The Chimneys of Lavington Manor

Jarrod Hill and Rob Thomson

Lavington Manor, Wiltshire, was completed in 1865 for Edward Pleydell-Bouverie (1818-1889), fourth Earl of Radnor and Member of Parliament for Kilmarnock, for the sum of £67,000. The designer was Ewan Christian (1814-1895), a respected architect of the Church Commissioners, later president of the RIBA and Royal Gold Medal winning designer of the National Portrait Gallery in London.

Rambling Tudor in style, the house has stout diaper-patterned brickwork facades with Bath stone features and dressings, surmounted by a steeply gabled and parapeted, polychrome Cornish slate roof. The building is topped by 11 fine decorative chimneys incorporating differing relief motifs, enrichments and monographs. These chimney stacks are constructed with cantilevered geometric brick crowns of an intricate, holly-leaf plan form, on unique circular, octagonal and composite shafts comprised of large brick blocks.

Lavington Manor, Wiltshire

DEFECTS

By 2005 the building was exhibiting typical signs of 140 years of weathering, including decayed mortar joints. Several of the chimney stacks were in need of comprehensive repair and the slate roof coverings needed to be replaced. Close inspection during a condition survey revealed that the chimney stacks were actually unstable and prompt action was needed to make them safe. Agreement was reached with the local authority for emergency works and three of the chimneys were immediately taken down following measurement and photographic recording. The units were individually marked to record their location and orientation in the structure and were set out from the disassembled chimneys at ground level by course and stack, for each block and brick to be scrutinised for defects. One chimney had partially collapsed before this process was complete, making the job of piecing together the units for this composite chimney extremely difficult. It was, in all, a 2,500-piece jigsaw.

Typical chimney defects were observed arising from wind loading, wetting and drying cycles, poor earlier repairs with inappropriate mortars, sulphate decay from flue gasses, frost action and differential weathering of the softer blocks. Some of the voids in the redundant flues were large enough to provide good bat roosts.

Although it was most unfortunate that it was not possible to repair all the chimneys in situ, taking three of them down provided the opportunity to study their construction and gather valuable information for their repair, which would not otherwise be evident. For example:

• It was discovered that the crowns contained a total of 18 different brick sizes over 12 courses. When CAD drawings were overlaid to compare these with the coursed units on the ground, it was evident that even this number of brick types was insufficient, and two particular forms needed to complete the bonding pattern had been missed out, resulting in a historical ‘fudge’. The problems were compounded by the later reconstructions, which relocated bricks within the crowns. The two missing units were subsequently commissioned to complete the scheme.
• The bonding plan was found to be poor, particularly on the composite chimneys with incomplete octagonal and circular shafts. This was remedied with stainless steel bed-joint reinforcement and simple brick repair techniques to the core.
• A consistent pattern of vertical stress failure was evident parallel to the face of the large clay units, which connected the weaker ‘necks’ between perforations – effectively making the blocks two separate units. In addition, vertical cracking was evident where one portion of the brick had been restrained by adjacent masonry and the other free to move. In some cases these were pinned and elsewhere reinforcement was carefully inserted into bed-joints to stiffen those chimneys which remained intact.
• The perforations of the blocks were un-grouted, leaving substantial voids for water retention throughout the structure and reducing the potential strength of the chimneys. The different grain, density and fissures in the block also aggravated decay by frost action.

MATERIALS AND PRODUCTION

Bricks from the 'holly leaf' crown of one of the chimneys showing typical deterioration

The materials of the chimneys were entirely local. Gault clays, which form the edge of the Salisbury Plain, were used by Holloway Bros (formerly Fox) at their local brickworks at Market Lavington. Lime came from Broadwell Kiln and aggregates from the adjacent greensand belt.

The local greensand mortars are typically very soft with a poor matrix of small spherical aggregate, and ordinary non-hydraulic lime is insufficiently durable for such an exposed location. As a result, most chimney joints had extensively weathered or had completely washed out. Clearly an ongoing problem, earlier campaigns of repair undertaken with thinly applied cementitious mortars had only aggravated the situation, leaving the joints with a brittle outer shell but little strength. Frost ‘jacking’ as trapped moisture expanded and contracted with the freeze-thaw cycle had caused the crown of one stack to shift almost 150mm (six inches) and corrosion of iron bed-joint reinforcement had made some chimney sections extremely unstable. During the 1950s several crowns had been re-built and one of the chimneys had had three of the shafts replaced with plain brickwork.

The chimney crowns comprise pentangle and bow-tie shaped bricks of varying height up to 400mm (16 inches) in length. The shafts contained 200mm (eight-inch) thick blocks of 300mm (12 inches) in height by typically 500mm (20 inches) in length. Manufacture of these larger units must certainly have tested the traditional brick-making techniques of the day and certain problems which demonstrate this are present on the chimneys.

The clay for larger, complex shaped units was thrown in several ‘handfuls’ rather than (as with a regular brick) in one go. Where, for example, clay was poorly blended (hard work by hand) or the clay was dry (not uncommon during a summer brick-making season), this could lead to seams of differential moisture content in the clay ‘loads’. Poorer workmanship could also have led to other problems including air pockets, unstable stone inclusions and variation in the grain of the fired clay where laid from different directions. Decorative detail on the mould face required clay to be very well pressed in when throwing the brick into the mould, whereas less force may be applied to the subsequently placed ‘loads’ causing variation in density and strength characteristics within the brick.

During drying, ‘green’ bricks can change shape and shrink differentially, so to aid stable drying and firing larger units are usually made with holes or voids through their core to increase the proportion of surface area and reduce the overall mass of the clay. However, with larger blocks this is hard to perfect and the span and height of units inevitably means that in some places the clay is much thicker than others. It is no doubt for this reason that the larger bricks found at Lavington have up to three holes of varying shapes, created with a tool similar to an apple corer and routinely cleared by hand. This action often dragged inclusions to create depressions and exposed the seams and air pockets discussed above.

A repaired brick stack

The temperatures in traditional clamp and up-draught kilns typically vary from 750oC to 1,200oC, so larger brick units were normally produced in a down-draught kiln which achieved the more even and consistent heat needed. Nevertheless, fluctuations in temperature were common and variations in quality were inevitable. Lighter coloured units, in this case those with a more orangey colour, generally indicate a softer ‘fireskin’, which is the dense protective surface of a brick caused by the partial vitrification of finer particles during firing. Bearing in mind that the core of any brick will be less well burnt than the brick face, the core of these lighter coloured units will also be weaker and more vulnerable to frost as a result. Those which were exposed to too great a temperature could, on the other hand, suffer from too much vitrification, making them brittle.

These characteristics, along with the differential densities, grain and inclusions of the brick units compromised their ability to cope with mechanical strains typical in a chimney stack.

CHOOSING NEW TERRACOTTA

As the original brickworks was no longer in operation and the local clay is no longer extracted, new special bricks for the repairs were sourced from a good geological match, produced by the Pitsham brickworks of Lambs Bricks & Arches in the traditional manner.

Any decision to vary the material or technique employed in the repair of historic fabric should never be taken lightly. However, for the larger brick units, changes in manufacturing techniques and a better understanding of the defects required the producer to reconsider the method for producing new units for the repair, and terracotta manufacturing techniques were employed which offer considerable benefits for units of such size:

• matching clays could be well blended and thoroughly screened to eliminate inclusions which could split the unit during firing
• the clay could be built up incrementally, in small quantities, and well pressed into a two-part moulds, thus avoiding the laminations and grain evident from larger quantities of thrown clay
• computer controlled drying and firing ensured dimensional stability and allowed for a more consistent and controlled kilning
• accurate predictions could be made about the shrinkage and movement during manufacture, and consistency of supply for a large number of units is achievable.

Nevertheless, the brick units from the existing chimneys characteristically vary individually in dimension and appearance (not only in strength and performance) and these valuable qualities require consideration in any new work. By comparison, one downside with new terracotta production is that these variations are less desirable, and manufacturing subtly different moulds for each block was not feasible in a small project such as this. As a result the quality of the new blocks are more uniform and regular than the original, but they retain significant similarity.

CHOOSING NEW MORTAR

The existing mortar with which the chimneys had been built had all but disappeared in many areas, leading to 100 per cent repointing on some of the chimneys that were not dismantled. Two different pointing mortars needed to be devised that matched well with the existing mortar but that should give an appropriate level of both strength and porosity. An improvement was needed on the purely greensand mortar that had originally been used. The mortar for the vertical sections of the chimneys was created using English weakly hydraulic lime (NHL2) from Hydraulic Lias Limes with a mix of Warmwell grit sand as the main aggregate, and greensand stone dust (from Wessex Dimensional Stone) as a finer aggregate and for colour. The crowns of the chimneys were felt to be particularly vulnerable, given the dramatic nature of their cantilevered construction, so for these the weakly hydraulic binder was replaced with a French strongly hydraulic lime (NHL5) from St Astier. It was felt that this would impart greater strength to this more vulnerable part of the construction.

On balance, any disadvantage caused by the more regular appearance of the new bricks is minor. Close inspection will, in the future, clearly reveal which elements are new, and no attempt has been made to deceive. Through careful refinement of the traditional brick manufacturing techniques, and reconstruction using sensitive details and more durable hydraulic lime mortars, these chimneys are likely to last for many years in the future.