an obvious capacity for alignment with distant features. Unlike long barrows, however, the ‘business end’ is more difficult to establish, so there are “no indications as to which of the two opposing alignments was considered important” (Loveday 2006, 132). In addition, they were overwhelmingly constructed in flat gravel landscapes “where skyline features were few and decidedly unimpressive … but where it is certain some stars will fall into line” (Loveday 2006, 132). Several cursuses are associated with rivers, particularly river confluences (Loveday 2006, 133). Excavating at Maxey in Cambridge, Francis Pryor found the cursus alignment there was manifestly illogical, aligned diagonally across an island which was later more practically divided by field boundaries at right angles to the waters edge; perhaps it was originally aimed at the River Welland, but the end was obliterated by the movement of the river (Pryor et al. 1985, 17-21). There has also been discussion about the length of time that cursuses were in use and it is difficult to tell when they were superseded. When henges were built, they were “usually” located some distance from cursuses (Harding 2003, 89-90), but at Dorchester a henge was built Chapter 2 53 alongside the cursus (Harding 2003, 89), while at Maxey the henge there was built over the cursus (Prior et al. 1985, fig 40), which presumably prevented its subsequent use completely. Loveday refers to the heavens as “a bran tub from which orientation can always be plucked, if the rising points of the sun and moon are too constrained, there are myriad stars to choose from” (Loveday 2006, 137). He goes on to say that claims for the Dorset cursus as an astronomical observatory (by Penny & Wood in 1973) are difficult to accept, as some elements of their observatory were not fixed. In addition, their claims for meaningful alignments for one part of the cursus were not reflected in the other arms of the same complex, which had no significant alignments at all (Loveday 2006, 137). There are other cursuses which are close to significant solar alignments, but “wilfully miss them” (Loveday 2006, 138), for example “the western section of the Dorchester cursus aligns broadly with midsummer sunset/midwinter sunrise, but it is awkwardly realigned halfway along its length, which is in itself odd, as the builders seem to have an impressive ability to build lengthy straight lines” (Loveday 2006, 138), although Barclay and Harding note that on occasion cursuses have only one ditch that is straight, the other meandering slightly (1999, 2-3). Other cursuses, at Maxey, Fornham, Buscot and Scorton are all aligned in a similar direction, roughly northwest/southeast, but also “miss the significant midsummer sunset/midwinter sunrise solar alignment, and the adjacent moonset alignments, by a few degrees” (Loveday 2006, 138). There are also a significant number of cursuses aligned north/south, for example at Rudston in East Yorkshire (Harding 1999, Fig 3.1 page 31), both of which are points where neither sun nor moon ever rise or set so make unlikely alignment focuses. There is a possibility that all these sites which are not aligned with the sun or moon are aligned towards a star, or group of stars, but since it is very difficult to see even the brightest star rise (Loveday 2006, 139), a point that is also emphasised by Ruggles (1999, 52) and Burl (2000, 205), a stellar alignment has to reflect a star’s transit rather than its rising or setting, which basically means that such an alignment could point to a wide arc of the sky whilst the star or group is above the horizon, making conclusions about the target impossible. Another difficulty is that due to polar motion, the position of the stars in the sky changes, varying by up to 22º in 1500 years (Ruggles 1999, 52; Burl 2000, 24), thus altering the night sky completely. Chapter 2 54 Loveday finally downplays the whole idea of intentional alignment of cursuses and avenues by concluding that “The fact that cursuses really were the first monuments to override regional patterns of construction and adhere to a type recognizable from Devon to Dumfries is of hugely greater significance than ultimately irresolvable questions about how precisely they were used” (Loveday 2006, 143). 3) Megalithic Monuments Megalithic monuments are the features about which most has been written on the subject of alignments. The comments by Jacquetta Hawkes concerning Stonehenge, quoted at the beginning of this chapter, are mirrored by an assessment by Aubrey Burl of the descriptions of the monument at Calanais on the island of Lewis in the Hebrides, with similar uses and properties ascribed to the monument by various writers over the last 300 years to those at Stonehenge (Burl 2000, 202). Many, if not most, of the alignments described as significant by writers such as Gerald Hawkins in Stonehenge Decoded (first published in 1965) and Alexander Thom on ‘Megalithic Astronomy’ during the 1960s (Thom 1967 and 1971) have been reassessed, and largely dismissed, by Clive Ruggles in 1999 in Prehistoric Astronomy in Britain and Ireland, and Aubrey Burl in the 2000 revised edition of The Stone Circles of Britain, Ireland and Brittany. Burl points out that astronomy is one of the most contentious aspects of the study of stone circles, and in the absence of evidence, modern man is denied “the likelihood of rediscovering the cosmology of the early people” (Burl 2000, 117-118), something which Ruggles refers to as a “search for lost knowledge” (Ruggles 1999, 3). Ruggles suggests that the difference of opinion between Archaeologists and Astronomers about the meaning of monuments arose because of “parallel but very different views of the same monument” and as simply seeing their own reflection in the past (Ruggles 1999, 8). During a prolonged discussion since the 1970s, each side has “largely argued past the other, starting from different tenets, addressing different aspects of evidence, using different methods, and has come to very different conclusions” (Ruggles 1999, 9). Gerald Hawkins arrived at two major conclusions in Stonehenge Decoded. Having identified several sun-based alignments, he wondered what other alignments might exist. To do this he decided that “we need the machine” (Hawkins 1970, 127- Chapter 2 55 128). Using an early IBM mainframe computer, a total of 120 pairs of points were assessed for azimuths and declination (1970, 136-137). This resulted in multiple alignments towards sun rise/set and towards the rising and setting moon in its extreme positions (1970, 140), but even with the computer, only fifteen of the eighteen possible alignment positions were identified (1970, 171). Figure 2.1 – Twelve of the eighteen major sun and moon positions (Ruggles 1999, 37) Figure 2.2 – The 18.6 year lunar cycle (Ruggles 1999, 36) Chapter 2 56 The second conclusion was that Stonehenge was used as an eclipse predictor. Hawkins wrote “There can be no doubt that Stonehenge was an observatory; the impartial mathematics of probability and the celestial sphere are on my side” (Hawkins 1970, preface), but as Ruggles demonstrates, errors in Hawkins’ mathematics reduced the chances of Stonehenge being an eclipse predictor to worse than evens (Ruggles 1999, 42-43). Due to the way the moon moves around the Earth, it can only be eclipsed when it is opposite the sun, in other words at full moon, and only then in the years when it is at the extreme ends of the range of rising and setting (when it is on the same plane as the Earth’s orbit around the sun, known as the ecliptic) (1970, 176). According to Hawkins, the 18.61 year cycle of movement of the rising and setting positions of the moon is most accurately made up of three figures – 19, 19 and 18 years – totalling 56, the number of Aubrey holes at Stonehenge, and that by the movement of six stones around the Aubrey holes in a certain way, Stonehenge “could have predicted – quite accurately – every important moon event for hundreds of years” (1970, 178). Alexander Thom proposed a consistent unit of length – the megalithic yard (2.72 feet) – which he said was in use in the Neolithic and Bronze Ages for the erection of stone monuments over a wide area (Thom 1967, 34-55). He also coined the term ‘Megalithic Astronomy’ which involved the use of stone circles as indicators of sun, moon or star rising and setting points. Apart from detailed claims that they were used for identifying the solstices and the major and minor limits in the rising and setting of the moon, he claimed to identify other functions at specific sites. For example, the Nine Maidens stone circle at Stanton Drew in Oxfordshire apparently marks a complete sequence of star risings and settings in the early morning hours at midwinter, “when any community wants to have a method of telling the time” (Thom 1967, 105), perhaps a rather overwhelming solution to such a small problem. The two principal problems identified by Burl and Ruggles about both of these works are those of spurious accuracy and the selective use of data, each of which is expanded below. In addition, they have noted other problems with interpretations of megalithic sites, such as post hoc justification, circular argument, and the tendency to emphasize those data which confirm a pre-conceived set of ideas while ignoring those that do not (Ruggles 1999, 41; Burl 2000, 63). Chapter 2 57 Spurious Accuracy The frequently identified connection between chambered tombs and the sun’s penetration on specific days is often counted as accurate to a few fractions of a degree by Thom, but in reality, is rarely accurate to better than 5º (Burl 2000, 60). Thom also favoured alignments using stones which indicated “slivers of the sun tracing hill slopes, or momentarily appearing in notches” at the solstices in 2000 BCE, but Burl is convinced that this level of accuracy is not possible using rough stones as markers, and felt that it is difficult to measure alignments from weathered stones to better than 2º (Burl 2000, 60) and much of Thom’s survey data is accurate to within one minute of arc, which is “so exact that it would take 180 years to collect the data to position the marker in the first place” (Ruggles 1999, 52). Some of this accuracy is further compromised by the fact that recent evidence about atmospheric scattering of light, particularly near the horizon, shows it to be more severe than Thom realised – which “make his stellar alignments particularly problematic” (Ruggles 1999, 50-51). The moon has particularly complex changes in its rising and setting positions over an 18.61 year cycle. Added to this complexity is the fact that many of these events occur in daytime, when the rising and setting could not be not seen, so it would take at least double the length of this cycle, in other words 37.2 years, to confirm the full details, with 13,577 risings and settings to record, with absolutely no evidence that records were ever kept (Burl 2000, 60). Even Hawkins acknowledged the difficulty that the Babylonians had with eclipse predictions, showing little success until a thousand years after Stonehenge (Hawkins 1970, 175). Selective use of data The selective use of data has been highlighted many times in relation to the work of Hawkins and Thom. Ruggles and Burl have both identified inconsistent treatment of similar monuments and the apparent working backwards from conclusions to determine where to survey in other cases. If specific aspects, such as stone outliers, were used to mark alignments, and there was communication between the various builders, which Thom insisted was present by the use of consistent measurement units (his megalithic yard (Thom 1967, 34-55)), it seems reasonable to assume that there would also have been a consistent approach to the use of the monuments. However, Thom proposed that of eighteen stone Chapter 2 58 circles in Cumbria, only nine had an astronomical use – four for the sun and the remainder had targets including the moon and several stars. So, not only did nine of the monuments produce no alignments at all, those that did were completely inconsistent in their application. The specific alignments at these nine sites were formed by outliers (three times), stone to stone across the circle (four times), circle to circle (ten times) and centre to entrance (twice) (Burl 2000, 117-118). In addition to the selective use of different markers at each of the sites, many of these make unsatisfactory foresights. Burl cites problems with outliers particularly and points out that parallel entrances are better; an entrance stone with a portal outside it creates an unequivocal sightline with no need to establish the centre of the ring, which in itself is often difficult (Burl 2000, 119). Care also has to be taken with interpretation of data. Clive Ruggles identifies groups of Scottish tombs which appear to show a regional pattern; the alignment of Clyde-Solway tombs is clustered around northeast, the Camster tombs in Caithness are clustered on due east and cairns in Shetland are clustered towards southeast. However, he goes on to show that there are other groups of tombs in the same areas within which the alignments are not clustered (those on the Isle of Arran and many of the Clava cairns of Inverness) (Ruggles 1999, 130), thus reducing the impact of those that are clustered together. Similarly, ignoring data that did not fit was also seen as a problem at Calanais in the Hebrides, where the southern orientation of the avenue was emphasised by Hawkins and Thom as a lunar standstill orientation with alignments to local mountains; this attempts to integrate the monument with its landscape and the broader sky, but its contextual reasoning ignores data that do not fit, such as other stones and other similar mountains, and is therefore guilty of data selection (Ruggles 1999, 136). Megalithic Monument Alignment: Conclusions A few alignments, particularly solsticial sunrise ones, can be shown to exist, but there needs to be a consistency of approach and method in order to prove that that was the intention of the builders, rather than just happenstance. Dispersion is a problem; similar attributes have been ascribed to most stone circles, some of which are recumbent, some with cairns; some have sightlines from circle to circle, others from circle to outlier, or from stone to stone or even from stone to hilltop, cleft in the skyline or a slope with a specific angle. All this “provides very weak evidence for a coherent calendrical Chapter 2 59 function” (Burl 2000, 61). However Burl does not dismiss the idea that circles had some function, by saying “If there was an astronomical function to any of the circles it can only have been a minor part” (Burl 2000, 61). If astronomy was one of the major functions of a stone circle, it would be reasonable to expect that most, if not all, of the stones would be involved in forming alignments, rather than the minority, or in some cases just single pairs of stones. Burl also asks the question: “If astronomical sightlines did exist, what are they for? Was it for scientific study of the heavens, or were the lines symbolical – if so, for the dead or for the spirits of an Other-World?” (Burl 2000, 59). Ruggles takes a similar view, by asking “If astronomy really was involved [at Newgrange] the possible reasons why it was important that sunlight around midwinter should light up the interior of a tomb need to be established. What is certain is that Newgrange was not an observatory … its chief function was as a tomb for the dead. Yet few people … have doubted that a powerful astronomical symbolism was deliberately incorporated into the monument” Ruggles 1999, 19). Apart from the site-specific problems of accuracy outlined by Burl and Ruggles, it has also been suggested by others that the accuracy required for some of the events has put them out of the reach of prehistoric peoples. For example, the 18.6 year period between some of the moon-based events that have been identified as the focus for certain alignments causes problems for Anthony Johnson (2008, 254), when he suggests that “the idea that Neolithic people had to wait up to half their lifetime to witness a particular lunar event, which may not even have been visible on the day, casts doubt on the whole idea”. John Oswin felt that the exact timing details of solar events were secondary, that it was more important to mark the main dates for the year, especially the farming year, and “it was enough to know that the sun had reached and passed its solstice” rather than to know exactly when it was (Oswin 2009, 37). Chapter 2 60 4) Burials Prior to the building of permanent religious buildings, alignment of burials and burial chambers seems to have been important since very early times. From the Palaeolithic to modern times some burials appear to have been associated with eastward alignments with purpose. Tombs in the Western Mediterranean The alignments of over 2,000 Neolithic and early Bronze-Age communal burial tombs and other megalithic features on the Iberian Peninsula and in the western Mediterranean have been measured by Hoskin (2001). With a few noted exceptions, particularly in Provence, where all 214 entrances faced sunset or west, 93% of the remainder (2,130 of 2,290) [author’s calculation from various graphs in Hoskin’s study], are aligned very close to the arc of sunrise between midwinter and midsummer, centred on east (Hoskin 2001, 21). Hoskin also comments that “this range was identical with the range of sunrise at the time when churches were constructed” (2001, 7-8). (This is actually not quite true, as the results of this survey will show; there are a small number of churches in this survey that are outside the local arc of sunrise, the furthest by 12º.) The placement of the vast majority of tombs within this arc cannot have happened by chance, and Hoskin infers a significance similar to that found in the alignment of the churches in Britain towards east, by imagining a future archaeologist finding the ruins of thousands of Christian churches and concluding that “sunrise must have played a fundamental role in church orientation, at least in the symbolism of the religion served by these churches” (Hoskin 2001, 7-8). It is disconcerting that, time and time again, turning the pages of Hoskin’s book reveals another graph of tomb orientations, each of which could easily be a graph of English church alignment, with similar ranges and distributions of alignments. In many cultures around the western Mediterranean there is a consistency of orientation over a wide geographical region and it seems that only in the sky can the explanation of such uniformity be found (Hoskin 2001, 16); for example, there are 177 seven-stone antas over a wide area of Portugal and Spain, every one of which is orientated east or southeast over a range of 60º, so Hoskin feels “confident, indeed certain, that we are engaged in archaeo-astronomy” (Hoskin 2001, 16), and that the same purpose was communicated over a broad area. Chapter 2 61 It seems reasonable that the Winter solstice could have been of great concern to early people because of its decisive importance in indicating the change of seasons. The Summer solstice and lunar standstills may also have been of interest (Hoskin 2001, 19); but just why prehistoric people should have found [equinoctial] alignments of interest is not obvious; the precise position is difficult to establish and “does not signify the sort of change that the solsticial changes of the direction of sunrise do” (Hoskin 2001, 18). This whole subject also raises the question of why should any change of season be important to the dead? Facing eastwards could be part of the light/rebirth idea that appears to have been important to the Egyptians, where Pharaohs were buried with their feet pointing east so that they would face the rising sun on revival (Parker Pearson 2003, 59). Mike Parker Pearson also points out that whilst the doorways of the Mediterranean tholoi, a subset of the tombs measured by Hoskin at Los Millares in southern Spain, face east or southeast, the tombs are never built to the west of the settlement, which would have allowed the dead to overlook the living [through the entrance of the tomb] (Parker Pearson 2003, 130), indicating that close consideration had been given to the precise situation of these tombs relative to the settlements, as well as to the orientation of their entrances. Chambered cairns of the central Scottish highlands Although not recorded in similar numbers to those in the western Mediterranean, the majority of the Highland chambered cairns, or passage graves, are aligned in a similar way to them, that is, within the quadrant centred on due east. Henshall and Ritchie consider that the fact that there are equal numbers to the north and south of east “reflects no more than a broad preference rather than one designed with a specific celestial movement in mind [although they are all within the arc of sunrise] (Henshall & Ritchie 2001, 119). Boyle Somerville claimed that the azimuths of two [Clava-type] cairns at Balnuaran near Inverness were aligned precisely with midwinter sunset and that this could “scarcely be a more convincing proof of orientation” (Somerville 1923, 207), although sunset later proved to be viewable in 1997 for “several days” before and after the actual solstice (Henshall & Ritchie 2001, 120); what remains of one of the two tombs is shown in Figure 2.3. Chapter 2 62 Figure 2.3 – Remains of one of the ‘Clava’ cairn pair at Balnuaran, from inside and outside, showing the width of the entrance passage allowing the sunset to be viewed for several days These results are similar to those at Maes Howe, on Mainland, Orkney, where the solsticial alignment is only a generalised one, the sun’s rays in fact strike the rear wall of the chamber for “three weeks either side of the actual solstice” (Foster 2006, 18), although others have noted this event as many as thirty-five days either side of the solstice (Henshall & Ritchie 2001, 121). Despite these results, the solstice was obviously still the focus, especially since the sunset appearances are equally spaced either side of the event itself, and given the time of year, an insurance that improves the Chapter 2 63 chances of the sun actually appearing. Far more of the Clava-type cairns face slightly further south than the two that convinced Somerville of midwinter-sunset orientation. The remaining seventeen appear to be targeted on major setting positions in the lunar cycle, according to Ruggles (1999, 130) and confirmed by Burl (2000, 238). Notwithstanding the possible sun-based or moon-based alignments for these tombs, there are others which appear to have been influenced in other ways. Some of them, particularly a group of long cairns, are thought to have been influenced by the topography of their site rather than any sky-based focus (Henshall & Ritchie 2001, 119), and Ruggles noted that, in most regions, there is a complete avoidance of aligning tombs close to north (Ruggles 1999, 130). Inhumations As far back as 60,000 years ago, the funerary practices amongst the Neanderthals seem to indicate a strong preference for west-east orientated burials. Mike Parker Pearson describes this as an “unlikely chance occurrence”, although the whole issue is still controversial and “there are still many unanswered questions” (Parker Pearson 2003, 149). In the early Bronze Age, inhumations in cemeteries in southeastern Europe were crouched and facing east. Men were placed on their right side with their head to the south, and women were placed on their left side with their head to the north, but “it is difficult to draw universal conclusions as there was an immense variation in practices within the same societies” (Parker Pearson 2003, 54). It has been argued that up to one quarter of all known Anglo-Saxon burials in Britain have relationships with older monuments, mostly Bronze-Age barrows (Williams 1998, 92). This relationship does not just consist of proximity, but in many cases involves alignment as well. At Mill Hill, Deal (Kent), all the Anglo-Saxon burials within the barrow ditch were aligned towards the centre and a significant proportion of the burials outside the ditch were laid with head pointing towards the barrow (Lucy 2000, 130). There are many recorded examples of alignment apparently being used to differentiate between burials of different types. In some instances, alignment appears to have been used to differentiate between adult and child burials. At Wittenham in Berkshire, most of the adult burials were aligned west-east or southwest-northeast, whilst the children were usually buried north-south. Similar patterns were found at West Heslerton in Yorkshire, where “orientation seemed to be quite heavily structured Chapter 2 64 by age, as well as assemblage” (Lucy 2000, 131-132). Elsewhere, orientation seems to have been used to differentiate by gender, as at Sewerby in Yorkshire, where “burials with jewellery were more likely to be orientated to the south or southwest, whilst burials with weapons seemed actively to avoid this direction” (Lucy 2000, 132). In a study of seventh- and eighth-century burials at Finglesham in Kent by Sonia Hawkes (1976, 33-51), it was argued that although generally west-east, 90% of burials were aligned between midsummer and midwinter sunrise azimuths (the arc of sunrise), and were therefore deliberately aligned with sunrise on the day of the burial, the variations in grave alignment reflecting the time of year of the burial (Hawkes 1976, 50-51). It was subsequently argued by Martin Welch (1992, 74-75) that, because of the lack of graves aligned at the southern end of the sunrise range, this model significantly under-represented the number of deaths that were likely between November and February, and that the variation in alignment was due to inaccuracies on the part of the burial party in assessing the direction of east, or to using a local referent [the barrows located close to the cemetery]. Welch also pointed out that west-east burials cannot all be assigned to Christianity, as this orientation seems to have been widely used before Augustine’s mission to England in 597 CE, and it is not known when Christian communities in Europe adopted west-east burial as their exclusive method, “but there is a suspicion that it was at a relatively late date” (Welch 1992, 74). With such variation in alignment of early Anglo-Saxon burials, it seems obvious that there were a number of different influences involved. Any number of reasons for specific alignments might have been considered, which could have been regional, local or personal, for example, towards the origin of his/her ancestors; towards former burial places; facing the deceased’s house or favourite place; towards a natural feature; following local existing practice; at a significant time of day (sunset etc); using the lie of the land; facing down hill; facing away from the prevailing wind; or facing away from the settlement (as with the tholoi mentioned earlier). As Ruggles puts it (1999, 89), “some of the possible methods are only significant to the deceased and their close family, others have a slightly wider significance – how can we know which of these situations applies for any given burial?” Excavations of Anglo-Saxon inhumations in Norfolk seem to have produced slightly less variable results. At Spong Hill, the 58 late fifth- to mid sixth-century graves excavated in 1974 were arranged approximately west-east, the majority within Chapter 2 65 20° of east, and almost all within the arc of sunrise with a mean direction slightly south of east (Hills et al. 1984, 21). At Sedgeford, the middle-Saxon burials were aligned either side of east, the later burials aligned south of east, although the earlier burials appear to follow the local contours. It was also considered that some burials may have been aligned towards a local referent on the summit of Dove Hill (Davies & Hoggett 2001, 18; Wilcox 2002, 38). Richard Morris suggests that the cemetery in early Christian times fulfilled some of the ritual functions that were eventually served by the church (Morris 1983, 33). Certainly by this time alignment eastwards had become important. This practice of eastward burial alignment continues today, although for the churchyards which are part of this survey, visible burial monuments (mainly post 1700) appear to be aligned with local ecclesiastical east, in other words, parallel with the church and other burials, irrespective of the actual alignment of the church. This is a pragmatic approach, avoiding the requirement for the gravediggers having to find true east, and one which was mentioned by Rahtz (1978, 1-14) as a consideration in relation to Saxon grave alignments. In one or two extreme cases in this survey on very steep slopes, the graves have been aligned parallel with the contours, rather than west-east. Despite the west-east pre-requisite for Christian burials, burials have still been used to indicate hierarchy; the lord of the manor was frequently buried in the church, the well-to-do buried on the south side of the church, close to the door, and the poor were often buried on the darker north side, which has always been associated in folklore with the devil and evil (Parker Pearson 1999, 14). Chapter 2 66 5) Mystical Alfred Watkins interpreted the straight lines he identified between objects and features in the landscape, such as horizon notches, as tracks used in prehistory linking ancient sites and he named them ley lines, but even he recognized that when these lines were close together and close to parallel, such as the convergences he noted at Stonehenge, that they could not have been tracks (Watkins 1925). The straight lines were reinterpreted by John Michel (Michel 1969) as lines of energy in the landscape utilised by prehistoric people (fitting Ruggles’ concept of lost-knowledge (Ruggles 1999, 3)). Subsequent careful statistical analysis has disproved the whole concept of the existence of these straight lines between ancient features as meaningful. Comparing the ‘actual’ ley markers identified on a particular 1:50,000 Ordnance Survey sheet, with a similar number selected for different reasons (named buildings beginning with certain letters), Forrest managed to produce more ‘ley lines’ with the second set (cited in Williamson & Bellamy 1983, 97). Similarly, Williamson and Bellamy conducted their own experiment with ley marker data, identifying all the supposed ley lines on a specific 1:50,000 Ordnance Survey sheet, then randomised each marker position within its 1km grid square, and found an equal number of six, seven, eight and nine-point lines. This led them to conclude that “there is thus not the slightest evidence for the deliberate alignment of ancient sites on the map” (1983, 101-102). In addition to the statistical dismissal of the idea, much of the original survey work by Michel was apparently subject to similar data selection problems to those identified earlier by Ruggles (1999, 41) in relation to stone circle alignments. Michel’s assertion that he only used prehistoric markers as primary evidence, and other features as confirmation, has been shown to be false. Later checking by Williamson and Bellamy showed that he not only used some later features as primary markers, but also ignored “innumerable large stones” as possible primary markers, but which did not fit into any of his lines (Williamson & Bellamy 1983, 104-115). The concept of Orthotonies, straight lines along which UFOs were supposed to have travelled, briefly became of interest in the 1960s; “but it was rejected [even within the fringe] because of its poor statistical basis” (Williamson & Bellamy 1983, 15). Chapter 2 67 ALIGNMENT: CONCLUSIONS There is clearly a general concentration of eastward alignments for each of the subjects examined in this section; some are practical whilst others are symbolic. In the case of house doorways, the selection appears to be practical, as it is easy to see why they should face east, to let in the maximum amount of light in the winter and face away from the prevailing wind, but the majority of eastward alignments are symbolic and seem to have been important for thousands of years. For many of the burial practices, facing east can be seen as part of a light/rebirth/resurrection focus, aimed at the rising sun, emphasising the positive, and easing the deceased’s passage into the next life. The interpretation of, and even the existence of, alignments at some monuments, however, such as at Stonehenge and Calanais, is more problematic. Some of the examples discussed here, such as many of the claims for megalithic astronomy and the multiple alignments in stone circles, have been shown to be either fanciful or grossly overexaggerated, but there are still many situations in each of the aspects of prehistoric life, such as burial tombs, where similar and specific alignments over a wide geographic area cannot have been arrived at by chance. If they were deliberate they must also have held meaning for the people who set them out, but it is that meaning which appears to elude modern man. Hoskin’s example of a future archaeologist encountering ruined Christian churches and arriving at a sun-based conclusion for their alignment highlights the problem. Of course, in a non-prehistoric context, the future archaeologist could use the written history to establish why churches were aligned eastwards, except, as will be demonstrated later, there is none. Chapter 2 68 CHAPTER THREE HISTORIOGRAPHY of CHURCH ORIGIN, CHURCH LOCATION and RURAL SETTLEMENT Bor, ha’ yew noticed yew hin’t far t’saarch in Norfick, if yew wanter see a chaarch? Jus yew go up ena little hill an yew’ll see savrel chaarches, that yew will. (Norfolk Churches – John Kett) INTRODUCTION This chapter examines the origin of churches and especially the pattern of hierarchical development, broadly described as the ‘minster model’, according to which, church building was spread down the social scale as time progressed. This chapter also explores earlier work on the location of churches and the reasons for the choice of site, ranging from the very broad plans of the early church hierarchy to incorporate earlier pagan religious sites, to the site-specific choices made by each local church builder. In addition, it also briefly outlines earlier explanations of the patterns of settlement development from the seventh century to the twelfth century, a time of much church building. Each of these subjects will be developed in the later chapters of this thesis by using the results of this survey to test whether the conclusions drawn by earlier writers can be expanded or refined, particularly concerning the location of churches and the timing of their building. CHURCH ORIGIN This section outlines the historiography of the hierarchy and origin of a range of rural church buildings, from minster church to field church. Defining either of these types of church is not simple, as the “lack of generally accepted terminology for religious institutions has caused terrible confusion” (Foot 1992, 216), and the term seems to have been used in several different ways over the years. In the early days of the Conversion, Chapter 3 69 minsters – to which large parochial territories were attached – facilitated the conversion of the population and their integration into the church by teams of clergy based at important early churches (Radford 1973; Blair 1988a, 2005; Morris 1989; Blair and Sharpe 1992; Foot 1992). Initially, their parochiae often seem to have extended to the whole Hundred, based on a seventh- or eighth-century Royal tun (Barlow 1979, 169170; Morris 1983, 64; 1989, 128). In recent times, “Minster has been used as a specialized sense by some scholars to refer to Anglo-Saxon institutions that lay behind the mother churches in Domesday Book” (Foot 1992, 215-216), but Franklin suggested that, in the most general terms, “a minster is a church that originally, or vestigially, had pastoral responsibilities for an area larger than a single village” (Franklin 1984, 69). Summed up neatly by John Blair, “the generally accepted view of how England acquired its rural churches was that Kings, and bishops under their patronage, founded churches in important administrative centres. By the mid eighth century all, or most, English Kingdoms had established a network of minster parochiae, typically covering between five and fifteen modern parishes, served by priests from the central church” (Blair 1991, 91). At the other end of the scale, field churches are defined by their lack of burial rights and therefore their lack of churchyard. King Edgar’s Law codes of 960 CE show ‘churches without graveyards’ as the last of the list of church categories (Barlow 1979, 187; Gem 1996, 22; Morris 1989, Blair 2005). This categorization implies that thegns’ bookland churches without graveyards would be included with field churches, along with chapels, such as monastery gate chapels and chapels of ease. There has been a wide range of influences on the origin of churches as well as on their location. The origin of churches and their position in the hierarchy, which in turn is determined by the position in the hierarchy of the sponsor/owner, are so closely linked and inter-related as to be difficult to separate. In other words, the origin of a church on a particular site is likely to be affected by its position in the hierarchy; whether it was an early minster church built on a mid-Saxon royal estate, or one of the thousands of lesser churches built later on smaller estates. Below minster-church level it is difficult to explain the background to a church, because there is little direct or specific evidence for the origin of most English parish churches (Owen 1971, 11-13; Morris 1985, 47; Blair 2005, 373). Exactly what constitutes the origin of a church on any site can also be difficult to assess, as the current church may have been preceded Chapter 3 70 by an earlier church, which may itself have been preceded by an earlier religious use of the site, such as a graveyard which could have been used as a preaching location, and which itself might have been associated with an earlier pagan ritual or burial site. Church Hierarchy There were two broad phases of church building in England, closely connected to the ecclesiastic hierarchical structure. The first phase, begun soon after the Conversion, established the ecclesiastic framework by setting up minster churches (Morris 1989, 93-139; Blair 2005, 79-290) and the second phase in the late-Saxon period saw the minster system begin to fragment, and the parochial system develop, as churches were built by those further down the lordly hierarchy (Morris 1989, 140-167; Blair 2005, 368-504). Late-Saxon Law codes mention church hierarchy several times. King Edgar’s Law codes of 960 CE list three categories of church – old minsters, thegns’ bookland churches with graveyards and churches without graveyards (Barlow 1979, 187; Gem 1996, 22). Archbishop Dunstan at the end of the tenth century classified churches as head, middling or lesser-minsters, with a fourth class of field churches with no burial grounds (Braun 1970, 38; Barlow 1979, 187). Some six decades after Edgar, the codes of Cnut (1014) and Ethelred (1020) list four church categories – head minsters, lesser minsters, still smaller minsters and field churches (Blair 2005, 368; Morris 1989, 129; Gem 1996, 22; Braun 1970, 38). John Blair (2005, 368) argues that the third of the four levels of church mentioned in the law code of 1014 must be the thegn’s estate church, redefining the other levels as cathedral, clerical minster and finally the field church, whereas Barlow referred to them as Episcopal minsters, other old minsters, manorial churches with burial rights and “the rest” (Barlow 1979, 187). By the tenth century there were many large rural parochiae, and tithe, churchscot and soulscot from local churches were still paid to the old minster (Morris 1989, 228), unless the church had a graveyard, in which case one-third of the Lord’s tithe went to his own church (Barlow 1979, 195; Morris 1989, 228; Jones A. 2002, 49). Subsequent devolution of power and the fragmentation of estates, referred to by Blair as “manorial fission” (2005, 372), along with the requirements of a rapidly expanding population, gradually increased the number of churches built on successively smaller and smaller estates. At the lowest end of the manorial spectrum, this expansion of Chapter 3 71 church building sometimes meant multiple churches in one settlement – a process seen in the extreme in East Anglia (Warner 1986, 45-46; Williamson 1993, 158-161; Groves 1995, 108-115) – in some cases with three or four churches built close together. There are seventy-nine villages in Norfolk alone that had more than one church (Batcock 1991, 10-11), and three neighbouring parishes in west Norfolk, Oxborough, Barton Bendish and Beechamwell, had at least three churches each (Rogerson 1987, 1). The position of a church (or patron) in the hierarchy is particularly important where church sites are concerned, and can be demonstrated to have had an effect in many cases – although these are frequently concentrated in urban areas. In York, for example, the only church that is not constrained by its site is York Minster, which is aligned at close to 90° (commented on by Harvey and Morris, noted in Chapter One), compared with all the remaining churches in the city, some of the which follow the Roman street grid (Holy Trinity, Goodramgate (36°), St Olave, Marygate, at (41°), St Mary, Bishophill (42°), and at right-angles to them is St Michael le Belfrey (127°), author’s measurements June 2001), and others which “took their cue from the nearby streets and properties of the Anglo-Scandinavian city, and are angled this way and that, like so many weathercocks” (Morris 1989, 209), the alignments of which range between St Cuthbert, Peasonholme, at 56° and St Helen, Stonegate at 83° (author’s measurements June 2001). In reality this seems to reflect the fact that the builders of the minster had the power to clear a large site, rather than be forced to build within the confines of the city’s existing topography. St Michael le Belfrey, adjacent to the minster but aligned at 127°, was rebuilt as late as the early sixteenth century, between 1525 and 153712 Pevsner & Neave 1999), but remains parallel with the Roman street alignment within the confines of its earlier site, shown below in Figure 3.1. 12 http://stmichaelsyork.org/cps/history.htm, (accessed 12th June 2008) Chapter 3 72 Figure 3.1 – Alignments of York Minster and St Michael le Belfrey (right foreground) compared13. At the foot of all the hierarchical lists, the field church seems to have been a classification that groups together the many types of small church that did not have a graveyard, including small estate churches, monastic gate chapels and chapels of ease. In medieval times they existed in considerable numbers, “especially in areas where the parochial system developed later, and provided places of worship for the convenience of growing and scattered populations” (Dymond 1995, 58), particularly in the winter months when travel might be difficult. In East Anglia, many were originally built as local churches by groups of freemen (Warner 1996, 197-198), and these have been suggested as one possible reason for the growth in shared churchyards (Williamson 1993, 159). 13 Source: http://york-tourism/images/36549 (accessed 12th July 2004) Chapter 3 73 What preceded the current church building? Although this thesis is principally concerned with existing church buildings, how the church site developed and what preceded the current building, whether it was an earlier church building or a cemetery, is important, as the fact that there was an earlier use probably determined the choice of the site for the current church. This aspect will be explored in the final chapters of this thesis. Early-Saxon cemeteries were usually separate from settlements (Hills 1979, 310; Rogerson 2005), but many of the ‘final-phase’ cemeteries which were founded on new sites in the seventh century and often integrated into the settlement of the period, have been obscured by later settlement, including churches and churchyards (Morris 1983, 53-55; Hoggett 2007, 314). It is not certain that the presence of a graveyard of the period 850 CE-1100 CE can necessarily be equated with presence of church, but if burial grounds that did not develop into churches were common, then more would be expected to have been found away from current churches, “therefore some, if not most graveyards developed into church sites” (Morris 1989, 153). There are few examples of churches where excavation has shown the extant church to have been rebuilt on the same site as an earlier one; Rivenhall and Asheldham in Essex (Rodwell 1973 & 1985), Raunds (Northants) (Boddington 1996), Barton on Humber (Lincs) (Rodwell 1981), Tong (West Yorks) (Ryder 1993, 119132), Wharram Percy (East Yorks) (Bell & Beresford 1987), St Helen in the Walls (York) (Rodwell 1981), St Marks, Wigford (Lincs) (Gilmour & Stocker 1986), Potterne in Wiltshire (Davey 1964), Llanelen on the Gower peninsular (Schlesinger & Walls 1996, 101-104), Iken (West et al. 1984) and Brandon (Carr et al. 1988) in Suffolk and four in Norfolk – All Saints, Barton Bendish (Rogerson et al. 1984), St Benedict, Norwich (Roberts & Atkin 1982), Thetford (Dallas 1993) and St Martin at Palace, Norwich (Beazley 2001). However, the development situation is bewilderingly complicated even when just these few sites are considered. At some, postholes or slots for an earlier wooden structure were discovered, in some cases within the area enclosed by the later stone foundations (St Martin Norwich, Thetford, Wharram Percy) and at others as separate buildings (Rivenhall, Wigford, Llanelen, and possibly at Shapwick in Somerset (Gerrard & Aston 2007)). At other sites earlier stone work or foundation trenches were found, again in some cases within the confines of the later building (Iken, St Helen York, Tong, Barton on Humber) and at others as a separate structure Chapter 3 74 (Raunds). At three of the sites – Barton Bendish, St Benedict’s Norwich and Potterne – the excavators felt that an earlier church was located close by in the churchyard, but was not discovered within the excavated area. The whole subject of earlier timber churches preceding the current building is further complicated by the possibility that postholes and beam slots within church buildings are, in some cases, for internal fittings or for construction scaffolding, such as those at Hadstock in Essex (Rodwell 1981, 114-115). At none of the churches noted above was a significant realignment noticed by the excavators. The discovered post holes and earlier stonework or foundation trenches were on the same, or similar, alignment as the current church, thus arguing against the comment by Richard Morris that there were changes in alignment in churches at village level between the Saxon and Norman periods as part of the “signs of a new concern with alignment during the eleventh and twelfth centuries” (1989, 208) that had been identified at Exeter, Wells and York when the cathedrals were rebuilt by the Normans in the eleventh century. Excavation shows few English parish churches with structures before 900 CE (Morris 1989, 163), although by the 940s, private churches were tipping the balance of pastoral organization in the countryside (Blair 1985, 119; Morris 1989, 196). Parish churches prior to 950 CE were normally fabricated in wood (Morris 1989, 165), and the widely held view is that the “great rebuilding” of churches in stone between 1050 CE and 1150 CE (Gem 1988; Morris 1989; Blair 2005) was a stamping of the new Norman manorial authority in England after the Conquest, rather than a material- or technique-led exercise. However, Alibert (1997, 180) argues that “Even in eleventhcentury Normandy, churches built in wood existed alongside those in stone, which [wooden churches] were traditional amongst the countries from which the ‘new Christians’ [Normans] came.” This seems to imply that the Normans had continued to build wooden churches long after their conquest of Normandy, even with access to excellent stone close to hand, at Caen. In Denmark, the rebuilding of wooden churches in stone peaked in the early twelfth century (Abrahamsen 1992, 293) and was therefore carried out in parallel with the same exercise in England, reinforcing the idea that rebuilding churches in stone in England was part of a contemporary process in many countries, rather than a specific exercise of the power of lordship. Chapter 3 75 CHURCH LOCATION There appears to have been a range of influences on the location of churches, operating either from the top down within a broadly over-arching spiritual plan of the church hierarchy, such as the possible incorporation of earlier ritual sites; or from the bottom up, with the simple practical local desire of each patron for his church to be located for his convenience and on a suitable secure site. The first two of these factors has been examined by earlier writers, particularly the reuse of earlier pagan sites, variously called “Christian substitution”, “Christianization” and “de-paganization”, with differing degrees of support. This section outlines and comments on the possible reuse of earlier types of religious site, but first addresses site suitability and siting for the convenience of the patron. Site Suitability Perhaps the most important reason for the choice of the location for a church, or almost any other building, has to be the suitability of the site for building on. The ground should have suitable load bearing qualities, and it should be both stable and accessible when needed. The successive floods of 2007 and 2008 have highlighted this issue, with frequent pictures in the media of Tewkesbury Abbey standing alone above the floodwaters, with the rest of the town submerged. Here, the choice of site must have been made after long familiarity with the characteristics of the flood patterns of the Rivers Severn and Avon, and is also a reflection of the patron’s position in the hierarchy and his ability to command the best site. Although the current abbey building was not started until 108714 (Verey 1970a, 256), the site for the town has Roman and Saxon origins, with an earlier Saxon monastery located on the same site15 (Aston 2000, 59); establishing a monastery is likely to have been sponsored and located by a person of high status who was able to donate the site. The suitability of the site is particularly marked when comparing the situation at Tewkesbury with pictures of the sites of neighbouring village parish churches, suffering in several feet of water in the same floods, where the owner did not have the same ability to choose the best site. They are pictured below as Figures 3.2 and 3.3. 14 http://www.tewkesburyabbey.org.uk/his.htm (accessed 17th Nov 2009) 15 http://www.tewkesbury.net/history.asp (accessed 17th Nov 2009) and http://www.theheritagetrail.co.uk/abbeys/tewkesbury%20abbey.htm (accessed 17th Nov 2009) Chapter 3 76 Figure 3.2 – Tewkesbury Abbey (Gloucestershire), during the floods of 2007, occupies the highest land in the town16. Figure 3.3 – St Michael’s, Tirley (Gloucestershire), three miles south of Tewkesbury, during the same floods – the porch gives a true indication of the depth of the water17. 16 17 Source: http://guardian.co.uk/artblog/2007/jul/23/week/ Tewkesbury.html (accessed 1st April 2008) Source: http://churchtimes.co.uk/uploads/images/Tirley-P/42537.html (accessed 1st April 2008) Chapter 3 77 The suitability of the site for building on is logically the most important criterion used in site selection for a church, but other influences must also have been considered. Some might be thought of as practical reasons, such as a convenient location close to the patron’s residence; and others as spiritual reasons, such as the incorporation or suppression of earlier pagan religious sites. These two are considered below. Convenience of the Patron Apart from site suitability, the main practical reason that is likely to have applied in all areas, urban or rural, is the siting of the church close to the patron’s seat for their convenience, referred to as “foundation by estate proprietors” (Blair 2005, 385), as “seigneurial considerations” (Morris 1989, 268), and as the Church/Hall focus (Dymond 1968, 29). This may be seen as being a development of the pagan custom of having an idol-room in the house (Morris 1989, 75), but there are few examples of churches which were absorbed into the house, such as at Cheddar, Somerset, and Deerhurst, Gloucestershire, “they are limited in number and a very elite practice” (Blair 2005, 386; Taylor & Taylor 1965). However, the close association of church and house seems to be applicable at all levels of the church hierarchy, from the early eighthcentury minster-church sites located close to Royal estates, such as at Bamburgh in Northumberland (Morris 1989, 131) or Rendlesham in Suffolk (Warner 1996, 115; Carver 2005, 494; Newman 2005, 478); the large parochiae representing Crown demesne manors, such as at Pickering, Pocklington and Driffield in East Yorkshire (Morris 1989, 133-135); countless lordly churches built on smaller estates during the tenth and eleventh centuries; to smaller field churches built by groups of freemen often located midway between hamlets or farmsteads (Warner 1996, 197-198; Williamson 1993, 159)18. The juxtaposition of the church and the manor house is commonly seen throughout most of the country. Richard Morris notes that although some houses next 18 One Norfolk example of a field church may have been misinterpreted in the past as an example of a church from the ‘lordly estates’ category - St Withburga’s at Holkham is an isolated church located on a large knoll now in the flat parkland of Holkham Hall, the seat of the Earl of Leicester, but in the thirteenth century was “located equidistant between three settlements in the parish” (Hassall & Beauroy 1993, 537). Chapter 3 78 to churches are of medieval date, most of the current buildings are post-medieval, and that some are either in castle baileys or next to mottes (Morris 1989, 248). In Herefordshire, 50 of the 102 mottes shown on Morris’ map (1989, 251) (author’s calculation from the map data) are “in close association” with the church – although this is taken as “within 200 yards”, rather than directly adjacent. He sees this as an indication of an early seat of power “often replaced by a hall later” (1989, 250), and if the Norman ownership, which begat the motte, was imposed on an earlier estate, then the juxtaposition is likely to be repeating an association between hall and church that was already in existence. In Suffolk, David Dymond has shown that 36 of 52 isolated church sites are next to manor houses and he feels that there is a possibility that many of these are successors to pre-Conquest halls (Dymond 1986, 29). Even in Norfolk, where Tom Williamson has suggested that this association is less strong (Williamson 1993, 167-168), the results from this survey show that of the 174 churches currently remote from their settlement, 31 are next to a Hall or moated site, and in addition there are 38 adjacent to just a single farm. Richard Morris has commented on the juxtaposition of farm and church as a possible indication of earlier manorial sites “dwindling in importance to the status of farms” (Morris 1989, 274). The building of the church prior to the motte has also emerged as a common model (Morris 1989, 252), particularly where Saxon burghs preceded Norman castles, for example at Cuckney in Northants, Corfe (Dorset) and Eye in Suffolk, where the churches are just outside the Saxon stronghold (Morris 1989, 262-4). The church was also built outside the ditched enclosure around the manor house at Raunds in Northants, circa 920 CE, and at Goltho in Lincolnshire, circa 1000 CE, which may imply a parochial use of the church, so that villagers could enter the church without having to enter the manorial enclosure (Blair 2005, 389). Chapter 3 79 Figure 3.4 – St Mary’s and Hall at North Aston, Oxfordshire – the ‘closest association’ between house and church found in this survey (although this version of the hall is of the seventeenth century (Sherwood & Pevsner 1974, 718)) John Blair, however, puts forward an alternative model to that of the church as an adjunct to the manor house. He postulates that there appears to be “an equally widespread pattern of church development as a part of structured peasant settlements, possibly explaining the location of thousands of churches” (Blair 2005, 395), which he attributes to absentee, particularly monastic, landlords (Blair 1991, 140-142). So in these cases, the church was located for the convenience of the congregation, rather than that of the patron. This is also hinted at in Lincolnshire, identified by Stocker and Everson, in that there is a recurrent pattern of locations of church and rectory within a row of settlement plots, rather than adjacent to a manorial site. The authors there thought it spoke either of public involvement in church planning or as lordly replanning of the settlement (Stocker & Everson 2007, Fig 3.28, 61 & 66). Settlement shift from possible Anglo-Saxon cores, still marked by the presence of the church, has been noted in several villages in southeast Somerset (Ellison 1983), Chapter 3 80 for example, at Aller (1983, 14), Beercrocombe (22), Seavington St Mary (84), Seavington St Michael (89), Stocklinch (98) and Broadway (104). Similarly, at Wharram Percy in Yorkshire, excavation has revealed that the site of the manor complex had been shifted twice, in the twelfth century and in the thirteenth century, each time moving north from the original Saxon settlement site, still marked by the church ruins (Beresford & Hurst 1989, 122-132 & Fig 25). In each of the cases noted here, it seems to point to the selection of the church site being made for the convenience of someone in the Saxon period, usually the patron. In Norfolk and Suffolk, a number of fieldwork investigations have been undertaken around church sites, each of which seems to point to Saxon-period church sites (Wade-Martins 1980a & 1980b; Rogerson 1997; Davison 1990 & 1995; Newman 2005; Martin 2005; Laverton 2001); these will be considered in more detail in Chapter Eight. Chapels of ease were also sited for convenience, but for the convenience of the congregation. Often located in positions which were closer to the settlements than the parish church, and easier to reach particularly during winter, many chapels of ease later became parish churches, although “this was often long delayed as it threatened the status and income of the mother church” (Dymond 1995, 59). Reuse of earlier sites The main spiritual reason for the choice of a particular location was arguably to incorporate the power or influence of an earlier site and to ease the transition into Christianity for the users of the site by its reuse. There are various possible situations involving such reuse and it has been considered important to differentiate between them (see Morris 1989, 57-58; Blair 2005, 183; Bradley 1993, Eaton 2000, Bell 2000, Hawkes 2003), although it is difficult to see much of a difference between the three categories. The site may have been reused for reasons of Christian substitution, by converting the site to Christianity and incorporating and suppressing the extant ritual uses; it may have been used as a Christian site after a break in use, noted as superimposition; or as an expression of Romanitas, which is seen as the emulation of the culture of the Roman Empire and early Christianity. This division seems to be somewhat artificial. Converting a pagan site to a Christian one is straightforward; incorporating the earlier use could ease the transition from one religion to another in Chapter 3 81 the minds of the participants. Re-adoption of a site after a break in use can either be made by chance (in which case it is not really a readoption but coincidence), or it can be made because the earlier use, although discontinued, was remembered or identifiable, in which case there is no real difference between converting a site and readopting it. Expressions of Romanitas appear to be similar to Christianization, in that it was hoped that the power of the stone, or the site’s earlier use and its connection to Rome and Christianity, could be incorporated (Eaton 2000, 96; Morris 1989, 28-29). Christian substitution was seen as important by the church in the earliest days of the Conversion. It was mentioned in the Pope’s letter to Abbot Mellitus in 601 CE (Grinsell 1986, 33; Morris 1989; Blair 2005), in order to assist in the conversion of the population to Christianity by using their existing religious sites. It was raised again much later as one part of the Canons of Edgar19 in 1005x8, which instructed priests to extinguish heathen practices, such as the worship of wells, trees and stones (Morris 1989, 60; Rattue 1995, 79; Harte 2008, 22), which, although it was not suggesting reusing the site, does confirm that these practices must still have been continuing in a widespread manner at this time. There are many examples of churches that have been built next to, or on, features that may represent earlier ritual sites, such as in the henge at Knowlton (Dorset) (illustrated as Figure 3.12 on page 93); in the centre of the IronAge hillforts at Whittlebury in Northamptonshire (Jones & Page 2006, 187), at Hanbury (Worcestershire), Tetbury (Gloucestershire), Aylesbury (Buckinghamshire) (Blair 1992, 234), and Breedon in Leicestershire (Morris 1989) (illustrated below); adjacent to the Iron-Age broch at Harray (Mainland, Orkney) (Petrie 1890, 93; Fraser 1923, 32), (illustrated below); alongside the megalithic remains at Stanton Drew (Avon) and Midmar (Aberdeenshire) (Morris 1983, 59) and possibly at Avebury (Wiltshire) in the Saxon burgh adjacent to the henge (Pollard & Reynolds 2002, plate 17 and 235-237) or next to the monolith at Rudston in East Yorkshire (shown as Figure 3.13 on page 93). There are also many examples of churches built on, or near, BronzeAge barrows, for example at High Wycombe, Buckinghamshire, which adjoins elite seventh-century barrow burials (Morris 1989, 256); Earsham (Norfolk) (Morris 1983, 60); Scartho (Lincolnshire) (Stocker & Everson 2007, 61); Ogbourne St Andrew (Wiltshire), and Fimber (East Yorkshire) (Morris 1983, 59; Grinsell 1986, 33). 19 Mis-named, according to John Blair, they should be attributed to Archbishop Wulfstan (Blair 2005, 444) Chapter 3 82 Figure 3.5 – St Nicholas’, Harray, Mainland, Orkney – the churchyard built around the remains of an Iron-Age broch, with a war memorial at its centre Figure 3.6 – St Mary & St Hardulf’s, Breedon on the Hill, Leicestershire – the church located in an Iron-Age hill-fort, also the site of a Mercian minster circa 680 CE (Morris 1983, 58) Chapter 3 83 Similarly there are churches built near, or over, springs (wells), of which there are many extant examples in this survey: Westwell and Binsey (Oxfordshire), Gilcrux, Kirkoswald and Castle Sowerby (Cumbria), Dereham (Norfolk), numerous churches in Cornwall and several in Pembrokeshire, for example at Burton, Dale, Lawrenny, St Ishmaels and Gumfreston (the last two are illustrated as Figures 3.7, 3.8, 3.9 & 3.10). Figure 3.7 – A spring emerging beneath the chancel at St Ishmael’s, Pembrokeshire Figure 3.8 – A spring enclosure in the nave north wall at St Lawrence’s, Gumfreston, Pembrokeshire Chapter 3 84 Figure 3.9 – Gumfreston Spring enclosure (detail) Figure 3.10 – view of the altar from the spring enclosure Springs, or wells (the O.E. form of spring (Gelling 1984, 30)), are the most obvious example of natural objects in the landscape which might have been the subject of veneration in earlier times – because of the magic of water issuing from the ground – and perhaps the most obvious candidate for Christianization. Jeremy Harte records 919 Holy wells, 201 of which are associated with a church and 67 are located in churchyards (Harte 2008, my calculations from his appendix “Analysis of sites”). James Rattue’s assessment of Holy wells though, seems to downplay the whole idea of Christianization, saying that “too much has been made of the various aspects of it”, particularly springs, because water is naturally at the heart of the Christian ritual, involving the use of holy water generally and particularly in the central act of baptism (Rattue 1995, 34). Few holy wells were recorded before Domesday; Rattue lists nine, and Jeremy Harte lists eight from Anglo-Saxon sources (charters, the Burghal Hideage and a will) (Rattue 1995, 63; Harte 2008, 17-18), although this may be an under recording, as charters usually concentrated on boundary features, so might not have included wells that were sited away from the boundaries. A few holywell-type placenames are mentioned in Domesday for the first time; a dozen according to Rattue (1995, 63) and three according to Harte (2008, 18). Both writers list many wells Chapter 3 85 Christianized at a later date, but Rattue goes on to say that “the simplistic model of Christianization is totally inadequate” (1995, 66), because it can be shown that the Church was fully prepared to create its own holy wells rather then convert pagan ones. He sees this as part of a larger pattern, where the Church was afraid of hydrolatry and other nature worship throughout the eleventh century, evidenced for example, by Edgar’s Canons of 1005x8, calling for extinguishment and destruction of pagan sites, mentioned earlier (Rattue 1995, 79-81), in addition to other mentions in the Northumbrian Priests’ Law of 1010 (Harte 2008, 22). These rules were reversed in 1102 by the Council of Westminster when cults of wells were brought under the Church’s wing, and were made subject to the authority of the local bishop (Rattue 1995, 79; Harte 2008, 23), which “attributed sanctity to the fountain” (Jones 1998, 76), when they “became part of the wider religious landscape” (Harte 2008, 23). Rattue goes on to suggest that this softening of attitude might have been because of the general spread of new churches, and that the authorities may have realised that “they were no longer living in a sea of sub-paganism” (1995, 79). If the Church had seen the Christianization of wells as a priority in earlier times, then the use of well sites close to churches would surely have been evident in a far greater number of cases. There are only a few examples of major churches associated with wells, notably at the eponymous Wells in Somerset, where the Holy well of St Andrew is close to the east end of the minster, with continuity of use on the site with a Roman mausoleum and a Saxon burial chapel (Rodwell 1981, 142-143; 2001). Overall, the idea of Christianization may be somewhat overstated, because as Rattue puts it “A very substantial number of wells, if not the absolute majority, were left alone by the church” (1995, 42). Harte supports this by suggesting that around two-thirds of all healing wells in several counties he surveyed – Devon, Herefordshire, Warwickshire, Worcestershire, Somerset and Gloucestershire – were never adopted by the church (Harte 2008, 93). Richard Morris also points out that the possible relationship between a parish church and a pre-Christian sacred place “becomes bewilderingly complicated because of the jumble of cults and the revival of rural heathenism in the ninth and tenth centuries, after the Scandinavian colonization” (1989, 56-57), particularly in the east of the country. The idea that there was a period after the eleventh century when the reuse of sites was considered more important than previously, is obliquely reinforced by John Blair’s observation that the examples of churches built close to Bronze-age barrows Chapter 3 86 which reflect a local continuation of ritual sites, is a pattern in which “the builders of the earlier [pre-eleventh century] minster churches had shown little interest” (Blair 2005, 376). On the other hand, John Blair has suggested that the model of ‘churches on sacred sites’ has not really been applied to England, where parish churches are normally seen as off-shoots of manor houses (Blair 2005, 374). He notes that several single-celled chapels of the tenth to twelfth centuries, mainly in Cornwall, have been shown, by excavation or by their association with above-ground remains, to have developed from pagan cemeteries or wells. “This suggests that there may be others that perpetuate different sorts of ritual features such as trees or standing stones which have left no trace” (Blair 2005, 376), an idea which will be developed by the application of the results of this survey in later chapters. The lack of permanence of some of these features presents a problem because possible sites such as sacred trees or groves are unlikely to be evident in the later landscape, as few trees in Britain live long enough. There are, however, some examples, such as at Ketton (Rutland, now Lincolnshire), where excavation has revealed burials around a tree which is close to a small church associated with further burials, although it is unclear which was there first (Blair 2005, 381). The yew is a tree of particular interest in this context, despite the fact that worship through nature was seen as the Kingdom of the Devil by the Church (Cornish 1946, 16; Ross 1967 33-35), as its evergreen foliage was seen as a symbol of everlasting life. It must also be remembered, that “in early times, the yew held a different position in peoples’ minds as it was almost the only evergreen tree in England and Wales” (Cornish 1946, 17). Although the yew can be particularly long-lived, Cornish felt that “it is unlikely that any yew tree growing in churchyards today was planted before the time of the great rebuilding, making them a Christian symbol rather than an earlier pagan one” (1946, 17). However, modern dating techniques have shifted the probable planting date of the largest yew trees back several hundred years. There are thirty-one examples in Bevan-Jones’ survey of trees which have trunks greater than ten metres in circumference, such as at Clun in Shropshire, which may be more than 1300 years old (Bevan-Jones 2002, 29 & gazetteer 189-193), firmly placing their planting in the middle-Saxon period. Bevan-Jones points out that only a few of the known Anglo-Saxon churches have ancient yew trees in their yards, but the trees’ Chapter 3 87 longer lifespan could make the yew a marker of seventh- or eighth-century AngloSaxon church sites which currently only contain a more modern church (2002, 29). A further 60 trees in his survey were measured at more than 26 feet in circumference, which Bevan-Jones classed as “more than 1000 years old” (Bevan-Jones 2002, 29 & gazetteer 189-193), including the tree, which is 28 feet in circumference, in the yard at All Saints’, Alton Priors in Wiltshire, illustrated below. Figure 3.11 – Yew tree in the churchyard at All Saints’, Alton Priors (Wiltshire) – 28 feet in circumference (A4-sized clipboard for scale) There remains the question of whether it was bishops, monks, or indeed anyone at all, who directed the incorporation of the earlier pagan sites, such as wells and barrows. John Blair has suggested that many small churches were built on older ritual monuments, where the sites were apparently selected for non-practical reasons and where they may represent sites of long-standing sacredness (Blair 2005, 376). But his observation, noted earlier, that the builders of the early minster churches, “frequently bishops” (Blair 1991, 91), had apparently shown little interest in commandeering earlier ritual sites (Blair 2005, 376), appears to indicate that they had other priorities, possibly the patron’s convenience. It may also be wrong to assume that, in the tenth and eleventh centuries, the church came after the manor; as Blair comments that Chapter 3 88 “placing his church on a perceived sacred site away from the manor, may be an example of the lord trying to gain control of a traditional source of power such as a well or tree” (Blair 2005, 382). This could represent an additional way that undeveloped cult sites were brought into the Christian fold, not specifically authorized or directed by senior members of the church hierarchy, but as a method of imposing local secular power or even as a way of gaining local acceptance. The in-depth studies at Shapwick in Somerset highlight the fact that detailed knowledge of any individual site can reveal so much extra information about its history, that previous assessments of the current church and its location do not necessarily reflect the true situation. In fact, the detailed knowledge provided by excavation, landscape analysis and interpretation at Shapwick (see Gerrard & Aston 2007, particularly 963-981) has raised additional questions which would not have been evident otherwise. In the absence of this research, the location of Shapwick church in the centre of the village would have been totally unremarkable; it is not located adjacent to a manorial centre nor is it close to a feature which could be interpreted as of earlier ritual importance. However, the extensive research there has shown that the final position of the church, and its relationship to the settlement and topographical features, is the result of a complex series of events. An earlier stone-built church, possibly a minster, thought to be of the ninth century and of at least two phases, was built close to a seventh- or eighth-century timber building, which was possibly an earlier church, or possibly a caput hall. Both the ‘hall’ and stone church were built close to a spring (although there were nine springs in the parish in total) and close both to Bronze-Age barrows and to the remains of a Roman building. In addition, there was another adjacent timber-framed building overlain by a stone building which was probably the site of the vicarage. Late-Saxon settlement shift saw the church and manorial curia left isolated and, approximately 250 years later, the manorial administrative functions were moved some 600 metres to a new moated site at the northern end of the current village at the end of the thirteenth century. The church site was moved into the village during the early fourteenth century, where a new larger church was built, albeit with an oldfashioned floorplan featuring a central tower. All this raises the following questions: if the timber building was a caput hall, and it preceded the church, was it sited close to the spring for practical reasons? Was the church built close by for the convenience of the hall residents, or because of the spring? If the earlier timber building was a church Chapter 3 89 and not a hall, was it located where it was because of the spring, the prehistoric barrows or the Roman building remains? Finally, was it only a desire for a larger building that caused the church site to be moved or was it to improve access for the patron and villagers, rather than leaving the church where it was - isolated, as so many others are? Investigating churches with Romanesque towers in Lincolnshire, Stocker and Everson (2007) have concluded that eighteen churches (27%) appear to have been located in relation to some natural or ancient feature in the landscape (thirteen of which were springs and twelve of these were in the Lindsey district), 31% were located in open spaces within settlements and 42% were located close to the manorial centre or due to some other episode of lordly planning or replanning (2007, 61-64). The high proportion of churches in Lindsey built close to a spring suggested to the writers that either the church was an early foundation that preceded the settlement, or that it was thought a more appropriate location within the settlement, rather than next to the manor (2007, 64), but there is no evidence as yet for any of these churches dating from a period when pagan wells were being ‘converted’ to Christianity (2007, 63). Subsequent phases of settlement development have clouded the association of the churches with natural features, so that at present, fourteen of the eighteen churches appear to be associated with lordly planning or open space within the settlement, but can be stratigraphically separated to show their original association (2007, 61 & 65). This work provides another indication that only very detailed analysis allows a true reading of the situation, in that without it, the importance of the springs to the original location of these churches would have been substantially diluted. Expressions of Romanitas An expression of Romanitas was the third type of possible spiritual reuse of earlier sites, where their incorporation was seen as a way to emulate the Christian culture of the later Roman Empire. There are almost 200 parish churches in Britain which are known to overlie or adjoin Roman sites, such as at Caister St Edmund (Norfolk), Castor (Cambridgeshire) and Bewcastle (Cumbria) and “it is plausible that some of these are developments of perceived ancestral burials or reuses of the earlier burial site” (Blair 2005, 377), although it is equally likely that they are examples of Chapter 3 90 Romanitas. Tyler Bell puts the number of churches associated with Roman structures at 160, but suggested that they “may represent only a fraction of the total” (Bell 1998, 2). Tim Eaton has identified the 25 churches located in Roman forts as the earliest acknowledged class of Roman structure consistently reused for Christian purposes, many as missionary churches early in the Conversion (Eaton 2000, 14-17). Bell calls the use of Roman sites a “distinct and purposeful reuse”, in many cases several centuries after the structures’ desertion and ruination (1998, 4). From the seventh century, Saxons saw Roman masonry buildings as synonymous with Christianity, and further, that there was an association between stone buildings and the church, indicated by the number of churches over Roman buildings throughout the country in settlements with placenames derived from stone (Bell 1998, 6; Rigold 1972, 38), for example Lullingstone (Kent), Stansted Abbots (Herts), Whitestaunton (Somerset) and Stanwix (Cumbria). This link is unlikely to be just an attraction to stone as a building material, as most churches were built in wood at this time (Thomas 1986, 121; Rodwell 1985). The reuse of Roman altars, however, seems to confuse the issue somewhat. More than 50 altars are known to have been reused in ecclesiastic contexts (Eaton 2000, 65), but, since some of them have been used in subservient positions, classified as “casual use” by David Stocker (Stocker 1990, 83-88), such as at the base of the chancel arch at Godmanstone (Dorset) (Eaton 2000, 106) and the three altars buried in the footings at St Mary’s, Bisley (Gloucestershire) (Eaton 2000, 101), it seems equally possible that these are examples of either the Christianization of pagan altars at Godmanstone, or merely the eminently practical use of the largest pieces of stone in the lowest courses of the wall at Bisley. Conspicuous sites The use of particularly conspicuous sites for churches can demonstrate elements of both practicality and spirituality. In many cases they are another example of the reuse of sites of earlier importance; many are in elevated positions, providing conspicuous locations such as knolls and hilltops, and have been noted by all writers for churches across the whole country, such as those at Edlesborough in Buckinghamshire, at Holme on Spalding Moor in East Yorkshire and at Godshill on the Isle of Wight. It has been observed that churches dedicated to St Michael frequently occupy hilltop sites and are an indication of spirituality (Morris 1989, 52-56; Stocker & Everson 2006, 82), Chapter 3 91 although the results of this survey show that St Michael’s churches are much more likely to be built on flat lowland sites. Richard Morris also refers to other “locally conspicuous sites” (1989, 69) and quotes many examples of pre-Conquest churches in the flat areas of East Yorkshire built on locally elevated sites, although he suggests that in these cases it was “most likely that the patrons also lived in such positions and in due course provided their eyries with churches” (Morris 1989, 267), and therefore amount to being sited for the convenience of the patron. Other sites which are conspicuous for reasons other than solely their elevation, which Richard Morris describes as having “indications of more than local status” (1989, 69), are an additional example of the sort of site that had earlier significant uses. The main example he quotes was the church at Knowlton in Dorset, which as well as being in the centre of a henge, was also in the later Hundredal meeting place, with a possible parallel at Thwing in East Yorkshire (Morris 1989, 74). In these cases, the sites would have had a prominent place in the minds of everyone over a wide area and their adoption as the site for a church possibly gained kudos for the patron, reflecting John Blair’s comment, noted earlier, that adoption of cult sites may have been an exercise by the patron of gaining the power of the site (Blair 2005, 382). Morris acknowledges that the Knowlton site is often used by others as an example of ‘de-paganization’ (Morris 1989, 72), but offers the possibility that the church-builders may just have adopted a convenient enclosure for burials adjacent to a now abandoned settlement (Morris 1989, 73). It is not possible to determine which of these reasons was behind the original choice of the site for the church, but it highlights the fact that care must be taken when making any assumptions. Chapter 3 92 Figure 3.12 – Knowlton church sited in the centre of the henge, which was also the Hundred meeting place, or was it sited there to be within a convenient enclosure for burials? Figure 3.13 – All Saints’, Rudston with monolith (chief measuring assistant for scale) The site of All Saints’ at Rudston in East Yorkshire is a similar case. Richard Morris makes the point that the church at Rudston was built next to the monolith, and asks “why are there not more churches built close to standing stones” (1989, 82)? One Chapter 3 93 possible explanation is that the Rudston site is also on top of a hill, so this ‘conspicuous site’ may have taken precedence over the fact that the monolith was there, or perhaps it reinforced the choice of the prominent site. The fact that there are few churches close to standing stones does suggest that they were not considered important enough to require Christianization and reaffirms the argument that Christianization may have been generally overstated as a concept. At Alton Priors in Wiltshire, however, two felled megaliths are set beneath the church floor, accessible by trapdoors (Surman 2008, 5-6), although whether the church was sited over them for the purpose of Christianization or not, and whether the stones have always been accessible, rather than buried, is unclear. These examples, particularly at Knowlton and Rudston, reiterate how complex the whole issue of church location is, particularly when coupled with the position in the hierarchy that the church patron occupied. This subject is developed in later chapters by using the results of this survey. CHURCH LOCATION: CONCLUSIONS The location of churches has been shown to be influenced by many, often inter-related, factors. In many cases the earliest churches were established on the ‘best’ sites in terms of topography, but whether that was for practical reasons, such as for the convenience of the patron who had already chosen to live there, for avoiding flood-prone land or for conspicuousness, is not clear. Writers have discussed a wide range of influences on the location of churches of every age, size and type, ranging from the incorporation of earlier ritual sites to the simple desire for the church to be close at hand. The incorporation of earlier ritual sites appears to have continued to operate over several centuries, although whether its operation, or focus, was as organized or as comprehensive as the church hierarchy appears to have originally intended, is unclear. John Blair feels that a fundamental distinction could be drawn between local churches which stand where they do because their sites had sacred significance, and those which grew from the contemporary locational needs of manorial settlement and had no roots in any ritual past (Blair 2005, 373). What is clear from the research at Shapwick and from some of the cases in Lincolnshire, noted earlier, is that detailed local work on a site-by-site basis can change the apparent relationship between church, manor and topography to such an extent that summary conclusions drawn without such Chapter 3 94 detailed analysis, especially at single sites or from very small datasets, should be treated with care. Statistical reliability of the results grows as the sample size grows, especially as it appears from hundreds of other sites in Somerset, Norfolk, Yorkshire and Lincolnshire that the moving of the church at Shapwick was an unusual exercise, and that the vast majority of churches stayed in their original location even when the settlement shifted away. Applying the results of this large survey in later chapters to the location of churches, and separating sites chosen principally for manorial reasons or for topographical reasons will be made more difficult when taking into account the many east-facing sloping sites that have churches built on them. Chapter 3 95 SETTLEMENT STUDIES This section outlines the historiography of rural settlement change, from the mid-Saxon period to about a century after Domesday, which was an important period for rural church building. It outlines the earlier views of settlement and discusses the more recent major writings about rural settlement, particularly the patterns that developed in the different regions based on the major landscape/agricultural classifications of ‘planned’ and ‘ancient’ countryside. It also touches briefly on the parallel development of open-field agriculture. Although the specific types, and patterns, of settlement that developed across the country are not the primary interest of this thesis, the timing of the appearance of the church in the settlement and its specific location within the settlement are matters of central importance, and will form a major part of the analysis in the final chapters. Formalizing terms “used by writers for several centuries”, Oliver Rackham’s definitions of ‘planned’ and ‘ancient’ countryside and the ‘highland’ area use different levels of settlement dispersion (in other words hamlets, villages & isolated farms) as one of the eight distinct differences between the regions (Rackham 1986, 4-5). Planned countryside has been described by others as ‘Champion’ or ‘Midland’, and ancient countryside has been described as ‘Woodland’ or ‘South-east’. Rackham notes an often sharp division between the types of countryside which do not reflect the local administrative boundaries and frequently cut parishes in half (Rackham 1986, 5). The three types of countryside are shown in Figure 5.1 below, along with the boundaries of the areas surveyed for this thesis. and it has been recognized that whilst “each area is broadly homogeneous, with many broad similarities in the types of rural settlement”, both in their formation and changes in them, there are also some variations within the areas, especially at the margins (Taylor 1983, 77-85; Williamson, 2003 5-6), and there are different opinions on the timing of settlement nucleation (Lewis et al. 2001, 201 & 214; Taylor 2002, 54; Williamson 2003). Different writers have offered slightly different versions of the boundaries of the two types of countryside, including those arrived at when measuring nineteenth-century dispersion of settlement in An Atlas of Rural Settlement (Roberts & Wrathmell 2002, 8-9), “but all agree that the Midland areas were [later] characterised by more nucleated patterns of settlement” (Williamson 2003, 4). Chapter 3 96 Figure 3.14 – Landscape classifications after Rackham 1986, showing the areas of churches surveyed for this thesis Chapter 3 97 Early settlement development Before the 1970s it was generally assumed that villages were established by the AngloSaxon people who colonized England in the fifth century, after the Romans left. The subject was considered in Hoskins’ publication The making of the English landscape in 1955 and he supported the idea by writing that “the English landscape as we know it today is almost entirely the product of the last fifteen hundred years, beginning with the earliest Anglo-Saxon villages in the middle decades of the fifth century” (Hoskins 1955, 20). He went on to describe several different village types – Green villages, Street villages and Crossroad villages, and he also referred to “mixed types” which he ascribed to “later changes” (1955, 60), but which may have been the polyfocal settlements later described by Christopher Taylor (1977) and Trevor Rowley (1978). Hoskins also referred to planned villages, which he thought were “mostly of eighteenth century date” (Hoskins 1955, 60). He felt that all these various types of villages reflected cultural and historical differences, but acknowledged that “the result may be due to successive changes” (Hoskins 1955, 60). Similar settlements to those noted by Hoskins (green villages and street villages) were built in Holland and northern Germany, and it “might legitimately have been expected that similar types were built here” (Rowley 1978, 72), except that there is a “fundamental flaw in this assumption in that there is no evidence that villages were built [anywhere] by the early Saxons” (Rowley 1978, 72). By this time, thinking had changed considerably and there were “many theories on village formation, but all were united in their opposition to the traditional beliefs” (Rowley 1978, 70). It was suggested that “the traditional view of nucleated villages [only] provided a comfortable picture of stability and antiquity in our landscape” (Taylor 1983, 110). In fact it was proposed that villages were “an aberration” in the landscape (Taylor 1983, 125; Aston 1985, 82) and that in many cases the hamlets and farmsteads that predated the villages were in some way a more normal form of settlement (Aston 1985, 82; Brown & Taylor 1989, 61), whether through forest clearance (Taylor 1983; Williamson 2003), continuity of the Roman pattern of scattered settlement (Hoskins 1955; Rodwell & Rodwell 1985; Williamson 2003), or through development in late Saxon times (Brown & Taylor 1989, 61; Warner 1991). Researchers in many areas across the country have “uncovered a continuous and often extensive change in medieval settlement, with the former idea of village Chapter 3 98 stability swept away” (Taylor 1992, 7), revealing a similarity of dispersed settlement types that were common to all areas before the communal systems of agriculture with their nucleated villages in the Midlands (Taylor 1983, 63-108; Williamson 1988, 9-10). Scattered early- to middle-Saxon sites, with several to a parish, have been discovered in many places, although it is not always clear whether these were contemporary or successive. For example, excavations at West Stow (Suffolk) and Mucking (Essex) were interpreted as shifting clusters of houses (Hamerow 1991; Taylor 1992, 8), whereas at Brixworth (Northamptonshire), it has been suggested that they may have been contemporary, or at least evidence for early nucleation (Jones & Page 2006, 88). Another example is Goltho in Lincolnshire, where excavation “provides the first real date for an English village on the site”, as being laid out in the tenth or eleventh century on a pattern unrelated to the eighth-century timber houses beneath it (Taylor 1983, 122; Beresford 1975), in a similar way to Bishops Waltham (Hampshire) and Brandon (Suffolk) (cited in Taylor 1983, 122-3). These examples have confirmed that there was “an early- to middle-Saxon pattern which was different from, and unrelated to, the later medieval nucleated villages” (Williamson 1988, 6; Taylor 1992). A pattern of dispersed settlements has also been found in Buckinghamshire and Leicestershire, with small outlying sites revealed by scatters of mid-Saxon pottery, but which seldom produce late-Saxon pottery on the same site (Lewis et al. 2001, 81), indicating their disappearance by then. During this period, a pattern of large estates, which contained many hamlets and scattered farmsteads, that may have perpetuated boundaries from at least Roman times, has been suggested in the past, particularly by Glanville Jones (Jones 1961 & 1979). The idea was developed and summarised by Mick Aston as “… a series of large units in the landscape from the seventh century onwards. These estates were centred on some of the most significant places in the landscape with all other settlements dependent on them and may be reflected later on in the lands of the old established monasteries and bishoprics, or in some of the Hundredal arrangements” (Aston 1985, 36). Examples were suggested by Glanville Jones across much of the country – Wales, Yorkshire, Hertfordshire, Sussex, Wiltshire, Cumberland and Durham (Glanville Jones 1961 & 1979; Aston 1985, 34), thus demonstrating that a similar form of dispersed settlement was common to all areas of the country, irrespective of soil type and region, Chapter 3 99 prior to the divergence of settlement patterns that later characterized the planned and ancient countryside areas. Despite Rackham’s comment, noted earlier, that there was often a sharp division between the two areas of countryside, many areas at the margins of the main landscape regions exhibit characteristics of both. As Christopher Dyer puts it “Village England runs down the middle of the country … areas of dispersed settlement lie on either side, although there are pockets of farms and hamlets in village-dominated regions and there is a scatter of nucleated villages in every corner of the country” (Dyer 1994, 47). Christopher Taylor points to “marginal areas of mixed settlement” (Taylor 1983, 77-85) and Roberts and Wrathmell describe “sub-provinces” in south Huntingdonshire and adjacent areas, in which a mixture of ‘Central’ and ‘SouthEastern’ settlement characteristics appear (Roberts & Wrathmell 1998, 102). Scholars broadly agree as to why nucleated settlements and open field agriculture appeared; shortage of pasture, innovations in ploughing technology, partible inheritance, rising population, administrative reasons such as tithes and parishes, as well as market surplus arising from improved techniques of open-field farming, have all been considered (Taylor 1983, 125-150; 2002, 53; Brown & Foard 1998, 82; Williamson 1988, 7; Lewis et al. 2001, 193), in addition to soil (Williamson 2003) and lordly power (Hamerow 2002, 87; Williamson 1988, 7; Dodgshon 1980). However, recent work in Northamptonshire has highlighted the substantial amount of grazing land remaining after the open fields were fully developed (up to 40% of the area), suggesting that re-assessment is needed of the idea that it was the shortage of pasture for grazing that led to the need for fallow fields, and therefore the ‘three-field’ system (Williamson in prep.). An additional problem is that the ‘pressure of population’ effect was not realised on all the areas of heavy clay. The claylands of Essex, Norfolk and Suffolk, in the ‘ancient countryside’ area, were not subject to the same settlement developments as the claylands of the midland counties (Williamson 1988, 10), even though the Domesday population density was greater there than in the Midlands (Darby 1976, 46). Christopher Taylor asks whether nucleation was an evolutionary process that started in one place and spread (Taylor 2002, 54), which is also hinted at by Carenza Lewis (Lewis et al. 2001, 214), but, if nucleation spread from the centre of the area, then villages at the edge would be expected to have nucleated later than those in the Chapter 3 100 centre (Taylor 2002, 55), a situation which has not been identified by any of these writers’ published researches. Chronology The chronology of the change is more problematic. The original theories of fifthcentury Saxons importing the village structure and agricultural technique have been shown to be wrong, and at the other extreme, Joan Thirsk put village formation as late as the twelfth or thirteenth centuries (Thirsk 1964), which has also been disproved; both from documentary sources, with Saxon charters referring to headlands, furlongs and yardlands (Pollard & Reynolds 2002, 219) and from field-walking finds of concentrations of earlier pottery (Jones & Page 2006, 81). There are two current schools of thought on chronology; firstly a middle-Saxon phase of nucleation of the late eighth or early ninth century (Taylor 1983, 130; Williamson 2003, 66), with “much of the nucleation completed before the middle of the ninth century” (Williamson 2003, 67), although some have suggested a seventh-century start, such as at North Elmham in Norfolk (Wade-Martins 1980a), and others have suggested a nucleation in the late seventh or early eighth century, with open fields started in the ninth or tenth century as part of the replanning of agricultural methods (Brown & Foard, 1998, 91). The second school of thought involves a longer chronology with a later beginning, suggesting settlement nucleation starting between the middle of the ninth century and the eleventh century, but not finishing in parts until as late as the thirteenth century (Lewis et al. 2001, 81; Taylor 1983), a period which has been called the “village moment” (Lewis et al. 2001, 191) and partly echoing Thirsk’s suggested timing, mentioned earlier. There is also considerable discussion over the issue of whether settlement change or replanning was a lordly effort or organized by the peasants through economic necessity. The majority of authors seem to favour the idea of lordly replanning of settlements and have suggested that it fits in well with the other changes during the period; of feudal transformation, with an increase in the number of local lords (Faith 1997, 1-15), coupled with urban growth, rapid population growth and a concomitant increase in demand for food production (Taylor 1983; Lewis et al. 2001, 194-195; Gerrard & Aston 2007). Bruce Campbell has doubts as to whether communities were sufficiently organized and capable of making the appropriate Chapter 3 101 decisions before the commonfield systems came into existence, and that "strong and undivided lordship” was more likely (Campbell 1981, 127), a position echoed by Richard Hodges, who considered that the “nucleation of villages cannot be ascribed to the collective zeal of the peasantry … it was surely the work of the manorial class” (Hodges 1989, 168). Carenza Lewis sees an obstacle to the assumption that lords were the principal agents in the planning of villages and fields by noting that “there are frequently differences between the boundaries of manor and township” (Lewis et al. 2001, 175), implying that the two are not contemporary features. Christopher Taylor, however, has also argued that nucleation could have been brought about by a conscious decision by a group of residents to occupy a single site, rather than being a function of lordship (Taylor 1983, 133), a similar view to that of Christopher Dyer concerning the development of villages and open field farming (Dyer 1994, 11). Development of open-field agriculture Some feel that there is a “strong likelihood” that open fields came into existence at the same time as the nucleation of villages (Lewis et al. 2001, 171), and Margaret Faull (1984) also saw village and agricultural development happening in parallel, between the ninth/tenth century and the twelfth century. There is pottery evidence of scattered Saxon settlements covered by later open fields (Foard 1978, 364), although this just appears to show that scattered settlements preceded the fields, but not by how long. In other words, this evidence could support either a contemporary change to both settlement and agriculture, or a later change in agriculture. Carenza Lewis mentions this as part of the problem with an evolutionary model of growth of nucleation, and suggested that once the change to nucleation was started in a particular vill it had to happen quickly, perhaps within a year, in order to avoid the conflict of land use between field and scattered settlement, clearly meaning that settlement nucleation and the creation of open-fields had to be contemporary (Lewis et al. 2001, 171). However, others felt that the open-fields may have been developed some time after the village nucleation process. As Tom Williamson puts it, “exactly what form that those open fields, as they existed by the twelfth and thirteenth centuries, took in late-Saxon times, is still open to debate” (Williamson 2003, 89), with suggestions by some that there may have been an intermediate form of communal agriculture, known as ‘long furlongs’ (Hall 1981, 36-37), although others have suggested that it was basically pastoral Chapter 3 102 (Brown & Foard 1998, 91-92). Whether there was a contemporary change in agricultural methods or not, there must have been “some reorganization of the lands that were attached to the hamlets as they combined” (Brown & Foard 1998, 80). At Shapwick in Somerset, the village nucleated in the tenth or eleventh century in the centre of a new larger parish, based on the lands of at least two earlier hamlets or manors (Gerrard & Aston 2007, 980-981). The new village was located in the centre of the new open fields, for which fieldwalking has uncovered a late-Saxon manuring scatter slightly in excess of 200 acres (Gerrard & Aston 2007, 980-981); what cannot be resolved is whether the creation of a new village and the introduction of open-field farming there were contemporary. ‘Ancient countryside’ settlement In ‘ancient countryside’ areas, for example on the clays of central Norfolk, early-Saxon settlement continued the dispersed character of Roman settlement, although more sparsely, but during the middle Saxon period, settlement began to be concentrated, usually only one settlement to a parish (Wade-Martins 1980b; Davison 1990, 16-19; Newman 2001, 7; Williamson 2002, 95). This area was characterized in later times by a variety of dispersed settlement forms, in many cases surrounding commonland, which “began developing in the eleventh, twelfth and thirteenth centuries” (Williamson 2003, 92; Martin 2001, 6; Warner 1991, 13). In the majority of cases where fieldwalking has been possible, similar foci to the middle-Saxon sites for settlement were retained in the late Saxon period (Wade-Martins 1980b; Davison 1990), although some of these identified middle-Saxon sites may only have grown into villages in the late Saxon period (Williamson 1993, 85). On the lighter soils to the east of the claylands in East Anglia, “a similar pattern is apparent … … the movement away from middle-Saxon sites to the common edge evidently began in pre-Conquest times”, and is “probably typical” (Williamson 2003, 98). Williamson feels that this indicates that there were strong forces operating in pre-Conquest times that prevented the growth of the nucleation of settlement, rather than settlement dispersing to the common edge (Williamson 2003, 98-99). Further south, in Suffolk and north Essex, where there were fewer large commons, most parishes had a cluster of settlement, but with numerous small isolated settlements often close to small greens (Warner 1991, 28-29; Williamson Chapter 3 103 2003, 101-2), a proportion of which became villages before Domesday, the rest remaining as subsidiary hamlets (Warner 1991, 29). Highland Zone settlement This zone has also been referred to as the “Northern & Western Province” (Roberts & Wrathmell 2002, 5). Settlement patterns in the Highland Zone are considerably simpler, basically remaining dispersed throughout the period of interest here, but with exceptions particularly along the coast and in the larger valleys (Roberts & Wrathmell 2002, 116-117), where some village nucleation occurred with sizeable associated township fields. Most of the area concentrated on an agricultural economy that was basically pastoral in nature, with settlement and farmsteads scattered along valley floors with associated upland grazing sites (Roberts & Wrathmell 2002, 162-170). SETTLEMENT: CONCLUSIONS There appears to have been a similarity in settlement patterns across much of the country at least until middle-Saxon times and possibly after; a dispersed settlement of farmsteads and hamlets, often assarted from woodland, which appeared, shifted, and disappeared, from the Roman period until the beginning of nucleated village settlements across most of the country during either the middle-Saxon or late-Saxon period, depending on which school of thought is followed. After this point a divergence in pattern began; in the ‘ancient countryside’ areas the nucleated settlements began to break up with some, if not most, settlement moving towards the edges of grazing land (the commons), during the post-Conquest period and possibly just before it; whilst in the ‘planned countryside’ areas, the nucleated settlements remained on the same site. There were some differences in patterns at the edges of the two countryside areas, with pockets of one type of settlement contained within the boundaries of the other. The question of when ‘planning’ started – in early or middle-Saxon hamlets, or only as part of the settlement nucleation process, or possibly when open-field agriculture started – is still debated. In addition, the question whether this was ‘lordly replanning’ or some form of peasant communal effort is also unclear. From a practical Chapter 3 104 point of view, it makes sense to locate the nucleated village close to the centre of the combined fields, rather than close to one boundary, leaving the question of whether settlement nucleation only happened on a new site, rather than by the growth of one of the constituent hamlets, when none of the hamlets was conveniently placed in the new combined lands? In later chapters, the results of this survey, particularly in Norfolk and Suffolk, will be used to assess whether a local graveyard was set up as a contemporary process with the nucleation of the village, and then the church was built later on the graveyard site, or whether the nucleation of the settlement was followed by the building of a church on a new site much later. Chapter 3 105 PART TWO SURVEY DESIGN AND BASIC RESULTS APPLIED TO EARLIER THEORIES Chapter 4 106 CHAPTER FOUR THE SURVEY METHOD AND BASIC RESULTS But churches face east, don’t they? (Anon: many times to the author, 2000 - 2008) INTRODUCTION This chapter describes the survey of 1,926 medieval rural parish churches conducted for this thesis. It explains the sample selection; outlines the survey methodology; presents the main results and provides an initial analysis of them. SURVEY SAMPLE In order to provide a large dataset and one which would allow an area-based analysis, the survey has a structured sample which covers a geographic spread of fifteen counties across the country (Cumbria, East Riding of Yorkshire, Shropshire, southern Lincolnshire, northern Cambridgeshire, Norfolk, northern Suffolk, northern Oxfordshire, Bedfordshire, Pembrokeshire, northern Somerset, southern Hampshire, East Sussex, eastern Kent and western Cornwall), shown in Figure 4.1. Some of these areas use county boundaries that predate the boundary reorganization of 1974, because some of the volumes of Pevsner’s Buildings of England, on which the sample was based, were written and printed prior to this date, and references to the churches will use these old County names and boundaries throughout this thesis. All of the rural medieval parish churches in the resulting areas were surveyed, except those described in the relevant volume of Buildings of England as having had their naves rebuilt in relatively recent times. The exclusion of these churches was to ensure that a postmedieval, especially Victorian, rebuild had not affected the alignment of the church. Where Pevsner noted that the nave was rebuilt above a particular level, for example above the window cills, and the earlier construction could still be seen, measurements were taken from the latter. In some counties (Cambridgeshire, Cornwall, Hampshire, Chapter 4 107 Lincolnshire, Kent, Oxfordshire and Somerset) only part of the county was surveyed due to limitations in the time available for the survey. In these cases, the survey was started at one end of the county and as many as possible of the churches that fitted the parameters were surveyed during the time available, after which a suitable north-south or east-west Ordnance Survey grid line was chosen as a cut off point, prior to the commencement of the analysis of the data. The part of northern Suffolk that was surveyed represents the extent of the Waveney Valley catchment in Suffolk which was the subject of the initial survey, undertaken for a Diploma in 1999, and subsequently incorporated into the larger survey. Figure 4.1 – Survey areas Chapter 4 108 Data were collected about each church building itself, about its immediate surroundings and about the wider landscape. For the church, these included the floorplan; the type of tower; the size of the nave and chancel; the alignment of the nave and chancel; the height above sea level; the latitude and longitude and an assessment of whether the building lay parallel to the churchyard’s principal boundary. For the churchyard, they included overall size; the closest proximity between church and yard boundary; the topography of the site and the angle and direction of any slope of the yard. For the wider landscape, an assessment was made of the surroundings, including the presence and location of any adjacent large house, the location of the village that the church served, and the current morphology of the village. For almost two-thirds of the churches (1,105), measurements were also made of the elevation of the eastern horizon; this was started when it became clear that the horizon elevation was important, particularly when dealing with sunrise. A completed example of the form used for the collection of the data is shown as Appendix 3 on pages 325-326. SURVEY METHOD To assist in standardizing alignment readings, especially on undulating walls or those built of materials such as cobble flint, they were taken with a Silva Type 15 compass fixed to a piece of hardwood 75 centimetres in length (with brass screws). Where possible, readings were taken inside the church, two on each side of the nave and two on each side of the chancel. If external readings were required due to lack of access, three were taken on both the north and south sides of the chancel, and of the nave, in an attempt to remove anomalies resulting either from local magnetic variations caused by iron in, or near, the walls, or caused by north and south walls not being exactly parallel. If there were differences of more than one degree between the readings for either part of the building, they were retaken at different places. A mean was taken of the results, to provide single readings for the nave and for the chancel. Finally, the magnetic compass readings were adjusted to True readings by deducting the contemporary magnetic declination in the area, as listed in Appendix 4 on page 327. The angular elevation of horizons of churches in eight of the counties in the survey, Bedfordshire, Cumbria, Hampshire, Lincolnshire, Norfolk, Oxfordshire, Pembrokeshire and Suffolk were taken with the same Silva compass as the church alignments. Placed on its side Chapter 4 109 and held against the south wall of the chancel, the horizon was sighted along the piece of wood on which the compass was mounted and the reading taken from the integral clinometer, in degrees. The slope direction of the churchyard was measured by taking a bearing of the lowest part from the highest part using a prismatic sighting compass. With practice it is possible to get repeat readings, taken blind at different times, within two or three degrees of each other, so, since the lowest and highest parts of a churchyard are usually easy to assess, the readings should be within two or three degrees of the actual slope direction. On difficult sites, a backsight was taken from the lowest part of the yard towards the highest, and an average reading taken. To reduce any possible bias in the readings, each was rounded to the nearest five degrees. The slope of the yard was assessed at the church itself. The actual drop in the land was measured on the church walls, and then calculated as a percentage slope measured against the length, width or diagonal measurement of the church as appropriate. Urban churches were omitted from the main survey in order to avoid buildings that were more likely to have been influenced in alignment by the character of their site. Much has been written about this subject in the past (for example: Biddle 1976b, 20-22, Morris 1989, 208-209; Rodwell 1984) but a supplementary sample of 70 urban churches (in Norwich, York (some of which follow the alignment of the original Roman street pattern noted in Chapter One) Lincoln and Stamford) was surveyed. They exhibit a completely different pattern of alignment from the rural sample, which appears to support the earlier writers’ assertions of alignment influence. The rural churches of this survey display a statistically “normal” result graph, described as a bell curve, whereas the urban sample displays no apparent pattern in alignment, being fairly equally spread across their range of alignments, which, despite the small numbers involved (70 churches), has exactly the same range as the considerably larger rural group of 1,926 churches, both varying by 90º between 38º and 128º True. The graphs of the alignment of urban and rural churches are shown at the same vertical scale below. Chapter 4 110 Alignment of Rural Churches 400 EAST 350 300 Figure 4.2 – Rural and Urban church alignment from this survey - comparison at the same vertical scale 250 200 150 100 Southeast Northeast Alignment of Urban Churches 50 50 EAST Northeast degrees from North 0 38-42 43-47 48-52 53-57 58-62 63-67 68-72 73-77 78-82 83-87 88-92 93-97 98-102 103-107 108-112 113-117 118-122 123-127 128-132 38-42 43-47 48-52 53-57 58-62 63-67 68-72 73-77 78-82 83-87 88-92 93-97 98-102 103-107 108-112 113-117 118-122 123-127 128-132 0 Southeas degrees from North THE SURVEY RESULTS FOR MEDIEVAL CHURCHES Analysis and results of an interim stage of this survey were written up as part of a Masters dissertation (Hinton 2003). At that time, the total survey amounted to 993 churches in nine counties. An additional 933 churches in a further six counties have since been surveyed, bringing the total to 1,926 churches in fifteen counties. The overall results, analysed here, closely confirm the earlier findings and bring an even greater statistical validity to the results. They have also allowed new conclusions to be Chapter 4 111 drawn about the causes of alignment variation due to the improved geographical coverage and the inclusion of counties further from the coast. The mean alignment for all churches in the survey is 86.1º (interim survey 85.8º). This confirms the results of earlier writers, discussed in Chapter One, where the mean alignments ranged from 85° to 90º, and had a mean value of 86.0º. None of these earlier surveys splits its results within the country, either because the sample was too small or because it concentrated on a single area. It has never been possible to compare results between studies due to the different methodologies used with unknown levels of accuracy. Consequently the differences in alignment between the east and west of the country, shown in the table below, have not been revealed previously. Table 4.1 Summary of church alignment by County western Cornwall Pembrokeshire Cumbria Shropshire northern Somerset East Riding Yorkshire northern Oxfordshire Bedfordshire southern Hampshire northern Cambridge southern Lincolnshire East Sussex Norfolk northern Suffolk eastern Kent TOTAL Number 72 77 74 104 91 110 130 96 76 123 103 104 549 125 92 Range 50-111° 48-116° 62-104° 55-126° 54-107° 51-111° 47-107º 38-107º 59-116º 57-121° 45-103º 54-118° 56-128º 65-119º 58-120° Mean 80.4°° 82.0° 82.3°° 82.4°° 82.5°° 83.0°° 83.1º 84.4º 85.6º 86.2°° 86.5º 86.8°° 88.9º 88.6º 92.4°° % North of East 76 77 76 71 74 76 70 75 62 67 69 57 56 55 40 1926 38-128º 86.1º 63.8 The results analysed by county clearly show a difference of 12° (80.4° - 92.4°) between the mean alignments of churches in Kent and Cornwall. The use of statistical confidence limits in this table is not applicable since in each county all the churches that fit the criteria were surveyed, so the ‘sample’ and the ‘whole group’ are one and the same. The additional counties surveyed since the interim survey (Norfolk, Hampshire, Lincolnshire, Bedfordshire, Oxfordshire and Pembrokeshire) were all selected to confirm or deny the pattern of alignment variation that had been observed in the interim survey across the country; in the cases of Oxfordshire and Hampshire, by Chapter 4 112 providing results for the centre part of the country which was not represented before; in the case of Norfolk, to provide a larger survey base in the east of the country to check the consistently higher mean alignment already observed (and because it was close to home); while Pembrokeshire was added to the sample to give additional weight in the west of the country. The results for each of these counties fit the previously identified pattern, with mean church alignments in Oxfordshire, Hampshire, Lincolnshire and Bedfordshire fitting numerically between those for Somerset and Cambridgeshire, and the mean alignment in Norfolk matching that from the previously surveyed areas of northern Suffolk. The results in Pembrokeshire are very similar to those from the other counties in the west. Broadly speaking, the numerical range between the most northerly and the most southerly individual church alignments in each of the counties surveyed is similar; it is the concentration of the numbers of churches aligned to one side of east or the other which causes the mean results to be different. In eastern Kent only 40% of churches are aligned to the north of east, shown in the final column in table 4.1, whereas almost twice as many, 76%, are aligned north of east in Cumbria and western Cornwall, and 77% in Pembrokeshire. Possible reasons for this pattern will be discussed in Chapter Six. Table 4.2 below shows the variation of the alignment results each side of the mean value for each of the categories in the survey, for example, the mean alignment value for each the categories of “Churchyard topography”; Flat, Almost flat, Sloping, Platformed and Knoll, varies by 0.4°, or less on one side of the mean figure, and by 0.7° on both sides added together - the overall mean is 86.1° and the mean for each of the categories is 85.8°, 85.9°, 86.1°, 86.4° and 86.5°. This stability of the results applies to all of the first eight factors in the table below, relating to churchyard topography and size; the planform, and size, of the church; or its number of aisles; whether the church is built parallel to its principal boundary and how close to it; as none of them varies by more than three degrees from the overall mean direction. It is not being suggested that any of these factors is actually likely to have influenced the alignment of the church when it was first built, (except perhaps the proximity to, or parallelism with, the principal churchyard boundary) but rather that a significant difference in one of them may have highlighted an element that required further investigation. Location of the church in relation to its village is a special case. Without taking into account the small Chapter 4 113 group of nineteen churches (out of a total of 1,926) with either ‘no village’ or a ‘scattered settlement’, the total variation in mean alignment falls from 4.6º to 1.4º. The remaining categories where the church is either located in, or on the edge of, the village, or is ‘remote’ or ‘isolated’ from the village, are far closer to the overall mean alignment. Table 4.2 - Variation in mean alignment by category When analysed by:Churchyard topography Season of the patron saint Church/churchyard ratio Church planform Whether parallel to boundary or not Number of aisles Proximity of churchyard boundary Size of church Location in relation to village (see text) Latitude Type of tower Height above sea level Longitude Maximum variation one side of mean value - in degrees 0.4 0.6 0.9 1.1 1.4 1.4 1.8 1.4 3.9 (0.7) 2.6 2.4 3.9 4.7 Total variation both sides of mean value - in degrees 0.7 0.7 1.4 1.5 1.8 2.3 2.4 2.6 4.6 (1.4) 3.9 4.2 5.4 8.7 In the analysis of the last four factors in the table, latitude, type of tower, height above sea level and longitude, church alignment varies by a considerably greater amount either side of the mean value. Both of the first two can be explained because of their relationship to longitude in the areas surveyed. The major variation in mean alignment between churches built on sites less than 30 metres above sea level (87.6º) and those built on sites above 110 metres above sea level (82.7º) (for the detailed figures, see table 4.9 on page 122), is explained by the fact that counties in the east of the country; Norfolk, Suffolk and Cambridge (where the mean alignment is high – 89º) have by far the highest proportion of churches built on lowland sites, whereas in the west of the country, Cumbria, Cornwall and Shropshire (where the mean alignment is low - 82º) have the highest proportions of churches built on upland sites. Similarly, the type of tower varies by longitude. The higher mean alignment for round-towered churches (87.9º) is because they are almost exclusively found in East Anglia, where alignments are generally higher (see table A5.2 on page 329), and churches with no Chapter 4 114 towers (mean alignment 83.7º) are over represented in the west of the country, particularly Cumbria and Pembrokeshire. The variation in mean alignments by latitude can also be explained by underlying differences in the longitude of the survey areas. The low mean alignments for the northernmost and southernmost categories (around 83º) are reflections of the fact that these survey areas are located in the west of the country, and the higher value (87.7º) for the area between latitude 52º 30′ and 53º North is a reflection of the large number of surveyed churches in the east of the country at this latitude, particularly in Norfolk. Quite why the variation in mean church alignment can apparently be described by variations in longitude is not at all clear. There is no obvious reason why any difference in longitude itself should cause, or be reflected in, differences in the alignment of churches. It must be remembered that the ‘mean alignment’ figure that has been used as an indicator of difference of alignment between categories, and between areas, is a reflection of the skewed nature of the data around east, in other words, a numerically lower mean alignment indicates that a greater number of churches in that category or area is aligned to the north of east (having a numerically lower value). This complicates the situation even further. Why should the difference in longitude between Cornwall and Kent cause twice as many church builders in Cornwall to align their church to the north of east compared with those in Kent? There may be another variable which could explain this difference and which is linked in some way to longitude, but it is none of the physical and topographical factors that have been surveyed here. These questions will be discussed in more detail in Chapter Six. Analysis of church alignment by each of the elements of the church and churchyard characteristics that were listed in table 4.2 is set out below in the same order as shown in the table. In a few cases, the data tables, and the analysis of alignment, for aspects of the church building itself, such as planform, number of aisles and type of tower are presented in Appendix 5 as the figures show little variation, are internally consistent, and add little to the analysis. The remaining category shown in the table – season of the patron saint – is discussed in more detail in the next chapter, where the survey results are used to test one of the earlier published theories of alignment, that churches were aligned with sunrise on the feastday of their patronal saint. Chapter 4 115 Churchyard topography Without any data, it might be assumed that a sloping site, particularly a steeply sloping one, would have an effect on the alignment of the church. The stresses within the building are considerably more complicated if a church is not built either up and down, or across, a slope. However, as the figures will show, there is almost no difference at all (0.2º) between the mean alignment of churches built on slopes and the overall mean, and with the same proportion of churches (63.8% - 64%) aligned to the north of east. One in five of all the churches in this survey was built directly on a slope of greater than two per cent (with no levelling of the site). This represents a slope of at least one in fifty, or approximately 60 – 100 centimetres over the length of the churches in this survey. There are 131 churches built on slopes steeper than one in twenty, and fourteen on slopes greater than one in ten, with the steepest slopes close to one in six; at the steepest site, St Issels in Pembrokeshire, the land falls 3.3m over the church length of 19.4m. The slope of two per cent was chosen as a cut off value so as to provide a large enough sample for analysis, but little variation occurs if only churches built on steeper slopes are examined. There are 373 churches in this survey that are built on sites with a slope of two per cent or more, with a mean alignment of 85.9º, 64% of them are aligned to the north of east. The 131 churches built on sites with a slope of five per cent (1 in 20) or more have a mean alignment of 84.0º, still only 2.1º from the overall mean, and 70% are aligned to the north of east, but they are disproportionately represented in the west of the country, where alignments are numerically lower. It appears, therefore, that neither gentle, nor steeper, slopes have had any real affect on church alignment. In many cases the slope of part of the churchyard has been artificially altered to create a more level surface on which to build the church. Described here as “platformed”, they were created either by raising the ground at the lower end of the slope; by building into the slope at the higher end, with retaining walls to hold back the land; or a combination of the two. What is difficult to assess is whether a platform relates to all phases of the building of the church, or whether it was created to accommodate a later extension, or remodelling, of the church. But either way, it is clear that any re-profiling of the churchyard slope, by the making of a platform, appears to have had no effect on the alignment of the churches built on them, with a mean alignment direction within 0.4° of the overall mean. Levelling part of a sloping Chapter 4 116 churchyard to enable the church to be built more easily also provided more flexibility in the setting out of the building, allowing the builders to align their church closer to a specific target if there was one, without having to deal with the constructional problems of building diagonally down a slope. However, the resulting 283 churches have almost as wide a range of alignments as the overall sample (between 50° and 128º), so they apparently chose not to use the opportunity of building on a level platform to align their churches any differently. Similarly, the desire, in some cases, to build churches on high points in the landscape, whether natural or artificial, also has an impact on the assessment of slope in relation to the church. In the majority of cases where churches are built on ‘knolls’, the land slopes away from the church in three directions and in many cases, in all directions. Even though there may be a slight overall slope from one end of the church to the other, with a rise in the centre, this has not been included in the analysis of slope. Even this relatively small group of 118 churches is aligned in a similar way to the whole sample, its mean being exactly the same as the overall mean. In summary, it appears from these figures that by itself the topography of the churchyard has had no measurable effect whatsoever on the alignment of the church. The presentation of 95% confidence figures in the following tables is to show that, for example from table 4.3 below, the mean alignment of all churches built on slopes greater than two per cent is likely to be within 1.3° of the mean figure of 85.9° measured in the survey. The small values for each of the confidence figures shown in the following tables indicates the general robustness of the survey figures. Table 4.3 – Overall results by topography of churchyard Slope of yard Slope >2% Flat Almost flat Knoll Platformed Chapter 4 No Range Mean 373 783 369 118 283 1,926 48-121 38-126 59-118 59-119 50-128 85.9 85.8 86.4 86.1 86.5 86.1 117 95% Conf. ±1.3 ±0.8 ±1.1 ±2.0 ±1.4 ±0.5 % N of East 64 65 63 64 61 63.8 Ratio of church floorspace to churchyard area It might also be assumed that the size of the churchyard would have had an effect on the alignment of the church, where, on a restricted site, a free choice of alignment might not have been available. Although urban churchyards often show signs of reductions in the size of their yard through encroachment, this does not often seem to be the case with the rural yards in this survey. In fact, a good proportion of them have been enlarged to cope with increased interments, and where an earlier boundary was obvious, such as a remnant hedge-bank, the earlier boundary was used for the calculations. Despite possible changes to the yard, this analysis is still worthwhile as it shows there is little difference in alignment irrespective of boundary proximity, as the table below demonstrates, with less than a 1° variation between the mean alignments of churches built in the smallest and the largest yards and the overall mean, and little difference in the proportions aligned north of east. Table 4.4 – Overall results by ratio of church floorspace to churchyard area Yard/church ratio <7 7 – 10.99 11-14.99 15+ Chapter 4 No range Mean 95% conf. 259 657 516 494 1,926 45-119 48-120 47-126 38-128 87.0 86.3 85.6 85.6 86.1 ±1.4 ±0.9 ±1.0 ±1.0 ±0.5 118 %N of East 58 65 63 65 63.8 Parallel with principal churchyard boundary Urban churches exhibit many examples of churches lying parallel with aspects of local topography, particularly street patterns. The alignment of churches with the grid patterns of towns with Roman origins is cited in studies noted earlier, covering much of the country, particularly churches in York, Winchester, Exeter and Chichester. Similarly, in planned and planted medieval towns, the churches tend to follow the street alignment (Butler 1976), especially those with grid-like street patterns, such as New Buckenham (Norfolk), Winchelsea (East Sussex) and Ludlow (Shropshire). Rural churchyards are usually larger and less formal in shape than urban yards, so assessing whether the church is parallel or not with its boundary becomes more difficult. For the sake of consistency, the side containing the main entrance to the churchyard has been used as the ‘principal’ boundary for this analysis. The fact that it contains the main entrance generally means that this part of the boundary is adjacent to a road. Churches parallel to their principal boundary have a slightly higher (more southerly) mean alignment, of 1.4º above the overall mean, but with a similar proportion of churches aligned to the north of east. This seems to indicate that, in rural churches, the fact that a church may be parallel to its principal boundary does not really affect its alignment. It is not easy to understand why a church and its churchyard boundary should be parallel, other than on a restricted site where space was at a premium, when in the cases where two churches built in the same yard, were, with one exception, never built parallel (this subject is discussed in more detail in Chapter Five). The argument that parallel is, in some sense, architecturally “good” does not seem reasonable when two of the most important buildings in a settlement were not built parallel to each other, apparently on purpose. Table 4.5 – Overall results by parallelism to the principal churchyard boundary Parallel to principal boundary Yes No Chapter 4 No Range Mean 95% conf. 346 1,580 1,926 45-126 38-128 87.5 85.7 86.1 ±1.3 ±0.6 ±0.5 119 % N of East 61 64 63.8 Proximity to the churchyard boundary Site restrictions do appear to have influenced the alignment of urban churches, but they do not appear to have affected the rural churches studied here. Although churchyards boundaries will probably have changed since the church was built, either because of an increase in size to accommodate the growth of interments, thereby hiding any possible earlier influence; or through a reduction in size by subsequent encroachment; this analysis is still worthwhile as it shows there is little difference in alignment, irrespective of boundary proximity. The range of alignments, between the most northerly and southerly in each of the groups in table 4.6 below, is almost exactly the same – a 74° range (45-119°) for those churches between two and four metres from their closest boundary, and 72º, 73° and 88° for the other large groups. Even the smallest group, where the boundary is closest to the church, has a range of 51°. The fact that this last group is smaller means that the mean alignment, although 1.8° higher than the overall mean, is less likely to be a significant difference. The figures clearly indicate little difference in mean alignment between those yards that are restricted and those which are not restricted, a strong indication that boundary proximity has never had a significant effect on church alignment. Table 4.6 – Overall results by proximity of closest churchyard boundary Closest Boundary <2 metres 2-4 metres 5-9 metres 10-14 metres 15+ metres Chapter 4 No Range Mean 93 349 677 546 261 1,926 62-113 45-119 47-120 38-126 56-128 87.9 85.6 85.5 86.6 86.3 86.1 120 95% Conf. ±2.3 ±1.3 ±0.9 ±0.9 ±1.3 ±0.5 % N of East 57 63 66 64 61 63.8 Size of the church There is no apparent reason why the size of a church building as it exists today after several hundred years of alterations and additions might have affected its original alignment. Apart from churches larger than 300 square metres, this is borne out by the fact that the remaining size categories have mean alignments within a degree or so of the overall mean alignment and have between 64 and 66% aligned to the north of east, compared with the overall mean of 63.8%. Larger churches are slightly under represented in the west of the country, particularly Cornwall, Pembrokeshire and Shropshire, where mean alignments are lower, again pointing to the ‘effect’ of longitude. Table 4.7 – Overall results by size of church Size of church <150 sq m 150-189 sq m 190-229 sq m 230–299 sq m 300 + sq m Chapter 4 No Range Mean 95% conf. 346 348 329 391 512 1,926 48-128 47-114 47-120 38-126 45-121 84.7 85.8 86.1 85.8 87.3 86.1 ±1.3 ±1.2 ±1.3 ±1.1 ±1.0 ±0.5 121 % N of East 66 64 64 65 61 63.8 Location in relation to settlements This is a particularly difficult category to interpret since the focus of so many settlements has shifted since the church was first built, particularly in East Anglia. However, since there is little variation between either the mean alignments of churches, or proportions of churches aligned to the north or south of east, for all the groups, apart from the small group of nineteen churches in scattered villages or which have no village at all, it is apparent that the location of the church had little effect on its alignment. It also means that any later changes in the relationship between church and village will not affect the results. Table 4.8 – Overall results by location of church in relation to its village Location In village Village edge Isolated Remote Scattered/No village No Range Mean 95% conf. 957 501 241 179 19 1,,926 38-121 54-126 47-119 48-128 74-107 86.0 85.4 86.8 86.7 90.2 86.1 ±0.7 ±1.0 ±1.3 ±1.7 ±4.7 ±0.5 % N of East 64 64 63 61 58 63.8 Height above sea level As mentioned earlier in this chapter, the higher (more southerly) mean alignments for churches built on sites below 30 metres above sea level (AOD) is a reflection of the predominance of this type of site in the lowland east of the country, where alignments are generally numerically higher. Conversely, the lower mean alignments, and higher proportions of churches aligned to the north of east, of churches built on sites above 70 metres AOD is a reflection of the predominance of these sites in the hillier west of the country. Table 4.9 – Overall results by height above sea level Height A.O.D. 0-10 metres 11-30 metres 31-50 metres 51-70 metres 71-100 metres 101+ metres Chapter 4 No 405 426 412 238 206 239 1,926 Range 45-118 54-128 38-120 54-126 47-110 55-107 122 Mean 87.0 88.0 86.9 85.3 83.1 82.8 86.1 95% conf. ±1.1 ±1.1 ±1.2 ±1.5 ±1.6 ±1.3 ±0.5 % N of E 62 58 59 69 68 75 63.8 Longitude and latitude As was shown in table 4.1 earlier, mean church alignments vary across the country by almost 12º when alignments in the east are compared with those in the west; a greater variation in alignment than within all of the other factors. The lower graph in Figure 4.3 below clearly shows that the mean alignments rise numerically (become more southerly) as one moves from west to east across the Country. The mean alignment for each longitude group is shown by the central bar, and the ranges at 95% and 99% confidence are represented by the edges of the shaded boxes. It can be clearly seen that at 95% confidence the lowest alignment in the east of the country is more than 6º higher than the highest alignment in the west. Even at 99% confidence, the difference between the lowest alignment of the range in the east is 5.4º higher than the highest alignment in the west (shown in table 4.10). By comparison, the differences in mean alignment by latitude are shown in the graph on the right of the Figure 4.3. Although there are variations between different parts of the country, unlike longitude, they do not form a linear pattern. The results of the Danish survey by Abrahamsen (1992), summarised earlier in Chapter One, serve to emphasise the east-west pattern in the change of mean church alignments by longitude, as the churches there have a mean alignment of 94°, some 2 to 4° more southerly than in the East of England. Although a relatively small number of churches was surveyed (204), and there was a substantial difference between the two areas surveyed (5.5°, even though they were close together), these results are higher (more southerly) than any of the results in England and they continue the pattern of increases in mean alignment by longitude eastwards so clearly displayed in this survey. The reason for this is not at all obvious. Chapter Six will consider factors which could vary with longitude and latitude, such as the spatial variations in the position of Magnetic North and west-east climatic variation, to see if longitude is masking a pattern that is caused by other factors. Chapter 4 123 Figure 4.3 – Rural church alignment by longitude and latitude Table 4.10 – Overall results by longitude Longitude (decimal) 1.70° - 1.00 º E 0.99ºE – 0.00 0.01ºW – 0.99º W 1.00° – 1.99º W 2.00° – 2.99º W 3.00º W + Overall Chapter 4 No Range Mean 95% 537 365 402 204 219 199 1,926 56-120 54-128 38-121 47-126 54-126 48-116 90.1 86.8 85.1 84.0 82.8 81.4 86.1 ±0.9 ±1.2 ±1.1 ±1.6 ±1.5 ±1.6 ±0.5 124 Range at 95% conf. 89.2 - 91.0 85.6 – 88.0 84.0 – 86.2 82.4 – 85.6 81.3 – 84.3 79.8 – 83.0 85.6 – 86.6 99% ±1.2 ±1.6 ±1.4 ±2.1 ±2.0 ±2.1 ±0.7 Range at 99% conf. 88.9-91.3 85.2-88.4 83.7-86.5 81.9-86.1 80.8-84.8 79.3-83.5 85.4-86.8 % N of East 50 60 72 67 72 78 63.8 Table 4.11 – Overall results by latitude Latitude (decimal) 53.00º N + 52.50°-52.99º N 52-00°-52.49º N 51.00°-51.99º N < 51.00º N Overall No Range Mean 95% 201 683 427 433 182 1,926 51-113 45-128 38-120 47-120 50-118 83.5 87.4 87.1 84.9 83.8 86.1 ±1.5 ±0.8 ±1.0 ±1.2 ±1.8 ±0.5 Range at 95% conf 82.0 – 85.0 86.6 – 88.2 86.1 – 88.1 83.7 – 86.1 82.0 – 85.6 85.6 – 86.6 % N of East 75 60 62 65 66 63.8 To remove the possibility that these differences have been brought about by a combination of longitude and latitude, the results for the churches in a ¼º wide strip across the country, including parts of Norfolk, Cambridgeshire, Lincolnshire and Shropshire, have been analysed and the results shown in table 4.12. Each longitude group in the strip shows similar results to the whole sample, with the mean alignment for each longitude group in the strip close to the value for the same group in the survey as a whole, with very similar proportions of churches aligned to the north of east, shown in the last two columns. Table 4.12 – Results by longitude for ¼º wide slice across the country, between 52° 30′ and 52° 45′ N Longitude (decimal) 1.70 - 1.00 º E 0.99ºE – 0.99ºW 1.00 – 1.99º W 2.00 – 2.99º W 3.00º W + Overall Chapter 4 No Range Mean 95% ±1.4 ±1.7 Range at 95% conf. 88.1 – 90.9 84.1 – 87.5 % N of East 51 65 197 166 0 59 0 424 56-111 45-121 89.6 85.8 55-126 82.5 ±3.5 79.0 – 86.0 73 45-126 87.1 ±1.1 86.0 – 88.2 61 125 Whole survey % N of East 50 66 67 72 78 64 SURVEY RESULTS: CONCLUSIONS This large-scale survey provides a statistically sound basis for the consideration of the general alignment of churches and for the analysis of the discovery that there is a significant difference in the mean alignment of churches across the country which has been identified here. None of the physical aspects of this survey, the churches themselves, their sites or their environments, appear to have been linked with their overall alignment, with one exception – their longitude. Possible reasons for this variation in alignment between the east and west of the country will be investigated in Chapter Six. The overall results, shown graphically in Figure 4.3 on page 124, seem to confirm that there was an intention on the part of church builders to align their churches roughly eastwards. Accuracy was not apparently paramount; an approximate direction appears to have been sufficient as demonstrated by the considerable variation between individual churches. The similarity of this pattern with those of prehistoric tombs and inhumations, discussed earlier, where the basic alignment direction was also eastwards, but with individual variations within the arc of sunrise, is striking, and begs the question of whether the alignment of churches is merely following an age-old tradition. However, if that were the case, it would not explain the significant variation in alignments between churches in the east of the country and those in the west. A simple desire to face east does not explain the fact that twice as many churches are aligned to the north of east in the west of the country than in the east, resulting in a difference of 12° between the mean alignment of the churches in Cornwall and of those in Kent. This difference between Cornwall and Kent is reinforced by the results for the other counties surveyed, which fit neatly in succession across the country, in between these extremes, confirming the east-west nature of the variation. Abrahamsen’s results appear to extend this pattern as far as Denmark. The possible reasons for this difference are explored in later chapters, where these results are expanded. Chapter 4 126 CHAPTER FIVE SURVEY RESULTS APPLIED TO EARLIER THEORIES I have measured quite a number of churches in different parts of England, and they suggest that what I have found in Oxfordshire [Patronal-saint sunrise alignment] is broadly typical of England as a whole. (Rev. H. Benson 1956, 212) INTRODUCTION This chapter applies the results of this survey to each of the five theories advanced in the earlier studies of church alignment outlined in Chapter One, in particular the ideas that churches were aligned with sunrise on the feastday of their patronal saint and that there was religious symbolism in the frequently found misalignment of nave and chancel at the same church. Additionally it applies them to the less widely published theories that suggest that the observed variation in the alignment of churches is due to the fact that they were aligned with sunrise on the day on which they were set out; that they were aligned towards sunrise at Easter, or that they were aligned towards Jerusalem. Since the variation in church alignment between the east and west of the country has been revealed for the first time in this survey, none of the earlier studies addressed this issue, as none was large enough, or suitably structured, to allow the pattern to be revealed, and each was solely focussed on the observed differences in alignment between, or within, individual churches. The considerably larger sample surveyed here allows a more statistically robust assessment to be made of each of the theories that have been proposed in the past. Chapter 5 127 1) ALIGNMENT WITH SUNRISE ON THE PATRON-SAINT’S FEASTDAY William Wordsworth’s reference to the alignment of Rydal chapel in a poem in 1823 was discussed in Chapter One, along with the survey details that show that the church apparently aligns with sunrise on one of the feastdays of St Mary. The poem itself, rather than any measurements on the ground, has been used by other researchers to validate the idea that all churches were set out this way. Even though the church does face a Marian festival sunrise, it was shown that the alignment might equally have been determined by practical concerns about the steep slope of the churchyard and the liturgical and building problems that the slope would introduce. The three main factors affecting the position of sunrise on a specific date; horizon elevation, calendar change and church rededication, were also discussed in Chapter One. There it was shown that the elevation of the horizon above the level of the church will delay sunrise and make it appear further south along the horizon, and that calendar change since medieval times has further complicated the issue due to the fact that, in some cases, accounting for the adjustment depends not only upon the year in which the church was set out, but also on the specific season of that year. If the time of the year when the initial alignment of individual churches was determined was spread throughout the year, then the differences noted above would tend to cancel each other out. In order to allow for the possibility that they might not cancel each other out, the general analyses presented in this chapter incorporate a range of 15° either side of the level horizon sunrise position. A separate analysis of the alignment of the churches in seven of the fifteen counties covered in this survey, where actual horizon elevation has been measured, is presented afterwards. The third issue affecting patron-saint sunrises involved church rededications. This was also discussed earlier in Chapter One, where it was concluded that although a variable proportion of churches in different parts of the country have been rededicated since they were first built, it is still important to examine whether churches face specific saint’s sunrises to place the earlier studies in context. The churches in the survey will firstly be analysed by their current dedication and broad conclusions drawn, then a summary analysis of dedications by their saints’ “season” will be presented and the results compared. Chapter 5 128 The 1,926 churches examined in this survey are dedicated to over 150 different saints. Most of the analysis that follows concentrates on the most common dedications and those with a single main feastday (based on the calendars of saint’s days in Cheney 2000 and Farmer 1991). Churches dedicated to St Mary have been excluded in the majority of instances because, without knowing which of the six major feast days was originally celebrated at specific churches, the introduction of so many additional possible sunrise points for a large number of churches complicates the issue. General analysis of saint’s day sunrise over a level horizon The curves in Figure 5.1 below show the position of sunrise throughout the year over a level horizon, for the dates shown on the horizontal axis, for the extremes of latitude in England – Cornwall, the red line (50°N) and Cumbria, the green line (55°N). They are shown in degrees from true north (the vertical axis), with east at 90° (the bold horizontal line in the centre), showing sunrises at the spring and autumn equinoxes where the curves cross the ‘east’ line. Superimposed are columns (in red) indicating the ranges of alignments in degrees for all the churches in this survey dedicated to a major saint that has a single main feastday, and are shown for each individual dedication on the date of the saint’s feastday (the horizontal axis). Chapter 5 129 Figure 5.1 – Sunrise azimuth and church alignment Chapter 5 130 It is immediately apparent from the Figure above that few churches are aligned towards sunrise on their patronal-saints’ day, shown by the fact that more than half of the columns fail to meet the sunrise curve at all, which shows that none of the churches dedicated to those saints face their sunrise. These include almost all churches dedicated to saints with festival days around midsummer (such as St. Botolph, St John the Baptist and St Margaret), or during the later autumn and winter (such as St. Andrew, St. Martin, St. Leonard, St. Nicholas and St John the Evangelist). The only dedications that have any churches facing their sunrise are those where the column crosses the sunrise line, mostly those dedicated to saints with feastdays closer to the equinox, such as St Bartholomew, St James, St Laurence and All Saints, although the sunrise line intersects with the extreme ends of most of these columns, in other words only close to the extreme end of the alignment range, thereby excluding the possibility that the vast majority of these churches, with alignments clustered around the middle (mean value) of the alignment range, face their sunrise. Most of the saint’s dedications have columns of roughly similar length and position, centred approximately on east, showing that almost all dedications have fairly similar alignment ranges, despite the fact that festival day sunrises vary between St John the Baptist and SS Peter and Paul, around 50°, and St Andrew and St Nicholas, at around 124°. Even the results for St Michael, whose feastday is September 29th, when sunrise is close to east, at around 95º, are inconclusive, as the alignments of the 110 churches dedicated to St Michael in this survey vary between 45° and 126°. The same information is shown in the table below and confirms how similar the mean alignments, and ranges of alignments, are across churches with different dedications. Only the small group of churches dedicated to St John varies by more than 2º from the overall mean, but with a small sample of twenty-nine churches. Chapter 5 131 Table 5.1 - Alignment of churches by dedication All Saints Holy Trinity SS Peter&Paul St Andrew St Bartholomew St Botolph St George St James St John Baptist St John St Laurence St Leonard St Margaret St Martin St Mary St Mary Magdalene St Michael St Nicholas St Peter Other Saints No saints day No dedication TOTAL total 256 27 58 147 23 17 22 38 52 29 41 20 78 25 399 16 110 70 148 297 27 26 1,926 Range of alignments 38-128 59-106 47-115 58-120 56-118 73-118 73-110 58-112 55-108 54-118 60-103 57-109 57-111 65-107 56-116 68-110 45-126 57-116 55-116 50-121 48-111 54-103 38-128 MEAN Alignment 86.5 84.9 87.3 87.4 87.9 84.5 87.0 86.9 84.5 82.3 84.1 84.4 87.8 84.1 86.9 87.8 85.9 86.4 86.2 84.3 84.3 84.5 86.1 95% conf. ±1.1 ±3.3 ±2.3 ±0.9 ±4.1 ±4.9 ±2.3 ±3.7 ±2.4 ±5.0 ±2.9 ±4.4 ±1.7 ±2.7 ±0.4 ±4.9 ±0.9 ±1.8 ±0.8 ±0.9 ±4.8 ±5.3 ±0.5 % North of East 64 65 55 56 59 80 65 52 63 72 71 74 49 74 63 63 61 63 64 66 67 54 63.8 Comparison of church alignment with saint’s day sunrise over a level horizon, and East The consistency in alignment eastwards, irrespective of dedication, rather than towards different sunrises, demonstrated in Figure 5.1, is emphasised by the figures shown in Table 5.2 below. Four of every five churches in the survey (80%) are aligned within ±15° of east, varying between 65%, for churches dedicated to St. Botolph, and 86% for those dedicated to All Saints, whereas only 18% of all churches are aligned within ±15° of their saint’s day sunrise position. In addition, half of all the churches in the survey (50%) are aligned more than 30° away from their saint’s day sunrise, but only 2% of churches are aligned more than 30° away from east. There are only two saints, St Bartholomew and St Michael – whose feastdays are close to the equinox – who have more than 50% of the churches dedicated to them aligned within ±15° of their feastday sunrise position, and even then, a greater Chapter 5 132 proportion of these churches (75%) is aligned within ±15º of east, suggesting that even with a feastday sunrise close to east, east itself was a greater focus for the church builder. Every dedication analysed has a greater proportion of its churches facing east than facing its sunrise. The fact that no churches dedicated to St Botolph are aligned within ±15° of sunrise, and only one church dedicated to saints Peter and Paul is within 15°, out of a total of 75 churches, whereas 59 of them (79%) face within ±15° of east, confirms the consistency with which churches of all dedications face generally eastwards rather than generally towards their patronal-saint’s sunrise point. If churches were intended to be aligned with their feastday sunrise, then it would be expected that any errors in setting out would mean that their alignments would vary either side of the saint’s sunrise position, rather than consistently varying around a point close to east. Table 5.2 - Church alignment compared with Saints day sunrise over a level horizon and Due East, by dedication degrees from Saints day sunrise degrees from Due East All Saints SS Peter&Paul St Andrew St Bartholomew St Botolph St George St James St John Baptist *20 St Laurence St Leonard St Margaret St Martin St Michael St Nicholas St Peter Other (with saints day) TOTAL Other (no saints day) St Mary No dedication total 256 58 147 23 17 22 38 51 41 20 79 25 110 70 148 369 1,474 27 399 26 ±15° No. % 30 12 1 2 7 5 13 57 0 8 36 5 13 5 10 14 34 2 10 9 11 3 12 74 67 2 3 9 6 95 26 16º30° 117 14 30 8 7 11 15 13 21 9 25 4 29 9 47 99 31+° No. % 109 43 43 74 110 75 2 8 10 59 3 14 18 47 33 65 6 15 9 45 45 57 18 72 7 6 59 84 92 62 175 47 277 458 739 18 50 ±15° No. % 220 86 48 83 120 82 19 83 11 65 17 77 31 82 37 73 34 83 15 75 62 78 17 68 81 74 57 81 120 81 289 78 1,178 22 333 19 80 82 84 73 16º30º 32 9 26 3 6 5 6 13 7 4 16 8 27 11 27 67 31+° No. % 4 2 1 2 1 1 1 4 1 3 1 2 1 5 1 1 2 2 2 3 1 1 13 4 267 4 64 6 29 1 2 1 The detailed results here seem to argue conclusively against the idea of patronal-saints’ day alignment. Some of the groups are small, but, if churches did 20 *One church was dedicated to St John the Baptist’s decollation (beheading), with a festival at the end of August, so was included in the “other” group Chapter 5 133 2 4 1 4 indeed face their sunrise, then all the churches of each dedication would tend to align closer to a single direction some distance away from east. It can be clearly seen, however, that churches of each dedication do not align in different directions, as is shown by the fact that around 80% of churches in each dedication are less than 15° from east. Added to this is the fact that churches in the “other saints” group, which consists of 369 churches dedicated to over 130 different saints, displays a similar range of alignments and has a similar mean direction to that of each of the individual saints, and has an almost identical proportion of churches facing within 15° of east (78%). Analysis by the season of the patron saint Since most churches are aligned close to east and most saint’s day sunrises are closer to the solsticial extremes, saint’s day sunrise alignment can be examined from the opposite perspective. Are the churches that are aligned furthest from east aligned towards specific saint’s days and could this display the remnants of a pattern that had existed in the past? For instance, it might be expected that churches dedicated to summer saints, such as St Botolph (June 17th), St John the Baptist (June 24th), Saints Peter and Paul (June 29th), St Margaret (July 20th), would be aligned well to the north of east, and conversely, churches with dedications closer to midwinter, St Martin (November 11th), St Edmund (November 20th), St Andrew (November 30th), St Nicholas (December 6th) and St John the Evangelist (December 27th), would be aligned well to the south of east. This is not the case, as is shown in table 5.3 below. Confirming the conclusions from the earlier detailed results, the churches dedicated to summer saints have the same relationship between the proportions of churches aligned towards northerly (summer) sunrises (16%) and southerly sunrises (13%) as do churches dedicated to “winter” saints (15 - 12%). Not only are the patterns of alignment between the two groups the same, but there are slightly more churches with “winter” saint dedications aligned towards summer sunrise (15%) than towards winter ones (12%). The mean alignment of each of the seasonal groups is within 0.4º of the overall mean for the whole survey and each of the seasonal groups has an almost identical alignment profile to that of the whole sample (shown in the fifth row), further confirming the complete lack of a seasonal pattern. Chapter 5 134 Table 5.3 – Alignment of churches by Season of saint’s dedication Church Alignment SUMMER SAINTS SS Botolph, Bartholomew, John the Baptist, Peter & Paul, Margaret EQUINOCTIAL SAINTS SS Lawrence, Michael, George, All Saints WINTER SAINTS SS Martin, Leonard, Edmund, Andrew, Nicholas, John the Evangelist OTHER Mary, No Saints day All churches in survey <75° No % 76-99° No % 100°+ No % 82 16 360 71 68 13 510 (100%) 86.5 92 16 428 72 72 12 592 (100%) 85.5 55 15 274 73 44 12 373 (100%) 86.4 55 12 341 76 55 12 86.5 284 15 1,403 73 239 12 451 (100%) 1926 (100%) Total Mean Align 86.1 These results show that it is certain that churches are not now aligned with their patronal-saints’ sunrise. Whether a large number of these churches has been rededicated so that they are still aligned towards their original patronal-saint’s sunrise is not known, but if that were the case, then all the churches that were rededicated would have to have been originally dedicated to saints whose feastday was close to the equinox (sunrise due east [90º]), such as St Michael, St Giles or St Bartholomew, since the majority of churches surveyed here are aligned within 10º of east. It would also mean that churches dedicated to some of the most popular saints whose feastdays are close to the solstices – for example St Andrew, St Nicholas, St John the Evangelist, St John the Baptist and St Peter – could never have been aligned towards their sunrise. It therefore seems almost certain, on this evidence, that the concept of churches facing their patronal-saint’s sunrise has never been true. Comparison using measured horizon elevation Two of the writers mentioned in Chapter One (Benson in 1956 and Muirden in 2005) held that taking the elevation of the horizon into account was crucial to the study of patronal-saint sunrise alignment. It is true that an elevated horizon does delay sunrise making the actual point of sunrise appear further to the south, as shown in Figure 1.1 Chapter 5 135 earlier. To take this into account, the horizons of 1,105 churches in seven counties, Bedfordshire, Cumbria, Hampshire, Lincolnshire, Norfolk, Oxfordshire and Pembrokeshire were measured, and calculations of their actual sunrise position made (See Appendix 6 for the formulae), and the results used to calculate the overall effect. Similar analyses to those shown in table 5.2 above (for sunrise over level horizons) are presented below using the actual sunrise point over the survey’s measured horizon elevation for these 1,105 churches. The results, in table 5.4, using the calculated sunrise positions, demonstrate the same alignment patterns as those in table 5.2 - within a few percent in every case. Overall, almost one in five churches (18%) is aligned within 15° of its saint’s day sunrise (c/w 18% over a level horizon), and half of all churches (52%) are aligned more than 30° away from the actual sunrise point (c/w 50% over a level horizon). In addition, more than four of every five churches (81%) are aligned within 15° of east (c/w 80%), with only eleven churches (1%) aligned more than 30° away from east (c/w 2% over a level horizon). Table 5.4 - Church alignment compared with actual saints day sunrise and Due East, by dedication, using horizon elevation (Churches in Bedfordshire, Cumbria, Hampshire, Lincolnshire, Norfolk, Oxfordshire & Pembrokeshire) total 154 37 98 13 12 12 23 29 47 60 43 85 235 degrees from saints day sunrise ±15° 16º31+° No. % 30° No. % 18 65 71 12 46 2 9 26 5 70 4 20 74 4 76 1 5 7 8 54 1 4 7 8 58 5 5 2 42 17 1 9 13 4 57 3 7 19 10 66 6 15 26 13 55 38 15 7 63 12 0 4 39 91 6 32 47 7 55 64 65 106 27 45 All Saints SS Peter&Paul St Andrew St Botolph St Edmund. St George St James St John Baptist. St Margaret St Michael St Nicholas St Peter Other (with saints day) TOTAL 848 149 Other- no saints day 8 - no dedication 10 - St Mary 239 OVERALL 1,105 Chapter 5 18 255 136 444 52 degrees from Due East ±15° 16º31+° No. % 30º No. % 133 86 17 4 3 29 7 1 78 3 82 16 0 84 8 5 0 62 12 100 0 0 10 2 0 83 21 2 0 8 22 7 0 76 38 9 0 81 42 17 1 70 2 36 7 0 84 70 15 0 82 187 81 43 5 2 690 7 7 203 907 81 88 70 85 82 147 1 3 35 186 11 0 0 1 12 1 1 1 To test more closely the overall impact that horizon elevation actually has on the results, table 5.5 below presents the results for the actual sunrise positions for the 1,105 churches, compared with their notional (level horizon) sunrise position. As would be expected, an elevated horizon which delays sunrise and therefore appears in a more southerly position on the horizon, means that summer sunrises (such as St John the Baptist and St Peter, close to northeast), when moved southwards, were brought closer to the alignment of the majority of churches, which are aligned close to east; whilst delaying a sunrise in winter (such as for All Saints, St Andrew and St Nicholas, already close to southeast) takes it further away from the alignment of most churches. This means that those sunrises that become closer to church alignments by taking the horizon elevation into account are offset by those that are shifted further away. The end result is that there is very little change in the overall pattern when horizons are taken into account. The number of churches aligned within 15º of their saint’s day sunrise decreases from 152 to 149, but stays at 18%, and the number aligned more than 30º from sunrise increases from 437 to 444, but stays at 52%. This clearly demonstrates that the assertion that the researcher has to calculate the effects of the delay of sunrise caused by elevated horizons, in order to test the patronal-saint sunrise theory, is not true. Whilst it is crucial at the level of an individual church, as soon as a larger sample is used the differences cancel themselves out. The 1,105 churches tested here show that there is no overall variation, and that improvements in alignment at some individual churches dedicated to summer saints are clearly offset by the opposite effect at others which are dedicated to saints with winter sunrises. Chapter 5 137 Table 5.5 – Comparison of notional and actual sunrise points by dedication (Churches in Cumbria, Pembrokeshire, Bedfordshire, Hampshire, Lincolnshire, Oxfordshire and Norfolk) saints in date order St George St Botolph St John Baptist. St Peter SS Peter&Paul St Margaret St James St Michael All Saints St Edmund. St Andrew St Nicholas St James Other (with saints day) TOTAL No dedication/day St Mary OVERALL total 12 13 29 85 37 47 23 60 154 12 98 43 23 235 848 18 239 1,105 degrees from LEVEL HORIZON saints day sunrise ±15° 16º31+° No. % 30° No. % 3 7 2 25 17 0 6 7 54 3 5 21 10 72 5 26 54 6 64 1 10 26 3 70 4 15 28 9 60 1 9 13 4 57 40 16 4 67 7 22 72 60 14 39 2 3 7 17 58 6 23 69 6 70 1 6 36 2 84 1 9 13 4 57 64 61 110 47 27 degrees from ACTUAL saints day sunrise ±15° 16º31+° No. % 30º No. % 5 5 2 42 17 1 5 7 8 54 3 7 19 10 66 6 32 47 7 55 2 9 26 5 70 6 26 13 15 55 1 9 13 4 57 38 63 15 7 12 18 12 65 71 46 1 4 7 8 58 4 20 74 4 76 0 4 39 91 1 9 13 4 57 64 27 65 106 45 152 149 18 259 437 52 18 255 444 52 Shared Churchyards Shared churchyards enable an additional view of church alignment to be taken, particularly of patronal-saint sunrise alignment, and they also enable reflection on the issue of ‘parallelism’. Parallelism was raised earlier when the alignment of churches was considered with reference to the principal boundary of the churchyard, where it was noted that there was little correlation between church alignment and boundary alignments, implying that a desire to build the church parallel with a close boundary was not an important consideration, and that visual neatness was not apparently an issue. Shared churchyards are a phenomenon which is concentrated in East Anglia with as many as 30 examples known, although most only have a single church now Chapter 5 138 (Batcock 1991, 10-11; Williamson 1993 158-161; Groves 1995, 108). There are only nine of those 30 shared churchyards remaining where sufficient parts of both churches still exist to enable accurate measurements of alignment to be taken (one of them South Walsham (Norfolk) is illustrated in the frontispiece). The churches in each of the pairs are dedicated to different saints but are aligned in similar but, with one exception, specifically not the same, direction. Each of the pairs includes one church dedicated to St Mary, but all the pairs are aligned within 11° of each other, whereas the sunrise positions for the festival days of the saints to which they are currently dedicated can be as much as 50° apart, depending on which festival of Mary was celebrated. If any of these churches has been rededicated and therefore still faces its original dedication sunrise, it would mean that each of these nine pairs of churches was dedicated to saints that had festival days just a few degrees apart as well as both being close to East. This would confine the possible dedicatees to festival dates only two or three weeks apart and close to the two equinox dates. Peter Warner considers that in the majority of these cases the church dedicated to St Mary was the second of the pair to be built, usually in the eleventh century, by groups of freemen (Warner 1986, 45), although it is difficult to see why this might have affected their alignments. The six major festivals celebrated for St Mary complicate the investigation of sunrise alignment as they cover most of the year. The sunrise point on Marian festivals in Norfolk varies between 54º and 123º, whereas the eight churches dedicated to St Mary are aligned at: 77º, 79°, 84°, 90°, 93°, 96°, 99° and 108º, with a mean of 90.8º. In each of the cases, except at Reepham illustrated in Figure 5.2 below, the churchyards are large enough to allow the church builders a free hand in selecting the alignment of both churches. The fact that each of these pairs of churches is aligned fairly closely together and that none is very different, implies that the alignment of the first of the two to be built has had some influence on the alignment of the later one. Since they do not appear to be aligned towards different saint’s day sunrises, it leaves the question – why were they not built on the same alignment? It is possible that alignment did not matter at all, but this is contradicted by the consistent general alignment eastwards, but it seems that the difference in alignment between the two churches did not matter enough to go to the (little) effort of aligning them in the same direction. Chapter 5 139 At Reepham, there was once a third church built in the same churchyard – All Saints’, Hackford – but which was largely destroyed by fire in 1549. It is shown on a plan drawn in 1750 (shown in Batcock 1991, 22), and appears to have been built on the same alignment as the other two. The situation in this yard is different from the others with more than one church as it is considerably smaller and the size of the yard seems to have forced the builders to align their churches in the same direction, as the two remaining churches actually overlap now, and the tower of St Michael’s, Whitwell had to be built on the south nave wall as it could not be located in its traditional place at the west end. Figure 5.2 – St Mary’s, Reepham and St Michael’s, Whitwell, built on the same alignment. The space to the left was originally occupied by All Saints’, Hackford Chapter 5 140 2) MISALIGNED CHANCELS AS RELIGIOUS SYMBOLISM The difference between the alignment of the nave and the chancel of churches has been discussed by many writers in the past. Although commonly known as weeping chancels, they have also been referred to as “crooked” chancels (Benson 1956; Muirden 2005), and “skewed” chancels (Cave 1950). It has been suggested that the misalignment represents religious symbolism, but applying the results of this survey will offer a more probable cause for the phenomenon. One in five of all churches is misaligned this way, and the theory holds that, particularly in cruciform churches, chancels align to the left of naves (more northerly), where the nave represents the body of Jesus on the Cross, with his head, the chancel, inclined to the left. This appears to be an extension of the proposal by William Durand, in The Symbolism of Churches, that the church represented the human body with the chancel as its head (Durand 1906, 17). Some of the writers found for the proposal (Benson 1956; Muirden 2005) and others against (Cave 1950; Hoare & Sweet 2000). In their introductory preface to Durand’s book, Neale and Webb (the kingpins of the CCS) also explained the misalignment as a reference to the Crucifixion, but commented that misalignment was “more frequently Figure 5.3 – St Andrew’s, Lamas, Norfolk. The chancel is misaligned 14° to the north ‘weeping’, and improving the alignment towards east (110° > 96°). One of the two most extreme northward realignment examples of 165 such churches in this survey Chapter 5 141 to the south” (Neale & Webb 1906, lxxxii), thereby undermining their explanation of its symbolism of the Crucifixion. Thompson, in the early years of the last century, referred to the “popular explanation” of the symbolism of the cross, commenting on the “general northward inclination of chancels of churches where the axes of the nave and chancel were different” (Thompson 1913, 131). He went on to say that “like most symbolical explanations, this is founded entirely on fancy” and noted that the phenomenon was not limited to churches with cross plans. According to Thompson, “others” had sought to explain the subject by suggesting the orientation of the chancel followed the direction of sunrise on the morning of the Patronal feast21. He also cited examples of realignments on sloping sites, where “masons kept as high on the slope as they could, so twisting the axis of the chancel” (1913, 132). Each of these issues will be considered using the results from this survey. Figure 5.4 – St Hermes’, St Ervan, Cornwall. The chancel is misaligned 7° to the south – not ‘weeping’, but improving the alignment towards east (81° > 88°). The most extreme southward realignment of the 150 such churches in this survey 21 This appears to be another example of a more widespread discussion of these issues than has been published - highlighted in Chapter One. Whilst Benson took up the idea in 1956 that crooked chancels reflected calendar drift it does not seem to have appeared in print before that. Chapter 5 142 Francis Bond, writing at a similar time, referred to the same general issues, but argued that misalignment, as ecclesiastical symbolism, “had escaped the notice of the ancient liturgists” (Bond 1914, 248). He also pointed out the images of Christ on the Cross were usually represented with his body and head in a straight line “until the twelfth century or later” (1914, 248), effectively excluding the possibility that this image might have been copied for churches set out before this time. The remainder of the points that he made concern large and urban churches, of the kind which are not included in this survey, firstly, where the choir was rebuilt in such a way as to encase the earlier choir, making it “impossible to see whether the axis of the new choir was being set out precisely in the line of the old nave” (1914, 249), and secondly, he mentioned the possibility that site restrictions in urban situations might be a potential cause of misalignment. Cave, in his pre-war survey, concluded that the small percentage of skewed chancels (16%), and the way that the numbers fell off as the skew became larger, indicated that they were accidental faults in setting out (Cave 1950, 51). More recently, Warwick Rodwell has referred to “setting out errors which equate to the thickness of one side wall, a very commonly found fault” (Rodwell 1981, 61) as an explanation for many of the strange contorted shapes that some church floorplans take. As far as the issue of alignment differences between nave and chancel is concerned, it will be shown here that the majority of these ‘errors’ can be better explained as having been a deliberate act and as having a particular purpose. The results from this survey Of the 1,926 churches in this survey, 377 (20%) have naves and chancels aligned differently by two degrees or more. Two degrees was taken as the cut-off point for the measurement of difference between nave and chancel as the compass used for the survey can only be guaranteed accurate to one degree. It is therefore possible (although unlikely) that a nave measured at 80° and a chancel at 82° could both really measure 81°. Despite the fact that with internal access to the church it can be seen that they are aligned differently, only external access was possible at some churches, where a difference was not necessarily so noticeable; so for consistency, all variations of less than two degrees have been excluded. Figure 5.5 shows an alignment difference of two degrees seen from inside a church, showing how obvious the misalignment is. Chapter 5 143 Figure 5.5 – An alignment difference of 2° between nave and chancel – at St Lawrence’s, Castle Rising, Norfolk, (left) and at Holy Trinity, Goodramgate, York (right) Sixty-two of the 377 misaligned churches in the survey have naves that are aligned close to east, or exactly east, so that they have naves that are aligned to the one side of east and chancels aligned the other side of east. Their alignments might be explained by the inability of the builders to locate east exactly, in which case they could have thought their realignment of the chancel was actually closer to east, rather than straddling it. Whatever the reason, they present a confused picture and are therefore only included in the total line of table 5.6 below and are excluded from the other tables and analysis in this section. Figure 5.6, below, illustrates the four possible situations involving churches where the nave and chancel have different alignments. It is immediately obvious that in only two of the four cases is the chancel inclined to the left - weeping (1 and 3), that in two of the four cases, the alignment of the chancel is closer to east (2 and 3) and in the last case (4), that the alignment is neither weeping nor closer to east. Chapter 5 144 Figure 5.6 – Possible variations in nave/chancel alignment Excluding the 62 churches mentioned earlier, the 315 remaining misaligned churches in this survey fit into one of the four cases shown in Figure 5.6 above; of these, 165 (52%) are in categories 1 and 3 (weeping), whilst 150 (48%) are in categories 2 and 4 (not weeping) which immediately destroys the idea of the folk tale, as in almost half the cases the chancel is misaligned the opposite way for the symbolism of the Crucifixion. The near equality of these two figures lends weight to the noted suggestion by both Cave and Rodwell that misalignments were accidental, as a roughly equal split would be expected of an accidental action. However, although on the surface it appears that the errors could be random, there are several factors that point to this having been a deliberate, or at least a non-accidental, act. Firstly, and most importantly, if the errors were random, or consisted merely of setting-out errors, then it would be expected that they would be the same through all Chapter 5 145 the ranges of nave alignment and therefore the numbers in each of the four cases in Figure 5.6 would be equal. They are not at all equal; case 1 has 80 churches, case 2 has 121 churches, case 3 has 85 churches and case 4 has 29 churches, strongly indicating that it was not random. There is a marked element of the alignment of the chancel being closer to east than that of the nave. In other words, chancels are aligned closer to east than their naves regardless of whether the alignment of the nave is to the north or south of east, with 206 chancels improving the alignment towards east (cases 2 and 3) compared with 109 churches with chancels further from east (cases 1 and 4). Secondly, it would have been easy to set out lines parallel to an existing nave for a new or extended chancel to enable them to be aligned in the same way, even with the most basic of equipment such as two pieces of string. For some reason, they were not set out in that way. It would also have been equally easy to correct an early error in the setting out, at the trench or footings stage, even as small an error as 2°, which would have been noticeable very early in the process. Thirdly, the number of churches involved adds weight to the idea that misalignment was not accidental; it seems very unlikely that easily avoided errors should affect every fifth church in the land. Lastly, the end result offends the modern eye and it seems reasonable to assume that it would have had a similar effect in medieval times. If there was no other intention behind the realignment, it would have been easier, and neater, to follow a single alignment either when originally setting out, or rebuilding part of, a church. Table 5.6 shows that of the 201 churches with their nave aligned to the north of east, the majority (121 - 60%) - are improving the alignment towards east, but not weeping (case 2), whereas the majority of churches with naves aligned to the south of east (85 of 114 – 75%) have chancels which are weeping but also aligned closer to east (case 3). The misalignment of the churches in the more extremely aligned groups is even more strongly biased towards east than that of the group of misaligned churches as a whole. In other words, the further the nave of a church is aligned from east, the more likely the chancel is to be realigned closer to east. Whilst 60% of all naves aligned to the north of east have chancels closer to east, this rises to 69% of churches where the nave is aligned north of 77°. Similarly for churches where the nave is aligned to the south of east, the overall proportion of 75% of chancels aligned closer to east rises to 78% of churches with naves aligned to the south of 103°. The full table is shown as table A7.18 in Appendix 7 on page 335. Chapter 5 146 Table 5.6 - Nave/chancel misalignment by alignment of nave All <77° All 78-87º All <88º (north) All >92º (south) All 93-102º All 103°+ Total exc 88-92º 88-92º See page 144 total 74 127 201 114 77 37 315 62 Improving No % 51 69 70 55 121 60 85 56 29 206 75 72 78 65 Not improving No % 23 31 57 45 80 40 29 21 8 109 26 28 22 35 Aligned left No % 23 31 57 45 40 . 80 85 56 29 165 49 75 . 72 78 52 . 79 Aligned right No % 51 69 70 58 60 . 121 29 21 8 150 13 25 . 28 22 48 . 21 The block-coloured cells in the table refer to Figure 5.7 below The overall proportions are shown pictorially in Figure 5.7 below. The left pie shows plainly that the majority of churches with naves north of east are misaligned right, therefore improving the alignment (40% left : 60% right), and the majority of those with naves aligned south of east have chancels that are misaligned to the left, both weeping and improving the alignment (75:25) (centre pie). Taken overall, as the right-hand pie shows, those misaligned left (weeping) are almost exactly balanced by those misaligned right, 52:48. If there was no particular reason for the misalignment, then it would be expected that the balance shown in these overall figures in the pie on the right, would be mirrored in the two groups that make it up. Instead there is a definite pattern of alignment of chancels closer to east than their nave. Naves north of east NOT improving Improving Naves south of east NOT improving Improving All Misaligned churches Aligned Left/Right Figure 5.7 – Proportions of all misaligned churches shown in table 5.6 – the chancel either improving, or NOT improving, the alignment towards east Chapter 5 147 Cave proposed that the number of misalignments “fell away” as the nave became aligned further from east, and he used this as a reason for his ‘accidental’ explanation for misalignment generally (Cave 1950, 53). However, the fact that the further a nave’s alignment is from east the more likely the chancel is to be realigned suggests strongly that Cave’s view is incorrect – this was not accidental, but implies both knowledge that the nave was not aligned eastwards and a desire to correct it. It is also reasonable to assume from these figures that there was not such a pressing need to deal with the problem if the church nave was originally aligned closer to east. Chancel rebuilding Many chancels were extended or completely rebuilt during the thirteenth and early fourteenth centuries for liturgical and space reasons after the Fourth Lateran Council of 1215 (Cook 1961, 42; Duffy 1992, 54). Were they realigned at this stage? Without the ability to deconstruct every church that has different nave and chancel alignments to establish construction sequences and with the problems of dating the earliest parts of church fabric from external examinations, published sources have to be relied upon. The assessments of church building period for this analysis have been taken from the relevant volume of Pevsner’s Buildings of England, and the results shown in Table 5.7. Whilst admittedly not perfect, the use of Pevsner’s assessment provides a level of consistency to the results, rather than using different, more local, architectural analysts in some of the areas, where their work is probably of differing standards. Date of Rebuilding Of the 315 churches in this analysis, Pevsner considered 101 chancels to be later than their nave; 83 naves and chancels to be of the same period; 124 churches where no comment is made on the period of either the nave, the chancel, or both; and seven churches where the naves are later than their chancels. Chapter 5 148 The chancels of all the 101 churches with later chancels are exactly equally divided between those that are misaligned left and those misaligned to the right (50:50), whereas almost three-quarters of them (72:29) are aligned closer to east. Table 5.7 – Alignments by relative dates of naves and chancels (excluding those aligned between 88 and 92º) All misaligned churches Total Imp. align Not Imp % improving alignment % aligned left Churches with later chancels Post medieval chancels 26 19 7 73 . 46 Medieval chancels 75 53 22 71 . 51 All later chancels 101 72 29 71 50 Nave/Chancel of Same period 83 49 34 49 59 Other churches 131 85 59 65 56 Total 315 206 109 65 52 The block-coloured cells in the table refer to Figure 5.8 below Within this group, the proportions of post-medieval chancels that are realigned closer to east is almost the same as churches with chancels rebuilt in the medieval period (73% compared with 71%), shown in Figure 5.8 below, although the sample is fairly small. However, the fact that both the medieval and post-medieval groups are split equally between those aligned left and right of the nave lends further weight to the arguments for the improvement of alignment towards east; if the samples were particularly biased due to their small size it would be reasonable to expect the bias to show in the left/right figures as well. It appears, from these results, that chancels that were rebuilt in the post medieval period were aligned slightly more accurately towards east than their medieval counterparts. Medieval rebuilt chancel NOT improving Post-medieval rebuild NOT improving Improving Improving Figure 5.8 – Proportions of medieval and post-medieval rebuilt chancels either improving or NOT improving the alignment to east, shown in table 5.7 Chapter 5 149 This leaves two further groups of churches that have different nave/chancel alignments. The first, shown in row 4 of table 5.7 above are those churches which, according to Pevsner, have naves and chancels built in the same period. Quite why a church should have a nave and chancel aligned differently if it was built in one campaign is difficult to envisage. It is just possible that these churches were actually originally built misaligned with the specific aim of aligning the chancel more closely to east, although this leaves the question of why the whole building was not aligned more ‘correctly’. A series of simple setting out errors would produce a random result, rather than one where 59% of the chancels are aligned closer to east than their naves. The more likely explanation is that the chancels were rebuilt shortly after the original build, but within the same architectural period, hiding the fact that there was a rebuild, and the opportunity to realign closer to east was used at this time. The ‘other churches’ group, shown in row five of table 5.7 – those where no assessment was made of the building periods of either one or both of the nave or the chancel, or have an earlier chancel – has 65% of its chancels aligned closer to east than the naves, 56% of them are aligned to the left and 44% aligned right of the nave, also indicates that Pevsner has probably not identified all the churches with rebuilt chancels. Post-medieval chancel rebuilds, particularly Georgian and Victorian ones, are much easier to identify, either from a stylistic point of view or through the use of different materials – particularly brick. Chapter 5 150 How was the realignment of chancels achieved? It has been shown that rebuilt and realigned chancels improve the alignment of the churches towards east, but this poses several further questions. Firstly, was the new alignment closer to True east or to magnetic east at the time of the realignment? Secondly, where was magnetic east when the church was realigned? Thirdly, as proposed by Benson and Muirden, was the new alignment closer to the position of sunrise on the patronal-saint’s day? Lastly, as suggested by Thompson, were the chancels misaligned or realigned because of problems with a sloping or restricted site. Realignment towards east using a compass The difference between the true direction and that shown on a magnetic compass, known as magnetic declination, was described in the Glossary on pages x-xii, as was the fact that the difference between the two directions was not first measured until 1576 CE. The possible effect that it may have had on overall church alignment is discussed in detail in Chapter Six on pages 179 – 184, alongside the details of when the compass may have first been used for churches. In this section the subject of interest is whether a compass may have been used in the realignment of chancels during rebuilding. The full analysis and tables are shown in Appendix 4 on pages 327-328, with the results summarised here. Overall, 47% of chancels rebuilt in the medieval period were realigned closer to Magnetic East (which was at an average of approximately 100° True during the medieval period (see Clark et al. 1988, 649)), whereas 71% were realigned closer to True East, strongly indicating that east was the focus and that a magnetic compass was not used. The post-medieval rebuilding of chancels, often a Georgian exercise, took place in a period when magnetic north was west of true north. In 1800, magnetic north was approximately 24° west of north (Clark et al. 1988, 649; Merrill et al. 1996, 46), so if a compass was used without adjustment for declination, it would result in alignments for east of 66°True. Forty-five per cent of the post-medieval chancel rebuilds were realigned closer to Magnetic East whilst 73% were realigned closer to True East, therefore, either a compass was not used for the realignment or appropriate adjustments were made to the readings to take declination into account. Chapter 5 151 Realignment towards patronal-saint’s sunrise Whether the whole church faces its saint’s day sunrise or not was considered earlier, along with the related issues and problems of horizon elevation, calendar drift and church rededication. Both Benson and Muirden concluded that, as the position of sunrise on the required saint’s day had shifted along the horizon due to calendar drift, the rebuilding of the chancel was used as a chance to realign the east end of the church to maintain the ‘correct’ alignment towards sunrise. The issue examined in this section is whether the results of this survey show that the chancels of misaligned churches are aligned closer to their patronal-saint’s sunrise than the nave. Only churches with a sufficient number dedicated to the same saint are included in this analysis. Table 5.8 shows the results for all churches with misaligned chancels, whether the relevant volume of Buildings of England noted them as rebuilt or not. Despite the variation in sunrise positions on the various saint’s days, between St Peter (50°) and St Andrew 116°), less than half of all chancels (41%) were realigned closer to their sunrise and 59% were realigned further away. More than half of the churches dedicated to St Michael, with a feastday close to the equinox (29th September), had their chancels aligned further from sunrise than their nave, demonstrating that sunrise on saint’s day played no apparent part in the ultimate alignment of the chancel. However, almost twothirds (61%) of the chancels of the churches dedicated to these saints were aligned closer to east than their nave. Table 5.8 – Alignment of chancels in all misaligned churches in relation to patronal-saint’s day sunrise Sunrise No. of position churches All Saints St Andrew St Margaret St Michael St Peter and SS Peter & Paul TOTAL Chapter 5 105° 116° 64° 92º 50º 52 22 16 23 44 157 Alignment to saints Day Alignment to EASTimprov further improv. further neither 17 35 28 20 4 11 11 13 7 2 8 8 9 6 1 9 14 14 8 1 20 24 26 17 1 65 (41%) 152 92 (59%) 90 (61%) 58 (39%) 9 Rebuilding the chancel gave the builder the opportunity to improve the alignment towards the intended target. Table 5.9 shows that his target was not the saint’s day sunrise, with an even smaller percentage (37%) of churches dedicated to All Saints and St Peter realigned towards their sunrise. The number of churches in the remainder of the dedications is too small to analyse. As with the analysis of all misaligned churches, a far greater proportion has been realigned towards east (63%). The numbers involved here are very small, but mirror closely the repeating pattern of realignment towards east, rather than towards sunrise. Table 5.9– Alignment of rebuilt chancels in relation to patronal-saint’s day sunrise and east All Saints St Peter TOTAL Chapter 5 Sunrise Position (approx) 105° 50º No. of churches 18 25 43 Align to saints Day Improv. further 7 11 9 16 27 16 (37%) (63%) 153 Alignment towards EAST Imp. further neither 9 7 2 16 8 1 25 15 (37%) (63%) Misalignment due to slope? It has been suggested that the misalignment of chancels may have been to compensate for, or take into account, the slope of the churchyard, either by building, or rebuilding, the chancel away from the axis of the slope (Thompson 1913, 31). Although particular problems are raised by building on sloping sites, the slope does not appear to have figured in the decision to rebuild the chancel in the churches here, as there is a similar proportion of churches with rebuilt chancels on slopes as there is in the whole survey sample. In order to test whether the slope of the site was a factor in either the misalignment of chancels, or their realignment, the proportion of misaligned churches and of all churches on sloping sites must be compared. If the slope had been a problem, then a greater proportion of misaligned churches would appear on sloping sites. Table 5.10 shows that 84 of 377 (22%) of all misaligned churches in this survey are built on a slope of more than two per cent, while the equivalent figure for all the churches in the survey is 19% - indicating that the slope itself was not a determining factor in the misalignment. Misaligned churches with rebuilt chancels do form a slightly higher proportion of all the churches with rebuilt chancels (30 of the 111 - 27%), so some of them might possibly occupy sites where the slope was a particular problem and had therefore been a factor in the rebuilding. Table 5.10 – Effect of slope on churches All Churches Churchyard slope: Flat or slope less than 2% Slope more than >2% Total No. 1,553 373 1,926 % 81 19 100 All misaligned churches No. % 293 78 84 22 377 100 All Rebuilt (later) chancels No. % 81 73 30 27 111 100 It is possible that the direction of the realignment of a rebuilt chancel could reflect attempts to deal with structural problems caused by sloping sites, as Thompson suggested earlier, where builders were rebuilding to avoid the slope (1913, 131). If the slope had caused problems in the building, any realignment to solve this would need to be closer to the direction of the slope to ease the stresses in the building. These 30 churches are equally divided between those that are realigned closer to the axis of the slope (or across it) and those which are aligned further from the axis of the slope, Chapter 5 154 indicating that the direction of the slope played no part at all in the need to rebuild the chancel, reconfirming that the slope had no effect on the misalignment of the chancel. Misalignment due to site restrictions? It has also been suggested that restrictions of the churchyard may have caused the church to have been built misaligned in an attempt to align the church towards east on a site with insufficient space to align the whole church the same way, a situation mentioned by Bond, particularly in relation to urban churches on cramped sites (1914, 249). Alternatively, it may reflect a situation where the ability to extend the chancel by the required length was prevented by site restrictions and that the realignment was necessary to fit the extended chancel into the site. This analysis is complicated by the fact that the churchyard boundaries will probably have changed over the centuries, although unlike churches in towns, rural churchyards in general are more likely to have expanded than to have contracted, to deal with the pressure of extra burials, and are more likely to have space around them to allow this. Table 5.11 – Effect of proximity of churchyard boundaries on churches Closest churchyard boundary: Less than 2 metres 2 – 4 metres 5 – 9 metres 10 metres or more Total All Churches No. 93 349 677 807 1,926 % 5 18 35 42 100 All misaligned churches As % No. % of All 17 5 18 84 22 24 131 35 19 145 39 18 377 100 20 All rebuilt chancels No. % 6 5 24 22 41 37 40 36 111 100 Those churches that are still close to one of their boundaries show no increased likelihood of being misaligned. As table 5.11 shows, similar proportions of churches are misaligned (around one-fifth (20%) – shown in column five of the table), whether they are in restricted churchyards, as indicated by the distance of the closest boundary, or located in larger yards. This confirms that site restriction has not played a part in the fact that these churches are misaligned. Neither does site restriction appear to have played a part during the rebuilding of chancels. The 111 churches with rebuilt chancels have the same profile of proximity to their boundaries as does the whole survey sample, with 5% in the closest category and around 40% in the largest. Chapter 5 155 Characteristics of Churches with misaligned chancels In order to determine whether there are any other specific differences between those churches with misaligned chancels and the whole sample, which might indicate a possible reason for the misalignment or highlight another avenue to research, the characteristics of the misaligned group as a whole is compared to the whole sample in tables 5.12 and 5.13 below. Their distribution between the counties surveyed is shown in Table 5.12 and demonstrates that the situation is a countrywide one, but is not equally distributed. Overall, one in every five churches is misaligned, varying between seven per cent in Cornwall, which is accounted for by the very high proportion of what John Betjeman called “Cornish Ends” (Betjeman 1968, Volume 1), where axial chapels are built flanking the chancel on both sides, thus preventing measurements being taken, and 36% of churches in East Sussex, which is much higher than in every other county and for which no explanation can be offered. Table 5.12 – Distribution of misaligned churches All churches Misaligned Bedfordshire 96 8 northern Cambridgeshire 123 24 western Cornwall 72 5 Cumbria 74 16 East Sussex 104 37 eastern Kent 92 22 southern Hampshire 76 12 southern Lincolnshire 103 22 Norfolk 549 93 northern Oxfordshire 130 28 Pembrokeshire 77 16 Somerset 91 17 Shropshire 104 25 northern Suffolk 125 23 East Riding Yorkshire 110 31 TOTAL 1,926 377 % 8 19 7 22 36 24 16 21 17 22 21 19 24 18 28 20 Apart from the difference in nave/chancel alignment, there appears to be no other differences between this group of churches and those with a single alignment constituting the remainder of the survey sample. When analysed against other factors, for example: size, floorplan, tower type and dedication, the proportions in each category are remarkably similar. They cover the full range of sizes, from the smallest to Chapter 5 156 the largest; they have similar floor plans in terms of the number of aisles and they have similar tower types. Table 5.13 – Comparison of misaligned churches with all churches in the survey % that misaligned Misaligned All churches in churches form of Category Churches Survey ‘all churches’ in the Number % Number % same category No Aisles 151 770 42 40 20 One Aisle 93 465 26 24 20 Two Aisles 133 691 32 36 19 Church <190 sq m 190-300 sq m >300 sq m 132 155 90 36 41 23 694 720 512 36 37 27 19 21 18 No tower Round tower Square buttressed 40 29 192 116 377 86.3° 10 8 50 31 100 227 140 1036 523 1,926 86.1° 12 7 54 27 100 18 21 19 22 20 Square unbuttressed MEAN ALIGNMENT The column on the right of table 5.13 shows that an almost identical proportion of each group of churches is misaligned. Overall, one in five churches is misaligned (20%) and this proportion is repeated in most of the table rows. For example: 20% of churches with no aisles; 19% of churches of less than 190 square metres in area; 21% of churches with a round tower, indicating no bias at all in the types of church involved. Finally, as a group, they are aligned almost exactly the same way as the overall sample – the mean alignment of the nave of misaligned churches is within 0.2º of the survey as a whole, at 86.3°. All of which appears to point to a random occurrence, which would be expected if random setting out errors were the cause of the misalignment. But that ignores the undeniable fact that for the majority of the misaligned churches, their chancels now face closer to east than their naves, implying that there was only one desire when realigning the chancel of the church – to face closer to east. Chapter 5 157 Misaligned chancels as symbolism: Conclusions Misalignment of chancels, or their realignment, cannot be explained by any of the topographical factors analysed here; patterns of churches which are misaligned or realigned are in most cases the same as those for churches as a whole. Patronal-saint sunrise also plays no part in the alteration of alignment. The only influence that can be shown to have occurred consistently throughout this analysis is east; the alignments of the majority of chancels are closer to east in each of the analyses, further confirming the overall intention in the alignment of church buildings. The fact that the symbolism of different nave and chancel alignments had evaded even William Durand must surely indicate that there was no contemporary symbolic intent in the minds of the builders. Durand found symbolism in every aspect of church doctrine, church buildings and church fittings. Symbolism towards such an important aspect of church teaching as the Crucifixion would have been the first, and most important, symbolic reference he would have identified. He lived until the end of the thirteenth century (Neale & Webb 1906, ix) during the last fifty years of which many of these chancels were being rebuilt and realigned. Had there been any intentional symbolic representation of the Crucifixion behind the different alignments of nave and chancel it must have come from the senior church hierarchy rather than from the individual builders themselves, so therefore it must have been discussed by the Church and would therefore have been known by Durand. Perhaps a simple realignment closer to east did not contain sufficient symbolism to mention. Chapter 5 158 3) ALIGNMENT WITH SUNRISE AT THE TIME OF THE SETTING OUT OF THE CHURCH The third theory to examine with the results of this survey is that stated in Chauncy’s quotation in the general introduction to this thesis – the assumption that churches faced sunrise on the day their foundations were laid. Chauncy assumed that churches that were aligned close to northeast were laid out near midsummer and those aligned close to southeast were laid out near to midwinter sunrise. The ritual setting out of the church foundations could have taken place at any time of year, but whilst this may explain the variation between the alignments of individual churches, it fails completely to explain the observed differences in mean church alignment across the country. Although church foundations could have been pegged out at any time of year, the previous sections have proved that this opportunity was not taken to align the church on its patronal-saint’s feast day. If churches were laid out throughout the year towards sunrise, a pattern of alignment that followed the movement of sunrise would be expected, where sunrise moves swiftly along horizon at the equinoxes, but slows to a standstill at the solstices. The result is that sunrise is only within 10º of east on 18% of days throughout the year (author’s calculations based on published sunrise data by the British Sundial Society (Davis 2004)), whereas it is more than 30º from east on 40% of days. The pattern of church alignment is the inverse of this, with 63% of churches aligned within 10º of east and only 2 % of churches aligned further than 30º from east, indicating again that there was purpose behind the alignment of churches as a whole, and that their general focus was generally eastwards. Table 5.14 Sunrise position throughout the year compared with church alignment Sunrise (days) Church alignment No % No % >30º from East 147 33 40 2 ±21-30º from East 86 160 23 8 ±11-20º from East 68 525 18 27 ±10º from East 64 1,208 18 63 Total 365 1,926 Chapter 5 159 There are certain seasonal issues surrounding the building of a church which need to be examined. The first action in the actual construction of a church would be the digging of the foundation trenches, which would tend to fix the alignment of the subsequent structure fairly closely. Without specialist labourers, the use of local labour would mean that the digging foundations for a rural church would have been likely to be more of a winter activity, when there was more free time from working on the land. Alignments fixed towards sunrise at the time of trench digging during the winter would be aligned well south of east. The great majority of foundation trenches cannot have been dug in the winter and aligned with sunrise, since almost two-thirds of all the churches surveyed here were aligned north of east, which would mean alignment towards sunrise during the spring, after the spring equinox, or in the late summer before the autumn equinox, both of which can be periods of intensive activity on the land. The second of these periods is discussed in more detail in Chapter Six in relation to harvest. Was the start of building above ground a more likely point to fix the alignment? Church builders had to take into account the likelihood of frost, especially where the local materials required the use of a considerable amount of mortar. Building contracts for churches in East Anglia often made specific references to start dates for building each year. At St Mary’s, Helmingham in Suffolk, for example, the contract for the tower specified that building (in flint) could only be undertaken between Whitsun (6 weeks after Easter - between May 5th and June 5th) and the 8th of September (Salzman 1992, 547-548), a season of no more than four months, and in some years, only a few days over three months. In order to maximise the amount of building in the first year, it would seem reasonable that building would start as soon as feasible, or as soon as the contract allowed, in that year. Sunrise at Whitsun in East Anglia is between 62° (5th May) and 54° (5th June). As only two of the 674 churches surveyed in Norfolk and Suffolk are aligned north of 62°, alignments with sunrise at the start of building are not indicated, at least in East Anglia. In areas where soil could be used instead of mortar to bed large non-calcareous stones, such as granite and slate in Cornwall, Pembrokeshire and Cumbria, or in areas in which the use of large limestone or sandstone ashlar blocks meant the use of considerably less mortar, late frosts would not be so important, which would mean that the start of building could be earlier in the year and allow the continuation of building later on in the year. As an example, the contract noted above Chapter 5 160 for the 60 foot tall flint-built Helmingham tower was for ten years, whereas the 100 foot tall ashlar-built tower of St George’s, Dunster (Somerset) was to be completed in three years (Salzman 1992, 548 & 514). Figure 5.9 – St Mary’s, Helmingham, Figure 5.10 – St George’s, Dunster, Suffolk, flint tower (10 year contract) Somerset, ashlar tower (3 year contract) If church building was started earlier in the year in Cornwall, Pembrokeshire and Cumbria, and the church aligned with sunrise at that time, this would result in alignments further south on average than in the remainder of the country. The results do not bear this out, as these counties exhibit numerically the lowest (most northerly) mean alignments of any of the areas surveyed. Detailed regional climate patterns in medieval times are unknown, but modern summary climate records obtained from the Meteorological Office22, especially of the number of frost days, and the timing of the last frost in the year, do not indicate any regional patterns, either north-south or east-west. In the three decades after 1960, Shropshire had the most frost days and Cornwall had the least, the latest frosts occurring in Kent and the earliest cessation of frost was shared by Cornwall, East Sussex and Suffolk. 22 http://met-office.gov.uk/climate/uk/averages/station.html (accessed 9th June 2001) Chapter 5 161 More recent, and more detailed, summary records published by the Meteorological Office (for 1971 to 2000)23 do show a slight east-west spatial variation in the number of frost-days, shown in Table 5.15 below. The counties closest to the sea seem to have the fewest frost-days, particularly those on the west coast, even if further north (Cornwall, Pembrokeshire and Cumbria) compared with counties on the east coast (Suffolk, Norfolk, Lincolnshire and East Yorkshire), but the counties in the centre of the country – Cambridgeshire, Shropshire, Bedfordshire and Oxfordshire – have the highest number of frost-days. Whilst the difference in the number of frost-days between the east and west coast might begin to explain the difference in alignment of churches there, the far greater number of frost-days in the centre of the country means that churches there would not be aligned between those in the east and west, as the later start due to the extra frost-days would mean an even more northerly alignment in the central area, rather than the reality of mean church alignment in the central area fitting numerically in between the alignments in the east and west. Table 5.15 Numbers of frost-days per year 1971-2000 County Cornwall Pembroke Cumbria Days of frost 0-100 10-110 20-140 20-120 exc. fells 25-110 25-130 Lincolnshire, 35-125 Sussex, Hampshire Kent Suffolk, Norfolk, East Yorkshire Somerset Cambridge Shropshire, Bedfordshire, Oxfordshire 40-120 70-125 102-144 Mean 50 60 70 67 77 80 80 97 123 The issue of general climatic differences between the east and west of the country, rather than just using frost-days as an indicator, is considered in more detail in the next chapter. 23 http://metoffice.gov.uk/climate/uk/averages/19712000/regions.gif (last accessed 16th Sept 2008) Chapter 5 162 Figure 5.11 – Examples of Meteorological Office Climate maps used for the frost days data shown in table 5.18 - for Cornwall and Shropshire (Although the Met. Office has used similar colours for these two counties, the values they represent are very different – the lowest values found in Shropshire (102 -108) are higher than the highest category found in Cornwall (80-100 – royal blue)) Sources: http://metoffice.gov.uk/climate/uk/averages/19712000/regions.gif last accessed 16th Sept 2008) Chapter 5 163 4) ALIGNMENT WITH SUNRISE AT EASTER The fourth theory that has been suggested may have caused the differences between individual church alignments is that churches were aligned with sunrise at Easter, and that the pattern of movement of Easter day is then reflected in church alignments built in different years. Easter is universally considered to be the most important festival of the church. Its date falls on the first full moon after the first Sunday after the northernhemisphere Spring Equinox, varying between March 22nd and April 25th. Sunrise on these days ranges between 90° and 67/69° in England, depending on latitude, with a mean of approximately 78°. This is between 2° and 14° more northerly than the mean direction of church alignment in any of the areas surveyed here, so it is not likely that the position of sunrise at Easter was the focus for the alignment of churches. The Roman and Western churches frequently celebrated Easter on different dates which could be reflected in a difference in church alignment between adherents of the two doctrines in the west and east of the country. The issue was discussed at the Whitby synod in 664 CE which fixed the calculation of the date of Easter using the Roman method and all but the Ionian church followed (Colgrave 1927, 9; MayrHarting 1972, 131-135). However, so few churches had been built by this time that any differences in the date of Easter cannot explain the observed alignment variations between east and west. 5) ALIGNMENT TOWARDS JERUSALEM The final theory to examine that has been suggested to explain the differing alignment of churches is that raised in the commonly heard tale24 that churches face east because they face Jerusalem, thereby following the Jewish tradition to face Jerusalem for prayer, although Davies alluded to facing east for prayer as a contrast to the Jewish tradition of facing Jerusalem (Davies 1972, 303). Unlike the other theories tested in this chapter, this one does not seek to explain why church alignment varies, merely suggesting that churches are aligned with Jerusalem. Apart from the more complex issue of the degree of medieval knowledge of the directional location of Jerusalem, this 24 Including the Council for British Archaeology website http://britarch.ac.uk/yac/leaderdocs/skillstraining/english_parish_church.pdf (accessed 13th Dec 2009) Chapter 5 164 theory can be simply dismissed on general grounds. There were many thirteenthcentury world maps which formed the basis of the medieval world awareness among the educated (Harvey 1996, 35). The Hereford Mappa Mundi is the largest and most detailed surviving example of these (Harvey 1996, 38) and, like the others, shows Jerusalem to be (geographically correctly) located to the southeast of Britain, rather than to the east. However, all the medieval maps show an island called “terrestrial paradise” – representing The Garden of Eden – at the eastern edge of the world, except the Psalter map, which is considerably smaller than the others (Harvey 1996, 34) and which shows paradise to be located within Asia, but still at the eastern edge of the world (Harvey 1996, 29). Therefore, in facing east, churches are unlikely to be facing Jerusalem, but appear to be aligned with paradise. It is also important to note that medieval map notation locates east at the top of the maps, rather than the modern location to the right, further emphasising the medieval importance of east. SUMMARY OF THE VALIDITY OF THESE EARLIER THEORIES Despite the certainty of the conclusions arrived at by many of the earlier students of church alignment, the data from this survey argue convincingly against them all. Many of these theories were first proposed over one hundred years ago and at least one of the writers on the subject was a vicar – Benson 1956 – who may have had a vested interest in attempting to prove a religious significance in his results. The large sample of this survey has enabled a more secure view to be taken of many of the statistical aspects of the results. Each theory has been shown to fail and in many cases a more pragmatic or logical explanation and conclusion is indicated. Alignment with Patronal saint sunrise Churches are not aligned with their patronal-saint’s feastday sunrises, nor with any other specific sunrises. This conclusion is confirmed by all the analyses here, ranging from the summary statistics of winter and summer saints, which indicate virtually identical alignment patterns for saints of all seasons compared with sunrise differences Chapter 5 165 of up to one-fifth of the horizon, to a specific analysis of the actual sunrise point of the 1,105 churches surveyed in seven counties, showing virtually no correlation at all between saint’s day sunrise and alignment. No amount of tinkering with the results, to take horizon elevation and calendar change into account, can alter the fact that exactly half of all churches are aligned further than 30º from their sunrise (table 5.2), whereas less than two percent are aligned more than 30º from east. Even the argument about the possible rededication of churches cannot help. Since the vast majority of churches are aligned close to east, rather than exhibiting a wider pattern of alignment towards multiple foci, spread throughout the year, it seems certain that they never did align with sunrise on the feastday of their patronal saint. Taken to its extreme, if the proposal that churches were aligned with the sunrise on their patronal-saint’s day was true, the overwhelming concentration of churches aligned close to east would have to mean that they were nearly all dedicated to saints with feast days close to either of the equinoxes (late March and late September) rather than the range of saints that have feast days widely spread throughout the year, with concentrations of feastdays closer to the extremes of sunrise, near the summer and winter solstices, where the majority of sunrises occur. Religious imagery of misaligned chancels This theory is also without foundation. Close to half of all the chancels that are aligned differently from their naves are aligned to the north of the nave (to the left when looking down the nave towards the chancel) and half are aligned to the south (right). The proposition that this misalignment represents religious imagery of the Crucifixion fails at that point, irrespective of the consideration that the representation of the Crucifixion was different in early times, both of which appear to have been ignored by the proposers of the theory. Further confirmation of the lack of intent for religious imagery is the fact that William Durand did not include this as a symbolic reference in his comprehensive treatise on church symbolism, despite finding symbolism in every other aspect of church buildings. The roughly equal division in misalignments left and right identified in the survey is what would be expected of a sample of this size if the distribution was a random one. However, within this apparently random result is hidden a more reasonable explanation. Firstly, where churches had their chancels rebuilt, threeChapter 5 166 quarters were realigned closer to east, probably as part of the rebuilding process, although it is possible that the new chancel was repeating an original misalignment. The improvement of alignment during rebuilding, even though it is usually only by a few degrees, is further emphasised by the fact that the later the rebuilding of the chancel, the more likely there is to be an improvement in alignment towards east. The eastward focus is also supported by the fact that the further the nave is aligned from east, the higher the proportion of chancels that are realigned more closely with east, even if the nave and chancel appear to be of the same architectural period. Overall, these results strongly suggest that misalignment was not an effort on the part of the builders to reflect religious imagery, neither was it an accident of setting out. Indeed, it appears to indicate a desire to achieve a more accurate eastward alignment, thus further confirming the importance of east, as well as indicating that it was appreciated that these alignments required correction. Finally, neither sunrise on the patronal-saint’s day, nor sloping sites, nor site restrictions appear to have been factors in misalignment or realignment. Analysis by sunrise on the feastday of the patronal saint shows that less than half the chancels were aligned closer to, and more than half were aligned further from, the relevant sunrise position. Similarly, half of the chancels appear to align closer to the axis of the slope of the churchyard and half further away. Lastly, churches in more restricted churchyards are no more likely to be misaligned than those in large yards. Alignment with sunrise at Easter The average date for the celebration of Easter between 900-1200 CE was April 9th (author’s calculation of Easter, based on the standard algorithm, set out in Appendix 8), when the sun rises at around 77° or 78º depending on the latitude. This is close to the mean alignment in the west of the country, but is several degrees different from the mean alignment in the east of the country. Although the position of sunrise at Easter varies depending upon the date of the festival (between 67° and 90º), this cannot be used as an explanation for the range of church alignments on the ground as the range is insufficient. Even the differences in dates between western and Roman Easter, which might have had an east-west impact on the figures, cannot be a factor, firstly because very few churches were built before the date of Easter was standardized, but mainly because western Easter was usually earlier than Roman Easter (Cheney 2000 47-54) Chapter 5 167 (for details - see table A8.1 in Appendix 8 on page 336). This would have resulted in more southerly sunrises on Easter Day in the west – which would mean that churches in western areas would have numerically higher mean alignments than those in the eastern areas – the complete opposite of the pattern actually seen on the ground. Alignment with sunrise when the church was built Similarly, delays in the start of springtime building due to climatic variations cannot explain the variations either between individual church alignments or the observed pattern of variations in church alignment across the country. Although the regional details of the medieval climate are unknown, modern climate frost records do not indicate any consistent pattern across the country, either north-south or east-west, as the centre of the country away from the sea has both more, and later, frost-days. If the mean direction of church alignment in each area was to reflect the most popular period of the year for the commencement of church building, it would require that building in Cornwall started later than everywhere else, and in particular, that twice as many churches had later building starts here than in Kent, in order that twice as many churches were aligned north of east. In addition, the fact that the mean alignment in Cornwall is even further north of east than in Cumbria, therefore aligned with a more northerly (later) sunrise in Spring, would mean later building starts in Cornwall than in Cumbria. This appears to confirm that sunrise at the time of building commencement was not a determinant of alignment. Churches aligned towards Jerusalem This theory can also be seen to have no basis, as Jerusalem was shown to the southeast of Britain, even on medieval maps. Southeast is 135º True and not one medieval church in the whole of this survey is aligned in that direction. Chapter 5 168 ALIGNMENT THEORIES: OVERALL CONCLUSIONS When tested against a large dataset, none of the earlier explanations for the variation in the alignment of churches can really be seen to have any validity. The earlier theorists have frequently used very small samples of churches which can provide misleading results; some have interpreted information to suit their arguments, such as Benson (1956) and Searle (1974) (discussed in Chapter One), while others have based their theories on little more than half of the churches they surveyed and have simply ignored the other half that did not fit their proposition (such as ‘weeping’ chancels). In general, it seems reasonable to conclude that churches were originally vaguely aligned eastwards but with a variance between the extreme alignments of 90º one quarter of the horizon. In some instances it is possible to interpret particular patterns, such as misaligned chancels (particularly rebuilt ones) as reflecting a need for realising a more accurate orientation over time, which was achieved when the opportunity arose through rebuilding. What is not clear is whether the intention of church-builders changed over time from a general sunrise alignment to one which aligned churches more generally towards east, or whether developing technology allowed a more accurate realisation of an original intention to align eastwards. Chapter 5 169 PART THREE ANALYSIS AND SYNTHESIS Chapter 6 170 CHAPTER SIX VARIATION IN CHURCH ALIGNMENT ACROSS THE COUNTRY Oh, East is East, and West is West, and never the twain shall meet. Rudyard Kipling, Barrack-room ballads, 1892. INTRODUCTION This chapter examines in more detail the east-west pattern of spatial variation in church alignments across the country that was established in Chapter Four and demonstrates that it also exists for the parish churches within the county of Norfolk, where the density of churches is so much greater than elsewhere and therefore provides a large sample in a relatively small geographic area. It explores a number of possible reasons for the differences in alignment; firstly that the use of a compass to set out churches may reflect the movement of magnetic north over the period in which the churches were built; secondly, that churches may have been aligned differently in different periods of church building for other reasons than the movement of magnetic north, such as shifts in belief or in the importance of specific directions. It had been hoped to examine another aspect of chronology by comparing the alignment of possible minster, or mother, churches with the remaining churches in the survey, which are more likely to have been built later. Unfortunately, it is not possible to identify a sufficient number of known minster churches to provide a large enough sample of alignments for a robust comparison, particularly when the data are sub-divided by longitude. Lastly, climate differences across the country are examined. Although climate differences in Spring, as measured by the number of frosts and their dates, have been shown not to have created the observed east-west pattern of alignment by affecting the date at which church-building began, the climatic differences across the country later in the year are investigated here to assess whether they may have affected harvest times which, combined with the prospect that an early successful harvest may have been seen as an auspicious time to build a church, may have caused the different patterns of alignment across the Country. Chapter 6 171 SPATIAL VARIATION IN CHURCH ALIGNMENT RESULTS In this survey, the mean alignment results for the counties in the extreme east and west of the country exhibit a difference of 12° (80.4° – 92.4°), as shown earlier in table 4.1 on page 112. The sample size in the individual counties is necessarily limited (as there were no more churches to survey), but since adjacent counties have very similar results, and the results were shown earlier to vary by longitude but not latitude, they have been re-analysed below by longitude. Due to the location of the counties used in the survey, there are natural gaps in the survey data at 4º West and 2º West. The greater number of counties surveyed in the east, and the greater density of churches within them, means that there are no other natural breaks in the data, but the two degree classification was continued for consistency. In order to ensure large enough groups for analysis and to provide more robust results, all the churches located at more than 2° west were amalgamated into a single group. At this level of aggregation, the difference between the mean alignment values in the east and west falls to 6.6° (82.2° to 88.8°), but the larger sample size in each group means that the difference between the extreme ends of the ranges for the east and west at 99% confidence, is still 4.2° (83.6° to 87.8°). In other words it can be stated with near certainty that the mean church alignments for these areas are significantly different and this difference did not happen by chance. The fact that the standard deviation for the mean figure in each of the areas, shown in the seventh column, (which indicates the spread of the data about the mean value), is so similar to the standard deviation for the overall figure, surely indicates that the same procedure for setting out churches was used everywhere, with the same sort of variability (hence the wide range of individual alignments) but subject to a systematic east-west bias. Table 6.1 – Summary of results by longitude Longitude (decimal) No Range WEST 418 48-126 2.00º W + CENTRAL 606 38-121 0.01ºW – 1.99º W EAST 902 54-128 1.70ºE - 0º E Overall 1,926 Chapter 6 Mean 95% conf Range at 95% Range at 99% Standard Deviation % N of East 82.2 ±1.1 81.1 – 83.3 80.8 – 83.6 11.45 75 84.7 ±0.9 83.8 – 85.6 83.5 – 85.9 11.35 70 88.8 ±0.7 88.1 – 89.5 87.8– 89.8 11.22 54 86.1 ±0.5 85.6 – 86.6 85.4 – 86.8 11.62 63.8 172 In Figure 6.1 below, the alignment of every church in this survey (vertical scale) is plotted against its longitude (horizontal). It also shows the calculated mean alignment across the country (the solid red line) which confirms the values shown in the county table (table 4.1, page 112) from around 80° in the west to 90° on the east. The apparent vertical lines in the data are because the longitude of each church was recorded to the closest 5′ (minute) interval, for example 1°, 1°5′, 1°10′, except in Norfolk, where 1′ intervals were used. Church Alignment by Longitude (including trendline) Longitude - degrees west 6.00 5.00 4.00 3.00 2.00 1.00 0.00 -1.00 -2.00 35 40 A lig n m e n t - d e g r e e s f r o m N o r t h 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 Figure 6.1 – Scattergram of alignment by the longitude of every church in this survey, including the calculated ‘trendline’25 25 A Microsoft Excel function which calculates moving mean values using the least squares fit through points. Chapter 6 173 Church Alignment by Latitude (including trendline) Latitude - degrees north 49.00 35 50.00 51.00 52.00 53.00 54.00 55.00 56.00 40 45 A lig n m e n t - d e g re e s fro m N o rth 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 Figure 6.2 – Scattergram of the alignment by latitude of every church in this survey, including the calculated ‘trendline’25 Figure 6.2 presents the same alignment data as shown in Figure 6.1, but analysed by latitude. The trendline closely reflects the longitude data in Figure 6.1, as the most southerly and northerly survey areas (at 50° and 55°N) are in the west of the country (with numerically low alignments) shown by the rises in the trendline at either end; the large concentration around 52.5° - 53°N is mostly Norfolk in the east (high alignment) which shows as the depression in the trendline, and even the small volume of Kent data, where churches also have a high numerical alignment, shows as a dip in the trendline at around 51°N. Taken together, these two diagrams confirm the generalised pattern shown earlier in Figure 4.3 on page 124, by illustrating a distinct east-west (longitudinal) pattern of alignment variation across the country. The northsouth (latitudinal) analysis does not show a similar trend up and down the country, but it does reinforce the longitudinal variation. Chapter 6 174 General discussion of possible causes for the spatial variation in alignment As demonstrated in the previous chapter, there appears to have been an overall desire to orientate churches generally eastwards, rather than towards the specific sunrises that were tested, whether that of the church patronal-saint’s day, Easter day, or on the day when building might have started. The majority of churches (63%) are aligned within 10º of due east. However, tables 6.1 and 6.2 above show clearly that medieval church alignments across the country exhibit significant differences between those in the east and those in the west. The statistical tests of confidence show that this is not a statistical fluke, but represents a real overall difference. This is confirmed by the fact that around three-quarters of churches are aligned to the north of east in counties in the west of the country, whereas church alignment is almost equally divided between those aligned north and south of east in counties in the east of the country. As an indicator of the size of the actual difference between church alignment in the east and west, almost one in every three churches in the west of the country – 239 in all – would need to have its alignment shifted from north of east to south of east to achieve the balance of alignments either side of east that occurs in the east of the country. This raises several broad questions. Since the intention of the builder appears to have been to align the church in a broadly easterly direction, what is the reason for the difference across the country in the overall mean alignment? There are four main possibilities, viz:– a) There was a single overall influence in church alignment, not exactly due east, but close to it, and with a shifting point of focus across the country, or b) There was a single influence, but with stronger or weaker effects in different parts of the country. This could work in either a positive or negative way; in other words, it could result in an increasing number of churches having a more northerly alignment from east to west, or alternatively, a reducing number having a more southerly alignment east to west, or c) There was a single overall influence but one which was counterbalanced, more in some areas and less in others, or d) There were two different influences, one acting in the east, and another in the west of the country. In addition to these possibilities, how did the influence(s) actually work? Was a conscious decision on the alignment taken when each church was set out, or was it a series of unconscious decisions of individual builders, whether influenced by external Chapter 6 175 factors or not, that taken together, realised the pattern? Further, if the difference was the result of a conscious decision was it ever written down or was it merely part of the unwritten ‘craft’ aspects of early masons and architects or some other group that was involved in church planning? The idea that this pattern was created by chance – that hundreds of individual church-builders have caused this configuration west to east across the country without any outside influence at all – seems impossible. Even in East Anglia, where the density of medieval church buildings is at its greatest, few churches are sufficiently intervisible to allow the alignment of one to influence that of another. It does raise the possibility of a local focus, for example something on the horizon, but variations in the actual point of focus would be expected to produce a random pattern, so this does not begin to answer the question of why the difference between alignments in the east and west of the country exists. If churches everywhere were intended to be aligned exactly towards east, and this was done with complete accuracy, then all churches would face east, irrespective of where they were in the country. The introduction of errors in setting out would produce a variation in alignments. If those errors were unintentional, then errors one side of the mean value would be mirrored by those on the other side, producing what statisticians call a ‘bell-curve’, but retaining the same mean value. As the number of errors increases, especially when they differ from the mean by a greater amount, the curve becomes flatter – that is, wider and less tall (resulting in a larger ‘standard deviation’), but if there are similar errors either side of the intended direction, then the same mean value is retained. The ‘bell-shaped’ nature of the church alignments in this survey is clearly illustrated in Figure 6.3 below, where the peak of each curve – red for those in the east of the country and green-dashed for those in the west shows a distinctly different value. This seems to indicate that either the first or last possibility listed earlier is correct, either that there were different points of focus for church alignment east to west across the country from a single influence, or that there were two different influences: one in the east and another in the west. The fact that the alignment curve of the churches in the centre of the country in Figure 6.3 (blue pecked line) sits between those of the east and west seems to point to the first of the two possibilities, in other words, a single influence with a shifting point of focus, rather Chapter 6 176 than introducing a third different influence in the centre of the country for the last possibility. This just leaves the problem of establishing what the influence was. Medieval Church Alignment by Longitude Group 30.0 EAST Percentage of churches 25.0 20.0 15.0 10.0 5.0 134-140 127-133 120-126 113-119 106-112 99-105 92-98 85-91 78-84 71-77 64-70 57-63 50-56 43-49 36-42 0.0 Degrees from North west central east Figure 6.3 – Alignment of medieval churches, summarised by the 2º longitude group used in table 6.1 Variation in church alignment within Norfolk The survey results for the alignment of churches within Norfolk provide, at a local level, another strong confirmation of the pattern of variation in mean church alignment by longitude. Rather than selecting a possibly arbitrary numerical division for this breakdown, the results for the county as a whole have been divided in half and are shown in table 6.2 below. The ranges of the mean results do not overlap at 95% confidence, indicating a likely statistically significant difference in the results, even within this short distance (fractionally over 100 km in total and only 50 km between the centres of these two areas). The different proportions of churches aligned to the north of east in the two halves of the county serve to confirm the real variation in the results – with 62% of churches aligned north of east in the western half of the county and only 50% in the east. Chapter 6 177 Table 6.2 - Alignment of churches in Norfolk by Longitude Longitude (decimal) (East ) 1.11º – 1.70ºE (West) 0.20º – 1.10ºE No. range Mean 279 270 549 56-120 64-128 90.3 87.4 88.9 95% conf. ±1.2 ±1.3 ±0.9 Range at 95% 89.1-91.5 86.1-88.7 Stand. Dev 10.5 11.0 10.8 %N of E 50 62 56 Norfolk Church Alignment by Longitude 30 EAST percentage of churches 25 20 15 10 5 134-140 127-133 120-126 113-119 106-112 99-105 92-98 85-91 78-84 71-77 64-70 57-63 50-56 43-49 36-42 0 degrees from North west east Figure 6.4 – Norfolk church alignment by longitude, shown in Table 6.2 The two curves clearly show that there is a greater number of churches in the west of the county with alignments north of east (numerically low), shown by the fact that the green-pecked line is further to the left than the solid red line on the left-hand side of the graph, and more churches in the east of the county with alignments to the south of east, shown by the solid red line being further to the right than the greenpecked one on the right-hand side of the graph. This reflects the similar pattern shown in the whole survey results, in Figure 6.3 above, but with a smaller difference between the two curves, which is probably due to the smaller geographical distance between the west and east of Norfolk, rather than between the west and east of the country. Chapter 6 178 POSSIBLE REASONS FOR THE SPATIAL ALIGNMENT VARIATION As demonstrated in the previous chapter, variations in specific sunrises cannot explain the noted difference between church alignments in the east and west of the country. A discussion of other factors that may explain the differences in church alignment between east and west are set out below; firstly, changes in the magnetic field of the Earth; secondly, possible chronological differences in church building, and lastly, variations in climate. 1) MAGNETIC VARIATION The idea that churches were set out magnetically towards east, and that the variations in alignment between individual churches built at different dates reflects the changes in magnetic declination (the difference between a magnetic direction as shown on a compass, and the true direction), seems, on the face of it, to provide a simple explanation for the variation in the alignment of churches. The subject of how churches were set out, and whether the compass was a known instrument at the time of much of the church building in England, has been discussed before (see Abrahamsen 1992, 292-303; Hoare and Sweet 2000, 167; Ali & Cunich 2001, 156-157). The results were inconclusive; both Abrahamsen and Ali & Cunich feel that some churches were set out magnetically, but Hoare and Sweet do not; but all acknowledge that the compass was known to the Church by 1187 CE from documentary evidence (Hoare & Sweet 2000, 167), and possibly long before (Ali & Cunich 2001, 156). North, as measured by a compass, is not a fixed point. Due to the constitution of the earth’s core, the magnetic field that the planet generates shifts in relation to the geographic North Pole. This difference is known as secular magnetic variation and has two components – magnetic declination, which is variation in an east-west direction, and magnetic inclination, which is variation in a north-south direction (Merrill et al. 1996, 5-6; Abrahamsen 1992). Magnetic declination was first measured in London in 1576 (Merrill et al. 1996, 3 & 46-47), and has been projected backwards in time by using datable objects which have thermo-remanent magnetism (Clark et al. 1988, 646), enabling a graph of the movement of the magnetic pole to be plotted (archaeomagnetic Chapter 6 179 calibration). These measurements and projections (in Figure 6.5 below) show a combination of magnetic declination (across the graph) and inclination (up and down the graph). M a g n e t i c I n c l i n a t i o n M a g n e t ic D e c l i n a t i o n (degrees from True North) (the figures on the chart line refer to magnetic north location at the beginning of that specific Century, that is 19 refers to 1900) Figure 6.5 – Archaeomagnetic calibration for Britain, normalized for Meriden in Warwickshire (the geographic centre of England) (Clark et al. 1988, 659) Figure 6.6 below shows the author’s extract of the declination element of the combined declination and inclination plot shown above, by year. It also shows the approximate positions of magnetic north, and therefore magnetic east, at the extreme positions in the movement of the magnetic pole over the last 1100 years. Chapter 6 180 Magnetic Declination - England, 900 - 2000AD -30 -20 2000 1950 1900 1850 1800 1750 1700 1650 1600 1550 1500 1450 1400 1350 1300 1250 1200 1150 1050 1000 950 1100 East -10 900 Degrees from true North East of north West of north . North 0 10 North East 20 30 40 Figure 6.6 – Magnetic declination of North between 900 CE and 2000 CE for Meriden in Warwickshire (Author’s extract of declination from the combined plot of magnetic declination and inclination shown in Figure 6.5) During the medieval era, apart from a roughly 70 year period during the fourteenth century, magnetic east was always south of true east, peaking at almost 30º south in 1000 CE. In 1100 CE it was approximately 20° south of true east and was 15° south in 1200 CE, resulting in readings for magnetic east in these years of 120°, 110° and 105° True respectively, compared with True east (90°). Any alignment, or realignment during rebuilding, of a church using a compass before the date when magnetic declination was first measured would have to have been made towards magnetic east. This would mean that churches built between 1100 and 1200 CE, and set out using a compass, would have been aligned between 15º and 20º south of east (between 105° and 110° True), rather different from the observed alignments in this survey, which are concentrated between 80° and 90° True. Chapter 6 181 Table 6.3– Approximate magnetic declination in the medieval era Year 1000 1100 1200 1300 1350 1400 1500 Declination Degrees east of North 28 20 15 7 -7 7 10 Magnetic East °True 118° 110° 105° 97° 83° 97° 100° It appears, therefore, that the magnetic variation not only occurs during the wrong period to explain the overall variation in church alignment, covering far too long a period between its maxima, but also has an insufficient medieval component to the west of north (when magnetic east was north of true east) to allow for the large number of churches aligned to the north of east. In addition, it also has an insufficient range. The overall range of some 55º in the position of north, and therefore east, is too small to explain the variation of 90º in the range of alignments of medieval churches, especially since more than 20º of the range of 55º has occurred since the middle of the seventeenth century, long after all the medieval churches were built. As well as varying over time, magnetic declination also varies by location. In 2002, a compass in Kent showed magnetic north to be 2° 30′ west of true north, whilst at the same time in Cumbria, the difference was around 5° 0′ 26, meaning that magnetic east was north of true east across the whole of the country. This variation across Britain in 2002 is shown in Figure 6.7 below. If churches were aligned according to the magnetic declination then, churches in Cornwall and Cumbria would be likely to face a similar direction, as would churches in Sussex and Norfolk. 26 http://geolab.nrcan.gc.ca/cgi-bin/geomag/MIRP/run_mirp – (accessed 5th April 2002) Chapter 6 182 Figure 6.7 – Actual magnetic declination for 2002 when magnetic north was west of true north. Source26 Figure 6.8 – Approximate magnetic declination during the medieval period when magnetic north was east of true north. The alignments of the churches in this survey do reflect this pattern, with the mean alignment direction varying in a similar, east-west, direction, with western areas exhibiting numerically lower (more northerly) mean alignments. However, although it is not possible to calculate the exact variation in declination across the country for periods well into the past27, it can be determined that any variation would have been in the opposite direction in earlier centuries. During most of the medieval period, when magnetic north was to the east of true north (the other side of true north from where it is now), local differences in declination would also have been reversed. In other words, slightly south of true east in Cornwall, becoming increasingly southerly as one moved east. Therefore, churches aligned with a compass during this period would all have been aligned to the south of east, but by differing degrees. It is also not possible to calculate whether the range of variations in declination across the country would have been more or less than the 2.5° difference currently exhibited. The one published study which has examined church alignment in relation to building date concluded that there was no link between the two. The study of Saxon 27 Pers. Com. October 2001, Larry Newitt, Canadian Geological Service, after being unable to refine his computer program for calculations of historical declination to better than a 10 to 12° range. Chapter 6 183 churches by Hoare and Sweet, summarised in Chapter One, shows that there is apparently no variation in mean alignment between early (seventh- to ninth-century) churches and late (eleventh- or early twelfth-century) churches, during which time magnetic east moved southwards by 20 degrees (Hoare & Sweet 2000, 173; Clark et al. 1988, 659). Overall, their sample of Saxon churches also has a similar mean alignment to all the other alignment surveys, which probably indicates that a similar method was used to set out churches both before, and after, the ‘discovery’ of the compass for church-building purposes (whenever that may have been), therefore precluding the possibility of any magnetic influence on church alignment. Chapter 6 184 2) POSSIBLE CHRONOLOGY OF CHURCH BUILDING The second possibility is that the focus for the alignment of churches changed over time and was nothing to do with magnetic changes, but reflected some other shift, perhaps in church thinking. It is not obvious why a chronological difference in building might result in a spatial pattern of alignment variation, but every avenue has to be explored and if a pattern of alignment variation did appear when analysing different periods of church building, then the cause could be investigated more closely. There are few published comments about a specific beginning of interest in, or indeed any changes in, church alignment. However, Richard Morris’ comment, noted earlier, indicated that there was a new concern with the alignment of churches by the Normans (Morris 1989, 208), implying either that the Saxons had had less interest in the specific alignment, or that the focus had changed. Morris also suggests that realignments were not necessarily limited to larger churches and that there were signs of similar changes in the concern with alignment at village level (1989, 208), although this was shown earlier not to be the case for churches where excavation had been able to identify earlier phases beneath current buildings. This idea was also at the centre of the church research in Hampshire, noted in Chapter One, in which Shore identified the allegedly Saxon or Norman origin of churches by their alignment (Shore 1886, 95-98), but his readings have been shown by the measurements for this survey to be suspect. Previous research by Hoare and Sweet, noted earlier, has indicated that there is no difference in the alignment of churches between the middle-Saxon and SaxoNorman periods, but they examined only a small sample of churches (188 in all, and inevitably only a few from the earlier period). This section re-examines the issue based on the larger sample of churches surveyed here which affords an opportunity to achieve a more statistically reliable result in examining the possibility that churches on early sites were aligned differently from those on later ones. It starts with a brief examination of the churches in this survey analysed by the date of their earliest fabric as noted in the relevant volume of Pevsner and in Anglo-Saxon Architecture by Taylor & Taylor, followed by a comparison, for all the counties in this survey except Norfolk, between the alignment of churches on minster church sites (either known sites or probable sites taken from published sources), with the alignment of the remainder of the churches in the survey. In Norfolk, where records prior to Norse occupation are virtually non- Chapter 6 185 existent (Dymond 1985, 77; Williamson 1993, 92), possible minster-church sites will be identified in two ways. Firstly by using the churches identified by Gerry Barnes in his investigation of church siting in Norfolk (Barnes 1996, 23-36), and secondly by identifying possible sites by combining a series of proxy indicators, then comparing the alignments of both these sets of possible minster-church sites with the remaining churches in Norfolk. It has to be recognised that the current church on a ‘minsterchurch’ site probably bears little relation to the original minster church, but if the current building is an extension or remodelling of the original, then the original alignment will have been retained. Even if the church has been rebuilt, it may have retained the same alignment by utilising all or part of the same foundations. Before addressing this problem, the results of the exercise will be examined to see if these aspects warrant further investigation. Church alignment by the age of the earliest fabric The graph below shows the result of comparing the alignment of all 1,926 churches in this survey, divided into those with an element of Saxon fabric, taken from The Taylors’ Anglo-Saxon Architecture and the relevant volume of Pevsner’s Buildings of England; those churches with elements of Norman fabric as their earliest part, and finally those with only later fabric, again taken from the relevant Pevsner volume. It is appreciated that the assessment in Pevsner’s work is only approximate and cannot take the place of detailed deconstruction of every building, but it is hoped that the evaluation is internally consistent. As Figure 6.9 below shows, there is virtually no difference between the pattern of alignment of churches allocated to the three periods, and it shows nothing like the variation in alignment as that when analysed by longitude, which was shown in Figure 6.3 on page 177. Here, the three curves are not only close together but their peaks are within a couple of degrees of each other, indicating little difference in alignment by the age of the building and therefore extending the conclusions drawn by Hoare and Sweet concerning the lack of alignment variation for Saxon church buildings. The very slight southerly shift in the Saxon curve (red-pecked) to the right is probably a reflection of the slight bias of Saxon buildings in the east of the country, where alignment is numerically higher. Chapter 6 186 Alignment by Age of Earliest Fabric 30 percentage of churches 25 EAST 20 15 10 5 Saxon Norman 134-140 127-133 120-126 113-119 106-112 degrees from north 99-105 92-98 85-91 78-84 71-77 64-70 57-63 50-56 43-49 36-42 0 Later Figure 6.9 – Church alignment by age of earliest fabric Comparison of the alignments of minster churches and other churches Minster churches outside Norfolk It had been hoped to use an additional indicator of church chronology to confirm that there was no real difference in alignment between early and late churches, by comparing the alignments of known minster churches which have been surveyed here with the alignments of the remainder of the sample, on the basis that a church on minster-church sites would have been built some time before the bulk of what are now parish churches. Unfortunately, the number of minsters that can be identified by examining earlier research has proved to be too small to provide a meaningful statistical comparison. Minster, superior or mother churches were identified from the following texts for the counties surveyed here: John Blair - The Church in Anglo-Saxon Society (2005); Richard Morris – Churches in the Landscape (1989); John Blair in Secular Minster Churches in Domesday Book (1987), Richard Morris – Churches, Settlement and the Chapter 6 187 beginnings of the parochial system 800-1100 (1983), M. Franklin in The identification of minsters in the Midlands (1984); N. Rushton - Parochialization and patterns of patronage in Eleventh-century Sussex (1999); P. Hase in The Mother Churches of Hampshire, (1988) and J. Croom - The Fragmentation of the Minster Parochiae of South-East Shropshire (1988). Each of these writers bemoans the fact that documentary evidence is thin and is completely missing in some areas (for example Blair 2005, 319; Morris 1983, 46-48). Several methods were used by the authors to identify minsters: documents where available, particularly charters; evidence of Royal ownership at Domesday; in some cases placenames or large glebe allotments were also employed. In all, this enabled 67 superior churches to be identified from the volumes above which have also been surveyed for this thesis. Many other minster churches are now in towns (and may always have been part of larger settlements) so have not been surveyed as a part of the rural sample for this thesis. Overall, there is little difference between the mean alignments of the two groups measured here – the 67 minsters have a mean alignment of 86.2°, compared with 84.9° for the remainder. Due to the small number of minsters, the statistical confidence in the results is so wide as to make it impossible to comment on the relationship between the two - the range of the mean values at 95% confidence overlaps almost completely. The fact that exactly 67% of both types of church are aligned to the north of east seems to indicate that there is unlikely to be much difference in their actual mean alignment. Unfortunately, the sample of minster sites is far too small to divide up by longitude to investigate whether the same spatial variation applies to them as that identified in the whole sample. Table 6.4 – Alignment of churches on probable minster sites in all counties in this survey except Norfolk Number Probable minsters Non-minsters Chapter 6 Range Mean 67 57-116 86.2 ±2.5 Range at 95% 83.7 – 88.9 1,310 1,377 38-128 84.9 84.9 ±0.6 ±0.6 83.3 – 85.5 84.3 – 85.5 188 95% % N of E 67 67 67 Minster church sites in Norfolk The east of England has few records of the pre-Viking church hierarchy (Williamson 1993; Scarfe 1987; Barnes 1997, 23), so attempts have been made to identify minsterchurch sites using other data as proxies. This was carried out firstly by Gerry Barnes in 1997 using the number of parish contacts (adjacent parishes), evidence of a royal manor or an archaic estate, and charter evidence (Barnes 1997, 23-36). As with the results for the other counties in this survey, the alignment of the minster churches identified that were examined in this survey, shows only a small variation from that of the remaining churches in the county. The sample of minster sites is too small, with too wide a range at 95% confidence level, to provide an accurate figure for comparison with the very much larger group of non-minsters, enabling no real conclusions to be drawn. Table 6.5 - Alignment of churches on possible minster sites in Norfolk (Barnes 1997) Number Possible minsters Non-minsters Range Mean 40 67-109 87.6° 95% conf. ±2.9 509 549 56-128 89.0° 88.9° ±1.0 ±0.9 Range at 95% conf 84.7 – 90.5 % N of E 63 88.0 – 90.0 88.0 – 89.8 55 56 Secondly, the same exercise has been undertaken as part of this thesis incorporating a greater number of proxies. The full details of the method of assessment are contained in Appendix 9 and the results summarised in the table below. Table 6.6 – Alignment of possible minster churches in Norfolk, based on the proxy assessment outlined in Appendix 9 Top 27 (possible minsters) Remainder Total Number Range Mean 88.3° 95% conf. ±4.7 Range at 95% conf 83.6 – 93.0 % N of E 61 27 67-110° 522 549 56-128° 88.9° 88.9° ±0.9 ±0.9 88.0 – 89.8 88.0 – 89.8 55 56 Table 6.6 shows that there is virtually no difference between the possible minsters identified here and the remainder of the churches surveyed in the county (0.6° difference in the mean alignment), although again the sample of minster churches is so Chapter 6 189 small that the 95% confidence range is very wide and completely envelops the mean range of the much larger ‘remainder’ group. Combining minster sites in Norfolk and elsewhere Combining the 27 identified possible minsters in Norfolk (table 6.6) with the probable minsters identified by the other writers in the other counties of this survey (table 6.4), results in the following overall assessment. Table 6.7 – Alignment of churches on probable or possible minster sites in all counties in this survey Probable & Possible minsters Non-minsters Number Range Mean 94 57-116 86.7 1,832 1,926 38-128 86.0 86.1 95% conf. ±2.2 Range at 95% 84.5 – 88.9 % N of E ±0.5 85.5 – 86.5 64 64 66 As with the figures for Norfolk alone, there is no real difference in the alignment results for the probable/possible minster churches and the remainder. The mean alignment figures are close, the proportions aligned to the north of east are almost identical and the ranges of alignments at 95% confidence for the two groups overlap completely, indicating that a significant difference is not likely. Consequently, the question as to whether the current church on the minster site bears any relation to the alignment of the original church, possibly dating back to before 800 CE, does not need to be addressed here. Chapter 6 190 CHRONOLOGY OF CHURCH BUILDING: CONCLUSIONS The investigations here into possible alignment differences through a chronological pattern of church building, by examining possible differences between minster or superior churches compared with daughter churches and by analysing the earliest fabric of the buildings, seem to confirm and extend the earlier findings for just Saxon churches by Hoare and Sweet, in that there was little difference in church alignment relating to the likely date of building. Little or no difference in alignment can be identified between churches that were likely to have been built at an early date and those likely to have been built later. Whatever method is used to try and establish which churches were minster churches, and whatever proxies are used, the resulting list of minster churches will never be long enough to allow a statistically sound comparison between their alignments and that of the remainder of churches. Different analyses may produce slightly different lists of minsters churches in Norfolk and elsewhere, but they will not produce a much longer list, therefore the same problem of having too small a sample still exists, and will continue to fail to provide a meaningful basis for comparison with the remainder of the churches. Chapter 6 191 3) CLIMATIC VARIATION The third possible influence on church alignment across the country is climate. The date of the last frost in spring and the number of days with frost were both considered in an earlier chapter when examining the possibility that churches were aligned with sunrise on the day that building started. Although the earlier weather data (1960-1989) did not produce any specific climate patterns, either north-south, or east-west, the later more detailed data (1971-2000) did show a slight variation between counties in the east and west of the country – those in the east having more frost-days – but they did not directly explain the spatial variation in church alignment because the incidence of frost days was considerably higher in the counties in the centre of the country than on the coast, either west or east, away from the warming influence of the sea during the winter. It is generally accepted that the west coast has a milder climate than the east coast and this relative mildness in the west owes much to the warm waters of the Gulf Stream and is evidenced by semi-tropical gardens all along the west coast of the Country, ranging from Tresco Abbey gardens on the Isles of Scilly in the south, to Inverewe in Scotland in the north, at latitude 57.8°N, which “contains many tender species from Australia and New Zealand despite being further north than Moscow or Hudson’s Bay in Canada” (King 2003, 567-568). The following section looks at the possibility that climatic differences across the country, rather than affecting the spring start of building, might affect the other end of the year, by examining the likelihood that climate variations result in different harvest times for the grain crop. Harvest is a particularly important time in any society that relies on local harvests for food availability. Thanksgiving for the successful harvest is a basic celebration for all concerned with the land; for the fertility of the soil, for the replenishment of stocks, for sufficient food for the following year and for seasonal employment for everyone, and it occurs in all arable economies (Evans 1971, 23-25; Baker 1974, 30). A successful harvest, particularly an early one, could have been a particularly good time for a Saxon land-owner to celebrate and express thanks to God by building a church on his estate. If there was a difference in harvest times across the country, could this have acted as an influence on church alignment, particularly the spatial element? Chapter 6 192 Harvest Celebrations All the major world religions, except for Christianity, have a fixed date for celebrating harvest completion28; for example, the Jewish calendar celebrates the Feast of Tabernacles (also known as the Feast of Ingathering) on the fifteenth day of Tishri, the seventh month (September or October on the western calendar)29. The Celtic festival of harvest commencement (Lughnasadh) is celebrated on August 1st, the same date as the Anglo-Saxon Hlaefmass (loafmass) or Lammas30 (Baker 1974, 126; Ross 1974, 289), but there never appears to have been an equivalent religious celebration after the completion of harvest. This was possibly due to the variability of harvest completion, thus making the establishment of a suitable fixed date difficult. As will be seen in the following analysis, harvest completion can vary by as much as three months in different areas of the country in different years. A fixed date that was reasonably close to harvest one year could easily occur early in the gathering-in during the following year, at great inconvenience. In the United States, Christian harvest festival is celebrated on October 31st 31, which, as will be shown, would allow for all but the very latest individual village harvests in the very worst years in England, but would be more than two months later than the earliest harvests in the best years. Local harvest time is fixed each year by a combination of long-term regional climate throughout the year and the annual local harvest-time weather conditions. Overall climate determines the general time when ploughing, sowing and harvesting can be considered: when the ground is dry enough to plough; when it is warm enough to sow; when the seasons are warm and wet enough for the crops to grow; and dry and warm enough for them to ripen. Local harvest weather conditions, however, determine whether or not the crop can be gathered in when it is ready for harvesting, and, as will be seen in the following analysis, can vary by several weeks between villages that are close together which must benefit from the same overall climatic influences. In other words, local weather conditions determine when the days are dry enough for the cutting and bringing-in to be carried out; whether the crop is standing, rather than flattened, which slows harvesting; if the quality is good enough with little mould and rot (before large-scale grain drying became feasible); and in some cases, whether harvest was 28 http://www.harvestfestivals.net/harvestfestivals.htm (accessed 27th June 2008) http://www.harvestfestivals.net/harvestfestivals.htm (accessed 27th June 2008) 30 http//:www.bbc.co.uk/religion/religions/paganism/holydays/lughnasadh.html (accessed 12th June 2008) 31 http://www.harvestfestivals.net/harvestfestivals.htm (accessed 27th June 2008) 29 Chapter 6 193 actually finished at all, when the seed heads of flattened crops could lie sprouting on the ground. For a robust statistical assessment of harvest times across the country to be made it would be necessary to collect thousands of records to ensure that the influence due to local harvest-time weather conditions can be identified and separated from any underlying variation caused by differences in climate. Sufficient harvest and local weather records do not exist for the eleventh or twelfth centuries (when many churches were being built) and it would be an immense task to collect sufficient later medieval harvest dates from bailiff’s accounts and manor court records, with little possibility of finding accurate weather records to accompany them. It was therefore decided to use late nineteenth-century harvest festivals as indicators. Local newspapers regularly reported harvest completion and harvest festival dates as well as comments on, and assessments of, the local weather at harvest time. This substitution is based on the assumption that any variations in overall harvest times across the country in the nineteenth century which are due to climatic differences between areas are likely to be similar to those during the eleventh century. The overall climate in southern England for these two periods, as measured by the Summer and Winter Indices32, was very similar. In both periods in southern England, on the Greenwich Meridian, there was a similar excess of wet summer months over dry ones, and a similar excess of mild winter months over cold ones (Lamb 1977, Fig 13.3, 35; and Table V4a & b, 562-5), so any variations in climate and local harvest weather between the areas studied will start from a similar base. Between these two periods, the climate was warmer during ‘the little optimum’ also known as the ‘early medieval warm epoch’, between 1150 and 1300, and colder during ‘the little ice age’, between 1550 and 1700 (Lamb 1977, 404408). Twentieth-century harvest festivals were not used because, although weather records would be more accurate and less subjective in character, harvest festival dates would be less indicative of the actual date of harvest, more an indication of the limitations of the local church calendar, particularly with the growing prevalence of team ministries, where services in small rural churches may only be held every second, third or even every fourth week. 32 Indices developed by H. H. Lamb to enable climate comparisons to be made for the last 1000 years (Lamb 1977) Chapter 6 194 Investigation of Harvest Festival dates in England Harvest festival dates have been collected and analysed for the thirty years between 1870 and 1899 for four areas of similar landscape which are part of this survey, in order to establish if there was a consistent difference in harvest dates between the areas: the Vale of Taunton in Somerset; the north and west coastal plain of Cumbria; Holderness in the East Riding of Yorkshire, and an area which covers parts of southeastern Norfolk, northern Suffolk and part of the eastern Norfolk Broads (known locally as the Isle of Flegg) [all of which will be called the “Norfolk Broads” for ease of reference]. The specific area harvest details are shown in Appendix 10 on pages 346350. Figure 6.10 – Areas investigated for Harvest Festival dates Chapter 6 195 A harvest festival was first recorded as a church service in 1843, both at Morewenstow in Cornwall33 and at East Brent in Somerset34 and it became a regular feature of the church calendar in most areas in the following decades. By the 1870s, church harvest festival services in many parishes were becoming the norm and in most areas had taken over from the village-based or farm-based private harvest homes which had gained a reputation for drunkenness. As the Beverley Guardian reported on 24th September 1870, “Many harvest festivals in Yorkshire churches in the last week show the supremacy gained by the church over the old system of village festivals where feasting and riot were too prominent”. These church-based festivals were frequently followed by large feasts or teas, but with no mention of alcohol. In some parishes in the early part of the period in the Broads area, private harvest homes were reported in the East Suffolk Gazette on specific estates/farms, particularly Sotterley in 1871, 1873, 1874 and 1878; Hales in 1873 and 1875; Earsham in 1875 and 1878 and Hedenham in 1879. When dates of these celebrations could be compared with services of thanksgiving in the parish church, there was a period of between four and seven days between the two, implying that harvest was completed just before the church service, which would have had to fit in with the regular church calendar with the approval of the rector or vicar. In Somerset, as in the Broads, private harvest homes appear to have gone out of fashion at the end of the 1870s, apart from on the estates of Viscount Bridport, particularly in Brent and Chard, and at an estate in Cricket St Thomas, with regular newspaper reports of 600 people sitting down to meals, right through into the 1890s, but where harvest homes were held and both dates recorded, there was a similar period between them as in the Broads. There were only two harvest homes recorded in the other areas studied, both in Cumbria at Ellerby, with none at all in Holderness. The quotation from the Beverley Guardian above, shows that the tradition of harvest homes had died out there by the beginning of the period studied. In a few of the larger parishes in the areas studied, particularly around Beverley (Holderness), Beccles (Broads) and Taunton (Somerset), announcements were made in the newspapers that the harvest festival would take place in the church on a particular date, usually on the next Sunday, presumably with the knowledge that harvest was almost complete and indicating that even in the largest, semi-urban, parishes there was still a tie between completion date and its more formal celebration. 33 34 http://www.harvestfestivals.net/harvestfestivals.htm (accessed 27th June 2008) Somerset County Herald, Notes and Queries, April 30th 1927 & Nov 20th 1934 Chapter 6 196 Harvest Festivals Examined In all, 2,691 harvest festival dates were recorded for the four areas between 1870 and 1899. For the Norfolk Broads, the dates of 634 harvest thanksgivings or festivals at 97 parish churches were gleaned from reports in the East Suffolk Gazette and the Yarmouth Mercury. In Somerset, harvest celebrations were recorded from the Somerset County Gazette & Bristol News from 1870 to 1878, which incorporated the Devonshire News from 1879. Covering an area of approximately 15-20 miles radius from Taunton, a total of 940 harvest celebrations in 164 villages and towns were noted. Harvest celebrations in Cumbria were extracted from the West Cumberland Times from 1874 (when it began publication) to 1899, and from the Cumberland Pacquet between 1870 and 1873. In the years when both papers were published, considerably more local news was recorded in the Times; consequently the entries from the Times were used, rather than from the Pacquet throughout the period, and a total of 567 harvest celebrations in 92 parishes was recorded. The tradition of church celebrations of harvest completion seems to have come later to Cumbria than to the other areas studied, with few harvest services recorded before the early 1880s. Newspaper reports show that some clergy felt the service to be pagan, akin to nature worship, with specific references over the years in the Cumberland Times, in Great Broughton in particular, of “refusing to decorate our church like a horticultural show”. Harvest festivals in Holderness were collected from the Beverley Guardian & East Yorkshire Advertiser, a total of 550 celebrations in 78 parishes. The national weather picture for the period studied indicates a series of wet years in the 1870s, particularly towards the end of the decade; very cold years in 1880 and 1881; while the weather in the middle years of the 1880s was good, with a series of very dry and sunny years in the 1890s, particularly in 1893, 95, 96, 98 and 99 (Lamb 1977, table V33, 623; Stratton 1978). Thus there was a range of overall weather conditions within which any local variations should be able to be identified. Comparison of harvest dates in Somerset and the Norfolk Broads The mean date of harvest festivals during the thirty-year period was four days earlier in Somerset, on 21st September, compared with 25th September in the Broads. The earliest annual mean date during the period is five days earlier in Somerset (12th Sept to 17th Sept) and the latest mean date of any year was the same – the nationally disastrous Chapter 6 197 harvest of 1879 – on the 8th October. In all, the Somerset harvest mean date was earlier than the Broads in 22 years, on the same date in one year and later in six years, with no comparison possible in 1883 (damaged microfilm). In each of the six years when harvest in the Broads was earlier, local weather appears to have severely affected the Somerset harvest, with local rain or storms. In these years, the Somerset County Gazette reports mention: Poor harvest in many cases due to poor drainage and storm damage (Somerset County Gazette 23rd Sept 1872). More than 8′′ of rain in ten days. Harvest delayed by rain (3rd Oct 1873). Prolonged harvest, longest in living memory (9th Oct 1875). Thunderstorms during harvest. Harvest disrupted due to tempestuous weather (14th Oct 1876). 2.35′′ rain on September 30th (7th Oct 1876). Heavy rain during harvest. Harvest disrupted by tempests – 8.1′′ of rain in September (6th Oct 1888). Crops beaten down very much due to incessant rain. Storms in the Taunton area (22nd Aug 1891). This severity is not reflected in the weather reports for the same periods in the Broads area. In eleven of the twelve years where there were no exceptional local weather conditions (1870, 1874, 1884-87, 1892-93, 1895, 1897-99) and where the national weather was described as good or dry, Somerset harvest was an average of five days ahead of that in the Broads, indicating a basic difference in climate between the two areas, allowing Somerset to finish harvest earlier. Examining the earliest celebrations of harvest in the two areas produces a similar relationship but with a greater difference. In Somerset, 23 harvest celebrations (2.4% of the total) were held in August, with nine of them in 1896, three of which were held as early as August 23rd. In the Broads, none was held in August, with only twelve (1.9%) held before the 10th September, the earliest held on September 4th in 1899. In Somerset, the early harvests occurred in 1876, 1887, 1892, 1896, 1897, 1898 and 1899, whereas in the Broads only four of these dates coincide, early harvests occurring in 1870, 1873, 1876, 1884, 1896, 1898 and 1899, indicating that local weather conditions play at least as important a part as overall climate in determining when harvest could be completed. In 1873 the mean Broads area harvest date was two days earlier than that in Somerset, where harvest was “delayed by rain” (Somerset County Gazette 3rd October 1873). Chapter 6 198 Comparison of harvest dates in Cumbria and the Norfolk Broads On average, harvest festivals were seven days later in Cumbria, with an overall mean date of October 2nd, compared with September 25th in the Broads. Despite this difference, the average harvest was actually earlier in Cumbria for five of the thirty years covered. In 1878, a very early harvest, the difference was nine days, and in 1879, a very late harvest, the average festival date was five days earlier in Cumbria, although only a few dates were recorded. In 1893, when very dry harvest weather was recorded in both areas, the mean date was two days ahead in Cumbria, whereas in 1896 the mean date in Cumbria was five days earlier, where the weather was dry, compared with a rain delayed harvest in the Broads. So, if the whole of the harvest period was dry in both areas, then the Cumbrian harvests were gathered in slightly before those in the Broads area. In two of these years, 1878 and 1896, the earliest harvest festivals in Cumbria were recorded in August, one on the 27th, two on the 29th and two on August 30th considerably earlier than the earliest festivals in any year in the Broads (with only three dates before September 7th, and none before the 4th September). In the best years, the earliest harvests in Cumbria were as much as a week ahead of those in the Broads. Comparison of harvest dates in Holderness and the Norfolk Broads On average, harvest celebrations were nine days later in Holderness, on October 4th compared with Sept 25th in the Broads area, two days later than in Cumbria. Also, in a similar way to Cumbria, despite the regularly later average harvest date, in four of the thirty years examined, the mean harvest festival date was earlier in Holderness than in the Broads - in 1893, it was four days earlier, three days earlier in 1887 and 1899, and one day earlier in 1884. In most of the remaining years, the weather during all, or at least part, of the harvest in Holderness was a problem, with many poor weather reports in the Beverley Guardian, for example: Severe thunderstorms to the east of Beverley flattened almost all the remaining crop to be harvested in that area (20th September 1873). … the disastrous effects that the heavy cold rains and fogs during August have had on the crops (18th September 1875). Severe thunderstorms with copious hail fell yesterday, severely damaging the standing crop (1st September 1877). Harvest operations brought to a standstill due to the heavy rain of the entire last week (18th September 1880). Chapter 6 199 Harvest operations continue extremely slowly, heavy rain has flooded all the low-lying land (3rd September 1881). There is still harvesting in some areas, the crops being laid down (November 25th 1882). Harvest in Holderness greatly interfered with since the weather broke, much has still to be reaped (15th September 1883). The cold northeast winds and absence on sun in August has delayed ripening in many areas (5th September 1885). A most unfavourable harvest month, with 31 days rain since St Swithin’s day [in 60 days since July 14th] (14th September 1889) But for excessive heavy rains, we would be well on with harvest (30th August 1890). In the nationally disastrous harvest of 1879, the latest recorded harvest festival was celebrated in Leconfield on November 2nd, with even later dates recorded in 1882 November 5th in both Walkington and Middleton. The quote from the Beverley Guardian noted above, shows that in 1882, in some parts of Holderness, harvest was still going on at the end of November where harvest festivals were not celebrated at all. In the years when harvest weather did not interrupt the gathering-in in Holderness – 1884, 1887, 1892 and 1899 – the mean harvest date was earlier than in the Broads. In only one of these years, 1892, was the harvest in the Broads delayed by weather, where there was “a late start due to rain” (East Suffolk Gazette 10th Sept 1892). The earliest individual harvest festival dates in Holderness also happened in these years – 2nd and 3rd September in 1887, 2nd September in 1899, 4th September in 1874 and 4th and 5th September in 1884, which are between three and eight days earlier than the earliest harvest festivals in the Broads area in these years, and earlier than the earliest harvests in the Broads area of any of the 30 years studied. Harvest-time weather and overall climate Variability in the annual mean harvest festival dates within each area further emphasises the difference in local harvest weather conditions between the areas. In Somerset and the Broads the harvest-time weather was fairly stable. In Somerset, the mean festival date for each year was more than seven days different from the overall mean date for the area for the 30 years studied in only five of those years – two earlier and three later. Similarly, in the Broads area the mean date only varied by more than seven days in four years – two earlier and two later. However, in Cumbria, the mean harvest dates varied by more than seven days in ten years – six earlier and four later – and in Holderness it was more than seven days earlier in six years and later in seven Chapter 6 200 years, indicating far less stable harvest weather conditions. The detailed dates for each of the areas are shown in Appendix 10 on page 346. When weather throughout the whole harvest period was equally good in Somerset and the Norfolk Broads, Somerset’s harvest was several days earlier than in Norfolk. Despite the fact that weather almost always intervened in the Cumbrian harvest, when it was equally good in Cumbria and the Norfolk Broads, as it was in 1893, then harvest in Cumbria in some parishes was five or six days earlier than the earliest in Norfolk. Similarly, when harvest weather was good throughout the whole period in Holderness, the earliest festivals were several days in advance of those in the Broads area. So, given good weather at harvest time, the underlying climate appears to benefit the west of the country particularly, and to a lesser extent, East Yorkshire, more than in East Anglia. Perhaps this is because colder winters or later frosts in the southeast of the country might delay planting or germination. Arthur Young, in his report to the Board of Agriculture on the agriculture of Norfolk, implies this by referring to the “North and northeast winds … and the climate consequently colder and more backward in Spring” and that these winds also “severely affect Suffolk” (Young 1804, 2). This assessment is confirmed by a later report to the Board for the County of Lancaster (which includes the southern part of the Cumbrian area of the church survey) where Holt referred to the Pennines “screening Lancashire from the ungenial eastern blasts and frosts … which infest the countries (sic) that border upon the German Ocean” (Holt 1813, 2-3). In 1899, the mean harvest date in Taunton, Cumbria and Holderness was earlier than in the Broads, when references were made to “one of the earliest harvests recorded”, but also to “the backward spring in the east” (Stratton & Houghton-Brown 1978, 129). So, in summary, when weather does not ‘get in the way’ of gathering-in, much, if not all, of the rest of the country enjoys earlier harvests than in East Anglia, certainly in the majority of years in Somerset, and at least in the occasional year in Cumbria and Holderness. Despite these underlying differences in climate, which indicate a difference of a few days in harvest date east-west across the country, the local harvest-time weather patterns were very variable at a micro level in all of the areas; something illustrated by the fact that harvest times vary considerably in the same year between villages located fairly close together. Even in the years when the earliest harvests occurred, other villages close by were recording harvest festivals up to several weeks later. For Chapter 6 201 example, around Taunton in 1896, harvest festival was celebrated at West Monkton (4km northeast of Taunton) on 23rd August and at Kingston St Mary (3km north of Taunton) on August 30th, whereas at Combe Florey (some 7km further to the west) harvest was celebrated on 20th September, on September 27th at Staple Fitzpayne (6km to the south of Taunton), and on October 4th at Hatch Beauchamp, six weeks after Monkton, but only 8km apart. HARVEST FESTIVAL DATES 1870-1899 (smoothed by pairing dates) 100 90 80 70 60 50 40 30 20 10 20 Aug 22-Aug 24-Aug 26-Aug 28-Aug 30-Aug 01-Sep 03-Sep 05-Sep 07-Sep 09-Sep 11-Sep 13-Sep 15-Sep 17-Sep 19-Sep 21-Sep 23-Sep 25-Sep 27-Sep 29-Sep 01-Oct 03-Oct 05-Oct 07-Oct 09-Oct 11-Oct 13-Oct 15-Oct 17-Oct 19-Oct 21-Oct 23-Oct 25-Oct 27-Oct 29-Oct 31-Oct 02-Nov 04-Nov 06-Nov 0 Somerset BROADS Cumbria HOLDNS Figure 6.11 – Comparison of harvest festival dates in Somerset, Norfolk Broads, Cumbria and Holderness Comparison of harvests dates in east and west Norfolk Harvest Festival dates were collected for the west of Norfolk and the adjacent silt fenland in the north eastern part of Cambridgeshire to allow a comparison of the harvest dates here with those analysed earlier in this section for the Broads area in the east of the county. They were taken from the Lynn News & County Press for the period 1880-1899, as there were too few recorded during the 1870s to allow a meaningful comparison. In all, 389 harvest festivals held in 103 villages were noted. Chapter 6 202 Overall, the mean festival date was five days earlier in the west of Norfolk – September 21st compared with September 26th in the Broads area for the period 18801899. The mean date was earlier in the west on fifteen occasions, occurred on the same date in three years (1881, 1888 and 1891) and was later in two years – 1886 and 1889. In 1886, the Lynn News reported on September 4th 1886 that “Harvest is gravely deficient with backward crops only now ripening”, whilst in 1889, on September 1st, that “harvest is well forward this year, with good weather it should be finished early”, followed on September 8th by a note of “Incessant rain for the last week has set back cutting and gathering in”. These same two years in the Broads were only average years, with mean dates of September 24th and 27th, and the newspapers in the east of the county reported no particular weather problems. There were six individual harvest festivals in the west of the county which were earlier than the earliest in the east of the county (September 4th), three of them were held in August (27th, 28th and 30th) and three others held at the beginning of September (1st and two on the 3rd). In the four years that these early harvests occurred (1887, 1895, 1896 & 1897) the earliest harvest in the Broads area was on Sept 10th, Sept 15th, Sept 6th and Sept 12th respectively, indicating a larger gap between the earliest dates than the mean dates for the same years. The harvest dates are consistently different between the west and east of the county throughout, not just between the mean dates and the earliest dates, but the overall pattern of harvest dates for the west of the county (shown in red on Figure 6.12) is consistently to the left of the curve (therefore earlier) than for the Broads area, shown by the blue pecked line. Thus, similar patterns exist within west Norfolk harvests as were noted in the four areas used in the main, national, analysis. Similar variation also exists in harvest dates at a micro level to that noted in the main analysis – even in the years of the earliest harvests, other settlements close by were suffering weather problems with much later harvest completion; for example, harvest festival was celebrated on 30th August in Terrington in 1896, but not until October 9th in Upwell, six weeks later but only some 10 km to the south. The earliest ten harvests occurred in ten different villages, with locations evenly balanced between Fenland, the lowland river valleys east of the fens, and the greensand scarp that defines the west Norfolk coast. The regularly later harvest dates for the east of the county also seems to confirm the earlier references by Holt and Young to the “backward climate” and “ungenial winds” Chapter 6 203 that affect the east coast, but do not appear to have such a great effect eighty kilometres further inland in west Norfolk. Harvest Festival Dates - West and East Norfolk 1880-1899 12.0000 percentage of harvests 10.0000 8.0000 6.0000 4.0000 2.0000 west norfolk 28-Oct 30-Oct 01-Nov 03-Nov 05-Nov 08-Oct 10-Oct 12-Oct 14-Oct 16-Oct 18-Oct 20-Oct 22-Oct 24-Oct 26-Oct 20-Sep 22-Sep 24-Sep 26-Sep 28-Sep 30-Sep 02-Oct 04-Oct 06-Oct 31-Aug 02-Sep 04-Sep 06-Sep 08-Sep 10-Sep 12-Sep 14-Sep 16-Sep 18-Sep 21-Aug 23-Aug 25-Aug 27-Aug 29-Aug 0.0000 BROADS Figure 6.12 – Harvest festival dates in west and east Norfolk Harvest dates and soil The local variability of harvest festival dates, noted in the previous paragraphs, does not appear to be reflected in differences between soil groups. The mean harvest festival date for the Broads area as a whole between 1870 and 1899 was September 25th; this is made up of 390 festivals in parishes which are on predominantly clay soil of the Beccles series, where the mean date was September 25th, 66 festivals on the Brown Earth/Loam of the Sheringham series on the Isle of Flegg, where the mean date was September 24th and 60 festivals in villages on predominantly clayey and sandy marine alluvium around the Broads themselves where the average date was also September 25th. In the west of Norfolk a similar consistency of mean harvest date is apparent irrespective of the soil. The overall mean festival date for west Norfolk was September 21st; on the “Good sands” of the Fakenham series, the average festival date was September 20th; on the brown earths and sands of the Rudham series, on the western scarp, the average date was September 21st and on the alluvial soils of the fens, the mean date was September 22nd, each of which is consistently earlier than harvests in the east of the county, indicating that climate and weather has a greater effect on the Chapter 6 204 date of harvest than the soil in which the crops were grown. The details are shown in the Harvest spreadsheet as part of Appendix 18. Availability of labour and the Husbandman’s year Although there is an argument that the setting-out of a church can be done at any time of the year, the digging of the foundation trenches and associated tasks would be best done at a time in the agricultural year when labour was available. In a good year, when harvest finished early, there was time left at the end of the farming year before Michaelmas, when labour could have been available for tasks such as digging foundation trenches and carting materials for foundations and walling. In fact, the earlier the harvest, the longer the period of labour availability. There were few slack times in the medieval agricultural year. Manorial records of the thirteenth century illustrate the agricultural cycles through custom obligations, and Homans described four distinct agricultural seasons (Homans 1970, 353-381). The periods he allocated to each of these seasons differ little from those used by Young and others in their reviews of agriculture in various counties some 500 years later, though Winter and Hocktide were amalgamated, and harvest was reduced from eight weeks to around six weeks (for example, Suffolk – Young 1813a, 223; Sussex – Young 1813c, 90). There seems little reason to assume that the agricultural cycle was much different in general terms even as far back as the ninth, tenth or eleventh century; the major slack times were at Christmas, Easter and Michaelmas. At Christmas the ground was likely to be too wet, sticky or hard to dig trenches and too wet for carting materials, especially if it was too wet or frozen to plough; Easter was traditionally the time that the peasants managed their own plots, planting their own vegetables and gardens and Michaelmas was considered to be the end of the farming year. As a continuation of this season in later times, Michaelmas was the time of hiring for labourers and servants in husbandry on annual contracts for much of the country, particularly in the south and midlands, whereas further north, hiring fairs were held on Martinmas (Nov 11th) (Kussmaul 1981, 51). Kussmaul remarked that “the date of change [of contracts] corresponds to the slack after grain harvest, or sometimes followed the autumn ploughing” (1981, 50), although as the investigation of harvest dates in the previous section shows, further north in England, harvest was often not completed until November, perhaps an additional reason for the later end to labour contracts. Chapter 6 205 CLIMATE AND HARVEST: CONCLUSIONS It is not possible to over-emphasise the importance to a small arable-based community of a good harvest for the following year, particularly before the international grain trade could even-out the fluctuations between good and poor years. In the early years of the fourteenth century, the beginning of what Lamb described as the “climatic worsening” of the later-medieval and early-modern period (Lamb 1977, 449-451), the grain harvest failed for at least three years in succession over almost all of Europe (Fischer 1996, 3541), the subsequent famine and disease resulting in a loss of over ten per cent of Europe’s population (Lucas 1930, 61; Fischer 1996, 37) when there were harrowing tales of peasants attempting to survive by eating cats, rats, insects, animal droppings and leaves (Kershaw 1973, 47-50; Fischer 1996, 35) and of parents killing their children for food and children killing their parents (Lucas 1930, 61; Fischer 1996, 37). An early harvest was usually a good harvest as it meant that the crop was ripe and almost certainly gathered-in dry; a particular advantage at a time before large-scale industrial grain drying and resulting in fewer problems with mould. The later a harvest was completed, the more likely the crop was to be wet, either because wet weather delayed ripening or it delayed gathering-in. This is particularly evident in Cumbria and Holderness. In Cumbria, the earliest harvest festivals were six weeks ahead of the mean harvest date for the area, whilst in Holderness the earliest recorded harvest festivals were nine weeks ahead of the very latest – 2nd September compared with 5th November. In these cases, an early harvest would have been a definite reason for celebration, when, in some of the worst years, harvest was not even fully gathered in. These early harvests could easily have been seen as an auspicious time by the estate owner in the pre-Conquest era, and a time to celebrate by the planning of his first church. If it was aligned with sunrise at this time, a church in Cumbria and a church in Somerset would be aligned in a similar direction, which would have been considerably to the north of those set out after the earliest, but later, harvests in the Norfolk Broads, with churches in Holderness aligned in between those in Somerset and in the Broads area. Local weather conditions played an extremely important part in the timing of harvest completion in all the areas studied, even within individual years in each of the areas, and appear to have played a greater part in the occurrence of an early harvest than other differences, such as between light and heavy soil. So much so, that the 23 Chapter 6 206 recorded harvest festivals in August in Somerset occurred in eighteen different villages spread all over the area; the ten earliest harvest festivals in Cumbria were in eight different villages; in Holderness, the ten earliest festivals occurred in nine different villages and in the Broads area, the ten earliest harvest festivals occurred in ten different villages which were widely spread across the area. This spread of early harvest occurrences means that it is reasonable to assume that over an extended period, perhaps of a century or so, most of the villages in an area would benefit from an early harvest completion at some time and therefore have a similar possible “target” for church alignment. CHURCH ALIGNMENT AND HARVEST Overall, in the best local harvest years, the earliest celebrations were held twelve days earlier in Somerset, eight days earlier in Cumbria and three days earlier in Holderness, when compared with the earliest years in the Norfolk Broads. To reduce the possibility that one or more of the areas had an abnormally early harvest during the period studied which could bias the results, the calculation of sunrise position at harvest time, shown below, is based on date of the fifth earliest harvest in each area, in other words the earliest 0.5% to 1% of the recorded harvest festivals, resulting in dates of August 28th in Somerset, Sept 1st in Cumbria, Sept 4th in Holderness and Sept 7th in the Broads. At this time of year, the position of sunrise at latitude 52°N is moving southwards at 4.5° per week, therefore the ten days difference between Somerset and the Broads would mean a 6.5° difference in sunrise position over a level horizon at this time of year35, with sunrise 6.5° further north along the horizon in Somerset. Due to the more northerly location of Holderness and Cumbria, where the daily change in sunrise position is slightly more rapid at this time of year, the seven days difference between Cumbria and the Norfolk Broads would mean a five degree difference in sunrise in Cumbria, and three days difference in Holderness would mean a two degree difference in sunrise position. The actual weekly positions of sunrise by latitude are shown in Appendix 2. 35 http://aa.usno.navy.mil/cgi-bin/aa_rstablew.pl US Navy website for sunrise times for any given location, (last accessed 21st June 2008) converted to horizon positions by formulae from British Sundial Society (Davis 2004) – shown in Appendix 6 Chapter 6 207 Consideration, earlier in this section, of possible differences between the agricultural year of medieval times and that of the Victorian period concluded that there was likely to be little difference between the seasons and crop cycles of the two periods. So, given identical conditions, seeds planted during the spring would take the same time to reach harvest whether in the eleventh or nineteenth century. As noted earlier, at the inception of the Julian calendar, the solar and calendar dates were the same, as they were again after the adoption of the Gregorian calendar in 1752. By the eleventh century, the calendar date had moved six days ahead of the solar date (Duncan 1999, 41-52), so harvest in the eleventh century would have been six solar days behind the same calendar date today. In other words, harvest completion on August 24th in the eleventh century would be the same solar day as August 30th after the calendar correction. Therefore six days have to be added to the medieval harvest date to be able to establish where sunrise occurred then, with reference to modern sunrise positions. < 6 days Harvest 24th Aug Spring > sowing < Spring sowing Harvest 24th Aug 6 days 11th Century > Solar date 30th Aug Post 1752 Figure 6.13 – Illustration of calendar drift for harvest times between the eleventh century and after the Gregorian calendar adoption The position of sunrise at the early harvest dates in the four areas, adjusted for calendar change, is within 2° in every case of matching the different mean values in church alignment found on the ground in these areas of the survey, and is shown in table 6.8 below; 82º in Somerset and Cumbria, 83º in East Yorkshire and 89º in northeast Suffolk and east Norfolk. More importantly, the 7º difference between the sunrise positions related to the early harvest dates in the east and west is the same as the measured difference between the mean church alignments in the same areas. Chapter 6 208 Table 6.8 – Sunrise at end of early harvest Location Somerset Cumbria Holderness Nflk Broads Early harvest Post Sunrise Mean Survey (earliest 5- top calendarposition – church Latitude 1%) change date level horizon alignment in (Medieval date) (+6 days) (see App. 2) County th rd 51ºN 28 Aug 3 Sept 82º 80º 54.5ºN 1st Sept 7th Sept 82º 82º th th 54ºN 4 Sept 10 Sept 83º 85º 52.5º N 7th Sept 13th Sept 89º 87º The church alignments in the actual villages where the harvest festivals were recorded (approximately 350, as some of the Somerset harvest villages were outside the church survey area) follow a similar pattern to the county-based results – Cumbria 81.6º ±3.8, Somerset 83.8º ±3.1, Holderness 84.2º ±4.0 and the Broads 88.6º ±2.1. A good case can therefore be made that sunrise after an early harvest provided the target for the alignment of a new church. Consequently, the different alignment curves shown by churches in the west and east of the country, as shown in Figure 6.1 earlier, indicate that the builders may well have had separate points of focus for the church alignment, with offsetting errors on either side of each mean value that provide the ‘bell-shaped’ results curves. Lending weight to the idea that the mean church alignment in an area represents the result of ‘focus’ based on early harvests there, is the fact that in the four counties used for the harvest analysis, around three-quarters of all churches are aligned within 10° of the mean alignment in their county – 77% in Cumbria, 71% in Somerset and 70% in Norfolk/Suffolk and in East Yorkshire. These proportions are the same as all the other counties in the survey except in Cornwall and Pembrokeshire, which at 60% and 62% respectively, have the lowest proportions of churches close to the County mean alignment value, and are probably the counties with the smallest proportion of arable farming, perhaps meaning that harvest provided less of a focus there. A particularly early harvest might not only be seen as a propitious time to start building a church in the mind of the land-owner, but in addition the fact that labour would have been available for a longer period after an earlier harvest seems to provide a happy coincidence of reasons why this would be the best time of year for such an action. Additionally, if the church was to be built in stone, annual building contracts for churches frequently came to an end in early September for climate-based reasons, Chapter 6 209 particularly in East Anglia (Salzman 1992, 499 & 547-548), so a mason could also have been available for the important task of setting out. Church alignment and harvest within Norfolk The apparent link between the countrywide pattern of sunrise at early harvest completion and church alignment can also be seen at county level within Norfolk - by using the same measures that were used in the national analysis. The fifth earliest harvests in the west and east of the county are on Sept 3rd and Sept 7th – a difference of four days. Adding six days for calendar correction, results in modern dates of September 9th and 13th which, at this time of year, when the sunrise position is moving southwards at 4.5° per week at Norfolk’s latitude36, or 0.65° per day, results in 2.6° difference in sunrise position for the four days between the dates, with sunrise in the west of the county 2.6° further to the north than in the east (84° in the west and 87° in the east – to the nearest degree). Mean alignments for all churches in the west and east of Norfolk, presented in table 6.2 on page 178, also showed a three degree difference between them (87.4° - 90.3°). Similarly, the mean alignments of churches in the villages where the harvest festivals were actually recorded also shows a three degree difference between west and east, (85.3° - 88.6°), shown in table 6.9 below. In both of these cases – all churches and the actual harvest churches – the difference in the mean alignments is a real one, rather than a statistical construct, as the different proportions of churches aligned to the north of east demonstrate, 66% in the west of the county and only 56% in the east. Table 6.9 – Church alignment in the actual villages where harvest festivals were recorded ‘West Norfolk’ ‘East Norfolk’ No Range Mean 79 73 67-108 67-109 85.3 88.6 36 95% conf ±2.2 ±2.1 Range at 95% 83.1-87.5 86.5-90.7 % North of East 66 56 http://aa.usno.navy.mil/cgi-bin/aa_rstablew.pl US Navy website for sunrise times for any given location, (last accessed 21st June 2008) converted to horizon positions by formulae from British Sundial Society (Davis 2004) – shown in Appendix 6 Chapter 6 210 The large number of monastic sites in Norfolk allows an additional examination to be made of church alignment within the county; in particular, whether the alignment of the naves of monastic churches presents a pattern which is similar to, or different from, that observed in the County’s parish churches, and, by inference, whether they might have been subject to the same influences on alignment. The detailed analysis is presented in Appendix 11 and shows that monastic churches do not display the same spatial pattern of alignment across the county; the naves of those in the east of the county are aligned similarly to those in the west, with similar proportions aligned to the north of east. This lends further, if negative, support to the theory that harvest time influenced the alignment of parish churches, because, although monasteries were large landowners with large agricultural interests, monastic churches are most unlikely to have had any decision about building made by an individual or group of individuals who were likely to have been influenced by crop cycles, or indeed by restrictions brought about by lack of labour availability. Religious belief after the Conversion It is impossible to know whether church builders were using sunrise after harvest as a way of finding east or whether the alignment of the church follows the specific sunrise as part of a deeper meaning concerning harvest thanksgiving. Although aligning a Christian church with harvest sunrise might be considered un-Christian, seen in the light of the “Christianized magic of nature-worship and propitiatory rights that survived into the tenth century” (Blair 2005, 169) and that the “rhythm of the seasons and propitiation of the Gods was still a primary focus at this time” (Morris 1989, 50), it seems reasonable to suggest that there may have been other influences involved as well. There are so many local churches of tenth-, eleventh- or twelfth-century date that appear to have been consistently aligned with early harvest sunrise, creating a pattern across the country that lends weight to the idea. There might have been many estate owners who reacted to a particularly good harvest in a Christian way, but such a consistent practice carried out over such a long period and across such a wide area by so many church-builders seems to argue for something such as a fusion with a longstanding pre-Christian religious belief, such as that represented by the celebration of the beginning of harvest at Lughnasadh or Lammas. Chapter 6 211 The details of the overlap of paganism and early Christianity, described by Blair and Morris above, especially at a local level, are difficult to assess. This is exacerbated by the lack of material evidence and the meagre contemporary historical sources for Anglo-Saxon paganism which has been commented on a number of times (e.g. Wilson 1992, 173-175; Branston 1957; Owen 1981; Hutton 1993). However, it has been recognized that the new Christianity was “shaped by the heathenism of the old religion” (Chaney 1960, 197), and there is evidence of apostasy and the switching back and forth between the two religions in Kent and Essex for the amelioration of plague or drought (Wilson 1992, 173-174; Blair 2005, 167) soon after the Conversion, and in Sussex and Northumbria for similar reasons (Wilson 1992, 175). There is also firm evidence of the continuation of other pre-Christian ideas and rituals involving aspects of the worship of nature, such as the Corn Dolly at harvest festivals made from the last sheaf cut, which in Pagan belief used to contain the Spirit of Harvest (Anderson 1995, 17-19; Baker 1974, 28), and the widely seen Green Man. Richard Morris refers to this overlap in the middle-Saxon period as Christianity “forming a crust on the surface of popular culture” (Morris 1989, 62). Such blurring of the religious boundary continued for centuries as the early eleventh-century calls for the extinguishment of other aspects of nature worship in Edgar’s Canons and the Northumbrian Priests’ Law indicate (Morris 1989, 60-62; Rattue 1995, 79-81; Harte 2008, 21). The conversion to Christianity in Estonia took place in the thirteenth century and incorporated a similar blurring of beliefs for a considerable period afterwards. Even five hundred years after the Conversion, some of the several hundred natural healing sites, consisting of groves, trees and wells, which existed alongside the churches, continued to be used, along with a second type of site which was thought to be capable of granting luck, success and welfare (Valk 2003, 572-575). This does seem to support the idea that a similar pattern of mixing Christian and earlier Pagan beliefs could have continued late into Anglo-Saxon England and affected church building. Chapter 6 212 SPATIAL ALIGNMENT VARIATION AND ITS POSSIBLE CAUSES: OVERALL CONCLUSIONS The spatial pattern of alignment variation across the country is both real and statistically robust. During medieval times church alignment in counties in the west of the country had a focus which was some 10º more northerly than that of churches in counties in the east of the country. The possibility that there was more than one factor involved in this difference was investigated, and it was concluded that there was more likely to be just a single influence that caused this spatial variation. It was shown that churches were not set out magnetically towards east because the variations in alignment between individual churches built at different dates did not reflect the known changes in magnetic declination over the same period. Investigations into the possible chronological differences in church building, firstly by comparing the alignment with the age of the fabric of the church; secondly by examining the alignment of early minster churches compared with ordinary parish churches, which are likely to have been built later, seem to confirm and extend the earlier findings of Hoare and Sweet; that there was little difference in the alignment of churches by the likely date of building. However, it can be demonstrated that climate has a definite effect on harvest. When local harvest-time weather conditions do not interfere with gathering-in, there is an underlying difference in harvest dates between the east and west of the country brought about by climatic differences. The position of sunrise on the horizon on the dates of early harvests in five areas in the east and west of the country closely mirrors the observed differences in mean church alignment in this survey. Whether the church builders were merely using harvest sunrise to ‘find east’, or whether harvest sunrise had real meaning to them, there is a strongly-suggested harvest-specific link with church alignment. Therefore a good case can be made that sunrise at this period was providing the focus for the alignment of a church. In addition, more than 70% of the churches in each of these counties are aligned within ten degrees of the local mean – a figure repeated in almost all the other counties in the survey, suggesting that perhaps harvest was an influence there as well. Unfortunately, the ultimate answer to the cause of the variation in alignment lies in the heads of those who actually undertook the tasks of setting-out the churches, and with no chance of regaining that information, we are Chapter 6 213 left only with the ability to work backwards from the results and calculate the most likely reasons. Whilst harvest thanksgiving may seem a little unlikely to the modern ear, it is currently the only explanation that comes anywhere near to explaining the spatial variation in church alignment across the country, whether physical, environmental or geophysical. The possibility that a fluid practice of religious belief was continued long after the Conversion, strengthens the prospect that the Saxons could have ‘hedged their bets’ when it came to applying the new religion to the process when churches were first built and could explain the use of harvest-time sunrise as a focus. In addition, the availability of medieval labour seems to be coincident: the earlier the harvest, the longer the period of labour availability. Further, if negative, support for the harvest-sunrise link is also provided by the fact that there is no alignment variation in the monastic churches built in east and west Norfolk, unlike the variation in the alignment of parish churches there, and, unlike local parish churches, monastic churches are far less likely to have had the building of their church affected, or constrained, by the interests of harvest completion or labour availability. Chapter 6 214 CHAPTER SEVEN ANALYSIS OF THE RELATIONSHIP BETWEEN SLOPE AND CHURCH ALIGNMENT When I see a slippery slope, my instinct is to build a terrace. Jon McCarthy - 1857-1943 (US Congress Member for Nebraska) INTRODUCTION This chapter analyses the survey results in more detail in two areas where the initial analysis of the results uncovered unusual patterns in the relationship between church sites and sloping land. Firstly, it examines the relationship between church sites and the direction of the slope on which they are built, particularly the fact that this association is very different depending on whether the church was built on a levelled platform on a sloping site or built directly on the sloping land. Secondly, it explores the fact that up to two and a half times as many churches that were built directly on sloping land were built on east-facing downslopes as were built on west-facing slopes. In an attempt to find explanations for both of these situations, the topography and environment of the church yards will be compared for both sloping and platformed sites by using data about the siting of the church in relation to the current location of the settlement that it serves and that of the ‘big house’, as well as data about the degree and direction of the slope of the churchyard, in an attempt to determine whether particular slopes may have prompted platforms to be constructed. The ages of the church buildings on the two types of site will be compared to test whether churches on platforms are more recent than those built on to the slope, which might suggest that platforms themselves were a later feature. In order to check whether the considerable bias of the selection of church sites on east-facing slopes could be partly or fully explained by a predominance of landscape that slopes that way, an analysis of the topography of one of the counties in the survey, Norfolk, is undertaken. Chapter 7 215 The importance of a sloping site is also considered by conducting a case study in Norfolk using the significant number of abandoned churches in the county, specifically those that were located in the same parish as an extant church, assessing whether the slope of their sites played any part in the selection of one church over the other for retention, the results are summarised here and the full details are presented in Appendix 12. First of all, as part of the investigation of the first of the survey’s findings – the difference between platforms and slopes – the problems of building on slopes, and the possible impact that the direction of the slope may have had physically on the building of a church, and its alignment, are considered. BUILDING ON SLOPES Building a church on a slope introduces an additional set of problems to those experienced by builders on flat sites, especially in areas of the country where freestone was not generally available as a building material. Mass walling in materials which do not lend themselves to coursing and bonding, particularly flint, where “even a simple wall demanded considerable skill” (Hart 2000, 5), makes dealing with the additional twisting stresses of building on a slope far more difficult to manage. “The generous quantities of mortar required … and the lack of mortar adhesion ... meant that only limited amounts of wall could be built at a time” (Hart 2000, 5). If a church is built directly on a slope, rather than on a levelled platform, aligning it directly up and down the slope, or directly across the slope, makes building for load-bearing and managing the stresses far simpler, usually requiring only the wall at the lower end of the slope to be buttressed, rather than building diagonally on the slope, where the whole building would be attempting to twist out of square. Earlier it was shown that the mean alignment of churches built on slopes was the same as that of the whole sample. It is possible that slope has had an effect on some particular church alignments, but that this has been hidden by compensating alignments of other churches within the group, without affecting the overall mean figure. To establish whether this is the case or not, an attempt has been made to measure the impact that the direction of the slope may have had on the alignment of each individual church by measuring the difference between the church alignment and the direction of the slope. If the alignment of the church is within ± 11° of parallel to, or perpendicular Chapter 7 216 to, the direction of the slope of the land, it was considered that the church could have been built that way to allow for the direction of the slope and was thus “affected” by the slope. This group, made up of four slices of 22º each, totalling 88º (one quarter of the full circle of 360º) is shown in orange on the upper panel in Figure 7.1 below. If the alignment of the church was outside this group (of ±11º from the direction of the slope) it was considered that the church was built on its alignment despite the slope of the land (shown as yellow in the lower panel in Figure 7.1), the slope was therefore classed as having had “no effect” on the church alignment in table 7.1. Figure 7.1 – Possible effect of the slope direction on church alignment Chapter 7 217 It can be seen from table 7.1 below that fractionally over one-quarter (27%) of churches on platformed sites are within this 88º segment – very close to what would be expected from a randomly distributed result (25%) – showing that the alignment of churches on platforms is definitely not affected by the slope direction. Churches on sloping land, however, are slightly biased towards this segment, with 130 (35%) of the churches built directly on slopes falling into this group, indicating that some slopes probably did influence the alignment of the church, causing them to be aligned closer to the direction of the slope, but, since a random distribution would result in approximately 25% (93 churches) this ‘influence’ is probably limited to around 10% of the churches built on slopes (30 or 40 churches of the total of 373). Table 7.1 - Church alignment possibly affected by the direction of the slope of the site No % SLOPING SITES No slope effect 243 65 Slope affected 130 35 Total 373 100 PLATFORMED No slope effect 208 73 Slope affected 75 27 Total 283 100 Having demonstrated that a slightly larger than random group of churches is built close to parallel with, or across, the slope, it would be reasonable to expect that this might appear in the overall alignment figures when churches that were “affected” by the slope were compared with those where there was “no effect”. However, the figures in table 7.2 below show this not to be the case. Of the 373 churches built directly on a slope of greater than two per cent (1 in 50), roughly one-third falls into the “slope affected” category described above; the mean alignment for this group is the same (within 0.1°°) as both the remaining churches built on slopes and the overall mean for the whole survey. The range of slope severity on both slope-affected sites and nonaffected sites is the same (between 2% and 15% - 1 in 50 to 1 in 7), as is the mean slope angle (4.7% - 1 in 22). Chapter 7 218 Table 7.2 - Church alignments on sloping yards, by “slope effect” Slope Effect No effect Slope affected All churches on sloping sites No 243 130 373 Platformed Yards 283 Other Yards 1,270 All churches 1,926 Slope severity Align. range (& mean) Range 2.0-15.5% (4.7%) 55-121 2.0-15.1% (4.7%) 48-120 (4.7%) 48-121 Mean 85.9 86.0 85.9 95% conf ±1.6 ±2.3 ±1.3 Range of mean at 95% 84.3 – 87.5 83.7 - 88.3 84.6 – 86.2 (4.6%) 50-128 38-126 85.5 86.0 86.1 ±1.9 ±0.6 ±0.3 83.6 – 87.4 85.4 – 86.6 85.8 – 864 It is possible that the excess of churches whose alignment is close to the slope direction is merely a reflection of the greater proportion of churches that are built on east-facing slopes (a phenomenon investigated later in this chapter), where churches would be expected to face eastwards. In these cases, the coincidence of directions would mean that the slope played no real part in affecting the church alignments even though they are built down the slope, as that was probably their preferred alignment anyway. However, when the individual church alignments are taken into account, the similar mean alignments between the categories, shown in red in the summary figures in table 7.2 above, do exhibit some differences within them, shown in Figure 7.2 below, which compares the overall alignment profile of the churches in the “slope affected” category with that of all churches built directly on sloping land. The blue pecked line shows the alignment profile for all churches built on slopes, with a similarly shaped curve to that of the survey as a whole, an approximately ‘bell-shaped’ curve, with no major bumps, and centred fairly close to east. The curve for churches in the “slope affected” category (red) shows two distinct clusters, one at each end of the curve, with increased numbers of churches aligned at 55°- 65° and 100°- 115° (circled on Figure 7.2), which have balanced each other out when the mean value was calculated, resulting in the same mean value as the ‘all churches on sloping sites’ group. A case can be made for the slope having had a real influence on the alignment of these particular churches, as the slope direction was some distance from due east. Although the number of churches in these two bumps is small (approximately 20), they do constitute the majority of churches aligned at these more extreme angles. Chapter 7 219 Alignment comparison by "slope effect" 18 EAST 16 percentage of churches 14 12 10 8 6 4 2 128-131 124-127 120-123 116-119 112-115 108-111 104-107 100-103 96-99 92-95 88-91 84-87 80-83 76-79 72-75 68-71 64-67 60-63 56-59 52-55 48-51 44-47 40-43 0 degrees from North Slope affected All sloping sites Figure 7.2 – Alignment comparison by slope effect (the black circles may highlight churches whose alignment was really affected by the slope – see text) Churches with the most extreme alignments Only 7% of all the churches surveyed have an alignment which is more than 20º from the mean for the county in which they are located, but they account for 13% of churches built directly on slopes, shown in table 7.3. Only a small number of churches is involved – and the difference between 13% and 7% amounts to just twenty-two churches. It is possible that only these twenty-two were truly affected by the slope of the site, and could be the same small group that was identified in Figure 7.2 above, as the two bumps in the graph. As this group of churches is only a notional excess, they cannot be identified. Chapter 7 220 Table 7.3 - Churches aligned more than 20º from their County mean alignment - by type of site and slope severity Sloping sites Platformed sites Other sites TOTAL No (%) total 48 (13%) 23 (8%) 74 (6%) 141 (7%) 373 283 1270 1926 1 in 50 – 1 in 20 30 9 - 1 in 20 – 1 in 10 17 12 - Over 1 in 10 1 2 - However, 30 of the 48 churches (with extreme alignments and built on slopes) are built on shallower slopes (between 1 in 50 and 1 in 20) leaving eighteen built on steeper slopes, where any influence of the slope direction would be likely to be greater. These eighteen churches can be identified. Although their alignment is further from east than the vast majority of the whole sample, there is little to distinguish them otherwise. They are spread across the country, located in eleven different counties, and although the numbers are small, making statistical conclusions difficult, half of them fall into the group that was classed as “affected” by the slope (rather than 35% of the total on slopes noted in table 7.1) but otherwise their profile remarkably resembles the profile of all churches built on slopes. They have a similar range of planforms, and ages of earliest fabric, their sites relate to both the village location and manor house in a similar way and the village name chronology is similar. The same applies to the fourteen churches on the steeper slopes but built on platforms, both the church sites and the churches themselves are similar to the whole sample of churches built on platforms, so there appears to be nothing that sets either of the two groups apart. A detailed analysis and additional tables are shown in Appendix 12 on pages 352-356. COMPARISON OF SLOPING AND PLATFORMED SITES There is a fundamentally different pattern between churches that are built directly on sloping land and those built on levelled platforms on sloping sites, in the way that they relate to the compass direction of the down-slope on which they are built. If levelled platforms were used as a way of aligning a church differently, by not being forced into building in a particular way to avoid the problems of building diagonally across a slope, it does not show in the results. Churches on platforms are also aligned within a fraction of a degree of the overall mean, and with a slightly greater range of Chapter 7 221 alignments, of 78° (between extremes of 50° and 128°), compared with the range of those built directly on the slope of 73° (between 48° and 121°). However, the way that the church relates to the direction of the slope of the land is completely different, depending on whether the church was built directly on the slope or on a levelled platform, shown below in table 7.3 below. A completely random distribution of slope directions would see 12.5% of the churches in each of the eight groups. Churches built on platforms are roughly equally distributed across sites with slopes in all directions (ranging from 7% to 19%), with the smallest proportions on slopes facing north and northwest. The distribution of churches built directly on slopes is quite different, with more than one in three churches (34%) built on an east-facing downslope, which is more than twice as many as the next largest group and almost three times more than the proportion that would result from a random distribution. Table 7.3 shows the numbers and proportions of churches built on sloping land, with the direction of the slope divided into 45º segments. The differences are shown even more clearly in Figures 7.3 and 7.4 below. Table 7.3 - Numbers of churches on platformed and sloping sites by direction of downslope, in 45° groups north northeast east southeast south southwest west northwest Total Chapter 7 Platformed No % 21 7 35 12 38 13 50 18 53 19 33 12 34 12 19 7 283 100 222 Sloping No 10 44 128 59 29 28 56 19 373 % 3 12 34 16 8 8 15 5 100 Figure 7.3 – Churches built on platformed sites, by direction of downslope Figure 7.4 – Churches built directly on sloping sites, by direction of downslope (at the same scale as Figure 7.3 above) Having established that the church locations on slopes and platforms are different in terms of the direction of the slope, it is important to see if there are any other distinctions between the two types of site - such as the severity of the slope, their Chapter 7 223 position in relation to the village and in relation to the manor house - which may help to explain the difference. 1. Distribution by location Despite fundamental differences in topography between the areas included in this survey, sloping sites are used for churches across the country. Almost one in five of all churches is built directly on a sloping site. There are broadly similar proportions in every county apart from Bedfordshire, where there are fewer, and Sussex, where there are more, and in every county bar one – Cumbria – churches built on east-facing slopes form the largest group. A slightly smaller number of churches, just over one in seven, is built on platformed sites on slopes, but a very different pattern emerges. As might be expected from the data in table 7.3, there is a more equal distribution, ‘east-facing’ is the largest, or equal largest, group in six counties – Bedfordshire, north Cambridgeshire, east Kent, south Lincolnshire, Norfolk and East Yorkshire. ‘west-facing’ is marginally the largest group in Cumbria and south Hampshire, and in eight of the remaining counties, north Cambridgeshire, west Cornwall, East Sussex, north Oxfordshire, Pembrokeshire, north Somerset, Shropshire and northeast Suffolk, ‘south-facing’ is the largest group. Despite all the hills and slopes mentioned, two-thirds of all the church sites in this survey are on flat or almost-flat sites, and even in the hilliest areas, the majority of the settlements are on coastal plains or near the bottom of river valleys. 2. Slope severity It was shown earlier, in Chapter Four, that there was little variation in the overall alignment of churches built on slopes of differing steepness. There are 373 churches in this survey that are built on sites with a slope of 1 in 50 or steeper, with a mean alignment of 85.9º; 64% of them are aligned to the north of east. The 131 churches built on sites with a slope of more than 1 in 20 have a mean alignment of 84.0º, and 70% are aligned to the north of east, still only 2.1º from the overall mean, but they are disproportionately represented in the west of the country, where alignments are numerically lower. Chapter 7 224 The results in table 7.4 below show clearly that the severity of the slope did not result in the need to create a platform on which to build the church. Sloping and platformed church sites are similarly distributed across the range of slope severity in this survey, with 61% of platformed sites on slopes of less than 1 in 20, compared with 65% of sloping sites; 35% of platformed sites are on slopes of between 1 in 20 and 1 in 10, compared with 31% of sloping sites and 4% of both sorts of site are on slopes exceeding 1 in 10, so there is virtually no difference in the distribution. Table 7.4 - Comparison of slope severity for sloping and platformed sites Sloping sites Platformed sites 1 in 50 – 1 in 20 65% 1 in 20 – 1 in 10 31% Over 1 in 10 4% Total 373 (100%) 61% 35% 4% 283 (100%) In addition, the direction of slope for church sites does not appear to alter by the severity of the slope, with similar proportions of churches on slopes of each direction irrespective of the steepness of the slope, except for the very small group of fourteen churches on the steepest slopes. For example just over half of all the churches built on slopes are on east-facing slopes, irrespective of the steepness of the slope and just under a quarter are built on west-facing slopes, again irrespective of the slope’s steepness (detailed figures are shown in table A7.12 in Appendix 7 on page 333). 3. Church sites in relation to current village and to the ‘big house’ (church/hall focus) The siting of the church in relation to the current location of the settlement centre which it serves is very similar whether the church is built directly on the slope or on a platform; they are equally likely to be in the centre of the current village, equally likely to be at the edge of the current village or isolated from it. Although the relationship between the current sites of church and village would not necessarily have a specific meaning if it were different for the two types of sloping site, a sizeable difference in one or more of the groups may have provided a pointer towards the need for further investigation (detailed figures are shown in table A7.13 in Appendix 7 on page 333). Similarly, the relationship between the location of the church and the manor house is also a complex one, which will also be developed in the next chapter. But, as Chapter 7 225 the results show, fairly similar proportions of churches currently close to the manor house are built on slopes (one in four) as are built on platforms (almost one in five) and consistent proportions are built on slopes facing in each direction within each group, ranging from 23% to 28% next to the hall on sloping sites and between 15% and 23% next to the hall on platformed sites. The detailed figures are shown in table A7.14 in Appendix 7 on page 334. COMPARISON OF THE CHURCHES BUILT ON SLOPING AND PLATFORMED SITES Having established that the sites of churches built on slopes and platforms are similar in terms of slope and location, what remains to be examined is whether there are any differences in the churches themselves on the two types of site. It is possible that the current churches built on platforms were built in a later period than those built directly on sloping sites, and that the reason that apparently drove the focus of church-builders in seeking out east-facing slopes for churches built on slopes had become less important, or that the sites that did satisfy this criterion were already in use. Without deconstructing every church, assessing the age of the earliest fabric in a church is difficult at best, as alterations and extensions over the years frequently disguise and cover the fabric of earlier phases of building. For consistency, the assessment of the earliest phase has again been taken from the relevant volume of Pevsner, accepting that accuracy is difficult and relying on the hope that the assessments are at least consistent. Since the results show that there is no real difference in the age of the earliest fabric of the churches on the two types of sloping sites, it is felt that the possible problems of accuracy are less important than if the results had highlighted a large difference in the ages of the churches, which would have required further investigation. The detailed tables for this comparison are presented in Appendix 7 (tables A7.15 – A7.17 on pages 334-335). They show that there are similar proportions of churches with their earliest fabric in each age group (11th, 12th, 13th, 14th centuries) whether they are built on platforms or slopes, and that this similarity extends to the distribution on slopes in different directions, despite the fact that some of the groups are small; about 1 in 20 of churches on slopes or platforms is of eleventh-century date, Chapter 7 226 approximately 40% are of twelfth-century date, and around one-third of them are of the thirteenth or fourteenth century. The churches on sloping or platformed sites are also similar in terms of size and floorplan. The results are as consistent as those for the age of the fabric, around 40% of churches are of less than 190sq m, whether on slopes or platforms, a similar proportion are slightly larger and around 20% are over 300 sq metres in floor area, whether on slopes or platforms. Around one in seven churches has no tower, three-quarters have a west tower and one in twenty has a central tower, irrespective of the type of site on which they are built. So it can safely be stated that there is little difference in the churches themselves, whether they are built on platforms or directly on the sloping site, offering no assistance in differentiating between the two types of site. COMPARISON OF CHURCH SITES ON EAST-FACING SLOPES WITH THOSE ON SLOPES FACING IN OTHER DIRECTIONS Having concluded that there is little difference between platformed and sloping church sites that could explain the fact that there is a substantial difference in the relationship between the church alignment and the direction of the slope on which they are built, and that there is also little difference between the churches on the two types of site, it still leaves the issue that two and a half times as many churches on sloping sites are built on east-facing slopes compared with west-facing ones. Are there any other differences between sites on east-facing slopes and those on slopes in other directions? Table 7.5 – East facing slopes compared with other sloping sites Slope of yard East-facing slopes Other slopes All Slopes No Range Mean 95% 203 170 373 48-120 54-121 48-121 86.8 84.9 85.9 ±1.7 ±2.0 ±1.3 Range at 95% conf. 84.7 - 88.5 82.9 – 86.9 84.6 – 87.2 %N of East 62 66 64 Again, the mean alignments of the two groups are within one degree of the overall mean, and with a 95% confidence limit of between 1.7 and 2.0, the difference is shown not to be significant as the ranges at 95% confidence substantially overlap, and Chapter 7 227 each has similar proportions aligned to the north of east. There is little difference between the sites, or their churches, on east-facing slopes and the remainder. Churches on east-facing slopes are, on average, slightly smaller than those on other slopes (45% in the smallest size group compared with 32% on other slopes); have slightly smaller yards (39% in the smallest yard group, compared with 26%); but they are located in similar positions in relation to the village and the manor house; have similar age profiles in terms of their earliest fabric; are dedicated to a similar range of saints by their season and are located in villages with a similar range of name origins. In other words there is little to distinguish the two types of sloping site, other than by the direction of their slope. The detailed tables are shown in Appendix 7 (tables A7.1 – A7.11 on pages 331-334). In case the greater proportion of churches built on east-facing slopes is merely reflecting the fact that there is a greater proportion of land that slopes that way, a detailed digital analysis of the topography of Norfolk, is presented below. DIGITAL ANALYSIS OF THE TOPOGRAPHY OF THE COUNTY OF NORFOLK As there appears to be no site-based or church-based explanation for the excess of churches built on east-facing slopes, it might be thought that this imbalance is brought about by a predominance of east facing slopes in the landscape. Norfolk contains a large number of churches built on slopes, and, at 2.5:1, a typical ratio of churches built on east-facing and west-facing slopes, therefore it provides a good case to analyse in detail. This can be done relatively simply using digital data.37 The degree of slope of the land, and the direction of that slope, has been calculated for a grid of points 100 metres apart across the whole county, resulting in 600-700 points in each of the 549 parishes in the county which has had a church surveyed for this thesis. The technical details and additional tables are set out in Appendix 13 on page 357. In order to compare the calculated landscape with the assessment of church sites already surveyed, the sampled points were separated into those which were located on land which had a slope of less than two per cent and those which had a slope of two per cent or above. In 37 The author is grateful to Bill Wilcox, a fellow PGR student, who obtained and organized the raw topographical data Chapter 7 228 each parish, sample points on land with a slope of two per cent or more were further divided into the direction in which the land slopes downhill in 45º segments; centred on north, northeast, east, southeast, south, southwest, west and northwest. Topography analysis details It is most unlikely that any natural terrain would have exactly the same proportions of land sloping in each direction although, as with the church alignment results, the larger the sample becomes, the more likely it is that small variations between directions would even themselves out. Overall, the results for Norfolk show that the general perception that landscape is “random” is more or less borne out by this exercise. Just over two-thirds of the land in Norfolk (67%) is shown to be either flat, or has a slope of less than 1 in 50 (2%). The remaining 33% has a slope of greater than two per cent and an equal distribution of this sloping land in all directions would amount to 4.1% for each of the eight directions used in the analysis, shown in the table below. Table 7.6 shows that the proportions of sloping land in each of the eight directions is fairly close to being equally distributed, with slightly lower proportions than 4.1% sloping to the east, west, southwest and northwest (3.6 – 3.9%), whilst there is marginally more land than average sloping to the north, northeast, southeast and south (4.3 – 4.5%). Table 7.6 – Results of topography analysis for the whole County of Norfolk All Norfolk Sites Proportion <2% slope North 67.0 4.5 Proportion of 2%+ slopes, by direction of slope North South South East East East South West West 4.3 3.9 4.4 4.3 3.8 3.8 North West 3.6 As might be expected, there is a greater proportion of sloping land in those parishes where the church is actually built on a slope (shown in table 7.7 below). Land sloping at more than two per cent rises from 33% over the whole county to 39.4% in these parishes; so the average proportion for each of the eight directions in these parishes would be 4.9%. As with county as a whole, the sloping land is roughly equally distributed, although in these parishes there is a slightly larger proportion with eastfacing slopes (5.3% compared with average of 4.9%). The slightly below average proportion of land with west-facing slopes (4.3%) compared with the higher proportion of east-facing land does slightly dilute the ratio of churches on east-facing Chapter 7 229 slopes to those on west-facing slopes, from the raw figures of 33:13 (2.5:1) to the adjusted ones of 30.67:14.89 (2.1:1). Table 7.7 - Topography of parishes with churches built on sloping sites - direction of down-slope <2% slope Landscape proportion Churches Number of churches standardised to 4.9%38 North North East East South East South South West West North West All slopes 60.6% 4.4% 5.5% 5.3% 5.4% 5.3% 5.4% 4.3% 3.8% 39.4% 4 7 13 2 3.71 6.38 14.89 2.59 1 6 1.12 5.37 12 33 30.67 10.94 78 The topography of the parishes with platformed church sites is the same as that of the sloping church site parishes (60.6% of the parish flatter than a two per cent slope in both cases) and although the detailed breakdown of the slope direction is slightly different, none of the directions is particularly above or below what would be expected for a random distribution. As was noted earlier, the distribution of churches built on platformed sites is far more equally spread so there is no topographical influence that has directed the choice of site. Overall, the topography of the county is similar to that described by Noël Coward as “awfully flat, Norfolk”, despite references to the Norfolk Mountain Rescue Service by the writer/comedian Mike Harding (Harding 1995, 123). Two-thirds of the whole county is flat or almost flat, the remaining one-third has slopes steeper than 1 in 50 and the slopes are almost equally distributed in all directions. The broad similarity between the amount of land that slopes east and the amount that slopes west demonstrates that the bias of churches built directly on the slope for east-facing slopes is not driven by a disproportionately greater area of eastern slopes, and that even when adjusted for the topography, there are still more than twice as many churches built on east-facing slopes, so another reason for this pattern has to be sought. There is no reason to assume that other counties in this survey would produce topographic results that are very different from those in Norfolk. 38 The number of churches in each landscape sector is divided by the actual proportion of landscape in its sector and multiplied by the standard proportion of 4.925% Chapter 7 230 Norfolk case study of lost churches Previous sections have demonstrated the apparent importance of east-facing sloping sites to the medieval church-builder. Whether or not a specific site for a church was chosen on this basis, contemporary with, or prior to, the location of the settlement there will be discussed in Chapter Eight, but the existence in Norfolk of seventy-nine settlements which had at least one church, where one has been lost, fallen into ruin or become disused, provides an additional way of examining the subject. If the slope of the site played a part in the retention of the one church over the other in these villages it might appear in the results. The full details of the analysis are shown in Appendix 14 on pages 358-361, but the slope of the site appears not to have featured in the decision of which of the two churches to retain, as very few of the churches in these parishes were located on east-facing sloping sites, whether they were abandoned or retained. The majority of the abandonments were relatively modern, many as a result of the sixteenth- and seventeenth-century upheavals of the Church (Batcock 1991, 180-184), rather than earlier when the siting on a slope may possibly have played a greater part. Chapter 7 231 SLOPE AND ALIGNMENT: CONCLUSIONS Trying to explain why churches built on sloping sites face predominantly downhill eastwards, whereas platformed sites do not, has not proved possible from the analysis of the factors contained in this survey. Sloping and platformed church sites appear to be the same – they are on the same slopes in terms of their steepness, they also have a similar relationship with both the current location of the village and the proximity of the manor house. Not only are the sites very similar, but the range of churches on the two types of site also appears to be indistinguishable. They are of similar age, according to the assessment in the relevant Pevsner’s Buildings of England; they are of similar size and have similar planforms. The only aspect of the two types of site that is distinctly dissimilar is the direction of the slope. Neither can the predominance of churches built on east-facing slopes when compared to churches on slopes facing in other directions be explained by differences in the sites, or by the churches built on them. The only major difference between the sites is the direction of the slope itself. The distinct tendency of choosing an east-facing slope when building on a slope is definitely not brought about by a predominance of east-facing slopes in the landscape, at least not in Norfolk. A detailed analysis has shown that here, the sloping sites are fairly evenly distributed in all directions, and even when adjusted for a slight topographical bias, there are still more than twice as many churches built on east-facing slopes as west-facing ones. The implications of the patterns of building churches on slopes are considered in the next chapter, particularly the possibilities that these east-facing slopes were either selected by Anglo-Saxons for religious reasons as sites for their graveyards, or that they reflect elements of Christian substitution on sites which had earlier pagan significance. Chapter 7 232 CHAPTER EIGHT LOCATION OF CHURCHES, ADOPTION OF LOCAL CHURCH SITES, DEVELOPMENT OF LOCAL BURIAL, AND CHRISTIAN SUBSTITUTION At Castle Sowerby in Cumbria, St Kentigern plunged his staff into the hillside (actually a west-facing slope) at the site where he had chosen to preach and a spring issued forth. Taking this as a sign, he commanded that a church be built on the spot. (Cumbrian legend) INTRODUCTION This chapter examines the relationship between the location of the church site and its settlement, in an attempt to determine whether the selection of the site had any effect on the location of the village which it served. The timing of the adoption of the church site, particularly in relation to the changing settlement patterns during Saxon times, is investigated by expanding on earlier landscape research. Locations where there is a coincidence between middle-Saxon settlement sites, identified by others, and church sites, are investigated in more detail by surveying the topography at each church and the possibility is considered that the current church site was selected early in the settlement process, perhaps just as a burial ground. It is frequently thought that many of the country’s small rural churches were sited by the manorial lord on his own land, often close to the manor house; but the considerable bias of slope-located churches towards east-facing slopes, noted earlier in the results of this survey, would be an odd one if the church sites were chosen after the siting of villages on these slopes, as eastfacing slopes are not an ideal location for an arable-based community – as the land is slower to warm, on both a daily and seasonal basis. Therefore, the possibility that the site, for at least some of these churches, was chosen first because of its suitability as an early graveyard, to be followed by the rest of the settlement, is examined. The development of local Christian burial, and the position in the late-Saxon church law codes of graveyards without churches, is also considered. Chapter 8 233 It is often thought that various types of earlier site, which may have been considered as having had a ritual use, have influenced the location of some churches – barrows, standing stones, springs and other features have all had churches built next to them, or even over them – with many examples listed in Chapter Two (Morris 1989, 50-91; Blair 2005, 183-195, 221-228, 374-383; Rattue 1995; Eaton 2000; Bell 2000). The incorporation of such sites was encouraged by the early church hierarchy under the broad heading of what has been called Christian substitution. The possibility that this type of site may have had a link with the east-facing slopes that have been shown here to have been so common for church sites is considered. Exactly what constitutes an easterly view in topographical terms is also discussed. In addition, an analysis is also made of other factors associated with the church or the site, in conjunction with the direction of the slope, in an attempt to establish whether a combination of elements might explain the imbalance of churches on slopes between east- and west-facing sloping sites. It was originally intended to make a detailed comparison between the sites of churches based on their position in the hierarchy, to enable an assessment as to whether the decisions taken over the siting of a minster church were the same as those taken when siting inferior churches. However, the generally accepted difficulty of identifying minster churches on the ground, highlighted by so many writers (Blair 2005, 319; 1987; 1992; Morris 1983, 46-48; Franklin 1984; Hase 1988; Rushton 1999), and the small numbers involved, has prevented the production of any meaningful results. Chapter 8 234 LOCAL CHURCH SITES IN THE ANGLO-SAXON PERIOD This section examines where the church site fits into the sequence of development of settlements during the Saxon era. It uses already published fieldwalking studies and settlement analysis to determine whether a site is from the middle-Saxon or late-Saxon period and analyses them in detail. It was noted in Chapter Three that the majority of the analysis of landscape changes throughout the Saxon period and early Norman period has concentrated on the changes in settlement and agricultural patterns. The rural church and its site, both in terms of when it was built and where it was built, have been very much a secondary consideration in relation to settlement change, and appear to have been viewed as part of the changes in lordship and land ownership in lateSaxon times. Richard Morris has referred to the fact that “there has been no systematic attempt to consider … village religion as part of village studies” (Morris 1985, 49); while others have referred to historical geographers’ study of settlement leaving “little room for the study of burial in the landscape” (Lucy & Reynolds 2002, 5). General studies that have considered when churches were built in villages seem to have concluded that it was usually a ninth- to eleventh-century process (Blair 2005, 368-374; Morris 1989, 140-167), although some detailed local studies (for example in Wade-Martins 1980b, 41 & 73; Newman 2005, 483; Taylor 1983, 153-157; Jones & Page 2006, 184-185), have identified situations which, the analysis here will suggest, point to a far earlier start for the selection of local religious sites, perhaps as early as the beginning of the settlement nucleation process. In general terms, if a religious site, particularly a graveyard, was chosen during the settlement nucleation process, then it seems likely that it would have been adjacent to, or within, the settlement, for two reasons; firstly for religious reasons, recognizing the importance of incorporating the dead into the community of the living (Blair 2005, 245; Morris 1989, 13; Penn 1996, 45; Thompson 2004, 170–206), and secondly for convenience of access. A location any further afield would have resulted in inconvenience, both for access to the site, and interference with any field layout, particularly if the graveyard conformed to the size of ‘God’s acre’. How then, was religion delivered to the residents of dispersed settlements immediately after nucleation? There was probably just one church serving an area, the minster church; with local facilities provided by a preaching site, perhaps with a cross; possibly using the local graveyard in, or close to, the settlement. Chapter 8 235 Any discussion of middle-Saxon ecclesiastical arrangements is inevitably dominated by the so-called ‘minster model’, in which the conversion of the population and their integration into the Church was guided by teams of clergy based at important early churches (the minsters), many on royal estates, to which large parochial territories were attached (Bassett 1992, 26-28; Blair 1985; 1988a; 1992; 1996; 2005; Foot 1992; Morris 1983; 1989). The development of the ecclesiastical system throughout the Anglo-Saxon period and its ultimate fragmentation into the parochial system of the medieval period has been discussed at length (see Everitt 1986, 196–224; Morris 1989; Cambridge and Rollason 1995; Hall 2000; Pestell 2004; Blair 2005; Hoggett 2007). Each of these writers identified the latter stages of this process as being the lordly churches built on smaller local estates in the tenth or eleventh centuries and after. However, there are several different forms of evidence that point to religious sites in settlements in the middle-Saxon period, some of them possibly containing churches, in many counties across the country, but particularly in East Anglia. East Anglia provides the strongest evidence because of the existence of two types of pottery. The first is Ipswich ware which was given its name because the only known kilns for its manufacture were in Ipswich but it was distributed throughout East Anglia. It was made between the second quarter of the seventh century until the mid-ninth century (Jennings 1981, 12), although some argue that production was not begun until the first quarter of the eighth century as it is not found in furnished burials up to circa 700 CE (Blinkhorn 1999, 8-10; Geake 1997, 90). The second pottery type is known as Thetford-type ware and was produced in large quantities at several town sites in East Anglia from around 850 CE (Jennings 1981, 14). The transition from one pottery type to the other provides a well-defined boundary for the change from the middle-Saxon to the late-Saxon period in the middle of the ninth century. Anglo-Saxon sites in Norfolk Substantial amounts of pottery from the middle-Saxon period have been found around the church in several parishes in Norfolk. In some of these cases the late-Saxon settlement site moved away from the church site, but in other cases, the spreads of middle- and late-Saxon pottery overlap, or are coterminous; and, with no ability to determine to which period the church or its site belonged, the automatic assumption has Chapter 8 236 been that the church site is related to the later of the two periods as that was the more likely timing for the building of the church. At Heckingham in the southeast of the county, a substantial number of finds of middle-Saxon pottery, with little late-Saxon ware, was discovered in a limited area around the church by Alan Davison with substantial finds of late-Saxon pottery 100 metres away (Davison 1990, 16-17 and figs 7 & 8), but the church site remained where the middle-Saxon settlement had been, leaving the church site isolated into modern times. A similar situation was discovered at Wormegay in west Norfolk, but with a far greater separation between the settlements of the two periods. Here, a large scatter of middle-Saxon ware was discovered adjacent to the church along with a very small volume of late-Saxon pottery which was taken to indicate that the settlement had moved to a new site where a much larger amount of late-Saxon pottery was found, shortly after the end of the middle-Saxon period, over 800 metres away to the west, again leaving the site of the church isolated (Silvester 1988, 143-150), which it still is. Peter Wade-Martins’ fieldwalking study of areas around churches in the Launditch Hundred in central Norfolk also discovered several sites of middle-Saxon activity which were close to, but occasionally distinct from, the sites of late-Saxon settlement (Wade-Martins 1980b, 25-91; Williamson 2003, 97). At Mileham, the current church is sited in a scatter of solely Ipswich ware (Figure 8.1 below), with later pottery spread to the north and west along the road, away from the church site (WadeMartins 1980b, 41). Wellingham, close to Mileham in Launditch (shown in Figure 8.2 below), also has middle-Saxon and late-Saxon pottery scatters which are distinct. The church is located on a slight spur of land in a high position, and is in the centre of the Ipswichware scatter (Wade-Martins 1980b, 72-73); the late-Saxon settlement (identified by Thetford-type pottery) spread down the slope away from the church site. In three other cases in the same study of Launditch, at Weasenham All Saints, Tittleshall and Horningtoft, the church site is located at the conjunction of overlapping middle-Saxon and late-Saxon pottery scatters (1980b, Fig 31, 62; Fig 28, 55; Fig 9, 26), therefore it is not possible to determine to which period the site relates, a similar situation to that at the DMV of Caldecote, a few kilometres to the west, where pottery of both periods was found “just to the south of the church site” (Wade-Martins 1980b, 80). Chapter 8 237 Figure 8.1 – Mileham – Saxon settlement (Wade-Martins 1980b, Figure 23) dark shading = middle Saxon, lighter shading = late Saxon Figure 8.2 – Wellingham (at the same scale as Mileham in Figure 8.1) showing the church site in the scatter of middle-Saxon pottery (Wade-Martins 1980b, Figure 39) Chapter 8 238 A combination of investigative methods has also discovered a middle-Saxon site at Bawsey in west Norfolk, where a substantial Ipswich–ware scatter shows that the settlement, or possible monastery, was located around the top of a hill (Rogerson 2003, 112-114; Taylor 1999, 67-73), where the current isolated Romanesque church ruins are located at the eastern edge of the hilltop, facing east up the Gaywood valley. Figure 8.3 – St Mary’s, Bawsey, Norfolk, at the eastern edge of the hilltop, surrounded by the 50′ contour (revised first series OS 1:10560 map) Figure 8.4 – St Mary’s, Bawsey, at the eastern end of the hilltop, seen from the north side of the Gaywood valley. Chapter 8 239 Middle-Saxon pottery was also found near several other village churches in the Launditch Hundred, sometimes in smaller amounts, often where “local conditions prevented much fieldwork near the churches” (Wade-Martins 1980b, 84): a middleSaxon scatter 100 metres across near the churchyard at Longham (1980b, 34); to the north and east sides of the church at Beetley (1980b, 17); adjacent to the churchyard at Kempstone (1980b, 30); and small amounts were found at Great and Little Dunham, “where fieldwork was particularly restricted” (1980b, 84). When taken together with the sites in the Hundred where greater amounts of pottery was found, this indicates that at least half of the current villages in the area were in existence as settlements in the eighth century, either including, or very close to, the current church site. Only one of the twelve scatters of middle-Saxon pottery discovered by Wade-Martins was not associated with a church site – at Sutton (1980b, 84). Sutton was one of three DMVs in what is now the parish of Tittleshall, and it was located only 400 metres to the south of the middle-Saxon scatter at Tittleshall (1980b, 53-57), which did contain a church site. Elsewhere in Norfolk, middle-Saxon pottery has also been found on the north side of the church at West Acre in west Norfolk (Davison 2003, 212-218), and around the neighbouring church of St Mary’s in Barton Bendish (Rogerson 1997, 21-22). In the far west of the county, in West Walton, a number of scatters were found close to the church (Silvester 1988, 88-96). At Wickmere and Mannington, in north Norfolk, the churches were associated with scatters of middle-Saxon pottery (Davison 1995, 166170), and at Witton, the greatest concentration of Ipswich-ware was found in the vicinity of the church (Lawson 1983, 70-72). In south Norfolk, a concentration of Ipswich ware and bone was ploughed up, identifying the site of the church of the DMV of Middle Harling (Davison 1983, 332-334); the position of the settlement was later confirmed by an excavated discovery of a hoard of middle-Saxon coins (Rogerson 1995, 121). Also, in south Norfolk at Loddon, a single mid-Saxon item was found at the edge of the churchyard, where the area around the church is completely built up, preventing further investigation (Davison 1990, 18; Williamson 1993, 90). At all of the sites mentioned in these paragraphs various amounts of late-Saxon pottery were also found, frequently in the same area as the middle-Saxon ware but occasionally in separate areas where the settlement moved in the late-Saxon period, such as at Heckingham, Wormegay, Mileham and Wellingham. In each case however, the church site remained where the middle-Saxon settlement had been, which seems to Chapter 8 240 point conclusively to the church site being associated with the earlier period. It seems reasonable to assume from this that the remaining church sites in settlements that did not move during the late-Saxon period are also related to the middle-Saxon period. All of the sites mentioned are shown in blue on Figure 8.5 below. In addition to the middle-Saxon sites identified by fieldwalking, records from the Norfolk HER indicate that middle-Saxon surface finds or excavated finds have been made, either pottery or metalwork, in 23 churchyards in the county. Seven of these are in the Launditch Hundred – East Bilney, Swanton Morley, Beeston next Mileham, Great Dunham, Great Fransham, Longham and Little Fransham, the remaining sixteen are spread all over the county – the find locations are shown in red on Figure 8.5 and are listed in Appendix 17 on pages 368-369. These finds indicate one of two possibilities; either that they are from an earlier use of the site, in other words pre-church or graveyard, or that they are yet another indication of middle-Saxon settlements around the church site. Figure 8.5 – middle-Saxon activity in and around Norfolk churchyards –sites in the Launditch Hundred are circled In all, there are 43 distinct associations between middle-Saxon items and current church sites shown in Figure 8.5 above. It is appreciated that the pattern of Chapter 8 241 finds through fieldwalking and the churchyard finds reflect the concentration of fieldwork in particular areas, particularly in the Launditch Hundred, hence the higher than average representation there. However, the presence of finds over most of the county suggests that county-wide fieldwork at the same intensity as in Launditch, where almost one half of all the current villages were in existence in the middle-Saxon period and related to a church site, would uncover a similar concentration elsewhere and therefore a similar pattern to that in Launditch can be inferred across the county, which sees a persistent model of association between the church site and the middleSaxon period. Anglo-Saxon sites in Suffolk Extensive fieldwork has also been undertaken in southeastern Suffolk by John Newman, covering many parishes either side of the Deben valley, close to Sutton Hoo. Summary results were published in 2005 (Newman 2005, 477-487) and displayed a pattern as definite as that found in Launditch in Norfolk, especially the relationship between church sites and middle-Saxon pottery finds. According to Newman, “All the major Ipswich-ware finds have been located near parish churches in the survey area … emphasizing the importance of these areas as nuclei around which later settlement grew” (2005, 483). “Of the twenty-seven parish churches in the area, twelve are associated with Ipswich-ware scatters [although thirteen are shown on his diagram including Foxhall with its lost church – Figure 8.6], and surveys around an additional six churches were not possible”, leaving only nine churches where such pottery was not found (2005, 483), but this does not necessarily mean that there is none there. In addition to the middle-Saxon sites identified here, Rendlesham has to be added to the list of sites that were associated with middle-Saxon finds, as it produced pottery from all periods, and is known as the site of the Anglo-Saxon Royal palace and church identified by Bede (Warner 1996, 115; Carver 2005, 494; Newman 2005, 478). The twelve [thirteen] churches associated with middle-Saxon finds are shown on Figure 8.6 below, along with Rendlesham, showing this author’s highlighting in red. According to Newman, “the remainder of the sites close to parish churches fall into a phase of mainly ninth- or tenth-century expansion” on areas of “less attractive soil, drier heathland and heavier boulder clay”, and he classes these as “daughter settlements, characterized by small quantities of middle-Saxon pottery as well as late- Chapter 8 242 Saxon pottery, a combination indicating a ninth-century origin” (2005, 483), highlighted in blue, below, although the presence of some pottery of the middle-Saxon period at these sites indicates a middle-Saxon origin for the settlement, albeit late in the period. Figure 8.6 – Survey results from the Deben Valley in Suffolk (Newman 2005, Fig 216b, 481) plus this author’s highlighting of identified middle-Saxon settlements adjacent to church sites in red, and late Saxon “daughter settlements” next to a church in blue Chapter 8 243 Fieldwalking evidence indicates that the population in the area “expanded through the late-Saxon period and up to the fourteenth century resulting in a dispersed settlement pattern along the lanes and footpaths in each parish, after which a large number of these small settlements were abandoned around the time of the climatic decline and the Black Death” (Newman 2005, 483). At least five of these churches are completely isolated now – Ramsholt, Melton, Grundisburgh, Great Bealings and Clopton. A settlement named Melton now exists about 1.5km to the southwest of the church and is a suburb of Woodbridge, whereas Ramsholt church has no buildings at all within the best part of a kilometre (see Figure 8.7), and the other three churches are isolated from small villages. In addition, Foxhall consists of just a single farm which incorporates part of the abandoned church as one of the outbuildings (Carver 2005, 493); Culpho church is accompanied by two cottages; Hemley is only close to Church Farm; and even at Rendlesham, the Saxon Royal seat, the church is now only close to the Old Rectory. The very fact that the current church is built where it is in all these cases confirms a continuity of religious use of the site since its middle-Saxon origin, as it would be an impossible coincidence if all of them were built on sites adopted later, but just by chance, amongst middle-Saxon pottery. Figure 8.7 – All Saints’, Ramsholt, in its isolated position. OS Explorer sheet 197 © Ordnance Survey Chapter 8 244 Figure 8.8 – All Saints’, Ramsholt, from the northeast Figure 8.9 – All Saints’, Ramsholt from the south Chapter 8 245 Further south in Suffolk, on the Shotley peninsular, “chance finds of middleSaxon Ipswich-ware are still being made in the churchyard at Stutton”, where the walls of St Peter’s church also contain several large pieces of late-Saxon carved stone, “either from a grave-slab or a cross” (Laverton 2001, 63-64), again indicating continuity of religious use of the site. Like so many others analysed here, this church is also remote from the settlement and is located on an east-facing slope – a steep one. To complicate matters further, there is a spring in the churchyard (Laverton 2001, 106), which may also indicate that the site was selected for the purposes of Christianization. Middle-Saxon pottery has also been found around some isolated churches in north Suffolk, although in smaller quantities than it has in Norfolk, and this suggests to Edward Martin that the churches there may only have ever been associated “with a small group of buildings – perhaps just a manorial complex” (Martin 2001, 5), although this pre-supposes that the manor and church were contemporary, and it could indicate a similar situation to the previous examples; an early local adoption of the site for religious purposes by the people of a small settlement. Whichever of these interpretations is correct in this instance, the pattern of association between middleSaxon pottery and the church site continues in yet another part of East Anglia. Anglo-Saxon sites elsewhere in Britain Christopher Taylor highlights two further examples of the coincidence of middleSaxon settlement and the church, both in Cambridgeshire. The first is Cottenham (identified by J. Ravensdale 1974, 121-123 & Fig 9; Taylor 1983, 157-159 & fig 60), where “the twelfth-century church is on the site of a sixth- to eighth-century village”, 800 metres away from the centre of the “ninth- to tenth-century planned (?) village” (Taylor’s question-mark), and not in the area of “the twelfth-century village expansion”, which stretches 1,500 metres away to the southwest. It appears here that the current church building is a replacement, built in the twelfth century (dateable from its architectural details, according to Taylor) when the village was expanded, but on the site of the original [mid-Saxon] church (although Ravensdale interpreted this as a resiting of the church in the Norman period (Ravensdale 1974, 123)). According to Taylor, if it were to have been first built during either of the two later periods of expansion (during the ninth/tenth or twelfth centuries), the church would have been Chapter 8 246 built several hundred metres away from its current site, in either of the centres of expansion (Taylor 1983, 159). The second of Taylor’s examples is at Burwell, where the church is located within the “original sixth- to eighth-century settlement area”, rather than in “the ninth- to eleventh-century settlement expansion”, or the later “planned medieval settlement” to the north (Taylor 1983, fig 57, 153), again pointing to the presence of a religious site, possibly a church, in the middle-Saxon period. Both of these sites appear to echo the situation at Heckingham and the other cases in Norfolk where the settlement shifted in late-Saxon times leaving the church site behind in its middle-Saxon location. All twelve of the churches in the Whittlewood study in Buckinghamshire/ Northamptonshire “were once integral to their medieval villages, despite three of them lying semi-isolated now” (Jones & Page 2006, 184-185). The authors also feel that the chronology of “village foundation followed rapidly by church building appears to apply in all of them, except possibly at Lillingstone Lovell”. Here, they postulated that the church “may have been located in a small nucleus to which the village was attracted” (Jones & Page 2006, 188). At Leckhampstead however, which they classed as “Whittlewood’s oldest settlement” (Jones & Page 2006, 91), their analysis shows that this was another settlement with a church located in “the pre-village nucleus”, an area of pre-850 CE settlement (Jones & Page 2006, Fig 32, 89), but they feel that the church came “after the nucleation”, which according to their analysis is “post-900 CE” (Jones & Page 2006, 85-91). Since the church is built in the earlier settlement area, it appears that at least the church site is contemporary with it, as it is unlikely that the church would have been built in a later period, but in a slot in the earlier part of the settlement, unless it was already ‘reserved’, perhaps as a graveyard. Jones and Page also suggest that since this church has “an apparent association with the manorial site, [it] might be assumed to have begun life as a lordly foundation” (Jones & Page 2006, 186-187), although the manor is not shown on their settlement plan. In Cornwall, at the western end of the peninsula, there are fourteen parishes, nine of which are classed as ‘superior’ and five as ‘inferior’ (Thomas 1989, 23). Figure 8.10 shows the churches associated with these parishes (in black), it also shows 26 preparochial chapels (highlighted in red), which were part of the early Christian landscape, either located in hamlets or as special-interest chapels, such as for sea-farers (Thomas 1989, 24). Thomas felt that these “proto-parochial” chapels and churches of the “sixth Chapter 8 247 to tenth century” (1989, 25), together with the burial grounds of the so-called lann model, also proposed by Thomas (1971, 49-51), formed a network of enclosed burial grounds prior to the formation of the parish structure with the noted fourteen parishes, after which many of the burial grounds and the chapels went out of use. The lann model has recently been called into question for the period before 800 CE by Petts (Petts 2002, 26-30; Turner 2003, 172), but it is only the date of the enclosure of the burial ground that is being questioned rather than the date or origin of early Christian burial sites separate from Monasteries (Petts 2002, 26 & 30), as Petts goes on to list several eighth-century West-Country and Welsh examples of unenclosed Christian burial grounds (Petts 2002, 30). Figure 8.10 – Western-Cornwall parish churches and associated pre-parochial chapels (after Thomas 1989, 25) – this author’s chapel highlights Coin hoards from the middle-Saxon period as well as those from the cusp of the middle- and late-Saxon periods have been discovered in churchyards in many parts of Chapter 8 248 the country39 (Morris 1985, 50). Their presence indicates that the graveyard must have been in existence well before this time in order to represent sufficient security for the deposition of valuables. What cannot be determined is whether the burial ground had become a churchyard by that time. In each of these areas examined, Norfolk, Suffolk, Northamptonshire/ Buckinghamshire, Cambridgeshire and perhaps Cornwall, there appears to be a firm recurring association between the settlements of the middle-Saxon period - established either from pottery scatters, metalwork finds or settlement analysis - and church sites. It seems certain that these cannot all be special cases and that this is part of a much wider pattern. The lengthy period between the apparent establishment of the site in the eighth or early ninth century, and the building of the current church probably in or after the eleventh century, allows for considerable change in the specific use and importance of the site and in the importance of the settlement itself. Whether a form of church building was present on any or all of these sites at an early date is unlikely ever to be established, except in western Cornwall where they have already been identified, but the reservation of the site for religious use during the middle-Saxon period seems a distinct probability, although perhaps just as a graveyard. If some of the settlements in Norfolk and Suffolk had not shifted from their middle-Saxon origins, then much of the fieldwork research would not have been possible, and the situation of the church in the centre of the village, still on its original site, would have been completely unremarkable, and also entirely undateable. The many examples listed here, particularly those in Norfolk and Suffolk, have highlighted a situation which may be repeated in many, if not most, other villages – a religious site, established at an early stage in the development of the settlement, possibly just as a graveyard. In situations where the settlement subsequently did shift, even as early as the later ninth century as at the sites in Norfolk, the investment made by the community in the generations of burials, and possibly a church building, effectively tied it to its original site. 39 http://fitzmuseum.cam.ac.uk/dept/coins/projects/hoards/index.list.html (accessed 9th April 2010) Chapter 8 249 DEVELOPMENT OF LOCAL CHRISTIAN BURIAL The abandonment of large early-Saxon cemeteries has been argued to be a direct result of the conversion to Christianity (Hoggett 2007, 268-271); therefore the subsequent coincidence of cemeteries and settlements might be seen as part of the same process. The proximity of the living and the dead was Christian practice throughout medieval Europe and is seen to mean that they were still an important part of the community, and formed a focus for Christian worship (Blair 2005, 245; Morris 1989, 13; Penn 1996, 45; Thompson 2004, 170–206). Being part of the settlement, rather than separated from it, is often explained as those buried there were waiting for Judgement Day along with the living residents, whose prayers would help those who were already dead into the afterlife (Blair 2005, 228–245; Penn 1996, 45; Petts 2002, 44; Boddington 1990; Turner 2006), and that the churchyard was a safe place to await resurrection (Gittos 2002, 195). The local cemetery appears to be part of a two-tier hierarchy at this time, where larger burial grounds were located in missionary stations, often in Roman forts, such as at Caister in Norfolk, where 3000-4000 burials dating from the seventh century were found (Darling & Gurney 1993; Penn 1996, 41), but local burials in individual settlements were endorsed in certain circumstances, frequently where distance was a problem (Morris 1983, 49–62; Blair 2005, 240–245; Boddington 1990), along rather similar lines to the later establishment of chapels of ease. Helen Geake also points out that final-phase cemeteries co-existed with early churchyard burials in several places, Norfolk included (Geake 2002, 151-152). How often did middle-Saxon cemeteries evolve into sites containing churches or chapels? The examples of cemeteries that have been discovered did not evolve any further, but that is why they are known, because they were separate from, rather than being obscured by, later churchyards and settlements. As the population grew during the middle-Saxon period, there must have been more burials than during early-Saxon times, but middle-Saxon cemeteries are rare discoveries in Norfolk. In all, 63 Saxon inhumation cemeteries have been discovered in Norfolk, of which only thirteen are from the middle-Saxon period (Myres and Green 1973, 258– 62; Hoggett 2007, 210214). Chapter 8 250 Figure 8.11 – Norfolk, early-Saxon and middle-Saxon cemeteries (data from Myers and Green 1973, 258– 62 and Hoggett 2007, 214) As can be seen from Figure 8.11, despite the small numbers of middle-Saxon cemeteries that have been discovered, both early-Saxon and middle-Saxon types are distributed across the whole county; from west to east and north to south, the major difference being the numbers discovered. Hoggett relates part of the reason for the extra discoveries of early-Saxon cemeteries as due to their greater ‘visibility’ with more metal finds, helped by the fact that they were remote from settlements (Hoggett 2007, 216), similar to the early middle-Saxon cemeteries such as at Morningthorpe, where up to a quarter of the graves contained weapons (Green & Rogerson 1987, 7). There are two known, excavated, middle-Saxon cemeteries in East Anglia that were integrated into their settlements; firstly at Sedgeford in northwest Norfolk, where a slight shift in the location of the late-Saxon settlement, left the earlier phase undisturbed and has enabled excavation of some of the middle-Saxon settlement, which included an inhumation cemetery of over 200 apparently Christian burials within the settlement (Cabot et al. 2004). Secondly in Suffolk, where excavation at Bloodmoor Hill, Carlton Colville, near Lowestoft, has discovered 24 graves integrated into the settlement, which suggested to the writers a Conversion-period cemetery which was separate from the earlier pagan burials on the crest of the ridge close by (Mortimer & Tipper 1998, 14; Chapter 8 251 Dickens et al. 2006, 74-76). The remainder of the middle-Saxon Christian cemeteries in East Anglia are probably still hidden beneath later settlements or churchyards which have not moved from their original site (Geake 1992, 86–7; Newman 1992, 26; West 1998, 317; Hoggett 2007, 319), as is strongly indicated by the number of churches that are located in middle-Saxon pottery scatters identified by the field walking, particularly by Alan Davison at various sites, by Peter Wade-Martins in Launditch – where half the current villages in the area were in existence as settlements then – and by John Newman in the Deben Valley, noted earlier. Under the minster-model, teams of monks based at the minster church would have visited local settlements to preach, and the burial rights for the area are likely to have been vested with the minster. These rights, and their income, must have been carefully observed and closely guarded. What does this say about local burials in villages close by? Such burials can hardly have gone unnoticed by the monks who preached in the area, especially if the graveyard also acted as the preaching site. Either local burial was overlooked completely, or a blind eye was turned, perhaps after payment, as these villages are hardly remote from burial facilities and therefore not able to claim justifiable relief from centralized burial in terms of distance, indicating that the tacit endorsement of local burials may have been much wider than previously thought. Archbishop Wulfstan’s later law codes hint at this by commanding that soulscot was to be paid to the minster to which it belonged, even if the body was buried elsewhere (Morris 1983, 65; Blair 2005, 444), which was the codification of a practice continued since at least the ninth century, when it was mentioned in charters (Gittos 2002, 201; Hadley & Buckberry 2005, 122-123). John Blair suggests that “lay burial in churchyards was exceptional in 650 CE”, but by 850 CE (the beginning of the late-Saxon period) was “starting to become the norm” (Blair 2005, 228), although referring specifically to burial at minster churches. The excavations of middle-Saxon graveyards at Sedgeford and Bloodmoor Hill confirm that the process of local burial probably became common somewhat earlier, and this is similarly suggested by the analysis of settlements outlined in this chapter, and that by 850 CE, instead of ‘starting to become the norm’, local burial may actually have been the norm, although not necessarily in churchyards. This is supported by Lucy and Reynolds who suggest that in the eighth and ninth centuries “many, if not most, were buried in rural cemeteries, perhaps unenclosed, but possibly quite Chapter 8 252 substantial if the example at Chimney in the minster parochia of Bampton is anything to judge by.” (Lucy & Reynolds 2002, 13). Hoggett goes as far as to suggest that “the vast majority of the population appears to have been buried in newly founded inhumation cemeteries situated in middle-Saxon settlements (Hoggett 2007, 322). John Blair also points out that ground used for burials does not need to have been consecrated at this time and there may have been a general belief that the actual ceremony of burial was more important than the specific location, and further, that village burial grounds in parts of Europe were often consecrated later (Blair 2005, 229). In addition, he notes that there were areas in northern Europe where village burial sites co-existed with churchyards (Blair 2005, 228-229), and points to local burials in neighbourhood or kindred cemeteries paralleling those in minster churches (2005, 180). In other parts of northern Europe however, particularly the Frankish parts, burial practices were different from those in Britain, in that furnished burials have been found in churchyards “in many cases”, whereas in Britain, churchyards “almost without exception” never contain grave goods (Geake 2002, 149); an answer for this difference is still sought, but could indicate that, in this country, churches were only built over Christian cemeteries, whereas ‘final-phase’ cemeteries may have been overbuilt in parts of Europe. Cuthbert, as Archbishop of Canterbury in the middle of the eighth century, authorised burials inside towns, and this “created cemeteries everywhere in England” (Morris 1983, 50; Allcroft 1930, 426-427), a situation noted in Hamwic, where ten separate burial sites have been discovered in the middle-Saxon settlement (Morton 1992, 68-77), but their short life suggests that they were encroached upon, “rather than remaining in use to a later period in the manner of rural cemeteries” (Lucy & Reynolds 2002, 13). Subsequently, burial in holy ground “was promoted by the Church as a privilege” (Morris 1983, 50-51), thus recognising that the majority of burials took place outside consecrated ground. This and the previous arguments support the proposition that, at an early stage in their formation, many, if not most, settlements in this country had areas within them that were reserved for burials, which were later formalized as church sites, whether soon after, or after a longer period, cannot be determined. Chapter 8 253 Churches without graveyards, graveyards without churches The analysis here has shown that it is probable that there were a large number of small graveyards in the middle-Saxon period around the country, no mention of which appears in the national picture of religious sites as painted by the various late-Saxon law codes. However, none of the categorizations of churches contained in these codes precludes the possibility that in remote rural areas, well before the tenth and eleventh century, people in small hamlets or proto-villages had reserved and used plots of lands for burial in the same way as Blair had noted in Europe (Blair 2005, 228-229). Penn refers to these sites in East Anglia as “unchurched cemeteries” (Penn 1996, 45). If this did happen at such a local level, were these sites considered at all by the law-makers when preparing the laws and defining categories? Since they may not have been consecrated, they were outside the listed categories and appear to have slipped completely under the radar. Once such sites were established by usage, it would be a natural progression when erecting a church, to put it on the same site, even if it was a new land owner establishing his new estate in the area. Could the building of a church there have helped legitimise the new ownership? It has been suggested that, as part of Christian substitution, the adoption of a site for a church that had had previous ritual significance could add kudos both to the church and to the adopter (Morris 1989, 74; Blair 2005, 382). Perhaps this kudos could also apply to a late-Saxon in-comer, by building a church on a site that had been used for some time by the residents for Christian burials. Of the thirty-five sites mentioned earlier with identified middle-Saxon connections in Norfolk, only one, at Saxlingham, has the hall next to the church (3% of the total), compared with one in seven (14%) of all churches in the county; this reinforces the possibility that these sites were selected for the church by the landowner because of the graveyard that was already established there, rather than build it next to his hall. As Helen Geake suggests, it is possible that in some cases burials may attract churches, rather than vice versa (Geake 2003, 266). In a slightly different context, another possible example of ‘legitimisation’ by the erection of a church could be the keeill at Speke Farm on the Isle of Man, investigated in 2007 by the Time Team, which was thought to be from the tenth century or after, when the Vikings adopted Christianity (Wessex Archaeology 2007, 21), but built centrally in a Christian graveyard which had Chapter 8 254 been in use since around 590 CE40, which itself was in an area used for Bronze-Age burials. This analysis appears to show that the delegation of burial rights was spread through practice rather than by law. The subsequent building of a church may, in some cases, have been effectively ratifying the earlier use of the site as a graveyard when both were consecrated at the same time, as has been noted when church sites have been excavated, when it has often been found that the church was preceded by unfurnished burials (Morris 1989, 152-153; Geake 1997, 267). An early adoption of local burial or the early distribution of burial rights would be likely to mean that folk religion played a much greater part in the process than if local burial was only sanctioned at a later date and imposed from above, when the reasons for, and processes of, burial would have been more focussed on the liturgy and teachings of the Church. This is discussed further when considering local religion of this period and the use of cult sites, later in this chapter. 40 Details from http://channel4.com/history/microsites/timeteam/2007_iom_found.html (accessed 26th Oct 2009) Chapter 8 255 Village then graveyard or graveyard then village Most of the middle-Saxon settlement sites identified by fieldwalking in Norfolk and Suffolk were found next to the church rather than surrounding it (Wade-Martins 1980b; Newman 2005). This could be explained in two different ways; firstly, that the restricted availability of land for fieldwalking prevented investigation on all sides of the church, so that other pottery around the church site might remain undiscovered. Very few churches are completely surrounded by arable land for fieldwalking, even the isolated churches in this survey have often been sited next to a building, frequently a farmyard. Secondly, it could reflect a real situation where the graveyard really was located at the edge of the original settlement. A detailed topographical assessment of each of the sites in Norfolk and Suffolk where scatters of middle-Saxon pottery were identified is outlined in the next section, where it shows that the majority of the sites are associated with east-facing slopes. If the graveyard was located in the settlement, it seems to offer the possibility that the chosen site with its east-facing slope actually determined the position of the rest of the settlement. It is possible that, during the period when settlements were beginning to nucleate in the late eighth or early ninth century, a particular site for both settlement and graveyard was chosen from a number of alternatives as it offered the opportunity to locate on an east-facing slope, whereas a graveyard located outside the settlement could represent a situation where the settlement location had already been established, after which a suitable piece of land was sought for the location of the graveyard close by; and, with a number of available sites around the settlement, the one with the most suitable east-facing slope was selected. Particularly in the first case, this would imply that the need for a site for the graveyard was occurring very early in the process of settlement nucleation. The idea of early small graveyards in Norfolk and Suffolk fits well with the conclusions of Rik Hoggett concerning the speed of the Conversion in East Anglia, in that he postulates that it happened both more quickly and down to a lower social level than has been previously suggested (Hoggett 2007, 328-331). After the Conversion the incorporation of the dead into the community took on a new importance, rather than burial in either the remote cemeteries of the pagan period, or the remote early Christian cemeteries in the Missionary stations. Perhaps the Conversion prompted the search for a new settlement-site or a burial-site with a suitable east-facing slope. If this was the case, it could date the cemeteries to the late Chapter 8 256 seventh or early eighth century. No early-Saxon pottery was found at any of the sites in Norfolk or Suffolk where middle-Saxon pottery scatters were identified next to the church (Newman 2005, fig 216b, 481; Wade-Martins 1980b), indicating that these were all new settlement sites, rather than the fixing of a settlement on a site used in the earlier period. Richard Morris also saw a similar pattern of local burial developing as that described above, but placed it within a later time frame (Morris 1983, 54). He suggests a sequence of remote seventh- to eighth-century graveyards developed adjacent to earlier pagan cemeteries, a process noted at Winchester by Martin Biddle (Biddle 1976, 69), followed by an eighth- to ninth-century transfer to Christian cemeteries or churchyards (Morris 1983, 54), although this second element is later referred to as an eighth- to eleventh-century exercise, which “usually belonged to proprietary churches attached to manorial centres” (Morris 1983, 62). Morris also suggests that graveyards were only rarely established in advance of churches (Morris 1989, 153), and adds that “If churchless burial grounds were a widespread phenomenon in the ninth or tenth century, one would expect a much larger number of incidences where the cemetery failed to evolve from graveyard to churchyard” (1989, 153). Although only two such cemeteries have been excavated in East Anglia, at Sedgeford and Bloodmoor Hill, it does indicate a considerable amount of burial activity, but where the settlement shifted or died out before the burial ground had a chance to be over-built with a church. However, the analysis here seems to point to numerous early cemeteries which became, and remained, the churchyard despite the fact that the original settlement moved, even if the shift in settlement was in the late ninth century, soon after the beginning of the late-Saxon period, as at Heckingham and Wormegay in Norfolk and Cottenham in Cambridge. What cannot be determined is whether the burial ground was still churchless at this time. Whether the early adoption of village burial sites proposed in this thesis occurs with the same consistency elsewhere as it appears to have done in parts of East Anglia, is unclear, although Richard Morris does recognize different patterns of early Christian cemetery development between the Midlands and East Anglia when commenting that “the Midlands appear to be richer in such examples [of pagan cemeteries close to medieval churchyards] than the East Anglian counties” (Morris 1983, 59). However, the settlement analysis by others, noted earlier, in some Midland counties (Cambridge Chapter 8 257 and Northamptonshire/ Buckinghamshire) (Taylor 1983; Ravensdale, 1974; Jones & Page 2006) appears to point to some church sites located definitely within the earliest part of the settlement, which could only happen with regularity if the founding of the cemetery was contemporary with, or soon after, the establishment of the nucleated settlement, rather than later, when a church was first built, otherwise the site would be located away from the earliest settlement centre. Circular, or sub-circular boundaries are frequently considered to be possible indicators of an early churchyard (Rowley 1972, 81; Thomas 1971, 51; Morris 1989, 455; Friar 1996, 121), although this is more likely to be an indicator that it was the first use to be established on a remote site without any neighbours, as a circular boundary is the shortest in length that can enclose a given area and is the easiest to construct if there are no other constraints, indicating perhaps a yard remote from, or on the edge of, a settlement. If an integral yard was set out as part of the establishment of a settlement, it would be most likely to have straight boundaries fitting in with the plots on either side, as did many of the churchyards analysed in Lincolnshire by Stocker and Everson (2007, 61-65). In addition, Richard Morris has identified at least three yards in Wales that have circular boundaries, but which were new in the twelfth century (Morris 1983, 58), confirming a first-use, rather than necessarily an early-use, of the site. In dispersed settlements, perhaps the location of the church or burial ground could be selected with more freedom than as part of a nucleated settlement, since it was not “bound” by the boundaries of the village. The location of freemen-built churches, which occur in East Anglia particularly, means that their actual siting was not determined by lordly locational requirements, but frequently determined by accessibility, pragmatically located between the farmsteads involved (Warner 1986, 43; Williamson 2006, 89), meaning that other factors, such as previously important sites, either an early graveyard located for the same reasons, or a prehistoric ritual site, could still come into play when selecting the actual church site. Chapter 8 258 Detailed landscape assessment of Norfolk and Suffolk middleSaxon settlement sites The analysis in this chapter has established a strong link between middle-Saxon settlement sites and current church sites in parts of East Anglia, possibly through early graveyards. Since a number of these sites (in south Suffolk) were outside the original church survey areas, the topography at all these sites has been subsequently surveyed in more detail in order to assess whether they are located on slopes and whether the link that was established between churches and east-facing slopes in the last chapter is continued in these early religious sites. Norfolk middle-Saxon sites The intensively studied Launditch Hundred, with many churches located next to middle-Saxon pottery scatters, or middle-Saxon finds in churchyards, provides a compact area in which to examine the landscape more closely. It consists of 30 modern parishes, covering an area of about 20km east to west by 15km north to south, and is located firmly in the Central Norfolk Claylands which are described as “poorlydraining stagnogleys formed on Boulder Clay, forming extensive level tablelands” (Williamson 2005, 8). These ‘tablelands’ are approximately 80–85 metres AOD in the west of the Hundred and 65-70 metres AOD in the east of the Hundred (OS Explorer sheet 238), meaning that overall, the land loses fifteen metres in height over a distance of 20 kilometres – flat by any standard. The eastern boundary of the Hundred is formed by the River Wensum, one of the County’s principal rivers, which is some 45 metres below the tableland-level at the bottom of a fairly steep slope. In all, sixteen of the parish church sites in the Hundred have an association with middle-Saxon finds, eleven sites identified by Wade-Martins (1980b) and the seven which had middleSaxon finds in the churchyard recorded on the Norfolk HER (two of which were also identified by Wade-Martins’ fieldwalking). The topographic survey results for this area show that there is a link between east-facing slopes and these sites, although the sample is too small to be able to declare it a significant link. Half of all sixteen sites were associated with east-facing slopes, whereas only one-third of the sites had down slopes in all the other three directions added together (16 of 48 possible slope directions). Just over half of the eleven middleSaxon settlement sites identified by Wade-Martins (1980b) are located on gentle eastChapter 8 259 facing slopes; North Elmham, Wellingham, Mileham, Kempstone Longham and Beetley, although they are too shallow to be included within the two per cent slope category in the main survey; a further four have level land to the east (Horningtoft, Weasenham, Great and Little Dunham), and at Tittleshall the land rises slightly to the east and falls to the west. Both the current church and the Anglo-Saxon Cathedral (Wade-Martins 1980a; Rogerson 2005) at North Elmham are on the 45 metre contour looking out over the Wensum valley to the east, the cathedral site has land falling away to the north and south as well as east. At Mileham, the church is on the 60 metre contour and has a sloping easterly view along the shallow Blackwater valley, a tributary of the Wensum. At Wellingham, the church is located on a small knoll above the village, close to the 70 metre contour, the land falls very gently to both the north and east. At Weasenham All Saints, the church is on the 80 metre contour, the land is level to the east, but falls away slightly to the north and west. The church at Horningtoft is on the 65 metre contour and the land is level for some distance in every direction. At Kempstone, the church, now ruined and in parkland, lies on a spur of land which falls away to the east, north and west, but rises slightly to the south; at Longham the land falls away slightly to the south and is level in the other three directions and at Beetley, the land falls to the east and north, is level to the west and rises slightly to the south. Of the seven sites associated with middle-Saxon finds in current churchyards noted on the Norfolk HER, two are located at the top of east-facing slopes (Swanton Morley and Beeston), Longham is on a gentle east slope and four are on land which is level to the east (East Bilney, Great Dunham, Great and Little Fransham). Swanton Morley is a particularly good example of a church and yard located in a position which maximises both the eastward view from the church and the view of the church from a distance (see Figures 8.12 and 8.13), and has almost the entire churchyard on an eastfacing slope. It is located on the top of a knoll between the River Wensum and a tributary, on the 40 metre contour, with a steep slope down to the Wensum which turns east at this point allowing a view along the valley. Chapter 8 260 Figure 8.12 – All Saints’, Swanton Morley church at the top of the slope, with its graveyard spread out down the east-facing slope Figure 8.13 – All Saints’, Swanton Morley from four kilometres to the east, viewed from the north side of the Wensum valley Chapter 8 261 Suffolk middle-Saxon sites The thirteen sites where current parish churches are located adjacent to middle-Saxon settlements, identified by pottery finds by John Newman (2005), have each been visited and assessments made of any slopes in the landscape. They are located either side of the River Deben. The lower part of the valley is on the southern edge of the coastal Sandlings (Williamson 2008, 29-34), some 20 metres AOD, and those in the upper valley are located on the southern edge of the clay plateau (Williamson 2008, 33), between 30 and 40 metres AOD (OS Explorer sheets 197 & 212). There are three times as many sites sloping down in both easterly and southerly directions than those sloping to the west or north. At nine of the thirteen sites (69%), there is a slope down to the east, three of which were steep enough to have been classified as a sloping site had they been part of the main survey (steeper than 1 in 50) and three of these nine are located on the east side of the river, despite the fact that there is a general westward slope down to the river. At ten sites (76%), the land sloped down southwards, at three (23%) the land had a northward slope and three had a westward slope. Middle-Saxon sites summary The settlement sites, with their possible graveyards, investigated here which are associated with middle-Saxon pottery, in both Norfolk and Suffolk, are dominated by east-facing slopes, even shallow ones, despite the fact that the sites in Norfolk are on essentially level ‘tablelands’ and three of the nine in Suffolk are on the eastern side of the River Deben, where the land generally slopes to the west, down to the river. In all, of the twenty-six sites in the two counties, the land slopes down to the east in fifteen cases (58%), thirteen to the south, seven to the west and six to the north. The opposite, which seems to confirm the importance of the east slope, is that the land only rises in an easterly direction in three cases, whereas there are six rising to the west, five to the north, although none to the south. Taken together, these topographies seem to indicate strongly that an east slope, however shallow, was important when the site was first selected, whether it was for a church at this time or for a graveyard. The fact that this link between east-facing slopes and middle-Saxon settlements and their possible graveyards is extended into the pattern of churches built on sloping sites over most of the country could imply that all of the churches on east-facing slopes are built on early graveyard sites. Chapter 8 262 Unfortunately, the need to identify the shallower east-facing slopes, (of less than 1 in 50) that showed up in the closer examination of middle-Saxon sites in Suffolk and Norfolk, was not recognized until well after the general survey for this thesis was completed and the data were being analysed, so the direction of the those shallow slopes, classed as ‘almost flat’ in the survey – at 369 sites across the country – was not recorded, so may hide a continuing association between church sites and east-facing slopes, however shallow, that was uncovered in the analysis here. Chapter 8 263 OTHER POSSIBLE LINKS WITH EAST-FACING SLOPES Having established in the previous chapter that a strong link exists between church sites and east-facing slopes, even some very gentle ones, which could point to a requirement for Saxon burial sites; the remainder of this chapter investigates other possible reasons why church sites might be located on easterly slopes. Firstly, the eastern view itself is discussed and how else the view might have been achieved; then Christian substitution is investigated as a possible factor in the selection of the site for the church, or whether it was only a coincidence of site requirements for both the prehistoric ritual site and a Saxon graveyard, that meant that a church was built there. An easterly view – how can it be achieved? The apparently disproportionate selection of east-facing slopes for middle-Saxon settlements (possibly graveyards) surveyed here, and for the churches in this survey which was identified in Chapter Seven, indicates that the direction of the slope was specifically chosen for these sites and therefore that it was important to the site’s function. This raises the question of how did those without such topography available to them manage to achieve a similar advantage? The ultimate distance of the eastern horizon can never have been the sole focus for the selector of the site at any church or graveyard, even those on east-facing slopes; otherwise all sites would be located at the top of the slope, or on the highest point possible in the parish, in order to gain the furthest view. Very few churches are built on the highest point in the parish. In Norfolk, only 58 churches (approximately 1 in 9) are sited at a height which is within 10% of the highest point in the parish (for example sited above 90 metres AOD, compared with a parish highest point of 100 metres AOD), whilst twice as many (124) are sited below 50% of the highest point in the parish (in other words below 50 metres AOD compared with a highest point of 100 metres AOD). Chapter 8 264 Table 8.1 – Comparison of highest point in the parish with the height of the church AOD, in Norfolk Number of churches: Number With level horizon Horizon of 2° + elevation 58 343 124 24 549 41 200 64 305 17 143 60 220 Within 10% of highest point Between 11% & 49% 50% of highest point, or below Fenland churches TOTAL However, any church that has a level eastern horizon, even one built on flat land, will have the longest view possible, as it is the elevation of the horizon that determines the distance of the horizon from the observer. Similarly, any church built at the top of a slope of any direction will have a reasonable eastern view, as by definition, being at the top of the slope means that the ground must either be level or downhill in the other directions, so even a church built at the top of a west-facing slope will have as good an eastern view as possible. Consequently, churches at the top of any slope but with a level eastern horizon should be added to those on the east slopes as also representing the best site possible for an easterly view, along with churches built on flat or almost flat sites which also have a level horizon. In order to calculate the number of churches in Norfolk with the best eastern view available, many churches need to be added: 47 churches on east-facing slopes; five of the 31 churches built at the top of slopes facing in other directions, 262 churches built on flat or almost flat sites with level horizons, and 34 of the 68 churches with level horizons that are built on platforms. This results in 348 churches of the total of 549 surveyed in the county (63% of the total), which might be said to have the ‘best’ eastern view possible. In the other counties in this survey where the elevation of the church horizon was measured, the equivalent figure of churches with the best eastern view is 65%, made up of: 68 on east-facing slopes, thirteen of the 61 built at the top of slopes in other directions, 256 churches built on flat or almost flat sites with level horizons and 46 of the 92 churches on platforms, adding a further 315 churches to those built on east-facing slopes. With two-thirds of all the church locations having a good eastern view (63% in Norfolk and 65% elsewhere), the fact that only around one in ten of all churches (9% in Chapter 8 265 Norfolk and 12% elsewhere) are located on an east-facing slope creates something of a paradox. Throughout this analysis, the assumption has been that the east-facing slope was chosen for its view eastwards. If a long easterly view can so easily be obtained from sites with different topographies, there must have been another reason why eastfacing slopes seem to have been sought out over slopes in other directions – the following sections examine some possible reasons. Cult Sites and Christian Substitution Various types of earlier ritual site have been thought to have influenced the location of some churches; barrows, standing stones, springs and other features have all had churches built next to them, or even over them, with many examples identified in Chapter Two. It has been argued that the incorporation of such sites was actively encouraged by the early church hierarchy, under the broad heading of Christian substitution. The possibility that this type of site may have had a link with the eastfacing slopes, and that the earlier use was the sole reason for the link between churches and eastward slopes has to be considered. Many of these sites are naturally associated with slopes. Barrows are well known for being placed near the top of a hill, on a ‘false crest’ presenting a prominent sightline from below; springs, although dependant on geological permeability, are usually associated with slopes; the other classes of site, such as standing stones, henges, trees or even Roman sites, can be associated with either flat or sloping sites. However, the direction of the slope for any of these sites, including barrows and springs, is not critical to their location - any slope direction will do, unless the direction of the slope was an integral part of the meaning of the site for the people who first developed it. The unanswerable question remains – did any of these sites require an east-facing slope? Sunrise at any time of year is seen in an approximately easterly direction, and the association with east has been important to most known religions, as outlined in Chapter Three, and it seems reasonable to assume that this also applied in prehistoric times, given the number of examples of eastward alignment noted earlier in prehistoric contexts in Chapter Three. Therefore, it is logical to assume that an eastfacing slope would have been chosen if there was a choice of sloping sites for an early ritual site. Chapter 8 266 John Blair sees the process of exerting control over what he describes as “openground cult places” as part of “the whole thrust of western European seigneurialization during the tenth and eleventh centuries” (2005, 382), but this firmly refers to the building of churches on these sites. A possible example of such a site is illustrated in Figure 8.14 below:, a remote church, built on top of a hill, serving a village named Ellough in Suffolk, which is named after a pagan site - an Old Scandinavian heathen temple (Ekwall 1989, 164; Mills 1991, 120). Figure 8.14 – All Saints’, Ellough, Suffolk, located at the top of a hill, remote from the small village, which is named after an Old Scandinavian heathen temple The results of the examination of churches known to be associated with middle-Saxon settlement in Suffolk and Norfolk, with their predominance on eastfacing slopes, appear to suggest that the process of ‘exerting control over old sites’, suggested by John Blair for the tenth- or eleventh-century development of churches (Blair 2005, 382), may have been pushed further back in time for the selection of early graveyard sites in the eighth or ninth centuries, well before the late-Saxon churchbuilding phase. It is impossible to determine, after 1,300 years, whether it was the requirements of an Anglo-Saxon Christian, or a prehistoric pagan, that meant that the Chapter 8 267 east-facing slope was selected. It is just possible that some of these slopes were once populated by sacred groves which have died out or have been cleared; or by springs which have dried up as groundwater levels have changed, neither of which would leave much of an archaeological record still visible. It also has to be appreciated that many prehistoric sites could still have had identifiable remains in the eighth or ninth century when establishing an early graveyard site, as the seven hundred years or so between the end of the Iron Age and 700 CE are within the lifespan of some hardwood trees, particularly the oak and yew. Additionally, 700 CE is almost as close to the end of the Bronze Age as it is to the present time. So from this point of view, it is impossible to tell whether transient prehistoric ritual sites determined the locations of some graveyards or churches, as any earlier features, which may have been more obvious then, are no longer visible. Looking at the issue from the other side, even if the church is built next to obvious remains, it is not possible to state with certainty that the church was built on the site because of the remains, rather than because both uses had similar site-selection requirements - an east-facing slope. So, despite the calls from the Pope at the beginning of the seventh century to incorporate local cult sites, proximity of a church and a prehistoric feature cannot necessarily be taken as proof of this process, and this situation is supported by the conclusions of James Rattue and John Blair, mentioned in Chapter Two, when for various reasons they feel that the extent of Christianization has been overstated. This applies particularly to the fact that pre-eleventh-century minster churches appear to have been sited with little interest in Christianizing sites (Blair 2005, 376), and that so many sacred wells or healing wells had been ignored by the church (Rattue 1995, 42; Harte 2008, 93). Richard Morris also suggests that despite paganism still being a force to be reckoned with [in the period up to the Conquest], the issue of churches commonly being built on pagan neighbourhood sanctuaries “is an open question” (Morris 1989, 92). Although Christianization may not have been high on the agenda for the upper levels of the late-Saxon church hierarchy, it does not mean that it was not a factor during the earlier stages of Christianity and at a lower level. The fact that there are a considerably larger number of churches built on east-facing slopes than would be expected by chance cannot be without explanation. On the face of it, the idea that this may have been part of the view that the dead were awaiting the Second Coming from Chapter 8 268 the east seems plausible, except for the fact that an extended eastern horizon is available from many different types of site, including flat ones, apparently making the specific selection of an east-facing slope unnecessary. However, the majority of the church sites based in middle-Saxon settlements in Norfolk and Suffolk, surveyed here, are located on east-facing slopes, which seems to leave the balance of probability pointing to some form of local Christian substitution as one of the reasons for the selection of east-facing slopes over others for the many sites that later contained village churches. The preaching of monks from minster churches at an early stage after the Conversion could have influenced the choice of site and fits in well with Hoggett’s conclusions about the strength, depth and speed of the Conversion in East Anglia mentioned earlier, when the Pope’s calls for the incorporation of earlier religious sites were still relatively recent. The inclusion of a possible insurance element of the assistance and influence of the previous ritual site (which may well have been their fathers’ place of worship) to the Judgement-Day reasons might well also have inclined the middle-Saxon villagers to pick such a site, over one which did not have this added value, as part of the fluid practice of belief after the Conversion that was discussed in Chapter Six on pages 211 and 212. The possibility that other factors, such as the proximity of the manor house and the proximity to the settlement, may also have had an effect on the siting of the church, and therefore on the choice of an east-facing slope over other slopes, was also investigated from the results of this survey and the detailed tables are presented in Appendix 7. The key fact from the analysis is that the numbers of churches on eastfacing slopes, whether as part of a church/hall focus or not, and independent of where they are located in relation to the village, exceed the numbers of churches on all the other slope directions put together, indicating that the direction of the slope was the principal concern and that none of these other factors examined appears to have anything like the same level of connection. All of the other factors considered; the age of the church fabric, the village name origin, the position of the church in relation to both the village and the manor house, whether close to them as at Stody or Fring in Norfolk (shown in Figures 8.15 and 8.16), or isolated from them as at Mundham in Norfolk (Figure 8.17), show little variation when analysed by the direction of the slope, which firmly places the direction of the slope as the key factor when selecting the site. Chapter 8 269 Fig 8.15 – St Mary’s, Stody, Norfolk, from the east - in the village at the top of the slope with its graveyard on the east-facing slope with views along the valley Fig 8.16 – All Saints’, Fring, Norfolk, located near the top of the slope above the rest of the village Chapter 8 270 Figure 8. 17 – St Peter’s, Mundham, Norfolk, at the top of the hill, isolated some distance from the village Minster church sites It had been intended to compare the details of minster church sites with those of churches lower down the hierarchy, in terms of location and site topography etc., to see if there may have been a different pattern of influences on site selection between the different classes of church. The same problems exist for this analysis as they did earlier in Chapter Six when comparing the alignments of minster and non-minster churches, in that it is difficult to draw conclusions from the small number of identified minster churches. The details are shown in Appendix 16 on pages 366-367, but there is little difference between the sites of the minster churches and the remainder, in that they are similarly likely to be built on slopes, and equally likely to be built on east-facing slopes. They are also equally likely to be built next to the big house, but minster churches are more likely to be located in the village and less likely to be isolated from their settlement. Even though the figures in many cases are close, no real conclusions can be drawn from this analysis as the extremely small numbers in some of the categories allow no certainty, and even a small variation in the figures would alter the proportions dramatically and therefore affect the conclusions. Chapter 8 271 LOCAL CHURCH SITES AND BURIALS: CONCLUSIONS It is widely accepted that Christian substitution is the explanation for the churches that are built on, or next to, extant prehistoric remains, and that this may illustrate a wider situation where other churches were also built on such sites, but where the evidence of the earlier use has disappeared. However, even if there are obvious remains next to a church, it is not possible to state with certainty that the church was built on the site because of the remains, or whether it was because the site selection requirements for both the earlier site and the church site were the same, particularly for prominent sites and east-facing slopes. Some writers have felt that the extent of Christianization has been overstated; pointing out that the Church ignored the majority of sacred wells (Rattue 1995, 42; Harte 2008, 93), so the process was anything but universal, and not even widespread. Richard Morris highlights the fact that there are few churches built close to monoliths (Morris 1989, 82), and this also suggests that, like wells, they were not considered important enough to require Christianization, or that Christianization was not as all-embracing as has been suggested. John Blair has pointed out that the process of Christianization appears to be a later one, in that the builders of the early minster churches ignored the Pope’s seventh-century call to incorporate earlier pagan sites (Blair 1991, 91; 2005, 376), and that much of Christianization was a tenth- or eleventh-century programme (Morris 1989, 91-92; Blair 2005, 382), rather than a grand headline process applied as part of the early church-building phase. The thirteen church sites in the Deben valley in Suffolk, the sixteen church sites in Launditch in Norfolk, as well as all the other sites identified as being associated with middle-Saxon pottery, seem to point conclusively to the proposition that these settlements were middle-Saxon in origin. Each of these sites indicates a recurring pattern of settlements fixing their position in the middle-Saxon period either containing, or adjacent to, the site that now contains the church. In some of these cases the settlement moved in the late-Saxon period, but in every case the church site remained where the middle-Saxon settlement had been, which seems to indicate strongly that the church site was part of the earlier settlement. It also confirms a continuity of occupation and a probable middle-Saxon religious use of the site as a graveyard, as it is not possible that all the churches currently built there had their sites chosen at a later date, but, just coincidentally, all in areas that had been occupied in the middle-Saxon period. This reinforces similar links identified in other counties across Chapter 8 272 the country that provide an excellent match between middle-Saxon settlement sites and current church sites. It seems reasonable to suggest that this situation applies to many villages in East Anglia and possibly throughout the country, where a site was reserved for religious purposes, perhaps just as a graveyard, very early in the settlement nucleation process. The identified sites cannot all be special cases, and the number in each of the areas studied implies a similar density elsewhere which has yet to be uncovered. Whilst there are records of corpse, or lych, roads being used to carry the dead to neighbouring settlements for burial until relatively recently, rather than burying them locally, these roads are confined to upland areas such as the fells in Cumbria, Northumbria and North Yorkshire where population was sparse (Hindle 1993, 57-60), rather than in the more densely-populated lowland areas. Many of the local burial sites identified here are associated with east-facing slopes, however gentle; perhaps this is, in part, due to the contemporary view of Judgement Day which meant that the dead needed to be able to sit up and face east on Judgement Day, which would be assisted by an east-facing slope and give them a better eastern view. Local burial also took on a new importance when incorporating the dead into the community meant that the prayers of the living surrounding them could aid their progression to the afterlife, rather than burial in either the separated cemeteries of the earlier pagan period, or the remote early Christian cemeteries in the Missionary stations or at minster churches. Richard Morris suggests that the cemetery in early Christian times fulfilled some of the ritual functions that were eventually served by the church (Morris 1983, 33); so these local sites appear to have been acting as the focus for religious activities, becoming the logical place later to build the estate, or village, church. The new church continued the tradition of the ritual use of the site, as well as local burial and, incidentally, the use of the east-facing slope. The excavated middle-Saxon burial sites in Norfolk and Suffolk which were integrated into the community, but are now completely deserted, such as at Sedgeford in Norfolk and Bloodmoor Hill in Suffolk, and the implied number of such sites where settlements have not moved to reveal them, also suggests that early local burial was taking place in many locations. Perhaps the “endorsement” of local burials “in certain circumstances” referred to by Morris and Blair (Morris 1983, 49–62; Blair 2005, 228– 229), was more widespread than originally thought, and the European examples of village burial sites quoted by Blair (Blair 2005, 228-229), which operated in parallel Chapter 8 273 with central churchyard burials, also applied widely in this country. The fact that, in Norfolk and Suffolk, these burials are in settlements that are close to their minster church indicates at least a tacit approval by the church hierarchy. The local graveyard may have been purely familial in its early stages, but, since it was following contemporary Christian thought in being part of the community, and in the majority of cases, it appears that the selected site was later definitively Christian due to the building of the local church there; it seems to confirm a continuity of purpose. There has to be a reason why the churches that are built directly on sloping sites and those built on levelled platforms on slopes have such a different relationship to the direction of the slope - overwhelmingly east-facing for sloping sites and almost equally distributed in all directions for those on platforms. The direction of the slope was obviously far more important for the builders of churches directly built on slopes. This could be explained if the slope and extended eastern view was important for the selection of the site as an early graveyard, and the subsequent late-Saxon or Norman church, built later on the same site, is merely coincidentally using the same slope that was selected by the middle-Saxon graveyard users; whereas churches built on platformed sites were the first religious use of their site. Churches built on platformed sites were perhaps part of later settlements that had had no earlier graveyard, and by this time, the middle-Saxon desire for an east-facing slope was no longer applicable. This would make the choice of, and particularly the topography of, the site considerably less important. Unfortunately, the current churches built on either sloping sites or platformed sites cannot assist with the dating of the earliest use of the site, as they are rarely earlier than the eleventh or twelfth century, and the only way to test this theory is by the complete excavation of the graveyard, as the earliest burials that may exist on the sloping sites are completely hidden beneath the church building and later interments. The fact that a distant eastern view is available at many sites, particularly at flat ones and even those at the top of slopes in other directions, slightly confuses the issue in relation to the idea of a Judgement-Day-driven reason for the selection of east-facing slopes. If a flat site could satisfy the contemporary requirements, why was a sloping site sought out? Although it might be considered that Christianization is unlikely at such a low social level, the incorporation of additional forces to assist with the passage of the dead into the after-life could easily have been part of the consideration of Chapter 8 274 someone at a local level, perhaps using folklore that had held sway for generations to reinforce the new religion. On balance, this seems to be the most likely reason for the use of an east-facing slope - incorporating an earlier use which had been located there and which had required an eastern view. A case can also be made that, in some situations, the often-noticed church/hall focus may have come about in the opposite way to the accepted order, in that the building of the first church could have been used to validate the lordship by utilizing the existing graveyard to achieve cachet, and the lord possibly further extended this advantage in some cases by building his hall next to the church, rather than vice versa. However, the link between manor house and church is less strong in Norfolk than in the other counties surveyed here. Whether this is due to a greater proportion of freemen-built churches in Norfolk; the greater number of small manors, many of which have been amalgamated or lost; or is due to the wider adoption of local burial where the church was built in the earlier graveyard and away from the manor house, is unclear. The small number of identified minster churches has prevented any real conclusions being drawn about whether the reasons for locating a minster church were different from those used when locating other churches. However, it is most likely that there are multiple influences on church location, so the end result cannot be a “one size fits all” solution. The main problem appears to be establishing which of the churches in this survey fit into each of the church categories. Minster churches are notoriously difficult to identify, as are the so-called ‘field churches’ at the other end of the scale, so identifying any difference in the factors at each church, or even which influences applied to the majority of churches at each particular level in the hierarchy, is not possible. Chapter 8 275 PART FOUR CONCLUSIONS Chapter 9 276 CHAPTER NINE CONCLUSIONS ‘The ultimate answer … is 42’ (Hitch-hiker’s Guide to the Galaxy) East as a focus, whether it was for monuments, burials or buildings, seems to have been with us for several thousand years. Prehistoric burials were aligned eastwards, some even as far back as 60,000 years ago; Mediterranean Bronze-age tomb entrances almost universally faced eastwards; Egyptian Pharaohs were buried facing east as well as countless millions of Christian interments, and it seems probable that Anglo-Saxon Christian cemeteries sought out slopes which faced east. The prehistoric examples can easily be explained as instances of a relationship with sunrise and aspects of light, good and rebirth, whereas Christian burials are probably related to resurrection and the supposed direction of the Second Coming. Whether facing east in Christian thought is related to sunrise intentionally as another representation of Christian substitution, or towards paradise, as shown on the medieval maps, or an age-old link to the rising sun, cannot be determined. This continuity may reflect separate but coincident foci on contemporary religious meaning, but it might also represent a continuation of the consistent prehistoric focus on east. Whatever the reason, it is the continuation of the use of east so widely which is important here. The focus on east is also seen in church buildings; particularly the overall general alignment of churches eastwards, and the apparent selection of east-facing slopes on which to site them, and more specifically in the rebuilding of chancels closer to east in the thirteenth and fourteenth centuries. Taken together, all these aspects of easterly focus provide an almost seamless pattern stretching back millennia which cannot be coincidental. Chapter 9 277 The large-scale survey of medieval rural churches undertaken for this thesis has resulted in four significant main findings. Each of them is being revealed for the first time due to the size of the sample and to its structured nature. Previous surveys have been considerably smaller in sample size, usually limited to churches in a single county or in a small typological group. Subsequent analysis of the survey results, in combination with other landscape surveys of settlements and archaeological finds, has helped to interpret part of the results for which no topographical explanation could be found. The first finding is that many of the previously published assertions about the reasons for the differences between the alignments of individual churches have been shown to be less than accurate. Suggestions that churches are aligned with sunrise on their patronal-saint’s day, or are aligned with sunrise on the specific day that building started, have both been proved to be in error, although a few individual churches will inevitably do so through chance. Some of the studies that have claimed to have proved these assertions to be true can be shown to have used suspect research logic as well as being based on small samples. The second finding is that the results show that there is a statistically significant difference, of approximately ten degrees, in the general alignment of churches between the east and west of the country; which is the same as the difference in the alignment of the two churches shown in the photo in the frontispiece to this thesis, the closer church representing the average alignment in the west of the country and the further church representing average alignment in the east. The countrywide variation in alignment seems to be convincingly linked to the direction of sunrise after particularly early harvests. A direct link cannot be established, but analysing thousands of harvest festival dates in parts of four counties spread across the country, and also within Norfolk, shows that the position of sunrise after the earliest harvests relates very closely (within a degree or two) to the mean values for church alignment in all five of the areas, in which the church alignments are several degrees apart. Other possible influences, such as topography, magnetic effects, the chronology of church building, or even the prevalence of different saints in different areas, have each been shown to be incapable of explaining the spatial variation in medieval church alignment. The third main discovery is the fact that a large proportion of the churches that are built on sloping land are built on slopes that face east – three times as many as Chapter 9 278 might be expected from a random distribution. The importance of this imbalance is reinforced by the fact that it does not occur amongst churches which have been built on artificially levelled platforms on the slope, where the distribution is much closer to a random one. Initially the pattern of bias towards east-facing slopes seemed to be explained best by the process known as Christian substitution – the adoption of sites for Christian worship that had been important to an earlier cult or religion. Suggested by the Pope in the early seventh century, and reinforced by later church pronouncements, substitution was designed to incorporate earlier religious sites in order to ease the transition to Christianity for the former worshipers there. These sites have, in many cases, left little archaeological evidence in the present, except perhaps for an extended eastern horizon which would have given a better eastern view, including that of sunrise. Several writers have suggested that Christian substitution was not as widely practiced as the church hierarchy may have hoped until several centuries after the Conversion; with the omission of many cult-type features, such as most wells and standing stones, and the reticence of the senior members of the Church hierarchy to locate their own churches on such sites, however, the following paragraph may offer a slightly different view of substitution. Lastly, the fourth main finding has led to this thesis arguing for a separate process for the creation of many local religious sites. As opposed to the broadly accepted minster model, where the building of churches, and presumably the selection of their sites, came later and later in the Saxon period as lower and lower levels of the church hierarchy and lordly hierarchy are considered, here it is being suggested that many burial sites were set up during the middle-Saxon period by residents for burial in, or adjacent to, their settlements, which later became the focus for the estate or village church (and possibly the manor house). The process of building the church formalized the position of the earlier graveyard when the church itself was consecrated. The early adoption of sites could have been due to the speed of the Conversion at a lower social level in East Anglia than had previously been suspected, when the incorporation of the dead into the community took on new importance, rather than burial in the remote cemeteries of the earlier periods. It is also possible that the development of a large proportion of these local cemeteries in the mid-Saxon period on east-facing slopes is either following contemporary thought about the location of the Second Coming, or reflecting the drive by the early church hierarchy for Christian substitution by using Chapter 9 279 earlier ritual sites, but again at a far lower level than previously supposed, perhaps with the happy coincidence of using the perceived force of the earlier use on the site as an additional factor to the need to face east on Judgement Day; this takes the aspect of Christian substitution one stage further back in the process. So, rather than Christian substitution being the driver of site-selection when the church was built, it is possible that it was applied when the graveyard sites were being identified. Then the incentive to build the later church on the same site meant that in many cases the fact that the church is sited on an east-facing slope is a coincidence and provides a possible explanation for the considerably greater proportion of churches built on slopes that face east. Whether the use of the east-facing slopes was prehistoric or Saxon, or both, is not really the issue; in either case this selection has to represent a specific choice for the sites in one or both periods, otherwise the numbers of sites on slopes in all directions would balance out. Those churches that are built on levelled platforms on slopes are roughly equally distributed between slopes of all directions and do not reflect the same bias towards east-facing slopes. This could be explained if platformed sites were the first religious use of the site, with no earlier graveyard, by which time perhaps the direction of the slope was no longer important, rather than churches built directly on to the slope being a continuation of the middle-Saxon use of the site as a local graveyard which had originally sought out the east-facing slope. The fact that the churches on sloping or platformed sites are of similar build-dates merely reflects the fact that they were all built during the main period for church building (or rebuilding), rather than reflecting the sites’ length of use for religious purposes. Whether the reasons behind the choices made in earlier times about the selection of one site over another for the building of a church can ever be identified is difficult to say. The ultimate answer can only lie in the heads of the people making the choices at the time; but the early adoption of a burial site in a village, whilst the new Religion was still in its formative infancy in the minds of the peasants, could easily have Christianized a site that may have had an earlier significance. In many cases the local tradition that was engendered by generations of burials on the site in the settlement later fixed the position of the village church. Given the size and structured nature of the sample used here, it seems unlikely that extending the survey would produce results that are very different from those outlined above, or indicate very different conclusions. It is possible that there is another Chapter 9 280 reason, which has not been considered here, explaining why churches are aligned in certain directions and that these vary across the country, but if it is not concerned with sunrise, magnetic changes, topography or climate, then it must have been an arcane reason, and it seems unlikely that such a reason would have been so influential over so much of the country. Therefore, since climate via early harvest dates appears to explain the variation so closely, and provides a plausible justification for the event of building the first church on an estate, then it can reasonably be presented as the answer. The analysis in this thesis, particularly concerning the early graveyards in villages, leaves some interesting questions, but ones which cannot be answered from the results of this survey or from further analysis of the other work discussed here. Firstly, was the local graveyard site fully Christian when it was adopted in villages, or was it merely a pragmatic approach to dealing with their dead? Secondly, if they were Christian, did the need for a local Christian graveyard site soon after the Conversion become a trigger for the change of settlement patterns of the period when nucleation of settlements began or perhaps even earlier, as the early-Saxon shifting settlements began to fix their position in the middle-Saxon period? Thirdly, did the burial ground only come after settlement nucleation? Or could nucleation have been more likely to occur around a settlement that already had a burial ground? Lastly, if the cemetery later became the focus for the building of a church (as part of the idea of the builder gaining the power of the site suggested by John Blair) in how many cases was the Manor House built next to the church as part of a similar process, in the opposite order from that normally accepted? Chapter 9 281 BIBLIOGRAPHY Abbreviations used (-)UP (City) University Press Antiq J Antiquaries Journal Arch J Archaeological Journal ASSAH Anglo-Saxon Studies in Archaeology and History BAR British Archaeological Reports CBA Council for British Archaeology EAA volumes in the “East Anglian Archaeology” series J Hist G Journal of Historical Geography JSAH Journal of the Society of Architectural Historians LALHS Lowestoft Archaeological and Local History Society Med Arch Medieval Archaeology MSRG Medieval Settlement Research Group NAHRG Norfolk Archaeological and Historical Research Group NNAS Norfolk and Norwich Archaeological Society OUCA Oxford University Committee for Archaeology OUDES Oxford University Department for External Studies POAS Proceedings of the Orkney Archaeological Society PSAS Proceedings of the Society of Antiquaries of Scotland PSIAH Proceedings of the Suffolk Institute of Archaeology and History SHARP Sedgeford Archaeological and Historical Research Project SMA Society for Medieval Archaeology 282 Bibliography Beverley Guardian & East Yorkshire Advertiser, Saturdays August – November, 1870-1899 Cumberland Pacquet, Tuesdays August - November, 1870-1873 Cumberland News, Tuesdays August - November, 1874-1899 East Suffolk Gazette, Saturdays August – November, 1870-1899 Lynn News & County Recorder, Saturdays August – November 1880-1899 Somerset County Gazette & Bristol News, Saturdays August – November, 1870-1878 Somerset County Gazette & Devonshire News, Saturdays August – November, 1879-1899 Somerset County Herald, Notes and Queries, April 30th 1927 & Nov 20th 1934 Yarmouth Mercury, Saturdays August – November, 1870-1899 Ecclesiologist, Volumes I – XXIX, 1841-1870, Cambridge Camden Society Abrahamsen, N., 1992 ‘Evidence for Church Orientation by Magnetic Compass in Twelfth Century Denmark’, Archaeometry 34 (2), 293-303 Addy, S.O., 1913, Church and Manor: a study in English economic history, London: George Allen Addyman, P. & Morris, R. (eds.), 1976, The Archaeological Study of Churches, York: CBA Research Report 13 Ali, J. & Cunich, P., 2001, ‘The Orientation of Churches: Some New Evidence’, Antiq J 81, 155-193 Ali, J. & Cunich, P., 2005, ‘The Church East and West: Orienting the Queen Anne Churches 1711-34’, J SAH 64, 56-73 Alibert, D., 1997, Chrétientés Médiévales VIIe – XIe Siècles, Neuilly: Atlande Allcroft, A., 1927, The Circle and the Cross: A study in continuity, Part 1 The Circle, London: MacMillan Allcroft, A., 1930, The Circle and the Cross: A study in continuity, Part 2 The Cross, London: MacMillan Anderson, A., 1898, ‘Some Hertfordshire Churches’, Bygone Hertfordshire, W. Andrews (ed.), 152-186 Anderson, M., 1995, History and Imagery in British Churches, Cambridge: CUP Armstrong, K., 2006, The Great Transformation: The World in the Time of Buddha, Socrates, Confucius and Jeremiah, London: Atlantic Books Arnold, C. & Wardle, P., 1981, ‘Early Medieval Settlement Patterns in England’ Med Arch 25, 145-149 Arnold-Forster, F., 1899, Studies in Church Dedications, Three Volumes, London: Skeffington Ashwin, T. & Davison, A. (eds.), 2005, An Historical Atlas of Norfolk, Third Edition, Chichester: Phillimore Aston, M., 1985, Interpreting the Landscape: Landscape Archaeology and Local History, London: Batsford Aston, M., 2000, Monasteries in the Landscape, Stroud: Tempus Ayres, B., 2001, ‘Excavations in St Martin-at-Palace Church, 1984-5’, Two Medieval Churches in Norfolk, O. Beazley & B. Ayres (eds.), Gressenhall: EAA 96, 65-97 283 Baker, M., 1974, Folklore and Customs of Rural England, Newton Abbot: David & Charles Baldwin-Brown, G., 1903, The Arts in Early England: Ecclesiastical Architecture in England, London: John Murray Barclay, A & Harding, J. (eds.), 1999, Pathways and Ceremonies: The cursus monuments of Britain and Ireland, Oxford: Oxbow Books Barley, M. (ed.), 1976, Plans and Topography of Medieval Towns, York: CBA Barlow, F., 1972, The Feudal Kingdom of England 1042-1216, London: Longman Barlow, F., 1979, The English Church 1000-1066: A history of the later Anglo-Saxon church, second edition, London: Longman Barnes, G., 1997, ‘Church Location in the Norfolk Landscape’, The Annual No 6, Norwich: NAHRG, 23-36 Bassett, S., 1992, ‘Church and Diocese in the West Midlands: the transition form British to Anglo-Saxon control’, Pastoral Care Before the Parish, J. Blair & R. Sharpe (eds.), 13-40 Batcock, N., 1988, ‘The Parish Churches in Norfolk in the 11th and 12th Centuries’, Minsters and Parish Churches: The local church in transition 950-1200, J. Blair (ed.), 179-190 Batcock, N., 1991, The Ruined and disused churches of Norfolk, Gressenhall: EAA 51 Beazley, O., 2001, ‘Excavations in St Martin-at-Palace Church, 1987’, O. Beazley & B. Ayres (eds.), Two Medieval Churches in Norfolk, Gressenhall: EAA 96, 1–63 Bede, (trans. Sherley-Price, L.) 1955, A History of the English Church & People, London: Penguin Bell, R. & Beresford, M., 1987, Wharram Percy Vol III: The Church of St Martin, SMA Monograph No 11 Bell, T., 1998, ‘Churches on Roman Buildings: Christian Associations and Roman Masonry in Anglo-Saxon England’, Med Arch 42, 1–18 Bennett, H., 1937, Life on an English Manor, Cambridge: CUP Beresford, G., 1975, The Medieval Clayland Village: Excavations at Goltho and Barton Blount, London: SMA Monograph No 6 Beresford, M. & Hurst, J., 1971, Deserted Medieval Villages, revised edition 1989, Gloucester: Alan Sutton Beresford, M. & Hurst, J., 1990, Wharram Percy: Deserted Medieval Village, London: Batsford Bettey, L., 1987, Church and Parish: An introduction for local historians, London: Batsford Bevan-Jones, R., 2002, The Ancient Yew: A history of taxus baccata, Macclesfield: Windgather Press Bewley, R., 2003, Pre-historic Settlements, Stroud: Tempus Biddle, M., 1976a, ‘The archaeology of the church: a widening horizon’, The Archaeological Study of Churches, P. Addyman & R. Morris (eds.), 65-71 Biddle, M., 1976b, ‘The evolution of towns: Planned towns before 1066’, Plans and Topography of Medieval Towns, M. Barley (ed.), 19-31 Binns, A., 1989, Dedications of Monastic Houses in England and Wales 1066-1216, Woodbridge: Boydell Press 284 Binns, A., 1995, ‘Pre-Reformation Dedications to St Oswald in England and Scotland: A Gazetteer’, Oswald: Northumbrian King to European Saint, C. Stancliffe & E. Cambridge (eds.), 241-271 Blair, J., 1985, ‘Secular Minster Churches in Domesday Book’, Domesday Book: a Reassessment, P. Sawyer (ed.), 104-142 Blair, J., 1987, ‘Local Churches in Domesday and Before’, Domesday Studies, J. Holt (ed.), 265-278 Blair, J., 1988a, ‘Minster Churches in the Landscape’, Anglo-Saxon Settlements, D. Hooke (ed.), 35-58 Blair, J., 1988b, ‘Introduction: From Minster to Parish Church’, Minsters and Parish Churches: The local church in transition 950-1200, J. Blair (ed.), OUCA Monograph No 17, 1-19 Blair, J., 1991, Early Medieval Surrey; Landholding, Church and Settlement before 1300, Stroud: Sutton Publishing Blair, J., 1992, ‘Anglo-Saxon Minsters: A topographical review’, Pastoral Care Before the Parish, J. Blair & R. Sharpe (eds.), 226-266 Blair, J., 1995, ‘Ecclesiastical organization and pastoral care in Anglo-Saxon England’, Early Medieval Europe 4(2), 193–212 Blair, J., 1996, ‘Churches in the Early English Landscape: social and cultural contexts’, Church Archaeology: Research directions for the future, J. Blair & C. Pyrah (eds.), 6-19 Blair, J., 2005, The Church in Anglo-Saxon Society, Oxford: OUP Blair, J. & Pyrah, C. (eds.), 1996, Church Archaeology: Research directions for the future, York: CBA Research Report 104 Blair, J. & Sharpe, R. (eds.), 1992, Pastoral Care Before the Parish, Leicester: LUP Blinkhorn, P., 1999, ‘Of cabbages and kings: production, trade, and consumption in middle-Saxon England’, Anglo-Saxon Trading Centres, M. Anderton (ed.), Glasgow: Cruithne Press, 4–23 Board of Agriculture, 1813, General View of the Agriculture of the County of Hertfordshire, London: Sherwood & Neely Boddington, A., 1990, ‘Models of Burial, Settlement and Worship: The Final Phase Reviewed’, Anglo-Saxon Cemeteries: A Reappraisal, E. Southworth (ed.), Stroud: Sutton, 177–199 Boddington, A., 1996, ‘Raunds Furnells: The Anglo-Saxon church and churchyard’, Archaeological Report 7, London: English Heritage Bond, C., 1985, ‘Medieval Oxford Villages and their Topography: a Preliminary Discussion’, Medieval Villages: a review of current work, D. Hooke (ed.), 101-122 Bond, F., 1914, Dedications & Patron Saints of English Churches, Oxford: OUP Bond, J., 2000, ‘Landscapes of Monasticism’, Landscape, D. Hooke (ed.), 63-73 Brandwood, G., 2000, ‘Appendix: A Camdenian Roll-Call’, ‘A Church as it should be’: The Cambridge Camden Society and its influence, C. Webster & J. Elliot (eds.), Stamford: Shaun Tyas, 359-454 Branston, B., 1957, The Lost Gods of England, London: Thames & Hudson Braun, H., 1970, Parish Churches: Their Architectural Development in Britain, London: Faber 285 Brine, J., 1990/1991, ‘The Religious Intentions of the Cambridge Camden Society and Their Effect on the Gothic Revival’, Fabrications 2/3, 4-18 Brook, D., 1992, ‘The early Christian Church east and west of Offa’s dyke, The Early Church in Wales and the West, N. Edwards & A. Lane (eds.), 77-90 Brown, A., 1983, ‘Grave Orientation: A further view’, Arch J 140, 322-328 Brown, E. & Taylor, C., 1989, ‘The origins of dispersed settlements: some results from fieldwork in Bedfordshire’, Landscape History 11, 61-82 Brown, P. (ed.), 1984, Domesday Book – Norfolk (2 vols), Chichester: Phillimore Brown, T. & Foard, G., 1998, ‘The Saxon landscape: a regional perspective’, The Archaeology of Landscape: Studies presented to Christopher Taylor, P. Everson and T. Williamson (eds.), 67-93 Burl, A., 1997, Prehistoric Astronomy and Ritual, Aylesbury: Shire Books Burl, A., 2000, The Stone Circles of Britain, Ireland and Brittany, revised edition, New Haven: Yale UP Butler, L., 1976, ‘The evolution of towns planted after 1066’, Plans and Topography of Medieval Towns, M. Barley (ed.), 32-47 Butler, L., 1986, ‘Church Dedications and the Cult of Saints in Anglo-Saxon England’ The Anglo-Saxon Church, L. Butler & R. Morris (eds.), 44-50 Butler, L. & Morris, R., 1986, The Anglo-Saxon Church, London: CBA Research Report 60 Cabot, S. (ed.), 2002, Sedgeford Historical and Archaeological Research Project Interim Report 2002, Sedgeford: SHARP Cabot, S., Davies, G. & Hoggett, R., 2004, ‘Sedgeford: Excavations of a Rural Settlement in Norfolk’, Land, Sea and Home, J. Hines, A. Lane, & M. Redknap (eds.), Leeds: Maney, 313–324 Cambridge, E. & Rollason, D., 1995, ‘Debate: The pastoral organisation of the AngloSaxon Church: A review of the ‘Minster Hypothesis’, Early Medieval Europe 4(1), 87–104 Campbell, B., 1981, ‘Commonfield origins – the regional dimension’, The Origins of Open-field Agriculture, T. Rowley (ed.), 112-129 Campbell, J., 1986, Essays in Anglo-Saxon History, London: Hambledon Carr, R., Tester, A. & Murphy, P., 1988, ‘The Middle-Saxon settlement at Staunch Meadow, Brandon’, Antiquity 62, 371–377 Caruth, J., 1996, ‘Ipswich, Hewlett Packard plc, Whitehouse Industrial Estate (IPS247)’, PSIAH 38(4), 476–479 Carver, M. (ed.), 1992, The Age of Sutton Hoo, Woodbridge: Boydell Press Carver, M. (ed.), 2003, The Cross Goes North: Processes of Conversion AD300-1300, Ipswich: Boydell & Brewer Carver, M., 2005, ‘Sutton Hoo in context’, A seventh-century princely burial ground and its context, M. Carver (ed.), Reports of the Research Committee of the Society of Antiquaries of London 69, 489-504 Catling, C., (News ed.), 2009, ‘A ‘Neolithic cathedral’ for the north of Scotland’, Current Archaeology 236, 6 Cave, C., 1950, ‘The Orientation of Churches’, Antiq J 30, 47-51 Chadwick, O., 1970, The Victorian Church - Part Two 1860-1901, London: Adam & Charles Black Chadwick, O., 1971, The Victorian Church - Part One 1829-1859, London: SCM Press 286 Chamberlain, W., 1809, General View of the Agriculture of the County of Oxfordshire, London: MacMillan Chamberlain, A. & Parker Pearson, M., 2001, Earthly Remains: The History and Science of Preserved Human Bodies, London: British Museum Press Chaney, W., 1960, ‘Paganism to Christianity in Anglo-Saxon England’, Harvard Theological Review 153 (3), 197-217 Chaney, W., 1970, The Cult of Kingship in Anglo-Saxon England: The transition from Paganism to Christianity, Manchester: MUP Chauncy, Sir H., 1700, reprinted 1826, The history and antiquities of Hertfordshire, London: J.M. Mullinger Cheney, C. (ed.), 2000, Handbook of Dates for Students of English History, revised edition, Royal Historical Society Guides and Handbooks No. 4, Cambridge: CUP Church, Dean R., 1892, The Oxford Movement: Twelve years 1833-1845, London: MacMillan Clark, A., Tarling, D. & Noel, M., 1988, ‘Developments in Archaeomagnetic Dating in Britain’, Journal of Archaeological Science 15, 645-667 Clark, R., 1992, ‘The Dedications of Medieval Churches in Derbyshire: their survival and change from the Reformation to the present day’, The Derbyshire Archaeological Journal CXII, 49-61 Clarke B., 1969, Church Builders in the Nineteenth Century, Trowbridge: David & Charles Cook, G., 1961, The English Medieval Parish Church, London: Phoenix House Cooper, J., 2000, ‘Church Dedications in the Colchester Archdeaconry’, Essex Archaeology and History 31, 161-168 Corcos, N., 2001, ‘Churches as Pre-Historic Ritual Monuments: A Review and Phenomenological Perspective from Somerset’, Assemblage 6, sourced from http//:www.assemblage.group.shef.ac.uk/issue6/Corcos_web.html (accessed 4th June 2008) Cornish, V., 1946, The Churchyard Yew and Immortality, London: Frederick Muller Cotton, S., 1980, ‘Domesday Revisited – where were the 11th century churches?’ NAHRG News 21, 11–17 Croom, J., 1988, ‘The fragmentation of the Minster Parochiae of South-East Shropshire’ Minsters and Parish Churches: The local church in transition 950-1200, J. Blair (ed.), 67-82 Cruden, A., 1769, A Complete Concordance to the Holy Scriptures, 3rd edition, London Cumbria County Council, 1998, Church of England Parishes in Cumbria, 3rd edition, Cumbria Archive Service Curl, J., 1995, Victorian Churches, London: Batsford Dallas, C., 1993, Excavations in Thetford by B.K. Davison between 1964 and 1970, Gressenhall: EAA 62 Darby, H., 1952, The Domesday Geography of Eastern England, Cambridge: CUP Darby, H. (ed.), 1976 revised edition, The New Historical Geography of England before 1600, Cambridge: CUP Darby, H., 1977, Domesday England, Cambridge: CUP Dark, K., 2002, Britain and the end of the Roman Empire, Stroud: Tempus 287 Darling, M. & Gurney, D., 1993, Caister-on-Sea Excavations by Charles Green, 1951– 5, Gressenhall: EAA 60. Davey, N., 1964, ‘A Pre-Conquest Church and Baptistery at Potterne’, Wiltshire Archaeological and Natural History Magazine 59, 116–123 Davies, G. & Hoggett, R., 2001, Sedgeford Historical and Archaeological Research Project Interim Report 2001, Sedgeford: SHARP Davies, J. (ed.), 1972, A New Dictionary of Liturgy and Worship, London: SCM Press Davies, R., 1984, ‘Church orientation in Rutland’, Rutland Record 4, 142-143 Davis, J., 2004, BSS Sundial Glossary: A sourcebook of dialling data, Second Edition, Ipswich: British Sundial Society Publications Davison, A., 1983, ‘The Distribution of Medieval Settlement in West Harling’, Norfolk Archaeology 38, 329–336 Davison, A., 1990, The Evolution of Settlement in Three Norfolk Parishes in Southeast Norfolk, Gressenhall: EAA 49 Davison, A., 1995, ‘The Field Archaeology of the Mannington and Wolterton Estates’, Norfolk Archaeology 42, 160–184 Davison, A., 2003, ‘The Archaeology of the Parish of West Acre, Part 1: Field Survey Evidence’, Norfolk Archaeology 44, 202–221 de Selincourt, E., 1939, The Letters of William and Dorothy Wordsworth: The Later Years, Volume 1 1821-1830, Oxford: The Clarendon Press de Selincourt, E. & Darbishire, H. (eds.), 1947, The Poetical Works of William Wordsworth – Volume 4, Oxford: The Clarendon Press Dickens, A., Mortimer, R. & Tipper, J., 2006, ‘The Early Anglo-Saxon Settlement and Cemetery at Bloodmoor Hill, Carlton Colville, Suffolk: A Preliminary Report’, ASSAH 13, 63–79. Dodgshon, R., 1980, The Origin of British Field Systems: An Interpretation, London: Academic Press Doggett, N., 1988, ‘Medieval Church Dedication in Hertfordshire’, Hertfordshire’s Past - Issue 25, 22-30 Doggett, N., 1989, ‘Medieval Church Dedication in Hertfordshire (concluded)’, Hertfordshire’s Past - Issue 26, 9-16 Duffy, E., 1992, The Stripping of the Altars: Traditional Religion in England c1400c1580, New Haven: Yale UP Duffy, E., 2001, The Voices of Morebath: Reformation and Rebellion in an English Village, New Haven: Yale UP Dumville, D., 1992, Liturgy and the Ecclesiastical History of late Anglo-Saxon England, Woodbridge: Boydell Press Duncan, D., 1999, The Calendar: The 5000 year struggle to align the clock and the heavens, London: Fourth Estate Durandus. W., (trans. J. Neale & B. Webb) 1906, The Symbolism of Churches and Church Ornaments, Third Edition, London: Gibbings & Co Dyer, C., 1994, Everyday Life in Medieval England, London: Hambledon Press Dymond, D., 1968, ‘The Suffolk Landscape’, East Anglian Studies, L. Munby (ed.), 17-47 Dymond, D., 1985, The Norfolk Landscape, London: Hodder & Stoughton 288 Dymond, D., 1995, ‘Chapels of ease and the case of Botesdale’, East Anglian Studies: Essays presented to JC Barringer on his retirement, A. Longcroft and R. Joby (eds.), 58-65 Dymond, D., 1999, ‘Churches and Churchyards’, An Historical Atlas of Suffolk, D. Dymond & E. Martin (eds.), 54 &197 Dymond, D., 2005, ‘Medieval and Later Markets’, An Historical Atlas of Norfolk, T. Ashwin & A. Davidson (eds.), 76-77 Dymond, D. & Martin, E. (eds.), 1999, An Historical Atlas of Suffolk, revised edition, Ipswich: Suffolk County Council Eeles, F., 1914, ‘The Orientation of Scottish Churches: Illustrated by an analysis of some examples in Aberdeenshire and Banffshire’, PSAS 48, 169-183 Easton, T., 1999, ‘Ritual Marks on House Timbers’, Weald & Downland Open Air Museum Newsletter, Spring 1999, 22-28 Eaton, T., 2000, Plundering the Past: Roman Stonework in Medieval Britain, Stroud: Tempus Edwards, N., 1992, ‘Identifying the Archaeology of the early church in Wales and Cornwall’, Church Archaeology: Research directions for the future, J. Blair & C. Pyrah (eds.), 49-63 Edwards, N., & Lane, A. (eds.), 1992, The Early church in Wales and the West, Oxford: Oxbow Monograph No 16 Edwards, N., & Lane, A., 1992b, ‘The Archaeology of the early church in Wales: An introduction’, The Early church in Wales and the West, N. Edwards & A. Lane (eds.), 1-12 Ekwall, E., 1989, The Concise Oxford Dictionary of English Place-names, fourth edition, Oxford: OUP Ellison, A., 1983, Medieval Villages in South-East Somerset, Bristol: Western Archaeological Trust Survey No. 6 Evans, G., 1971, The Pattern Under the Plough, London: Faber & Faber Everitt, A., 1986, Continuity and Colonization: The evolution of Kentish settlement, Leicester: LUP Everson, P., Taylor, C. & Dunne, C., 1991, Change and Continuity: Rural Settlement in North West Lincolnshire, London: HMSO Everson, P. & Williamson, T. (eds.), 1998, The Archaeology of Landscape: Studies presented to Christopher Taylor, Manchester: MUP Faith, R., 1997, The English Peasantry and the Growth of Lordship, Leicester: LUP Farmer, D., 1991, The Oxford Dictionary of Saints, Oxford: OUP Faull, M., 1976, ‘The location and relationship of the Sancton Anglo-Saxon cemeteries’, Antiq J 56, 227-233 Faull, M., 1983, ‘Roman and Anglo-Saxon Settlement Patterns in Yorkshire: A computer generated analysis’, Landscape History 5, 21-40 Faull, M., 1984, ‘Late Anglo-Saxon Settlement Patterns in Yorkshire’, Studies in Late Anglo-Saxon Settlement, M. Faull (ed.), 129-142 Faull, M. (ed.), 1984, Studies in Late Anglo-Saxon Settlement, Oxford: OUDES Fernie, E., 1980, The Architecture of the Anglo-Saxons, London: Batsford Fischer, D., 1996, The Great Wave: Price Revolutions and the Rhythm of History, New York: Oxford University Press 289 Foard, G., 1978, ‘Systematic fieldwalking and the investigation of Saxon settlement in Northamptonshire’, World Archaeology 9, 357-374 Foot, S., 1992, ‘Anglo-Saxon Minsters: a review of terminology’, Pastoral Care Before the Parish, J. Blair & R. Sharpe (eds.), 212-225 Foster, S., 2006, Maes Howe and the Heart of Neolithic Orkney, Edinburgh: Historic Scotland Franklin, M., 1984, ‘The identification of Minsters in the Midlands’, Anglo-Norman Studies vii, 69-87 Fraser, J., 1923, ‘Some Antiquities in Harray Parish’, Proceedings of the Orkney Antiquarian Society 1, 31-37 Friar, S., 1996, The Companion to the English Parish Church, Stroud: Chancellor Press Geake, H., 1992, ‘Burial Practice in seventh- and eighth-century England’, The Age of Sutton Hoo, M. Carver (ed.), 83–94 Geake, H., 1997, The Use of Grave Goods in Conversion-Period England, c600-850, Oxford: BAR British Series 261 Geake, H., 2002, ‘Persistent Problems in the Study of Conversion-Period Burials in England’, Burial in Early Medieval England and Wales, S. Lucy and A. Reynolds (eds.), SMA Monograph 17, 144-155 Geake, H., 2003, ‘The Control of Burial Practice in Middle Anglo-Saxon England’, The Cross Goes North: Processes of Conversion in Northern Europe AD300-1300, M. Carver (ed.), 259-270 Gelling, M., 1983, Place-names in the Landscape, London: J.M. Dent Gem, R., 1988, ‘The English parish church in the 11th and early 12th centuries: a great rebuilding’, Minsters and Parish Churches: The local church in transition 950-1200, J. Blair (ed.), 21-30 Gem, R., 1996, ‘Church Buildings: cultural location and meaning’ Church Archaeology: Research directions for the future, J. Blair & C. Pyrah (eds.), 1-6 Gerrard, C. & Aston, M., 2007, The Shapwick Project, Somerset: A Rural Landscape Explored, SMA Monograph No 25 Gilbert Scott, Sir G., 1881, A History of English Church Architecture, London: Simpkin Marshall Gittos, H., 2002, ‘Creating the Sacred: Anglo-Saxon Rites for Consecrating Cemeteries’, Burial in Early Medieval England and Wales, S. Lucy and A. Reynolds (eds.), 195-204 Gooch C., 1813, General View of the Agriculture of the County of Cambridge, London: Sherwood & Neely Green, B. & Rogerson, A., 1978, The Anglo-Saxon Cemetery at Bergh Apton, Norfolk, Gressenhall: EAA 7. Green, B., Rogerson, A. & White, S., 1987, The Anglo-Saxon Cemetery at Morning Thorpe, Norfolk, Gressenhall: EAA 36. Griffith, J., 1908, ‘The origin of Advent and other three-week celebrations’, Nature LXXIX, 36-37 Grinsell, L., 1986, ‘The Christianization of pre-historic and other pagan sites’, Landscape History 8, 7-38 290 Groves, N., 1995, ‘Two Sisters: Two Churches’, East Anglian Studies: Essays presented to J. C. Barringer on his retirement, R. Joby and A. Longcroft (eds.), 108-115 Hadley, D. & Buckberry, J., 2005, ‘Caring for the Dead in Late Anglo-Saxon England’, Pastoral Care in Late Anglo-Saxon England, F. Tinti (ed.), Anglo-Saxon Studies (6), Woodbridge: Boydell, 121-147 Hall, D., 1981, ‘The origin of open-field agriculture: the archaeological fieldwork evidence’, The origins of open-field agriculture, T. Rowley (ed.), 22-38 Hall, T., 2000, Minster Churches in the Dorset Landscape, Oxford: BAR British Series 304 Hamerow, H. 1991, ‘Settlement Mobility and the “Middle-Saxon Shift”: rural settlement and settlement patterns in Anglo-Saxon England’ Anglo-Saxon England 20, 1-17 Hamerow, H., 2002, Early Medieval Settlements: the archaeology of rural communities in Northwest Europe 400-900, Oxford: OUP Harding, J., 1999, ‘Pathways to new realms: cursus monuments and symbolic territories’, Pathways and Ceremonies: The cursus monuments of Britain and Ireland, A. Barclay & J. Harding (eds.), 30-38 Harding, J., 2003, Henge Monuments of the British Isles, Stroud: Tempus Harding, M., 1995, Hypnotising the cat, London: Robson Books Harte, J., 2008, English Holy Wells: a sourcebook, 3 volumes, Loughborough: Heart of Albion Press Harvey, J., 1974, Cathedrals of England & Wales, London: Batsford Harvey, P., 1996, Mappa Mundi: The Hereford World Map, Hereford: Hereford Cathedral Hase, P., 1988, ‘The Mother Churches of Hampshire’, Minsters and Parish Churches: The local church in transition 950-1200, J. Blair (ed.), 45-66 Hawkes, J., 1967, ‘God in the Machine’, Antiquity 41, 174-180 Hawkes, S., 1976 ‘Sunrise dating of death and burial in an Anglo-Saxon cemetery in East Kent’, Archaeologia Cantiana 92, 33-51 Hawkins, G., 1970, Stonehenge Decoded, London: Fontana/Collins Hayfield, C., 1986, ‘Wawne, East Riding of Yorkshire: a case study in settlement morphology’, Landscape History 8, 41-68 Hayslett, H., 1973, Statistics Made Simple, second (revised) edition, London: W H Allen Heilbron, J., 2001, The Sun in the Church: Cathedrals as Solar Observatories, London: Harvard UP Henshall, A. & Ritchie, J., 2001, The Chambered Cairns of the Central Highlands, Edinburgh: EUP Hesse, M., 2003, ‘Domesday settlement in Suffolk’, Landscape History 25, 45-57 Heywood, S., 2005, ‘Round-Towered Churches’, An Historical Atlas of Norfolk, T. Ashwin & A. Davison (eds.), 60-61 Hill, R., 2008, Stonehenge, London: Profile Books Hills, C., 1979, ‘The archaeology of Anglo-Saxon England in the pagan period: a review’, Anglo-Saxon England 8, 297–329 Hills, C., Penn, K. & Rickett, R., 1984, The Anglo-Saxon Cemetery at Spong Hill, North Elmham. Part III: Catalogue of Inhumations, Gressenhall: EAA 21 291 Hindle, B., 1993, Roads, Tracks and their Interpretation, London: Batsford Hinton, I., 2001, ‘The alignment of churches in the Waveney Valley’, The Annual No. 10, Norwich: NAHRG, 3-16 Hinton, I., 2003, A regional study of the alignment of medieval rural churches, unpublished MA dissertation, University of East Anglia Hinton, I., 2004, ‘Do Chancels Weep: Does the often noticed difference between the alignments of nave and chancel actually mean anything?’, Church Archaeology 5&6, 42-54 Hinton, I., 2006, ‘The Rural Church in the Landscape: further considerations of aspects of alignment and location’, Antiq J 86, 206-226 Hoare, P. & Sweet, C., 2000, ‘The Orientation of Early Medieval Churches in England’, J Hist G 26 (2), 162-173 Hodges, R., 1989, The Anglo-Saxon Achievement: archaeology and the beginnings of society, London: Duckworth Hoggett, R., 2001, The Origin and Early Development of Sedgeford, Norfolk, unpublished M.A. dissertation, University of Bristol Hoggett, R., 2007, Changing Beliefs: An Archaeology of the East Anglian Conversion, unpublished PhD thesis, University of East Anglia Holt J., 1813, reprinted 1969, General View of the Agriculture of the County of Lancaster, Newton Abbot: David & Charles Holt, J. (ed.), 1987, Domesday Studies, Woodbridge: Boydell Homans, C., 1970, English Villagers of the Thirteenth Century, New York: Harper & Row Hooke, D. (ed.), 1985, Medieval Villages: a review of current work, OUCA Monograph No 5 Hooke, D. (ed.), 1987, Anglo-Saxon Settlements, Oxford: Basil Blackwell Hooke, D. (ed.), 2000, Landscape: the richest historical record, Society for Landscape Studies supplementary series 1 Hoskin, M., 2001, Tombs, Temples and their Orientations: A new perspective on Mediterranean pre-history, Bognor Regis: Oscarina Books Hoskins, W., 1955, The making of the English Landscape, revised edition 1974, London: Book Club Associates Insoll, T., 2004, Archaeology, Ritual, Religion, London: Routledge Jay, E., 1993, The Literary Associations of Rydal Church, Rydal: Armitt Trust Jennings, S., 1981, Eighteen Centuries of Pottery from Norwich, Norwich: EAA 13 John, C., 2001, The Saints of Cornwall: 1500 Years of a Christian Landscape, Padstow: Tab House Johnson, A., 2008, Solving Stonehenge: the new key to an ancient enigma, London: Thames & Hudson Johnson, M., 2007, Ideas of Landscape, Oxford: Blackwell Johnson, S., 1983, Burgh Castle, Excavations by Charles Green 1958–61, Gressenhall: EAA 20. Johnson, W., 1912, Byways in British Archaeology, Cambridge: CUP Jones, A., 2002, A Thousand Years of the English Parish: medieval patterns and modern interpretations, London: Cassell Paperbacks Jones, F., 1998, Holy Wells in Wales, University of Wales Press 292 Jones, G., 1986, ‘Holy wells and the cult of St Helen’, Landscape History 8, 59-76 Jones, G., 2003, Saints of Europe: studies towards a survey of cults and culture, Donnington: Shaun Tyas Jones, G., 2006 ‘TASC’ website at http://www.le.ac.uk/elh/grj1/database/data.html, last accessed 23rd Aug 2008 Jones, G., 2007, Saints in the Landscape, Stroud: Tempus Jones, G. J., 1961, ‘Settlement patterns in Anglo-Saxon England’, Antiquity XXXV, 221-232 Jones, G. J., 1989, ‘Multiple Estates and early Settlement’, English Medieval Settlement, P. Sawyer (ed.), 9-34 Jones, R. & Page M., 2006, Medieval Villages in an English landscape: Beginnings and Ends, Macclesfield: Windgather Press Kaler, J., 1996, The Ever-Changing Sky: A Guide to the Celestial Sphere, Cambridge: CUP Keble, J., 1888, The works of that Learned and Judicious Divine Mr. Richard Hooker, seventh edition, Oxford: OUP Kershaw, I., 1973, ‘The Great Famine and Agrarian Crisis in England 1315-1322’, Past & Present 59, 3-50 Kett, J., 1973, Tha’s a Rum’un Bor: Norfolk Verse, Norwich: Bertram Books Knox, E., 1933, The Tractarian Movement 1833-1845: A study of the Oxford Movement, London: Putnam Kosminsky, E., 1956, Studies in the Agrarian History of England in the Thirteenth Century, translated R. Kisch, Oxford: Basil Blackwell Kussmaul, A., 1981, Servants in husbandry in early modern England, Cambridge: CUP Laing, L., 1977, The Archaeology of Late Celtic Britain and Ireland, London: Book Club Associates Lamb, H., 1977, Climate, Present, Past and Future: Volume 2 Climatic history and the future, London: Methuen Lancaster-Brown, P., 1976, Megaliths, Myths and Men: An introduction to astroarchaeology, Poole: Blandford Press Laverton, S., 2001, Shotley Peninsular: The making of a unique Suffolk landscape, Stroud: Tempus Lawrie, W., 1859, History of Free Masonry and The Grand Lodge of Scotland, Edinburgh Lawson, A., 1983, The Archaeology of Witton, Gressenhall: EAA 18 Levison, W., 1956, England and the Continent in the Eighth Century, Oxford: OUP Lewis, C., Mitchell-Fox, P. & Dyer, C., 2001, Village, Hamlet and Field: Changing Medieval Settlements in Medieval England, revised edition, Macclesfield: Windgather Press Linnel, Rev. C., 1962, Norfolk church dedications, York: St Anthony’s Hall Lloyd, T., Orbach, J. & Scourfield, R., 2004, The Buildings of Wales: Pembrokeshire, New Haven: Yale UP Lockyer, Sir N., 1909, Stonehenge and Other British Stone Monuments: Astronomically Considered, London: MacMillan Longcroft, A. & Joby, R. (eds.), 1995, East Anglian Studies: Essays presented to JC Barringer on his retirement, Norwich: University of East Anglia 293 Loveday, R., 1999, ‘Dorchester on Thames – ritual complex or ritual landscape’, Pathways and Ceremonies: The cursus monuments of Britain and Ireland, A. Barclay & J. Harding (eds.), 39-48 Loveday, R., 2006, Inscribed across the Landscape: The Cursus Enigma, Stroud: Tempus Lucas, H., 1930, ‘The Great European Famine of 1315-1317’, Speculum 15, 57-66 Lucy, S., 2000, The Anglo-Saxon Way of Death, Stroud: Tempus Lucy, S. & Reynolds, A. (eds.), 2002, Burial in Early Medieval England and Wales, SMA Monograph No 17 Lucy, S. & Reynolds, A., 2002, ‘Burial in Early Medieval England and Wales: Past, Present and Future’, S. Lucy & A. Reynolds (eds.), Burial in Early Medieval England and Wales, 1-23 Margeson, S, Ayres, B, & Heywood, S., 1996, A Festival of Norfolk Archaeology, Norwich: NNAS Martin, E., 2001, ‘Rural settlement patterns in medieval Suffolk’, MSRG: Annual Report 15, 5-7 Mayr-Harting, H., 1977, The Coming of Christianity to Anglo-Saxon England, London: Batsford Meaney, A., 1964, A Gazetteer of Early Anglo-Saxon Burial Sites, London: Allen and Unwin. Meaney, A., 2003, ‘Anglo-Saxon Pagan and Christian Attitudes to the Dead’, The Cross Goes North: Processes of Conversion in Northern Europe AD3001300, M. Carver (ed.), 229-242 Merrill, R., McElhinny, W. & McFadden, P., 1996, The Magnetic Field of the Earth: Paleomagnetism, the Core and the Deep Mantle, San Diego: Academic Press Messent, C., 1934, The Monastic Remains of Norfolk & Suffolk, Norwich: H W Hunt Michel, J., 1969, The View Over Atlantis, London: Sago Press Mills, A., 1991, A Dictionary of English Place-names, Oxford: OUP Morris, R., 1979, Cathedrals and Abbeys of England & Wales, London: J M Dent Morris, R., 1983, The Church in British Archaeology, York: CBA Research Report 47 Morris, R., 1985, ‘The church in the countryside’, Medieval Villages: A review of current work, D. Hooke (ed.), 48-60 Morris, R., 1989, Churches in the Landscape, London: Phoenix Press Mortimer, R. & Tipper, J., 1998, ‘Excavation of early/middle Saxon settlement at Bloodmoor Hill, Carlton Colville, Suffolk’, Annual Report 31, LALHS, 1416 Morton, A., 1992, ‘Burial in middle Saxon Southampton’, Death in Towns: Urban Responses to the Dying and the Dead 100-1600, S. Bassett (ed.), Leicester: LUP, 68-77 Moshenska, G. (ed.), 2004, Annual Report, Sedgeford Historical and Archaeological Research Project Muir, R., 2007, How to Read a Village, London: Ebury Press Muirden J., 2003-2005 (published 2005), ‘Crooked Churches and Saintly Sunrises’, Church Archaeology vols 7-9, 33-43 Muncey, R., 1930, A History of the Consecration of Churches and Churchyards, Cambridge: W. Heffer 294 Myres, J. & Green, B., 1973, The Anglo-Saxon Cemeteries of Caistor-by-Norwich and Markshall, Norfolk, London: Society of Antiquaries. Mytum, H., 2000, Recording and Analysing Graveyards, York: CBA Neale, J., 1841a, A Few Words to Church Builders, Cambridge Camden Society Neale, J., 1841b, A Few Words to Churchwardens on Churches and Church Ornaments: No. 1, Suited to Country Parishes, eighth edition, Cambridge Camden Society Neale, J. & Webb, B., 1906, ‘Introductory Essay: Sacramentality - A Principle of Ecclesiastical Design’, The Symbolism of Churches and Church Ornaments, William Durand, third edition, London: Gibbings & Co Newman, J., 1908, Apologia Pro Vita Sua: Being a History of His Religious Opinions, London: Longmans Newman, J., 1969, The Buildings of England: North East & East Kent, London: Penguin Newman, J., 1992, ‘The Late Roman and Anglo-Saxon Settlement Pattern in the Sandlings of Suffolk’, The Age of Sutton Hoo, M. Carver (ed.), 25-38 Newman, J., 2001, ‘The landscape of dispersed settlement – change and growth in south-east Suffolk’, MSRG Annual Report 15 - 2000, 7-8 Newman, J., 2005, ‘Survey in the Deben Valley’, A seventh-century princely burial ground and its context, M. Carver (ed.), 477-488 Newman, R., 1984, ‘The Problems of Rural Settlement in Northern Cumbria in the PreConquest Period’, Studies in Late Anglo-Saxon Settlement, M. Faull (ed.), 155-175 Northeast, P., ‘Moving the Signposts: Changes in the Dedication of Suffolk Churches after the Reformation’, East Anglian Studies: Essays presented to J.C. Barringer on his retirement, R. Joby and A. Longcroft (eds.), 201-205 Nuttall, A., 1974, Common Sky: Philosophy and the Literary Imagination, Berkeley: University of California Press O’Carragain, T., 2003, ‘A Landscape Converted: Archaeology and Early Church Organisation on Iveragh and Dingle, Ireland’, The Cross Goes North: Processes of Conversion in Northern Europe AD300-1300, M. Carver (ed.), 127-152 O’Sullivan, D., 1984, ‘Pre-Conquest Settlement Patterns in Cumbria’, Studies in Late Anglo-Saxon Settlement, M. Faull (ed.), 143-154 Orme, N., 1996, English Church Dedications: with a survey of Cornwall and Devon, Exeter: EUP Oswin, J., 2009, ‘Stones and Solstices on Scotland’s Remotest Island’, Current Archaeology 227, 30-37 Oosthuizen, S., 2002, ‘Medieval Greens and Moats in the Central Province: Evidence from the Bourne Valley, Cambridgeshire’, Landscape History 24, 73-88 Owen, D., 1971, Church and Society in Medieval Lincolnshire, Lincoln: History of Lincolnshire Committee Owen, G., 1981, Rites and Religions of the Anglo-Saxons, Newton Abbott: David & Charles Park, R., 1988, Geological Structures and Moving Plates, London: Blackie & Son 295 Parker Pearson, M., 1993a, ‘The Powerful Dead: Archaeological Relationships between the Living and the Dead’, Cambridge Archaeological Journal 3, 203–229 Parker Pearson, M., 1993b, Bronze-Age Britain, London: Batsford/English Heritage Parker Pearson, M., 1999, The Archaeology of Death and Burial, Stroud: Sutton Publishing Parker Pearson, M. & Richards, C., 1994, Architecture and Order: Approaches to social space, London: Routledge Parsons, D., 1986, Liturgy and Architecture in the Middle Ages, Leicester: LUP Patton, M., Nash, G., Finch, O. & Rodwell, W., 2002, ‘La Hogue Bie’, Current Archaeology 183, 125-129 Pearce, S., 1985, ‘The early church in the landscape: the evidence from North Devon’, Arch J 142, 255-275 Pearce, S., 2003, ‘Saintly Cults in South-Western Britain: A Review’ Saints of Europe: Studies Towards a Survey of Cults and Culture, G. Jones (ed.), 261-279 Penn, K., 1996, ‘The Early Church in Norfolk: Some Aspects’ A Festival of Norfolk Archaeology, S. Margeson, B Ayres & S. Heywood (eds.), 40–46. Penn, K., 2000, Excavations on the Norwich Southern Bypass, 1989–91. Part II: The Anglo-Saxon Cemetery at Harford Farm, Caistor St Edmund, Norfolk. Gressenhall: EAA 92. Pestell, T., 2004, Landscapes of Monastic Foundation: the establishment of religious houses in East Anglia c.650-1200, Woodbridge: Boydell Press Petrie, G., 1890, ‘Notice of Brochs and Large Round Towers of Orkney’, Archaeologia Scotica, 71-94 Petts, D., 2002, ‘Cemeteries and boundaries in western Britain’, Burial in Early Medieval England and Wales, S. Lucy and A. Reynolds (eds.), 24-46 Pevsner, N., 1951, The Buildings of England: Cornwall, London: Penguin Pevsner, N., 1958a, The Buildings of England: North Somerset, London: Penguin Pevsner, N., 1958b, The Buildings of England: Shropshire, London: Penguin Pevsner, N., 1967, The Buildings of England: Cumberland and Westmorland, London: Penguin Pevsner, N., 1969, The Buildings of England: North Lancashire, London: Penguin Pevsner, N., 1970, The Buildings of England: Cambridgeshire, London: Penguin Pevsner, N., 1988, The Buildings of England: Bedfordshire Huntingdon and Peterborough, London: Penguin Pevsner, N., 1991, The Buildings of England: Suffolk, London: Penguin Pevsner, N. & Harris J., (2nd edition revised by N Antrim), 1998, The Buildings of England: Lincolnshire, London: Penguin Pevsner, N. & Lloyd D., 1967, The Buildings of England: Hampshire and the Isle of Wight, London: Penguin Pevsner, N. & Nairn, I., 1999, The Buildings of England: Sussex, London: Penguin Pevsner, N. & Neave, D., 1999, Buildings of England: York and the East Riding of Yorkshire, London: Penguin Pevsner, N. & Wilson W., 2002a, The Buildings of England: Norfolk, Northwest & South, New Haven: Yale UP 296 Pevsner, N. & Wilson, W., 2002b, The Buildings of England: N.E. Norfolk & Norwich, New Haven: Yale UP Pickford, C., 1994, Bedfordshire Churches in the Nineteenth Century: Part I Parishes A-G, Bedfordshire Historical Records Society Vol 73 Pickford, C., 1998, Bedfordshire Churches in the Nineteenth Century: Part II Parishes H-R, Bedfordshire Historical Records Society Vol 77 Pickford, C., 2000, Bedfordshire Churches in the Nineteenth Century: Part III Parishes S-Z, Bedfordshire Historical Records Society Vol 79 Pickford, C., 2001, Bedfordshire Churches in the Nineteenth Century: Part IV Appendices and Index, Bedfordshire Historical Records Society Vol 80 Pitts, M., 2000, Hengeworld, London: Arrow Books Plymley, J., 1813, General View of the Agriculture of the County of Shropshire, London: Sherwood & Neely Pollard, J. & Reynolds A., 2002, Avebury: The biography of a landscape, Stroud: Tempus Pope, R., 2008, ‘Roundhouses: 3,000 years of prehistoric design’, Current Archaeology 222, 14-21 Pounds, N., 1994, ‘The Churches of Canterbury’, Proceedings of the 140th Summer Meeting of the Royal Archaeological Institute (Supplement to the Archaeological Journal Vol 151, N. Pounds (ed.), 27-30 Prescott, A., 2008, ‘Masonic Papers’, Pietre-Stones Review of Freemasonry, sourced from http://www.freemasons-freemasonry.com/ prescott10 .html (accessed 28th April 2008) Preston-Jones, A., 1992, ‘Decoding Cornish Churchyards’, The Early Church in Wales and the West, N. Edwards & A. Lane (eds.), 105-124 Prout, D., 1989, ‘‘The Oxford Society for Promoting the Study of Gothic Architecture’ and ‘The Oxford Architectural Society’, 1839-1850’, Oxoniensia liv, 379391 Pryor, F., French, C., Crowther, D., Gurney, D., Simpson, G. & Taylor, M., 1985, The Fenland Project 1: Archaeology and Environment in the Lower Welland Valley, Cambridge: EAA 27 Pugin, A. W., 1853, The True Principles of Pointed or Christian Architecture, London: Henry Bohn Rackham, O., 1986, The History of the Countryside, London; J M Dent Rahtz, P., 1978, ‘Grave Orientation’, Arch J 135, 1-14 Rattue, J., 1995, The Living Stream: Holy wells in historical context, Woodbridge: Boydell Ravensdale, J., 1974, Liable to Floods: Village Landscape on the edge of the Fens, AD 450 - 1850, Cambridge: CUP Rees, E., 2001, Celtic Saints in their Landscape, Stroud: Sutton Publishing Reynolds, A., 1999, Later Anglo-Saxon England: Life and landscape, Stroud: Tempus Reynolds, A., 2002, ‘Burials, boundaries and charters in Anglo-Saxon England: a reassessment’, Burial in Early Medieval England and Wales, S. Lucy and A. Reynolds (eds.), 171-194 Reynolds, A., 2003, ‘Boundaries and settlements in later sixth- to eleventh-century England’, Anglo-Saxon Studies in Archaeology and History 12, 98-136 297 Ridgard, J., 1999, ‘The structure of a Medieval Manor’, An Historical Atlas of Suffolk, D. Dymond & E. Martin (eds.), 86-87 Rigold, S., 1962, ‘The Anglian Cathedral of North Elmham, Norfolk’, Medieval Archaeology 6, 67–108 Roberts, B., 1977, Rural Settlement in Britain, London: Hutchinson Roberts, B., 1982, ‘The Anatomy of a Village: Observation and Extrapolation’, Landscape History 4, 11-20 Roberts, B., 1992, ‘Dating villages: theory and practice’, Landscape History 14, 19-30 Roberts, B. & Wrathmell S., 1998, ‘Dispersed settlement in England: A national view’, The Archaeology of Landscape: Studies Presented to Christopher Taylor, P. Everson & T. Williamson (eds.), 95-116 Roberts, B. & Wrathmell, S., 2002, Region and Place: An Atlas of Rural Settlement in England, London: English Heritage Rodwell, W., 1981, Archaeology of the English Church, London: Batsford Rodwell, W., 1984 ‘Churches in the Landscape: Some Aspects of Topography and Planning’, Studies in Late Anglo-Saxon Settlement, M. Faull (ed.), 1-23 Rodwell, W., 1986, ‘Anglo-Saxon Church Building: Aspects of design and construction’, The Anglo-Saxon Church, L. Butler & R. Morris (eds.), 156175 Rodwell, W., 1989, Archaeology of Religious Places: Churches and Cemeteries in Britain, revised edition, Philadelphia: UPP Rodwell, W., 2001, The Archaeology of Wells Cathedral: excavations and structural studies 1978-1993, London: English Heritage Rodwell, W. & Rodwell, K., 1973, ‘Excavations at Rivenhall Church, Essex: An interim report’, Antiq Jl LIII (2), 219-231 Rodwell, W. & Rodwell, K., 1985, Rivenhall: investigations of a villa, church and village, 1950-1977, York: CBA Research Report No 55 Rogerson, A., 1995, A Late Neolithic, Saxon and Medieval Site at Middle Harling, Norfolk, Gressenhall: EAA 74 Rogerson, A., 2003, ‘Six Middle Anglo-Saxon Sites in West Norfolk’, Markets in Early Medieval Europe: Trade and ‘Productive’ Sites 650-850, T. Pestell & K. Ulmschneider (eds.), Macclesfield: Windgather Press, 110-121 Rogerson, A., 2005, ‘Middle Saxon Norfolk (c.AD650–850)’, An Historical Atlas of Norfolk, T. Ashwin & A. Davison (eds.), 32–33 Rogerson, A., Ashley, S., Williams, P. & Harris, A., 1987, Three Norman Churches in Norfolk, Gressenhall: EAA 32 Rogerson, A., Davison, A., Pritchard, D. & Silvester, R., 1997, Barton Bendish and Caldecote: fieldwork in south-west Norfolk, Gressenhall: EAA 80 Ross, A., 1974, Pagan Celtic Britain: Studies in Iconography and Tradition, London: Sphere Rowley, T., 1972, The Shropshire Landscape, London: Hodder & Stoughton Rowley, T., 1978, Villages in the Landscape, 1994 edition, London: Orion Books Rowley, T. (ed.), 1981, The Origins of Open-field Agriculture, London: Croom Helm Ruggles, C., 1999, Astronomy in Prehistoric Britain and Ireland, New Haven: Yale UP Rumble, A. (ed.), 1986, Domesday Book: Suffolk, 2 volumes, London: Phillimore Rushton, N., 1999, ‘Parochialization and patterns of patronage in 11th-century Sussex’, Sussex Archaeological Collections 137, 133-152 298 Ryder, P., 1993, Medieval Churches of West Yorkshire, West Yorkshire Archaeology Salmon, J., 1981, Saints in Suffolk Churches, Ipswich: Suffolk Historic Churches Trust Salmon, M., 1991, A History of St John’s Church, Newton Arlosh, privately published Salzman, L., 1992, Building in England down to 1540: A documentary history, revised edition, Oxford: Clarendon Press Sawyer, P., 1976, Medieval Settlement: Continuity and Change, London: Edward Arnold Sawyer, P. (ed.), 1985, Domesday Book: a Re-assessment, London: Edward Arnold Scarfe, N., 1986, Suffolk in the Middle Ages, Woodbridge: Boydell Scarfe, N., 1987, The Suffolk Landscape, Bury St Edmunds: Alastair Press. Schlesinger, A., Walls C., Kissock, J., Lovegrove, C., Pollard, K. & Wright, N., 1996, ‘An Early Church and Medieval Farmstead Site: Excavations at Llanelen, Gower’, Arch J 153, 104-147 Scott, Sir G.G., 1881, An Essay on the History of English Church Architecture, London: Simpkin Marshall & Co Searle, S., 1974, ‘The church points the way’, New Scientist, Jan 3rd 1974 and subsequent correspondence Shepherd Popescu, E., 2009, Norwich Castle: Excavations and Historical Survey 1987– 1998, Gressenhall: EAA 132 Sherwood, S. & Pevsner, N., 1974, The Buildings of England: Oxfordshire, London: Penguin Shore, T., 1886, ‘Orientation of Churches in Hampshire’, Walford’s Antiquarian 10, 95-98 Short, E., 1955, The House of God – A History of Religious Architecture, London: Eyre & Spottiswood Silvester, R., 1988, The Fenland Project Number 3: Marshland and the Nar Valley, Norfolk, Gressenhall: EAA 45 Somerville, H.B., 1923, ‘Instances of Orientation in pre-historic monuments of the British Isles’, Archaeologia 73, 193-224 Somerville, H.B., 1927, ‘Orientation’, Antiquity 1, 31-41 Stamp, G. (ed.), 1995, Personal and Professional Recollections: The autobiography of the Victorian architect Sir George Gilbert Scott, Stamford: Paul Watkins Stancliffe, C. & Cambridge, E. (eds.), 1995, Oswald: Northumbrian King to European Saint, Stamford: Paul Watkins Stenton, Sir F., 1971, Anglo-Saxon England, 3rd edition, Oxford: Clarendon Press Stocker, D., 1990, ‘Rubbish recycled: a study of the reuse of stone in Lincolnshire’, Stone: Quarrying and Building in England, AD43-1525, D. Parsons (ed.), Chichester: Phillimore, 83-101 Stocker, D. & Everson, P., 2006, Summoning St Michael: Early Romanesque Towers in Lincolnshire, Oxford: Oxbow Books Stratton J. & Houghton-Brown J., 1978, Agricultural Records: AD 220 -1977, London: John Baker Strzygowski, J., 1928, Early Church Art in Northern Europe: With special reference to timber construction and decoration, London: Batsford Surman, R., 2008, Secret Churches: Ecclesiastical Gems from Britain & Ireland, London: HarperCollins 299 Tatton-Brown, T., 1987, ‘The Anglo-Saxon Towns of Kent’, Anglo-Saxon Settlements, D. Hooke (ed.), 213-232 Tatton-Brown, T., 1988, ‘The Churches of the Canterbury Diocese in the 11th Century’, Minsters and Parish Churches: The local church in transition 950-1200, J. Blair (ed.), 105-118 Taylor, A., 2001, Burial Practice in Early England, Stroud: Tempus Taylor, C., 1977, ‘Polyfocal Settlement and the English Village’, Medieval Archaeology 21, 189-193 Taylor, C., 1983, Village and Farmstead: A History of Rural settlement in England, London: Book Club Associates Taylor, C., 1992, ‘Medieval rural settlement: changing perceptions’, Landscape History 14, 5-18 Taylor, C., 2002, ‘Nucleated settlement: a view from the frontier’, Landscape History 24, 53-72 Taylor, H.M., 1978, Anglo-Saxon Architecture- Vol 3, Cambridge: CUP Taylor, H.M. & Taylor, J., 1965, Anglo-Saxon Architecture - Vols 1 & 2, Cambridge: CUP Taylor, T., 1999, Time Team 1999: The site reports, London: Channel Four Television Thacker, A. & Sharpe, R. (eds.), 2002, Local Saints and Local Churches in the Early Medieval West, Oxford: OUP Thirsk, J., 1964, ‘The Common Fields’, Past & Present 29, 3-39 Thom, A., 1967, Megalithic Sites in Britain, Oxford: OUP Thom, A. & Thom, A.S., 1971, ‘The astronomical significance of the large Carnac menhirs’, Journal for the History of Astronomy 2, 147-160 Thomas, A., 1971, The early Christian architecture of north Britain, Oxford: OUP Thomas, C., 1986, ‘Recognizing Christian Origins: an archaeological and historical dilemma’, The Anglo-Saxon Church L. Butler & R. Morris (eds.), York: CBA Research Report 60, 121-125 Thomas, C., 1989, ‘Christians, Chapels, Churches and Charters: Proto-parochial provisions for the pious in a peninsular (Lands End)’, Landscape History 11, 19-26 Thomas, G., 2005, ‘Bishopstone: in the shadow of Rookery Hill’, Current Archaeology 196, 184-190 Thompson, A., 1913, The Ground Plan of the English Parish Church, Cambridge: CUP Thompson, V., 2004, Dying and Death in Later Anglo-Saxon England, Woodbridge: Boydell. Tilley, C., 1994, A Phenomenology of Landscape: Places, Paths and Monuments, London: Berg Tooley, B., 2002, North Cove & Barnby: A History of Two Villages, privately published Trubshaw, B., 1989 ‘Church Orientation’, Mercian Mysteries No 5, 17-31 Turner, S., 2003, ‘Making a Christian Landscape: Early Medieval Cornwall’, The Cross Goes North: Processes of Conversion in Northern Europe AD3001300, M. Carver (ed.), 171-194 Turner, S., 2006, Making a Christian Landscape: The countryside in early medieval Cornwall, Devon and Wessex, Exeter UP Tyrell-Green, E., 1929, ‘The Ecclesiology of Anglesey’, Y Cymmrodor XL, 43-118 300 Valk, H., 2003, ‘Christianisation in Estonia: A Process of Dual-Faith and Syncretism’, The Cross Goes North: Processes of Conversion in Northern Europe AD300-1300, M. Carver (ed.), 571-580 Vancouver C., 1813, Reprinted 1969, General View of the Agriculture of the County of Devon, Newton Abbot: David & Charles Veasey, S., 1986, ‘Discussions: church orientation in Rutland’, Rutland Record 6, 215216 Verey, D., 1970a, The Buildings of England: Gloucestershire 1: The Cotswolds, London: Penguin Verey, D., 1970b, The Buildings of England: Gloucestershire 2: The Vale and the Forest of Dean, London: Penguin Vinogradof, Sir P., 1932, The Growth of the Manor, London: George Allen Wall J., 2006, ‘Church Orientation’, Bulletin of the British Sundial Society 18 (i), 16-17 Wade-Martins, P., 1980a, Excavations in North Elmham Park 1967-72, Gressenhall: EAA 9 Wade-Martins, P., 1980b, Fieldwork and Excavations on Village Sites in Launditch Hundred, Norfolk, Gressenhall: EAA 10 Warner, P., 1986, ‘Shared churchyards, freemen church builders and development of parishes in eleventh century East Anglia’, Landscape History 8, 39-54 Warner, P., 1991, Greens, Commons and Clayland Colonization: The Origins and Development of Green-side Settlement in East Suffolk, Leicester: LUP Warner, P., 1996, The Origins of Suffolk, Manchester: MUP Watkins, A., 1925, reprinted 1974, The Old Straight Track: Its Mounds, Beacons, Moats, Sites and Mark Stones, London: Abacus Webb, S., 2006, Suffolk Church Dedications and the Landscape, unpublished Diploma dissertation, University of East Anglia Webster, C. (ed.), 2003, ‘Temples … worthy of His presence’: the early publications of the Cambridge Camden Society: the complete texts of eight important pamphlets published between 1839 and 1843 with a critical analysis, Reading: Spire Books Webster, C. & Elliot, J. (eds.), 2000, ‘A Church as it Should be’: The Cambridge Camden Society and its influence, Stamford: Shaun Tyas Welch, M., 1985, ‘Rural settlement in early and middle Anglo-Saxon periods’, Landscape History 7, 13-26 Welch, M., 1992, Anglo-Saxon England, London: Batsford/English Heritage Wessex Archaeology, 2007, ‘Speke Keeill, Mount Murray Hotel, Isle of Man: Archaeological Evaluation and Assessment of Results’, sourced from: http://scribd.com/doc/12247576/speke-keeill-isle-of-man/ (accessed 26th Oct 2009) West, S., 1998, A Corpus of Anglo-Saxon Material From Suffolk, Ipswich: EAA 84. West, S., Scarfe, N. and Cramp, R., 1984, ‘Iken, St Botolph, and the Coming of East Anglian Christianity’, PSIAH 35(4), 279–301 Whiffen, M., 1947, Stuart and Georgian Churches: The Architecture of the Church of England outside London 1603-1837, London: Batsford White, J., 1962, The Cambridge Movement: The ecclesiologists and the gothic revival, Cambridge: CUP White, W., 1845, reprinted 1969, History, Gazetteer and Directory of Norfolk, Newton Abbot: David & Charles 301 Wilcox, W., 2002, ‘Grave Alignment at Sedgeford’, SHARP Interim Report 2002, S. Cabot (ed.), 36-39 Williams, H., 1998, ‘Monuments and the past in early Anglo-Saxon England’, World Archaeology 30, 90-108 Williamson, T., 1988, ‘Explaining Regional Landscapes: Woodland and Champion in Southern and Eastern England’, Landscape History 10, 5-14 Williamson, T., 1993, The Origins of Norfolk, Manchester: MUP Williamson, T., 2003, Shaping Medieval Landscapes: Settlement, Society, Environment, Macclesfield: Windgather Press Williamson, T., 2005a, ‘Soil Landscapes’, An Historical Atlas of Norfolk, T. Ashwin & A. Davison (eds.), 8-9 Williamson, T., 2005b, ‘Place-name Patterns’, An Historical Atlas of Norfolk, T. Ashwin & A. Davison (eds.), 34-35 Williamson, T., 2006, England’s Landscape: East Anglia, London: English Heritage Williamson, T., 2008, Sutton Hoo and its Landscape: The context of monuments, Oxford: Oxbow Books Williamson, T. & Bellamy, L., 1983, Ley Lines in Question, Kingswood: World’s Work Williamson, T. & Skipper, K., 2005, ‘Late Saxon Population Densities’, An Historical Atlas of Norfolk, T. Ashwin & A. Davison (eds.), 38-39 Wilson, D., 1992, Anglo-Saxon Paganism, London: Routledge Worgan G., 1815, General View of the Agriculture of the County of Cornwall, London: Sherwood & Neely Wrathmell, S. and Roberts, B., 2002, Region and Place: A study of English Settlement, London: English Heritage Young, A., 1804, reprinted 1969, General View of the Agriculture of the County of Norfolk, Newton Abbot: David & Charles Young, A., 1813a, reprinted 1970, General View of the Agriculture of the County of Lincolnshire, Newton Abbot: David & Charles Young, A., 1813b, reprinted 1969, General View of the Agriculture of the County of Suffolk, New York: Augustus Kelley Young, A., 1813c, reprinted 1970, General View of the Agriculture of the County of Sussex, Newton Abbot: David & Charles 302 APPENDICES Appendix 1 303 APPENDIX 1 – Eighteenth- and Nineteenth-Century new-build Churches: Ecclesiologists, Survey Results and Analysis. This appendix presents the results of a parallel survey of the alignment of 400 newbuild eighteenth- and nineteenth-century churches in the same areas as the medieval church sample. It examines the writings of the various nineteenth-century ecclesiological and architectural groups to see if they had any influence over church alignment during the period, and tests this using the results of the survey. It also assesses the impact of the eighteenth- and nineteenth-century survey results on the possible explanations for the alignment variations amongst medieval churches presented in the body of this thesis, and attempts to establish whether the same influences were still in force, or whether there were different influences in the two periods. Nineteenth-Century Church builders The nineteenth century saw a considerable church-building- and church-restorationprogramme for several reasons; rapid population growth and the creation of new parishes, mostly through urban expansion; the degradation of medieval churches, especially chancels, since they were disused after the Reformation and particularly during the Commonwealth period; and an upsurge in high-church feelings. Accepted wisdom is that churches were neglected during the Georgian period, although this is probably influenced by Victorian propaganda; Steven Curl put it most strongly, “Anglican churches had been shamefully neglected” since the time of the Reformation (Curl 1995, 47). However, despite the suggested Georgian ‘neglect’, new churches were being built at least 30 years before the start of Victoria’s reign. An Act of Parliament was passed in 1803 to “Promote the building, repairing and otherwise providing churches and chapels”, and amended in 1811. It was followed in 1818 by an Act for promoting the building of additional churches in populous parishes (Chadwick 1971, 84-85; Curl 1995, 21). The Church Building Society was formed in the same year and incorporated in 1828. The Act allowed for state funding to be made available for church building, resulting in buildings known as Commissioner’s Churches. In all, 214 large churches were built, often known as “preaching-boxes” (Curl 1995, 21) – a theme often used by Pugin. In full flight in his lectures about church architecture, he said that “A room full of seats at the least possible cost is the present idea of a church”. He went on to refer to “Government preaching-houses, called Churches” which he likened to “Bethel Houses and Socialist Halls” (Pugin, 1853, 48). There are very few of these churches in this survey as the vast majority of them were town or city churches, rather than rural ones. This was also a time of growing Dissent and Catholicism, and attendance at the established church, especially in cities, was falling rapidly; in Birmingham, 75% abstained from worship (Curl 1995, 22). The Oxford and Cambridge Ecclesiastic Societies and their writings In the early Victorian period, two groups were set up to influence the building and restoration of churches, with the aims of reintroducing churches in the High-church mould, rather than the Evangelical mould. The Oxford Society for Promoting the Study of Gothic Architecture (OSPSGA) was formed in 1839, two months before the Appendix 1 304 Cambridge Camden Society (CCS) (White 1962, 38). The CCS was setup to study “Gothic Architecture and Ecclesiastic Antiquities” and gave rise to the term Ecclesiology, whereas the remit of the OSPSGA also included secular buildings (White 1962, 43). The OSPSGA has to be distinguished from the Oxford Movement, which was predominantly about liturgy and belief, rather than church buildings. In his book on the Oxford Movement, the Dean of St Paul’s, R. W. Church, wrote a 24 page chapter entitled ‘The ideal of the Christian Church’ where church buildings are not mentioned at all (Church 1892, 360-384). It has been suggested that the Oxford Movement was interested in orientation of churches, as Johnson wrote in 1912, quoting Victorian authors, that the “practice of orientation had grown lax in the years preceding the founding of the Oxford Movement in 1833” (Johnson 1912, 206 citing Murray in 1895), although the results here will show this not to be true, as alignment towards east in Victorian times was rather less rigorous than during the medieval period, indeed the only examples of alignments towards north or south are from the Victorian period. Although the OSPSGA has been seen as less influential than the CCS, and was described in an article in the Eclectic Review in 1849 (quoted in White 1962, 24) as “more academic and antiquarian rather than religious crusaders”, the OSPSGA was the first of the two societies to produce concrete advice on church building. After requests for advice, the OSPSGA published a set of working drawings in 1840 for the building of a church which was “a monument to ecclesiological principles” and “a great example of how to build in a ‘correct’ Gothic style”, preceding any comparable advice by the CCS (Prout 1989, 381/2). One of the aims of the CCS was the recording of existing churches and it published a “blank form for the description of a church” (White 1962, 54). This ‘form’ grew to 260 items by its 4th edition in 1843 and measurements of the building were considered particularly important. It was published as A Few hints in the practical study of Ecclesiastic Antiquities for the use of the CCS between 1839 and 1843, culminating in the Handbook of English Ecclesiology in 1847,which consisted of 266 pages of text and an additional 118 pages of Appendices (White 1962, 58). The fourth edition of A Few hints in the practical study … contained two completed examples of church surveys - St Mary & St Michael, Trumpington and St Andrew, Cherry Hinton, both in Cambridgeshire. Interestingly, neither has any details entered for the orientation of the church, despite the remainder of the form being comprehensively completed (reprinted in full in Webster 2003, 115-126). In 1844, the Society published ‘The Orientator’ (The Orientator: A simple contrivance for ascertaining of the orientation of churches) possibly to assist church recorders in obtaining the information lacking in the examples published a year earlier. It consisted of a rectangular card with an attached disc showing the position of sunrise on particular saints’ days, to determine “the point of the compass to which a church is directed, and more particularly whether that point be the suns place of rising at the festival of the saint in whose honour the church is dedicated” (Orientator, 1 - cited in White 1962, 59). It also included a table of saint’s days to establish “whether the supposed rule of orientation was adhered to” (White 1962, 60). The instructions for use were “to place the card parallel to the wall of the church, establishing north with a compass placed on it, then to observe the spot where the sun appeared at daybreak”. The Society seems to have lost interest in the subject without establishing any conclusive results (White 1962, 60), presumably because of the lack of data due to the difficulty in its collection, which requires the observer to be at the church at the time of sunrise on the feastday of the Patron saint (soon after 3:30 a.m. GMT for saints such as John the Baptist, with a close to midsummer feastday), Appendix 1 305 and in addition, for there to be a clear sky in order to observe the actual azimuth of sunrise. The first CCS document to give advice on building was A few words to churchbuilders, written by John Neale and first published in 1841 (Neale 1841a). It contained 58 paragraphs aimed at church-building committees rather than architects directly (Webster 2003, 128), presumably with the aim of persuading the clients to specify the desired forms of building to the architect. George Gilbert Scott noted that prior to his “conversion” by the CCS and Pugin in 1841 (Stamp 1995, 87), he built at least six churches in 1839, when “he had no idea about ecclesiastical arrangement” (Stamp 1995, 86). Paragraph sixteen of A few words to church-builders contains the only mention of orientation, but it is more descriptive than prescriptive, accepting that churches were not aligned east-west in the past; and apparently accepting that some modern churches were built north-south. It appears to have identified the foundation date of the church (consecration day) with the patronal-saint’s festival, rather than as separate dates in most cases, and its only criticism on the subject of alignment was reserved for churches that have their altar at the west end. Paragraph sixteen says: The orientation, that is the precise degree of inclination of the church towards the East, is the next point. It is well known that the direction to the due East was not thought necessary by our ancestors: they used to make the church point to that part of the horizon in which the sun rose on the day of the foundation of the church, the day also, it should be remembered of the Patron saint. But many modern churches are built directly north and south, in total defiance of the universal custom of the Church in all ages: and some, as if out of pure perverseness, though they stand east and west, have the Altar at the west (Neale 1841a, 10). It did not suggest that the modern churches should be built accurately east-west, except to say that north-south is “in defiance of the universal custom”, but without specifically criticising it. Given that a major proportion of the new churches of this period were being built in towns and cities, perhaps a prescriptive statement that churches ‘must’ be aligned east-west was seen as difficult to comply with on restricted urban sites. In another pamphlet written by Neale in 1841, this one to Churchwardens (Neale 1841b), he encouraged them to maintain churches in the highest order, with practical suggestions for removing mould and damp, and requested them “to resist every kind of change if you would not have your church spoilt” (Neale 1841b, 6-9). In the middle of all his suggestions is the isolated line “There is no one but knows that every old church is built east and west” (Neale 1841b, 7). It is part of his description of churches generally, which defined nave, chancel and aisles, but does not bear at all on what he went on to say about maintenance of the building. The CCS published a magazine entitled the Ecclesiologist, which commenced publication in November 1841, and ran for 29 years in all (White 1962, 49/50). By 1843, Benjamin Webb, “the driving force behind the CCS” (Brandwood 2000, 447), said that the CCS had had a great influence on church building - “the calm and steady diffusion of the views and principles advocated by the society, and especially the growing adoption of them by professional architects, are highly satisfactory (Ecclesiologist II, 88; White 1962; 183). Almost a quarter of a century later, Hope stated “we have turned minds upside down as to the outside and general fabric of the church , … and so have given new life … to the worship in the Church of England” Appendix 1 306 (Ecclesiologist XXV, 209; White 1962, 183). The CCS’ concern seems to have been almost entirely to do with particular aspects of church architecture, such as which form of pointed architecture was the most suitable, and mostly concentrated on specific parts of the building, particularly the chancel. Colour and music also featured strongly (White 1962, 187). The redundancy of chancels during the Puritan period was seen as a problem to be redressed in both new and restored buildings, and restrictions by the Church authorities on the numbers of steps into the chancel and the elevation of the altar were seen as arbitrary by the Ecclesiologists (White 1962, 184). Conversely, the archdeacons of Middlesex and London described the effects of Ecclesiology as “under happier circumstances would have been safe and harmless” and that it had “a tendency to heighten Romanizing fever” (White 1962, 184). The ideal church, as originally conceived by the CCS, should take the form of “an exquisite village church” (White 1962, 186), but various churches in London were criticised for being “too like a town church”, or had “too much of a country-church look”. Writing in 1850 (Ecclesiologist XI, 229-231), G.E. Street set out “six points of essential importance for a town church, but not all equally necessary in the country”, among these were “avoid rusticity”, shallow roofs, clerestories, “regularity of parts”, “height was of immense importance”, but not necessarily a spire – again no mention was made of orientation. The main focus of the two groups was to increase the number of churches in which Tractarian services, which bordered on the Anglo-Catholic, could be performed. Tractarians encouraged the building of “fortress-like” churches (Curl 1995, 66), such as St Peter Kirkgate in Leeds (1839-41 by Chantrell), enabling a high-church return of processions and generally awe-inspiring services. All Saints’, Margaret Street, Westminster, built by Butterfield between 1849 and 1859, was also designed on Ecclesiological lines to provide Tractarian services and is generally seen as the beginning of the High-Victorian phase of the Gothic Revival (Chadwick 1971, 168172; Curl 1995, 66). This type of architecture was seen as the High Church desire to express confidence in the future of Anglicanism after the defections to the Catholic Church, of John, later Cardinal, Newman from the OSPSGA, and Augustus Pugin from the CCS, the most prominent examples (Curl 1995, 65). Between 1840 and 1900, 6,000 churches were built in England (Morris 1989; Curl 1995, 50), mostly in towns where the growing population required them, and many thousands were “restored” with greater or lesser sympathy, as Curl put it, “The tyranny exercised by The Ecclesiologist in criticism of architects, must not be underestimated, as it was a Journal of immense power and influence” (Curl 1995, 50). The Ecclesiologist fully embraced Pugin’s ideas about architecture – that Decorated Gothic was “the true Christian Style”, and this so called “second- or middle-pointed” form of gothic architecture took over (Scott 1881, 87-89; Ecclesiologist X, 204) and “was often brutally imposed on real medieval churches, whether Early English or Perpendicular” (Curl 1995, 49). At the same time, there was also a rise in the Evangelical Party, which saw symbolism, imagery and decoration as “idolatrous gewgaws” and “superstitious practices” (Curl 1995, 48). They were referring to chancels with piscinae, aumbries, credence tables and Easter sepulchres as survivals from the dark ages and a threat to Protestantism. On the other side, the Ecclesiologists argued that the Evangelists had allowed churches to fall into disrepair, allowed liturgy and sacraments to fall into disuse and the young to grow up ill-informed. They concluded that “the High Church Appendix 1 307 required a decorated church with a chancel, and architects and builders had to learn Gothic” (Curl 1995, 49). At the start of his History of English Church Architecture, written in 1881 after he had “left” the Ecclesiologists, Sir George Gilbert Scott wrote that “We have broken the tradition which maintained the continuity of art history and made each successive style the natural outcome of its predecessor. Everywhere we meet with reproductions of ancient styles and the attempted revivals of lost traditions” (Scott 1881, 2). These church “traditions” which the ecclesiologists were trying to uphold, or re-create, did not apparently stretch to the subject of church alignment, merely to the form and decoration of the buildings themselves. The main analysis of the results of the post-medieval church survey follows in later paragraphs, but an initial examination is made here to test whether the extensive influence that the Ecclesiological Societies had, also extended to the issue of the alignment of churches, despite the lack of its direct mention. Initial survey results analysis – possible influence of the Ecclesiological Groups Although the CCS and OSPSGA wrote nothing that referred directly to alignment, any indirect influence in the alignment of churches built after the two groups became well established has to be checked. In other words, was the upsurge in the support for “true principles” that formed the central tenet of their writings reflected in a change in the alignment of newly built churches closer to East? Allowing for the lead-time for the commissioning, planning and building of a church, and time for the two societies, particularly the CCS, to grow in size after their inception in 1839, any influence that they may have had on church alignment is not likely to have manifested itself until the late 1840s. The table below shows the relative alignments of churches built before and after 1850. Table A1.1 New build Post-medieval church alignment summary by date of building No. Range Mean 95% 95% Range % N of conf East Pre 1850 144 43º-122º ±2.4º 77.8º-82.6º 80.2º 73 Post 1850 251 36º-126º ±2.3º 81.0º-85.5º 83.3º 64 TOTAL 395 82.2º 67 Although alignments are 3° closer to east after 1850, it is not a function of the influence of the CCS, as table A1.2 below, and its analysis will show. The variation in alignment is very close to being a statistically significant one, as the 95% confidence ranges of two mean alignments only just overlap. This time, the difference cannot be explained by an imbalance of new builds between the west and the east of the country, in the “high” and “low” alignment areas. There is almost exactly the same balance of locations for pre- and post-1850 churches - around 30% in Cumbria, around 23% in Norfolk, Sussex, Lincolnshire and Kent and exactly 48% in the other counties, in both groups. The movement of magnetic north during this period may have had an influence and will be discussed in later paragraphs. Appendix 1 308 Analysis by architect A list of the 1243 known members of the Cambridge Camden Society, and their professions, was provided as an appendix in A Church as it Should be … (Brandwood 2000, 359-452). It identified 82 architects, many of them of international repute, such as William Butterfield, R.C. Carpenter, Benjamin Ferrey, Anthony Salvin, George Gilbert Scott, G.E. Street and S.S. Teulon. For new-build churches in this survey (as opposed to restorations), where the relevant volume of Buildings of England identified the architect, churches have been divided into members and non-members of the Society in the table below. Although some of these designs from the better-known names probably did not come from the named architect themselves, Gavin Stamp wrote that Gilbert Scott “was responsible … for pioneering the modern architectural office, producing work in a characteristic ‘house style’ which is not always from the hand of the nominal senior partner” (Stamp 1995, c), so any influence that the Cambridge Camden Society may have had on the principal might reasonably be assumed to have filtered down to the hand that actually drew the plans. Table A1.2 - New-build post-1850 church alignment, summary by architect No. Range Mean 95% 95% %N conf Range of East C.C.S. members 61 44º-115º ±4.5º 79.2º-88.2º 83.7º 64 not C.C.S. members 155 36º-126º ±3.1º 80.1º-86.3º 83.2º 63 Architect unknown 35 50º-110º ±4.8º 78.1º-87.7º 82.9º 66 TOTAL 251 83.3º 64 There is no meaningful difference between the alignments of churches designed by identified architects, whether they were members of the Cambridge Camden Society, shown in the first row, or those who were not, shown in the second row, with no more than a 0.4° difference between any of the groups and the overall mean value. Similarly, churches where the architect was not identified in Buildings of England vary little from the mean alignment of the whole sample. The proportion of churches that are aligned to the north of east hardly varies at all between the three groups. As might be expected, the churches designed by unidentified, presumably local, architects tend to be located in areas far away from London. This applies particularly in Cumbria, where almost 50% of the ‘architect unknown’ churches in the table above are located, and also where church alignments are generally more northerly. It seems fair to say that, from both the similar mean alignments of the churches in each of the groups, and the similar wide range of alignments between individual churches in each of the groups, the alignment of the churches built during this period was not influenced at all by the writings of the various religious and architectural organisations of the period. Nineteen of the 25 post-medieval churches which were excluded from the analysis because of their extreme alignments, were built by known architects, according to the relevant Buildings of England. Only one of them was designed by a CCS member - Benjamin Ferrey designed St Barnabas’, Swanmore in Hampshire in 1846, but did not become a member until 1858 (aged 48), although he was listed as a member of the Oxford Architectural Society (the successor to the OSPSGA) in 1845 (Brandwood 2000, 391 & 361). The church is aligned almost due south-east. The photograph below was taken at 9.10 a.m. GMT (the church clock is showing 10.10 BST), when the sun is almost due south-east, but the sun still just shining on the ‘north’ Appendix 1 309 wall. The church was built on a large flat site, at least ten metres from any of its yard boundaries and nowhere near parallel with any of them, so the alignment appears to have been chosen specifically. Twelve of the remainder of the 25 are built within two metres of a churchyard boundary in restricted yards where the alignment of the church does seem to have been influenced, if not determined, by the site itself. Figure A1.1 – St Barnabas’, Swanmore (Hampshire), aligned to the south-east. Possible influences on alignment, by someone other than the architect The possibility that the architect was not the only determinant of the alignment of some Victorian churches has to be considered. In certain circumstances the ‘owner/sponsor’ of a new church, to put it in medieval terms, may have had, or wanted to have, an influence on the siting of the new church - for example when the church was to be built close to the ‘big house’, affecting the way that the church was to be seen from the house. Thirty-eight of the post-medieval churches in this survey are built next to a manor house/hall, or in adjacent parkland, where they might have been built in such a way as to maximise the beneficial view of the church for the occupants of the house. If the church is aligned close to east, it is not possible to separate any possible ‘visual’ factors from the general practice of eastward alignment. However, eight of the thirtyeight churches are aligned more than 30° from east – five of these can be shown to have been influenced by other factors, such as down the slope of the land or built parallel with the closely adjacent road or boundary, so they were not necessarily rotated away from east for the benefit of the view from the house, but three probably were – Kirkandrews in Cumbria, Leaton in Shropshire and Eastville in Lincolnshire. At Kirkandrews in Cumbria, the current church was built in 1776, but with older plate and vessels (Pevsner 1967, 147), so is not the first church on the site. The current church is aligned almost due south-north at 357°, and is therefore at rightangles to the view from the house some 400 metres to the west, rather than presenting an end-on view of the nave if the church was aligned east-west. Appendix 1 310 Figure A1.2 – Kirkandrews church, in its large yard, aligned south-north for the improved view from the modern house, close to the pele tower seen in the background. Compare the church alignment with the east-facing gravestones. The ‘big house’ at Eastville however, is the Vicarage, where Pevsner noted “the [Vicarage] doorway with Gothic detail is in axis with that of the church tower” (Pevsner & Harris 1998, 267), which contains the entrance to the church via the west door, so the Vicarage door and church entrance are opposite each other. Pevsner did not comment on the building date of the Vicarage, but since the church was built soon after the drainage in this part of the fens was completed (Pevsner & Harris 1998, 65), it must be assumed to be contemporary with the church. The church was built in 1840, and is the most extremely aligned church in this survey, at 200º, the chancel is twenty degrees west of due north-south, but is built roughly parallel with the adjacent road, although at least ten metres away, so nowhere near close enough to have been forced to do so, leaving only the presence of the Vicarage as the unlikely determinant of the alignment. At Leaton in Shropshire, the church is aligned at 46° (within a degree of due north-east) and is built at right-angles to the house which is close-by, thereby giving a better view of the church from the house. It is built on a flat site at least ten metres from the closest boundary, so neither slope nor a small site can be used to explain an alignment which is shared by less than 0.5% of the post-medieval churches in this survey, some 36° north of the county mean value. There are a further five churches built close to a manor house or Hall which are aligned between 15° and 30° from east. Two of them, at Child Ercall and Fodesley, both in Shropshire, would have been seen better from the house if they were built eastwest, rather than as they were at 69° and 115° respectively. The alignment of the remaining three can be shown to have been influenced by factors other than sightlines, where each of them is built close to a road, and the relationship between the church and Appendix 1 311 house has not been improved by the alignment of the church – at Fauls and Peplow in Shropshire and at Sewerby in East Yorkshire. It seems safe to say that only two, possibly three, out of a total of over 400 postmedieval churches in this survey, were set out with the intention of providing a specific view of the church from the house, but on the other hand, there are at least two churches where the view from the big house would have been better if the church was aligned closer to east-west, rather than with their actual alignments. There may have been other examples where the siting and alignment of the church contributed to the house occupant’s view but cannot be separated from other influences, but overall, it seems an insignificant factor on alignment as far as the overall sample is concerned. As well as the church at Eastville, there were several other Anglican chapels built in the Lincolnshire fens between 1816 and 1821, soon after the land was drained (Pevsner & Harris 1998, 213). Five of these were built by a local architect - Jeptha Pacey, under the sanction of the Fen Churches Act 1812 (Pevsner & Harris 1998, 65). Three of them, at Carrington (1816), Langrick (1818) and Whaplode Drove (1821) are built specifically not parallel to the churchyard boundaries, so their alignments were presumably set out on purpose, rather than just conforming to the site. The other two, Midville (1819) and Frithville (1821) are roughly parallel to the adjacent road. All five were built on flat sites (on drained fenland) and are obviously aimed eastwards, but without specific accuracy – 94º, 83º, 90º, 96º and 97º, which seems odd given that the alignment of at least the first three of them (in bold) was apparently ‘chosen’, in that there were no other obvious influences such as a boundary to align with. This variation seems to argue against the same method of setting out being used at each site, or at least the use of accurate equipment such as a compass (unless the setter-out was not competent). The five churches built by Pacey are a good example of small village churches that are remote from any possible influence of the Lord of the Manor, or indeed anyone else, who might have tried to influence alignments to improve the view of the building from the village or a specific house, and are part of the great majority of post-medieval churches which are aligned close to east. Ironically, the last of the six Fen Act chapels was at Eastville, mentioned earlier, and was built by Pacey’s pupil, J.C. Carter (Pevsner & Harris 1998, 267), some 20 years later and 110° different in alignment. Post-medieval church alignment Despite there being several surveys of the alignment of medieval churches, discussed in Chapter One, there are no published surveys of churches of the Victorian era – a period of much new church building. There is one published study of Queen Anne period churches (early eighteenth-century), but its sample was too small to be statistically useful and only included urban sites; the details are listed below, followed by the results of the survey for this thesis. Queen Anne churches A small group of eighteen churches on the eastern edge of London was surveyed by Ali & Cunich in 2005 and showed a similar range of alignments to the larger medieval set analysed here. In all, even for this small sample, the variation in alignment was 58º, between 57º and 115º. Since many of these churches were built on urban sites, it is Appendix 1 312 impossible to draw any conclusions about alignment intentions, despite Ali & Cunich’s assertion that when Sir Edmond Halley joined the Church Commissioners, he “must have affected alignment with his scientific knowledge” (2005, 67). There are no rural churches of this period in the survey for this thesis. Survey results for eighteenth- and nineteenth-century new-build churches In all, 420 new-build churches, as opposed to a rebuilding on a medieval church site, built between 1721 and 1902 were surveyed in ten of the counties used for the survey of medieval churches. Ninety-three percent of them were built after 1800, and they are all hereafter referred to as ‘post-medieval’ for simplicity. The results seem broadly to confirm medieval alignment aims, focusing generally eastwards. Medieval church alignments exhibit a range of exactly 90º between the most northerly and southerly aligned churches (between 38º and 128º). Post-medieval churches exhibit an even wider range of alignments – in excess of 220º. One church chancel, mentioned earlier, at Eastville in Lincolnshire is aligned 20º west of south, Forest Row (East Sussex) faces due south, whilst two others are approximately 20º east of south; (Madeley, Shropshire and Nenthead, Cumbria). Two chancels (Wendy, Cambridgeshire and Brathay, Cumbria) face due north, with five others aligned even further from east, actually to the west of north (Kirkandrews (mentioned earlier), Frizington and Ainstable in Cumbria, Tilstock in Shropshire and Ellerby in East Yorkshire). In these cases the actual compass reading is in the region of 350º, which causes problems when mean values are being calculated as they are ‘only’ 100º away from east, rather than the 260º implied by the numerical difference between 350 and 90º. In the calculations of means and other statistical analyses, the alignment of these churches has been taken as -10º, which better reflects their actual alignment in relation to east. Post medieval church alignment 60 EAST number of churches 50 40 30 NORTH SOUTH 20 10 degrees from north Figure A1.3 – Alignment of Post-medieval Churches Appendix 1 313 203-207 193-197 183-187 173-177 163-167 153-157 143-147 133-137 123-127 113-117 103-107 93-97 83-87 73-77 63-67 53-57 43-47 33-37 23-27 13-17 3-7 353-357 343-347 333-337 0 However, as can be seen from Figure A1.3, the vast majority of these churches, like their medieval counterparts, are aligned close to east. Overall, there are twenty-five churches with alignments that are so far from east that they appear to be pragmatic solutions to site specific problems (or site-specific influences), in most cases these are narrow sites that would have prevented alignments closer to east-west. In order to compare the Post-medieval results with the medieval set, these twenty-five churches have been excluded from the majority of the analyses below, using only the remainder under the heading “less extreme churches”. Since the twenty-five churches are spread at both extremes, close to north and south, their removal from the analysis hardly alters the overall picture. As table A1.1 shows, the overall mean alignment is reduced by 0.6º, but the proportion of churches aligned to the north of east remains the same, at 67%, and the overall statistical confidence in the mean results is improved as measured by the 95% confidence level, which reduces the range from ±2.4º to ±1.7º. Table A1.3 - Post-medieval churches built on new sites ALL No. 95% Range Mean conf Cumbria 115 -24 - 162° 77.4°° ±4.5° Shropshire 66 -14 – 160° 82.2°° ±6.6° North Somerset 24 43 - 120° 82.6°° ±7.6° East Yorkshire 42 -14 - 117° 78.7°° ±6.6º North Cambridge 17 ±12.5º 1 - 106° 76.1°° South Hants 50 36 – 144º 87.9º ±7.0º East Sussex 33 ±10.8º 24 - 181° 88.2°° Norfolk/East Suffolk 33 ±5.0º 55 - 117° 87.2°° South Lincolnshire 28 50 – 202º 94.3º ±8.8º East Kent 7 ±9.7º 58 - 102° 90.1°° OVERALL 420 -24 – 202º 82.8°° ±2.4°° Less extreme* 395 36-126º 82.2° ±1.7°° * excluding the 25 churches mentioned in the text above Mean range % N at 95% of East 72.9-81.9° 84 75.6-88.8° 70 75.0-90.2° 67 72.1-85.3° 79 63.6-88.6° 65 80.9-94.9° 48 78.0-98.0° 52 82.2-92.2° 59 85.5-102.5° 36 80.4-99.8° 43 80.4 – 85.2° 67 80.5-83.9° 67 Overall, the number of post-medieval churches in each county is smaller than the number of medieval churches, except in Cumbria, where a large-scale postmedieval church-building programme followed industrialization and the creation of new parishes by the division of the large medieval ones (Cumbria C.C. 1998), but the east-west pattern of difference noticed in the medieval Figures is still exhibited in the mean alignment Figures for the post-medieval churches; lower in the west (Cumbria, Shropshire and Somerset), whilst higher in the east (Suffolk, Norfolk, Lincolnshire and Kent). Overall, there is an eight degree mean alignment difference between churches in the east and in the west, shown in table A1.5, with a difference of 2º between the closest ends of the mean ranges at 95% confidence level. The results are still significant even at a 99% confidence level, as the range of variation in the mean values for alignment in east and west still do not overlap. As with the medieval churches, the difference is caused by the alignment curve being displaced to one side of the mean – 77% of churches aligned to the north of east in the west of the country and 58% in the east (compared with 75% and 54% respectively, for their medieval counterparts). Appendix 1 314 Why does this east-west difference still exist? The presence of churches that are aligned north-south or south-north clearly demonstrates that post-medieval builders were obviously less constrained than their medieval forebears, on occasion taking a practical view of the site constraints and working within them. In these cases the Victorian builders were obviously not obligated to align as close to east as possible, so either the focus on alignment was no longer so rigorously applicable, or the situation never arose in medieval times. None of the 1,926 medieval sites in this survey was as small, in other words as narrow in an east-west direction, as some of the post-medieval ones, so the issue of being forced to align north-south did not arise in the medieval period. Whether or not it was the same reason that caused medieval church-builders and post-medieval church-builders to align more churches to the north of east in the west of the country, and fewer to the north of east in the east of the country, the pattern clearly continued. In all, slightly more than two-thirds of post-medieval churchbuilders aligned their churches to the north of east, varying from 84% in Cumbria to 43% in Kent. The same patterns in alignment east-west across the country in postmedieval times are shown in Figure A1.4 on page 320, as were shown in the medieval alignments in Figure 6.1 in the main text. The smaller sample means that the curves are less smooth, but they patently shows the same variation. Table A1.4 - “Less Extreme” Post-medieval churches built on new sites 95% Range Mean Number conf Cumbria 111 74.0-79.6° 44 - 118° ±2.8° 76.8°° Shropshire 62 77.3-86.5° 46 – 126° ±4.6° 81.9°° North Somerset 24 75.0-90.2° 43 - 120° ±7.6° 82.6°° East Yorkshire 41 ±4.3º 76.3-84.9° 43 - 117° 80.6°° North Cambridge 16 ±9.2º 71.6-90.0° 45 - 106° 80.8°° South Hants 47 36 – 126º ±6.3º 78.3-90.9° 84.6º East Sussex 27 ±6.2º 81.7-94.1° 37 - 115° 87.9°° South Lincolnshire 27 50 - 108º ±4.4º 85.9-94.7° 90.3º Norfolk/East Suffolk 33 ±5.0º 82.2-92.2° 55 - 117° 87.2°° East Kent 7 ±9.7º 80.4-99.8° 58 - 102° 90.1°° OVERALL 395 36 -126º 80.5-83.9° 82.2°° ±1.7°° %N of E 84 69 67 76 63 51 48 38 64 43 67 Table A1.5 - ‘Less extreme’ Post-medieval Church alignment summary by longitude No. Range Mean West 197 (2º W+) Central 114 (0.01º – 1.99º w) East 84 (0º - 1.70º E) TOTAL 395 Appendix 1 43126 36126 37117 79.1 84.2 86.8 95% Range at conf 95% . ±2. 76.83 81.4 ±3. 80.84 87.6 ±3. 83.44 90.2 82.2 315 Range at 99% % N of East 76.0 82.2 79.8 88.6 82.3 91.3 77 Medieval Mean 95% %N of E 82.2 ±1.6 75 64 85.0 ±1.0 70 58 88.8 ±0.7 54 67 86.1 63 Impact of the post-medieval church alignment results Alignment variation by longitude Since the spatial variation in the alignment results of medieval and post-medieval churches is similar, there are two possibilities – firstly that the same influences were in force in both periods, or secondly that two different influences were in play that happened to have similar effects on the results. If the same influence was in force in both periods, then it removes the possibility that either the short medieval annual building programme, or sunrise at the start of building, had an effect on the alignment of medieval churches, since the same factors cannot apply to the post-medieval builders. Technology, building materials and building methods had altered and improved, with a far greater use of brick which requires considerably less mortar, thereby removing the need for very short annual building campaigns, particularly in the areas of predominantly flint construction in the south-east of England. Similarly, if there was a single influence, then it cuts out the possibility of magnetic determination of alignment in both periods – since magnetic north in the eighteenth and nineteenth centuries was at least 40º away from where it was in the eleventh and twelfth centuries. The positioning of eighteenth-century sundials either on contemporary, or medieval, churches shows that the issue of variation from east was both understood and capable of being corrected fairly accurately by this time. To allow for the fact that the south wall of a church did not face due south, either the face of the sundial was set at an angle to the church wall, or the gnomon was offset from vertical (for a more detailed explanation, see Wall 2006, 16-17). Similarly, labour availability in post-medieval times would be unlikely to place the same restrictions on building as it had in medieval times; particularly from the mid-Victorian period onwards when full-time labourers were available, and probably employed directly by the builders, rather than the medieval method of using feudal labour during slack periods in the farming year. Therefore labour was likely to be available at any time, and the only thing likely to determine when foundation digging could not happen would be when the ground was too wet or too frozen to dig, otherwise it could be undertaken at any time of the year. The similarity between medieval and post-medieval results could increase the possibility that the spatial difference is brought about unconsciously. Medieval builders might have been influenced by any number of factors, either liturgical or craft-based, but this is less likely in post-medieval times, especially for the Victorians, as they seem to have taken a pragmatic approach to alignment on many sites, so why would they have consciously followed some other influence on alignment elsewhere?, thus increasing the possibility of an unconscious influence. Earlier in this chapter it was shown that the writings of the members of the CCS and the OSPSGA had a major influence over the architecture of the churches of the time, but they wrote nothing about the reasons for, or necessity of, aligning churches in specific directions. The only reference to alignment in this period being that in Wordsworth’s poem, mentioned earlier in Chapter One. As there was no influence by the CCS on the alignment of individual churches towards east, it cannot have had any influence on the differences in alignment in churches between those in the east and west of the country. Neither have there been any apparent medieval written instructions on alignment, perhaps indicating that it was so widely appreciated that it did not require committing to paper, or that it was a craft- or trade-secret so closely guarded that nothing has been revealed to outsiders. By its very nature, such a ‘craft secret’ is difficult to investigate. However, Appendix 1 316 the History of Masonry explained that Freemasonry, as it is known now, was effectively re-invented in Edinburgh in the eighteenth century (Lawrie 1859, 12-14), although Prescott placed the re-invention in London by the formation of the first Grand Lodge in 1717, from which 450 Lodges across the country were authorised by 1757 (Prescott 2008, 2), so continuous links with medieval masons, when there may have been a genuine craft element to the building process, are very unlikely. The position of sunrise still has to be considered as the possible cause of the variation in the alignment of churches in both periods. For the position of sunrise to reflect the fact that the mean post-medieval church alignments are numerically a little lower than medieval ones, area for area, sunrise would need to be a little further north on the horizon, in other words either a little later in the spring or a little earlier in the late summer, than the medieval equivalent. In an earlier chapter, Victorian harvest festival dates between 1870 and 1900 were used as a proxy for medieval harvest times. They showed that an apparent underlying climatic difference across the country closely reflected the actual difference in mean church alignment. Is it possible that the same influence played a part in the planning of post-medieval churches and that a particularly early harvest could still influence church-builders in Victorian times? Perhaps the use of large-scale mechanization meant that harvest was gathered in more quickly in later Victorian times41, and therefore was finished earlier, shifting the date forwards a little and making sunrise a little further north. However, the possibility that all post-medieval builders were equally influenced by harvest completion seems unlikely. Despite the fact that there was still a strong link between the harvest and rural people, even continuing into the early twentieth century when school records in (very) rural Suffolk reflected this intimate connection, recording children absent from school as “they were helping to bring in the harvest” (Tooley 2002, 70), the previously close link between harvest and church-building was lost. Even in the most rural of areas, those who commissioned the building of a church in post-medieval times had to go through a lengthy period of Episcopal approvals and committees, which removes the close timing between the individuals’ decision to build a church and the start of its building, which could have applied in medieval times. It also firmly divorces the date of the inception of the idea to build a church from the day that the plans were drawn up as part of the building contract, which is the point at which the alignment of the postmedieval church was effectively fixed, rather than the more direct medieval action of the marking out and digging of foundation trenches, post holes or ground-beam slots. Taken together, all these arguments seem to remove completely the possibility that the same influence existed on church-builders over a period of several hundred years between medieval and Victorian times, whether it was a magnetic one; dictated by the building programme; prompted by an early harvest; a craft-based tradition or even merely an unconscious influence. Therefore it seems that there had to be two influences on alignment, one in medieval times and another in post-medieval times, but which happened to have similar effects. Whilst early harvest seems to provide a strong guide to church alignment in medieval times, it is less supportable for the postmedieval building process. The possibility of magnetic influence for the alignment of medieval churches was dismissed earlier, as the variation in magnetic declination was in the opposite direction from the church alignment differences. In the nineteenth century however, the variation in magnetic declination was in the same direction as the difference in church alignment between the east and the west of the country. 41 The Beverley Guardian reported several large scale tests of self-binding harvesters on farms in the area in the early 1880s, which left the farmers “much impressed” Appendix 1 317 The United States’ Government Geomagnetic Service website (USGS) can calculate magnetic declination for any point on the globe back to the year 1900, but can go no further into the past as there are insufficient contemporary measurements on which to base the calculations42. The following calculations from the website were made for Cornwall in 1900 (18.5° west of true north) and East Norfolk (15.5° west of true north) – a similar range to that measured in 2002 and listed in Chapter Eight (5° west in Cornwall, and 2.5° west in East Norfolk) and likely to have been a similar range for periods before 1900, although each end of the range would have higher values. These declination values would mean that if a compass was used to set out the churches, there would be a 3° difference between the churches in the extreme east and the extreme west of the country, those in the west being aligned further north of east. The actual mean alignment of post-medieval churches measured in this survey was 79.1° in the west of the country, which is 10.9° north of east, and 86.8° in the east of the country, which is 3.2° north of east, a difference of 7.7°, which is greater than the 3° difference in declination, but the variation at this period is in the same direction. Table A1.6 – Variation in Magnetic north and church alignment by longitude Magnetic east: Mean Church Church alignment: degrees north of true alignment degrees north of east at 1900 ± 95% confidence true east West 18.5° (Cornwall) 79.1° ± 2.3 10.9° (8.6 – 13.2) East 15.5° (E Norfolk) 86.8° ± 3.4 3.2° (0 – 6.6) If correct adjustments were made for declination then all churches would face due east, which obviously did not happen as the east-west difference in alignment is still evident. Declination was part of public knowledge, at least to the educated public (as evidenced by the adjusted sun-dials, mentioned earlier), but it is possible that Victorian architects assumed that the degree of magnetic declination was consistent across the whole country, at the level in London where it was first measured, rather than having a variable value which depended on location. This is supported by the “scarcity of contemporary declination measurements before 1900”, that the USGS website mentioned. If the value for London (approximately 16° in 190043) was applied to all the compass readings across the country, then the result for true east, in the west of the country would be 2.5° north of where it should be (18.5° – 16°); whilst in the east, the result would be 0.5° south of where it should be (15.5° - 16°), so there would be a 3° difference in the results between those in the east and those in the west. So, if magnetic declination was ignored, or was used but applied incorrectly, the alignment of churches in the west of the country would be 3° more northerly than those in the east, in both cases. Alignment variation throughout the nineteenth century In addition to the spatial variation across the country for the whole sample, the postmedieval results also display a change in mean alignments as the nineteenth century progressed, irrespective of longitude. The dates for the groups for this analysis were chosen in order to give similar sized groups, unfortunately, the dataset is far smaller than the medieval set, and cannot be expanded, so the statistical confidence that can be expressed in the results is not quite as forceful. However, at 90% confidence levels, 42 43 http//:www.ngdc.noaa.gov/geomagmodels/struts/calcGRFWMM (24th March 2009) http//:www.ngdc.noaa.gov/geomagmodels/struts/calcGRFWMM (2nd April 2009) Appendix 1 318 churches built before 1850 have a significantly different mean alignment from those built towards the end of the century – 80° rising to almost 85°, shown in table 8 and Figure 9 below. As with the medieval results, the proportion of churches aligned to the north of east confirms the pattern, reducing from 73% to 59% over the same period, demonstrating that the change in alignment is not merely a statistical sleight of hand. Despite the significant difference in the mean alignments, the actual range of observed alignments is similar in each of the three periods, 43°-122° for churches built before 1850, 44°-126° for those built between 1850 and 1869, and 36°-126° for churches built after 1870. Table A1.7 New-build post-medieval church alignment summary by date of building No. Range Mean 90% 90% Range %N conf of East Pre 1850 141 43º-122º ±2.0º 78.2º-82.2º 80.2º 73 1850-1869 130 44º-126º ±2.6° 79.6°-84.8° 82.2º 70 1870 & 124 36°-126° ±2.8° 82.0°-87.6° 84.8° 59 after TOTAL 395 82.2º 67 As was discussed in an earlier chapter, the movement of the magnetic north pole reached its maximum position to the west of true north at around the turn of the nineteenth century, after which the position of magnetic north, and therefore east, moved closer to their true positions at a fairly constant rate of approximately 1° per decade. The post-medieval church alignments appear to show a close link to the changes in magnetic directions as time progressed. There is a much closer match between the values involved here than with the spatial variation results, as, over the period of post-medieval church building, magnetic east shifted 6° southwards, whilst the mean church alignment shifted 4.6° southwards. This is a pattern which did not appear at all in the medieval dataset, when it was shown that church alignment did not vary by date of building, whereas the position of magnetic north moved a similar amount to that in the nineteenth century, albeit in the opposite direction. Table A1.8 – Change in Magnetic north and church alignment in the Nineteenth century Date Pre 1850 (avge date 1825) 1850-1869 (1860) After 1870 (1885) Appendix 1 Magnetic east: degrees north of true east 24° Mean Church alignment ± 90% confidence 80.2° ± 2.0 Church alignment: degrees north of true east 9.8° (7.8 – 11.8) 20° 18° 82.2° ± 2.6 84.8° ± 2.8 7.8° (5.2 – 10.4) 5.2° (2.4 – 8.0) 319 POST MEDIEVAL CHURCH ALIGNMENT by DATE OF BUILDING 30 EAST Percentage of churches 25 20 15 10 5 0 38-47 48-57 58-67 63-67 73-77 78-87 88-97 93-97 103-107 108-117 118-127 128-137 Degrees from North PRE 1850 1850-1869 1870+ Figure A1.4 – Post-medieval church alignment by date of building The alignments of the churches built in the early 1800s by Jeptha Pacey, the local man who designed and built the five Fen-Churches-Act chapels, mentioned earlier, confuse the issue, because although his churches were all built close together, both in location and in time, his results vary either side of true east, apparently at random - 94º, 83º, 90º, 96º and 97º - (mean value 92°). Although Pacey’s alignments appear to be aimed towards true east, they give no indication of the method he was using to achieve them. If he was making consistent errors, either with the compass itself, or the subsequent calculations of declination, one would expect a consistent numerical error in the alignment of his churches. But his variations pale into insignificance when compared to the range of alignments of churches built by every single one of the named architects who built five or more churches in this survey during the period, for example Blomfield 60-117°, Butterfield 79-106°, Cory 69-112° (+ another at 162°), Ferguson 44-98°, Ferrey 69-115° (+ another at 133°), Gilbert Scott 47-85°, Haycock 56-107° (+ two others at 16° & 157°), Paley & Austin 58-105°, Salvin 55-106° and Street 64-109°. When all is said and done, it is this random variation that apparently causes the wide overall range of alignments, either side of the mean value, in the whole church sample. It is not easy to see how the movement of magnetic north should have had an effect on the alignment of post-medieval churches. Whilst the difference in alignment is numerically similar to the difference in the apparent position of east, the actual values are different, and could not have been achieved by using a compass. If a compass had been used to align the churches and no adjustment made for the declination, then the church alignments should share the actual values of magnetic east. Appendix 1 320 If a compass was used and declination adjustment was incorrectly applied, then there would be a 3° difference in alignment, but if a compass had been used and correct adjustments made for declination, then each of the churches should face true east. The results on the ground appear to show elements of all these cases – a shift in mean alignment that mirrors the change in magnetic east over the same period, but with values that are far closer to true east than magnetic east. Unlike all the other analyses of variations of alignments in both the medieval and post-medieval elements of this survey, the variation of post-medieval church alignment by date of building cannot be explained by the predominance of churches located in either the east or west of the country, in other words in the areas of ‘low’ or ‘high’ alignment values. In this instance, churches in the west (the ‘low’ area) form the majority of the cases in all three of the building periods shown in table 8 and Figure 4 above, thereby cancelling out any specific influence of longitude. Rebuilt chancels Another example of Post-medieval church alignment can be illustrated by the rebuilding of ruined medieval chancels. As mentioned in Chapter Five, when considering churches with naves and chancels aligned differently, it appears to have been equally important to realign a chancel during its rebuilding in the eighteenth or nineteenth centuries as it had been to do so during the medieval period. However, there is a good chance that some of the chancels rebuilt in the eighteenth and nineteenth centuries had already been rebuilt in thirteenth century. It cannot be determined whether these later rebuilds followed the earlier foundations or whether they were newly re-aligned as the result of the same desire to point closer to east. Post-medieval rebuilt chancels are usually much shorter and often with thinner walls built in brick, which would have made realignment on the original wider foundations much easier. Of the 730 churches surveyed in Norfolk, Lincolnshire and Hampshire, Pevsner assesses the age of both chancel and nave in 389 cases. Of these 389, 150 have a later chancel, of which 113 are medieval rebuilds and 37 are eighteenth- or nineteenthcentury rebuilds. Of the 113 medieval rebuilds, 33 chancels (29%) are aligned differently from the nave, two-thirds of them (22) were aligned closer to East and onethird (11) further away. The chancel alignments ranged between 67º and 111º - all bar one of the most extremely aligned were realigned (one at 111º was not). Of the 37 eighteenth- and nineteenth-century rebuilds, eleven (30%) were realigned, ten of these were aligned closer to east and one was not. Interestingly, when rebuilding offered the chance to realign the chancel closer to east, the opportunity was not always taken, even when the alignment was far from east. Seven of the 26 chancels which retained the same alignment as the nave on rebuilding had more extreme alignments than the eleven that were realigned. Those that were realigned were between 82 and 106º, whereas the whole group was aligned between 56 and 108º. This seems to be counter to the greater desire to re-align the chancel when the original alignment was further from East, which was highlighted in Chapter Six. Without written records at each site, it is not possible to say why realignment wasn’t effected, but at the two most extreme sites, Thwaite and Surlingham (both Norfolk) and also East Bilney (Norfolk) the churches were built facing directly down slopes; two others, Frettenham and Great Plumstead (both Norfolk) were built on flat sites but within two metres of the closest churchyard boundary, so considerations of slope and site restriction could have made realignment Appendix 1 321 more difficult and may have outweighed the benefits of realigning closer to east. At the remaining two sites, however, at Needham and Burston (both Norfolk) the sites are large and flat so there were no restraints to realigning the chancel when it was rebuilt. Conclusions Almost 400 eighteenth- and nineteenth-century churches built across the country appear to have been aligned in a similar pattern to their medieval counterparts. Apart from a few churches, most of which are a pragmatic solution to the problems of difficult sites, they are aligned basically eastwards, but with wide variations. Two significant results in the alignment of post-medieval churches are apparent; firstly a similar spatial variation across the country to that of the medieval church set, and secondly a difference in alignment depending on when the church was built, which varies apparently in parallel with the change in the position of magnetic north over the same period. What caused the second pattern is unclear, as the extensive writings of the time by the Cambridge Camden Society and others, despite being forceful on other aspects of ‘true’ church construction, hardly mentioned orientation. The CCS produced the Orientator as part of their church recording exercise, but apart from the brief mentions in the pamphlets to church-builders and church-wardens, had not only little to say about the subject of orientation but even less influence over it. There is no difference in alignment between the churches designed by architects who were members of the CCS and those who were not, which clearly confirms the absence of any direction in this matter from the Society. Why then, do eighteenth- and nineteenth-century churches apparently follow the same spatial pattern of variation in alignment as the medieval examples? This analysis has proved that the same influence cannot have been in action across so long a period, causing both early medieval, and post-medieval, church builders to build their churches with similar alignment variation between the east and west of the country. This strongly suggests that there were two different influences in the two periods that resulted in similar patterns of alignment. As far as the medieval period is concerned, early harvests seem to point to an auspicious time to build a church which was aligned towards sunrise then. The close link between decision to build a church and its actual building was broken during the post-medieval period, both from the point of view of the extended intervening time and the number of extra intermediate processes, such as committees, parochial church councils, architects, plans, building contracts etc., thereby removing the possibility of a post-medieval link between harvest and the date of building of the church. Magnetic influences cannot explain the variation by longitude over both periods, as magnetic north was rapidly shifting in opposite directions, but during the nineteenth century the movement in magnetic north and the east-west variation in church alignment were changing in the same direction. The absolute values are different, but the trend is the same, and appears to be the only realistic explanation for the variation in alignment across the country. It is difficult to see what other specific influence could have affected post-medieval builders, since nothing has apparently been written down, or reached modern times through ‘folklore’. Unlike the medieval church sample, which was shown not to vary in alignment by date of building, post-medieval church alignment did alter over time, with mean alignments shifting closer to east as the nineteenth century progressed. Like the postAppendix 1 322 medieval variation in alignment across the country, this change also reflects the rapid changes in magnetic declination over the same period. However, unlike all the other elements of this survey, an unequal balance in the sample between the east and west of the country, the ‘low’ and ‘high’ areas of alignment, cannot be used to explain this difference, as churches in the west of the country dominated each of the three postmedieval periods analysed – pre-Victorian, early Victorian and late Victorian. The fact that both of the observed elements of post-medieval church alignment variation - across the country, and across the period, parallel the changes in the position of the magnetic north pole during the period, emphasises the probability that magnetic variation was at the root of both of the differences, despite the fact that the actual values of the compass bearings of the position of magnetic east are not copied in either the east-west differences across the country, or the early- and late-Victorian mean alignment Figures, but the trends in the movement of magnetic east are closely mirrored in both cases. Although the concept of magnetic declination had been understood since it was first measured in the sixteenth century, it was still misunderstood three centuries later, as is demonstrated in the survey of churches in Scotland by Eeles in 1913 (outlined in Chapter One), where his interpretation of magnetic changes was erroneous (Eeles 1913, 180), indicating that this may have been at the root of the Victorian errors. Appendix 1 323 APPENDIX 2 - Sunrise azimuth by latitude Assumes a level horizon, and is shown in degrees from North. Calculated from the formulae shown in Appendix 6 50 deg 51deg 52 deg 53 deg 54 deg 55 deg 89.4 85.1 80.9 76.9 73.0 69.4 66.0 63.1 60.5 58.3 56.6 55.4 54.7 54.5 54.8 55.6 56.9 58.7 61.0 63.6 66.7 70.1 73.8 77.7 81.8 86.0 90.2 94.5 98.7 102.8 106.7 110.3 113.7 116.7 119.3 121.5 123.3 124.5 125.3 125.5 125.3 124.5 123.2 121.5 119.3 116.6 113.6 110.2 106.6 102.7 98.6 94.4 89.4 85.0 80.7 76.6 72.6 68.9 65.5 62.5 59.8 57.6 55.9 54.6 53.9 53.7 54.0 54.9 56.2 58.0 60.3 63.1 66.2 69.6 73.4 77.4 81.6 85.9 90.3 94.6 98.9 103.1 107.0 110.8 114.2 117.3 120.0 122.2 124.0 125.3 126.0 126.3 126.0 125.2 123.9 122.1 119.9 117.2 114.1 110.7 106.9 102.9 98.8 94.5 89.4 84.9 80.5 76.3 72.2 68.5 65.0 61.9 59.2 56.9 55.1 53.9 53.1 52.9 53.2 54.1 55.4 57.3 59.7 62.5 65.7 69.2 73.0 77.1 81.4 85.8 90.3 94.7 99.1 103.4 107.4 111.2 114.7 117.9 120.6 122.9 124.7 126.1 126.8 127.1 126.8 126.0 124.7 122.9 120.5 117.8 114.6 111.1 107.3 103.2 99.0 94.6 89.3 84.8 80.3 76.0 71.8 68.0 64.4 61.2 58.5 56.1 54.3 53.0 52.3 52.0 52.4 53.2 54.7 56.6 59.0 61.8 65.1 68.7 72.6 76.8 81.2 85.7 90.3 94.8 99.3 103.7 107.8 111.7 115.3 118.5 121.3 123.7 125.5 126.9 127.7 128.0 127.7 126.8 125.5 123.6 121.2 118.4 115.2 111.6 107.7 103.5 99.2 94.7 89.3 84.7 80.1 75.6 71.4 67.4 63.8 60.5 57.7 55.3 53.5 52.1 51.4 51.0 51.5 52.4 53.8 55.8 58.2 61.2 64.5 68.2 72.2 76.5 81.0 85.6 90.3 94.9 99.5 104.0 108.3 112.2 115.9 119.2 122.1 124.5 126.4 127.8 128.6 129.1 128.6 127.7 126.3 124.4 122.0 119.1 115.8 112.1 108.1 103.9 99.4 94.8 89.3 84.5 79.8 75.3 70.9 66.9 63.1 59.8 56.9 54.5 52.4 50.9 50.0 49.8 50.1 51.0 52.5 54.9 57.5 60.4 63.9 67.7 71.8 76.2 80.8 85.5 90.3 95.1 99.8 104.3 108.7 112.8 116.6 119.9 122.9 125.3 127.5 129.2 130.1 130.6 130.1 129.2 127.6 125.3 122.8 119.8 116.4 112.7 108.6 104.2 99.6 94.9 Appendix 2 324 LATITUDE/ DATE 21st March 28th March 4th April 11th April 18th April 25th April 2nd May 9th May 16th May 23rd May 30th May 6th June 13th June 20th June 27th June 4th July 11th July 18th July 25th July 1st August 8th August 15th August 22nd August 29th August 5th September 12th September 19th September 26th September 3rd October 10th October 17th October 24th October 31st October 7th November 14th November 21st November 28th November 5th December 12the December 19th December 26th December 2nd January 9th January 16th January 23rd January 30th January 6th February 13th February 20th February 27th February 5th March 12th March APPENDIX 3 – Survey Form Example survey form of the writer’s local church - photos from north and south follow: Appendix 3 325 St John the Baptist’s, Barnby from the south (above) and the north (below). Figure A3.1 – St John the Baptist, Barnby, Suffolk Sunrise over the level horizon from the church on St John the Baptist’s feast day, June 24th 2002 @ 4:38 a.m. BST Appendix 3 326 APPENDIX 4 - Magnetic Declination 1999-2008 Calculated by the Canadian Geological Service website – http://geolab.nrcan.gc.ca/geomag/e_cgrf.html accessed April 2000, July 2001, April 2002, June 2004, Jan 2005 renamed: http://geomag.nrcan.gc.ca/apps/mdcal_e.php/ accessed June 2006, July 2007, June 2008 now located at: http://geomag.nrcan.gc.ca/apps/mdcal-eng.php (last accessed 18th June 2010) Survey Areas Lat. Waveney Valley (NE Suffolk) East Yorkshire North Cambridgeshire Cumbria East Sussex North Somerset Shropshire East Kent West Cornwall E Norfolk W Norfolk N Oxfordshire Bedfordshire Pembrokeshire S Hampshire S Lincolnshire Long. 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 52° 30’N 1° 45’E 2°48’w 2° 33’w 53° 55’N 1° 5’W 4° 5’w 3° 56’w 52° 55’N 0° 3°20’w 3° 12’w 54° 30’N 3° 5’W 5° 5’w 4° 55’w 51° 00’N 0° 3° 5’w 2° 57’w 51° 15’N 2° 30’W 52° 45’N 2° 45’W 4° 46’ w 3° 58’w 4° 14’w 51° 25’N 1° 20’E 2° 22’w 50° 15’N 5° 25’W 4° 48’w 4° 39’w 52º 45’N 1º 20’E 2º 25’w 2º 16’w 2º07’w 1º 40’w 52°45’ N 0° 45’E 2º 40’w 2º 31’w 2º 23’w 1º 57’w 51º 50’N 1º 30’W 52º 30’N 0º 30’W 51º 45’N 4º 50’W 4º 10’w 50º 45’N 1º 30’W 2º 38’w 52º 50’N 0º 05’W 3º 24’w 3º 15’w 2º 47’w 2º 26’w 2º 17’w 2º 08’w A deduction for declination was made from each measured church alignment for the relevant year and area above, to the nearest whole degree Details of Magnetic declination calculations for realignment of chancels As table A4.1 shows, fewer churches (43%) are realigned closer to Magnetic East (at an average of approximately 100° True during the medieval period (see Clark et al. 1988, 649), than are aligning closer to True East (65%), strongly indicating that east was the focus and that a magnetic compass was not used. Appendix 4 327 Table A4.1 – All Misaligned chancels – closer to true east or magnetic east? (excluding naves of 88-92º for true east, and 98-102º for magnetic east) Improving Not improving No. % No. % Chancels realignment to magnetic east Chancels realignment to True east Total 135 43 180 57 315 206 65 109 35 315 The same analysis for churches where the chancel was rebuilt in the medieval period, shows an even greater bias towards True East, with 71% aligned closer to True East and 29% further away, compared with 47% realigning closer to Magnetic East at the time and 53% further away. Table A4.2 – Medieval rebuilt chancels – closer to true east or magnetic east? (excluding naves of 88-92º for true east, and 98-102º for magnetic east) Improving No. % Chancels realignment to magnetic east Chancels realignment to True east Not improving No. % Total 35 47 40 53 75 53 71 22 29 75 The post-medieval rebuilding of chancels, often a Georgian exercise, would have taken place in a period when magnetic north was west of true north. In 1800, magnetic north was approximately 24° west of north (Clark et al. 1988, 649; Merrill et al. 1996, 46), meaning that magnetic east at that time was at 66°True. If a compass was used without adjustment for declination, it would result in alignments for east of 66°T. Of the 33 churches in this survey with chancels rebuilt in the post medieval period (7 of which were excluded in the earlier analysis, with naves aligned between 88 and 92º), fifteen were realigned closer to Magnetic East (45%) and eighteen were aligned further away. Since it was shown earlier (table 5.7 above) that nineteen of the 26 churches (73%) were aligned closer to True East (90°) and only seven further away, either a compass was not used for the realignment, or appropriate adjustments were made to the readings to take declination into account. Table A4.3 – post-medieval rebuilt chancels – closer to true east or magnetic east? (excluding naves of 88-92º for true east, and 78-82º for magnetic east) Improving Not improving No. % No. % Chancels realignment to magnetic east 15 18 45 55 Chancels realignment to True east 19 7 73 27 Appendix 4 328 Total 33 26 APPENDIX 5 – Additional church alignment tables These results are presented here for completeness, rather than in the main text, as they show clearly that these particular factors have not affected the alignment of the churches themselves. Church planform There is little variation in alignment apparent between churches with different floorplans - the mean alignment of churches with either no aisle or two aisles is within half a degree of the overall Figure. Churches with a single aisle have a lower mean alignment, mainly because they are over-represented in Cornwall and Somerset, where alignments are generally lower. For some reason, these are the two counties that have a preponderance of single south aisles. In all the other counties in the survey, singleaisled churches are roughly equally split between a single north aisle and a single south aisle, but in Somerset, the split is 64:36 and in Cornwall it is even more marked at 68:32. The overwhelming majority of churches in this survey have a west tower, the other groups are considerably smaller which means that the 95% confidence limits in the results are wider, but even so, there is little difference in the alignment of churches in each of the tower groups. Churches with a central tower are over-represented in Oxfordshire and Somerset, whilst churches with no tower are particularly overrepresented in Cumbria, Pembrokeshire and Shropshire. Each of these counties has a lower mean alignment, contributing to the lower means for churches with either no tower or a central tower. More than one in five of all the churches in the survey with an ‘other tower’ are located in Kent, contributing to the higher mean alignment for this group, reconfirming the apparent effect of longitude. Table A5.1– Overall medieval results by church planform No Range Mean Plan Form Central tower 76 57-121 85.1 West tower 1,508 38-126 86.4 Other tower 115 45-116 87.0 No tower 227 48-128 83.7 1,926 86.1 95% conf. ±2.7 ±0.6 ±2.2 ±1.6 ±0.5 % N of East 68 62 61 70 63.8 Type of church tower Round-towered churches have a higher mean alignment because they are found almost exclusively in Norfolk and Suffolk, where mean alignment is higher, whereas square unbuttressed towers are over-represented in Cornwall and Shropshire where alignments are lower. Table A5.2 – Overall medieval results by type of tower No Range Mean Tower type None 227 48-128 83.7 Round 140 56-109 87.9 Square buttressed 1,036 38-126 86.9 Square unbuttressed 523 50-121 84.9 1,926 86.1 Appendix 5 329 95% conf. ±1.6 ±1.7 ±0.7 ±1.0 ±0.5 % N of East 70 60 62 65 63.8 APPENDIX 6 – Sunrise Position Calculation Formulae Azimuth of Sunrise on Saint’s Day (position on horizon in degrees from True North) – assuming a level horizon Step 1 W= 2*PI()*SAINTS DAY NUMBER/365 (Saints Day number in year, Jan 1st = 1 etc.) Step 2 Suns Declination = DEGREES(0.006918-(0.399912*COS(W))+ (0.070257*SIN(W)) -(0.006758*COS(2*W))-(0.000907*SIN(2*W))) Step 3 Azimuth of sunrise on Saints day = 180-DEGREES(ACOS(-SIN(RADIANS(SUNS DECLINATION))/ COS(RADIANS(LATITUDE)))) (Latitude in decimal degrees) __________________________________________________________________________ Azimuth of Sunrise on Saint’s Day (position on horizon in degrees from True North) – taking horizon elevation into account (AZ = Azimuth) Steps 1, 2 & 3 as above Step 4 Hour Angle of Sunrise = DEGREES(ACOS((SIN(RADIANS(HORIZON ELEVATION))-SIN(RADIANS(SUNS DECLINATION)) *SIN(RADIANS (LATITUDE)))COS(RADIANS(LATITUDE))/COS (RADIANS (SUNS DECLINATION)))) (Measured Horizon Elevation in degrees) Step 5 COS AZ = (COS(RADIANS(SUNS DECLINATION))*SIN(RADIANS(LATITUDE)) *COS(RADIANS(HOUR ANGLE OF SUNRISE))-SIN(RADIANS(SUNS DECLINATION))*COS(RADIANS(LATITUDE))) /COS(RADIANS(HORIZON ELEVATION)) Step 6 SIN AZ=COS(RADIANS(SUNS DECLINATION))*SIN(RADIANS(HOUR ANGLE OF SUNRISE))/ COS(RADIANS(HORIZON ELEVATION)) Step 7 Horizon adjusted sunrise azimuth on Saint’s Day=180-DEGREES(ATAN2 (COS AZ, SIN AZ)) ___________________________________________________________________________ Sources: Davis, J., 2002 pers. comm. Davis, J., 2004, BSS Sundial Glossary: A sourcebook of dialling data, Second Edition, Ipswich: British Sundial Society Publications Appendix 6 330 APPENDIX 7 - Detailed tables of church alignment and slope Detailed tables of churches built on slopes, referred to in Chapter Seven Table A7.1 – Churches on slopes by size and direction of slope Church floorspace <190 sqm 190-299sqm 300+ sqm East facing slopes No. 90 70 40 % 45 35 20 Other slopes No. 56 71 46 % 32 41 27 All sloping sites 146 141 86 Total 200 100 173 100 373 Table A7.2 – Churches on slopes by churchyard size and direction of slope Church/yard ratio <7 7-10.99 11+ Total East facing slopes No. 26 77 97 200 % 13 28 49 100 Other slopes No. 17 51 105 173 % 10 29 61 100 All sloping sites 43 128 202 373 Table A7.3 – Churches on slopes by location and direction of slope location In village Isolated Vill edge East facing slopes No. 107 42 51 % 54 21 25 Other slopes No. 74 50 49 % 43 29 28 All sloping sites 181 92 100 Total 200 100 173 100 373 Table A7.4 – Churches on slopes by age of fabric and direction of slope Earliest nave fabric Not known 11 or 12th C 13th C + East facing slopes No. 37 98 65 % 18 49 33 Other slopes No. 33 87 53 % 19 50 31 All sloping sites 70 185 118 Total 200 100 173 100 373 Table A7.5 – Churches on slopes by church/hall focus and direction of slope No focus Church/hall Total focus East facing slopes No. 150 50 200 % 75 25 100 Other slopes No. 130 43 173 % 75 25 100 All sloping sites 280 93 373 Table A7.6 – Churches on slopes by patronal-saint’s season and direction of slope Saints season Winter Equinox Summer No date St. Mary East facing No. 42 69 45 7 37 slopes % 21 35 23 3 18 Other slopes No. 41 42 54 5 31 % 24 24 31 3 18 All sloping sites 83 111 99 12 68 Appendix 7 331 Total 200 100 173 100 373 Table A7.7 – Churches on slopes by village name origin and direction of slope O.E. name O.N. name Other Total East facing slopes No. 160 14 26 200 % 80 7 13 100 Other slopes No. 131 16 26 173 % 76 9 15 100 All sloping sites 291 30 52 373 Table A7.8 - Church/Hall focus on sloping sites by direction of slope, age of fabric by area Non- Norfolk Norfolk All Survey areas Focus Sloping sites/ age of fabric Slope East 11/12th C 13-15th C Not known 11/12th C 13-15th C Not known Total all sloping sites Slope other 21 13 6 26 10 4 80 Not % with focus focus 59 29 26 48 28 25 215 Focus 6 5 1 1 13 26% 31% 19% 35% 26% 14% 27% Not % with focus focus 13 18 5 11 14 4 65 Focus 27 18 6 27 11 4 93 32% 22% 8% 7% 17% Not % with focus focus 72 47 31 59 42 29 280 27% 27% 16% 31% 21% 12% 25% Table A7.9 - Church/Hall focus on platformed sites by direction of slope, age of fabric by area Non- Norfolk Norfolk All Survey areas Focus Not % with Focus Not % with Focus Not % with Platformed sites/ focus focus focus focus focus focus age of fabric Slope East 11/12th C 13-15th C Not known 11/12th C 13-15th C Not known Total all platformed sites Slope other 9 2 3 17 10 5 46 22 21 9 55 39 23 169 29% 9% 25% 24% 20% 18% 21% 1 1 3 5 6 14 3 14 19 7 63 7% 7% 14% 7% 9 3 3 18 13 5 51 28 35 12 69 58 30 232 24% 8% 20% 21% 18% 14% 18% Table A7.10 - Church/Hall focus on sloping sites by direction of slope, village name origin by area Non- Norfolk Norfolk All Survey areas Focus Not % with Focus Not % with Focus Not % with Sloping sites/ focus focus focus focus focus focus name origin Slope East OE origin ON origin Other Slope other OE origin ON origin Other Total all sloping sites Appendix 7 34 1 5 31 4 5 80 85 7 22 66 13 22 215 29% 13% 19% 32% 24% 19% 27% 7 3 1 2 13 332 31 5 24 4 1 65 18% 38% 8% 17% 41 4 6 33 4 5 93 116 12 22 90 17 23 280 26% 25% 21% 26% 19% 18% 25% Table A7.11 - Church/Hall focus on platformed sites by direction of slope, village name origin by area Non- Norfolk Norfolk All Survey areas Focus Not % with Focus Not % with Focus Not % with Platformed sites/ focus focus focus focus focus focus name origin Slope East OE origin ON origin Other Slope other OE origin ON origin Other Total all platformed sites 12 2 20 4 8 46 36 4 12 69 6 42 169 1 3 1 5 25% 14% 22% 40% 16% 21% 22 1 37 3 63 13 2 23 5 8 51 4% 8% 25% 7% 58 5 12 106 9 42 232 18% 14% 18% 36% 16% 18% Table A7.12 - Comparison of slope severity and slope direction for sloping and platformed sites in 90° groups Slope severity 1 in 50 – 1 in 20 – Over Total 1 in 20 1 in 10 1 in 10 Sloping sites East 58% 50% 29% 54% South 15% 18% 36% 17% West 21% 23% 36% 22% North 6% 9% 7% 242 (100%) 117 (100%) 14 (100%) 373 (100%) ALL SLOPING SITES Platformed sites East 39% 23% 20% 33% South 28% 45% 60% 35% West 22% 17% 20% 20% North 11% 15% 12% 173 (100%) 100 (100%) 10 (100%) 283 (100%) ALL PLATFORMS Table A7.13 - Sloping church sites by location of church In vill isolated Vill edge No % No % No % SLOPING SITES 177 47 92 25 100 27 East facing 104 51 45 23 51 25 South 32 51 16 26 13 21 West 35 42 21 25 27 33 North 6 24 10 40 9 36 PLATFORMED East facing South West North Appendix 7 146 52 48 29 17 51 57 48 51 50 66 16 22 17 11 23 17 22 30 33 333 70 24 30 10 6 25 26 30 18 17 other No % 4 1 3 1 1 1 0 0 0 0 1 0 1 0 0 1 0 1 0 0 Total No 373 203 62 83 25 % 100 100 100 100 100 283 92 100 57 34 100 100 100 100 100 Table A7.14 - Sloping church sites by church/hall focus Church/hall No church/ Total together hall focus No % No % No SLOPING SITES 93 25 280 75 373 East facing 51 152 203 25 75 South 16 46 62 26 74 West 19 64 83 23 77 North 7 18 25 28 72 PLATFORMED East facing South West North 51 15 15 13 8 18 16 15 23 23 232 77 85 44 26 82 84 85 77 77 283 92 100 57 34 Table A7.15 - Sloping church sites by Age of earliest fabric in current church (Pevsner) Age N/K 11th C 12th C 13th C 14th C TOTAL No % No % No % No % No % No % SLOPING SITES 69 16 170 46 67 51 18 4 18 13 373 100 East facing South West North PLATFORMED 38 17 9 5 50 55 25 13 7 18 10 2 2 2 13 63 13 13 13 5 90 27 40 13 111 53 16 24 8 39 East facing South West 16 19 11 32 38 22 6 3 1 46 23 8 30 42 24 27 North 4 8 3 23 15 33 9 21 4 61 49 13 31 6 22 33 38 22 20 22 11 14 8 PLATFORMED East facing South West North Appendix 7 92 25 19 10 124 38 42 28 16 63 17 13 7 44 31 34 23 13 71 24 38 8 103 37 40 18 8 334 50 17 27 6 36 36 39 17 8 40 13 26 7 56 17 18 11 10 48 63 14 22 2 17 42 36 18 20 14 10 13 4 Table A7.16 - Sloping church sites by Size of church <190 sq m 190-299 sq 300+ sq m m No % No % No % SLOPING SITES 146 141 86 39 38 23 East facing South West North 32 7 11 1 47 15 30 8 20 30 32 20 18 203 62 83 25 283 92 100 57 34 283 54 17 22 7 100 33 29 21 92 100 57 8 34 12 TOTAL No 373 203 62 83 25 % 100 54 17 22 7 100 33 35 20 12 35 20 Table A7.17 - Sloping church sites by planform Central No tower West tower tower No % No % No % SLOPING SITES 16 48 4 13 292 78 East facing South West North PLATFORMED East facing South West North 11 3 2 14 3 5 4 2 69 19 13 5 21 36 29 14 25 8 10 5 43 11 15 9 8 52 17 21 10 15 26 35 21 19 163 48 65 16 56 16 22 5 202 71 69 72 38 23 34 36 19 11 Table A7.18 - Nave/chancel misalignment by alignment of nave Improving Not improving total No % No % Nave <62º 13 9 69 4 31 63-72º 28 19 68 9 32 All < 73º 41 28 68 13 32 73-77º 33 23 70 10 30 78-82º 60 36 60 24 40 83-87º 67 34 51 33 49 All 73-87º 160 93 58 67 42 All <88º (north) 201 121 62 80 38 All >92º (south) All 93-102º 93-97º 98-102º 103º+ Total exc 88-92º 88-92º see main text – page 144 Appendix 7 114 77 51 26 37 315 62 85 56 39 17 29 206 74 72 77 65 78 65 29 21 12 9 8 109 335 26 28 24 35 22 35 Other tower No % 17 5 4 3 6 4 24 9 8 6 1 TOTAL No 373 21 18 35 21 203 62 83 25 8 283 38 33 25 4 92 100 57 34 % 100 54 17 22 7 100 33 35 20 12 Aligned left No % 4 31 9 32 13 32 10 30 24 40 33 49 67 42 80 40 Aligned right No % 9 69 19 68 28 68 23 70 36 60 34 51 93 58 121 60 85 56 39 17 29 165 49 29 21 12 9 8 150 13 75 72 77 65 78 52 79 25 28 23 35 22 48 21 APPENDIX 8 – Calculation of Easter Standard Easter calculation algorithm (Cheney 2000, 5) Also listed on various websites – for example http://webexhibits.org/calendars/calendar-christian-easter.html (accessed 23rd April 2007) Step Action Example for 902AD result 1 Remainder (Year/19) 902/19 47, remainder 9 2 Remainder (Year/4) 902/4 225, remainder 2 3 Remainder (Year/7) 902/7 128, remainder 6 4 (19 * (step 1)) + 24 (19 * 9 )+ 24 195 5 Remainder ((Step 4)/30) 6, remainder 15 6 (2*step2)+(4*step3)+(6*step5)+5 7 Remainder ((step 6)/7) 8 (step 5)+(step 7) 195/30 (2*2)+(4*6)+(6*15)+5 (123/7) (15)+(4) 9 IF((step 8) > 9) then ((step8)-9) = APRIL easter date 10 IF ((step 8) <10) then ((step 8)+22) = MARCH easter date 123 17, remainder 4 19 (19) – 9 =APRIL 10th Western and Roman Easter dates Differences in the methods of calculation of Easter between the Western church and the Roman church both before and after the consolidation by the Synod of Whitby of 664 CE, meant that in a period of almost 400 years, the date of Easter only coincided on 154 occasions (Cheney 2000, 47-54). In 150 years Celtic Easter was earlier than Roman Easter and was only later in 63 years. Table A8.1 – Comparison of Celtic and Roman Easter dates between 400 and 779 A.D. Celtic Easter Easter on Celtic Easter later by:earlier by 7 same date days 7days 14 days 21 days 28 days Pre Whitby synod 81 147 20 16 Post Whitby synod 69 7 25 2 TOTAL 150 154 0 0 45 18 Appendix 8 336 APPENDIX 9 – Calculations to establish possible Norfolk Minsterchurch sites Method of calculating overall ranking Twenty-six of the 549 Norfolk parishes where medieval churches were surveyed for this thesis were not mentioned in Domesday. The ranked position for the remaining 523 parishes (1 – 523) was calculated for each of the factors for each parish. In some calculations, where several parishes had the same score, they were each given the same rank and the following parish in the sequence was given the rank score that it would have been given if each of the parishes in the group above it had been given sequential rankings. The rankings for each of the factors for each parish were then added together, resulting in an overall “score”, which provides a direct comparison with all the other parishes in the analysis, and leads to a list of parishes that, on these criteria, are the most likely to have had minster churches. Since measuring the absolute value for a factor might lead to erroneous conclusions, for example, the size of one of the many large fen-edge parishes would appear to have the same importance as a large parish elsewhere in the county where parishes are generally smaller, an alternative measure was developed whereby the size of the subject parish was compared with the sizes of all the parishes that surround it. In this case, the areas of all the parishes surrounding the subject were added together and an average size for them was calculated. This was then compared with the size of the principal parish. If the principal is larger than its contacts, the resulting score is greater than 1, and below 1 if the average size of the contact parishes is higher than the principal. The rankings for each parish were then assessed based on these scores This method was used for the majority of the factors considered. Finally, each factor was assigned a weighting to take into account the fact that some of the factors are more important than others in assessing the medieval importance of the parish – for example, whether the parish bears the same name as the Hundred that it is in, or the village name suffix, compared with whether a Domesday church was mentioned or the earliest recorded fabric in the present church building. Factors Used 1 2 3 4 5 6 7 8 9 10 11 12 13 Parish named after a Hundred – indicating most important parish in Saxon times Parish name suffix – indicating relative name chronology Relative soil quality between parish and surrounding parishes – indicating possible preferred area for primary settlement Area of glebe land at Domesday – indicating wealth of church endowment/ endowers Relative Domesday total population – indicating relative importance compared with neighbouring parishes Relative Domesday feudal population Relative Domesday numbers of sokemen Acreage of influence over other parishes, or influence from other parishes, at Domesday (outliers, jurisdiction etc.) – indicating earlier, and continuing, importance. Relative parish size in acres Relative number of surrounding parishes – the number of parish ‘contacts’ for the principal compared with the same scores for its surrounding parishes. Presence of Saxon or Norman Monasteries – perhaps indicating the best estates that were left to the richest Normans Landscape assessment of the church site – the most prominent sites first Relative value from the Norwich Ecclesiastical Taxation of 1254 – relative to the value of its neighbours Appendix 9 337 14 15 16 Earliest church fabric – combination of assessments by H.M. Taylor’s Anglo-Saxon Architecture and the Buildings of England volumes for Norfolk by Pevsner & Wilson Number of Domesday churches, or implied presence of a church Floorspace of the existing church 1 Name (1) – This is based on the assumption that that the Saxon administrative Hundred was named after the most important parish in the area at the time. Parishes with the same name as the Rural Deanery were also considered, but as they are later in date (some time after reorganization of the Diocesan structure in 1072) it was decided not to class them with those with the same name as the Hundred. (There is a very strong argument for altering the rank attributed to the vast majority of parishes that do not bear the Hundredal name, as there is insufficient differentiation between the ‘1’ assigned to the 10 parishes that appear in this factor and ‘11’ assigned to the remainder. It was subsequently increased to 200. Rank Parish with Hundredal Name Hundred name 1 None 11 (200) 2 Village Name - Based on the village name suffix, which is taken as an indicator of the relative age of its foundation, based on the following breakdown:Name Score Rank ham (primary settlements- 99) 1 1 ingham, ing, ton, by, kirk and others not included in 1 or 3, including burgh or borough 2 100 (secondary settlements - 389) thorpe, toft, wick, ley, ling, thwaite 3 489 (tertiary settlements - 35) 3 Soil Taken from the National Soil Survey map, the proportion that each of the soil series formed of each parish was measured. Based on the following list, the soil series information was converted to an assessment of the soil quality for each parish. Soil Series Quality Sheringham (541), Freckenham (551), Burlingham (572) GOOD Gresham (711a), Blackwood (821), Hockham (552) Fakenham (581), Evesham (713) OK Others not included elsewhere Other Sand (110a), Newmarket (343), Beccles (711b&c) POOR The proportion of each soil quality in each parish (GOOD, POOR etc) was summed from the data and a single measure resulted - based on the largest single quality group. For example, in a parish where 50% of the area was soil series 541, 25% was 572, 20% was 711b and 5% was “other” – since both the first two are categorised as GOOD, the third as POOR and the last as other, the assessment for this parish would be 50% + 25% = “75% GOOD”. In order to obtain a measure of the comparative soil quality between a parish and its neighbours, the percentages of each soil quality of all the surrounding parishes was added together and divided by the number of parishes, resulting in a single measure on the same basis as the example above, which could be compared directly with the quality assessed for the principal parish. They were compared on the following basis:- Appendix 9 338 Comparison of Soil Quality between a parish and the average of its neighbours Category Assessment Rank Change from POOR to GOOD (29 cases) Much Better 1 Change from POOR to OK/other, Better OR from OK/other to GOOD OR an increase of 50% or more in the same 30 category– e.g. from 30-60% (50% decrease if POOR) (70 cases in all) Increase or decrease of less than 50% in the same Same category, e.g. from 50-70%, or vice versa 100 (337 cases) Change from GOOD to OK/other, Worse OR from OK/other to POOR OR a decrease of 50% or more in the same 438 category– e.g. from 60-30% (50% increase if POOR) (77 cases in all) Change from GOOD to POOR (8 cases) Much Worse 516 4 Glebe Land – John Blair uses 1 hide of glebe as an indicator of superior church status over the country as a whole (Blair 1987). This level has been reduced slightly here, since there were only five churches in Norfolk endowed with more than 120 acres. Churches with 120+ acres glebe 60-119 acres glebe Less or no glebe land Rank 1 6 100 5 Relative total population recorded in the Domesday Survey The Domesday population was counted for each parish, and a comparison made with the average population in the parishes surrounding it. The resulting ratio for each parish was ranked from 1 to 497. The 24 parishes not mentioned in Domesday were all assigned rank 498. 6 Relative total feudal population recorded in the Domesday Survey The Domesday population of villagers, smallholders and slaves was counted for each parish, and a comparison made with the average numbers in the parishes surrounding it. The resulting ratio for each parish was ranked from 1 to 476. The 46 parishes either not mentioned in Domesday, or had no feudal population were all assigned rank 477. 7 Relative total sokemen (freemen) population recorded in the Domesday Survey The Domesday population of sokemen (freemen, as opposed to Free Men) was counted for each parish, and a comparison made with the average numbers in the parishes surrounding it. The resulting ratio for each parish was ranked from 1 to 299. The 223 parishes which were either not mentioned in Domesday, or had no sokemen, were all assigned rank 300. 8 Domesday outliers The acreage of influence that each parish had over other parishes, or was affected by others, was calculated from the Domesday survey. This included estate “outliers”, land “appertaining to” and land “in the jurisdiction of” another parish. Acreage was Appendix 9 339 summed, and the net Figure of influence was ranked. The 88 parishes with a net positive “ownership” elsewhere were ranked from 1 to 88, the 191 parishes with a net influence from elsewhere were ranked 331-523, and the 244 parishes with no net influence, were all ranked 89. 9 Relative Parish size - (in acres) taken from 1844 Whites Directory, there were no missing values. The average size of surrounding parishes was calculated and compared with the subject parish. The subsequent ratio of parish size to average surrounding parish size was ranked from 1 to 523. They ranged from Wymondham (which is 6.26 times the average size of its neighbours) to Waterden (which is 0.10 the size of its neighbours). 10 Relative number of Surrounding Parishes - The number of parishes whose boundaries touch the subject parish was measured and ranged from 19 to 3. The relationships between the principal parish and the average number calculated for its neighbours were also calculated and ranged between 3.23 (Wymondham) and 0.28 (Crownthorpe) – these were ranked 1 to 522 11 Presence of Saxon/Norman Monastery – taken as a measure of the importance of the parish. Presumably, the best estates were left to the most important (and richest) Normans, who endowed monasteries. In addition to Pestell’s classification, stylus finds and “productive” sites have been added to the Saxon list Rank Saxon/Norman Monastery Saxon monastery/stylus/productive sites 1 Norman monasteries 20 None 100 12 Landscape Assessment Landscape at the church was assessed from the 1:25000 OS series (in the case of “assembled parishes”, the landscape at the “senior” site was used – e.g. Great Dunham, rather than Little Dunham. The following categories were used, and ranked thus:Landscape Category Rank Knoll/Island/Promontory (19 cases) 1 1 Valley side, upper slope (10metres + above water) (168) 2 20 Valley side, lower slope (<10metres above river/stream) 3 189 Valley floor (95 with Cat 3) 4 189 Lowland - flat/interfluve (25m or less AOD) (91) 5 284 Highland – flat/interfluve (30metres or more AOD) (147) 6 375 13 Relative Norwich Ecclesiastical Taxation of 1254 – Provided the value of the ecclesiastic property in each parish. Twelve parishes with missing values were assigned the Mean value of the remainder. Relative tax values were also arrived at by calculating an average tax value for all surrounding parishes and comparing the two figures. They ranged between Wymondham (worth 16.11 times the value of the average of its neighbours) and Wrenningham (worth 0.04 times the average of its neighbours). The ratios were ranked from 1 to 523. 14 Earliest church fabric – data taken from the relevant volume of Pevsner, adjusted by Taylor’s Saxon churches where the two volumes fail to agree. (There may be an argument for shifting back to the “Saxon” period, those parishes where a church was mentioned in Domesday. It was almost certain that, in most cases, Appendix 9 340 it is not the current church building but even then it is does indicate an earlier church presence) Rank Saxon fabric 1 Norman fabric 50 Other 189 15 Domesday churches - The presence or absence of a church in Domesday was recorded, along with some “assumed” churches, where mention was made of priests or glebe land, or land belonging to the church, but not a specific church building. Rank 178 Parishes with a church present (or assumed) 1 27 churches with Saxon fabric, but not mentioned in 1 Domesday No church mentioned, or no Saxon fabric 206 Since there is a well-known under-recording of churches in Domesday, this factor has been modified to take into account parishes that have churches with Saxon fabric, but no mention in Domesday, as there obviously was a church in these parishes prior to the Domesday record in 1086. 16 Floorspace of existing church. One of John Blair’s comments about minster churches is that “the church is very large for a village of this size”. Whilst this is a well-known phenomenon in East Anglia with large “cloth” churches, is there also a picture of large churches in other villages too? Churches were ranked from the largest to the smallest, 1 – 523 MULTIPLIERS Each of these sixteen categories was then assigned a multiplier to take into account the fact that some of the factors were assessed as being more important than others, in other words, the early name origin of the parish was deemed considerably more important than the current floorspace of the church. The rank scores for the factors were assigned a multiplier on the following basis:- Parish named after a Hundred Parish name suffix Relative soil quality between parish and surrounding parishes Area of glebe land at Domesday Relative Domesday total population Relative Domesday feudal population Relative Domesday numbers of sokemen Acreage of influence over other parishes at Domesday (outliers, jurisdiction etc.) Relative parish size Relative number of surrounding parishes Presence of Saxon or Norman Monasteries Landscape assessment of the church site Relative value from the Norwich Ecclesiastical Taxation of 1254 Earliest church fabric Number of Domesday churches Floorspace of the existing church Appendix 9 341 Multiplier 10 10 8 8 6 6 6 6 5 5 5 4 4 2 2 1 Ranking Results The results of this analysis produced a ranking of the 523 parishes that were mentioned in Domesday. There was an apparent ‘natural’ break in the list after the first 27 parishes. The first 27 had larger gaps between the parish scores, with consistently smaller gaps from 28 onwards. The first 27 parishes were, in order :Wymondham, Holt, North Walsham, Aylsham, RUDHAM, Reedham, Loddon, SOUTH WALSHAM, SHOULDHAM, RAYNHAM, Tunstead, Thetford, Happisburgh, BURNHAM, Hempnall, Earsham, Old Buckenham, Diss, Ludham, Heacham, Taverham, BARSHAM, Cawston, Foulsham, North Elmham, Dereham, ACRE (the capital letters indicate parishes which were amalgamated for this analysis as described earlier). Generally, the top 27:• had churches that were more than double the size of the remainder, • were in parishes that were 2.5 times the size of the remainder • had twice as many neighbouring ‘contact’ parishes • their churches were valued at 3 times the rate of the remainder in 1254 • had a majority of primary settlement names, with no tertiary names • there was little difference in the church site landscape • had four times the Domesday population of the remainder • had 2.5 times the feudal population at Domesday • little difference in relative soil quality with the neighbours • had twice as much glebeland at Domesday • had a slightly greater number of churches with early fabric • had many times more Saxon and Norman monasteries • 21 became Market towns (Dymond 2005, 76) Summary Results The following tables summarise the results based on the groups that arise from the rank order classification of all sixteen factors. Top 27 have been separated out because the ranking data showed a natural break at this point, with larger gaps between the parish scores between 1 and 27 and generally smaller gaps from 28 onwards. Alignment Top 27 Remainder Total Number 27 522 549 Church Floorspace <150 sq ft No % Top 27 Remainder Total 0 94 94 22 21 Range 67-110° 56-128° Mean 88.3° 88.9° 88.9° 150-189 No % 0 114 114 21 21 95% ±4.7 ±0.9 ±0.9 190-239 No % 1 111 112 4 21 21 Range @ 95% 83.6 – 93.0 88.0 – 89.8 88.0 – 89.8 240-299 No % 300+ sq ft No % Average size 3 86 89 23 120 143 473 sqf 245 sqf 264 sqf Larger churches in top group and smaller churches in bottom group. Appendix 9 342 % N of E 61 55 56 11 16 16 85 23 26 Latitude & Longitude Average Latitude Top 27 52.70° N Remainder 52.67° N Total 52.67° N Average Longitude 1.02° E 1.10° E 1.09° E Overall, top churches are located in same areas as the others Area of Parish Top 27 Remainder Total Number 27 496 523 Size Range (acres) 2167 – 10600 353 - 12953 Average 4771 1673 1935 Larger parishes in top group and smaller parishes in bottom group, although the largest parish in the county (Methwold) is in the “remainder” group. Comparative Area of Parish Number Top 27 27 Remainder 496 Total 523 Average 2.46 0.93 1.00 The “relative area” data confirms that the parishes in the top group are on average just over twice the size of their neighbours and those in the bottom group are more or less average in size. Parish Contacts Number Range of contacts Average Top 27 27 7 – 18 9.81 Remainder 496 3 - 11 5.72 Total 523 6.00 More contacts (neighbouring parishes) in top group and fewer contacts in bottom group Comparative Parish Contacts Number Top 27 27 Remainder 496 Total 523 Average 1.71 0.93 0.97 Parishes in the top group have more contacts than their neighbours, and vice versa. Church Taxation 1254 Number Top 27 27 Remainder 489 Total 516 Tax Range (£) 10 – 100 0.5 – 53.3 Average 31.9 10.4 11.53 More valuable churches in top group and less valuable churches in remainder group. Appendix 9 343 Comparative Parish Tax Value Number Top 27 27 Remainder 489 Total 516 Average 3.04 0.97 1.01 Parishes in the top group had higher value than their neighbours. Parish name (suffix) 1 (Primary) ham Top 27 17 63% Remainder 77 16% Total 94 18% 2 (Secondary) ingham, ton etc 10 37% 381 81% 391 75% 5 (Terciary) thorpe, toft etc. 16 4% 37 7% Primary “Hams” in top 27 and no tertiary “other/Thorpe”. Parishes with same name as Hundred Hundred Top 27 3 (11%) Remainder 9 (2%) Total 12 (2%) Total 27 494 523 Higher proportion of “named” parishes in the top group. Landscape Top 27 Remainder Total 1 Knoll, Prom. 2 Upper valley 2 7% 73 15% 75 14% 18 67% 151 30% 169 32% 3/4 Lower valley/ val floor 2 7% 105 22% 107 20% 5 Lowlnd flat 6 Highlnd flat total 2 7% 150 30% 152 36% 3 11% 17 3% 20 4% 27 496 523 Higher proportions of the top group in “upper valley”, and lower in the “lowland flat” group Domesday Population (total) Number Top 27 27 Remainder 470 Total 497 Range 51 – 974 0 – 146 Average 137 39 44 Range 30 – 208 0 – 97 Average 59 24 28 More people in top parishes. Domesday Population (feudal) Number Top 27 27 Remainder 470 Total 497 Higher feudal pop in top parishes Appendix 9 344 Domesday feudal pop as % of the total (parishes with >20 feudal pop) Less than 20 Over 80% 50-79% Below 50% feudal Top 27 9 33% 16 60% 2 7% 0 Remainder 84 18% 119 25% 33 7% 234 50% Total 93 19% 135 27% 35 7% 234 47% Total 27 470 497 Higher proportions of parishes with large feudal pop in top group Soil – largest group GOOD Top 27 13 48% Remainder 279 56% Total 292 56% OK 3 11% 25 5% 28 5% other 4 15% 74 15% 78 15% POOR 7 26% 116 23% 123 24% Virtually no difference in soil quality between the groups. Domesday churches & glebeland Vills with church No. of mentioned churches Top 27 14 52% 24 Remainder 156 32% 182 Total 169 32% 206 Avge no. of churches 1.71 1.16 1.22 Glebeland (acres) 801 3,663 4,464 Avge. glebeland 33.8 20.1 21.7 Slightly higher mentions of churches and more churches per parish and larger areas of glebe in top category Earliest Fabric (Pevsner/Taylor) Top 27 Remainder Total 11th C 6 22% 43 9% 49 9% 12th C 8 30% 146 29% 154 29% 13th C 7 26% 104 21% 111 21% Very slight bias towards early buildings in the top group. Saxon/Norman Monastery in Parish Saxon Norman Top 27 6 22% 4 15% Remainder 11 2% 16 3% Total 17 3% 20 4% Total 27 496 523 Higher proportion of early monasteries in top group Saxon Dedication for Parish Church Saxon Total Top 27 2 7% 27 Remainder 33 7% 496 Total 35 7% 523 No difference between the groups Appendix 9 345 14th C 4 15% 125 15% 129 25% 15th C 2 7% 29 6% 31 6% Not known 0 49 10% 49 9% APPENDIX 10 – Harvest Festival Details Average Harvest date by area Year 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 Avge Taunton Average date 13 Sep 22 Sep 21 Sep 21 Sep 18 Sep 22 Sep 12 Sep 22 Sep 22 Sep 8 Oct 22 Sep 20 Sep 24 Sep Blue Red 19 Sep 23 Sep 22 Sep 15 Sep 4 Oct 19 Sep 23 Sep 1 Oct 23 Sep 18 Sep 22 Sep 24 Sep 12 Sep 16 Sep 16 Sep 15 Sep 21 Sep Broads Difference Cumbria Difference Hold’nss Difference Average Taunton ~ Average Cumbria ~ Average Hold’ness date Broads date Broads date ~ Broads - 4 days 28 Sep + 11 days 17 Sep 22 Sep + 5 days 24 Sep -2 days 29 Sep + 5 days 1 Oct + 7 days 30 Sep + 13 days 1 Oct + 14 days 17 Sep + 4 days 19 Sep + 2 days 5 Oct + 16 days 5 Oct + 16 days 22 Sep - 4 days 6 Oct + 14 days 23 Sep + 1 day 19 Sep +3 days 28 Sep + 9 days 23 Sep + 4 days 25 Sep - 13 days 4 Oct + 9 days 29 Sep + 4 days 26 Sep - 4 days 4 Oct + 8 days 10 Oct + 14 days 22 Sep 9 days 1 Oct + 8 days 13 Sep 3 Oct - 5 days 25 Oct + 17 days 8 Oct 25 Sep - 3 days 29 Sep + 4 days 11 Oct + 16 days 23 Sep - 3 days 23 Oct + 30 days 12 Oct + 19 days 28 Sep - 4 days 12 Oct + 14 days 11 Oct + 13 days 26 Sep 11 Oct + 15 days 11 Oct + 15 days 20 Sep - 1 day 25 Sep + 5 days - 1 day 19 Sep 26 Sep - 3 days 7 Oct + 11 days 8 Oct + 12 days 27 Sep - 5 days 10 Oct + 13 days 10 Oct + 13 days 23 Sep - 8 days 24 Sep + 1 days - 3 days 20 Sep +1 day 9 Oct + 6 days 3 Oct 16 Oct + 13 days 24 Sep - 5 days 4 Oct + 10 days 4 Oct + 10 days 28 Sep - 5 days 7 Oct + 9 days 5 Oct + 7 days 30 Sep +1 day 9 Oct + 9 days 7 Oct + 7 days 28 Sep - 5 days 12 Oct + 14 days 13 Oct + 15 days 26 Sep - 8 days 24 Sep - 2 days 22 Sep - 4 days 29 Sep - 7 days 4 Oct + 5 days 7 Oct + 8 days 25 Sep - 1 day 1 Oct + 6 days 1 Oct + 6 days 21 Sep - 9 days 16 Sep - 5 days 26 Sep + 5 days 22 Sep - 6 days 29 Sep + 7 days 7 Oct + 15 days 23 Sep - 7 days 1 Oct + 8 days 30 Sep + 7 days 23 Sep - 8 days 22 Sep - 1 day 20 Sep - 3 days 25 Sep - 4 days 2 Oct + 7 days 4 Oct + 9 days more than 7 days earlier than the area average harvest festival date more than 7 days later than the area average harvest festival date Appendix 10 346 Year Local weather comments Local weather comments Avge date (Taunton) Avge date (Cumbria) 28 Sep 1870 Dry harvest, largely 13 Sep complete by 10 sept. Appendix 10 Local Weather comments (Holderness) Avge date Dry harvest – crops 22 Sep on hills and valleys all ready together 1 Oct 347 1871 Very wet Sept -5.6” 22 Sep rain 29 Sep 1872 Fourteen days good 21 Sep weather, followed by storms – damage, 25 days with rain 1873 Delayed by rain 21 Sep More than 8” in 10 days 1874 Very dry season 18 Sep 30 Sep 1 Oct Frequent heavy 5 Oct rains delaying harvest Severe wind a 6 Oct problem in some areas Disastrous cold 23 Sep winds and heavy rain in August Severe thunderstorms 5 Oct east of Beverley 1875 Prolonged harvest – 22 Sep longest in memory. 3.8” rain in Sept 1876 Thunderstorms during 12 Sep harvest, disrupted due to tempests 1877 Delayed by rain. 22 Sep 1878 Late completion in 22 Sep some areas due to heavy rain Appendix 7 4 Oct Severe 4 Oct thunderstorms and copious hail damaged standing crop Good harvest 13 Sep weather 347 23 Sep Heavy cold rains and 28 Sep fogs have had disastrous effects on the corn crops 29 Sep Severe thunderstorms 10 Oct with copious hail in some areas in early September have delayed harvest 1 Oct Local weather comments (Broads) Avge date Dry harvest, in early. 17 Sep Autumn tillage begun early Oct. Dry Aug, wet Sept, 24 Sep Wet harvest but in good 17 Sep condition Late sowing. harvest Good 19 Sep Most in before rain 22 Sep Weather spoilt harvest 19 Sep 25 Sep 26 Sep Early start, heavy rain 22 Sep during harvest National weather* Dry year, excellent crops. Good grain harvest weather – some storms – result variable. Wet year, average crops, somewhat damaged. Fine grain harvest, badly laid and mildewed Wet summer, poor harvest. Fine grain harvest – some damage Showery summer, poor harvest. Poor seed time and low yields Year of good harvests. Wheat excellent, wet later Very wet summer, crops suffered. Wheat and barley below average. (First grain from USA) Poor sowing weather, good at harvest time. Low yields but good grain Wet summer, poor yields. Wheat & barley below average Warm wet summer, Wheat crop above average – better on higher land. Poor autumn sowing for 79 Appendix 10 Year Local weather comments Local weather comments Avge date (Taunton) Avge date (Cumbria) 1879 Gloomy prospects, 8 Oct 3 Oct much damage to standing crops. Few cattle at Michlm. Fair, farmers still at harvest Local Weather comments (Holderness) Avge date Late start to harvest 25 Oct due to weather conditions Local weather comments (Broads) Avge date Marshes flooded in Aug. 8 Oct Some late imp. Late harvest, worst for 40 years 1880 Early start to harvest 29 Sep Brought to standstill 11 Oct due to weeks of rain Good progress, mildew 1881 Continued rain 20 Sep delaying harvest. 1.9” on 22 Aug. Nine days rain in late Sept. Very late harvest, 23 Oct weather delays throughout Heavy rain. 12 Oct Late commencement, low lying country flooded Light lands good harvest 1882 Late finish, large 24 Sep amount of corn still in ground mid Sept. Gloomy prospects 12 Oct – a full weeks rain at end of Aug, another the following week Great damage by 11 Oct storm in north of county 25 Sep Nov 25 - still 11 Oct harvesting in some areas, crops being laid down Wheat below some blight Very late harvest Cold North east 8 Oct winds in August and lack of sun have delayed ripening of all grain crops in the area, delaying harvest. 348 22 Sep 1883 1884 Rain at end Aug after 19 Sep 4 weeks good weather. Aug rain 0.9” 1885 Later than normal 23 Sep start to harvest Appendix 7 348 7 Oct Greatly delayed 11 Oct harvest since weather broke Very dry – only 14” 19 Sep rain this year to the end of August. some 25 Sep 23 Sep average, 28 Sep 26 Sep 20 Sep Very dry July (lowest 26 Sep temp 32º) National weather* Unusually wet, bad harvests, cold winter, poor sowing, backward spring, very wet JunAug. In Aug – pastures flooded like winter. Harvest very late (First refrigerated beef from Australia) Good weather year but yields light, backward spring. Heavy barley crop Wet summer, moderate harvest. Worst snow of C19.Hard frost Jun 9th. Frost Jul 28th cut down beans. Below average crop and damaged. Wet year, deficient harvests. Frosts mid June and September. Crops poor – little autumn sowing. Unsettled, crops below average. Harvest stormy bur some good harvesting days Warm summer, good harvests Good harvest weather, good crops and yields Dry year, harvests below average. Mild winter but backward spring. Worst root harvest in many years Year Local weather comments (Taunton) Avge date 1886 Unusually late but 22 Sep catching up. Big storm 10 Sept. Appendix 10 1887 End Aug rain 15 Sep interfered with Corn 1888 Disrupted by tempests 4 Oct 8.1” rain in Sept 257 1889 Average harvest, good 19 Sep root crops. 1890 Rain and wind 23 Sep interfered with harvest. Catching up 349 1891 Early start, much 1 Oct beaten down by incessant rain. Late finish 1892 90% crop forecast. 23 Sep Rain early Sept stopped harvest Appendix 7 Local weather comments Local Weather comments (Cumbria) Avge date (Holderness) Avge date 10 Oct The rain spate was 10 Oct unparalleled, resulting in widespread flooding on already sodden fields 24 Sep 20 Sep Slow progress, 9 Oct heavy rain prevented cutting and flattened crops Hard frosts 22/9/89 4 Oct Severe 7 Oct thunderstorms for several days at end of August Rainfall in August 9 Oct 2.5 times average. Crops severely damaged and flattened Started under poor 12 Oct weather, floods in north of county. Sept wild and rough, storms and gales 349 16 Oct Local weather comments (Broads) Avge date Harvest progressing well, 27 Sep little damage to standing crop. National weather* Several stack fires Very dry year, good harvests (only 13” rain in Fens) 16 Oct Backward spring. Heavy crops Wet cool year, poor harvests. Wet summer, reasonable harvest weather. 23 Sep Ripening at last, little 3 Oct mildew, exceptional year for weeds. Most unfavourable 4 Oct weather, 31 days rain in last 2 months Prolonged wet 5 Oct weather Wet Sept – lowest temp 24 Sep 29º Storms in parts of E 7 Oct Yorks have severely damaged crops Busy harvest when 30 Sep weather permits – little mildew. Later than avge Rain at end of August 13 Oct have delayed reaping Late start due to rain 28 Sep 28 Sep Changeable year, crops above average. Late and cold springmuch winter corn killed Average yields but some blight. Showery year, average harvests. Good sowing weather - showers throughout harvest Showery year, crops below average. Damaging frost 30th June, wet summer, good harvest weather Wet summer but good crops. Thames frozen in Jan. Backward spring. Good but some late harvests Dry year – poor disastrous harvest. Late dry spring. Heavy rain in Aug, good harvest month Year Local weather (Taunton) comments Local weather comments Local Weather comments Avge date (Holderness) Avge date Avge date (Cumbria) Appendix 10 1893 Unsettled weather 18 Sep (end Aug) slowed harvest Reasonable harvest 24 Sep in spite of rain 1894 Poor year. Needs 22 Sep good weather for avge crop. Some sprouting 1895 Good harvest weather 24 Sep Wheat finished 7 Sept Thunderstorm 14 Sep damaged barley 1896 Early start, some 12 Sep delayed by rain 350 1897 Rain, gales. Gloomy 16 Sep delayed harvest. Only 19.7” rain in year, longterm avge 29.6” 22 Sep Local weather comments (Broads) Avge date National weather* Very dry year Drought year, crops below average. Fine spring very dry April (50 day drought in Weald) Poor grain crop Showery year, harvest above average. Mixed grain yields and quality Hot dry year, deficient harvest. Thames blocked by ice in Feb. Very hot, dry summer with storms. Wheat much destroyed by cold winter, late sown barley failed to ripen Dry year wet harvest, but crops good. Very dry May,Jun,Jul. Very wet Sept, but wheat good 26 Sep 4 Oct 7 Oct 1 Oct 1 Oct Wet stormy Aug 26 Sep Very dry Aug, harvest 21 Sep earlier than usual 7 Oct Very dry July, Aug. Early 22 Sep start and finish to harvest Fine through most 16 Sep of harvest-time. Early in many places Good early 29 Sep harvesting, poor later 1898 Harvest never so 16 Sep bountiful. Over early in West –fine weather 1 Oct 1899 Early finish. Articles 15 Sep about reduced rain – is it a permanent problem? 22 Sep 29 Sep 25 Sep 30 Sep 23 Sep Only 0.7” rain in 20 Sep August c/w 3” average Dry June, July. Early 23 Sep harvest finish * from Stratton J. & Houghton-Brown J., Agricultural Records, AD 220 – 1977, 2nd edition 1978, London: John Baker Appendix 7 350 Dry until harvest, then stormy. Wet spring, vegetation forward – hard frost 12th Jun. Wheat below average, but good quality Hot dry summer good crops. Cool and damp spring. Good harvest above avge, though some damaged in May Hot dry summer, average crops. Mild winter, backward spring, hot dry harvest. One of earliest harvests on record APPENDIX 11 – Norfolk Monastic sites – alignment A total of 96 monastic sites can be identified in Norfolk – 40 Norman sites and 56 postNorman sites, from a combination of Pestell’s Landscapes of Monastic Foundation (2004), for the Norman period monasteries and Messent’s The Monastic Remains of Norfolk and Suffolk (1934) for the post-Norman monasteries. The church element of the monastery could be identified in fifty-two of them, either as visible remains from the first- or second-series large-scale Ordnance Survey maps, or as sub-surface foundations through aerial photography, by parch- or crop-marks, using the 1940s “RAF” series held as part of the Norfolk HER at Gressenhall. The alignment of the nave was measured directly from the map where visible, but the alignment was measured from the aerial photograph in every case. In those cases where comparisons of the readings were possible, the figures were, with one exception, within 2º of each other. To ensure comparability, the readings from the aerial photographs were used throughout this analysis. Two point five degrees was subtracted from the measured readings, as an average figure for the County, to adjust the figures to True north from the Grid-north based measurements that the aerial photographs used. Overall, the alignment figures are similar to the rural parish churches in the larger parish church survey. The mean alignment is just 1.8º lower (87.1º to 88.9º); and the proportion of churches aligned to the north of east is very similar, at 54%, compared with 55.8% for the medieval parish churches. Comparing the monastic results across the county with the east-west pattern of changes in mean alignment already observed in parish churches both across the country and within Norfolk, shows that the same pattern does not exist within the monastic sample. Although the sample is small, the mean alignments west and east of the longitude-divide are very close and the ranges of alignments at the 95% confidence level overlap almost completely. Emphasising this is the fact that similar proportions of the alignments are to the north of east – 53% in the west of the county and 55% in the east (compared with 62% and 50% respectively amongst parish churches). So although the overall mean alignment figures fit well with the rural parish church results as a whole, the lack of difference in the proportions of monastic churches aligned to the north of east across the county confirms that the same spatial pattern in alignment found in parish churches does not exist within Norfolk’s monastic churches. Table A11.1 - Summary alignments of Monastic sites in Norfolk by longitude Monastic sites Norfolk – west of 1.10° East Norfolk – east of 1.11° East Total Appendix 11 No 19 33 52 Range 63.599.5 71.5110.5 Mean 95% conf Range at 95% %North of East 86.7 ±4.4 82.3-91.1 53 87.3 ±3.7 83.6-101.0 55 87.1 ±2.8 84.3-89.9 54 351 APPENDIX 12 - Using the survey results to assess other possible factors linked with sloping church sites It was established earlier in this thesis that although there was a significant difference in the numbers of slope-built churches built on east-facing slopes, there was little difference in the alignment of the churches, whether they were built on slopes, flat sites or on platforms. The aim of this section is to see if there is a difference in other factors, in addition to the slope direction, particularly the close presence of the big house (the manor house or lordly hall), especially those built on east-facing slopes, which could be part of a form of Christian substitution outlined earlier, rather than merely a lordly foundation of a church next to the house. As was noted earlier, there are more than twice as many churches sited on eastfacing slopes than would be expected in a random sample – 201 (54% of all on sloping sites) compared with a figure of 93, which is 25% of the total of 373 which would result from an equally distributed sample – an apparent excess of 108 churches. Churches on platformed sites do not exhibit this same degree of bias – with 90 (32%) on east facing slopes, an excess of only nineteen over the 71 which an equally distributed sample (25% of 283) would produce. Are there two concurrent patterns in these figures? Does this ‘excess’ of 108 churches, represent churches which were built in earlier graveyards or built to incorporate earlier ritual sites which have no other obvious remnants now, whilst the remainder of the east-facing group, along with slopes in the other three directions, are lordly private foundations that follow the random pattern of landscape slope which might be expected? Table A12.1 Churches on sloping and platformed sites SLOPING SITES No % East facing 201 54 South 61 16 West 85 23 North 26 7 Total sloping sites 373 100 PLATFORMED East facing South West North Total platformed 90 101 57 35 283 32 36 20 12 100 The similarity in the number of churches on south facing slopes with the numbers of churches on west-facing slopes might support this assertion and also serves to highlight how few churches are sited on slopes which face north. A north-facing slope would be the last choice of location for an agricultural community in a land where some growing years are marginal – shown by harvest figures earlier in Chapter Six, when harvests were not always entirely gathered in, and weather comments were made specifically referring to cold summer seasons delaying ripening and therefore harvest. From an agricultural point of view when the warmth of the land and its speed of warming up are important, if the land in an area generally sloped in a single direction, then south would be first choice, followed by West, then East then North. This pattern is almost mirrored by churches built on platformed sites, except for a small Appendix 12 352 excess on east-facing slopes, with 36% south facing, 20% west facing, 32% east facing and only 12% facing north. However, for churches built directly on sloping sites, the equivalent figures are 16% south, 23% west, 54% east and 7% north facing. There are two possibilities – firstly, as outlined above, that churches built on platformed sites represent the norm – a fairly balanced selection of sites on slopes of all directions, and that the excess of churches built directly on east-facing sloping sites represents the use of earlier graveyard or ritual sites, previously unknown; or secondly, that all platformed sites are part of the later phase of lordly church building (late-Saxon or Norman) and are built on new sites, rather than in an earlier graveyard, and show that for later churches the east-facing slopes of mid-Saxon times were no longer considered important for a church site. As table 12.2 below shows, there is little difference between the ages of the current churches, whether built on slopes or platforms, and it is not possible to examine the original churches on these sites, or establish whether there is an earlier pre-church graveyard beneath a church, without thorough excavation. Table A12.2 Churches on sloping and platformed sites by age of earliest fabric SLOPING SITES No % th th 11 , 12 century 185 50 13th – 15th century 118 32 Not known 70 19 Total sloping sites 373 100 PLATFORMED 11th, 12th century 13th – 15th century Not known Total platformed sites 124 109 50 283 44 39 18 100 Detailed comparison of sites where church and hall are together, with those where there is no hall Overall, as shown in table 12.3, Norfolk has fewer churches next to the hall (at 13%) than the rest of survey area (at 21% of all churches) and Norfolk would need almost to double the number of church/hall focus sites (an extra 50 churches, making 120 in all) to achieve the same proportion as the rest of the survey sample, so the difference is substantial. However, the numbers become much smaller in some cases when the Norfolk results are broken down into the topographic groups, despite the fact that Norfolk provides between one quarter and one-third of all the churches in the whole survey (29%), but there is still a difference between Norfolk and the remainder (NonNorfolk) in every category of yard topography, where the proportion of churches next to the manor house is lower in Norfolk, whether the yards are sloping, flat, on a knoll or platformed. Is this smaller number of church/hall groups in Norfolk due to a higher proportion of field churches in Norfolk, built by groups of freemen where no manor was involved (Warner 1986, 43; Williamson 2006, 89); does it reflect changes in the manor system in Norfolk, which saw many small manors amalgamated with the loss of many manor houses over the centuries (Williamson 1993, 164; Barnes 1997); or is it due to a greater number of middle-Saxon sites that predated the manor system, where the church was built later but in an existing early graveyard rather than next to manor Appendix 12 353 house? Whatever the cause, there is a slightly higher proportion of churches next to halls on sloping sites (27%) than in any of the other possible yard topographies. Table A12.3 Church/Hall focus by Topography of churchyard by area Non- Norfolk Topography Focus Sloping yards Norfolk 80 Not % with focus focus 215 27% Flat/almost flat 148 628 Knoll 12 Platformed yards Total Focus All Survey areas 13 Not % with focus focus 65 17% Focus 93 Not % with focus focus 280 25% 19% 51 325 14% 199 953 17% 79 13% 2 25 7% 14 104 12% 46 169 21% 5 63 7% 51 232 18% 286 1091 21% 71 478 13% 357 1569 18.5% Church/hall focus may be more likely on sloping sites, but it does seem to matter in which direction the land slopes. There is no difference between east-facing slopes (26%) and slopes in other directions (28%), when all the survey areas outside Norfolk are taken together, as shown in table 12.4. Although this is not the case in Norfolk, where the numbers become very small when subdivided - if only three churches changed from an east-facing to an ‘other’ slope, then both would be the same percentage. It is safe to say that there is no connection between the direction of the slope and the proximity of the church and the manor. Table A12.4 Church/Hall focus by Direction of slope of yard by area Non- Norfolk Topography and slope direction Sloping Yards - East - other directions Total Platformed yards East - other directions Total Focus 40 40 80 14 32 46 Norfolk Not % with focus focus 114 26% 101 28% 215 27% 52 117 169 21% 21% 21% Focus 11 2 13 1 4 5 All Survey areas Not % with Focus focus 36 23% 29 6% 65 17% 23 40 63 5% 9% 7% Focus 51 42 93 Not % with focus focus 150 25% 130 24% 280 25% 15 36 51 75 157 232 17% 19% 18% Examining the church/hall location in relation to its proximity to the location of the village, shows that, again, there is little difference between the groups, except for isolated churches, which at 27% are slightly more likely to be part of a church/hall focus in areas outside Norfolk, although Norfolk itself still has smaller proportions of churches in each of the three categories than do the Non-Norfolk sites, as shown in table 12.5. Table A12.5 Church/Hall focus by Location of church in relation to village by area Non- Norfolk Norfolk All Survey areas Focus Not % with Focus Not % with Focus Not % with focus 18% focus focus focus focus 16% focus Church in village 129 608 25 211 154 819 11% Church isolated 76 203 32 141 108 344 27% 18% 24% Church at vill edge 81 280 14 126 95 406 22% 10% 19% Total 286 1091 71 478 357 1569 18.5% 21% 13% Appendix 12 354 Analysing the church/hall location by the age of the earliest church fabric also shows little difference between the categories, although the churches with earlier fabric (eleventh- and twelfth-century) have slightly higher proportions next to the hall at 21 24%, than the later churches of the 14th and 15th centuries at around 15%, as shown in table A12.6. Again, Norfolk has a lower proportion in every single fabric age category than the remainder of the survey, between 9% and 19%, compared with the range 15% and 24% for the churches in other areas. Table A12.6 Church/Hall focus by Age of church fabric by area Non- Norfolk Norfolk Oldest fabric Focus Not % with Focus Not % with focus focus focus focus 11th century 11 42 6 25 21% 19% th 12 century 134 432 23 141 24% 14% 13th century 61 242 19 103 20% 16% th 14 century 20 103 14 127 16% 10% 15th century 8 45 3 31 15% 9% Not known 52 227 6 51 19% 11% Total 286 1091 71 478 21% 13% All Survey areas Focus Not % with focus 20% focus 17 67 157 573 22% 80 345 19% 34 230 13% 11 76 13% 58 278 17% 357 1569 18.5% There are similar proportions of churches next to the hall in each of the categories of name origin, whether the church is in Norfolk (at around 13%) or outside Norfolk, at around 21% - shown in Table 12.7. When these last two analyses, by age of fabric and by name origin, are further split into slope direction, there is little variation between each of the categories, whether built directly on the slope or built on a platform, so the tables were shown in Appendix 8. Table A12.7 Church/Hall focus by name origin of village by area Non- Norfolk Norfolk O. E. Name O. E. Process/feature O. N. Name O. N. Process/feature Other Total All Survey areas Focus Not % with focus focus Focus Not % with focus focus Focus Not % with focus focus 69 159 13 9 36 286 267 519 64 53 188 1091 29 27 8 6 1 71 165 233 47 27 6 478 98 186 21 15 37 357 432 752 111 80 194 1569 21% 23% 17% 15% 16% 21% 15% 11% 17% 18% 14% 13% 18% 20% 16% 16% 16% 18.5 % Churches with church/hall focus have been shown to be more likely to have been built on a slope (table 12.3) as well as being in an isolated location (table 12.5). Table 12.8 below shows that this applies whether they are built on east facing slopes or on a slope in another direction (43% on east-facing slopes and 37% on other slopes). The figures are similar for other village locations too, whether on east-facing- or other slopes. At this level of disaggregation, the Norfolk figures are too small to be meaningful, as a change between categories of a single church alters the balance of the figures substantially. Appendix 12 355 Table A12.8 - Church/Hall focus on sloping sites by direction of slope, location of village by area Non- Norfolk Sloping sites Slope East – in village - isolated - village edge Slope other – in village - isolated - village edge Total all sloping sites Norfolk All Survey areas Focus Not % with focus focus Focus Not % with focus focus 16 12 11 16 14 11 80 70 16 28 46 24 31 215 5 3 3 1 1 13 16 11 9 12 11 6 65 19% 43% 28% 26% 37% 31% 27% 24% 21% 25% 8% 14% 17% Focus 21 15 14 16 15 12 93 Not % with focus focus 86 27 37 58 35 37 280 20% 36% 27% 22% 30% 24% 25% The key fact throughout this section is that the numbers of churches on east-facing slopes, whether with church hall focus or not, and independent of where they are located in relation to the village, outnumber the churches on all the other slope directions put together, indicating that the direction of the slope was the principal concern and that none of these other factors examined here appear to have anything like the same level of connection. For example, 39 churches have church/hall focus on east facing slopes, compared with a total of 41 on the three other slope directions added together; 114 do not have church/hall focus on east slopes compared with 111 on all other slopes. All of the other factors considered - the age of the church fabric, the village name origin, the position of the church in relation to both the village and the manor house, show little variation when analysed by the direction of the slope, which firmly places the direction of the slope as the key factor when selecting the site. Appendix 12 356 APPENDIX 13 - Technical details of Norfolk Topography Analysis Computer-based topographical assessment of Norfolk - Methodology Digital (vector) county and parish boundary data for Norfolk, for the year 1851, were obtained from the UKBorders website via the Athens website under the Combined Higher Education Software Team (CHEST) agreement. Nineteenth century parish boundaries were used in preference to modern ones as they are more consistent with early boundaries. Digital (vector) contour data for Norfolk were obtained from the Ordnance Survey at a scale of 1: 50,000 via the Athens website – Panorama Digital Elevation Model (PDEM). The county and parish vector data were imported into ArcGIS and converted into raster layers with a resolution of 50m. The contour (vector) data was imported into ArcGIS and rendered into a Triangular Irregular Network (TIN) – ‘a wire network’ for ground slope and aspect using the 3D analysis function (create/modify TIN from features). Both the ground slope TIN and aspect TIN were subsequently converted into raster layers with a resolution of 50m using the 3D analysis function (convert TIN to raster). Land was categorized as ‘flat’ if the slope was below 2 degrees and hence the ground slope was divided accordingly. This was achieved using the 3D analysis function (reclassify). A regular grid of 59,743 points at 300m spacing was generated using Hawths tools within ArcGIS (sampling tools – generate regular points). For each of the 59,743 points in Norfolk, the ground slope (above or below 2 degrees), aspect and parish name were added using the sample function (spatial analysis tools – extraction – sample points over raster layers) and a Microsoft Excel data table was produced. Using basic data sort routines, simple ‘sum’ routines, etc, within Excel, the percentage of landscape (over 2 degrees ground slope) within each parish in Norfolk for each of the eight directional slope aspects was determined. To check the findings in greater detail, within the restriction of 64,000 sampled data points, the Norfolk parish boundary (vector) data was reduced in number to around 50 parishes at random using the Editor (delete) function in ArcGIS. The above analysis process was then repeated but the resolution of the sampling points was 100m (to fit within the 64,000 sampling limit of the program), which produced an average of about 600 – 700 data points within each parish. Detailed text and table omitted from the main text Parishes where the church is built on a flat or almost flat site have a generally flatter topography; with approximately 70% of the land in the parish having a slope of less than two per cent, very slightly above the average for the whole county (shown in table 7.3). In these parishes the churches are not built on slopes, but even here what sloping land there is slopes roughly equally in all directions. Table A13.1 – Topography of parishes which have churches built on flat sites Parishes with Proportion Proportion of 2%+ slopes by direction of slope churches on sites N S S N which are:No. <2% slope North East East East South West West West Flat/Almost Flat 376 69.9 4.3 4.1 3.7 4.2 3.8 3.4 3.0 3.6 Knoll sites 27 71.2 2.9 2.9 2.0 2.6 3.4 4.4 4.1 2.8 All Norfolk Sites 549 67.0 4.5 4.3 3.9 4.4 4.3 3.8 3.8 3.6 Appendix 13 357 APPENDIX 14 - Case study - abandoned churches in Norfolk: was the choice affected by sloping sites In the majority of these settlements with two or more churches, one church fell into disuse, and in most cases has been abandoned completely. Was the decision as to which of the churches to retain, and which to abandon, influenced by the topography of the two sites involved? Was a church that was sited on an east facing slope favoured for retention over a church located differently? There are seventy-nine parishes in Norfolk where there were two or more medieval churches, identified by Batcock (1991, 186), at least one of which fell into disuse and either remains as ruins of varying completeness or has disappeared completely. To assess whether the church site which disappeared, particularly its slope, was different from the one that remains, each of these sites has been visited and measurements of slope and horizon elevation made. Fourteen of the seventy-nine settlements had two churches either in the same yard or adjacent yards - at Antingham, Barnham Broome, Bedingham, Blo Norton, East Carleton, Gillingham, Great Melton, Little Snoring, Rockland, South Walsham, Snetterton, Stiffkey, West Dereham, and Wicklewood. In addition, the churchyard in Reepham, which still contains the parish churches of St Mary, Reepham and St Michael, Whitwell, used to contain All Saints, Hackford as well, prior to the fire in 1549 (Batcock 1990, 22), illustrated as Figure 5.2 earlier. Despite being built in the same (very small) churchyard, the churches were actually sited in three different parishes, and the parish boundaries still meet there. The arable lands of the two South Walsham parishes, illustrated in the frontispiece, were intermingled until quite late and the current parish boundary divides the churchyard (Williamson 1993, 158). None of these fourteen sites has been included as part of this case study as two churches built in the same, or adjacent yards will obviously experience the same topography, so in these cases there can have been no landscape-based influence on the selection of the church that is still in use over the one that was abandoned. Fifty-eight of the remaining sixty-five settlements which had two or more churches located in different parts of the parish have been surveyed and the results presented below. The final seven, where the specific site of the lost church cannot be identified (Batcock 1991, 53-55), have not been surveyed as the site-specific landscape was not able to be assessed accurately. In addition to these sites, there are fifteen settlements that still retain both churches in use, but the parishes that contained them have been subdivided. In many cases these are the larger parishes in the west of the county. This gives rise to the peculiarly Norfolk and Suffolk pattern of naming the parish by adding the church dedication to the settlement name, such as Terrington St Clement and St John, Weasenham All Saints and St Peter, and Walpole St Andrew and St Peter. In some cases the parish was split and renamed, such as at Wheatacre, which became Wheatacre and Burgh St Peter (Batcock 1990, 10). Since all these churches are still in use or have only recently been declared redundant, they have not been included in this analysis. Topography The results here indicate strongly that topography played no part in the decision as to which of the two churches to retain. The sites of those churches that disappeared are topographically very similar to the 549 churches that still exist in Norfolk and form part of the main survey. Forty-one (71%) of the lost churches were located on flat, or Appendix 14 358 almost-flat sites, compared with 68% of the existing churches in the Norfolk survey; four (7%) of the lost churches were located on east facing slopes, compared with 8% in the whole of the county and thirteen (22%) were located on other slopes, compared with 24% of those in the whole county. Thirty-eight of the 58 settlements surveyed had both churches located on sites with the same topography. As with the cases of two churches in the same yard, topography obviously was not a factor in the decision as to which of these pairs of churches was retained. In 29 of the 38 cases, both churches were on flat sites, in five cases they were both built on almost-flat sites and in four cases they were both built on sites sloping in the same direction – one east-facing, one southeast-, one south- and one west-facing. Seven sites were on similar topography – four cases where the church on the flat site remained at the expense of the other on an almost-flat site, and three cases the other way round. Table A14.1 - Topography Analysis – One settlement : two churches (same or similar topography at both sites) 50150350550750Distance apart 100m 300 500 700 900 Same topography Both sites flat 3 10 2 4 4 Both sites almost flat 1 1 1 Both sites sloping 3 1 (in same direction Similar topography Flat site retained – 1 2 almost flat site lost Almost flat site retained 1 1 –flat site lost 5 13 5 7 5 950+ total 6 2 - 29 5 4 1 4 1 3 10 45 Lastly, there were thirteen settlements where the two churches were built on sites with very different topography. The numbers involved are small and there is no pattern of retention or abandonment of the churches built on the slope. Three involve a church on an east-facing slope; one was retained at the expense of a flat site (Fincham, lost in the eighteenth century (Batcock 1991, 158)); and two were lost, one in favour of a flat site (Burnham Thorpe, fourteenth century (Batcock 1991, 53)) and one in favour of an almost-flat site (Swainsthorpe, sixteenth century (Batcock 1991, 54)). The remaining ten sites have slopes other than east-facing; in seven cases the church on the sloping sites was retained, and in three cases the church on the sloping site was the one that was lost. Table A14.2 - Topography Analysis – One settlement : two churches (sites with different topography, involving East facing slopes) 50150350-500 550-700 750Location Distance apart 100m 300 900 (& when lost) East facing sloping site retained E platform retained – Fincham 1 flat site lost (lost in 18th C) East facing sloping site lost East slope LOST – 1 Burnham Thorpe almost flat site retain (lost in 14th C) East slope LOST – 1 Swainsthorpe flat site retained (lost in 16th C) Appendix 14 359 Table A14.3 - Topography Analysis – One settlement : two churches (sites with different topography, involving OTHER slopes) 50150350550-700 750950+ Distance apart / 100m 300 500 900 Site type retained SE-facing slope retain. 1 al. flat 1 flat N facing slope retained 1 S slope 1 flat W-facing slope retain. 1 W plat 1 S slope 1 flat Knoll retained 1 al. flat 1 al. flat, 1 flat 1 3 5 1 2 1 TOTAL tables 14.2 & 14.3 Total 2 2 3 3 13 Period of /reason for disappearance Topography apparently did not affect the decision to abandon one church in a settlement in favour of another, but the figures in table 14.4 below, show that almost half of all the lost churches were abandoned in the sixteenth century, facilitated by the Act of 1535/6 permitting the consolidation of parishes. The religious upheaval during this time contributed to either reductions in, or cessation of, maintenance of church fabric generally (Duffy 1992, 424-428). In several instances, the fabric of one of the two churches was used to repair the other; Batcock recorded such a plea to the Diocese in Norfolk for the church at Guist (Batcock 1991, 11). In addition, some churches where located in manors where the Lord was a religious institution affected by the Dissolution, such as at Kirby Bedon. Secular abandonments were also growing at this time in East Anglia as parkland was being established, for example at Wolterton and West Wretham (Batcock 1991, 12), and there was also continued growth in the development of large sheep runs, such as at Godwick (Healy 1982, 59-67), Bawsey and Leziate (Allison 1955, 136). Table A14.4 - One settlement : two churches - Period of disappearance Century Unkn 12th 13th 14th 15th 16th 17th Same Yard 1 3 1 5 1 Other sites - assessed 1 2 3 5 29 5 - not locatable 1 3 1 2 TOTAL 1 4 3 7 6 36 6 18th 2 8 10 19th 1 1 20th 5 5 total 14 58 7 79 Proximity must have had an effect on church abandonments. If two churches were close, there would have been no reason to maintain both after the historic imperative of attending the church of your manor had disappeared. In addition to the fourteen churches in the same yard, 39 of these lost churches were within 650 metres of the other church in the settlement, and only eleven were more than one kilometre from the other church. Even in these days of increased car-use, these distances would not be considered too far to walk to church, in medieval times such distances would not have been considered an issue at all. Case Study Conclusion The slope of the site appears not to have featured in the decision of which of the two churches to retain in the cases of these Norfolk abandonments, with very few churches sited on east-facing slopes involved at all. Of the four churches on east-facing slopes, two were retained at the expense of the other church in the settlement, and two were Appendix 14 360 lost. During the period of the majority of the abandonments, it appears that pragmatic decisions relating to relative state of repair of the two buildings, and the proximity of the churches to the remaining settlement were more important. Whether this was the case for those (few) abandonments in the twelfth and thirteenth centuries, when slope may have been considered, is more difficult to determine, as half of these sites cannot be identified closely enough on the ground to be able to make an assessment of the topography at the church. Appendix 14 361 APPENDIX 15 - Site-specific landscape assessment of all churches in Norfolk built on sloping land The landscape context of the sites of all the churches that were built directly on slopes, and those built on levelled platforms on slopes in Norfolk has been assessed with a view to establishing if it was possible to determine whether specific churches were built where they are for a particular reason or not. Norfolk is a good choice for this exercise as it exhibits similar patterns of slope-built churches to almost all the other counties, despite the fact that the topography is fairly gentle, and secondly it is close to the author’s home. The reason for the selection of a specific church site could have been one or more of many - to use the slope and its extended horizon; to incorporate a site of earlier importance; the close proximity to the manor or the proximity to another feature, such as a spring or a river-crossing. Consideration was also given to whether the church was built in such a place as to ensure that it could be seen from a great distance, which also reflects the fact that there is an extended eastern view from the church. Norfolk In Norfolk, there were 47 churches on east-facing sloping sites, nine on south-facing slope, nineteen on west-facing slopes and three on north-facing slopes. On platformed sites, there were 24 on east-facing slopes, 23 on south-facing slopes, sixteen on westfacing slopes and five on north-facing slopes. Norfolk churches on east-facing slopes Of the 47 churches on east-facing sloping sites, thirteen (27%) appear to have been sited solely for the slope and the extended view it offers, in that there is no close manor now (and no evidence of one) and in these cases the site is apparently the best in the locality for its eastern view, as the remainder are either lower down the slope or have additional reasons for the chosen site, such as an adjacent Hall. The comments on their siting are as follows:Table A15.1 – Norfolk church sites on sloping land Name Grid Ref Comments Hunworth TG064355 Village currently along SW-NE valley on SW side of River Glaven. Church at top of steep hill – no sign of manor close by. Castle Hill (motte) 800m to SE. Claxton TG328032 Near top of slope down to R Yare marshes. 1 mile from Claxton castle and manor. Isolated, rest of village on marsh edge, next to castle East Tuddenham TG085115 West side of small steep-sided valley, to east of current village. Isolated, village moved to top of hill (common?) on flat site. No sign of manor. Colton TG104094 At top of slope above wide shallow valley. Village now to N, away from river, adjacent to common. No sign of manor. Brandon Parva TG070082 High on slope above village, which is located in valley bottom, opposite Barnham Broome on other side of R Yare. Manor Farm and Monk’s Hall 200m lower down slope. Most of the parish slopes the same way, but church is closest to hill top. Appendix 15 362 Honingham TG115113 On Norwich-Lynn road, 1 mile from village now located at river crossing. Church halfway up steep slope from river, Completely isolated, but visible from Honingham Hall, 1k to N across park, on other side of river valley. Wighton TF941399 Village split either side of crossing of N-S River Stiffkey. Not close to manor, but in best place for east view, and highly visible from the now silted up river-port and directly opposite Cley, its port and grand church, across the river estuary. Fring TF726348 Above village at top of steep-sided narrow valley (Figure 8.16). Village at valley bottom on X roads and river crossing. Church adjacent to and above large farm, but Fring Hall is on other side of river. Bexwell TF632034 At edge of small hamlet, on west edge of small tributary valley. Hall located 100m to east. If church was next to it, the slope would be in the opposite direction, i.e. on a west-facing slope. Stody TG056352 Village located at X roads, on spur at confluence of two tributaries of R Glaven. Church at the edge of the slope, on the only site in the village that could give such a lengthy east view and visibility from a distance. Graveyard on east-facing slope (Figure 8.15). Swanton Morley TG019173 Huge church at top of valley side of R Wensum (Figure 8.12). Village at X roads at bottom of steep-sided tributary valley to S. Extensive east views (Figure 8.13), visible for miles. Morley Castle with moats, sited 1k to east on the valley floor. Mundham TM325980 Originally, one of two churches in vill. This one better sited for slope, although it is next to parish boundary with Seething. Church is now far closer to Seething than Mundham. Excellent east views (Figure 8.17). The “Old Hall” is at bottom of slope. Pulham St. Mary TM197861 On west side of small N-S tributary of R Waveney. Church at E end of village, near river crossing and 1k W of Hall, which is even more remote from village and isolated. Village at top of hill, but much flatter topography In addition there are twelve churches (25%) located on an east-facing slope, but in parkland settings close to the ‘big house’. In these situations it is not possible to speculate as to which came first, the house or the church, therefore it cannot be determined whether the church was located close to the house for lordly reasons, rather than the sloping site being chosen first for other reasons. South Pickenham West Barsham East Raynham Runton Anmer Thorpe-nextHaddiscoe, Shelton Spixworth Appendix 15 TF856042 TF905337 TF879256 TG179428 TF737296 on the edge of Pickenham Hall Park. in the parkland of West Barsham Hall. in the parkland of Raynham Hall adjacent to a large house, now a hotel. in the parkland of Anmer Hall. TM437981 TM221910 TG240148 located 50m from Thorpe Hall. close to Shelton Hall. on edge of Spixworth Hall Park. 363 with a 2nd church, ruined, between it and both The Old Hall and The Manor House, both with parkland and 100m to the north. Letheringsett TG060389 adjacent to Letheringsett Hall Denton TM286874 next to Denton Hall West Lexham TF843172 isolated, on edge of Lexham Hall Park Apart from these churches, there were an additional fourteen (29%) built on east-facing slopes where the church could have been sited over a wide area, as all of the village, or much of the parish, sloped in a similar direction and therefore any specific selection reasons for a particular site are masked by the general availability of similar conditions – these were:Surlingham TG305065 Diss TM118800 Brampton TG220245 Caston TL960976 Burlingham TG372083 Briningham TG038344 Marsham TG196237 Hackford TG060024 (also close to a spring) Newton Flotman TM213984 (also close to the river crossing) Loddon TM364988 (also close to middle-Saxon pottery noted earlier) Thwaite TM333950 Shelfanger TM107837 Ovington TF924026 Bridgham TL957857 Garboldisham TM005816 There are also eight others on east-facing slopes with no specific features to indicate why the particular site was chosen. Norfolk churches on slopes in other directions Of the 31 churches built on slopes other than east-facing, four are located in parishes where much of the land has an eastern slope and therefore would have been better located for the view east if built elsewhere in the parish, so the site was definitely chosen for a reason other than a lengthy east view; three are built next to the big house and are probably sited for lordly reasons; three are built at the top of their respective slopes, so have the benefit of a long east view anyway. The remaining nineteen have no obvious reasons for their location. Table A15.2 - Summary of Norfolk sloping sites Slope-built churches Platform-built churches – slope main reason for site - additional possible reason for site (hall etc) - similar slope over larger area 13 8 12 6 14 2 no identifiable reason 8 8 East-facing slope Appendix 15 364 Other slopes - other specific reason for location - Rest of vill same slope 3 5 4 8 - No identifiable reason 19 24 - Sited away from vill with better east view - Sited away from places with east slope TOTAL 1 1 4 6 78 68 Appendix 15 365 Appendix 16 - Comparison of minster church sites with other church sites It had been intended to compare the details of minster church sites with those of churches lower down the hierarchy, in terms of location and site topography etc., to see if there may have been a different pattern of influences on site selection between the different classes of church. Unfortunately, there are two major problems – firstly, every writer on the subject has commented on how difficult it is to establish which churches should be classed as minsters, due mainly to the lack of surviving documentary evidence of the relationship between minster- and daughter-churches (Blair 2005, 319 & 465; Blair 1987 & 1992; Morris 1983; Franklin 1984; Hase 1988; Rushton 1999), but also to the many subsequent alterations and extensions to churches over the centuries which have blurred the difference between the generally larger ‘superior’ churches, and the generally smaller ‘inferior’ churches; and lastly, even when minster churches can be established, many of them are located in towns and therefore fall outside the parameters (of rural churches) for this survey, so the data required to compare many of those sites have not been collected. The tables below compare the 67 probable minster churches identified either in general studies by Blair (1987) and Morris (1983); or in specific county based studies by Franklin 1984 (The Midlands), Hase 1988 (Hampshire) and Rushton 1999 (Sussex), and which have been surveyed for this thesis, plus the possible 27 minster churches in Norfolk identified earlier in Chapter Six, with the remaining (non-minster) churches in this survey. Table A16.1 – Alignment of churches on probable or possible minster sites in all counties in this survey Number Range Mean 95% Range at % N of conf. 95% East Probable & Possible 94 57-116 ±2.2 84.5 – 88.9 66 86.7 minsters Non-minsters 1832 38-128 ±0.5 85.5 – 86.5 64 86.0 1926 64 86.1 There is no real difference between the alignment results for the probable/possible minster churches and the remainder of the churches in the survey. The mean alignment figures are close, the proportions aligned to the north of east are almost identical and the range of alignments at 95% confidence for the two groups overlaps completely, confirming that a significant difference in alignment is not possible, but this could still disguise the fact that their sites were chosen by a different process. Tables analysing several factors for the two groups of churches are shown below. Overall, minster churches are slightly more likely to be built on slopes than the remainder of churches – 24% to 19%, but both types of churches built on slopes are equally likely to be built on an east-facing slope - 52% of minsters and 54% of nonminsters. They are also equally likely to be built next to the big house (at around 20%), but minster churches are more likely to be located in the village (68% to 50%) and less likely to be remote from their settlement (11% to 24%). They are also equally likely to be in a village with an Old English derived name, at around 75%. Even though the figures in many cases are close, no real conclusions can be drawn from this analysis as the extremely small numbers in some of the categories of minster churches allow no Appendix 16 366 certainty whatsoever, and even a small variation in the figures would alter the proportions dramatically and therefore affect the conclusions. Table A16.2 - minster/non-minster churches by site topography ‘minster’ churches Other churches No. % No. % Sloping sites 23 350 24 19 Flat/almost flat 51 1101 54 60 Knoll 10 108 11 6 Platformed sites 10 273 11 15 TOTAL 94 100 1832 100 Table A16.3 minster/non-minster churches built on slopes by slope direction ‘minster’ churches Other churches No. % No. % Slope facing EAST 12 189 52 54 Slope facing WEST 8 77 35 22 Other slopes 3 84 12 24 TOTAL 23 100 350 100 Table A16.4 minster/non-minster churches by church/hall focus ‘minster’ churches Other churches No. % No. % No church/hall focus 75 1494 80 82 Church/hall focus 19 338 20 18 TOTAL 94 100 1832 100 Table A16.5 minster/non-minster churches by church location ‘minster’ churches Other churches No. % No. % In village 64 909 68 50 isolated 10 442 11 24 Village edge 20 481 21 26 TOTAL 94 100 1832 100 Table A16.6 minster/non-minster churches by village name origin ‘minster’ churches Other churches No. % No. % O.E name origin 70 1398 74 76 O.N name origin 10 217 11 12 Other 14 217 15 12 TOTAL 94 100 1832 100 Appendix 16 367 APPENDIX 17 - Anglo-Saxon finds in Norfolk Churchyards A catalogue of the Anglo-Saxon finds in churchyards discussed in Chapter Eight FIELD HER Name Early Middle Late Details DESCRIPTION The record number in the Norfolk HER. The name of the church. Whether any Early Saxon pottery or metalwork has been discovered. Whether any Middle Saxon pottery or metalwork has been discovered. Whether any Late Saxon pottery or metalwork has been discovered. Details of the finds and whether they were surface or excavated finds. HER Name 157 425 Norwich, St Benedict Norwich, St Michael at Plea Titchwell Warham, St Mary Barmer Syderstone Hindringham Walton, West Massingham, Little Massingham, Great Runcton Holme Southery Oxborough, St Mary Bilney, East Swanton Morley Guestwick Saxlingham Anmer Congham, All Saints Gayton Pentney Narford Lexham, West Newton-by-Castleacre Lexham, East Beeston-next-Mileham Dunham, Great Fransham, Great Shouldham, St Marga. Hilgay Barton Bendish, St Mary Barton Bendish, St Andrew Houghton-on-the-Hill Necton Threxton Pickenham, South Weeting Harling, Middle Harling, East Harling, West Blakeney Erpingham 1389 1853 1990 1991 2110 2210 2344 2345 2432 2590 2628 2828 3014 3131 3201 3513 3562 3770 3941 4015 4019 4053 4074 4093 4178 4206 4290 4453 4513 4514 4625 4642 4686 4717 5639 6033 6049 6051 6167 6720 Early Middle Pot Pot Pot Pot Pot Pot Pot Pot Pot Pot Pot Pot Pot Pot Pot Metal Pot Pot P&M Pot Excavated find. Metal Pot P&M Pot Pot Pot Pot Pot Pot Pot Pot Pot Pot Pot Pot Pot Pot P&M Pot Pot Pot Pot Pot Pot P&M Metal Pot Pot Pot Pot Pot 368 Excavated find. Excavated find. Pot Pot Pot Pot Pot Pot Details Surface find. Excavated find. Excavated find. Surface find. Excavated find. Excavated find. Surface find. Excavated find. Surface find. Surface find. Coins - Surface find. Excavated find. Surface find. Surface find. Surface find. Excavated find Excavated find Surface find. Brooches- Excavated find. Surface find. Surface find. Surface find. Surface find. Surface find Surface find. Surface find. Brooch. Surface find Cremation urn. Exc. find. Excavated find. Pot Pot Bold = parishes with middle-Saxon finds Appendix 17 Late Pot Pot Surface find. Surface find. Shears. Surface find. Coin. Surface find. Surface find. Excavated find. Surface find. Surface find. Excavated find. Surface find. APPENDIX 17 (cont) - Anglo-Saxon finds in Norfolk Churchyards A catalogue of the Anglo-Saxon finds in churchyards discussed in Chapter Eight FIELD HER Name Early Middle Late Details DESCRIPTION The record number in the Norfolk HER. The name of the church Whether any Early Saxon pottery or metalwork has been discovered. Whether any Middle Saxon pottery or metalwork has been discovered. Whether any Late Saxon pottery or metalwork has been discovered. Details of the finds and whether they were surface or excavated finds. HER Name 7120 7277 7297 7313 7471 7475 7583 7695 7912 8393 8457 8517 8523 8987 8989 9047 9064 9065 9067 9646 10072 10104 10115 10212 10265 10280 10464 10793 10913 11118 Hempton, St Andrew Longham Fransham, Little Tuddenham, North Reedham Witchingham, Little Felmingham Hautbois, Little Costessey Hickling Ludham Walsham, South Burlingham, North Rockland, St Peter Stow Bedon Hockham Breccles Shropham Snetterton, All Saints Thorpe St Andrew Wacton, Little Tasburgh Saxlingham Thorpe Bedingham, St Andrew Blofield Buckenham, Old Sisland Quidenham Roydon Earsham Early Middle pot Pot Pot Pot Pot Pot Pot Pot Pot Pot Metal Pot Pot Pot Pot Pot Pot Pot Pot Pot Pot Pot Pot Bold = parishes with middle-Saxon finds Appendix 17 Late 369 Pot Pot Pot Pot Pot Pot Pot Metal Pot Pot Pot Pot Details Surface find. Surface find. Surface find. Surface find. Surface find. Surface find. Spindle Whorls. Exc. find. Strapend. Surface find. Excavated find. Excavated find. Excavated find. Excavated find. Surface find. Surface find. Surface find. Surface find. Surface find. Excavated find. Surface find. Excavated find. Surface find. Excavated find. Excavated find. Excavated find. Iron knife. Surface find. Surface find. Surface find. Surface find. Surface find. Cremation urn. Exc. find.