St John the Evangelist, Shobdon
A Conservation Case Study
|The Church of St John the Evangelist, Shobdon,
Herefordshire (All photos: Frederick Gibson)
|View of the church’s Rococo Gothic interior looking towards the chancel with its pendant arch and, to either side, north and south transept pendant arches|
St John the Evangelist, Shobdon, is a small Grade I rural Anglican parish church adjoining the former parkland of the Shobdon Estate. The tower is the only element of the original 12th-century church to be incorporated into an extensive rebuilding (1752-1756) in a Rococo Gothic style by Richard Bateman. Bateman was a friend of Horace Walpole and a member of the ‘Committee of Taste’, then exploring the use of different styles of architectural decoration for Walpole’s own house, Strawberry Hill. The direct connection to Walpole strongly influenced the design of the church at Shobdon, which survives, subject to some re-ordering in 1907, as the most notable Rococo Gothic style church in England.
CONDITION PRIOR TO REPAIRS
The nave and chancel roof were extensively stabilised and repaired in 2002. However, insufficient funds meant that guttering and roof coverings to the transepts had not been attended to, and the parapets of the chancel and transept were leaning inwards. Water ingress had led to the breakdown of the decorative plasterwork and provided perfect conditions for the development of rot and beetle infestation. This, in turn, affected the structural support offered by the ceiling joists and laths.
Localised opening up showed that movement in the parapets was caused by:
- rot to damp wall-plates
- movement in the roof structure
- inadequate support from the unconsolidated core of the wall below.
Without intervention, the crenellated parapet could have collapsed inwards, causing considerable damage and placing the public at risk. During the exploration of the parapet it became apparent that the bearing of the chancel’s principal rafters onto decayed wallplates was affecting the integrity of the roof and ceiling structure. So too were defects in the ceiling joists, coving ribs and ashlar pieces (the short vertical timbers that rise from the inner surface of the wall to meet the common rafters).
A major constraint was that the internal plaster finish of the church was judged of greater historic significance than the structure that supported it. This resulted in a strong presumption for the repairs to be carried from the external side only in order to conserve the plaster in situ.
A measured survey was carried out of the chancel, transepts and crossing at roof level to provide detailed plan and section information. The parapet elevations were photographically recorded and rectified images were produced to provide a detailed record of the structure and to assist with specifying the repairs.
An ecologist was appointed because there was a high likelihood of bats being present and their habitat would be disturbed by the proposed works. Ironically, it transpired that the roof space had been free from bats but while the roof was left open during inspection work (beneath a temporary roof) three lesser horseshoe bats moved in. As a result, a European Protected Species licence was required and a temporary screen had to be erected to provide a bat sanctuary for the duration of the work. A bat access slot also had to be provided to compensate for the loss of access once the slate roof was reinstated.
|The finished work viewed from the tower showing the reconstructed parapets around the nave, transepts and chancel, new leadwork to the transept roofs with improved ventilation, and a fall protection system fixed around the roof to aid maintenance|
The structural engineer’s initial investigation of the parapet included dismantling part of the transept wall down to cornice level. The investigation established that the 250mm wide x 75mm deep oak wall-plates to the roof timbers had been built into the wall below the parapet level and were heavily decayed. A second decayed timber bulk was found built into the wall at the level of the ceiling joists and internal cornice. The parapet was also partially built off poor fill between outer and inner wall faces.
These conditions were causing the tilt in the wall. Measuring the verticality of the wall showed that, in most cases, the centre of gravity was close to or extended outside the middle third of the wall. The structure was consequently deemed unstable. The structural engineer was also concerned that the wall-plate would continue to decay and the wall would continue to move.
Roof timbers and ceiling support
A survey of insect damage to the accessible roof timbers was commissioned to devise a management plan for current beetle activity. When wood-boring insect flight holes are discovered in structural timber it is often assumed that the remainder of the timber will succumb to attack. Without chemical modification of the timber by fungal decay, however, only the sapwood is susceptible to insect attack at moisture contents over approximately 16 per cent. The heartwood remains impervious to attack due to the various chemicals it contains and that confer its durability.
The visible damage in the roof timbers was confined to the sapwood edges of the historic oak. The structural significance of the damage therefore depended on the proportion of sapwood in the timber section. There was clearly a history of perceived problems with insect infestations as many already eroded sapwood edges had been treated with insecticidal paste.
Historic powderpost beetle attack (which tends to occur in the first 25 years after construction) was recorded in the sapwood of a significant amount of timber. The insects produce large quantities of bore dust that is not compressed into frass pellets, and the resulting increase in volume causes the surface of the timber to rupture. The damage is often mistaken for furniture beetle infestation because both produce small round emergence holes, but powderpost damage is quite distinctive, especially when the absence of pellets is demonstrated with a hand lens.
|The north transept roof before work began: note the parapet’s pronounced inward lean|
The survey for beetles strongly suggested that a widespread deathwatch population no longer existed although there could be residual populations in the lower wall-plates where moisture levels were higher. The roof was generally dry and where timbers remain dry deathwatch beetle populations will not become established. Replacing the damp wall-plates (using oak with the lowest possible proportion of sapwood) would therefore remove the main beetle populations while natural predators such as bats and spiders would control any remaining beetles. It was decided that chemical treatment of the timbers was not appropriate or necessary.
Thermal imaging was used to survey the pendant arch and beam above the south transept opening where cracks and deformation in the plasterwork were evident. The survey was reasonably successful in determining the lay-out of a timber ‘goalpost’ structure with two intermediate hanging posts and associated framing to create the pendants. However, localised opening up of the plain plaster was still necessary to check beam ends where built into damp masonry, and the junctions between the hanging structures and the main crossbar.
Micro-drilling of the main oak timber beam ends over the south transept arch established that the cross-sectional area of heartwood was sufficient for bearing and shear, despite the considerable degree of beetle attack to the outer sapwood. The proposed strengthening of the beam ends with steel splices was therefore unnecessary.
The south transept pendant arch framing was found to be a mixture of oak, lime, sweet chestnut, beech and poplar of such variable quality that some elements had to be replaced due to beetle attack. Iron fixing cramps to the goalpost structure had pulled out of the wall and the hanging framework forming the pendants had mortice and tenon joints to the goalpost crossbar with defective pegs. To ensure transference of loads, these joints required mechanical fixings, as did the goalposts to the external walls. More extensive opening up to fully inspect all timbers and joints within the hanging pendant arches was considered. It was a difficult decision to make because of the intrinsic importance of the decorative plaster elements and because some of the lath was weakened. Lath and plaster can provide significant strength and stiffness to a structure and it was decided, based on observations during the limited opening up, that more extensive opening up might do more harm than good.
|A detail of the foot of the chancel roof following
removal of the parapet, cornice and dentil course,
which shows extensive decay to the outer section
of the double wall-plate (bottom), and the poor
connection between the principal rafter leg and the
|The same detail after repair, showing the new single
wall-plate with a stainless steel bracket to the
principal rafter leg to ensure a positive structural
connection: new ashlar pieces (the vertical studs which
rise from the wall plate to support the roof plate) can
be seen to the right of the principal rafter leg
|The inner leaf of the south transept parapet was rebuilt in brick and hydraulic lime. A new oak wall-plate was fitted and the ends of the rafters were cut back to prevent them being built into the parapet.|
|New timber framework and stainless steel lath trays
were added behind the existing cornices. Suspension
wires were fixed to steel flats above and plaster of
Paris was poured through the mesh to bond with the
existing plaster nibs.
The findings in the south transept were provisionally used as a guide to the condition of the other two pendant arches. However, once work had commenced on site it was discovered that various parts of the south pendant arch and the beam above had been previously opened up and repaired, and that the north pendant arch had been completely reconstructed, possibly 40 years ago, in steel angle frame and fibrous plaster. The chancel pendant arch, which was constructed wholly from oak, was found to be in particularly good condition.
A specialist plasterwork survey was commissioned to assist in the specification of repairs. The plaster nibs which key plaster to laths were generally intact throughout most areas of coving and flat ceilings, although it was noted that they were set a little too close together in places. The major problem was the extent of disintegrated lath and the consequent lack of support of up to 75 per cent in some areas. A system of lightweight supports to reintroduce structural integrity was developed.
A CCTV survey was carried out to assess the drains and soak-aways. This showed that the below ground drainage was in poor condition with various fractures, misaligned joints and build-ups of debris. One soak-away was full of water and another was in poor condition.
WORKS CARRIED OUT
The aim of any repair to a historic building is to conserve its historic fabric with minimum disturbance. The extent of intervention has to be carefully considered in each case, assessing each element of structure or fabric both individually and as part of the whole building. At Shobdon the repair was essential to the structural integrity of the building, and the decay to the chancel roof at wall-head level necessitated much greater structural intervention than initially envisaged to ensure that the loads from the roof were distributed more evenly onto the walls. While the distinct character of the building is more evident internally, the castellated parapet is a typical feature of the Gothic style and the extent of reconstruction necessary was a concern.
Parapets and wall-plates
Various methods for stabilising the parapets were reviewed. One option that would have minimised loss of fabric was to inject grout into the voids formed by the rotten timbers and stitch the masonry with vertical stainless steel rods and resin anchors. However, as the injection would be done ‘blind’, its effectiveness could not be assured. Another option was to open up internal finishes, cut out defective timbers and replace them with masonry, but this would not address the parapet’s tilt and would cause considerable damage to the decorative plasterwork.
A third option was to dismantle the parapet, repair the wall-plates and reconstruct the parapet using as much original stone as possible. However, the parapet had been constructed from a local siltstone which was susceptible to expansion in its horizontal beds, and the more weathered stones were already exhibiting horizontal delamination. Dismantling the parapet would release pressure from the stones, exacerbating delamination and increasing the amount of replacement. However, this approach offered full access to the wallplate timbers with minimal disturbance to the internal plaster and was consequently adopted.
Opening up the structure revealed double wall-plates supporting the chancel roof, with the principal rafters rebated into both the inner and outer plates. The timbers were in a far worse state than anticipated and required complete replacement. Reconfiguration of the wall-head using a single wall-plate ensured that the timber would have air all around it and that the parapet above could be built off the solid masonry of the outer half of the wall.
The configuration of the roof and the parapet suggested that the roof had originally been intended to have overhanging eaves, and that the parapet and parapet gutter had been an afterthought as the 1756 reconstruction proceeded. This could explain why the truncated rafters terminated on a roof-plate along the top of the ashlar pieces above the inner wall-plate, and why the outer timber wall-plate was built into the parapet.
The replacement of the wall-plates in the chancel and transepts required the roof structure to be supported while new timber was inserted and the wall-heads consolidated. Initially it had been hoped that all temporary support of the timber roof structure would be from above to avoid interference with the historic plasterwork. This approach was feasible over the flat-roofed transepts, which have a lattice work plaster ceiling, as a steel beam could be propped off the transept wall-heads with temporary steel hangers attached to each ceiling joist and rafter. However, replicating this approach on the pitched chancel roof was unrealistic. Here the main trusses had to be supported on scaffold props from below. This meant that the ceiling had to be punctured, but only through flat plaster, and on three sides of the chancel the coving needed to be replaced in any case.
Replacement of the chancel wall-plates was carried out in sections to ensure that the roof structure was always connected to the walls along at least 66 per cent of its length. As a precautionary measure the decorative plaster frieze, pendentive cornices and ceilings were fully supported from an internal birdcage scaffold for the duration of the works. In rebuilding the parapets the opportunity was taken to replace the delaminated and eroded dentil course and various cornice stones which had lost their weathering capability. Their function in shedding water and the retention of the architectural lines provided by these important features were considered more important than conservation of the decayed stone.
The parapets were rebuilt with new brick internally and maximum reuse of the original stone facing. Each stone was numbered and built back as close as possible to its original location. The original stone, a local calcareous siltstone with poor weathering qualities, is no longer quarried. Second-hand stone was not available in large enough quantities or sizes. Ideally, a stone with a similar petrological profile would have been used for replacements but none was available. Another consideration was the need for a resilient stone for parapet copes and cornices. It was therefore decided to use Forest of Dean blue and green/grey stone which will gradually tone in with the existing.
The original ashlars have a rough hand-dressed texture, in some cases axed. To integrate the new machine-cut stone a palette of four stone-dressing variations was developed by the masons, not replicating the original but in the same character. A furrowing technique to match existing cornice and blocking course stones was also developed. Masonry work was commenced in lime putty but contractor liquidation prevented completion before winter, and the work had to be completed early the following spring when weather dictated the use of a hydraulic lime mix.
Roof timbers and ceiling support
Structural timber repairs to the perimeter of the chancel roof mainly consisted of the replacement of the wall-plate, introduction of additional ashlar pieces to support the eaves roof-plate and the installation of stainless steel brackets to all timber connections. In cases where the ceiling joists were no longer supported at their ends (the internal plaster coving had become the structural support), hangers were designed to suspend the joists from purlins above. These all improved the structural load paths down into the external wall.
|The north transept beam encasement after partial opening up. Existing plaster was retained wherever possible.||The beam encasement with the pendentives reinstated|
Structural timber repair to the transept roofs consisted of splicing new timbers to rotten rafter ends, doubling up beetle-damaged ceiling joists, reducing ceiling joist ends to enable support without being built into the masonry, and again the installation of stainless steel brackets to all existing timber connections.
Where the ceiling lath had been attacked by beetles, new support was provided by threaded stainless steel rods suspended from stainless steel flats spanning between the ceiling joists. The lower ends of the rods were attached to the ceiling with bronze gauze washers bedded in plaster of Paris. Where several adjacent laths were rotten, expanded metal lath trays were also introduced, bedded into the plaster, to provide additional support.
|The chancel coved ceiling (west elevation) with suspended threaded stainless steel rods and expanded stainless steel laths set in plaster of Paris, bonded to the rear of the existing plaster nibs|
Cornices had generally lost support from their timber framework and lath. These were reinforced from inside the hollow core using horizontal reinforcement rods attached to expanded metal lath cages bedded in plaster of Paris on the back of the existing plaster. The rods were supported by suspension cables fixed to stainless steel flats spanning between the ceiling joists above.
Water penetration through the lead roof had caused rot to the timber wallplates supporting the transept beams and the resultant settlement caused cracking in the lath and plaster beam casing. Original pendentives were salvaged for reuse while a new moulded cornice was run in situ, to match the existing. Some pendentives could not be salvaged but interestingly revealed a lump of charcoal at their core, presumably used as a lightweight filler around which to mould the decorative embellishment.
Roof coverings and rainwater disposal
The works necessitated the removal of the roof slates, GRP parapet gutters and transept lead roofs to allow access. The lead roofs to the transepts were recast and reinstalled in bay sizes to current industry norms (they were previously up to one metre wide).
The narrow parapet gutters around the chancel have limited height in which to fall and, to avoid a complete reconfiguration of the transept oak roof structure to create more fall, the gutter was reinstated using terne-coated stainless steel sheet. This can be laid in longer lengths than lead and without the need for joints or steps, and has a far longer life expectancy than plastic. Vents were added to the flat roofs and at the pitched roof eaves to improve ventilation in the roof space. These will remove moisture-laden air, keep general fabric moisture levels in check, and help to control any surviving deathwatch beetle.
The rainwater outlets from the transept roofs were widened to reduce the risk of blockage. Overflow pipes were installed to the transept gutters in case of blockage at the parapet outlets or hoppers, and trace heating was installed to prevent them becoming blocked by ice. Repairs were carried out to the rainwater pipes and one plastic pipe was replaced in lead. The drainage system was renewed and re-routed to nearby lakes, thus avoiding the need for soak-aways in the heavy clay soil.