Strings Magazine

Jan/Feb 1994

Caption: Gregg Alf and Joseph Curtin compare the plaster mold, the copy, and the original Guarneri instrument in their Ann Arbor workshop.

Violinist Elmar Oliveira recently acquired a beautiful example of Guarneri's early work, the "Lady Stretton" of 1726, once part of the Hottinger collection. this marked a transition in his concert work away from Stradivari; he previously owned two violins from the maker's early period. The new acquisition prompted him to commission a copy of it from us, replacing our replica of the "Booth" Stradivari that he had been using as one of his concert violins since 1990.

An initial visit with the "Lady Stretton," to take photographs and measurements, was set up at a hotel near Lincoln Center in Manhattan. (The city was still echoing from the bombing of the World Trade Center earlier that day when Mr. Oliveira dropped by between rehearsals with the New York Philharmonic.) Back in Ann Arbor, while my memory was fresh, and using the photographs taken in New York for reference, my partner, luthier Joseph Curtin, and I selected several sets of wood for the replica. The top (rather wide-grained spruce) was easy, but the back and ribs were more difficult. The two-piece back is of beautiful but narrow-flamed maple - not much sought after for new instruments, so we didn't have a lot to choose from. The ribs were only faintly figured in places. Although we maintain an extensive wood inventory, we need to keep reminding our selves that it is the plain, the knotty, and the weakly flamed pieces that are so often required when doing exact copies.

When it came to the scroll, however, we were prepared. During a recent shop ski trip, we sat around the fire talking about violin making and the upcoming year's work. As we joked about how hard it is to find scroll blocks plain enough for copies, someone pulled a burning piece of maple out of the fireplace and said, "No problem." that piece, extinguished and transported back to Ann Arbor with several others from the wood pile, made its way on to the short list for the "Lady Stretton." In the end, a better match (European, with lighter medullary rays) was found, but someone in the Berlin Philharmonic Orchestra now owns and loves the fireplace scroll!

With the wood selected, top and back joints prepared, and tools sharpened, we got the original Guarneri del Gesu in June of 1993. The first step was documenting the setup - bridge and post position - so we could put everything back in the same place later.

Next we began preparing top and back outline templates. We avoid using photographs or outside tracings to generate these. Distortion due to parallax, as well as 250 years of wear, is difficult to compensate for (and looks dreadful when transferred to the purfling outline). Instead, we use the purfling as a line of reference, as it is more or less unaltered by time. To copy the purfling outline, we cut a hole in a sheet of clear acetate large enough to accommodate the arching, so the acetate lay smoothly over the purfling and edges. the purfling and edge outline were transferred directly onto the acetate by means of a series of tiny pinpricks. Separate templates were prepared for the top and the back, which differed significantly on the "lady Stretton," as has been the case with all the instruments we have copied.

We created an unworn edge outline by drawing a line four millimeters out from the purfling outline. This was transferred to the piece of wood, which was then cut out and purfled in the normal manner (though it is quite a trick to get the variations in width and cut so characteristic of Guarneri's work). Later, the edges and corners were worn down; we usually are conservative at that stage, leaving them a little fuller than the original.

The majority of our copy work is done from palster casts of the instrument. On previous occasions, we had used lead foil as a separator to protect the varnish from the plaster. This time, however, I wanted to capture an impression of the entire, including a few millimeters of the rib garland. This would not be possible with a plaster mold due to the overhang of the edges. however, by pouring an eight-millimeter-thick RTV (room-temperature vulcanizing) rubber mask over the arching, around the edges, and down a little past the ribs, we were able to create a cast flexible enough to later clear the undercut areas. A conventional plaster mold was poured over the RTV, then used to hold the rubber in its original shape once it was peeled of the insrument.

With this technique one can create castings of the entire scroll. Care must be used in selecting a product that doesn't adhere to or affect the varnish, is tough enough to be peeled back from the edges without tearing, has an acceptable set-up time, and cures with minimal shrinkage. Another potential problem is that some types of rubber emit alcohol-based fumes on curing, or are inhibited in drying by contact with certain residues present on the surface of the instrument or found in many brands of the modeling clay used in setting up the casting. (Special noninhibiting modeling clay is available.)

The rubber casting technique is not completely new. A colleague kindly introduced me to an alginate gel which violin makers have used for some time in scroll and undercut castings. Bt, being water based, the gel shrank rapidly after the instrument was removed, was good for only one positive cast, and required a thin layer of oil to be applied over the instrument to protect its finish. RTV rubber, being completely inert, gets by these shortcomings quite nicely.

Once the castings were complete and checked against the original for accuracy, we began developing working templates. Using casts spares the original a great deal of unnecessary handling. For example, the palster cast made from the RTV mold had hard f-hole outlines, and it was easy to cut f-hole templates by using clear acetate temporarily adhered to the plaster cast with spray glue - not to be done on the original! Similarly, scroll templates and arching guides could be made without worry.

Another note on protecting the original: when they are not in use, we keep instruments in a special vault originally manufactured to protect magnetic media. Normal fire safes are designed to keep paper below its flash point. The lower temperature rating required for protecting magnetic media from heat erasure is much better for the safe keeping of fine stringed instruments.

We found other benefits in working from castings. For example, with a plaster cast that extends past the top and back edges to include a cross section of the rib garland, we could reproduce and exact rib outline, which of course differs for top and back plates. Rib templates were prepared in clear acetate, then aligned one atop the other, so as to reproduce the relationship found in the original. A reference hole was drilled through both templates and into stock prepared for making the form. The two templates were then transferred to either side of the rib-form stock (itself three layers of wood, aligned with pins). Finally, the form was finished to blend the top and back outlines into one another thus reproducing the distortion, buckling, shrinkage, variations in overhang, and other idiosyncrasies seen in the original ribs. (We used a three-part form, the outer parts of which are removable, to allow installation of the linings while the rib garland is still supported by the form.Contrary to what one is taught in lutherie school, removing the rib garland, both sets of linings in place, is not difficult as long as the back is not glued on.)

The amount of distortion to be reproduced in the plates is something to be considered carefully before the arching is undertaken. In copying the "Lady Stretton," we attempted to capture most of the back distortion. Because of this, the ribs had to be carefully fitted to the varying planes that resulted. The top was more or less true to a plane, and the soundpost distortion relatively small. We made no particular effort to copy exactly the graduation of the original. Instead, we went by weight and tap tones, in an effort to match our plates to the acoustical properties of the original.

The process of reproducing the indescribably beautiful finish of the original violin more resembles painting an oil portrait than varnishing a new instrument. It is tremendously important to have wood that matches the reflectivity of the original, along with the curl and medullary rays.

A basic palette of Indian Yellow, Ultramarine, Rose Madder, and Quinacridone Orange (versions of which are commercially available as artists' oil paints) proved adequate for pigmenting the varnish. The varnish itself was based on cooked pine resin and Florentine linseed oil, the final finish consisting of three coats, variously worn, chipped off and dented. Many layers of "fake dirt" (we used water based pigments and inks) were put on and partially removed to give the patina so characteristic of antiques. When doing this kind of work it is necessary to consider the precise color value and transparency of this dirt, which varies as much as the color of the varnish itself.

We find copying varnishes to depend much more on one's eye for color and texture than on the particular mediums used, though of course a happy selection of materials an immensely simplify the process. Similarly, no amount of template work will guarantee convincing woodwork. In the end, it is the minute adjustments one makes by eye that creates a first-rate copy.

And this is how it should be. We try to do an exact copy about once a year, because we know no better way of training our eyes, discarding misconceptions, and forcing ourselves to reconsider stale workshop habits. Our copy of the "Lady Stretton" was finished in early September and played on stage by Oliveira a week later. Delighted and exhausted, we got back to our regular jobs, each making our own instruments, using personal interpretations of the models developed while copying, and luxuriating in the freedom that comes with not having an original on the bench!

back

1221 Prospect Ave., Ann Arbor, MI 48104

(734) 665-2012   fax: (734) 665-4623

email: violins@alfstudios.com