The examination and restoration of a colonial-era sofa.
Long hidden secrets of an Australian colonial-era sofa were revealed during its preparation for display in the Colony exhibition at NGV Australia in mid-2018. Investigative work had commenced at the time of its acquisition in 1996, however without a display date, the project was put on hold. The Colony exhibition provided a valuable opportunity to learn more about the sofa’s construction and the craftsman who made it, the different upholstery schemes throughout its 160-year history, and to undertake repair of the timber framework and reinstate the upholstery.
Removal of the modern ticking upholstery revealed more recent upholstery padding on the arms, but original padding on the backrest.
Background
Dated to around the 1840s, the sofa was acquired by the NGV as a gift from Michael and Traudl Moon. Uncovering the hand of the craftsman through the process of examination was one the most rewarding aspects of its conservation prior to display. Tool marks on an artwork can help date its manufacture, such as irregular saw lines indicating the use of hand tools as steam-powered saws were not in use until the 1850s. The sofa’s craftsman had used materials at-hand, with rough-hewn timber struts used to brace the front and back side uprights, even re-using a piece of timber previously cut with a dovetail joint*. Score lines* are visible on the carved front faces of the red cedar frame, which would have guided the carver. Refinements in the design were revealed too, with shallow channels cut into the side seat rails* to incorporate and secure the upholstery webbing*.
View of the side of the sofa, towards the front face, showing the heavily worn tacking margin* and revealing a block secured with hand-forged nails to brace the rough-cut stretcher bar.
Examination of sides in raking light revealed three recesses cut into the curved end rails to incorporate webbing*. Similar notches were found on the front and back tacking margins* of the armrests.
Examination and research
Examination of the surfaces and structure of the sofa, methodically from one end to the other, revealed past repairs and changes made over the years. These included a higher re-positioning of the backrest, modern wire nails used to repair a broken back foot, and a relatively thick modern-style coir (coconut fibre) upholstery padding stapled over the armrests.
The aim of the sofa’s treatment was to return it to its original concept, without erasing the evidence of its past use. This would mean the reupholstering of the sofa, a transformative process and one that requires significant detective work. As a functional object, furniture items can often be reupholstered several times, demonstrating layers of provenance while also often presenting more questions than answers. Under microscopic examination at least five different remnants of top cover upholstery fabric were found; a black horsehair cloth, two different types of black coated cotton (i.e., types of oilcloth), a modern brown vinyl (faux leather), and the ticking or striped fabric (both old scraps adhered to the timber frame, and more modern from the most recent upholstery scheme). There were also several scraps found of a modern mustard-coloured velvet.
Horsehair cloth, also known as haircloth, is an extremely durable material often used in the 19th century for the upholstery of high-use furniture. Made from the hair of horse’s tails it is woven in the shorter direction with a base weave such as cotton or linen in the longer length. It can be woven plainly or with patterns to create a contrasting satin-like shine. Of course, the width of haircloth is limited by the length of a horse’s tail, so this constrained the furniture on which the cloth could be used to those with a depth up to around 65cm. There are techniques such as joining with a seam, or adding panels called ‘pull cloths’ to extend the width, but these are less likely to have been used on this sofa considering the shallowness of the backrest padding with very little space to hide such extensions. Haircloth remnants were found under hand-forged tacks on both the sofa’s backrest and the seat rails, several of which were positioned under later repairs making it clear they are earlier. Considering the production of haircloth has been in decline since the 20th century and is now extremely expensive and rare, the haircloth remnants are likely to be one of, if not the, earliest upholstery choice.
View of the underside of a tack head showing a woven cloth with the aligned black-brown glossy horsehair fibres running in one direction, and the thicker cotton weave in the opposite.
The analytical technique Fourier-transform infrared spectroscopy (FTIR) was used to investigate the black oilcloth samples and to identify the binder* in the coating. This technique produces a range of spectral peaks that are characteristic of different known materials so we can match what the FTIR identifies the material as with the reference library of materials. Samples of each type of fabric were carefully removed from the sofa frame and examined. The first sample identified cellulose nitrate used as a binder*, which was a product introduced in the 1850s, unlikely to have been in common use in Australia until the later 1800s and was used up until the 1950s. The second oilcloth sample produced a characteristic peak for shellac, a natural resin secretion from a beetle. Shellac is a less common ingredient to make oilcloth, usually reserved for flat inflexible surfaces so this finding is curious, however considering several similar remnants were found tacked over one of the haircloth tacks, it must be later. The other upholstery remnant was a vinyl faux leather which was undoubtably modern. Therefore, this left the horsehair cloth as the likely original upholstery fabric.
During colonial times, imports of materials were slow and often unreliable, so supplies were sourced from as close to the shop as possible. The trade of timbers around the colonies for use in furniture making, such as red cedar from NSW to Adelaide, is well documented, however new specialist tools and materials were harder to come by. Examination of the sofa’s original backrest revealed the padding was constructed with curled horsehair and additional refinements such as the curved outer edges packed with cut straw. Royal Botanic Gardens of Victoria’s Herbarium staff were able to identify the straw as probably a Hordeum species (a kind of barley), a common food staple at the time and the other construction methods confirmed the backrest was original. Stuffed in amongst the curled horsehair padding on the backrest, we also found several textile rags – a common method, even now, to fill out upholstery quickly and cheaply.
Densely packed and cut straw creates the curved tips of the original backrest upholstery padding.
It’s common even today to stuff upholstery with what’s at hand, in this case some cotton rags can be seen in the backrest’s original padding.
The red cedar timber frame was in generally good condition, and the front carved face was preserved very well. However, the tacking margins* were badly damaged and splintered from so many tacks entering and being removed from it over its lifetime. Treatment was undertaken to stabilise these areas by applying an adhesive to re-join splintered timber and adhering paper strips to the tacking margins* to hold it all together prior to any reupholstering processes.
X-radiography revealed that the position of the sofa’s backrest had been raised in the past, with three scraps of timber nailed (with modern wire nails) to the bottom edge. While the additions were obvious to the naked eye, they would be challenging to remove as the x-ray revealed that each piece was glued and nailed to the other.
X-ray of the backrest, revealing the layered additions of timber joined with modern wire cut nails into the tack-filled original lower edge.
The timber varnish was analysed to understand its history. When viewed in ultraviolet light most surfaces had a bright orange appearance, indicating a likely shellac-based varnish. This was also found in areas which were hard to access and therefore less likely be over-varnished in a restoration, lending weight to the theory that shellac was the original varnish resin.
Defining the original upholstery scheme
The next step was to define how the upholstery padding of the seat and arms should look according to the period and origin of the sofa – should the seat padding be soft and cushy, flat and hard, or thick or thin? It is rare to find original upholstery and upholstery padding on furniture made prior to 1850s due to the functional nature of furniture, so one useful resource is to look at paintings and drawings from the time, but this is only of so much help as artistic license can hamper efforts for authenticity. Our search revealed few furniture examples to aide our restoration, with the most notable being the extraordinary Tasmanian-made Hamilton Inn sofa from the Tasmanian Museum and Art Gallery (TMAG) This sofa, with its steeper sides, has numerous wooden slats to support the seat which is padded with a loose mattress-like cushion. In contrast the NGV’s sofa has only two curved stretcher bars* and evidence of webbing* use, and the more shallow and undulating curves of the sides lend itself more to a seat to perch, rather than to recline on.
The carved ornament of pinecones on the NGV sofa points to a Germanic origin with classical revival scroll work, perhaps late Biedermeier style. It’s enticing to think that one of the two larger communities of German immigrants in the 1840s to both South Australia and the Western District of Victoria could have been the makers of the NGV’s sofa.
The Conservation profession is a collegial one and it is common to consult colleagues around the world who can assist, but also might be interested in seeing more examples than those in their own collection. Specialist Upholstery Conservator Nancy Britton from the Metropolitan Museum of Art, New York, confirmed that the coir padding currently on the arms (that we knew wasn’t original due to staples being used) was far too high and should be more hard and flat, similar to the sofa’s backrest. The seat rail, with its chamfered tacking edge, lends itself to a tacked-on seat pad rather than a loose one as seen in TMAG’s sofa. Of particular interest was the front edge of the seat, which would have had quite a sharp-angle and very little loft*, being generally flat and hard. Certainly not the soft relaxing comfort most of us would think of today. Britton was also able to share that the technique of using straw for shaped edges as seen on the backrest padding was more common in the 18th century, and rarely seen by the 1840s, so she surmised that the maker was likely to have been an older gentleman – therefore adding a personality to the story of our sofa. This research and consultation across different departments and specialists meant we had a direction with which to create the new upholstery layers.
The reupholstering process
The upholstering of furniture in a museum setting is a specialisation and not commonly undertaken by generalist conservators. A commercial upholsterer was engaged who worked on-site in our studio, and understood both the discussions we would need to have at every step of the process, and the attention to detail required to get the desired result.
The aim for the seat and side padding was to create a historically correct shape with the least amount of damage to the wooden frame, and to use the same materials as the original padding where possible. All the padding layers were secured with a staple gun, creating the least damage to the already severely weakened timber tacking margins*. Unlike the repetitive hammering of a tack to secure the upholstery, staples only require one impact and are much finer so cause less damage to the wood. So this modern method was utilised with stainless-steel staples due to their corrosion resistance.
The undercloth* was applied to the frame, loose coir gently massaged into position and secured with a stitched down overcloth*, then regulated* to create an even padding. A rolled edge was created to ensure the front edge of the seat cushion would be box-like, and a skimmer* of flock* was applied to flatten any undulations. The final preparatory layer before the top cover was a calico cloth pulled tightly down.
Upholsterer Keith Rogers worked with the NGV Conservation department staff to create the upholstery padding for the seat and arms. Here he is applying the flock skimmer over the hand-stitched rolled edge.
Selecting the top or show cover upholstery was a complicated process as we were not able to source a black horsehair fabric in the width we needed to accommodate the depth of the seat. After an exhaustive search, a black calendered* linen was found, with its surface polished to produce a sheen. In effect, the calendered surface gave a similar shine to the original haircloth.
The application of the top cover was slowly undertaken to ensure the correct appearance of the padding, and also included the making of bolsters for either end, considering details such as how many tucks or folds, piping/no piping, button or tasselled ends. The choice of trimmings was extensive, yet surprisingly few were high quality and appropriate in style for our sofa. After consulting our research for historical examples, the decision was made to use a simply woven gimp* braid, to match the sofa’s restrained upholstery and carved decoration. This was glued into position to avoid further damage from tacking or stapling to the fragile timber.
NGV Curator David Hurlston, NGV Conservator Suzi Shaw, and Upholsterer Keith Rogers discussing the seat profile and armrest padding prior to final attachment.
The careful removal of the later timber added to the backrest allowed us to reposition it lower down, lending the entire piece a very different aesthetic and feel. The backrest was upholstered, with the valuable original padding remaining beneath. The modern brackets that had been glued on to secure the backrest into the back face of the armrests, were carefully removed by sawing, chiselling, then planing them down to the original timber. The backrest’s re-positioning meant that the straw-packed curves on the backrest now lined up with the flat padded armrests.
The broken foot was repaired so that the join was tighter than the previous repair and structurally stable. This was done by removing the old glue joins, extracting the crossed nails, cleaning off the old glue, then re-adhering it in a more tightly fitting position. The foot was then adhered in place with a protein glue, and losses filled and inpainted to disguise the break. NGV Senior Art Technician Eamon O’Toole designed and welded up a metal structural auxiliary support to take the weight off all legs in both storage and display situations.
Conclusion
All the textile evidence of the previous upholstery schemes remains under the new upholstery, to allow for further investigations in the future if required. The research, investigation and consultation on this sofa is documented in a now substantial dossier, along with any samples that were taken from the object. The sofa itself has revealed its secrets, and is an important record of colonial furniture making techniques and materials. Through meticulous and extensive teamwork, the restoration of the colonial sofa revealed the hand of the craftsman, returning the piece to the handsome sofa it once was to be enjoyed for future generation to come.
The restored sofa, complete with lowered backrest, reinstated upholstery and bolsters, and a steel support structure to take the weight off the repaired leg.
Acknowledgements
Keith Rogers, Upholsterer at Alexander J. Cook Aust Pty Ltd
Nancy Britton, Conservator (Upholstery specialist) at the Metropolitan Museum of Art
Val Stajsic, Identifications Botanist at the Royal Botanic Gardens of Victoria
Glossary of terms
- Binder – a word used to describe the medium that holds particles in suspension. In the case of oilcloth, this is an oil or resin into which pigments are added.
- Calendered fabric – fabric that has been crushed and compressed as part of its manufacturing process, to create a very smooth and often shiny surface.
- Dovetail – a type of wood joint.
- Flock – a soft upholstery padding material, usually cotton, for the upper layers of a seat pad. This also prevents other coarser lower padding layers such as coir from pushing through the show cover.
- Gimp braid – an ornamental woven band of cotton, wool or silk woven tape that can be simple or highly decorative, applied to the edges of the upholstered sections to hide any tack heads.
- Hessian – a woven cloth or webbing made from jute.
- Loft – the height of the cushioning on a seat.
- Overcloth – the cloth applied over the main padding.
- Regulator – a long fine metal tool used to poke into the upholstery padding material and move it about to create an evenly padded cushion.
- Score lines – fine lines shallow cut into the timber to help guide the carver.
- Seat rail – the horizontal wooden framework that is at seat height, onto which the upholstery is attached.
- Show cover – also called the top cover, it is the final upholstery fabric applied to cover the chair padding layers.
- Skimmer – a layer of cotton flock used to even out any undulations in cushion heigh in prior to the final show cover being applied.
- Stretcher bars – a structural part of the wooden framework used to add stability and strength, such as timber bars passing from the back to the front of a sofa seat area. They also support the upholstery padding applied above. These are normally jointed to the seat rails, as opposed to slats which are often just placed.
- Tacking margins/tacking edge – the recessed cut edge of the seat or backrest, where tacks are nailed into the timber recess to secure the upholstery.
- Undercloth – the cloth, usually loosely woven hessian, that is applied first onto the seat frame after the webbing, but prior to applying the padding layers.
- Webbing – woven bands of hessian, that are tacked onto the wooden framework running side to side and back to front, creating the suspension onto which the main seat padding is applied.