The Scratch Dials of Kent

333 The Scratch Dials of Kent chris h.k. williams Scratch dials are primitive medieval sundials. Extant examples are almost exclusively confined to churches. As they were used to determine the time of devotional services, they are also known as mass dials. Scratch dials survive in a wide variety of forms, principally in the period of the day covered and the extent of the associated markings (Fig. 1) – some are very ‘busy’, others the exact opposite. Although most medieval churches no longer have a surviving scratch dial, many have more than one, some several. As well as examining the pattern of surviving scratch dials an assessment is made of their original prevalence and appearance. It will be appreciated that surviving scratch dials have endured both natural elements and church rebuilding for centuries. Nevertheless, it can be estimated that scratch dials were once widespread, most churches having multiple dials. It will also be seen that the use of scratch dials and their appearance, including their probable decoration, have both evolved in ways reflecting religious change and the availability and adoption of more sophisticated timekeepers. Scratch dials first aroused the interest of antiquarians and ecclesiologists during Victorian times. Kent does not appear to have participated in this initial pioneering stage.1 The first substantive study of scratch dials (based on an exhaustive survey of Somerset churches) was by Dom Ethelbert Horne, published in 1917.2 Another study, based on a partial survey of Hampshire churches by Green, appeared a decade later.3 In the 1930s Cole listed by county – including Kent – some 1,300 churches with scratch dials.4 Whilst by no means comprehensive or systematic in coverage, Cole established the widespread and extensive survival, and earlier use, of scratch dials. Over the last twenty years The British Sundial Society has recorded some 5,000 English scratch dials.5 Scratch dials have also been noted from France to Poland.6 Prior to 2006, in terms of published work, Kent enjoyed neither the complete Somerset and Rutland,7 nor the less extensive Hampshire and Herefordshire,8 recording of its scratch dial heritage. Furthermore Kent is not at the forefront of the British Sundial Society’s mass dial database in terms of coverage or detail. However information in the public domain CHRI S H.K. WILLIA MS 334 F ig. 1 Examples of Kentish Scratch Dials. Notes: (1) Line drawings by G.J.W. Winzar (see Table 1 note (e)). (2) Scale varies between dials. See Table 7 for dial sizes. (3) For relative frequency of dial survival see Tables 6 & 8. (4) Dials were painted, annotated and decorated (see main text). (5) For dating of when in use see Fig 4. THE SCRA TCH DIAL S OF KEN T 335 is a very poor reflection of both the study and research, as well as the extent, of Kent’s scratch dial heritage. Kent has been partially surveyed on several independent occasions. These range from listing only to full technical and photographic recording. Including hitherto unpublished work Kent is amongst the better surveyed counties. Moreover some of its recordings are of a quality and level of detail permitting statistical analysis not previously undertaken – placing Kent at the forefront of scratch dial research. The inspiration for this paper flows from Pat and (posthumously) Gerald Winzar. In the 1970s Gerald began recording Kent scratch dials with a degree of professional and technical rigour matching, if not exceeding, today’s best practice. Although failing health prevented Gerald realising his full ambition, his work in progress represents a contribution of national importance. His widow Pat has most generously made all Gerald’s work available to the author and agreed to donate his papers to the British Sundial Society archive. I nterpreting Scratch Dials To interpret scratch dials we must abandon the modern approach to timekeeping. In medieval times sunrise and sunset separated day and night – each being of 12 hours duration throughout the year.9 Hours varied in length both seasonally and diurnally.10 As summer approached day hours lengthened and night hours shortened (vice versa for the approach of winter). Only at the spring and autumn equinoxes would the modern (60 minute) equal hour pertain. Strange as such a system appears to us now, in an age with very limited artificial illumination and most economic activity geared to agriculture, such a simple system worked – indeed it had practical advantages. The Church (most emphatically in its monastic rules) stipulated the need for prayer/specific devotional duties at particular times.11 During the day these were mid morning (i.e. at the end of the third hour), midday (sixth hour) and mid afternoon (ninth hour) – Terce, Sext and None – and gave rise to the canonical hours. Terce had special ecclesiastical significance in being the accepted time for celebrating Mass. A scratch dial reflecting the medieval approach to time was (apparently) simple to construct. A horizontal stick, peg or rod gnomon12 inserted in a south facing wall would cast a shadow that moved through a semi-circle during the day. Subdivision of the semi-circle divides the day into its constituent parts. See Fig. 2 for a simple schematic scratch dial showing daytime hours, canonical hours and the mass line. The words ‘simple schematic’ are used as the Fig. 2 dial would only be strictly accurate – on its own contemporary medieval time terms – if on the equator and on an exactly east-west aligned wall. Adjustments would be necessary for any CHRI S H.K. WILLIA MS 336 misalignment of the wall and for latitude.13 Whilst such technicalities were known to astronomer mathematicians, it is most improbable those who made or used scratch dials became aware of such issues before the sixteenth century, when a number of treatises on the art of (sun)dialling were published.14 Although students of scratch dials have long recognised their association with the medieval time system,15 empirically based examination of the link was by example dials rather than statistical investigation – i.e. a selected few, rather than a large representative sample of all, dials. The medieval time system was eventually displaced by the modern equal hour. This was very much later than the invention of clocks and was a protracted and progressive affair;16 the two time systems co-existed for several centuries.17 The eventual triumph of the modern equal hour system depended on the spread and ownership of clocks/scientific sundials becoming sufficiently commonplace – a fact whose timing is both uncertain and variable geographically. Certainly by the eighteenth century Fig. 2 Simple Schematic Scratch Dial showing Medieval Time System and Canonical Hours. N ote: C onstruction involves: – inserting (at 90°) a horizontal rod gnomon in a south facing wall; – incising the midday noon line – established when the gnomon shadow is shortest (this is the only time of the medieval day in accordance with modern time keeping); and – positioning other lines at the required angle (from easily made templates). H our lines are 15° apart. THE SCRA TCH DIAL S OF KEN T 337 public turret clocks/scientific sundials and private clock ownership was not insignificant and rising rapidly. The scratch dials on Kent churches are some 300-800 years old.18 I ncidence of Surviving Scratch Dials Kent’s scratch dials have been at least partially surveyed on several occasions (Table 1). A combined listing by parish is at Annex. The first published listings were in the mid 1930s. Cole broadened the pioneering work of Horne and Green by extending listing onto a national scale – including Kent. Cole collaborated on a county basis and for Kent he cites Mr F.H. Worsfold and (unnamed) others;19 the latter may have included Mr George F.J. Rosenberg.20 With the benefit of subsequent work we can see the 1930s listings were not comprehensive. Their chief value today lies in providing a base against which to assess scratch dial loss during the twentieth century. TableABLE 1. KENT SCRATCH DIAL LISTINGS L isting Publ. R ecording Period N o. churches with dials N o. of dials C olea Yes Pre 1935 58 - Meeb Yes Pre 1936 10 - Wellandc Yes Pre 1979 104 - Groved Yes Pre 1979 13 28 Winzare N o 1970s 62 149 C hambersf N o 1990s 94 172 BSSg N o 1988-06 73 152 N otes: a T.W. Cole, Origins and Use of Church Scratch-Dials, 1935. b A a. Mee, The King’s England, 1936. c L l.G. Welland, ‘Scratch Dials in Kent’, Bygone Kent, 1979. d L l.R.A. Grove, ‘Scratch Dials in the Neighbourhood of Chislet’, in K.H. McIntosh, Chislet and Westbere, Villages of the Stour Lathe, 1979. e G.J.W. Winzar. ����������������������������������������������������������������Personalrecordandfieldbooks.DonatedtoTheBritishSundialPersonal record and field books. Donated to The British Sundial Society archive 2006/7. f R r.H. Chambers, 19 Sept 2000, letter to A.O. Wood. g The British Sundial Society records June 2006. The 1970s witnessed a revival of activity. Winzar began his detailed technical recording. Grove published a thorough, but very localised listing. Welland published what remains the largest single individual listing – a listing based on visiting 310 churches. The extent to which CHRI S H.K. WILLIA MS 338 he recorded, as opposed to listed, is not known.21 Neither is it known who assisted him in his searches.22 Interestingly Grove cites (in addition to Winzar and Welland) credits for Peter Lambert, R.J. Spain and Mrs C. Young. Clearly not all the detailed researches of the 1970s have yet found their way into the public domain. Hopefully these have not been irretrievably lost and await (re)discovery. During the 1990s Chambers began his as yet uncompleted photographic recording. The British Sundial Society has also received ad hoc reports and recorded individual dials and churches. Extensive and valuable as all this work is, it does not yet constitute a completed county recording – a task The British Sundial Society has completed for about a dozen counties. C ombining the individual listings, scratch dials have been noted on 135 separate Kentish churches (Table 2 and Fig. 3). Table 2 also shows the prevalence of multiple dials. Whilst at first sight such redundancy might appear odd, over the centuries there would have been several triggers TableABLE 2. KENTa Churches with Scratch Dial(s) Diocese C hurchesb with dial(s) Scratch dials per churchc 1 2 3 4 5 6 7 8 9 10+ R och. 28 19 6 2 1 - - - - - - C by 107 50 22 15 9 5 3 - 2 - 1 Total 135 69 28 17 10 5 3 - 2 - 1 N otes: a The pre-1832 historic county of Kent as mapped by the Institute of Heraldic and Genealogical Studies. See C.R. Humphrey-Smith, Atlas & Index of Parish Registers, 1984, 1995, 2003. b Source: Table 1 and Annex listings. c For each church the listing with the highest number of dials is taken. 13 churches have no quantified listing – these are included in 1 dial category.precipitating dial redundancy. A new dial might have been necessitated by the building of a porch or growth of trees; a new location might simply have become more convenient; a different type of dial might have been considered more appropriate (to be discussed later); or some priests might have wished to leave their own mark. Scratch dials are heavily concentrated in Canterbury diocese reflecting both a higher survival of medieval churches and a higher scratch dial likelihood per church (Table 3). Comparison with listings of those counties for which complete surveys of eligible churches have been published indicates that Kent is not atypical (Table 4).23 The similarity suggests that although Kent does not yet have a full recording of its dials, its listing is (virtually) complete.24 THE SCRA TCH DIAL S OF KEN T 339 Fig. 3 Kent Parish Churches with Scratch Dial(s), (Prepared by E. Connell). CHRI S H.K. WILLIA MS 340 TableABLE 4. COMPARISON OF KENT INCIDENCE OF SCRATCH DIALS w ith Somerset and Rutland (per cent) C ounty Scratch dials per church N o. of medieval churches 0 1 2 3 4 5 6+ Somerseta R och. dio.b C by dio.b R utlandc 68.8 71.1 54.3 45.1 19.6 19.6 21.4 27.4 7.9 6.2 9.4 13.7 2.2 2.1 6.4 5.9 1.1 1.0 3.8 3.9 0.2 - 2.1 2.0 0.2 - 2.6 2.0 465 97 234 51 N otes: a A author’s analysis of data in E. Horne, Primitive Sun Dials or Scratch Dials. Containing a list of those in Somerset, 1917. b F from Tables 2 and 3. c A author’s analysis of data in R. Ovens & S. Sleath, Time in Rutland. A History and Gazetteer of the Bells, Scratch Dials, Sundials and Clocks of Rutland, 2002. O riginal Prevalence of Scratch Dials Today we can only see the distribution of scratch dials surviving 300-800 years of church rebuilding and weathering. Although an obvious statement, its implications (and their estimation) are far from obvious, but of profound importance to a genuine appreciation of scratch dials. It might be thought the current high frequency of churches without a dial reflects the fact many never had one.25 Reflection concludes such a hypothesis is not credible. Scratch dials were a simple low cost device. Surviving examples alone are sufficiently numerous and geographically TableABLE 3. PROPORTION OF MEDIEvAL KENT ChuRCHES WITH Scratch Dial(s) N o. of churchesa No. churches with scratch dial(s)b R ecorded incidence (per cent) R och. dio. C by. dio. 97 234 28 107 28.9 45.7 Totalc 331 135 40.8 N otes: a Based on the author’s analysis of Pevsner (J. Newman, West Kent and the Weald, 1980 & North East and East Kent, 1983). b F from Table 2. c See Table 2, note (a). THE SCRA TCH DIAL S OF KEN T 341 dispersed to indicate know-how, capability and adoption must have been universal. It is difficult to conceive of a diffusion model that results in the surviving patchy dalmation distribution (Fig. 3). Replication of practice in the adjacent parish was too easy!26 Such considerations are reinforced by multiple dials per church. Might these indicate a degree of personalisation and ‘throw-away’ culture? Any tendency for non adoption by individuals would manifest itself in the number of dials per church, rather than their complete non adoption over half a millennium. Over the centuries churches have been subjected to continuing repair and rebuilding – climaxing with Victorian restoration. The impact on scratch dials is evidenced by inside the church, not on a south facing wall, inverted and/or sideways dials. Such obviously moved dials represent approximately 15 per cent of listed dials.27 The variable and scattered fate of moved but still visible dials suggests many, if not most, moved dials are now ‘invisible’ or lost – equivalent to a significant supplement to the number of listed dials. The difference between east and west Kent incidence of scratch dials may well in large part be due to the greater rebuilding, redevelopment or extension of churches as London is approached. Turning to the impact of weathering on scratch dials, the effects of repeated freeze-thaw cycles and atmospheric/precipitation pollution can be expected, if the fate of medieval carving and statues is a guide, to have been devastating. It is salutary to examine the condition of Ordnance Survey benchmarks – mainly cut in the third quarter of the nineteenth century – yet scratch dials are 3 to 6 times older and typically less deeply incised. About 20 per cent of listed dials are in a seriously distressed condition.28 These are merely the latest batch of dials to be on the verge of extinction – indeed approaching 10 per cent of the Kent dials listed in the 1930s have not subsequently been relisted (Annex). There can be no doubt the combined effect of church rebuilding and weathering has been substantial; it cannot however be independently or precisely quantified. The most illuminating way to progress matters is to calculate, within the context of the foregoing discussion, the implied original prevalence of scratch dials for a range of assumed scratch dial loss. Table 5 is based on the level of loss reconciling the existing distribution of surviving scratch dials with all churches originally having a dial: its implications are stark.29 Scratch dials were once common; scratch dials have, quite literally, been decimated. What Did Scratch Dials Look Like? This section and the next are based on Winzar’s detailed recordings. As these are of the highest quality,30 account for half of Kent’s churches with scratch dials, and are spread across the whole county, they are statistically representative. CHRI S H.K. WILLIA MS 342 Medieval churches were often limewashed externally. The scratched lines or pock holes facilitated dial reinstatement after limewashing or repainting if redecoration was otherwise called for.31 That scratch dials were painted has obvious implications for their original appearance – not all lines may have been scratched; painted and scratched line lengths might differ; dials might have been annotated or decoratively embellished. Surviving scratch dials are a colourless partial skeletal image of their original appearance. A threefold scratch dial categorisation has been examined32 (Table 6 and Fig. 1). Half of Kent’s dials are of the quarter, or ‘morning only’ type – a much higher proportion than in other counties.33 About one fifth of Kent’s dials span the night period in that they include lines above the sunrise-sunset horizontal. Table 6 is based on the appearance of surviving dials: as during painting it would have been an easy matter to extend daytime markings into the night, some surviving half dials may originally have been circle dials. Painting a circle dial from a scratched quarter dial appears an altogether less probable event. Dial size shows a not unexpected progression by dial type (Table 7), but TableABLE 5. COMPARISON OF KENT’s SurvivING AND ORIGINAL Scratch Dial Incidence Distribution of scratch dials A pprox. proportion of churches without scratch dial(s) Mode/median dials per church A pprox. proportion of churches with 6 or more dials Survivinga ½ - ⅔ 0 (all churches) 1 (churches with dials) N egligible O riginalb 0 4 ⅓ N otes: a F from Table 4. b A author’s estimates assuming: - all churches originally had scratch dials (see main text). - original frequency distribution of dials/church profiled in relative terms to listed surviving dials/church (Table 2). - lost dials dependent on parameter combinations bounded by values shown below. Proportion of moved dials lost ¾ ½ Proportion of dials weathered out ⅔ ¾ The above dial loss is that which renders the distribution of surviving scratch dials consistent with all churches originally having one or more dial. A lower dial loss is inconsistent with all churches originally having had a dial; a higher dial loss would increase the original number of dials. THE SCRA TCH DIAL S OF KEN T 343 the full story needs to be teased out. The surviving small dial incidence probably overstates their original incidence; surviving small dials might have been weathered down in size, or alternatively enlarged when originally painted. It would appear quarter dials are relatively large, there being a considerable overlap in size distribution with half dials. A significant proportion of half dials might originally, when painted, have been circle TableABLE 6. CATEGORISATION OF IDENTIFIABLE SCRATCH DIALS IN Kent and Three OtoTher Counties (per cent) Type of scratch diala Quarter H alf C ircle Kentb 54 29 17 Somersetc H erefordshired R utlande 10 31 25 53 66 50 37 3 25 N otes: a Based on number of medieval hours spanned by the dial (Fig. 2). - quarter dials span the first six hours of the day to noon (Fig. 1D and E) - half dials span the twelve day hours – sunrise to sunset (Fig. 1C) - circle dials (Fig. 1A and B) include the twelve night hours (which clearly had no functional use). b Author’s analysis of Winzar recordings. Percentages are of total dials capable of categorisation. 21 per cent of dials are remnants insufficient to be categorised. c A author’s analysis of categorisable photographs in The Ethelbert Horne Archive (Somerset Studies Library, Taunton; copy in The British Sundial Society archive). d Author’s analysis of categorisable line drawings in R. and C. Botzum, Scratch Dials, Sundials and Unusual Marks on Herefordshire Churches, 1988. e Author’s analysis of categorisable line drawings in R. Owens and S. Sleath, Time in Rutland. A History and Gazetteer of the Bells, Scratch Dials, Sundials and Clocks of Rutland, 2002. TableABLE 7. Size OF KENT’s SCRATCH DIALS (PER CENT) Typea of dial Maximum dimension in inchesb >0-1 >1-2 >2-3 >3 -4 >4-5 >5-6 >6-7 >7-8 >10-11 Quarter 6 9 38 29 9 9 - - - 100 H alf 4 - 32 12 16 16 16 - 4 100 C ircle - - 6 11 33 11 22 17 - 100 N otes: a See Table 6, note (a). b Author’s analysis of Winzar recordings. CHRI S H.K. WILLIA MS 344 dials, given the overlap in size distribution.34 An upper limit on the painted size of dials is imposed by circle dials; enlargement beyond the scratched circumference would be limited to annotation. Pulling the strands of the previous paragraphs together, two broad categories of scratch dial, the functional versus the symbolic, emerge. Most of the more functional dials probably had a painted/annotated size of 3 to 6 inches; with half dials only marginally larger than the quarter dials. The more symbolic circle dials had a typical decorated size of 5-9 inch diameter. Such dials were, of course, just as functional as other dials during the day. They are termed symbolic because of their non-functional elements – nocturnal hour lines and their obvious associated symbolism. Their actual decorative appearance and colouring can only be conjectured at this juncture – involvement of the sun, both literally and metaphorically, being the most obvious candidate. What was the pattern of functional versus symbolic scratch dials on Kent churches? The surviving pattern of dial type on churches (Table 8) is consistent, after allowing for weathering and rebuilding (see Table 5), with most, if not all, churches having had, at some stage, a symbolic dial. The lower survival rate of circle dials is suggestive of an earlier style. Might circle dials be an early scratch dial form (hence subjected to above average loss) that became less fashionable in the sixteenth century as traditional religious symbolism was called into question? Thereafter half dials held sway – whilst newly scratched dials would be half dials, the old scratched circle dials would be converted, via painting, into half dials. TableABLE 8. PROPORTION OF MEDIEvAL KENT ChuRCHES WITH Different TypYPes of Scratch Dial (per cent) A ny type of scratch diala Type of scratch dialbcd Quarter H alf C ircle 40.8 22.7 21.2 11.3 A pprox. 30e A pprox. 20e F unctional Symbolic N otes: a F from Table 3. b See Table 6 note (a). c Calculated as the product of the proportion of all churches with a scratch dial(s) having each type of dial (Table 10) and the proportion of churches with any type of scratch dial(s) (Table 3). d The overall proportion of churches with scratch dial(s) is exceeded by the sum of each type of dial because some churches have more than one type of dial. e As half the half dials equal or exceed circle dials median size (Table 7) and daytime radial lines (Table 9) many might originally have been painted circle dials (main text). THE SCRA TCH DIAL S OF KEN T 345 R adial Lines on Scratch Dials The number of day lines (Table 9) on half dials is bimodally distributed; again suggestive many might originally have been painted circle dials. This is further corroborated by circle dials often having few or no marked nocturnal lines (Fig. 1A and B) – suggestive of their having been painted in. Quarter dials are decidedly different, and not only because they are limited to the morning. They have markedly fewer radial lines than other dial types’ morning markings, and are associated with a higher likelihood of multiple dials (Table 10). Turning to the pattern of a scratch dial’s radial lines (Table 11) it is clear that half and circle dials were laid out on the medieval time system – 4 out of 5 dials have a tolerably accurate mass line and accurate noon line;35 4 out of 5 of all lines on such dials accord with medieval hour TableABLE 9. NuMBER OF RADIAL LINES ON KENT SCRATCH DIALS, b y TypePEa (per cent) Morningc Dayd N ighte N o.b Quarter H alf C ircle H alf C ircle C ircle 2 21 - - - - - 3 12 18 7 - - 13 4 27 24 33 6 7 13 5 30 6 7 6 7 20 6 3 35 20 29 - 7 7 3 6 33 - 7 - 8 - 12 - 12 13 13 9 - - - 6 13 - 10 3 - - 12 13 - 11 - - - 6 13 13 12 - - - 18 - - 13 - - - - 27 20 14 - - - 6 - - Totalf 99 101 100 101 100 99 N otes: a See Table 6 note (a). b Author’s analysis of Winzar recordings. Only radial lines (or pockmark to gnomon hole equivalents) included. Circumferential line(s) omitted. c The morning period has 7 hour lines (Fig. 2). d The day period has 13 hour lines (Fig. 2). e C clearly nocturnal lines have no practical use. To facilitate symmetry comparison the sunrise and sunset lines have also been included to give 13 hour lines – the same as the day. f Deviation from 100 due to rounding. CHRI S H.K. WILLIA MS 346 lines.36 As a group, quarter dials do not accord with the medieval time system. Many of those who made them must have been meeting a newly perceived requirement – the obvious candidate being the modern equal hour time system. As this (and its associated clocks and scientific sundials) spread, the implications of the traditional system would have become increasingly apparent. Many quarter dials show alternative mass lines for different times of the year on the modern equal hour system (Table 11 TableABLE 10. KENT’s SCRATCH DIAL INCIDENCE by TyPE ON Churches with Dials (per cent) Type of scratch diala Scratch dials per churchb 1 2 3 ≥4 Total Quarter H alf C ircle 29.7 40.8 20.4 12.9 11.1 5.5 5.6 - 1.9 7.4 - - 55.6 51.9 27.8 N otes: a See Table 6, note (a). b Author’s analysis of Winzar recordings. Percentages calculated from 54 churches with at least one categorisable scratch dial. TableABLE 11. CORRESPONDENCE OF RADIAL LINES ON KENT SCRATCH Dials to the Medieval Time SystYSTem (per cent) Type of scratch diala R adial line testedb Dials with: L ines not on hour linesd Mass linec noon lined Quarter 32-74e 65 40 H alf and circle 44-81 81 23 N otes: a See Table 6 note (a). b Mass, noon and hour lines as shown in Fig 2. Author’s analysis of Winzar recordings. c Tolerance range up to ±3º and ±5º. d Tolerance up to ±3º. e Judged against the modern time system, 9 a.m. Mass results in a scratch dials mass lines moving seasonally between 11º and 48º to the sunrise line. Allowing ±3º tolerance the presence of mass lines is shown below: N o. of mass lines 0 1 2 3 4 Total Per cent dials 9 34 38 13 6 100 90 per cent of quarter dials have a mass line(s) – over half two or more. THE SCRA TCH DIAL S OF KEN T 347 note (e) and Fig. 1E). Many of these quarter dials have few lines (Table 9) and contain only variant mass lines. Sometimes a church has several different quarter dials (Table 10) mapping out the variant mass lines.37 Such quarter scratch dials represent a most particular juxtaposition of time and place, of social and economic circumstance. They bear testimony, somewhat paradoxically, to both the need to adopt modern equal hour time, but also to the absence of the associated technology – clocks or scientific sundials. The rapid take off in Kent’s domestic clock ownership during the second half of the seventeenth century38 suggests it was then that equal hour time became well nigh universal. Poorer parishes did the best they could – they adapted scratch dial technology.39 Most probably therefore the quarter dials showing variant seasonal mass lines on modern time were made in the second half of the seventeenth century and constitute the last type of scratch dial made. They convey the impression of scant regard to decorative appearance. Individual quarter dials are rarely bounded with a scratched circumferential arc, scratched line lengths are often very variable, whilst multiple dials can appear as a jumble. Quarter, morning only, dials also bear testimony to afternoon religious services no longer requiring their time to be indicated. As some quarter dials are on the medieval time system (Fig. 1D), the eclipse of afternoon services probably commenced in the sixteenth century. The quarter medieval time dials are of a neater appearance compared with the modern time ones and do not occur as multiple dials on a church. The Evolution of Scratch Dials This paper’s deduced reconstruction of the use and appearance of scratch dials is consolidated in Fig. 4.40 More than half a millennium of stability was followed by two centuries of change and decline culminating in the obsolescence of scratch dials. Prior to the sixteenth century scratch dial use would have been almost universal and their appearance unchanging.41 During the sixteenth to eighteenth centuries the use and appearance of scratch dials were significantly modified by three factors:- - religious change – accounted for the demise of the symbolic dial (as part of the questioning of traditional imagery) and the introduction of the quarter dial (with the decline of afternoon services); - adoption of new technology – as clocks and scientific sundials progressively spread, the proportion of churches using scratch dials declined; and - universal adoption of the modern time system – led to new scratch dials on the new time system being made and used until the new technology could be procured. CHRI S H.K. WILLIA MS 348 F ig. 4 Evolution of Kentish Scratch Dials. (1) Focus is on original (painted/decorated), not surviving, appearance. Jagged lines indicate major uncertainties. (2) Focus is on the evolving pattern of dials in use, not their surviving pattern. Older dials have a lower survival rate. (3) The narrowing of the scratch dial funnel reflects the adoption of the clocks/scientific sundials progressively rendering scratch dials obsolete. (4) Line positioning reflects author’s best estimates on currently available data. Additional scratch dial recordings and technology adoption evidence might refine their precision. Readers with alternative interpretations of any of the data used can mentally move the lines as required and assess the implications for the coherence of the resulting implied evolution. Scratch dials would have conformed to contemporary decorative conventions and style. Pre-Reformation dials can be expected to have been painted, many with a coloured decorative appearance rich in symbolism. Thereafter their painting is likely to have been less exuberant and limited THE SCRA TCH DIAL S OF KEN T 349 to functional adequacy. The last scratch dials made appear to have paid scant attention to appearance. The (monochrome and skeletal) evolution of scratch dials, as they appear today, is chronicled via examples in Fig. 1. C oncluding Thoughts Scratch dials are the earliest time keeping artefacts to survive in any number. Vernacular rather than professional in construction, they are rich in social context. Their demise mirrors the arrival of the modern age. Yet they are poorly understood, not widely appreciated and little researched. Some readers may already have diagnosed the cause. Research in this area has a voracious appetite for data; data that is not easily or quickly garnered. Although this paper has shed a new and more robust and detailed insight on the subject and the Kentish scene in particular, it too has been constrained by data availability. The next step will be to better understand the regional and national context through comparative analysis of a dozen well surveyed and recorded counties in The British Sundial Society’s mass dial database. There can be no doubt Kent’s scratch dial heritage is a rich one, despite only a fraction of it having survived the rigours of centuries of church rebuilding and weathering. Kent’s recording is not yet complete. Time is running out, particularly for the oldest scratch dials. We are one of the last generations of antiquarians with the opportunity to adequately record this aspect of our heritage – a heritage whose ecclesiological, horological and socio-economic significance we have barely commenced to decipher. acknowledgments Tribute of the highest order is due to Gerald Winzar’s meticulous recordings: as is appreciation of Pat Winzar’s generosity in encouraging their use and availability. Thanks also to Dick Chambers for permitting use of the interim listing of his photographic survey; Tony Wood for making The British Sundial Society’s mass dial database available and other advice; and my daughter Philippa for diligently managing this paper’s production. endnotes 1 L listings, recordings and discussion took place in the publications and proceedings of the county archaeological societies. No substantive reference to scratch dials has been found in Archaeologia Cantiana for this period. 2 E e. Horne, Primitive Sun Dials or Scratch Dials. Containing a list of those in Somerset, 1917. See also E. Horne, Scratch Dials. Their description and history, 1929 – an expanded version, excluding the Somerset listing. 3 A a.R. Green, Sundials, Incised Dials or Mass-Clocks, 1926. CHRI S H.K. WILLIA MS 350 4 T.W. Cole, Origin and Use of Church Scratch-Dials, 1935. 5 The British Sundial Society was founded in 1988 and has a mass dial group dedicated to recording scratch dials. 6 See for example Horne, 1929, op. cit.; R.J.R Rohr, Les Cadrans Solaires Anciens d’Alsace, 1971; E. Zinner, Alte Sonnenuhren an Europäischen Gebauden, 1964; T. Przypkowski, ‘The Art of Sundials in Poland from the 13th to the 19th Centuries’, Vistas in Astronomy, Vol. 9, 1967. 7 R r. Ovens and S. Sleath, Time in Rutland. A History and Gazetteer of the Bells, Scratch Dials, Sundials and Clocks of Rutland, 2002. 8 R. and C. Botzum, Scratch Dials, Sundials and Unusual Marks on Herefordshire Churches, 1988. 9 A system of time subdivision that can be traced back to at least Babylonian times. 10 A fact well known to astronomer mathematicians of the time who could not use the prevailing conventional approach (with its ‘flexible’ hours) for the scientific study of astronomical phenomena. 11 This provided a practical stimulus for the measurement of passing/successive hours in addition to sundials – both for cloudy weather and the determination of night time devotional offices. Monasteries initially used water clocks for this purpose and went on in the late thirteenth century to develop the mechanical clock. See C.F.C. Beeson, English Church Clocks 1280-1850, 1977; and L.A.A. Romeyn, Torenuurwerken. Tijd voor ledereen, 2005. 12 As no original scratch dial gnomon survives this was once a matter of hot debate. Most authorities firmly believed the rod gnomon was horizontal, but Green, op. cit., postulated a bent rod – with severe implications for the interpretation of scratch dial markings. Documentary evidence has since been found confirming the horizontal rod gnomon – John Hovedon (d.1275) Practica Chilindri. 13 For a detailed technical discussion of this and related issues see A.A. Mills ‘Seasonal-Hour Sundials’, Antiquarian Horology, Vol. XIX, Winter 1990. In brief allowing for latitude results in the hour lines: - converging on a point vertically above the horizontal rod gnomon. - no longer being equiangular but bunching around midday. 14 Such texts outlined the scientific and practical basis for the angled gnomon, the bunching of hour lines around noon, and the relationship between traditional seasonal hours and equal (equinoctial) hours – the latter, of course, becoming the modern time system. This paper refers to such dials as scientific sundials. 15 The first categorically rigorous statement was J.R. Findlay, ‘The Construction and Use of Wheel Dials’, Antiquarian Journal, Vol. 7, 1927. See also Cole, op. cit. 16 The first clocks were in monasteries and would have kept seasonal time as did their water clock predecessors; see C.B. Drover ‘A Medieval Monastic Water-Clock’ Antiquarian Horology, Vol. 1, Dec. 1954. There is evidence of equal hour clocks as early as the fourteenth and fifteenth century in Europe. But these were confined to major cities where civic and commercial needs rivalled those of the church. Documentary evidence suggests the two time systems ran in parallel. See E.L. Edwards, Weight-driven Chamber Clocks of the Middle Ages and Renaissance, 1965. 17 Additional evidence is provided by the dial makers of Nuremberg and Augsburg who exported widely in Europe. Their sixteenth and seventeenth century portable compass sundials include dials marked to show both equal and seasonal hours. See M. Lennox-Boyd, Sundials; History, Art, People, Science, 2005 and M. Cowham, ‘Portable Sundial Making in Nuremberg’ Antiquarian Horology, Vol. 29, Sept 2005. 18 There were of course dials before this but, reflecting the fate of the structures on which they were mounted, very few have survived. Surviving Anglo-Saxon dials are of very high THE SCRA TCH DIAL S OF KEN T 351 quality compared to medieval scratch dials. See M. Bowen, ‘Saxon Sundial in the Parish Church of All Saints, Orpington’, and R.I. Page ‘Note on the Inscription’, Archaeologia Cantiana, lxxxii, 1967; also D. Scott, ‘Sundials in Anglo Saxon England, Part 4, The Late Period – Aldbrough and Orpington’, The British Sundial Society Bulletin, Vol. 12(i), Feb 2000. 19 T.W. Cole, Scratch-Dials on Churches. Interim List, 1934. 20 Green, op. cit. Rosenberg taught at the King’s School Canterbury and read papers to the Canterbury Archaeological Society. 21 Welland states ‘a complete set of drawings in book form is under consideration’ and invited prepublication purchase. Did drawings exist for the entire list or was (intended) completion contingent on sufficient pre-publication interest? 22 Welland cites no credits. Gerald Winzar informed him of his findings (Pat Winzar pers. comm.) and Welland’s listing includes almost all dials recorded by Winzar. It also includes all dials listed by Grove. That said, Welland’s compilation contains some 25 churches listed for the first time. 23 F for those concerned that three counties is too limited a sample, initial statistical analysis by the author of the (unpublished) British Sundial Society database for some dozen counties with complete to comprehensive surveys of eligible churches, indicates the three counties are not unrepresentative. See The British Sundial Society Bulletin forthcoming. 24 Also supported by the correspondence of Welland’s 310 churches visited with 331 churches with medieval components. Whilst additional discoveries can never be ruled out, any such can be expected to be few and not materially affect the picture shown in Tables 2, 3 and 4. 25 The other possibility – unknown survivals – does not arise for the Kent, Somerset and Rutland data under consideration (main text and note 24). If all churches have not been surveyed, relating listed scratch dials to the total eligible churches will result in an underestimate of survival. This is avoided if listed scratch dials are related to the number of churches surveyed (assuming the latter to be statistically representative). This is not always possible because of the natural tendency only to list positive findings. Statistically ‘non-finds’ are just as important! Non reporting of churches without dials is why the published partial surveys of Hampshire (Green op. cit.) and Herefordshire (Botzum op. cit.) are not included in Table 4. In contrast the Somerset and Rutland surveys are both complete and list churches without a scratch dial. Non reporting of churches without scratch dials also affects The British Sundial Society mass dial database. Fortunately such information can be retrieved and the database is in the process of being enhanced to permit statistical analysis (note 23). 26 The appearance of scratch dials and the nature of incisions are not the work of professional stone masons. Most likely they were made by (or supervised by) the local priest, sextant or verger. 27 A author’s analysis of Somerset (Horne 1917, op. cit.), Rutland (Ovens and Sleath op. cit.) Hampshire (Green op. cit.) and Herefordshire (Botzum op. cit.). Kent information has not been systematically recorded – but numerous moved dials are known. The cumulative impact of repairs, re-use and replacement is vividly detailed in J.F. Potter ‘Anglo-Saxon building techniques: Quoins of twelve Kentish Churches reviewed’, Archaeologia Cantiana, cxxvi, 2006. Two of these churches have scratch dials on the quoin stones – one obviously, the other possibly, moved. 28 H horne (1917), op. cit., describes 19 per cent of Somerset dials as ‘doubtful’. Ovens and Sleath, op. cit., describe 8 per cent of Rutland’s as ‘very poor or faint’ and 42 per cent ‘poor’. Botzum, op. cit., describes 25 per cent of Herefordshire dials as variously ‘very weathered, eroded, remnant, flaking or indistinct’. Of the Kent dials recorded by Winzar 21 per cent are too indistinct to be categorised (Table 6). CHRI S H.K. WILLIA MS 352 29 Sufficient perhaps to prompt some to question their validity. For Table 5’s original prevalence of scratch dials to be considered suspect, argumentation and evidence (outweighing that presented) is required to both, overturn the presumed universality of scratch dial diffusion, and indicate scratch dial loss to be overstated. This dual requirement is a logical consequence of the methodology employed. It is interesting to note Ovens and Sleath, op. cit., in the introductory discussion to their recording of Rutland’s scratch dials, favour the presumption they were originally present on all churches, and intuitively speculate less than 30 per cent have survived. 30 Winzar’s photographs, line drawings, measurements of dimension, line angles and position constitute a benchmark unsurpassed nationally. His thoroughness is exemplified by the fact that on a quarter of the churches surveyed he has recorded the most dials (Annex). 31 Church limewashing and dial painting renders the estimated original prevalence of multiple dials less inexplicable. Only one or two would be in active use, the rest being limewashed out. 32 More complex categorisation (based for example on gnomon hole – between stones or separately drilled, lines or pock marks, presence of a circle) has appeared in the literature. Horne 1929, op. cit., has twelve. There are severe practical and methodological difficulties with such taxonomies. The more multivariate the greater the challenge of defining a taxonomy that is both exhaustive and mutually exclusive i.e. capable of statistical analysis. Moreover the taxonomies developed in the literature are biased towards surviving, rather than original, dial appearance. A simpler categorisation orientated toward broad original, rather than the detailed vagaries of surviving, appearance, offers the prospect of more meaningful insight. 33 A additional counties, especially those near to Kent, need to be examined before the/any regional context can be determined. Kent also appears to have far fewer dials with scratched enumeration around their edge compared to other counties. 34 H half and circle dials share a common maximum dimension – the diameter of a circle. 35 O ovens and Sleath, op. cit., report similar results for Rutland: 76 per cent with mass, and 84 per cent with noon lines. Compared with this paper their figures contain offsetting effects – the inclusion of quarter dials and the adoption of ±10º tolerance. 36 Missing mass and noon lines can be attributed to the weathering out of the scratched line or it being painted only, not also scratched, on the original dial. None hour lines can be attributed to badly set out dials, walls badly off east-west alignment or the time/event indicator for various activities undertaken by the priest. 37 A although the radial lines of individual scratch dials have been examined in the literature, systematic statistical consideration is rare. An exception is C.M. Lowne, ‘An Analysis of Some Mass Dials of Sussex and Kent’ The British Sundial Society Bulletin, Vol. 97(ii), April 1997, analysing 70 scratch dials from East Sussex and south-west Kent. Comparison is hindered by differing methodologies (a single stage discriminatory analysis versus this paper’s staged segmented approach) and sample size (only half Winzar’s) both contributing to over a quarter of Lowne’s dials being statistically indistinguishable by time system. As however two-thirds of Lowne’s distinguishable dials conform to this paper’s Fig. 2, the broad thrust of his results is not inconsistent with this paper. I t is appropriate to note L.G. Welland ‘Scratch Dials and a Suggestion of their Working’ Bygone Kent, 1981 principally because of Welland’s contribution to listing Kent scratch dials. However the paper is neither definitive nor conclusive, believing ‘… the whole puzzle may be insoluble’. 38 See C.H.K. Williams ‘Charing Clocks, Clockmakers and Clockkeepers’ Archaeologia Cantiana, cxxv, 2005, and C.H.K. Williams, ‘Seventeenth & Eighteenth Century Clock Demand, Production and Survival: An Economic and Statistical Analysis’, Antiquarian Horology, Vol. 28, March 2005. THE SCRA TCH DIAL S OF KEN T 353 39 Many rural parishes in Kent never had a church turret clock and did not procure a scientific sundial until well into the eighteenth century. At Selling a half dial has been re-incised with bunched hour lines and bears evidence of a wedge gnomon. Such conversions of scratch dials to pseudo-scientific sundials are rare. 40 Scratch dial literature often mentions the presence of deliberate later dial alteration; an issue warranting more conclusive consideration than hitherto in the literature. The greatest source of comfort is that it is still possible to discern patterns. (Any) dial alteration has been insufficient to obliterate the underlying picture. A particular concern has been artificial circle dials – yet this paper has advanced sound reasons why we should not be surprised by them. If there was a fashion to make artificial circles why apparently spare quarter dials from such attention? Why should there be a fixation to make circles to deceive? Why not a range of artistic vandalism? Yet we do not see it. It is difficult to conceive of the past activities of miscreants being of a scale and systematic nature sufficient to materially mislead us today. It will have been noted the dating of scratch dials receives little assistance from the building dates of churches, primarily because of their great age and subsequent rebuilding. The most useful information is the dated building of porches that cast scratch dials into permanent shadow. 41 The only exceptions would have been those communities sufficiently wealthy and cosmopolitan to procure clocks/scientific sundials early – these were very few in number. Clocks are recorded in Canterbury, Dover and Hythe by 1292, 1404, and 1412 respectively; see M. Pearson, Kent Clocks & Clockmakers, 1997. Scratch dial incidence is lower in Canterbury, Rochester and the Cinque ports, consistent with an early discontinuance of use. However this is too small a sample to be categoric or to infer the relationship between surviving incidence and the time when scratch dials were displaced. annex Kent Scratch Dials by Parish Key: √ Scratch dial(s) listed: 1,2,3… no. of scratch dials listed. Source: see Table 1. Parish L isting C ole Mee Well. Grove Winz. C hamb. BSS* A disham √ √ 5 5 3 6 A ldington √ 3 1 4 A ppledore 1 Ash (Sandw.) √ √ 1 1 1 A sh (Wrot.) √ √ 1 A shurst 1 Bapchild √ √ 1 1 Barfreston √ √7 8 6 8 Bekesbourne √ 1 Bethersden 1 Bexley √ 3 3 Bidborough √ Biddenden √ 1 1 Blean 2 Bobbing √ √ 2 2 CHRI S H.K. WILLIA MS 354 Parish L isting C ole Mee Well. Grove Winz. C hamb. BSS* Borden √ √ 1 1 Boxley √ 1 Brabourne √ √ 5 6 Brenzett √ 1 1 1 Bridge 2 Brook √ 1 Brookland √ 1 1 1 Burmarsh √ 2 3 2 C apel √ 3 C apel/Ferne √ 1 C halk √ C hallock √ √ 1 2 C haring 1 1 C hillenden √ √ 3 1 5 C hislet √ √ 1 1 C obham √ 1 1 1 C ooling √ 1 C rayford √ √ 1 C rundale √ √ 1 1 C udham 1 Dartford √ Denton √ √ 4 2 2 Doddington √ √ 2 1 2 Dymchurch √ √ √ 1 2 3 E ast Langdon 1 2 E astry √ √ √ 6 4 1 E bony √ 1 E lham 1 E lmsted √ √ 5 3 3 E lmstone √ 1 E ythorne √ √ 1 F olkestone √ F oots Cray √ √ 2 1 F rindsbury √ G’nstone(F) 3 G’nstone(W) 1 Goudhurst 1 Great Chart √ H ackington √ √ 1 1 Harbledown √ 1 2 H artley √ 1 1 H arty √ √ 1 1 H astingleigh √ THE SCRA TCH DIAL S OF KEN T 355 Parish L isting C ole Mee Well. Grove Winz. C hamb. BSS* Hawkinge** √ √ 1 Hawkhurst 3 H ayes √ 1 1 H erne √ √ 1 1 1 H inxhill √ √ 2 1 H oath √ 1 2 2 H orton Kirby √ 2 Hothfield √ 1 3 1 H ythe √ 4 3 3 I ckham √ √ 2 3 1 Kemsing 1 Kenardington √5 1 1 Knowlton 1 L amberhurst √ 1 1 L enham √ √ 1 1 L ittlebourne √ √ 2 4 2 Lwr Halstow 1 L yminge √ √ 1 1 1 L ynsted √ √ 4 2 2 Mersham √ √6 8 5 6 Milton Grav.) 1 Milton Regis √ √ 2 Minster(Tht) 1 Molash √ 3 1 Monks Hort’n √ 1 2 3 Monkton √ Newington 1 New Romney √ √ 1 1 1 N onington √ √ 2 N orthbourne √ 3 Northfleet 2 O are √ O tford √ 1 1 1 Paddlesworth √ √ 1 1 2 Patrixbourne √ √ √10 10 11 13 Postling √ 1 Preston(Wng) √ √ 5 5 2 Ringwould √ √ 3 2 R olvenden √ √ √ 2 2 3 R uckinge √ √ 2 2 3 Saltwood √ √ 1 1 1 Sandhurst √ 1 Seal √ 1 CHRI S H.K. WILLIA MS 356 Parish L isting C ole Mee Well. Grove Winz. C hamb. BSS* Sellindge 1 Selling √ √ 1 1 1 Sevington √ 2 2 Sheldwich √ √ √5 4 2 4 Smeeth √ 4 1 2 Speldhurst √ Southfleet 1 Staple √ √ 3 4 2 Staplehurst 1 Stodmarsh √ √ 2 2 2 Stourmouth √ √ 2 Sturry √ Swanscombe √ √ 2 Swingfield √ 2 2 2 StMaryMarsh √ 2 1 1 Temple Ewell √ 2 2 4 Tenterden √ 1 2 2 Teynham √ 1 1 1 Tilmanstone √ √ 2 Tonbridge √ √ 1 U pr Hardres √ √ 1 3 Walmer √ 2 Waltham √ √ 2 2 2 Warehorne √ √ 1 1 1 Westbere √ Westcliffe 1 Wickhambr’x √ 1 Willesboro √ √ 2 Wingham √ 2 2 Womenswold √ 1 Woodchurch √ √ 1 1 Wootton √ √ 1 2 4 Worth √ √ 3 5 5 Yalding √ 1 4 * The British Sundial Society. ** Converted to a private house.