Archaeological Recording at St. Stephen’s Church, West Putford, Devon 1. Introduction This document has been produced by South West Archaeology (SWARCH) at the request of Mr Graham Edwards, Jonathan Rhind Architects, on behalf of the West Putford PCC (the Client). Consolidation work on the roof structures at St Stephen’s had been deemed necessary, and the intervention of conservation building specialists allowed opportunity for archaeological recording of built structures and investigation of accumulated debris deposits known to exist between the ceilings and the roofs of nave and transepts. This report is the final phase in a programme of work that was set out in a Written Scheme of Investigation (WSI) produced by SWARCH and set out the methodology for archaeological recording to be undertaken prior to and during the proposed repair works. 2. Background St. Stephen’s is a rural parish church with what has been described by Pevsner as “an early plan” comprising a west tower, Nave, Chancel, North and South Transepts and a south Porch. (Cherry & Pevsner 1989, 902) Pevsner suggests that the building dates largely from the 14th century with the tower and the transepts being of a later date. Restoration in the 19th century gave the chancel a new roof structure (English Heritage file reference no: 04/001/091971). 3. Summary Monitoring and recording work at St. Stephen’s Church, West Putford, Devon, revealed a number of interesting details shedding further light on the development of a relatively littlestudied rural parish church. The nave and north transept roof were revealed to have rather finely constructed double archbraced trusses, whilst the south transept was of poorer quality with the lower arch braces’ function replaced with a vertical ashlar piece, and evidence of this roof either having been altered during its construction or to have been a re-used structure brought in from elsewhere. Other details recorded.The existing lath-and plaster barrel ceilings were demonstrated not to have been original features but have been later insertions in what was formerly an exposed roof. The east window recorded in the south transept had been subject to alteration but was likely to have been of late 13th century origin. This was an earlier date than had been previously suggested for much of the fabric within St Stephen’s. 4. Methodology The work carried out consisted of three principal elements. The first was the archaeological investigation and recording of the deposits accumulated between the roof and ceilings that earlier structural investigation had identified. Where the deposits warranted such treatment they were to be excavated archaeologically utilising the same principles and methods as would be applied to below-ground excavation. The second element involved the recording of the roof structures of the nave and transepts, with the aim of understanding the phasing of each structure, and identifying any repairs. The first part of the work carried out after the slates had been removed entailed a visual survey of the entire roof structure in order to identify those parts requiring more detailed recording. At that stage it was decided to undertake the following recording: a written record of each truss in nave and transepts, including notes and a entry for each joint and member on pro-forma sheets; a measured elevation drawing at a scale of 1:20 of a sample truss from nave and each transept (Figs 3, 4 & 5); and a photographic record of specific details as necessary to supplement the information in written and drawn records. The chancel roof, was outside the scope of the present programme of recording. In this report for the sake of consistency the same numbering scheme and abbreviations for roof trusses as adopted by Jonathan Rhind Architects has been used throughout, (NR= Nave Roof, STR = South Transept Roof, NTR = North Transept Roof) (see Fig 2) The third element involved the archaeological recording of the south transept window to allow its construction and subsequent repair to be understood. For both the interior and exterior of this window an annotated measured drawing at a scale of 1:10 was made, and a photographic record of the interior and exterior elevations were made. Recording took place before the removal of glass and dismantling of the ferramenta. 5.0 Results 5.1 Debris Deposits Removal of the roof slates enabled the examination of the debris identified in the earlier Opening up Report. The accumulated debris between the ceilings and roof of nave and the south transept was revealed to be of three sorts. The first and upper layer, as much as 0.5m thick in places was the result of bird-nesting activity. It was very dry and consisted primarily of twigs, a high proportion of which were of ash (fraxinus excelsior) indicating that the nests were likely to be of a corvid species which, given the location, made Jackdaws the most likely culprit. Also included in these deposits were fine dry soil, small stones and grit, hair, animal fur and skin, feathers, bones of avian spp rat and mouse, and small diameter tunnels which indicated rodent nesting. A certain amount of modern rubbish including food wrappers and cigarette ends were also present, presumably scavenged by the birds and rodents. The bulk of this material was revealed in the narrowing space between ceilings and roofs immediately above the wall tops. Relatively little was found in the larger spaces above the centre of the ceilings or in the valleys formed between transept and nave roofs. This material was checked for artefacts (not an uncommon occurrence in Corvid nests) and was then removed being of no archaeological significance. The second layer was of building/demolition debris, in resting on top of the wall tops, filling the spaces between the rafters of the trusses, between the roof slates/boards and above the barrel of the lath and plaster ceilings. This material included small and medium fragments of roofing slate some with mortar adhering to one face, fragments of much decayed timber boarding, corroded iron nails, a few partial tapering oak pegs of approximately 2cm diameter and one fragment of ceramic crested ridge tile, (gravel tempered fabric, green glazed exterior, mid 15th century, ) with lime mortar adhering to its underside. Some of the roofing slate fragments also had traces of thin mortar adhering to on face. The depth of this deposit was quite variable; averaging 0.25m but in places as much as 0.5m thick. The third and lower layer of material consisted of a loose build of stone and a loose lime mortar matrix. The lower part of this material was predominantly a mixture of uncoursed stone rubble of small and medium sizes in a loose crumbling lime mortar. It also contained fragments of roof slate, wood and ferrous objects. In places this build included localised use of coarse cob (see Plate 1). In order to allow access to the lower parts of the trusses this build was removed from between the trusses of alternate bays. In the north transept, there was no loose build of mortar and demolition material on the wall top between the rafters. Some bird nest accumulation, similar to that described above was present. In several places in the nave and both of the transepts internal plaster was seen to survive adhering to the inner facework of this loose build, above the level of the existing barrel ceiling (see Plate 2), in the case of the second nave truss (NR 2, see Fig 3) as high as the point where the lower arched braces sprang from the principal rafter. In the south transept internal plaster on the inside face end (south) wall was also seen to rise above the level of the lathand-plaster barrel ceiling. Once the loose build from between the rafters was removed it was apparent that it had been in large part responsible for accelerating the deterioration of the bases of the truss rafters, and wall plates. Many of these timbers bore a distinct line of rot that corresponded with the upper level of the infill build debris deposit prior to its removal. 5.2 Roof Structures The roof structure was revealed to be in a much poorer condition than had been understood prior to the stripping of the slates. Many of the timbers were in places, especially at the wall tops, badly decayed. The Nave Roof: Structure The Nave roof was constructed on 26 trusses. For the purposes of identification these were numbered from west to east. Truss 1, therefore, stood against the east wall of the tower, and truss 26 formed the western end of the chancel roof. The roof was of a double arch-braced construction in oak, with a wagon/barrel lath and plaster ceiling directly below and following the line described by the lower and upper arched braces. An example truss (NR 2, western elevation) is illustrated in Fig 3 The principal rafters were an average of 4.9m in length, and the apexes of the trusses were an average height of 3.6m above the level of the wall tops. The base of each truss was formed by a sole plate 0.8m long lying across the thickness of the wall tops. This sole plate was notched to lie over two wall plates, one running along the centre and one along the inner face of the wall. The lower end of each rafter was tenoned into a mortice in the outer end of each sole plate. The inner end of each sole plate was tenoned into a mortice cut into the lower end of each lower arch-brace. The two rafters in each truss were joined at the apex by a halved joint fixed with a single peg 2cm in diameter. In addition they were joined by a horizontal collar located 3.5m up the rafter or 2.3m above the wall tops. Each end of the collars were tenoned into mortices extending through approximately half the thickness of the rafters, and fixed with pegs. Describing an arc between each collar and rafter were a pair of upper arched braces. Each was 1.7m long and the ends formed tenons jointed into mortices in collar and rafter stopped with small right angled butts. All joints were pegged, although the number of pegs varied (see discussion below). The butts accommodating the lower ends of the upper arch braces were an average of 2.15m up the undersides of the rafters from the wall top. The upper end of each lower arch-brace was butt-stopped in a similar way to the upper arch braces, springing from an average of 0.45m below each upper arch-brace, or 1.7m up the rafters from wall top. The bottom of each lower arch-brace was tenoned into a mortice in the inner wall plate and itself morticed to receive the tenon at the inner end of each sole plate. These lower joints were fixed with 1 peg. In combination the inner faces of the lower archbraces, rafters, upper arch-braces formed a smooth arc and the centre underside of the collars were cut in a very shallow V to continue this in a single curve spanning the full width of the nave. The lower members and joints of most trusses were particularly badly affected by rot, especially where in contact with the loose rubble build that filled the voids between rafters and lower arch-braces. In some cases those joints between lower arch-braces and wall plate and sole plate, and between rafter and sole plate were partly or completely absent. To the east of the truss drawn (NR 2) the loose rubble build infilled between the trusses was seen to have plaster adhering to its vertical inner face, above and behind the level of the barrel/wagon ceiling of lath and plaster (see Fig 3, Plate 2). This observation was also made elsewhere, for example on the gable wall of the south transept. Nave Roof: Other Observations Various deviations from the standard form in the nave roof were observed The easternmost seven trusses (NR 19-25) of the nave roof were of a slightly different form. These trusses lay at the crossing of the north and south transepts. Here the lack of side wall to the nave necessitated a different arrangement to support the base of the principal rafters. Thus for these trusses there was no sole plate, or wall plates. Instead, the void formed by the openings of the transept was spanned by a bressummer, a beam with moulding on the inner face, into the top of which the base of the rafters and lower arch-braces were tenoned. In trusses NR 23-25 the trusses had slipped down and to the south due to decay in the south bressummer. Here the principal rafters on the north side were overlain by additional softwood rafters, clearly added as a repair to preserve the roofline. One area of difference between trusses was that of the number of pegs utilised in fixing the joints of the arched braces. The joint between lower arched brace and rafter, upper arch-brace and rafter, and upper arch brace and collar was in some cases fixed with two and in other cases three pegs. No pattern could be observed in this: some timbers had the same number of pegs used at each end, whilst in other cases the numbers were different, in some cases the upper and lower arch-brace used similar numbers of pegs, in others they were different. One anomalous feature observed in the majority of nave trusses were a series of pegholes that bore no obvious relationship to any existing part of the roof structure. These holes were found in the eastern face of the principal rafters 1.2m from the bottom of the member, just above the top of the loose rubble build on wall tops on both north and south sides of the nave in all trusses apart from NR 5, 6 and 14. There was one hole in each timber, and each was horizontal and penetrated approximately half the rafter’s thickness. There was no evidence for these holes having formed a part of any joint with other extant timbers. They were of 2cm diameter, similar to other pegholes used for the truss joints (see Plate 2). These seem to be typical examples of those somewhat mysterious features identified by Brunskill as “problem holes” or “rafter holes” believed to either be concerned with the rigging and aligning of the two rafters in a truss, or as the fixing point for sprockets (Brunskill, 172). The North Transept: Structure The roof of the north transept was constructed on 9 main trusses. For the purposes of identification these were numbered from north to south. Truss 1, therefore, stood against the north wall of the transept whilst 9 was the one closest to the nave. These equated to NTR 4-12 in the numbering schema used in the records of Jonathan Rhind Architects. The roof was of a double arch-braced construction in oak, with a wagon/barrel lath and plaster ceiling directly below and following the line described by the lower and upper arched braces. An example truss (NTR 11, south elevation) is illustrated in Fig 4 The principal rafters were an average of 3.4m in length, and the apexes of the trusses were an average height of 2.3m above the level of the wall tops. The base of each truss was formed by a wall plate lying along the inner face of the wall top, and a course of pad-stones lying along the top of the outer face of the wall. These pad stones were moulded and overhung to form a cornice. The lower end of each rafter rested directly on the pad-stones. The full width of the lower end of each lower arch-brace was fitted into a mortice cut into the top of the wall plate Unlike the nave roof where the base of rafter and lower arch-brace were joined by a soleplate, in the north transept there were no sole-plates. Instead there was a horizontal tie beam 0.3m above the wall top, one end of which was tenoned into a mortice in the underside of the rafter, the other end tenoned into a mortice in the lower arch-brace (see Plate 3). Both these joints were fixed with a single oak peg. The two rafters in each truss were joined at the apex by a bridle joint fixed with a single oak peg. In addition they were joined by a horizontal collar 1.15m in length, located 2.5m up the rafter or 1.7m above the wall tops. Each end of the collars were tenoned into mortices extending through approximately half the thickness of the rafters, and fixed with single pegs. Describing an arc between each collar and rafter were a pair of upper arched braces. Each was 1.25m long and the ends formed tenons jointed into mortices in collar and rafter. All joints were pegged, although the number of pegs varied (see discussion below). Unlike in the nave trusses the joints at the ends of the arch braces were not stopped with right angled butts. Instead at their upper ends both arch-braces in each truss butted up against a small timber filet inserted between the two under the collar. At their lower ends each upper arch-brace butted against the upper ends of the lower arch-braces. At this point the lower arch-braces were tenoned into a mortice in the underside of each rafter 1.1m above the wall tops. Each lower arch-brace was 1.25m long and the lower ends were tenoned into the wall plate. In combination the inner faces of the lower arch-braces and upper arch-braces formed a smooth arc, this single curve spanning the full width of the transept. Attached to the lower face of the arch formed by each truss, was a lath-and-plaster barrel / wagon ceiling The lower members and joints of most trusses were affected by rot. The extent of the rot, whilst perhaps not as severe as that seen in parts of the nave and the south transept, was sufficient that the bottom ends of some of the principal rafters had completely disintegrated. North Transept Roof: Other Observations Various points of deviation from the standard form of the roof trusses were observed. The north end of the transept roof was not supported on free-standing trusses. Instead it was carried on three pairs of modern sawn softwood rafters laid directly over the thickness of the end (north) wall. These members (NTR 1-3) were not recorded in detail. At the time works commenced it appeared that the ridge line of the slate transept roof continued unbroken right up to the nave. Removal of the roof slates revealed that internally the transept roof, and its ceiling formed a discrete entity separated from the nave roof by the timber bressummer which continued the line of the nave wall and carried the base of the principal rafters of the nave. The lath and plaster ceiling of the transept curved down from truss NTR 9 to the level of this bressummer, forming a valley between the transept and nave. In order to continue the ridge of the transept roof proper formed by NTR 4-9, a series of additional structures had been constructed. They took the form of five partial trusses (NTR 13-17), cross members and a ridge piece, all of sawn softwood. These structures were apparently modern, and not recorded in detail. It was unclear what form the earlier roof line may have taken. One possibility is that some other timber structure carried a continuous ridge from transept to nave. The other is that an open valley existed between transept and nave. No physical evidence of structures was observed that could confirm either interpretation, although it may be considered likely that the presence of a valley may have left some remains of flashing. The absence of such evidence lends rather more weight to the idea of a continuous ridge from transept to the flank of the nave roof. One area of difference between trusses was that of the number of pegs utilised in fixing the joints of the arched braces. The joint between lower arched brace and rafter, upper arch-brace and rafter, and upper arch brace and collar was in some cases fixed with two and in other cases one peg. No pattern could be observed in this: some timbers had the same number of pegs used at each end, whilst in other cases the numbers were different, in some cases the upper and lower arch-brace used similar numbers of pegs, in others they were different. In the truss drawn (NTR 11) there were a series of anomolous horizontal peg holes evident in the faces of the lower 0.5m of the principal rafters. In the west rafter there were six such holes, and in the east, seven. They did not appear to relate to any other part of the surviving roof structure. The South Transept: Roof Structure The roof of the south transept was constructed on 9 main trusses. For the purposes of identification these were numbered from south to north. Truss 1, therefore, stood against the southern end wall of the transept whilst 9 was the one closest to the nave. These equated to STR 3-11 in the numbering schema used in the records of Jonathan Rhind Architects. The roof was of different structure to that of nave and north transept. Although the upper part of the trusses had a similar arrangement of collar and upper arched braces, the function of the lower arch brace seen in nave and north transept, was here replaced with an ashlar piece (Plate 4). The north elevation of a sample truss (STR 10) is illustrated in Fig. 5 The principal rafters were an average of 3.2m in length on the west side, and 3.3m long on the east. and the apexes of the trusses were an average height of 2.4m above the level of the wall tops. The base of each truss was formed by a sole plate 0.75m long lying across the thickness of the wall tops. This sole plate was notched to lie over two wall plates, one running along the centre and one along the inner face of the wall. The lower end of each rafter was tenoned into a mortice in the outer end of each sole plate. The inner end of each sole plate was morticed to accommodate a tenon formed at the end of a vertical post or Ashlar piece that rose 0.45m before being tenoned into a mortice in the underside of the principal rafter. All joints here were fixed with single oak pegs. The two rafters in each truss were joined at the apex by a bridle joint fixed with a single oak peg. In addition they were joined by a horizontal collar 1.4m in length, located 2.2m up the rafter or 1.55m above the wall tops. Each end of the collars were tenoned into mortices extending through approximately half the thickness of the rafters, and fixed with single pegs. Describing an arc between each collar and rafter were a pair of upper arched braces. These were much more crudely shaped than in the nave and north transept roof. Each was 1.25m long and the ends formed tenons jointed into mortices in collar and rafter. All joints were pegged with a single oak peg. Unlike in the nave trusses the joints at the ends of the upper arch braces were not stopped with right angled butts. In combination the inner faces upper arch-braces and rafters formed a rather irregular arc, this single curve spanning the full width of the transept. Attached to the lower face of the arch formed by each truss, was a lath-and-plaster barrel / wagon ceiling. Rather than beginning at the wall top as in nave and north transept, this curve began at a higher level. The sides of the ceiling were vertical being attached to the ashlar pieces that continued the vertical line of the inside face of the wall 0.5m above the wall top. The lower members and joints of most trusses were badly affected by rot, especially where in contact with the loose rubble build that filled the voids between rafters and ashlar pieces. Particularly badly affected were the bottoms of the rafters and outer ends of the sole plates in some trusses the joint between these members being partly or completely rotted away. South Transept Roof: Other Observations Various points of deviation from the standard form of the roof trusses were observed. When the slate was initially stripped from the south transept roof, boarding was still in situ beneath (Plate 5). This boarding had peg holes that suggested the former presence of wooden shingles. This evidence was not seen in other parts of the church roof. Comparable to the construction of the north transept roof, the south end of the south transept roof was not supported on free-standing trusses. Instead it was carried on two pairs of modern sawn softwood rafters laid directly on the masonry over the thickness of the end (south) wall. These members (STR 1-2) were not recorded in detail. STR 2 was a flat board rather than a rafter, that member on the west side being sinuous. At the time works commenced it appeared that the ridge line of the slate transept roof continued unbroken right up to the nave. Removal of the roof slates revealed that internally the transept roof, and its ceiling formed a discrete entity separated from the nave roof by the timber bressummer which continued the line of the nave wall and carried the base of the principal rafters of the nave. The lath and plaster ceiling of the transept curved down from truss STR 9 to the level of this bressummer, forming a valley between the transept and nave. In order to continue the ridge of the transept roof proper formed by STR 3-11, a series of additional structures had been constructed. They took the form of two partial trusses (STR 1213), supporting a square section oak ridge purlin that ran all the way fro STR 10 to the nave roof. In addition these trusses and the purlin were reinforced with a series of bolstering props and struts bodged in to the other timbers of nave and transept. The joints on these timbers were makeshift, largely consisting of simple Vs and notches, fixed with iron nails. These structures were clearly additions to the transept roof proper, although perhaps of some antiquity and not so obviously modern as those of the north transept roof-valley. It was unclear what form the earlier roof line of the transept may have taken. One possibility is that some other timber structure carried a continuous ridge from transept to nave. The other is that an open valley existed between transept and nave. No physical evidence of structures was observed that could confirm either interpretation, although it may be considered likely that the presence of a valley may have left some remains of flashing. The absence of such evidence lends rather more weight to the idea of a continuous ridge from transept to the flank of the nave roof. Other anomalous structures were observed in the south transept roof. On the upper surface of each collar in truss STR 11 and STR 5 the decayed remains of a single vertical oak post, bridle jointed over a notch cut in the centre or the collar were present. The collars in other trusses showed notches suggesting that more such posts may have been present. The post present on STR 7 (see Plate 6) was itself notched as if for a bridle joint at its surviving upper limit approximately 0.25m above the collar. Another feature was a facing pair of empty mortices seen in each pair of rafters. These were located on the inner faces of the rafters 0.6m below the apex of each truss 0.15m above the existing collars. 5.3 The Window of the South Transept The east window in the south transept was 2.4m in width and 3.8m in height. The window subdivided in the form referred to as intersecting Y tracery. It consisted of three equal principal lights each lancet shaped 0.6m in width and 2.9m in height, above these the tracery extending the curving lines of the lancet tops and intersecting to form three small four sided lights. The internal and external elevations of the window were drawn (see figs 6 & 7). Each of the principal lights was fitted with an iron ferramenta (see Fig 6). Each was of similar proportions consisting of a continuous vertical bar with three evenly spaced horizontal crossbars. The members fitted together by means of each cross bar having a central loop which fitted over the upright bar, and was then held in the correct position with iron wedges. The ends of most of the ferramenta’s members were corroded away (Plate 7). Parts of the stonework of the tracery were in poor condition. There was some mis-alignment of component stones, in particular the central X member, and surrounding components of the upper tracery, probably as a result of settling. These movements had been identified in the architects drawing 377/08. Some of the stonework had previously been repaired with mortar, and in places this mortar obscured the detail of joints between the components. There was no evidence that any of the stone components had been replaced, although some damage to the tracery corresponded with the position of the members of the ferramenta, suggesting that either it had been re-set, or that the exfoliation of its ends had damaged the stone settings (Plate 7). The exterior of the window (see Fig 7) revealed an important structural detail completely unapparent from the inside where internal plaster concealed the wall masonry and splay of the window opening. The external elevation provided clear evidence that the window had been altered by the blocking of its lower part thus reducing the overall height of the window. The original height of the external window opening would have been 4.5m. The masonry lower cill of the former opening was still visible, the infill taking the form of a single mortared course of roughly dressed stone blocks fitted snugly into the splay of the window (Plate 8). Discussion There was no evidence of the stonework of the east window of the south transept having been replaced, making it entirely possible that it represents and original feature. Intersecting Y tracery has been generally accepted as being characteristic of the 13th century (Friar 286) a date which provides interesting contrast with the usual characterisation of West Putford church as largely 14th century by Pevsner (1989, 902) Spurr (1983, 72) and and other authors who have clearly followed their lead (e.g. Slater 1999, 213). No evidence suggested a motive for the blocking of the the lower part of the window. The equivalent window in the north transept, was of similar tracery form, but was much shorter than either the full or blocked height of the window recorded. 6. Conclusions The archaeological monitoring of the restoration and repair works at West Putford Church allowed the recording of a considerable amount of interesting detail. In respect of the ceilings and the roof structures of nave and transepts, as well as the window of the south transept this has shed some new light on the matter of the development of St Stephen’s. The Internal ceiling and Roofing Material. The vertical plaster faces seen rising above the level of the barrel of the lath-and-plaster ceiling in the side wall of both nave and gable wall of the south transept was a clear demonstration that the ceiling was a later addition, post-dating the plastering of the wall tops. This strongly indicates that originally the roof timbers were not designed to be hidden, but rather were exposed within the church. The debris incorporated in the rough build between the rafters provided evidence of the former roof covering. The presence of slates with mortar adhesion, and ceramic ridge tile suggested that an earlier roof had consisted of a torched slate roof with glazed decorative roof furniture. In the south transept boards with pegoles had suggested the former presence of wooden shingles, but no evidence for these was seen elsewhere. It remains possible that one phase of all or part of the church’s roofing was in shingles, but certainly much of the church seems likely to have been roofed in slate at an early date as fragments of such roofing material is incorporated into the build between rafters. It seemed apparent that the addition of the wagon lath-and-plaster ceilings to the underside of the existing roof arrangements had substantially contributed to the acceleration or rot within the roof members. The Roof Structures. The timber structures of nave and each transept were different, strongly suggesting that each represented a discrete phase of building. In the north transept the small tie between rafter and arch brace above the level of the wall top and fulfilling the function of the sole plate in nave and south transept was similar to that seen at Tuckers Hall Exeter dated to c.1470 and that in the north aisle at Hatherleigh church described as late medieval (Westcott PDAS 50, 80). The north transept was much more skilfully constructed than the south. The south transept roof had a number of peculiarities. Two of these, the empty mortice holes on the rafter undersides as if for a higher collar, and the traces of upright posts were particularly interesting. These two anomalies suggest that either the roof structure was significantly altered during construction, or that the roof represents a re-use of roof timbers (or even an entire roof structure) brought in from elsewhere. The construction of the poorer-quality south transept roof, and the way in which rather ‘bodged’ carpentry had been used to bridge the gap between each of the transepts and the nave roof certainly emphasised their status as secondary added structures. It was not clear which of the transepts was the earlier, the poorer quality roof structure of the south being conclusive evidence for neither interpretation. (see below). The rafter holes or “problem holes” seen in the nave trusses are an interesting feature. If the interpretation of them as the fixing points of sprockets to carry an eaves-extension to the roof (Brunskill 172) is correct then the presence of them along the whole length of the nave indicates this was roofed in its entirety in a consistent way. This adds further weight to the notion that the transepts were added to an existing plan, presumably necessitating many alterations to the roof arrangements, (such as the insertion of the bressumers to carry the rafters over the transept entrances). One change occurring at that time may have been the removal of the sprockets along the nave eaves, leaving the redundant peg holes. The East Window of the South Transept. If the east window of the south transept , was as it appeared to be, an original feature, and the method of dating on the grounds of architectural style in windows is held to be reliable, then this provides a date of before c. 1300 for the construction of that transept. That is of considerable interest as it is an earlier date than is usually held for the church as a whole, and if as Pevsner claimed, the transepts are additions to the existing nave (certainly an interpretation indicated by the physical evidence of the roof structures seen during the works), this pushes the possible date for the earliest parts back to an even earlier date. The similarity of window tracery in north and south transepts may suggest similar or closely separated dates for their construction, although this offers little help in explaining the markedly different roof structures. Fig List 2 Plan of roof showing position of trusses 3 West Elevation of nave roof truss NR 2 4 South elevation of north transept roof truss STR 10 5 North elevation of south transept roof truss NTR 11 6 Interior elevation of east window in south transept 7 Exterior elevation of east window in south transept Plates 1 (Jon jpeg18) South side of nave, build of rubble and cob on wall top 2 (Jon 28) north side of nave, rubble build between rafters with plaster showing at top. "Problem Holes" visible in rafters above rubble. From NE 3 (jon 11) North transept, bottom of truss with and modern softwood "rafters" over N gable wall thickness. From SE 4 (Jon 16) South transept rafter, sole plate and ashlar piece from SSW 5 (jon 25) boarding beneath slates on south transept, from SSE 6 (Jon 04) S transept. collar and vertical post From N 7 (jon 57) Ferramenta in N light of east window in south transept. View from W 8 (Jon 59) , blocked base of east window in south transept exterior view from E References Westcott K. A. 1992: The Spire and Roofs of Hatherleigh Church, Proceedings of the Devon Archaeological Society, No. 50. Brunskill, R.W. 1994: Timber Building in Britain. London. Cherry, B. & Pevsner, N. 1989: The Buildings of England: Devon. London. Friar, S. 1998, A Companion to the English Parish Church. Godalming. Slater M. 1999: The Old Parish Churches of Devon, Folly Publications, Malvern Spurr D. 1983: Devon Churches, Merlin Books, Braunton