Restoring The 1761 Kirkman Double Manual Harpsichord By Tom Strange.
The nameboard of the 1761 “Queen Anne” Jacob Kirkman harpsichord.
When the Carolina Music Museum was awarded the Marlowe A Sigal collection in 2019, it was known that a number of extraordinary keyboards and woodwinds were among the holdings, and plans were immediately made to mount an exhibit of the gems of the collection in late Spring or Summer of 2020, under the banner of ‘Sensational Sigal: Highlights from the Musical Instrument Collection of Marlowe A. Sigal’. Most of the instruments to be exhibited needed only cleaning and minor repair, but among the most anticipated harpsichords was the 1761 Kirkman double, said to have been the wedding present of King George III to his bride Charlotte, and possibly the instrument that the nine year old Mozart played upon in his debut before the royal court. It is in very good original condition, retaining most of the original jacks and tongues, but with all cloth and perishable materials replaced. However, on inspection and comparison with photos made in 1987 at the Christies auction, it was clear that the move to Boston and the dry winters there had been hard on the instrument, and part of the soundboard had come loose from the belly rail, and under tension from the case was popped up, blocking the 4 foot strings and several of the jacks. Additionally, it had a ghostly fog over the key well and marquetry, unlike our 1758 marquetry Kirkman, a twin sister, which is still very clear.
Full view of the 1761 Kirkman harpsichord
Under museum controls, every approach to an instrument for any reason should have a purpose and a plan even if very simple. For our museum, where a fair number of the instruments are in playing condition and maintained as such, this example of ‘artifacts in use’ leads to some straightforward record keeping and protocol, introduced specifically with the acquisition of Sigal.
- Short demonstrations or a check of tuning condition are not recorded
- Practice for recitals and/or performances are logged with time played and artist
- Service calls to tune or to service quills or similar are logged with the action taken
- Repairs to fix some limited problem are logged with action taken and list of materials used
- Partial restoration or large scale repair requires a consultation with select subject matter experts, consideration of restore or conserve, a written plan of action if the work is to proceed, and a written log of actions taken and any observations made.
- Full restoration requires all the above with concurrence from our board of advisors and the museum board
- Reclamation (involving replacement of large sections of missing parts) involves all of the above plus a budget and a written reason why an instrument should be a candidate for such thorough treatment. The plan should show how the original engineering intent is to be preserved.
After a thorough examination of the Kirkman, we determined that the level of intrusiveness fell between a repair and a partial restoration. As such, we developed a short plan of attack which essentially reversed several actions taken in 1911 at Broadwood and Sons and addressed the problems that came up in Boston. Specifically, in addition to restringing and bringing together or shimming a few soundboard cracks, Broadwood had shellacked the entire soundboard and the keywell marquetry and jack rails. They did not add anything to the outside of the case, and a wooden block added inside to the cheek corner to reverse or halt cheek warp was left in place, as it was doing that job admirably, and had not intruded on anything historical.
Soundboard cracks as found
The strings were measured and it was clear Broadwood used their own thinking on scaling and material choice, as the strings changed diameter at points different than the very clear gauge marks on the nut, and even the brass was started at an odd point. The strings were removed and new wire from Stephen Birkett was ordered to replace it, sensitive to the intent of Kirkman. With the harpsichord on its spine, a single cut was made in the middle of the bottom boards where two natural cracks had left a large potential opening exposed, just ahead of the middle bottom brace. This allowed access into the harpsichord without the disruption of removing the entire bottom, and could be closed up with original material and the integrity of the bottom restored, but now leaving an access hatch for future use. The engineering of an English harpsichord requires the bottom to be on and intact for full structural integrity.
Once inside, the separation of the soundboard from ribs and belly rail were clearly seen. At this point the top rim of the harpsichord was padded and 2 X 6 pine timbers were clamped to bridge across the instrument, allowing the soundboard to be pressed back into place with a series of wedges and blocks acting against the timbers. Hide glue was used to rewet the surfaces and, working one crack at a time, the soundboard was brought back together. Historically, the seams were usually covered from the inside with glue soaked linen strips, which we still use today. However, for cracks that might have enough stress to reappear, I have taken the liberty of using carbon fiber cloth soaked in hide glue to form the patch. Carbon fiber will not oxidize and degrade with time, it adds no more material weight than linen, and it is tremendously strong. Two pieces of soundboard so glued up make a formidable joint, but it can be reversed with warm water and removed completely without trace. Once dry, the sonic properties are good as well.
With the soundboard repaired and the bottom closed up again, we turned our attention to the foggy marquetry decoration. A cursory examination showed the presence of shellac, a material known at the time but never used then for a finish coat. This was clearly added without taking the action out; you could see where it sagged and ran past key levers. It had interacted poorly with the original finish and had ‘bloomed’ or pulled slightly away leaving the dull fogged look. On applying alcohol to a test spot below key level, the shellac came up leaving the original finish showing cleanly underneath. With the action out, it was decided to remove this shellac and small areas were worked at a time. The original oil varnish is insoluble in denatured ethyl alcohol, but spirit varnishes like shellac come up nicely. Once the shellac was off, the original finish and fillers were clear and nice again, making it hard to understand why they shellacked things at all? A single coat of microcrystalline wax was all that was needed to bring the look together.
Before and after cleaning
The shellac on the sound board was not applied very nicely, but it was intact, the soundboard had remained almost stain free, and we decided to leave well enough alone there, after a thorough cleaning with soft damp terry towels. The shellac does protect the wood, it does not hurt the sound, and it is now part of the history that does not detract from the look of the instrument. It did have 110 years of dust and dirt on the soundboard, which was not only unattractive, but accumulated grime is corrosive and damaging to both metals and wood. I do not like to see an instrument scrubbed, varnished, and shined up to some impossible state of brightness, but a really dirty instrument is in trouble from a conservation point of view.
Kirkman marquetry harpsichords with their burl wood case veneer have a lower molding of veneer-covered oak that must have been the tastiest eating for wood boring beetles that exists, as these moldings are invariably found under some level of attack. The 1761 has only light attack, and had minimal loss of veneer. That said, nothing encourages additional veneer loss like a bit missing, so new material in European walnut or laburnum was cut to size, marked with date and initials on the back, and applied to fill the small gaps. From 6 feet it is seamless, but from 6 inches you can see the replacement, and as it is also marked, it will not distract a future historian.
With the case complete and the action cleaned, key lever interference addressed (small corrections made by heating and gently bending the offending key lever to stop it rubbing a neighbor) and the action installed again, we began stringing. I am using P-wire by Stephen Birkett of the University of Waterloo, Canada which I think has the closest historical basis and which sounds extremely fine as well. Many builders are now using Birkett wire exclusively, and as always, Stephen will work with the team on scaling considerations to achieve the correct diameter for a given gauge marking from a given time. The use of iron and brass are clear from the nut markings, but Stephen was of material help with the transition from yellow to red brass as needed.
I make hitch loops with the wire doubled back on the twisted area in a figure 8 pattern, and the tag end cranked loose, leaving the finished end smooth to prevent ‘finger pricks’. Such a hitch will not unwind but looks far neater than the little flags that Broadwood had left, which do poke the person that touches the hitch rail. Broadwood had also drilled every tuning pin and I largely ignored this, preferring a hand wrap to beckets and such. I did use the holes for a few of the brass wires where the roughness of the tuning pin could initiate a premature break at tension. The original blacking was still evident on the tuning pins so I did no more than brush off loose grime and made no further attempt to clean pins.
The last step is returning the jacks to the registers from the custom holder I keep for holding jacks during work. The jack should fit smoothly in the register, and then we can address the quilling. I like to achieve the original sound of crow quills by using crow quills. They are available readily and shaping a quill is not difficult. One flight quill will service 3-5 jacks.