JQURN

In This Issue:

Results of CASJ Readers' Survey

A Closer Look at the String Player's Bowing Gestures
Wood Testing by Tap-Tone
Tap Tones of Old Italian Top Plates
Special Feature: Historic Instruments
Measurements of Principal Resonance Modes of Historic Instruments
Reverse Graduation in Fine Cremonese Violins
Documentation of a 1604 Violin by Hieronymus Amati, Cremona, Italy
Early Tuning Pegs and Tailpieces
Curtate Cycloid Arching in Golden Age Cremonese Violin Family
Instruments
Working Methods of Early Classical Violin Makers, as Implied by the Results
of Recent Dendrochronological Studies

Catgut Acoustical Society

To increase and diffuse the knowledge of musical acoustics
and to promote construction of fine stringed instruments

Vol. 4, No. 7 (Series II) May 2003

 Feeling the Elephant: It's What We Do Best

The problem of investigating string instruments is a lot kke the famous story about a group of inquisitive but vision-impaired
men who
attempt to describe an elephant for the first time. They approached the giant beast with outstretched hands to find out its nature. The one
feekng the elephant's side said "a wak," the one feekng the elephant's leg said "a tree trunk," the one feekng the elephant's tail said, "a
rope," and so on. Each bekeved he was the only one who spoke the truth.

The story could just as weU describe a group of able CAS members who are set loose to figure out how stringed instruments work.
One
determined density and elastic modulus of tonewood samples, while another measured frequencies of free plates on a shaker table,
another concentrated on effects of surface coatings on damping, and others painstakingly measured arching curves and thickness graduations.
Each bekeved they had found the truth, and they ak experienced a sense of puzzlement when it became clear that others had different
ideas.

If we were to provide a solution to the original story, we could envision a keen observer sitting up in a
tree with a spyglass, who, viewing
ak parts of the elephant, saw the whole beast clearly and accurately. Only this observer had the perspective necessary to appreciate the
detaks offered in each of the men's descriptions.

In this issue, readers are invited to poksh their spyglasses, whke a series of selected authors describe many aspects of historic
instruments.
We selected a series of articles on such things as how classic historic instruments are shaped, the nature of their accessories, and at what
frequencies they buzz when a small glass tube touching their surface is briskly rubbed. These articles represent solid, reproducible
research on some of the world's most esteemed instruments, especiaky Cremonese viokns, violas, and cekos. The value they
represent
for makers and researchers is that they describe a considerably different "animal" than what has been described in the usual viokn making
mythology. Readers are invited to put the whole story together for themselves.

In addition to articles about historic instruments, we acknowledge Joan Maker's years of dedication to CAS, we learn the results of
the
recent readers' survey, Knut Guettler shares insights into bowing gestures and describes ISMA '02 Mexico City, and two technical notes
investigate the subject of tap tones. We hope you enjoy and take something rewarding from this issue.

Good Reading!

Jeffrey S. Loen

The CAS Journal is published twice a year by the Catgut Acoustical Society, Inc. a non-profit organization which aims to
increase and diffuse knowledge of musical acoustics and to promote the construction of fine stringed instruments.

The annual fee for membership in the Catgut For membership and back-issues contact the CAS
Acoustical Society is: Office:

Individual US: $55 55 Park Street

Students US: Montclair, NJ 07042, USA
Institutions US: Phone: (973) 744-0371
Fax: (973) 744-0375
An additional postage and handkng charge of US E-mak: [email protected]
applies to memberships outside the USA Web Site: www.catgutacoustical.org
 C^Ao Journal
May 2003

Vol. 4, No. 7 (Series II)

The Catgut Acoustical Society is known for fostering pioneer research in musical acoustics and the appkcation of these principles to the
making of fine stringed instruments. To fulfill its mission, the Society supports publications, meetings for researchers and makers,
musical compositions, lectures, and concerts.

Authors 2
EDITOR Award
Jeffrey S. Loen Distinguished Service Award given to Dr. Joan E. Miker 4
[email protected]
Letters to Editor
ASSOCIATE EDITORS Sound Power of Mezzo Violin, by Carleen M. Hutchins 4
Gregg Alf Reconstruction of a King's Lyre, by Lloyd Craighill 4
[email protected] Reflections on Making a 21st Century Rebec, by Lloyd Craighill 5
Virginia Benade Comment and Response: Drawing Body Outknes 7
[email protected] Al-Bl Question, by Thomas Cox : 7
va V
Additional References on Modal Analysis, by David Bell 7
[email protected]
Results of CAS J Readers ' SutV^ 8
Bob Schumacher
[email protected] From the Contributing Editors
Look at the String Player's Bowing Gestures, by Knut Guettler 12
CONTRIBUTING EDITORS _ A, Closer
. ,__
Technical Notes
Charles Besnainou
George Bissinger Wood Testing by Tap-Tone by Thomas M. Cox 17
Xavier Boutillon Tap Tones of Old Itakan Top Plates by Jeffrey S. Loen 20
„ .
loseph Curtin
Knut Guettler
Feature: Historic Instruments
oo

Martin Schleske Measurements of Principal Resonance Modes of Historic Instruments, by A. Thomas King..23
Jim Woodhouse Reverse Graduation in Fine Cremonese Violins, by Jeffrey S. Loen 27
Documentation of a 1604 Violin by Hieronymus Amati, Italy, by Jeffrey S. Loen
PRODUCTION MANAGER
and A Thomas Kin& 40
Deana Campion
[email protected] Early Tuning Pegs and Tailpieces, by Eric Meyer 42
Curtate Cycloid Arching in Golden Age Cremonese Violin Family Instruments,
EDITORIAL ADVISORY BOARD .
by
.n/ .
(Juentin rlaytair
j x. jj
4o
t"Q
A. Thomas King
loseph Regh Working Methods of Early Classical Viokn Makers, as Impked by the Results of Recent
Okver E. Rodgers Dendrochronological Studies, by John Topham 59
Fan-Chia Tao __«
„««"__
Research Briefs oo
Meetings and Exhibits 70
Impressions From ISMA '02 — Mexico City, December 9-13, 2002, by Knut Guettler 72

CAS Journal (ISSN 0882-2212) is pubkshed semi-annuaky by the Catgut Acoustical Society, Inc., 55 Park Street, Montclair, New Jersey
07042. Neither the Society nor the Journal's editorial staff is responsible for facts and opinions expressed in articles or other materials
contained in the Journal.
Copyright 2003
Thomas M. Cox is a retired aerospace engineer. He is an amateur viokn maker, vioknist, and bassist.

Knut Guettler (born 1943 in Oslo, Norway) is a former principal double bassist of the Oslo Philharmonic. He has performed as soloist
and lecturer in U.S.A, Japan and a number of European countries. He is author of "A Guide to Advanced Modern Double Bass
Technique" (Yorke Edition, London). Presently, KG. devotes most of his time to research on the bowed string, often through the use of
computer simulations.

A. Thomas King is a vioknmaker in Potomac, Maryland. He is a past president of CAS and past editor of the CAS Journal.
Jeff Loen is a Seattle scientist/luthier who edits the CAS Journal.
Eric Meyer, BA, Univ. Of Notre Dame 1970. Apprenticed 1973-77 with luthier Jeffrey Elliot. Owner/ operator Twelfth Fret Guitar Shop
1978-86, restored, repaired and built plucked stringed instruments. Built five 18th-century French Baroque guitars with guidance of
lutemaker and historian Robert Lundberg. Machinist and set builder for Will Vinton Animation Studios 1987-89. Presently owner/
operator of Finetuning, making pegs and tailpieces for stringed instruments. http://www.vanzandtviokns.com/FineTuning.htm

Quentin Playfair has been working with bowed instruments since the early 70's, initiaky in the U.K., and then in Toronto. For the last
15 years he has been working on his own, deakng mainly with professional musicians from Ontario and northern N.Y. state. A growing
interest in baroque performance and instruments is reinforced by his continued fascination with History, his degree subject.

John Topham is a viokn maker and restorer who studied at the Staatkche Fachschule fur Geigenbau in Mittenwald, Germany. After
working in London for nearly ten years, he began working on his own and has been restoring instruments ever since. About eight years
ago, after gaining a degree at the Open University in the UK, he began investigating the technique of dendrochronology appked to the
viokn and, in association with a cokeague, Dr. Derek McCormick of Queen's University, Belfast, pubkshed two papers on Engksh and
Itakan instruments in the Journal of Archaeological Science.

2 CASJ Vol. A, No. 7 (Series II), May 2003

 "" O. .;.- :_ , ' - .. B '■' ■„ "■ ;.. -. . \ : 0 '"'
!
"■->■ ''''■'■"' S .\ o''*o V " W' "!""!'.

MMMMMMM . ., ..
M M M
M^^^^^^^^^^^^^^^^^^^^M^^^MMMMMMMMMMW________i

Joan E. Miller
CAS Distinguished Service Award

Joan Miker holding silver bowl awarded

for distinguished service to CAS.

I am very pleased to announce that the Catgut Acoustical Society has presented a Distinguished Service Award to Dr. Joan E. Miker. The
presentation was made during the Awards Dinner of the November 2002 VSA (Viokn Society of America) Convention, at which time the
VSA also presented her with a Distinguished Service Award for contributions to their organization. Joan has been very active in both CAS
and VSA, so the awards are wek deserved.

Joan is a past-President and Board member of CAS. In addition to serving in these roles, Joan took a very active part in supporting the day-
to-day operations. It was clear that Joan accepted substantial, time-consuming jobsthat others hesitated take on.
to the first
Joan brought
computers into the CAS Office, trained the staff, and continues to provide support. She set up the database for membership records (along
with design of forms and reports), wek
as as the financial management and reporting software. Further, she took charge of the materials
for the CAS Library at Stanford, gathering materials from Montclair, then cataloging the entire collection. Any of these tasks, alone,
deserve our thanks.

Many of you may know that Joan, a retired Bek Labs employee, runs JEM Software, which produces the SAND software for the spectral
analysis of viokns. Her work for CAS, however, is often done in the background, and she makes a great effort to help the Society run
smoothly. We are pleased for the opportunity to thank and congratulate Joan for her contributions.

Jukus J. VandeKopple
President, CAS

CASJ Vol. A, N0.7 (Series II), May 2003 3

 Sound Power of Mezzo Violin

Carleen M Hutchins

In their excekent study, Wang and Burroughs [I], state, "the mezzo Burroughs in the low mezzo range transmitted to the sound heard
violin does not appear to radiate more acoustic power than the from the wholeinstrument?" I hope that some aspiring researcher
standard instruments." However, listening tests of mezzo violins wik take on this challenge.
as well as 40 years of player experience indicate that the 30 or so
mezzos now being played are perceived to be louder than average References
normal good violins. [1] Wang, L.M., and Burroughs, C.8., 2002, Comparison of
radiated power from structurally different violins: Catgut
Perceived sound is measured in phons on the Fletcher-Munson Acoust. Soc. Journal vol. 4, no. 6 (Series II), p. 65-72.
curve (see [2] p. 1-7), which shows the way the average human ear [2] Berendt, R.D., Corliss, E.L.R., and Ojalvo, M.S., 1976,
weights sounds of equal levels but differentfrequencies, being more "Quieting: A practical guide to noise control": National
sensitive in the mid-range and faking off at either end. This means Bureau of Standards Handbook no. 119, US Environmental
that the human ear perceives an increase of 2-3 dB in the low G- Protection Agency, Office of Noise Abatement and Control,
string range as twice the loudness in phons. The big question is Washington, DC 20460, 115 p.
"How is the extra 10 dB sound power as measured by Wang and

Reconstruction of a King's Lyre

Lloyd Craighill
11 Cranberry Lane
Amherst, Massachusetts 01002 USA

Maryly and I were attending lectures on Beowulf with readings of

the original by Robert Creed, who has reconstructed the poetics of
the work. John Bokard accompanied him on an instrument based
on British Museum drawings of fragmentary remains of a lyre buried
with an Anglo-Saxon king who died c. 625 CE. I thought the
reconstruction had inadequate acoustics, and built John a version
that is more playable and thus historically perhaps more accurate.
For example, I added sound holes and gave more volume to the
soundbox, which is simply a hollowed out maple plank with a
sounding board. The corner escutcheon plates were reconstructed
by jeweler Jonathan Glatt, and the rest is my own concoction. My
cokeagues plan to record ak the original Beowulf (on the order of
three CD's) using Robert's prosodic researches, and John's musical
inventions for accompaniment.
Hypothetical reconstruction of a King's lyre (ca 625),
based on drawings at the British Museum.

4 CASJ Vol. A, No. 7 (Series II), May 2003

 Letfers to fhe Editor

Reflections on Making a 21st Century Rebec

Lloyd Craighill
11 Cranberry Lane
Amherst, Massachusetts 01002 USA

I have some thoughts about the extremely varied group of edges at the widest part of the plate, because there would be no
instruments known to musicologists as the "rebec" that appeared liners and sturdiness was a major objective. The walls had been
in Europe, perhaps as early as the tenth century. Webster's Cokegiate ground to about 7 mm except at the bottom where 10 mm was used
kmits its definition to "an ancient bowed usuaky 3-stringed musical because of end grain concerns. A very kght bass bar was glued in,
instrument with a pear-shaped body and slender neck." located to run under the bass foot of the bridge and converging a
kttle towards the centerkne, viokn fashion. Later a sound post was
My experiences are of a practical nature. I wanted to make a sturdy set. The late rebecs might have had one or both of these features
bowed instrument that could be played cello-style but that would and the medieval fiddle was clearly moving in this direction, but
be small enough to take car camping. The Arabic rabab and its history (in my project in any case) was clearly giving way to
multiform European cousins provided a model, but clearly my practicality. I concocted a scimitar and bird head soundhole
instrument was to be a 21st Century one and not a repkcation of —
pattern much more interesting than the back-to-back dumb-bunny
something that might have existed in the 15th Century. Thus, I C-holes of latter day gamba makers.
don't have the materials for a proper research paper, and instead
provide an account of a small extension of the historical process —
The bridge in this case, a smak abandoned ceko bridge with new
that began with the invention of the horsehair fiddle bow, probably —
feet glued on, and othermodifications wouldhave to be high enough
somewhere in central Asia and perhaps in the first mikenium B.C. for the bow to clear the boat-shaped body. My underlying modulus
was based on tenor viokn making. A roughly 600 mm string length
REFLECTIONS ON MAKING A 21ST CENTURY REBEC design could use existing cello strings and cello fingering. I
Visions of the medievalrebec had danced through my head for personally would not be able to manage the smaker placed-on-the-
years, and the time to bukd a reconstruction came in the summer of forearm instrument of northern Europe of, say, the 16th Century.
.
2001 I got a chance to be a paying guest on a 38-foot sloop about to With bridge kept as low as possible while considering bowing
cruise Maine's Casco Bay, and the skipper was, among other clearances, the fingerboard (3/4 ceko size) would still need a very
accompkshments, a professional musical entertainer. He suggested fat wedge, unless it were propped up with a couple of smak blocks —
bringing an instrument that could be stowed on board. the birth of the A-frame, or "peek-a-boo" neck assembly.

The boat was leaving in three weeks, so my project had to be done The best angle to and over the bridge could be replicated from
fast. What should the design be? The latest Grove's dictionary [1] cekos or tenors, and clearly a saddle at the bottom of the resonator
provides endless possibilities: a decorated body of Javanese —
would provide intolerable compression of the bridge so how about
hardwoodwith a stretched sound table ofbuffalo-bladderparchment a hitching post? Better, support it with a smak block at the bottom
... I had a log of maple that could produce a three-foot wooden of the resonator to counter the bending moment from the tak gut.
spoon, but that wouldbe a bitch to carve into a big, spoon-shaped
rebec. At the lumberyard I found a ten-foot plank of maple miked The bottom end was shaping up, but what about the head? My neck
smooth to 20 mm. I cut it into three pieces, and bandsawed them lamination plan had included a peg-board at the top, but pegs stuck
into two tennisrackets and a canoe paddle, the paddle blade thinned into the top of a guitar-kke head did not hold modern string tension.
to about 5 mm except at the edges. Laminated, they became the big Rebecs in later times had gone to the sickle-shaped pegbox and I
spoon I was after, easily drum-sanded into clean surfaces. The had better go to the pegbox myself. (By now, the boat was about due
upper half was a 60 mm wak that was turned into a fiddle neck by to leave port, and I wasn't done because of my backing, fiking, and
two long slightly converging saw cuts, producing about 28 mm head scratching). I stuffed a tenor into a sea bag and the skipper
thickness at the bottom and about 20 mm near the head, with the verykindly found a way to lash it under the foredeck next to his
resulting handlerounded to resemble a very fat viola neck, the slant guitar.
to the head being about 5 degrees from horizontal.
On return and with a kttle more time for my project I carved a lady-
The table was glued up from red spruce wedges, sawed out and skpper pegbox with a figurehead scrok based on a marble portrait
skghtly arched with a graduation of 3.5 mm around the bridge, a bit bust of a Roman matron in the Smith Cokege Museum collection.
thinner north and south of that, but thickeningto about 7 mm at the A bit of class for my poor-man's campfire-oriented three-stringed

CASJ Vol. A, N0.7 (Series II), May 2003 5

Letters to the Editor

ceko. The head had an extension below the pegbox that was scarf- viola, I would say. The timbreis out of our world—East of the viola
joined to the upper end of the neck. and West of the Chinese erhu—maybe somewhere in the Seljuk or
Ottoman Empire, perhaps on the Black Sea shores where woods
I finished the instrument with egg white sealer and oil varnish dried were plentiful, and skin heads not needed. It sings wek with a kght
under a plant-growing fluorescent light fixture, and strung it up ceko bow, but a palm-up gamba bow played with some wrist-snap
with leftover Super Sensitive tenor strings. A, D, and G worked comes closer to the soul of the instrument—cross your ankles, set it
well, and the little resonator, no bigger than a small dishpan, between your knees and set the (optional) neck clamp against the
supported the G as fully as it did the D and A (an E, D, A alternative middle of your chest for firm fingering. Strung with gut I am sure
did not go as wek, and experiments with four strings were quickly this instrument could be played in pubkc by the early music people
abandoned). without visible embarrassment. However, is it a rebec?

This product of the Tight-Bond age of lamination needs no mold, Look at the incredible multiplicity and diversity of often home-
corner and end blocks, bent ribs and bent liners, carved back or made rababs, rabobs, ribabs, rababas, etc. that extended across the
neck mortise. An efficient fiddlemaker with plans in hand could Muslim world, and their transformation in half-literate Europe in
knock it off in a week, maybe less. However, how well does it the tenth or eleventh century, resulting in the rebec, rebet, ribeca,
work? The sides are too thick to radiate useful sound but the bottom rabe, rebesbe, ribiblc.and you can see that this workshop
does work as a tympanum or wooden drumhead, and showed a experiment is not without cousins or precedent. It's just one more
clean circle mode on the shaker table. The top showed a clean adaptation.
circle mode too, but not much else and the nodal knes ran off the
confines of the plates, top and bottom. REFERENCE
[1] Remant, M., 2001, "Rebec", The New Grove Dictionary of
The assembled instrument has surprising acoustical efficiency with Music and Musicians, London.
—
the wound steel-core strings I am using as much noise as a cheap

Figurehead scrok inspired by a marble sculpture of a Roman matron in

the Smith College Museum collection.
Craighik playing the reconstructed rebec. The
strut attached to the neck makes playing easier
by fixing the position from the instrument to
the chest.

6 CASJ Vol. A, No. 7 (Series II), May 2003

 Letters to the Editor

Comment and Response: Drawing Body Outlines

Comment by David Golber
About the article by Mark French [I], readers interested in using modern computerized methods for instrument design might want to

read the "CAD Notebook" series of articles by Dana Bourgeois in "American Lutherie"(issues 65 to 71), and my own article [2].

References
[1] French, M., 2002, Instrument body descriptions using polar coordinates: Catgut Acoust. Soc. Journal, vol. 4, no. 5, p. 8-10.
[2] Golber, D., 2002, The right CAD curve: Beziers, not spknes, are truly smooth: American Lutherie, no. 72, Winter, 2002, p. 50-53.

Kesponse by Mark French

While spknes wik almost always give a precise result in an appkcation such as this one, they are not part of common PC software suites.
Luthiery is certainly practised more widely as a hobby than a profession, so it would seem reasonable to seek the most accessible
approach possible. The purpose of this article was to develop a simple, closed form solution. The result is compact and can be implemented
easky using a universaky avakable spreadsheet program. It is true that the slope of the proposed curve is not exactly zero at the bottom
of the instrument. Rather, it is approximately 2.25°. In practice, this smak ridge is sanded off during the construction process. There is
nothing inherent in the method that forbids a zero slope; it is just a matter of developing a different curve fit. The central idea of the article
is not the quakty of the curve fit, but the idea of transforming from cartesian to polar coordinates. Finaky, I did indeed add an extra point
to try to enforce a zero slope at the bottom of the instrument since the software I was using didn't akow me specify the slope directly.

Al-B1 Question by Thomas Cox

CAS has pubkshed evidence [1,2] for a definite correlation between Al-Bl frequency delta and the viokn's acoustic power. Measurements
on my own viokns fak to agree. I measure Al per CAS May '94 page 26, using surgical rubber tubing between the transducers. I measure
Bl per CAS Nov. '95 page 16, using a phono pickup set to sense vertical motion. I am unable to differentiate between Bl- and 814-. From
CAS papers, they are both strong radiators and pump air in the corpus, so either mode should interact with Al. Therefore, I consider the
delta to be the difference from Al to whichever is the stronger mode between450 and 600 Hz. However, this gives inconsistent results.
This problem frustrates me. I would appreciate help.

[1] Schekeng, J.C., 1963, The viokn as a circuit: Journal Acoust. Soc America, vol. 35, no. 3, p. 326-338.
[2] Hutchins, CM. and Rodgers, 1992, Methods of changing the frequency spacing (delta) between the Al and Bl modes of the viokn:
Catgut Acoust. Soc. Journal, vol. 2, no. 1 (Series II), p. 13-19.

Additional References on Modal Analysis by David R.Beu

Many readers of this journal are wek-versed in the mathematical physics of sound. Nonetheless, it was remiss of me to omit from the
Modal Analysis article in the last issue (CAS Journal, Vol. 4, No. 6, p. 7-13) some suggestions for either those who have forgotten what they
may once have known or have only hazy notions about the topic.

The best single short reference I know of is a classic work:

Mathematicsfor the Nonmathematician: Morris Kkne, 1967/1985, ISBN 0-486-24823-2, Dover edition paperback in print.
Chapters 18 and 19 introduce the relations between trigonometric functions and oscikatory motion and the analysis of musical sounds by
these means.

A second more comprehensive reference is:

Fundamentals of Musical Acoustics: Arthur H. Benade, 1976, 1990, ISBN 0-486-26484-X; Dover edition paperback in print.
This book covers not only the physics of sound but also deals with psychoacoustic matters and the functioning of musical instruments
themselves. It is a rich and authoritative work.

CAS J Vol. 4, N0.7 (Series II), May 2003 7

 Results of CASJ Readers' Survey

Jeffrey S. Loen
[email protected]

SUMMARY MAIN INTERESTS

A survey returned by 192 members suggests that a major change My main
interests are (multiple responses allowed):
has taken place in the interests and orientations of CAS members.
Percent of
Years ago, CAS included mainly scientists, engineers, and players Responses (%) Respondees
who were interested in applying basic science to unravel the Acoustical Research
121 (31.5) 64.4
mysteries of string instruments. However, results of a mak-in survey Instrument Making 138 (35.9) 73.4
confirm that viokn making is now the leading interest, and a number Repair/restoration 69 (18.0) 36.7
of the craft-related readers state that the Journal is "too technical" Playing
52 (13.5) 27.7
and question the use ofmath and equations. Main instruments of Other 4 (1.0) 2.1
interest are violin, viola, and cello, which constitute 63% of
responses, although bass, new viokn family, guitar, and others are 'otals 384 (99.9) n= 188
also favored. The peer review process, which tries to "weed out"
incorrect information, is valued by 89% of respondents. The most Responses to the question about main interests (Fig. 1) suggest that
popular subjects, chosen from a kst of 47 topics, are ak about methods instrument making is the leading interest,
followed by research,
of improving tone. repair/restoration, playing, and other interests (such as simply
kstening). Many readers checked more than one interest. Research,
which once was the predominant interest of CAS members, has
INTRODUCTION declined proportionally as research funding has dwindled and
In November 2002, a two-page survey form was mailed to ak CAS people of other orientations have joined CAS.
members with their annual renewal notice. The purpose of the
survey is to keep the CAS Trustees and the editorial staff informed DEGREE OF TECHNICAL DETAIL
about our members' interests and opinions, so we can make Recent issues of the CAS
Journal have been:
informed decisions about charting the future course of CAS, and
choose interesting and pertinent articles for the Journal. We Responses Percent
received 192 responses by press time. Too technical y 14.2
About right 142 84.0
In the early years of CAS, the membership was mainly physicists, Not technical enough 3 1.8
acoustical scientists, engineers, and technical-oriented players.
Today, vioknmakers are the largest group, and many members are Totals 169 100.0%
involved in repairs, playing, or can simply be described as "string
enthusiasts." Yet we retain a core of ardent researchers with Most (84.0%) responding readers feel the
Journal has about the
impeccable credentials and boundless creative zeal. Rather than right degree of technical detak (Fig. 2). It is interesting that the 24
aiming ak articles at one group or the other, our goal is to choose a readers who think the
Journal is "too technical" are all instrument
blend of articles that reflects the interests and orientations of all makers, whereas the three who think it is "not technical enough"
members. This is a difficult task, because we are, in the end, kmited are all acoustical researchers. You obviously can't please all the
by what authors submit to us, and what passes peer review. people ak the time! One perceptive reader stated, "The ideal would
Moreover, writers are more willing to share detailed results of be to have a good mixture that covers a broad spectrum." That is
acoustical experiments than they are to describe workshop practices indeed our goal. We regret that some articles are beyond the grasp
that lead to commercial success in viokn shops! Results of the of non-technical readers, whke other articles might seem simpkstic
survey are as follows: to advanced readers. However, we hope that everyone will find
something rewarding in each issue.

8 CASJ Vol. 4, No. 7 (Series II), May 2003

 Results of CASJ Readers' Survey

INSTRUMENTS OF INTEREST
Figure 1. Pie diagram showing main interests of 192responders to The instruments that interest me most are (multiple responses
CAS survey. Many readers chose more than one answer, giving a allowed):
total of 384 repkes. Percent of
Main Interests Responses (%) Respondees
Violin 155 (34.1) 82.9
of CASJ Readers other
Viola 96 (21.1) 51.3
Playing _ n=4, 1%
} Cello 84 (18.5) 44.9
~=yy, yy \.-
Acoustical Bass 35 ( 7.7) 18.7
Research New violin family 32 (7.0) 17.1
n= 12
Repair/ Guitar 39 (8.6) 20.9
Restoration
n=69. 18% Other 13 (2.9) 7.0

otals 454 (99.9) n = 187

Instrument CAS members are interested in a variety of instruments (Fig. 3), the
Making
most popular of which are viokn, viola, and cello. We also have
many enthusiastic bass players and makers. The New Viokn Famky
has many dedicated adherents, as do fretted instruments including
Figure 2. Readers' assessment of degree of technical detak of CAS guitar and mandokn. Some members state that they are drawn to
Journal. Based on 169 responses. viols and other early string instruments, while others prefer
keyboard and wind instruments.
Technical Detail
"Not Technical
Enough" of CAS Journal IMPORTANCE OF PEER REVIEW
n=3,
The peer review states of articles is:

Responses Percent
Very important to me 70 43.2
Somewhat important to me 74 45.7
Not important to me 18 11.1

Totals 162 100.0%

"About Right"
Responses to the question about peer review (Fig. 4) suggest that
readers want to retain the process of having papers evaluated by
professional peers before pubkcation, rather than simply accepting
and printing submittals with few questions and conditions. Some
Figure 3. Preferences of musical instruments, based on 454 readers express confusion about the purpose of "peer review", which
responses from 192 members. "NVF" is New Viokn Family. one reader accurately summarized as "to weed out crank stuff which
is plain wrong, and causes problems later."
Other
Instruments
Guitar n=.3,3% of Interest One instrument maker, who feels the CAS Journal is "too technical"
n said that peer review is very important, "but only if peers are viokn
makers, not academics or engineers." This brings up a vakd point,
NVF Violin
because few "viokn making" articles seem to meet the approval of
n=32, 7% 33%
highly technical reviewers who are non-makers. Authors of violin
Bass making articles often form conclusions based on personal judgment
n=3s, 8% and years of experience, and they often support their assertions
with little or no numerical evidence and few kterature references.
Such papers are kkely to be rejected when judged on a strict technical
Cello basis. For this reason, we try to send papers to reviewers of a
n=84, 19% Viola similar professional orientation. Regardless of the author's
n=96, 21%
professional persuasion, it is important to us to print articles in

CASJ Vol. 4, N0.7 (Series II), May 2003 9

Results of CASJ Readers' Survey

Table 1. Detailed survey results.

Responses Ratin Responses Rating
5) Varnish, & Adhesives (continued)
1) Stringed Instrument History and Documentation
c) Acoustical effects 174 4.5
a) Development & evolution of stringed instr. 167 3.4
f) Adhesives, behavior over time 161 3.8
b) label authentication 164 2.7
Averages 165.7 4.0
c) Age determination, e.g. dendrochronology 166 3.1
d) Acoustical measurements 183 4.4
6) Strings
c) CT scans, archings, & thickness graduations 177 .4.2 a) Design, composition, & physical properties 166 3.7
Averages 171.4 3.6
b) Dynamic properties & sound production 166 4.0
c) Effects of string tension 160 3.9
2) Wood & Other Materials
d) Selection to fit instrument properties 164 4.2
a) Selecting, harvesting, curing tonewood 170 3.9
c) Player preferences 159 3.5
b) Species variation 169 3.7 Averages 163 3.8
c) Practical measurement and test techniques 177 4.4
d) Pre-treatment of wood 168 3.9 7) Bows
c) Synthetic materials (e.g. carbon-fiber) 163 2.8
a) History & development 161 3.0
Averages 169.4 3.7
b) Physics of the bow/string interaction 168 3.7
c) Materials 162 3.4
3) Shop Practices & Manufacturing
d) Evaluating bow performance 167 3.7
a) Assembly techniques, hand vs. machine 169 3.5
c) Technical function of a bow 166 3.6
b) Shop layout and tools 163 3.1
Averages 164.8 3.5
c) Influence on acoustical properties 173 4.2
d) Antiquing 163 2.7
8) Acoustical Research
Averages 167.0 3.4
a) Mathematical models 163 3.5
b) Computer simulations (e.g. finite element meth.) 167 3.7
4) Repair & Adjustment c) Experimental tests and interp. of results 172 4.3
a) Effects of repairs on acoustical response 171 4.1
d) Effect of changes (bridge, post, etc.) on sound 179 4.6
b) Effects of set-up on acoustical response 176 4.6
c) Measuring equipment 172 4.2
c) Methods of improving tone 176 4.7
f) Useful software and instrumentation 172 4.0
d) Diagnosing & solving tonal problems 179 4.7
Averages 171 4.0
c) Artificial "playing-in" 163 3.7
f) Instrument fatigue 165 3.8
9) Performance
Averages 171.7 4.3
a) Psychoacoustics, tone quakty perception 175 4.2
b) Effects of design changes on performance 174 4.2
5) Varnish, & Adhesives
c) Effects of changing component materials 166 4.0
a) Chemical & physical properties 161 3.9
d) Physical functions as instrument is played 164 4.1
b) Varnish recipes 165 3.7
c) Methods of recording & analyzing sound 168 3.9
c) Application methods 166 3.8
Averages 169.4 4.1
d) Wood surface preparation 167 4.1

Figure 4. Readers' opinions about peer review. Based on 162 which assertions are logically supported. This ekminates much of
responses. the opinions, anecdotes, and blind adherence to tradition that
masquerades as "information" in the viokn world.
"Not Importance of
Important" Peer Review
11%
SUBJECT SURVEY
We asked for ratings of interest of nine subject categories, with 47
'Very subtopics. The rating system ranges from one to five, with five
Important" being the highest level of interest and one being "no interest."
43% Readers responded differently to the survey, with some taking the
time to fill in all blanks, and others checking only their favorite
"Somewhat
Important" subjects. One thoughtful reader, rather than filkng in blanks, simply
46% stated, "I have no strong preferences—all these topics are appropriate
and interesting if well done."

10 CASJ Vol. A, No. 7 (Series II), May 2003

 Results of CASJ Readers' Survey

The fokowing subjects have the highest ratings (in decreasing order Please offer specific suggestions for improving the CAS
of interest): Repair and adjustment (4.3), performance (4.1), Journal:
acoustical research (4.0), varnish, ground, and adhesives (4.0), strings Rather than offering specifics, many readers took the opportunity
(3.8), wood and other materials (3.7), string instrument history and to make general statements, some critical, some compkmentary. A
documentation (3.6), bows (3.5), and shop practices & manufacturing common theme is the degree of technical detak. The fokowing is a
(3.4). The subtopics with the highest ratings are about the effects of sample of responses:
various types of changes on instrument sound. For example, the
maximumvalue of 4.7 (Table 1) was given to "methods of improving
tone" and "diagnosing & solving tonal problems". A skghtly lower "Are we writing only for scientists?"
value of 4.6 was given to "effects of changes on sound" and "effects
of set-up on acoustical response", and 4.5 was given to "acoustical
effects of varnish". "As I am not a scientist, you loose me when it is too
mathematical and scientific. Perhaps you could put an
Please list additional topics you would like to see presented in emphasis on a more practical approach ofinstrument making,
the CAS Journal: try to popularize the more scientific articles."
We thought we had developed an exhaustive kst, and some readers
agreed, saying "If you can cover ak the above, you are doing very "Keep scientific approach central! Other journals kke JVSA,
well!" The following are suggestions for additional topics: Strad, Strings, and American Lutherie are covering general
interest and workbench procedures."
"Effects of different types and quakties of bow hair" "CAS Journal is already very good!"
"Debunking con men who claim to have discovered the secret "I'd prefer to have technical articles be less abstract, and
of Stradivarius and who get pubkcity from the naive media more closely related to instrument properties (playability,
(who else is quakfied to do this?)" sound character, etc.). I'dkke more why and how, and where
this leads in the article. Sometimes it seems like people
"Superiority of old violins, and commercial interest" measure things just for the heck of it."
"Baroque-Modern transitions and acoustical differences" "The last few issues of the Journal have been more than
"Acoustical mode measurements of classical concert viokns/ —
fabulous Excellent Work!"
violas/cellos" "Are we writing only for scientists? Violin maker/players
"Objective methods of tone testing in competitions" do not generaky have the background in physics and higher
"More down-to-earth information for instrument builders math to benefit from articles fiked with math formulas!"
who are not privileged to possess high-tech testing "I enjoy it the way it is."
equipment." "Reject highly technical articles—refer them to academic
"Any studies that help to relate what the researcher can journals"
measure with what the player and kstener perceive." "Thank you for a most interesting Journal!"
"Measurement of historic violins, with thickness of plates."
"More emphasis on the grandest of ak viokns — the bass." CONCLUSION
The readers' survey illustrates that our membership has changed
"Advice for the players based on physical function."
over the years. Viokn makers now dominate a group that began in
"Figure out why those old viokns are so kghtweight!" 1964 as a handful of skilled researchers and inquisitive players.
"Playing technique as it appkes to tone" Some of our current group express discomfort with technical tools
and methods that are the essence of the scientific approach of
"Would someone design a useful stiffness caliper? A experiment, data analysis, and explanation. Survey results tek us
"stiffness graduation map" of plates of great instruments that the burning desire of many current members is to obtain non-
would be very helpful." technical suggestions for improving tone. In contrast, the motivation
"Explanations for the non-scientist/non-mathematician that of original CAS members was to understand the workings of stringed
supplement the scientific presentation of studies and instruments. Wanting to improve tone is understandable. However,
experiments" it is doubtful that we can learn how to improve tone unless we first
seek, and attain, a basic technical understanding of how instruments
work. We are not there yet!

CASJ Vol. A, N0.7 (Series II), May 2003 11

 A Closer Look at the String Player's Bowing Gestures

Knut Guettler
Norwegian Academy of Music
P.B. 519Q Majorstuen, N-0302 Oslo, Norway
[email protected]

ABSTRACT Does the timbre change when only the bowing position is
This presentation seeks to give an overview of present knowledge changed?
and theories concerning bowed-string transients and tone coloring. Regard Figure 2: Here the "rounded corner" causes the string release
Bowed instruments are quite unique in the way tone production to spread out over a smak transition interval before the fuk negative
results from a considerable number of physical parameters, many velocity is reached at slip. Accordingly, a comparable transition
of which are manipulated simultaneously by a single hand: bow takes place at capture. (In practice the transition at capture is often
—
"pressure" and speed; the bow's position on the string as wek as shorter than the one at release.) It can be shown that as long as these
the string's position on the bow hair; the bow-hair angle and impact transitions are independent of the bowing position, the force signal's
trajectory, to name the most important ones only. However, such spectral slope (envelope) wkl remain entirely unaffected by J3 (see
complexity makes a two-edged sword since there are often narrow Fig. 3). However, with the bow speed held constant, moving the
margins between perfection and blunder for the musician trying bow from one position (1) to another(2) wik change the sound level
to get the most out of his/her instrument. by an approximate factor /3Jfi2 ~ Av2/ Azo. Adjusting for this, the
spectral alterations will generaky be related to the low-leveled "node
BOW FORCE, POSITION, SPEED, AND THE SCHELLENG frequencies" of the "old" and "new" bowing positions (i.e., nf/fi
DIAGRAM and nfJJ32, where n-1, 2, 3... andf is the fundamental frequency).
Both Raman and Schekeng had the musician's perspective in mind These will, however, not contribute to an overall raise of higher
when performing some of their most intriguing analyses. While partials as the bow is moved towards the bridge, although fewer
Raman [1] partly collected empirical data by means of his and fewer ("node") partials wik be missing as the distance to bridge
mechanical bowing machine—where the bowing gesture in terms gets smaller. In Figure 3, spectral differences between simulations
of speed, normal force, and string position were ak controkable— with six different /? are compared:
Schelleng [2] utilized Raman's theoretical analysis on reflected
waves and used the same three parameters for mapping the
requirements for maintaining the Helmholtz motion. Figure 1: Diagram from Schelleng's JASA paper. Given a fixed
bow speed, the triangle sets the borders for maintaining Helmholtz
But Schelleng added one more important aspect: timbre. At two motion in a bowed string.
separate places within the Helmholtz-mode area in his diagram,
the expressions "brilkant" and "sul tasto" are marked as indications
of timbre. ("Sul tasto" kterarky means "by the fingerboard", but is RELATIVE POSITION OF BOW, /S
here, presumingly, a reference to soft tone color.) .01

t_J
It is not immediately apparent from the figure whether these timbre v
c_
differences are caused by change of bowing position (the two 0.100
v.
expressions are marked at different bow positions within the ID
Helmholtz area), or by change of bowing force (expressions are >
b.OlO
positioned at different values with respect to the ordinate). In the <-J
IxJ
discussion that followed, Schelleng showed how the waveform of cc
the string velocity under the bow is changing with bowforce. Later
0.7 1.4 2.8 4.2 7 14
Cremer [3] analyzed this more in depth, and estabkshed the theory BRIDGE TO BOW, CELLO A CM
of "the rounded corner". However, the remaining question is:

12 CASJ Vol. A, No. 7 (Series II), May 2003

 Knuf Guettler — A Closer Look at the String Player's Bowing Gestures

How much does the timbrechange as function ofbow force? axis roughly through the thumb and the middle finger. This means
Pickering [4] has shown how each violin strings has an "area of that gravity contributes a considerable amount to the bow force in
maximum bow-force sensitivity, and that this varies with the string bow positions close to the frog, and much less close to the tip. For
impedance and the core material. For a certain kght-gauge aluminum a normal violin bow, the change could be estimated to be
A-string with Perlon core, changing the bow force from 200 to 500 approximately 3 N, provided that the path of the bow is
mN resulted in a gain of more than 15 dB for the harmonics 5 approximately horizontal as when playing on the G-string. The
through 17 (bowing speed and position unaltered). Bowing with a major part of this change occurs within the last third of the bow
normal force so high that pitch flattening occurs typicaky produces toward the frog. The compensation is achieved by balancing the
a shrik sound with no real brilkance, and emphasis on the "near- bow with the index and kttle fingers on top of the bow stick on both
node" frequencies. This is more so for some string types than others. sides of the pivot point."
Nevertheless "excess bowforce" (with left-hand pitch compensation
in combination with a wide vibrato) is often used in expressive CHANGE OF DYNAMICS
passages, e.g., on the viokn G-string. Askenfelt further presents measurements of a series of strokes (their
transients excluded) where the most accurate player kept variations
Does the timbre change if only speed is changed? in bow force less than 0.2 N (around 50%) when playing at the
Yes, it does! And the reason is that changing the bow speed (without dynamic levelpiano, whke less than 0.7 N (around 33%) inforte. It
changing the bow force) alters the transition characteristics at is interesting to notice that for one of the two players, the piano
release/capture. Higher bow speeds give longer skp intervals. When strokes were performed with greater speed in piano (29 cm/s at a
moving the bow towards the bridge, the speed is usuaky reduced, bow-to-bridge distance 42 mm, and bow force 0.5 N) than in forte
and that is what causes the noticeable greater brkkance. Increasing (21 cm/s, 18 mm, 2.3 N), whilefor the otherplayer the corresponding
the bow "pressure" wik of course add to the effect. One should also figures were 21 cm/s (42 mm, 0.4 N) and 24 cm/s (20 mm, 1.7 N),
notice: with a fixed bowing position, when it is musicaky desirable respectively. The general tendency thus seems to be altering position
to maintain tone color best possibly while playing a diminuendo, andforce rather than speed in order to create the major part of the
both force and speed should be reduced simultaneously, because dynamic difference. The dynamic differences produced by the two
reducing the bow force alone would imply softening the tone color. players during these tests were about 9 dB for each of them. These

Double stops
When playing large-interval double stops, the bow force should
(normally) be balanced so that the string of lowest sounding pitch Figure 3: Spectral changes as the bow is movedfrom ft — 1/12 to ft
feels the greatest "pressure"—regardless of which string "carries = 1/7 while keeping other bowing parameters unaltered (the first
the melody". This prevents the pitch of the highest sounding string harmonics normalized to 0 dB). Apart from local deviations, the
to be flattened due to its significantly higher ft. (To count on left- high-end spectral decay fokows the same slope in ak the simulations.
hand compensation is not the greatest idea here!)
"Quasi 'plastic' friction" is a model where changes of value in the
Keeping the bow force even hyperbokc velocity-dependent friction curve are slowed down by a
time constant. The spectral tendency illustrated in this figure is,
In order to achieve an even sound throughout a stroke, the player
however, seen with a variety of friction models.
must actively be changing the fingers' action on the bow stick.
Askenfelt [5] writes: "The bow acts as a lever that pivots around an

Figure 2. Example of string velocity underthe bowfor two different

ft, but with the same bow speed and identical transition functions
from stick to skp and vice versa. (T = the fundamental period; Ay
= K^J B, whke B — ratio between the bow-to-bridge distance and
BOW ' ' °
the total string length.)

5 10 15 20 25 30 35 40 45 50
Time Time Harmonic number

CASJ Vol. 4, N0.7 (Series II), May 2003 13

Knut Guettler — A Closer Look at the String Player's Bowing Gestures

Figure 4: Comparison of (simulated) slip patterns with three Figure 5: Spectral changes resulting from the different bow speeds
different bow speeds. The string velocity is shown relative to the utikzed in the simulations of Figure 4: The ampktudes of the first
bow velocity in each case. At high bow speeds, transitions from harmonics—as well as all harmonics of the reference—are
stick to fuk flyback velocity and back last longer than at lower bow normakzed to 0 dB. Reduced bow speed results in brighter sound.
speeds.

4
2
"I* , 0
> —1-2
v. ,

~ *1
■E -4

"E -8

a: 1U

12
-0,2 0 5 10 15 20 25 30
Harmonic number

last figures point at the role of timbre as means for expressing vioknists to prefer a tilted hair ribbon may, however, be found in
loudness. Remember that raising the bow velocity alone would the fact that tilting facilitates more gentle stroke onsets resulting
decrease the relative content of high partials, which to some extend from a more gradual string contact. Playing with "all the hairs on
would counteract the perception of increased loudness. We shall the string" could also give rise to noise caused by partial skpping,
neither forget that there is a practical side to the choice of bowing since the string's velocity relative to the bow will differ across the
position too, as forte notes can be played longer when bowed close hair ribbon [7]. Close to the bridge, this might even cause spiky
to the bridge, but in the above tests that was not an issue. skpping sounds [B].
Crescendo-diminuendo A pronounced feature of hair-ribbon tilting is the change of
The theoretical maximum range of wave energy in the string due to frequency for the bow's natural bounce rate, as for instance
differences in bow force aloneis 7T/V6 ~ 2.16 dB, whichreflects employed in ricochet. In a violin, a normal tilting angle of ca 30°
the ratio between a true sawtooth- and a pure (1. harmonic) sine causes the natural bounce rate to drop between 1 and 2 Hz [9]. The
wave. This means that by halving ft and doubkng the bow force, the point of impact on the string plays a role too, as playing closer to
energy of the signal could be increased up to a maximumof some 8 the bridge raises the rate somewhat. For rapid spiccato, playing with
dB, while the perceived difference in loudness would probably be the hair flat should hence give the best response, while for playing
significantly greater due to the change in timbre. A gradual halving down-stroke ricochet, progressively changing the hair angle from
of the ft, by the way, is best performed as illustrated in Figure 6 flat to tilted, is sometimes recommendable, as this to some extent
(particularly for instruments with long strings), since this technique counteracts the increased natural bounce rate closer to the tip (see
prevents introducing longitudinal friction forces during the move. below), making it easier to terminate bouncing at the end of the
stroke.
TILTING THE BOW-HAIR RIBBON
Vioknists often claim that tilting the bow provides greater brilliance. STARTING THE HELMHOLTZ MOTION
Pitteroff [7] found through measurements of string waveforms some In a recent paper [10], the wave buildup of the Helmholtz motion in
indications pointing in this direction. Pitteroff used a bowing bowed strings is described. It is shown that in order to start this
machine to ensure identical strokes, while spktting the hair bundle pattern as quickly as possible, the bow must be accelerated within
so that larger or smaker widths of hair had contact with the string. certain limits during the transient, unless substantial loss is present
In these experiments he found that the slipping intervals became by the nut. That is, under normal conditions the stroke cannot start
progressively shorter as the hair-ribbonwidth was decreased, albeit abruptly, "switched on". The tolerable acceleration values are
the effect was small (Askenfelt, Schoonderwaldt, and the present proportional to the fundamental frequency and inversely
author recently repeated this experiment with a thinner string where proportional string's wave resistance (or, if you prefer: to the string's
these spectral effects were quite noticeable). The main reason for vibrating mass). Acceleration must also increase with increasing

14 CASJ Vol. 4, No. 7 (Series II), May 2003

 Knut Guettler — A Closer Look at the String Player's Bowing Gestures

bow force. When starting a bow stroke, the string player has in
principle three options at hand: Figure 6: By angling the bow (Z - a), the contact point can be
moved along the string without introducing longitudinal friction
forces. From Guettler/Yorke Edition [6]
(1) starting the bow "from the air", i.e., with a certain bow
speed, but from zero bow force
(2) starting the bow "from the string", i.e., with some bow
force, but from zero bow speed
(3) starting the bow from zero bow speed and zero bow force.

The first method is always employed when crossing strings under a

slur. Violinists and viola players also frequently utilize it in order
to create gentle attacks, particularly at phrase openings.

Nonetheless, in a study performed by Guettler and Askenfelt [11]

advanced listeners discriminated strongly against attacks with
extensive nonperiodic slip/stick triggering, preferring few or no
irregular periods for the transients of "neutral attacks". To faciktate
smooth accelerations, many players advocate letting the arm/hand
describe a figure eight or a. narrow oval (the bow's contact point on
the string remains nearly unaltered) rather than moving the frog
back and forth restricted to a perfectly straight line [12]. The
movement thus becomes continuous instead of ceasing at the end of
each stroke. The small excursions from the straight line could be
Figure 7: In spiccato and ricochet, the bow-stick rotates around a
performed in the z-plane (verticaky with respect to the instrument's
point near to the thumb on the frog. During a spiccato this point is
body), or in the x-plane (parakel to the strings). In skiked smooth
bow changes, the bow's deceleration is normally seen to be of
moved diagonally up and down kke shown in panel a), while the
rotation (see arrow at the tip) takes place with twice the frequency.
greater magnitude than the acceleration that follows. Simulations
In ricochet, the frog is moved in one direction only, forwards or
show that decelerationis less likely to produce noise from extra
backwards. Panels b) and c) display trajectories resulting from good
skps than is acceleration of the same magnitude.
and bad timing between the two movements during spiccato.
Spiccato and Ricochet bowing Trajectories as shown in panel c) will involve scratchy attacks.
Both spiccato and ricochet involve a combination of rotational and
translational movements [13] (see Figure 7). With correct timing
between these two movements, very clean attacks can be produced
in spiccato, in which case the first release ofeach stroke takes place
just before the normal force reaches its peak. After each jump, the
bow should land on the string while still moving in the "old" bowing
direction, to ensure a quick noise-free damping of the fading waves
of the last stroke before an "on-the-string" bow change takes place.
In ricochet, where the orientation of the Helmholtz corner remains
unaltered, damping is not required, and clean attacks are hence
much easier to achieve. When moving a bow from the frog to the
tip, the natural bouncing rate varies from ca 6 to 30 Hz (provided b)
the bow in contact with the string [9]). A fast spiccato is best
performed at a point where its natural rate kes one or two Hertz
above the driving (spiccato-note) frequency.

PLAYING "HARMONICS" (FLAGEOLET TONES) AND

MULTIPHONICS
''Harmonics" (i.e., higher-pitched tones produced by touching the
string lightly, thus filtering out certain partials) should be excited
r> — w-.-iMSSMSI
:f,

with bowing parameters suitable for that same pitch fingered on the
same string (i.e., stopped where the highest node of the "harmonic"

CASJ Vol. 4, N0.7 (Series II), May 2003 15

 Knut Guettler — A Closer Look at the String Player's Bowing Gestures

is situated). The effective ft is always relative to the highest node of REFERENCES

—
the "harmonic" not to the any other node that happens to be [1] Raman, C. V. (1920-21) On the mechanical theory of the
fingered. As a rule of thumb, compared to the open string, the bow vibrations of bowed strings and of musical instruments of the
speed (and the transient acceleration) of a "harmonic" should be violin family, with experimental verifications of the results:
increased by a factor n = fHARM /f0 , where /HARM and f0 are the Part II—Experiments with mechanically played viokns, Proc.
fundamental frequencies of the "harmonic" and the open string, Indian Assoc, for Cultivation of 6. 19-36.
respectively [14]. [2] Schelleng, J. C. (1973) The bowed string and the player, J.
Acoust. Soc. Amer. 53(1), 26-41.
Multiphonics can be achieved by bowing on the nut side of a kghtly [3] Cremer, L. (1972 and 1973) The influence of "bow pressure"
touched "harmonic"node. Swift changes between the pitches of the on the movement of a bowed string. Part I and 11. NL. Catgut
open string and the "harmonic" will occur. Acoust. Soc. #18 pp. 13-19and #19, 21-25.
[4] Pickering N. C. "String tone quakty related to core material",
UP-BOW VS. DOWN-BOW CASJ. 1(5) Ser.ll, 23-28, 1990.
Spectral differences between up-bow and down-bow in steady state [5] Askenfelt, A. (1989) Measurement of bowing parameters in
have been claimed, but not convincingly documented. If any, the violin playing. II: Bow-bridge distance, dynamic range, and
main difference kes in the attack transient, where dissimilar dynamic limits of bow force.
properties at the tip and the frog (e.g., the bouncing rates stated [6] Guettler, K. (1992) A Guide to Advanced Modern Double
above) play a role in shaping the bow-force envelope. Also, a strong Bass Technique, Yorke Edition, London.
coupling between a mode of the hair in the z-plane and wood- [7] Pitteroff, R. (1995) Contact mechanics of the bowed string.
dominated "bouncing modes" of about 130 and 150 Hz [15], might Doctoral dissertation, Univ. of Cambridge, U.K.
influence the timbre skghtly when playing some 40 cm away from [8] Mclntyre, M.E., Schumacher, R.T., and Woodhouse, J. (1981).
the frog, but even here sokd documentationof spectral changes is Aperiodicity in Bowed-String Motion. Acustica 49(1), 13-32.
lacking. [9] Askenfelt, A., and Guettler, K. (1998) "The bouncing bow—
An experimental study" CASJ Vol. 3, No 6 (II), 3 - 8.
AN ADDITIONAL REMARK [10] Guetder, K. (2002). On the Creation of the HelmholtzMotion
To complete the picture, it should finally be mentioned that the in Bowed Strings. Acustica - Acta Acustica 88(6), 970-985.
player has one more important timbre-shaping device at hand. To [11] Guetder, K. & Askenfelt, A. (1997) Acceptance kmits for the
be more precise, at the left hand: the finger pad. The reflection duration of pre-Helmholtz transient in bowed string attacks.
properties of the finger pad highly influence the output spectral J. Acoust. Soc. Am. Vol. 101, No 5, Pt. 1, 2903-2913.
envelope. A soft finger-pad tissue wik filter out, or reduce, the higher [12] Williams, C.E. (1985). Violin bowing skill analysis: the
partials, thus reducing brilliance. It should be enough to mention mechanics and acoustics of the change in direction
the timbre difference between open and stopped strings. The finger- (Unpublished doctorate dissertation: University of
tip angle and firmness of pressure might play roles here. On the Melbourne, Australia), see from page 54.
double bass, vibrato played with two or more fingers is sometimes [13] Guettler, K. and Askenfelt, A. (1998). On the kinematics of
seen. In those cases the adjacent, lightly touching, "extra fingers" spiccato and ricochet bowing. Catgut Acoustical Society
are used mainly to modulate the spectrum, not the fundamental Journal, 3(6) Ser. 2, 9-15.
frequency. [14] Guetder, K. (2002). On playing "harmonics" (flageolet tones)
Catgut Acoustical Society Journal, 4(5) Ser. 2, 5-7.
[15] Guetder, K. (1997). Bow notes. Proc. Institute of Acoustics
(ISMA '97), Vol. 19Part 5. 1-8.

16 CASJ Vol. 4, No. 7 (Series II), May 2003

 Wood Testing by Tap-Tone

Thomas M. Cox
3667 Woodland Dr
Baldwinsville, NY 13027
[email protected]

SUMMARY scientific publications express E in metric units, the most common

The maker can get a good indication of sound quality potential being the Pascal (Pa, the metric unit of pressure; 1 psi=6B92 Pa).
from the dimensions, weight, and a single tap-tone from each of [Note: Metric units are expressed here as MKS (meter, kilogram,
two smak test strips cut from each piece of tonewood. Two strips and second) and CGS (centimeter, gram, and second). Both MKS
measuring 10 x 2.5 x 0.2 cm are suspended by taut string at the and CGS units are commonly used in measuring wood properties,
quarter points and tapped to get the lowest frequency as a free bar. and hence are used here. Conversion factors are as follows: 1 Pa =
The ratio of the frequencies provides a measure of the ratio of the 0.1026 kg/m 2 (MKS) = 0.01026 gm/cm 2 (CGS); 1 MPa = 106 Pa =105
—
two stiffnesses along the grain and across the grain. Fundamental kg/m 2 = 104 gm/cm 2 ].
free-bar tap tones (F) are used to calculate E (modulus of elasticity),
c (sound velocityin the wood), and the figure ofmerit (c/p). Target To get E in terms of f we use
values to be met or exceeded are:
E (Pa) = 0.946 p if length / thickness
4 2
(Equation 1) [2]
E = 10,000 MPa
c = 4500 m/sec pis density (gm/cm2). p (MKS) = 1000 p (CGS).
c/p = 1.2 milkon cm 4/gm sec (CGS units)
= 12 m4/kg sec (MKS units) Sound Velocity
To get sound velocity c in terms of f we use
Bi-tri rplate toningo is faciktated using
o
values of E. , ./Ecross . that
with grain
are less than 15 (less than 15 times weaker). Bi-plate tuning is f = 1.028 c thickness/length2 (Equation 2 ) [3]
favored by values of less than 30.
Note that constant 1.028 remains the same whether MKS or CGS
INTRODUCTION units are used. We can get E in terms of c by rearranging Equation
A simple wood test involving measuring the tap tone of wood 2 and substituting in Equation 1.
samples enables the builder to predict if the choice of wood will
attain certain target values related to wood quakty. Previous papers c = 0.9728 f length2 / thickness (Equation 3)
[1,2,3,4] have shown that it is favorable to attain certain values for c 2 = (0.9728)2 (fo 2 length4 / thickness2 ) = (0.9728)2 (E / 0.946 p) =
the modulus of elasticity (E), sound velocity in wood (c), and the 0.946 E / 0.946 p°
figure of merit (c/p). Some studies have used ultrasonic meters,
oscilloscopes, and other specialized equipment for such tests, Hence E=c2 p ....(Equation 4) [Units must be c (m/s), E (Pa), p (kg/
although it is possible to obtain useful values by tapping wooden m 3)] /
test strips.
To calculate E, first find c = 0.9728 f lengtiV/thickness (cm/sec)
THEORY (Equation 4)
We will use fundamental free-bar tap tones (f) to calculate elastic
modulus (E), sound velocity in the wood (c), and stiffness ratio. Then divide by 100 to convert c to m/sec.
Test strip dimensions and weight are used to calculate density (p).
Figure of Merit
Modulus of Elasticity The popular figure of merit (c/p is obtained using Equation 3 to
The modulus of elasticity (E) is a measure of specific stiffness of a calculate c. Leave in cm/sec, divide by p, where p=weight/length
material. Its Engksh units are pounds/square inch (psi). However, x width x thickness (CGS), to get c/p. To get high plate tap tones on

CASJ Vol. 4, N0.7 (Series II), May 2003 17

 Thomas M. Cox — Wood Testing by Tap-Tone

a viokn top weighing under 70 gm, c/p should be greater than 1.2 Measuring the tap tones of our wood strips, we obtain f = 1000
million cm4/gm sec =12 m4/kg sec. Hz and fcross = 250 Hz. The stiffness ratio, calculated using Equation
_
s>is5> is Emth / E s = (fw th / f ros/ = (1000/250)
__o.
v
' 2 =42= 16 o(Note: this
Please note that I have not included Q (The ratio of height to width value is OK for bi-tri tuning).
of a resonance peak in the frequency domain) in the figure of merit
(as c Q/p). I am unsure of its value to preliminary wood tests _
We obtain sound velocity c usingo Equation 3: c -with = 0.9728 fwith
J
because of the extreme variabikty of plate-wood Q to that of the lengthVthickness = 0.9728x1000x100/0.2 = 486500 cm/sec = 4865
finished instrument's resonances.. See [I], page 24. m/s (Note: this is a typical value for good spruce). The speed of
sound for the cross-cut piece is c =1216 m/sec (Note: cis pro-
Stiffness Ratio portional to f for same length and thickness. Hence c =c ,/4).
Q
We can calculate the, stiffness ratio based on E, or by squaring the
ratio of the tap tone frequencies. We note in Equation 1 for the two We obtain elastic modulus E using Equation 4: E. = c2 p = 4865 2 x
strips p, length, thickness are common and cancel out, leaving the 400 kg/m3 = 9.47x109 Pa = 9470 Mpa (Note: thisvalue should exceed
ratio 10000Mpa). E for the cross-cut piece is E^ = 591 Mpa (Note: Ewjth
/Ecross =9470/591 = 16).
- '
../Ecross = 0.946p (fwith/
E with >■ .J2 L 4/T2
= (f .J{cross'f (Equation 5)
' ' \ with'

0.946p1 (f . 2 4
/T
v cross) L 7
2 We obtainthe figure of merit using c^Jp = 1.22 x 106 (CGS) = 12.2
m 4/kg-sec (MKS) (Note: fine spruce exceeds about 12 m4/kg-sec
[3]). The figure ofmeritfor the cross-cut piece is c /p = 3.04 x 105
It has been observed that bi-tri plate tuning is difficult to achieve if
this ratio is greater than 15. Bi-tuning is difficult if this,, ratio is (cgs) = 3.04 (mks). (Note: this value should exceed 4 m4/kg-sec).
greater than 30.
COMPARING RESULTS
When the first set- of data is obtained, it can be used as a base-kne
TEST PROCEDURE
from which subsequent test strips can be evaluated, if the original
Cut two test strips to same dimensions from both top and back
dimensions are duplicated accurately. This is done by simply
tonewood; one with grain lengthwise and one with grain crosswise.
comparing the new tap tones with the base-kne ones.
Strips as small as 9 cm x 2.5 cm x 0.2 cm work fine. With each pair
of strips, match thickness carefuky. A 10% difference in thickness
wik cause a 20% error in stiffness ratio. Giving the original, parameters the subscript " ", we get c/c = f/f
(since length and thickness are unchanged). Hence c= c f/f.
Next, make a smak frame to hold two loops of string, each with a Next, p is calculatedfrom the new weight, which gives a new figure
of merit (c/p). Equation 5 is calculated direcdy from the taprtones.
rubber band in series to keep in tension. Rest the strip on the strings
[Note that if wood were isotropic it would be good practice to .
Also, if it is desired to know E, then E= Eg (c/cj 2 f/f However,
avoid cutting length-to-width ratios to integer ratios to obviate Equation 5 -and c/p are sufficient to determine the suitabikty of the
tonewood.
harmonics that could confuse identifying the frequency. However,
there is no need for doing this for spruce].

A simple way check tap tone frequency is with a piano keyboard. Figure 1. Cantilever method for testing wood
Even toy electronic keyboards are remarkably accurate since they
use quartz crystals, simkar to those used in digital clocks. Mark the
key frequencies on tape and set behind the keys. [Editor's note:
Another way to capture the tone is to use a microphone and the Micrometer
CoolEdit computer program (see article by King, this issue)]

SAMPLE CALCULATION
As an example, let us assume that we have prepared two wood
strips of same dimensions: length =10 cm, width = 5 cm, CLAMP
thickness=o.2 cm, and weight=4 gm (CGS units). One strip is cut A

with the grain, and one is cut across the grain. Volume equals 10
cm3 (10 x 5 x 0.2), so p equals 0.4 gm/cm 3 (4gm/10cc = 0.4 gm/cm 3 =
400 kg/m 3). [Note: the fokowing calculations can be done using a
scientific calculator (select [FSE, S for scientific) or the calculator Clampstring with strip at
bundled with Windows computers (select View/Scientific). Large
end
t P ( gm)
numbers will be displayed as powers of 10].

18 CASJ Vol. 4, No. 7 (Series II), May 2003

 Thomas M. Cox — Wood Testing by Tap-Tone

VERIFICATION OF THE TAP-TONE METHOD [3] Meyer, H,G, 1995, A practical approach to the choice of tone
The tap tone method was checked against a mechanical method of wood for the instruments of the violin family: Catgut Acoust.
measuring E similar to [4], but simpler to implement. The test strip v
Soc. Journal, 01.2, n0.7 (Series 2), p. 9-13.
is set up as a cantilever beam (Fig. 1), clamped at one end with a [4] Wood, C, 1996, Wood testing: Southern California
fixed weight at the other. A 34 cm strip was reduced in 3 cm steps Association of Violin Makers Bulletin. Also at
from 30 to 12 cm free length, with f measured at each step, f is http://www.scavm.com/Woods.htm
.
proportional to 1/length2 The measured value was compared to
the computed value at each step. Mean error was +/- 2.8%. E was REVIEWER'S REAdARKS:
calculated both ways. The six test mean value for c = 4033 m/s and "Very simple to do with precisely cut pieces from corners of the wood
from this, E = 9.92x10 9 Pa. Mechanical test values were c = 4024 blanks. A.nd it providesprecisely the information about the wood that is
m/s (E = 9.73xl09 Pa). most important. "

REFERENCES "I would do that test on everypiece of wood in storage as an essential bit
[1] Schelleng, J.C., 1963, The violin as a circuit: Journal Acoust.
of data to helpme choose decent wood."
Soc America, vol. 35, no. 3, p. 326-338.
[2] Hutchins, M.A., 1983, Physical measurements on sampling of
European spruce and maple for violin top and back plates:
Catgut Acoust. Soc. Newsletter No. 40, p. 28-30.

CASJ Vol. A, N0.7 (Series II), May 2003 19

 Tap Tones of Old Italian Top Plates

Jeffrey S. Loen
[email protected]

The practice of measuring tap tones is embedded in violin making In summary, these data on tap tones of old masters seem to confirm
tradition, even among makers who have little interest in more Hutchins' [1,5] statements that the tap-tone resonance (mode no. 5)
advanced acoustical techniques. Makers estimating tap tones on of professional violin top plates should have a frequency of 360-370
their own instruments often wonder about the tap tones of fine Hz. However, it should be pointed out that the effect of tap tones,
Italian instruments, although little such data have been published. and free-plate tuning in general, on the resulting frequency spectrum
Table 1 gives tap tones (from author's unpublished files) measured of the violin has been questioned [6]. It is useful to remember
for 1 1 top plates of old Italian instruments played by professionals. Curtin's remarks [7] that "The trouble with plate tuning is that it
Estimation of the frequency of the "ring mode" (mode 5 [1]) was offers seductively precise data whose meaning is only partially
done by professional violin makers and restorers in shops in the US understood." Additionalresearch is needed. If there are any useful
and Europe while the instruments were in for repairs. relationships between tap tones and instrument performance, we
are unlikely to discover them without compiling comprehensive
Results show that plates with bass bar generally give an F sharp descriptive and acoustical data from a range of instruments.
note (about 370 Hz). Notes without bass bar are 55-60 Hz lower,
and in one case a plate with an old bass bar is 45 Hz lower than the REFERENCES
same plate with a new bass bar.
[1] Hutchins, CM., 1983, Plate tuning for the violin maker: Catgut
Variables that control tap tones include outline, graduation, arching, Acoust. Soc. Newsletter no. 39, p. 25-32.
physical properties of the wood, and surface coatings such as ground [2] Hutchins, CM., Stetson, K.A., and Taylor, PA., 1971,
and varnish [1,2,3]. In addition, many of these instruments have Clarification of "free plate tap tones" by hologram
had extensive repairs including sound post patches, doubled edges, interferometry: Catgut Acoust. Soc. Newsletter no. 16, Nov.
and chest patches. Graduation values are known for these 1971, p. 15-23.
instruments (table 1), although arching character and wood [3] Strobel, H.A., 1988, Useful measurements for violin makers:
properties are not recorded. A few instruments having greater A reference for shop use (4th Ed.): Henry A. Strobel,
thickness values between the ff-holes seem to have slightly lower Aumsville, OR, 46 p.
tap tones. [4] Hutchins, CM., 1970, More about plate tap tones: Catgut
Acoust. Soc. Newsletter no. 14, p. 15-17.
Previous work gives some additional data on tap tones of old Italian [5] Hutchins, CM., 1982, Stradivarius plate tests: Catgut Acout.
violins, and a variety of recommendations for tap tones of top plates. Soc. Newsletter no. 37, p. 30.
Hutchins [4] presents resonance curves for a 1727 Stradivari violin [6] Schleske, M., 2000, Eigenmodes of vibration in the working
showing a tap tone at D sharp (315 Hz). Hutchins et al [2] list values process of a violin: Catgut Acoust. Soc. Journal vol. 4, no. 1
of D to D sharp (304-307 Hz) for "well-tuned plates", although they (Series II), p. 90-95.
present data showing that a Guarneri del Gesu violin top rings at F [7] Curtin, J., 1993, The trouble with plate tuning: Violin Society
sharp (363 Hz). The 1713 "Wirth" Stradivariviolin (Note: this is a of America, Proceedings of 20th Convention (Part I), vol. 12,
differentviolin from the 1713 Strad listed in Table 1) gave a tap tone no. 3, p. 21-34.
of F sharp (360 Hz) [s]. Strobel [3] gives a representative value of F
(349 Hz) for violin.

20 CASJ Vol. 4, No. 7 (Series II), May 2003

 Jeffrey S. Loen — Tap Tones of Old Italian Top Plates

Table 1. Tap tones of Old Master plates. Tap tone notes are as reported (frequencies in Hz are approximate), "nd" indicates no date.

Name, place, date Tap tone of Thickness (mm)

(mm)
top plate (frequency, Hz) between ff-holes

Andrea Guarneri, Cremona, c 1677 F# (370) 1.8-2.5

Joseph Guarneri filius Andrea, Cremona (nd) F# (370) 2.1-2.7

Antonio Stradivari, Cremona, 1700 slightly flat F (340) with new bar 2.2-2.9

Antonio Stradivari, Cremona, 1707 F# (370) 1.8-2.5

Antonio Stradivari, Cremona, 1712 F#+(37s) 1.9-2.7

Antonio Stradivari, Cremona, 1713 1/8 note above F (355) 2.2-2.5

Peter Guarneri of Venice, Venice D (294) without bar 2.7-3.3

F (349) with new bar

Peter Guarneri of Venice, Venice 1729 D# (311) withoutbar 2.1-3.0

F# (370) with new bar

Guarneri del Gesu, Cremona, 1742 F (349) 2.2-2.8

G.B. Guadagnini, Taurin, 1782 low E (325) without bar 2.9-3.3

low G (385) with new bar

G.B. Guadagnini, Taurin, 1785 skghtly flat E (325) with old bar 2.6-3.2
F# (370) with new bar

CASJ Vol. 4, N0.7 (Series II), May 2003 21

 Historical Instruments
The Catgut Acoustical Society has long been known for detailed slowly extricating itself from the quagmire of secrecy, mythology,
technical studies of vibration, resonance, and tone. However, other and reliance on opinions of a few strong personalities in which it
topics arise as the composition of our membership shifts has been mired for 250 years. Rather than relying on highly filtered
increasingly from researchers to violin makers (see results of information from such authorities as the Hills or Sacconi (some of
Readers' Survey, this issue). In the violin making world, the ultimate whom had their own agendas), we are finally allowed to examine
measure of success is attaining favorable visual and acoustical raw data and make our own decisions. The best makers have been
comparisons with the famous icons of 18th Century violin making highly successful at applying this new wealth of knowledge.
constructed by Stradivari, Guarneri, and other classic makers. Such Conversely, we might ask "Is it likely that equally skilled makers
success, which previously was rarely attained, is now enjoyed by who are uninformed about classic instruments can construct one
an increasing number of makers worldwide. How has such unlikely that looks or performs like an old master?" I think not.
success suddenly been grasped by so many makers (to the obvious
glee of players fortunate enough to obtain their productions)? And
how can CAS help more makers attain this success? ..
". the good, the bad, and the sometimes ugly
facts..."
Advances in acoustical understanding, in which CAS has been a
leader for 40 years, have played a role in this success. However, the
real story is that successful makers have paid close attention to This issue of CASJ departs somewhat from our traditional
historical research on classic instruments. Makers have done their concentration on acoustical research and focuses on new findings
own research, and also take advantage of the fruits of a recent by some of the world's leading researchers on classic instruments.
blossoming of organized research and publication including These papers should be like a breath of fresh air to instrument
exhibitions on Stradivari, Guarneri Del Gesu, Vuillaume, Venetian makers, hungry for first-hand descriptions of what classic
(and other) makers, Pollens' book on Stradivari's violin forms, new instruments are really like. These papers report the good, the bad,
work on Guadagnini, and the availability of good technical and the sometimes ugly facts about instruments that often are not
documentation (e.g. Strad Poster Series, and Pringle's technical quite as close to perfection as we have been led to believe. Acoustical
drawings from the Ashmolean and National Music Museum). researchers can also benefit from reading such studies, inasmuch as
Equally important, investigators of topics such as classic instrument this information might help them to calibrate their experiments
design, acoustics, varnish, arching, graduations, fittings, and and computer models to real world examples.
dendrochronology have been willing to share data and conclusions
with the violin making community. As a result, the profession is

Ornamental back of a Brescian violin, c. 1640 Scroll on "Harrison" Strad (National

(National Music Museum, photo by A. T. King). Music Museum, photo by A. T. King).

22 CASJ Vol. 4, No. 7 (Series II), May 2003

 Measurements of Principal Resonance Modes of Historic
Instruments

A. Thomas King
[email protected]

ABSTRACT [6]. I have glued a small piece of thin leather to one end of the tube
Frequencies of principal modes (AO, 80, 81, Tl, and C2) were to protect the instrument's varnish. Place the tube lightly on the
measured using the Fuhr technique for 19 historic instruments instrument, moisten a thumb and finger and grip the top of the tube
ranging in size from a violino piccolo to a tenor viola. The lightly between them. Slide the fingers downthe tube. With a bit of
instruments include 14 museum specimens (made in Italy, Germany, practice, you will hear distinctive noises, something like the honking
Austria between 1564 and 1824), and five 18th century Italian violins of a goose. The tone you hear is an eigenfrequency of the violin
played by performers and teachers. Generally, body length shows corpus.
a negative correlation with modal frequency, with the tiny violino
piccolo giving very high frequency modes, and the large violas giving Each of the corpus modes, 81, Tl, and C2, is activated by placing
low values. Full-sizeviolins gave "shotgun" patterns for most modes, the glass tube at specific locations suggested by Schleske (personal
and the values compare favorably with previous work. The clearest communication, 2002). 81, the main corpus mode, is activated by
and most definitive modes are associated with the finest placing the tube near the inner edge of the bass f-hole a little below
professional instruments, including Italian violins made by Carlo the notch. Tl is activated by placing the tube about 30 mm in from
Bergonzi, Domenico Montagnana, Francesco Stradivari, and Carlo the edge and near the middle of the treble C-bout on the back plate.
Antonio Testore. C2 is activated by placing the tube near the purfling near the middle
of the treble C-bout on the back plate.
INTRODUCTION
For the past several years I have been collecting modal information DATA CAPTURE AND ANALYSIS
for the better instruments that I have had the opportunity to examine. I record the sounds of the modes using a small microphone attached
I was inspired to do this by Martin Schleske's path-breaking work to a laptop PC running the program "CoolEdit" [7]. The program
on making "tonal copies" [1,2]. If one hopes to make tonal copies, allows the user to record, edit, and analyze the frequency spectrum,
then information about modes is necessary. Unfortunately, little and identifies peaks in cycles per second (Hz; Fig. 1).
such information is readily available.
SOURCES OF ERROR
PROCEDURE Please note my study methods are simple and subject to uncertainty.
To collect modal information (frequencies of AO, 80, 81,T1,C2), Recordings took place in the ambient space with no ability to control
we need a way to activate the modes ofinterest, a way to capture the echoes or other reverberations. Although one becomes proficient
data, and a way to analyze them that allows us to identify at generating the honking noises, it is possible to move the
frequencies. frequencies about somewhat by pressing down more or less firmly
and squeezing the glass tube more or less. Usually, I generated each
MODE ACTIVATION mode several times, giving me the opportunity to exclude spurious
To activate the main air or "Helmholtz" (AO) mode, I blow lightly trials. Finally, errors may exist in the analysis and identification of
across the bass f-hole. various modes. As a rule, values for AO and BO are reliable. Bl is
usually strong and definitive. Tl and C2 are somewhat less certain.
To activate the body (BO) mode, I hold the instrument near the
purfling at the widest point of the lower bout with my left hand INSTRUMENTS ANALYZED
thumb and middle finger. With the bridge resting lightly on my I have been able to test several instruments over the past two years,
chest (to dampen the strings), I tap the scroll gently with my other usually as part of the work I have done with JeffLoen on measuring
hand (see [3] for more information on determining these modes). instrument graduations [B]. The quality of the instruments ranges
from fine soloist-grade violins by Carlo Bergonzi and Domenico
To activate all the other modes, I use the Fuhr technique [4,5], using Montagnana, to somewhat lower-level instruments by makers of
a thin glass tube about 1.5 mm in diameter and about 150 mm long lesser renown.

CASJ Vol. 4, N0.7 (Series II), May 2003 23

 A. Thomas King — Measurements of Principal Resonance Modes of Historic Instruments

Figure 1. Example of determination of frequency of air cavity Figure 2. Cremonese violin made by Hieronymus Amati in 1609
mode (AO) using Cool Edit [7]. Frequency response outlined in (National Music Museum).
white was recorded using microphone while blowing across bass
ff-hole. Analysis screen (lower right) identifies AO frequency as
264 Hz.

Figure 3. Fine Cremonese violin made by Carlo Bergonzi in 1732

Fourteen instruments were studied at America's National Music

Museum in Vermillion, SD (I thank Museum Director Andre Larson
for permitting me to examine these instruments). These include
one tenor viola, two standard violas, nine full-size violins (body
lengths 328-365 mm; Fig. 2), one half-size violin, and a violino piccolo
(Table 1). The instruments were made in Cremona, Brescia, Milan,
and Florence, Italy, and in Vienna, Austria and Markneukirchen,
Germany. Several of these instruments are in unmodified condition
(see Table 1). These instruments were made by important European
makers, although it was impossible to evaluate their tone.

In addition, I studied five privately owned full-size Italian violins.

Two of these, a Montagnana and a Bergonzi (Fig. 3), are owned by
professional performers. Three others, a C.A. Testore, an F.
Gagliano and a J. Gagliano are owned by teachers who also perform.
All except the J. Gagliano are judgedsubjectively to have outstanding with results of previous work, for example, Schleske determined
tonal qualities. 268 Hz for a 1740 Montagnana violin [I], and 286 Hz for a 1712
Stradivari violin [2].
FINDINGS
In general, mode frequencies show a negative correlation with The body (BO) mode ranges from 178 to 352 Hz, and the values for
instrument body length (Table 1; Figs. 4,5). A priori, this inverse full-size violins range from 240 to 330 Hz. Air and body modes are
correlation would be anticipated for the air mode (AO) because of poorly coupled in most of these instruments (Fig. 4), and values
larger air volumes in longer bodies. A "shotgun" pattern is evident differ by as much as 54 Hz. Instruments showing the best AO-BO
within the data spread for full-size violins. matching (6 Hz or less) are the PJ Rogeri violin, the Montagnana
violin, the F. Gagliano violin, and the Zanetto viola. Previous work
The main air mode (AO) shows a broad range (172-343 Hz) related [9, 10] suggests that AO-BO matching has a positive influence.
directly to the size of the air cavity, with the violino piccolo showing
the highest frequency and the tenor viola showing the lowest value. The Bl (main corpus) mode has been described as a "tonal
Full-size violins gave values in the range 253-307 Hz. This agrees barometer" of the instrument, with values below 510 Hz being

24 CASJ Vol. 4, No. 7 (Series II), May 2003

 A. Thomas King — Measurements of Principal Resonance Modes of Historic Instruments

Figure 4. Relation of main air mode (Helmholtz mode; AO) and Figure 5. Relation of higher frequency corpus modes to body
body mode (BO) to body length. No V value was recorded for length. No V values were recorded for values lying on X-axis.
value lying on X-axis.

0 100 200 300 400 500 600

Body length (mm)
0 100 200 300 400 500 600
Body length (mm)

REFERENCES
characteristic of "soft" violins and values above 550 Hz having a [1] Schleske, M., 1996, On making "tonal copies" of a violin:
harsher, brighter sound [2]. Full-size violins in the study cover a Catgut Acoust. Soc. Journal, vol. 3, no. 2 (Series II), p. 18-28.
broad range of Bl frequencies (479-580 Hz). This range is somewhat [2] Schleske, M., 2002, Empirical tools in contemporary violin
broader than therange of500-565 Hz given by Schleske [2] for 1 1 old making—Part I. Analysis of design, material, varnish, and
violins. All are higher thanthe value of475 Hz given by Schleske [1] normalmodes: Catgut Acoust. Soc. Journal, vol. 4, no. 5 (Series
for a 1740 Montagnana violin. Schleske gives a value of 524 Hz for II),p. 50-63.
the Bl of a 1712 Stradivari violin [2]. [3] Hutchins, CM., and Voskuil, D., 19934, Mode tuning for the
violin maker: Catgut Acoust. Soc. Journal, vol. 2, no. 4 (Series
The Tl mode, the lower of the two corpus modes, causes strong II),p. 5-9.
sound radiation due to pumping movements and vibrations [2]. [4] Fuhr, X., 1958, Die Akustischen Ratsel der Geige (The
Values recorded range from 334 to 527 Hz. Values for full-size acoustical riddles of the violin): Verlag Friedrich Hofmeister,
violins range from 418 to 494 Hz, compared to 448 Hz for a 1712 Frankfurt am Main, 187 p.
Stradivari violin describedby Schleske [2]. The C2mode, the higher [5] Wall, E., 1993, A trial of Fuhr's method of testing violins and
of the two corpus modes, is related to strong twisting of the body, violin plates: Catgut Acoust. Soc. Journal, vol. 2, no. 4 (Series
and this moderadiates sound weakly [2]. Full-size violins gave C 2 II),p. 15-17.
values of 351-421 Hz, compared to a value of 409 Hz for a 1712 [6] A source for such tubes is Warner Instrument Corp., Tel.
Stradivari violin [2]. 203-776-0664 Fax.203-776-1278
[7] Cool Edit, developed by Syntrillium Software Corp. See
An additional observation based on this work is that the clearest http://www.syntrillium.com.
and most definitive modes are shown by the finest quality [8] Loen, J.S., and King, A.T., 2002, Thick and thin: The Strad,
instruments, such as the Bergonzi, Montagnana, and Testore violins, Orpheus Publications, London, vol. 113, no. 1352, p. 1354-
which are owned by professional players. The F. Stradivari violin -1359.
also had notably strong clear modes, although we were unable to [9] Spear, D.Z., 1987, Achieving an air-body coupling in violins,
evaluate its sound. violas and cellos: A practical guide for the violin maker: Catgut
Acoust. Soc. Newsletter no. 47, p. 4-7.
[10] Hutchins, CM., 1990, Some of the effects of adjusting the AO
and the BO modes of a violin to the same frequency: Catgut
Acoust. Soc. Journal, vol. 1, no. 5 (Series II), p. 35-37.

CASJ Vol. 4, N0.7 (Series II), May 2003 25

A. Thomas King — Measurements of Principal Resonance Modes of Historic Instruments

Table 1. Modal frequencies recorded for European violins and violas. Body lengths are caliper measurements. NMM numbers refer to
catalog numbers at National Music Museum, Vermillion, So. Dakota.

Body Frequency (Hz)

Maker (location) Date Instrument Length (mm) AO BO Bl Tl C 2 Comments

Bros Amati (Cremona) 1613 violino piccolo 266 343 329 761 527 n/a NMM No. 3361;
unmodified condition

Lorenzo Storioni (Cremona) 1793 small violin 299 309 352 n/a n/a n/a NMM No. 3359

Hieronymus Amati (Cremona) 1609 violin 342 284 240 502 432 372 NMM No. 3364

Brescian School (Brescia) c 1630 violin 355 287 n/a n/a 450 n/a "NMM No. 3363; Bl is
very weak, hard to
determine"

Brescian School (Brescia) c 1640 violin 352 284 293 522 445 380 NMM No. 3358

Pietro J. Rogeri (Brescia) 1715 violin 328 253 248 580 427 n/a NMM No. 3362

Composite N. Italian 1719 violin 332 307 297 512 454 399 Francesco Stradivari
belly; NMM No. 3414;
clear, lively modes
Carlo Bergonzi (Cremona) 1732 violin 349 282 253 535 459 393 Soloist instrument

Domenico Montagnana (Venice) 1740 violin 355 267 261 518 439 385 "Dutchess of Cleveland"
Concertmaster instrument

Carlo Antonio Testore (Milan) 1748 violin 355 288 240 544 451 421 Well-regarded; present
owner purchased in
preference to a Strad

Tomaso Carcassi (Florence) 1759 violin 356 262 306 497 436 369 NMM No. 4899; weak Bl

Ferdinando Gagliano (Naples) c!784 violin 346 274 269 528 434 351 Warm, dark sound. Not
strong solo instrument,
but very beautiful

Joseph Gagliano (Naples) 1791 violin 358 280 248 479 418 354 weak sound

J.G. Hamm (Markneukirchen) 1796 violin 365 287 267 517 468 415 NMM No. 5204;
unmodified condition;
Bl weak relative to Tl

Giovanni B. Ceruti (Cremona) 1801 violin 350 290 247 543 450 407 NMM No. 4900; weak Bl

M.I. Brandstatter (Vienna) 1824

1824 violin 360 276 330 536 494 413 NMM No. 6097
unmodified condition

Peregrino di Zanetto (Brescia) 1564 viola 438 204 200 369 334 276 NMM No. 3367

Andrea Guarneri (Cremona) 1664 tenor viola 482 172 134 n/a n/a n/a NMM No. 3354;
unmodified condition

Nichola Bergonzi (Cremona) 1781 viola 410 225 178 436 397 374 NMM No. 6046

26 CASJ Vol. 4, No. 7 (Series II), May 2003

 Reverse Graduation in Fine Cremonese Violins

Jeffrey S. Loen
[email protected]

ABSTRACT large differences in the overall structures of back and top plates.
Many classic Cremonese violins that are in demand by the best Back plates invariably define concentric patterns having the
modern players have top plates that are carved in the reverse of the maximum thickness near the center. Concentric contour lines vary
usual pattern. Rather than being thick between the ff-holes (normal from circular "bull's eye" patterns to longitudinal patterns (see fig.
graduation), these plates display minimumvalues (1.5-2.3 mm) near 1). In contrast, top plates often have a structure that, in the broadest
the center and thicken towards the edges (reverse graduation). An sense, has been described as a "membrane" [B]. However, when
evaluation of contour maps of 105 Cremonese violin top plates by examined in detail the area between the ff-holes is often different
ten Golden Age makers suggests that most (60%) of the plates are from the thickness of the upper and lowerbouts. Adequate data on
reverse graduated. Only nine plates (including four by Nicolo the area between the ff-holes are required to classify these areas as
Amati, none by Stradivari and three by Guarneri del Gesu) are thicker (normal graduation) or thinner (reverse graduation) than
normally graduated, similar to the pattern recommended in modern the upper and lower bouts.
violin making schools, books, and articles. Reverse graduated
violins include some of the best preserved Strads ("Messiah", "Muir- On published maps showing thickness data (for example, the Strad
MacKenzie") and del Gesus ("Alard", "Soil"), suggesting that this poster series [9]), measurements are often omitted in the critical
graduation pattern is not solely the result of regraduation. In area between the ff-holes, although data are sufficient to suggest
contrast, top plates of most Italian and German violins in the that Nicolo Amati's 1666 violin is normal graduated and Stradivari's
database made after the close of the Golden Age around 1750 are 1694 "Muir-Mackenzie" [9] is reverse graduated. Biddulph's volume
classified as having normal graduation. on Guarneri del Gesu [10] gives adequate data to show goodexamples
ofboth normal and reverse graduation. Additional graduation maps
INTRODUCTION of Cremonese instruments were obtained from research in
Modern violin makers are taught to make spruce top plates thick in museums, and donations from research collaborators in the US and
the center and thinner toward the edges ([1,2,3,4]; in roughly the Europe. My collaborators chose to remain anonymous, and ask
same way that maple back plates are graduated). The famous that instrument names be kept confidential. Clearly, some
"Canon" violin by Guarneri del Gesu, which was played by Italian publishers, instrument owners, and researchers do not want to be
soloist Nicolo Paganini, is often cited as a clean example of a associated with the stigma of "too thin" plates.
"normally" graduated Cremoneseviolin top plate (fig. 1) that should
be emulated. The implication is that the Canon is typical of Incidentally, it is important to use contour maps for distinguishing
Cremonese concert violins that have not been modified, and that graduation types, rather than the thickness graduation plots that
modified violins are somehow inferior acoustically and structurally. appear in most published sources of graduation data. Contour
However, before accepting such a notion we should ask, "how maps make distinguishing reverse vs. normal graduation patterns
common is 'normal' graduation in working Cremonese concert as easy as reading a topographical map.
violins?" Is the Canon example a general case, or is it a glaring
exception?
"How common is 'normal' graduation in
The overall style of plate graduation can be defined based on visual working Cremonese concert violins?"
examination of contour maps that show lines of equal thickness (in
millimeters; [see 5,6,7 for descriptions of the current project, and
computer methods of contouring]). In this paper 105 historic
Cremonese top plates are evaluated and "normal" and "reverse" NORMAL GRADUATION OF TOP PLATES
graduation patterns are distinguished. A contour map of a normally graduated plate resembles a
topographic map of a mountain peak or ridge, in which contour
DEFINING GRADUATION PATTERNS lines decrease concentrically away from the center (Figs. 1,2). The
Classic Italian violin plates show a variety of graduation patterns thickest values between the ff-holes are often 3.5 mm or more, and
on computer-generated contour maps, and it is common to see the wood pinches down towards minimum values of about 2.5-

CASJ Vol. A, N0.7 (Series II), May 2003 27

 Jeffrey S. Loen — Reverse Graduation in Fine Cremonese Violins

Figure 1. Thickness Contour maps of the 'Canon' violin, Guarneri del Gesu, 1743 [10]. The graduations are thought to be unmodified.
Top and back plates show normal graduation. Measured with magnetic thickness gauge. Black dots indicate locations of measurement
points. Contour interval 0.25 mm. Viewed from outside. Note: Caliper measurements by Candi (1937 [22]) suggest that the top plate is
actually thicker (4.1-4.3 mm in the center) and the back plate is thinner (max 5.6 mm rather than 6.2 mm) than is suggested by modern
measurements using magnetic thickness gauges. A possible explanation for the differenceis the large margin of error of the single-magnet
thickness gauge at thicknesses greater than 4.0 mm.

Upper bout 168 mm

Middle bout 112 mm Upper bout 168 mm
Lower bout 207 mm Middle bout 111mm
Rib thickness 1.3-2.0 mm Lower bout 206 m m
Mensur 197 mm Arching 15 mm
Arching 15 mm Rib thickness 1.3-2.0 mm
Two piece quarter maple

3.0 mm in the upper and lower bouts. Contour lines often are (Figs. 3,4). The thickest parts of reverse graduated Cremonese tops
irregular, sometimes containing closed contours of higher or lower are often at the edges, which attain maximum values around 4.0
values. mm. In addition, the areas between the ff-holes and the C-bouts
often show the maximum values (3.0-3.6 mm) of the part of the top
REVERSE GRADUATION OF TOP PLATES plate inside the linings. Patterns are often highly irregular. Contour
The opposite situation, in which contour values increase away from maps of these top plates often show a closed semicircular contour
the center, produces a map that superficially looks like a of minimum value, 20-35 mm in diameter, positioned at or slightly
topographical valley or basin. Reverse graduation patterns have higher than the bridge position (Figs. 5,6). In some cases (Fig. 7), the
minimum thickness values (1.5-2.3 mm) near the center between contours for minimum thickness values define a longitudinal trough
the ff-holes, and larger thickness values further away from the center extending from the upper to the lower bouts.

28 CASJ Vol. 4, No. 7 (Series II), May 2003

 Jeffrey S. Loen — Reverse Graduation in Fine Cremonese Violins

Figure 2. Thickness graduation maps of a violino piccolo, Brothers Amati, 1613, in unmodified condition [11]. Top and back plates show
normal graduation, in which contour lines decrease away from the center. Note: this is not a concert violin, although it is an important
example of an early Cremonese instrument. Measured with magnetic thickness gauge. Measurements are shown in millimeters (contour
interval 0.25 mm). Viewed from outside.

Body 266.0 mm
Upper bout 124.4 mm
Middle bout 85.3 mm
Body 266.2 mm
Lower bout 148.4 mm
Arching 10.5 mm Upper bout 125.7 mm
Middle bout 84.8 mm
spacing 34.9 mm
Rib height 22.5-24.2 mm Lower bout 150.0 mm
Arching 12.1 mm
Rib thickness <1. 0-1.1 mm
Weight 207 One-piece slab maple
Weight 207 g

CLASSIFYING GRADUATION PATTERNS OF GOLDEN AGE concentric contour lines in the upper and lower bouts. Top plates
VIOLIN TOP PLATES that have no discernable concentric contour patterns are classified
A sample population of 105 violin top plates by ten major Cremonese as "undefined", and top plates lacking measurements in the area
makers was systematically evaluated to determine the relative between the ff-holes are classified as having "insufficient data."
abundance of normal and reverse graduation patterns (Table 1).
The age of the instruments ranges from 1564 to 1737. Plates having Results (Table 2) suggest that 60% of these Golden Age violins are
concentric contour lines that generally decrease away from the reverse-graduated, 23% are undefined, 9% display normal
center are classified as "normal", whereas plates with contours that graduation, and 8% have insufficient data. The earliest Cremonese
increase away from the center are classified as "reverse". Reverse makers, Andrea Amati, the Brothers Amati, and Nicolo Amati,
graduated plates commonly have extreme minimum thickness each have a few normally graduated plates. Nicolo Amati has the
values of as little as 1.4 mm in the center and 1.0 mm in the bouts largest proportion of normally graduated plates, at four (50%) of
(Table 1). In many cases, particularly with violins by Stradivari, a the sample of eight plates. Guarneri del Gesu was the only other
chest patch (to add reinforcement and correct arching deformities) maker for which normal graduated tops are identified. For 38
obscures the graduation pattern in the ec-bout area, although a pre- Guarneri del Gesu violin top plates, three (the 'Canon,' 'Haddock,'
existing reverse graduation pattern can be inferred based on the and an unnamed 1742 violin) are classified as normal, 27 are

CASJ Vol. 4, N0.7 (Series II), May 2003 29

Jeffrey S. Loen — Reverse Graduation in Fine Cremonese Violins

Figure 3. Thickness graduation maps of the 'Kreisler' violin, Guarneri del Gesu, 1733, the career violin of American soloist Fritz Kreisler
[12]. Top plate shows reverse graduation, and back plate shows normal graduation. Measured with magnetic thickness gauge. Measurements
are shown in millimeters (contour interval 0.25 mm). Viewed from outside.

Body 353 mm
Upper bout 166.5 mm
Middle bout 108.5 mm
Lower bout 203 mm Upper bout 166 mm
Arching 14.5 mm Middle bout 108 mm
Weight 434.8 g Lower bout 203.5 mm
(w/boxwood chin rest Arching 14.5 mm
and accessories) Two-piece quarter maple

classified as reverse, and eight are undefined. Del Gesus thought to instruments, including the 1716 "Messiah," which is in nearly
be unmodified include the "Canon" (Fig. 1; normal graduated), the pristine condition, the 1694 "Muir-Mackenzie," which Hargrave [9]
"Alard" (reverse graduated), which Chaudiere [21] calls "the best calls "a rare and beautiful example of a 'clean' Stradivarius", and
preserved of all the 'del Gesu' instruments", and the "Soil" (reverse the 1693 "Harrison" (Fig. 8), considered "the greatest surviving
graduated), which Hargrave [9] describes as "in an almost perfect concert instrument of the 17th century."
state of preservation".
ADDITIONAL CONSIDERATIONS
The majority (60%) of 43 plates by Antonio Stradivari are classified Effects of Graduation Pattern on AcousticalPerformance
as reverse graduated, and none are normally graduated. Almost Golden Age violins known for high performance include both
one quarter of the Stradivari plates are undefined, probably in part normal and reverse graduated examples, although reverse patterns
because of effects of extensive repairs and patching. Violins with dominate. The fact that most of these reverse graduated instruments
reverse-graduated plates include many of Stradivari's most famous are working modern concert violins indicate that they are

30 CASJ Vol. 4, No. 7 (Series II), May 2003

 Jeffrey S. Loen — Reverse Graduation in Fine Cremonese Violins

Figure 4. Thickness contour maps of the 'D'Egville' violin (Guarneri del 1737 [10]). Top plate shows a well-developed reverse
graduation pattern, and back plate shows normal graduation. Measured with magnetic thickness gauge. Black dots indicate locations of
measurement points. Contour interval 0.25 mm. Viewed from outside.

Body 350 mm
Upper bouts 165 mm Upper bouts 164 mm
Middle bouts 109 mm Middle bouts 109.25 mm
Lower bouts 203.5 mm Lower bouts 204.5 mm
Arching 12.5 mm Arching 14.2 mm
2-piece maple, quarter

acoustically successful and, in highly desireable. The arching and wood quality remain equal. This is verified, in part, by
performance of some of the normally graduated del the extremely low weight (usually 355-385 g) of reverse graduated
especially the is likewise unassailable. For example, luthier violins in the database. Vigdorchik [23] comments that Stradivari's
Sesto Rocchi [22] stated "its maker used thicknesses then considered thin plates "flex in the opposite fashion" from plates of Nicolo
out of the average. I am convinced that those thicknesses are greatly Amati that were made thicker in the middle section. Vigdorchik
responsible for the fact that the violin has retained its power and [23] concludes "...it is this significant difference in the flexibility of
quality of sound, equaled in this only by very few other instruments." the plates that determines the characteristic timbres of these
This quote tells us that the Canon's thick graduations were unusual instruments..." I speculate that the acoustical consequences of
even when it was new. minimal thickness (as little as 1.5 mm) in the bridge area might
include larger vibration amplitude related to reduced mass loading.
Such major differences in thickness distribution would be expected
to cause differences in weight and flexibility, if other things such as In summary, examples of both normal and reverse graduated violins

CASJ Vol. 4, N0.7 (Series II), May 2003 31

Jeffrey S. Loen — Reverse Graduation in Fine Cremonese Violins

Table 1. Classification of graduation types, top plates of Cremonese violins, 1564-1738. "G" refers to "Guarneri." Grads is number of
thickness measurements on top plate. Class (Graduation class) is as follows: n, normal; r, reverse; i, insufficient data; undefined. Length
refers to plate length (caliper distance, mm). Arch is maximum arching height, mm. "Anon" source indicates contributor wishes to
remain anonymous. Mm (minimum), Mean (average), Max (maximum), and sd (standard deviation) are statistics on the thickness
measurements for each plate, n/a is "not available".

Maker Date GradIs Class Name Length Arch Source Mm Mean Max Sd
Amati, Andrea 1564 26 i Charles IX 343.5 16.0 [9] 2.10 2.80 3.30 0.29
Amati, Andrea 1574 135 n n/a 340 15.3 [11] 2.00 2.60 3.50 0.29
Amati, Andrea 1577 135 v n/a 351.5 15.3 [11] 1.20 2.11 3.30 0.45
Amati, Bros 1604 115 r n/a 340.5 18.0 [11] 1.00 1.74 3.30 0.49
Amati, Bros 1609 136 v n/a 342 14.5 [11] 1.40 2.13 3.10 0.40
Amati, Bros 1613 93 n n/a 266 11.5 [11] 1.50 2.04 2.70 0.34
Amati, Nicolo 1628 200 n n/a 352 14.2 [11] 2.00 2.61 3.30 0.30
Amati, Nicolo 1649 26 v Alard 352 16.5 [13] 2.50 2.97 4.00 0.29
Amati, Nicolo 1650 133 n n/a 353.5 n/a [14] 1.90 2.81 4.30 0.36
Amati, Nicolo 1654 125 r Brookings 353.5 15.9 [12] 2.00 2.58 3.70 0.35
Amati, Nicolo 1655 53 n n/a n/a n/a Anon 1.70 2.47 3.20 0.34
Amati, Nicolo 1666 46 n n/a 353.5 16.0 [9] 2.00 2.57 3.50 0.37
Amati, Nicolo 1667 72 r n/a n/a n/a Anon 1.70 2.34 3.00 0.33
Amati, Nicolo n/a 23 i n/a n/a 17.0 [15] 2.50 2.91 3.20 0.21
Bergonzi, C 1731 37 i n/a n/a n/a [9] 2.20 2.85 3.60 0.31
Bergonzi, C 1732 145 r n/a 349 14.2 Anon 1.60 2.53 3.50 0.33
Bergonzi, C 1734 55 r n/a n/a n/a Anon 2.00 2.60 3.20 0.35
G del Gesu 1727 46 r Dancla 354 16.8 [10] 2.10 2.83 3.90 0.41
G del Gesu 1728 89 r n/a n/a n/a Anon 1.80 2.56 3.30 0.29
G del Gesu 1729 46 r Stretton 351.5 15.0 [10] 1.70 2.72 3.70 0.44
G del Gesu 1730 140 r Goldberg 353 n/a [14] 2.10 2.80 3.80 0.33
G del Gesu 1730 134 r Kreisler 353.5 15.0 [12] 2.00 2.74 3.80 0.39
G del Gesu 1731 46 v Baltic 349 16.5 [10] 2.20 2.71 3.40 0.36
G del Gesu 1733 146 r n/a n/a n/a Anon 1.50 2.39 3.40 0.35
G del Gesu 1733 45 r Soil 352 15.5 [9] 2.10 2.62 3.50 0.36
G del Gesu 1734 46 n Haddock 349 16.5 [10] 2.30 2.71 3.60 0.28
G del Gesu 1734 63 r n/a n/a n/a Anon 2.00 2.41 2.90 0.19
G del Gesu 1734 127 r n/a 350.3 20.5 Anon 2.00 2.56 3.40 0.32
G del Gesu 1734 46 r V. dv Diable 349.8 15.5 [10] 2.10 2.65 3.20 0.26
G del Gesu 1735 46 r D'Egville 349 12.5 [10] 2.20 2.73 3.50 0.31
G del Gesu 1735 47 r King 350 13.0 [10] 2.70 3.13 4.20 0.28
G del Gesu 1735 27 v n/a n/a n/a Anon 1.90 2.48 3.00 0.26
G del Gesu 1735 46 r Plowden 350 13.5 [10] 2.20 2.83 3.50 0.29
G del Gesu 1736 19 r n/a n/a n/a Anon 2.50 2.80 3.20 0.19
G del Gesu 1737 46 r Joachim 351.5 15.8 [10] 2.50 2.86 3.40 0.22
G del Gesu 1737 46 r King Joseph 351 16.0 [10] 2.20 2.69 3.30 0.29
G del Gesu 1737 46 r Stern 352.5 14.7 [10] 1.80 2.73 3.40 0.36
G del Gesu 1738 126 r Adam n/a n/a Anon 1.80 2.62 3.70 0.25
G del Gesu 1738 46 r Kemp 352.8 15.2 [10] 2.30 2.77 3.40 0.25
G del Gesu 1739 46 r Kortschak 351.5 17.4 [10] 1.60 2.51 3.60 0.37
G del Gesu 1740 46 v Heifetz 354.5 15.8 [10] 2.20 3.02 3.50 0.25
G del Gesu 1740 45 r Ysaye 352.8 16.8 [10] 2.00 2.61 3.50 0.29
G del Gesu 1741 46 r Koch anski 351.5 15.0 [10] 2.20 2.99 3.80 0.42
G del Gesu 1741 46 v Vieuxtemps 352.8 15.0 [10] 2.50 2.85 3.50 0.31
G del Gesu 1742 42 r Alard 352.4 14.2 [9] 2.30 2.69 3.10 0.23
G del Gesu 1742 60 n Canon 353 15.0 [9] 2.50 3.11 3.70 0.24
G del Gesu 1742 46 v Lord Wilton 353.5 14.0 [10] 2.30 2.85 3.80 0.30
G del Gesu 1742 34 n n/a n/a n/a Anon 1.80 2.52 2.90 0.25
G del Gesu 1743 48 r Carrodus 351 16.0 [10] 2.20 2.87 3.70 0.30
G del Gesu 1743 46 r Sauret 351 17.4 [101 2.40 2.78 3.50 0.27

32 CASJ Vol. 4, No. 7 (Series II), May 2003

 Jeffrey S. Loen — Reverse Graduation in Fine Cremonese Violins

Maker Date Grads Class Name Length Arch Source Mm Mean Max Sd
G del Gesu 1744 48 v Doyen 350.5 15.5 [10] 2.00 2.58 3.20 0.27
G del Gesu 1744 46 v Ole Bull 351 16.1 [10] 1.60 2.71 3.20 0.34
G del Gesu 1745 46 v Leduc 353.5 17.0 [10] 1.90 2.62 3.70 0.33
G del Gesu 1739-41 40 r n/a n/a n/a Anon 1.90 2.59 3.20 0.24
G del Gesu 174 x 26 r n/a n/a n/a Anon 1.60 2.36 2.70 0.28
G del Gesu & 17xx 30 v n/a n/a n/a Anon 1.80 2.46 3.50 0.38
Jos Guar fil Andrea
Andrea 1677 46 r n/a n/a n/a Anon 1.50 2.05
G Joseph fil. 1705 64 r n/a n/a n/a Anon 1.50 2.29 3.00 0.40
G Joseph fil. 1710 34 r n/a n/a n/a Anon 1.50 2.03 2.70 0.34
G Joseph fil. 1714 32 r n/a n/a n/a Anon 1.60 2.08 3.00 0.40
G Joseph fil. 17xx 86 v n/a n/a 15.5 [15] 1.80 2.87 4.50 0.75
Ruggieri, F 1684 23 v n/a n/a n/a Anon 1.75 2.39 3.10 0.31
A c 1670 51 v n/a n/a n/a Anon 1.70 2.30 2.90 0.31
A 1679 123 v Hellier n/a n/a [14] 2.20 2.87 3.80 0.27
A 1682 230 v n/a n/a n/a Anon 2.00 2.76 3.80 0.24
A 1683 43 r Cipriani P n/a 15.0 [13] 2.00 2.60 3.25 0.24
A 1685 36 r n/a n/a n/a Anon 1.10 2.11 3.20 0.46
A 1693 200 r Harrison 362 15.5 [11] 1.65 2.38 3.30 0.35
A 1694 57 r Muir MacKen 362.5 16.0 [9] 2.30 2.61 3.00 0.18
A 1697 83 r n/a n/a n/a Anon 2.20 2.70 3.90 0.41
A 1699 94 r Castelbarco 357.0 17.0 [12] 1.90 2.35 3.00 0.23
A 1700 44 r n/a n/a n/a Anon 2.00 2.49 3.20 0.25
A 1700 87 v Ward 351.0 15.9 [12] 1.60 2.31 3.10 0.38
A 1701 36 i n/a n/a n/a Anon 1.70 2.15 2.70 0.23
A 1702 27 r n/a n/a n/a Anon 1.60 1.91 2.60 0.25
A 1703 21 i Alsager n/a 16 [16] 1.50 2.15 2.70 0.35
A 1703 190 r n/a n/a n/a Anon 1.90 2.44 3.20 0.21
A 1704 94 r Betts 353.0 16.3 [12] 2.10 2.57 3.40. 0.28
A 1705 75 r n/a n/a n/a Anon 2.00 2.52 4.00 0.45
A 1707 75 r n/a n/a n/a Anon 1.80 2.42 3.60 0.30
A 1708 30 r n/a n/a n/a Anon 1.40 1.92 3.00 0.41
A 1709 121 v n/a n/a n/a Anon 1.70 2.25 3.10 0.36
A 1710 33 i n/a n/a n/a Anon 1.80 2.65 2.90 0.23
A 1711 17 r n/a n/a n/a Anon 1.90 2.41 2.90 0.29
A 1712 48 r n/a n/a n/a Anon 1.80 2.41 2.90 0.24
A 1713 44 v n/a n/a n/a Anon 1.80 2.48 3.00 0.27
A 1714 51 r n/a n/a n/a Anon 2.00 2.66 3.40 0.31
A 1715 22 v Cremonese 355.5 15.4 [17] 2.30 2.76 3.50 0.34
A 1716 69 r Mediceo 356.0 15.4 [18] 2.00 2.74 4.00 0.36
A 1716 48 r Messiah 356.0 15.5 [13] 2.00 2.71 4.00 0.40
A 1717 54 r n/a n/a n/a [19] 1.90 2.30 3.30 0.32
A 1717 45 r n/a n/a n/a Anon 2.00 2.50 3.00 0.24
A 1718 130 r n/a 352 16.6 Anon 1.50 2.32 3.20 0.37
A 1719 78 r n/a n/a n/a Anon 1.70 2.36 3.50 0.28
A 1719 64 r n/a n/a n/a Anon 1.50 2.23 3.00 0.31
A 1724 42 v n/a n/a n/a Anon 1.80 2.39 2.90 0.22
A 1726 35 r n/a n/a n/a Anon 1.70 2.23 3.00 0.43
A 1727 20 v n/a n/a n/a Anon 1.70 2.50 4.40 0.74
A 1727 40 r n/a n/a n/a Anon 1.80 2.41 3.10 0.36
A 1727 24 v n/a n/a n/a Anon 1.10 1.84 2.50 0.41
A 1728 42 i Milanollo 352 16.7 [9] 1.90 2.42 3.20 0.35
A 1731 19 i n/a n/a n/a Anon 1.70 2.43 3.20 0.41
A c 1733 177 r Pushilov n/a 18.0 [20] 1.70 2.67 4.50 0.50
A 1734 26 i n/a n/a n/a Anon 2.00 2.42 3.00 0.18
A 1736 185 v Oistrakh 358.0 14.0 [20] 2.00 2.60 3.60 0.31
Stradivari, F 1719 139 r n/a 332 13.0 [11] 1.85 2.72

CASJ Vol. 4, N0.7 (Series II), May 2003 33

Jeffrey S. Loen — Reverse Graduation in Fine Cremonese Violins

Figure 5. Thickness graduation maps of a fine concert violin by Antonio Stradivari, 1707 (instrument name withheld by request of
contributor). Top plate shows reverse graduation with a prominent semicircular minimum thickness contour at bridge position, and
back plate shows normal graduation. Measured with caliper. Measurements are shown in millimeters (contour interval 0.25 mm).

are known for exceptional sound, although the reverse graduated free plate tuning were developed based on what produced "good
ones are much more abundantand the high performance might be sounding instruments", although this was done empirically and
explained by lighter weight, reduced mass loading, and increased with little regard for the nature of successful Cremonese violins.
flexibility (although this needs verification). However, it appears
that there are important exceptions, and no single explanation exists In contrast, results of this study will be no surprise to Schleske [25],
for the structure and performance of the finest violins. As stated by who constructs tonal copies based on fine historic instruments,
the Hills [24], "many points have to be considered: thicknesses, including specimens by Stradivari and Montagnana. He clearly
model, dimensions, wood, and, last but not least, varnish all play recognizes that some violin top plates, such as the "Kreisler" del
their part..." Gesu (Fig. 3), become thinner towards the middle. Modal analysis
allows Schleske to make "resonance sculptures" of fine instruments
Relation to Plate Tuning and Modal Analysis in which differences in materials and varnish are adjusted to obtain
Results about reverse graduation of Cremonese violins are difficult the desired acoustical signatures.
to reconcile with rules developed for free plate tuning [1,2,4].
Thicknesses for most historic top plates are apparently too thin to Possibility of Regraduation and Thinning
attain principal modes at frequencies that have been defined as It is possible that many examples of reverse graduation were caused
favorable. However, data are almost completely lacking on the by regraduation and thinning, years after the violins left their makers'
subject of principal modes of historic instrument plates. Rules for benches, although we have little evidence besides opinions and

34 CASJ Vol. 4, No. 7 (Series II), May 2003

 Jeffrey S. Loen — Reverse Graduation in Fine Cremonese Violins

Figure 6. Thickness contour maps of the 'Stern' violin (Guarneri del Gesu, 1738 [10]), one of the career violins of American soloist Isaac
Stern. Top plate shows reverse graduation, and back plate shows normal graduation. Measured with magnetic thickness gauge. Black dots
indicate locations of measurement points. Contour interval 0.25 mm. Viewed from outside.

Body 352 mm
Upper bouts 164.5 mm Upper bout 163mm
yiddle bouts 119.5 mm Middle bout 119.25 mm
Lower bouts 204 mm Lower bout 202 mm
Arching 14.7 mm Arching 13.0 mm
Rib thickness 1.1-1.2mm Rib thickness 1.1-1.2 mm

anecdotes to support this idea. The idea of regraduation is deeply considerably, making a thickness pattern that works well for a
engrained in the violin literature, for example Weisshaar and particular violin work poorly for an instrument with different
Shipman [26] state that regraduation "must have been done on a arching. Arching and plate thickness both influence plate stiffness.
regular basis, for today we seldom find an instrument which has There is considerable variation in arching heights of historic
not been somewhat altered." They conclude that "detrimental instruments (Table 1), for example, 26 Guarneri del Gesu violin top
changes in graduations were made by irresponsible makers, dealers, plates range from 12.4 to 17.4 mm (aye, 15.4mm), and ten Stradivari
and dilettantes who hoped to improve the sound of instruments by violin top plates range from 14.0 to 18.0mm (aye. 15.9mm). However,
thinning the soundplates." the type of top graduation shows no correlation to arching height.
Also, the argument that archings are significantly different does not
Such reports are countered by the fact that several reverse-graduated recognize Playfair's investigations of arching shapes [27, and this
instruments (including Stradivari's 'Muir-Mackenzie', 'Messiah,' issue], which conclude that many Cremonese instruments follow
and del Gesu's "Soil" and "Alard") are some of the best-preserved the same family of mathematical curves (Note: we were able to
examples of classic Cremonese violins. However, regardless of verify Playfair's findings in our own investigations of Cremonese
who did the graduating or when it was done, reverse graduated violins).
violins dominate the modern concert stage.
Normal Graduation in Post-Golden Age Violins
Influence of Arching The violin graduation database suggests that normal graduation
Another consideration is that the arching shapes might vary becomes the rule, rather than the exception, in instruments made

CASJ Vol. 4, N0.7 (Series II), May 2003 35

Jeffrey S. Loen — Reverse Graduation in Fine Cremonese Violins

Figure 7. Thickness graduation maps of a fine concert violin by Carlo Bergonzi, 1732, the career violin of an American soloist
(instrument name withheld by request). Top plate shows reverse graduation with a strong longitudinal trough, and back plate shows
normal graduation. Measured with magnetic thickness gauge. Measurements are shown in millimeters (contour interval 0.25 mm).

Upper bouts 161.5 mm

Middle bouts 103 mm
Lower bouts 197 mm Upper bouts 162 mm
Arching 14.2 mm Middle bouts106 mm
Lower bouts 198 mm
spacing 38.5 mm
Rib thickness <1.0-1 1mm Rib thickness <I.Q-l.lmm
Rib height 29.5-31.1 mm
Rib height 29.5-31.1 mm

Table 2. Summary of top plate graduation classifications.

Number Insufficient
Maker Name Plates Normal Reverse Undefined Dat

Andrea Amati 3 1 0 1 1
Brothers Amati 3 1 1 1 0
Nicolo Amati 8 4 2 1 1
Carlo Bergonzi 3 0 2 0 1
Guarneri del Gesu 38 3 27 8 0
Guarneri del Gesu & 1 0 0 1 0
Joseph fikus Andrea
Joseph Guarneri fikus 4 0 3 1 0
Andrea
Francesco Ruggieri 1 0 0 1 0
Antonio Stradivari 43 0 26 11 6
Francesco Stradivari 1 0 1 0 0

Totals 105 9 62 24
Percent 100 9 60

36 CASJ Vol. 4, No. 7 (Series II), May 2003

 Jeffrey S. Loen — Reverse Graduation in Fine Cremonese Violins

Figure 8. Thickness graduation maps of the 'Harrison' violin, Antonio Stradivari, 1693 [11]. Top plate shows reverse graduation, and
back plate shows normal graduation. Measured with magnetic thickness gauge. Measurements are shown in millimeters (contour
interval 0.25 mm). Viewed from outside.

Body 361.5 mm
Upper bouts 161.7 mm
Middle bouts 109 mm
Upper bouts 161.5 mm
Lower bouts 202.3 mm
spacing 40 mm Middle bouts 108.4 mm
Lower bouts 202.5 mm
Arching 15.5 mm
Arching 15.1 mm
Rib height 30.4-31.6 mm
Two-piece quarter maple
Rib thickness <1.0 mm
Weight 386

after the Golden Age ended in about 1750. Good examples are the difficult to argue with both acoustical success and high instrument
Gagliano family (Naples, Italy), other Italian makers (e.g. C.F. values. Perhaps we should reevaluate the idea that thin top plates
Landolfi, G.B. Gabbrielli, L. Storioni, A. Gragnani, G.B. Ceruti, N. are predestined to be acoustically and structurally unsuccessful.
Bergonzi, P. Pallotta, P. Mantegazza, G. Rocca, G. Fiorini, S.
Scarampella, R. Antoniazzi, C.C. Bruno, P. Badalassi, G. Ornati, The most surprising finding of this study is that little support could
and F. Garimberti; Fig. 9), and various German makers (e.g. L. be found for the concept of "normal" graduation, as recommended
Maussiel, J.G. Meisel, Georg Klotz, J.G. Voigt, J.G. Hamm, and by violin making schools, books, and articles, and exemplified by
Mathais Hornsteiner). del Gesu's "Canon". Only nine (9%) of 105 Golden Age plates are
classified as normally graduated, whereas 62 (60%) are classified as
CONCLUSION reverse graduated. One could argue that many of the reverse
The purpose of this article is to shed some light on the shady topic graduated plates are the result of regraduation, which may be true
of the modern configurations of Cremonese violin plates. Spruce but is acoustically irrelevant. If we want to learn something about
soundboards of many of the world's best Cremonese concert violins instrument acoustics, we should focus on current configurations of
from the Golden Age are shaped opposite the way that modern extant Golden Age concert violins, which have been carefully chosen
violin makers are taught to carve their tops. Rather than swelling to for their remarkable abilities to sing in the hands of the world's best
greater thickness in the center between the ff-holes, the top plate players. These violins are quite unlike the Canon, and appear to
pinches down to minimum thickness in this area. Despite the fact have different characteristics of weight, mass loading, flexibility,
that these instruments break many of our modern-day rules, it is and plate tuning frequencies.

CASJ Vol. 4, N0.7 (Series II), May 2003 37

Jeffrey S. Loen — Reverse Graduation in Fine Cremonese Violins

Figure 9. Thickness graduation map of a violin by Ferdinando Garimberti, Milan, 1967. Top and back plates show normal graduation
patterns.Measured with magnetic thickness gauge. Measurements are shown in millimeters (contour interval 0.25 mm). Viewed from
outside.

Body 353.5 mm
Body 355.0 mm
Upper bouts 169.0 mm
Upper bouts 169.0 mm
Middle bouts 115.5 mm
Lower bouts 209.0 mm Middle bouts 114.5 mm
Lower bouts 209.0 mm
Arching 15.9 mm
Rib thickness 1.0-1.1 mm Arching 15.6 mm
Two-piece quarter maple

Finally, I should state that I favor conservatism with regard to making [2] Brown, E.S., and Campbell, E.C., 1990, Chimneys violin
new instruments, and I do not recommend that modern makers maker's workshop, Book II: Edward C. Campbell, Boiling
reverse graduate their plates without experimental verification of Springs, PA, 48 p.
the acoustical virtues of thin plates. [3] Strobel, H.A., 1997, Violin making, Step by step (second
edition): Henry A. Strobel, Aumsville, OR, 78 p.
ACKNOWLEDGMENTS [4] Johnson, C, and Courtnall, R, 2000, The art of violin making:
I thank the Smithsonian Institution, National Music Museum, U.S. Robert Hale, London, 253 p.
Library of Congress, and Metropolitan Museum of Artfor access to files [5] Loen, J.S., 2001, Thickness graduation mapping: Methods &
and instrument collections. Data contributors are thanked for their goals: Catgut Acoust. Soc. Journal, vol. 4, no. 4 (Series II), p. 5-
generosity. 6.
[6] Loen, J. S., 2001, Thickness graduation systems ofviolin family
REFERENCES instruments: Preliminary statistics and conclusions: Catgut
[1] Hutchins, CM., 1983, Plate toning for the violin maker: Catgut Acoust. Soc. Journal, vol. 4, no. 4 (Series II), p. 42-44.
Acoust. Society Newsletter no. 39, p. 25-32.

38 CASJ Vol. 4, No. 7 (Series II), May 2003

 Jeffrey S. Loen — Reverse Graduation in Fine Cremonese Violins

[7] Loen, J.S., and King, A.T., 2002, Thick and thin: The Strad, [20] Zimin, P., 1926, A uniform system for the measurement of the
December 2002, vol. 113, no. 1352,p. 1354-1359. thicknesses of the plates of bowed stringed instruments (see
[8] S.F., 1979, The "secrets" of Stradivari: Libreria del Vigdorchik, 1982).
Convegno, 285 p. [21] Chaudiere, F., 2002, Silent witness: The Strad, published by
[9] The Strad poster series, Orpheus Publications, London. Orpheus Publications, London, vol. 113, no. 1346, June 2002,
[10] Biddulph, P., Carlo C, Dilworth, J., Hargrave, R, Klein, P., p. 618-625.
Pollens, S., Rosengard, D., and Wen, E., 1998, Giuseppe [22] Giordano, A., 1995, The Cannon and typical features of
Guarneri del Gesu: Peter Biddulph, London, 339 p. Guarneri's instruments: in Paganini's violin: Dynamic Sri,
[11] National Music Museum, Vermillion, S.D. Italy, 105 p.
[12] U.S. Library of Congress, Washington, D.C. [23] Vigdorchik, 1., 1982, The acoustical systems of violins of
[13] Ashmolean Museum, Oxford, England. Stradivarius and other Cremona makers: Bradenton, Florida,
[14] Smithsonian Institution, Washington, D.C. 169p.
[15] Vitachek, E., 1926, Thicknesses and their distribution on the [24] Hill, W.H., Hill, A.F., and Hill,A.E., 1963, Antonio Stradivari:
top and back of instruments of the ancient craftsmen: Russian His life & work (1644-1737), New York, Dover Publications,
Commission on Musical Instrument Making: State Institute Inc., 31 5 p.
of Musical Science (GIMN). Reprinted in 1952 as "Essays on [25] Schleske, M., 2002, Emprical tools in violin making (Part I):
the history of bow instrument making". Analysis of design, material, varnish, and normal modes:
[1 6] Goodkind, H., 1972, Violin iconography ofAntonio Stradivari, Catgut Acoust. Soc. Journal, vol. 4, no. 5, p. 50-63.
1644-1737: New York, 780 p. [26] Weisshaar, H, and Shipman, M., 1988, Violin restoration: A
[17] "Strumenti di Antonio Stradivari" (1991), Ente Triennale manual for violin makers: Weisshaar-Shipman, Los Angeles,
Interazionale Degli Strumenti ad Arco. CA, 261 p.
[18] Carlson, B. (cd.), 2001, Antonio Stradivari, Violino 1716, [27] Playfair, Q., 1999, Cremona's forgotten curve: The Strad,
Mediceo: Cremonabooks S.r.L, Italy. Orpheus publications, London, vol. 110, no. 1315, Nov. 1999,
[19] Metropolitan Museum of Art, New York. p. 1194-1199.

CASJ Vol. 4, N0.7 (Series II), May 2003 39

 Documentation of a 1604 Violin by Hieronymus Amati,
Cremona, Italy

Jeffrey S. Loen and A. Thomas King

The Witten-Rawlins collection of northern Italian string Collection in London, England. The violin appears to have been
instruments at the National Music Museum in Vermillion, South actively played for part of its history, based on erosion at the bridge
Dakota includes a disassembled violin made in Cremona, Italy by location.
Hieronymus Amati in 1604. The violin (no. 3423) is small size
(body length 341 mm) and includes top, back, and rib garland (neck The top plate (Fig. 1) is made of two pieces of spruce that appear not
is absent). Examination of these parts is instructive regarding wood to match. It weighs 46 g, withoutbass bar, and the maximumarching
preference, design, dimensions, weight, and free plate tap-tones for height is about 20.0 mm (18-19 mm if adjusted for plate warpage).
a 400-year old Cremonese violin. The label reads "Hieronymus Basic dimensions are as follows: body, 340.5 mm; mensur, 190 mm;
Amatus Cremonen Andreae fil. F. 1604". No information is upper bouts, 156 mm; middle bouts, 103.3 mm; and lower bouts,
available about the acoustical quality of violin, or the names of 188.5 mm. The ff-holes are 73-74 mm long and the upper holes are
previous owners, besides having once been in the E.M.W. Paul spaced 34.8 mm apart. All edges are doubled and numerous cracks

Figure 1. Top plate of violin by Hieronymus Amati, Cremona, Italy, 1604. Left shows photo of outside of free plate, right shows thickness
graduation map (thickness in millimeters; contour interVal 0.25 mm)

40 CASJ Vol. 4, No. 7 (Series II), May 2003

 Jeffrey S. Loen and A. Thomas King — Documentation of a 1604 Violin by Hieronymus Amati, Cremona, Italy

are present. Purfling is about 1.55 mm wide, and is situated 3.5 mm cycloid templates that the arching conforms to this family of curves.
from the edges. We measured tap tones (although they may not be The thickness pattern is concentric, with a maximum value of 4.3
meaningful because of the cracks): Mode I, D-Eb (73-78 Hz); Mode mm located near the 50% point (171 mm from bottom edge) of the
11, C-C# (131-139 Hz). The plate is too fragile to obtainmode V, the back plate, coinciding with a small pinhole. The upper and lower
ring mode. The thickness graduation pattern of the top plate includes bouts have large thin (1.05-1.7 mm) areas, with a minimumvalue of
large areas that are extremely thin (1.0-2.1 mm). Minimum values 1.05mm in the upper right quadrant. Edges range from 3.25-3.5 mm
of 1.0 mm occur between the ff-holes, and slightly below the middle in the upper and lower bouts, to 4.0-4.1 mm in the middlebouts, and
bouts. Thicker areas (>2.5 mm) occur between the middle bout overhang is 2.5 mm
edges and the ff-holes, and at the upper block location. The
maximum thickness of 3.3 mm is at the upper block location. We The rib garland weighs 31 g. Rib heights are 25.6-27.7 mm and rib
have no indication of whether these are original or modified thickness is 0.9-1.1 mm. The lower rib is one piece. The lower
graduations. block, made of willow, is 52.2 mm wide and 12.5 mm deep. The
upper block is 56 mm wide and 15 mm deep. Corner blocks are
The back plate (Fig. 2) is made of two pieces of maple cut almost on also made of willow, and their lengths are 19.75 mm (upper) and 21
quarter. The two-piece back is joined such that the flames have a mm (lower). Willow linings are 5 mm high and about 1.6mm wide.
consistent slant upward from the treble side to the bass, resembling The linings on the middle bouts are not inserted into the blocks, but
a one piece back in appearance. The plate weighs 82 gm and the are merely butted up to the blocks.
maximum arching height is 15.5 mm. Basic dimensions are as
follows: body, 341.0 mm; upper bouts, 156 mm; middle bouts, 106 ACKNOWLEDGMENT
mm; and lower bouts, 195 mm. Modes I, 11, and V are estimated at We thank Dr. Andre Earson, Director of the NationalMusic Museum,
112, 133, and 350 Hz. We confirmed to our satisfaction using curtate for access to the collection and permission to present our results in CAS].

Figure 2. Back plate of violin by Hieronymus Amati, Italy, 1604. Left shows photo of outside of free plate, right shows
thickness graduation map (thickness in millimeters; contour interval 0.25 mm)

CASJ Vol. 4, N0.7 (Series II), May 2003 41

 Early Tuning Pegs and Tailpieces

Eric Meyer
6837 N.E. Alameda Drive, Portland, OR 97213 USA
[email protected]

ABSTRACT SOURCES OF DATA

This study is based on examples of early pegs and tailpieces in
museums. Styles of wooden tuning pegs have "evolved" from Ashmolean Museum, England
myriad shapes and designs during the 18th Century to the few classic Boyden [2] indicates that fittings are not original for most, if not all,
styles we are familiar with today. Early makers working in violins and violas in the Hill collection. Almost all are replacements
individual shops were separated by distance and lack of made at W.E. Hill and Sons. The work is often beautiful and done
communication, whereas in modern times most peg production is in the spirit of the instrument, but one wonders what happened to
automated and mass distributed. Tailpieces were originally straight the original fittings. I asked the curator Dr. Whitely if any
wedges of wood with simple holes. Strings were tied lute fashion, information exists on this or any old pegs in the stacks. He was
using the loop of the string to anchoritself. They have been improved sympathetic to my quest but had no additional information other
by the addition of a nut, the use of black ebony rather than dyed than Boyden [2]. I include a photo of possibly very old pegs mounted
maple (which stained the chin), the smoothing and rounding of the on a Gasparo da Salo viola (Fig. 1). In addition, an Andrea Amati
lower end for greater comfort, and the invention of an underside viola of 1574 is mounted with an old, possibly original, tailpiece
attachment system for the tailgut. It is a worthwhile endeavor to (Fig, 2).
examine this evolution as it tells us of the attempts of the early
makers to effect changes in acoustics by modifying tailpieces. NationalMusic Museum, Vermillion, So. Dak. USA
Assorted early pegs, tailpieces, and other fittings and parts are kept
INTRODUCTION in boxes in a storage room. Few of these parts have any sort of
Practically every aspect of violin family instruments has been
investigated in detail except for fittings. Since fittings wear out, Figure 1. View of back of pegbox and tuning pegs on viola by
perhaps some have thought that they were expendable, and were Gasparo da Salo, Brescia, Italy, before 1609 (Ashmolean Museum,
therefore unworthy of attention. However, upon close examination, Oxford). This set appears to be made of giuggiolo (Ziziphus. jujuba
the early fittings, particularly pegs and tailpieces, tell fascinating [3]), which was a common peg wood employed by the early Italian
stories of the same sort of careful artisanship, practical functionality, makers, but could be plum wood.
elegant design, and gradual improvement that is shown by the
instruments themselves. In a scenario reminiscent of the Cinderella
story, would we have the present controversy over authenticity of
Strad's Messiah if Vuillaume had simply saved the pegs? As regards
tailpieces, there is little doubt that they also serve an important
acoustical function [I], although this aspect needs more research.

Regrettably, considerable historical information has been lost

because original fittings were not preserved or properly catalogued.
During the past several years I have searched for collections of
early fittings in the US and Europe. Assortments of old fittings are
often in the possession of museums but are not on display. In this
paper, I summarize theresults of my detectivework on the scattered
examples of early violin pegs and tailpieces.

42 CASJ Vol. 4, No. 7 (Series II), May 2003

 Eric Meyer — Early Tuning Pegs and Tailpieces

identification. At least one instrument on display, the Brothers

Amati violino piccolo, 1613, is in original condition and appears to Figure 4. Original (?) maple tailpiece on violino piccolo by Brothers
be fitted with very old pegs (Fig. 3) and a tailpiece, probably original Amati, 1613 (National Music Museum; A.T. King photo).
(Fig. 4). The collection includes a variety of pegs (Fig. 5). Early
ebony tailpieces are shown in Fig. 6. Access to the collection was
courtesy of Director Dr. Andre Larson

Stradivari Museum, Italy

The Stradivari Museum includes a set of eleven unfinished pegs
(Fig. 7) that supposedly came from Stradivari's shop (although they
are not listed in Sacconi's inventory of the collection [4]). One of
these pegs (Fig. 7A, second from left), with ivory collar and nib, is
depicted in the Hill's book on Stradivari [5, Fig. 50]. My
understanding is that these pegs came from the Cherani collection
and can be traced back to Count Cozio di Salabue. These pegs are
mostly unfinished, including three just as they came off the lathe
before the sides were cut off and flattened. Dimples in the top of the
head are the marks left by the tailstock of the lathe. I believe that
bone pips were added as a decoration to disguise these holes.
Figure 5. Variety of peg designs at National Music Museum.
Perhaps the separate white collar that we see in the lighter colored
boxwood peg (Fig. 7A, second from left) was added to compliment A) Early pegs with large nibs or pips; center peg is probably a
cello
this accent. Most appear to be made from giuggiolo, except for two peg,
peg on right probably for viola. B) Violin pegs (note steep taper
on center peg). Wood type is probably boxwood in most cases,
of boxwood (Fig. 7A, left two pegs). One of these boxwood pegs
(Fig. 7A, leftmost), which may have been intended for use on a
although some may be fruitwood. C) Violin pegs, center peg has
lathe dimple in center of pip, appears to be rosewood. D) Violin
pegs, possibly hornbeam. E) Finally crafted violin pegs, possibly
Figure 2. Tailpiece on the Figure 3. Pegbox of violino pernambuco.
Charles IX viola by Andrea piccolo by Brothers Amati, 1613,
Amati (1574) at the Ashmolean showing early if not original
museum, Oxford. Brass plate boxwood tuning pegs (National
anchors metal wire anchoring Music Museum; A.T. King
system, most likely added after photo).
1750. Maple with decorative
purfling.

CASJ Vol. 4, N0.7 (Series II), May 2003 43

Eric Meyer — Early Tuning Pegs and Tailpieces

Figure 6. Ebony tailpieces at the National Music Museum. Left Figure 7. Pegs in different stages of completion, reputed to be
tailpiece is early example of the "rooftop" design. Right tailpiece from Stradivarius shop (Stradivari Museum, Cremona). A) View of
has modern keyhole-type slots but lacks a modern-style attachment set of eleven pegs. Right three pegs are round, as if directly from
system and saddle, indicating that development of tailpieces was lathe. Two on left have been filed to finished head thickness and the
not linear. Note variety of attachment mechanisms. The first two central group has been sawed flat but not finished to final thickness.
are earlier gut loop types, but the first is sunken, to avoid abrasion. Note rectangular, tapered ends remaining on pegs 4-6 from right.
Both theroof top style and the sunken loopholes are evolved features This implies the use of a collet with a rectangular tapered hole for
but the squared bottom is not. torque at the headstock; B) Oblique view from a different angle.
C) Close-up of five pegs in center of fig. 7A. Note saw marks on
faces of left three pegs. Wood appears to be Zisiphus (guigolo) [3].

A.

pardessus, is partially stained black. This stain was likely

accomplished by using a solution of logwood (Haematoxylon
Campeachianum) [9] and verdigris (copper acetate)[10]. I was able
to photograph these pegs courtesy of museum director Prof. Andrea
Mosconi.

Karl Schreinzer Germanisches Nationalmuseum,

Nuremberg, Germany
Karl Schreinzer (1884—1969), was professor and musician in Vienna,
Austria. He collectedold fittings and other parts that were discarded
when instruments were converted to modern set-up. The collection
was described in 1961 based on a talk for the Galpin Society [6]. In
1967 it was presented to the Germanisches Nationalmuseum. When
I was there the collection appeared different from photos in the
Galpin Soc. publication. Some of the pieces displayed in the original
photos seemed to be missing and, at the time I visited, the trays
were unsorted. I was able to photograph part of the collection in
1995, courtesy of curator Klaus Martius. The present curator Dr.
Frank Bar recently informed me that the collection is still not on
display but is available for research purposes.
to allow for strings and tail gut, as well as more elaborate systems
This collection is significant because early examples of pegs and that developed. There are transitional tailpieces that were attached
tailpieces are labeled. However, the significance of the labels is by metal wire emerging through the top of the piece and soldered to
open to question; names recorded most likely indicate the makers an ornamental plate, as seen in the Amati viola in the Ashmolean
of the instruments from which the fittings were taken. Here we can
(Fig. 2). One tailpiece shows the chin wear that removed black
see various early tailpieces, simple slabs of maple with small holes stain, which encouraged the use of ebony as a tailpiece material.

44 CASJ Vol. 4, No. 7 (Series II), May 2003

 Eric Meyer — Early Tuning Pegs and Tailpieces

Figure 8. Violin pegs in the Schreinzer collection. A) Violin pegs. Figure 9. Tailpieces from the Schreinzer collection, Germanische
Attached labels from left say "Joh. Jos. Stadlmann, 1750(?)", museum (note different scales). A) Photo ofinside of simple early
"englishe", and "FridoliKama". Stadlmann (1720-1781) was a violin tailpiece. Writing appears to say "Beethoven yon". Metal attachment
maker in Vienna [7]. B) Labels from left say "Salzbg", Joh. Jos. connects to top plate as seen in Figure 2 B) Early tailpieces with
Stadlmann, 1767", "aus Mailand yon", and "(unintelligible) Mailand". scalloped upper end (note black dye worn off left tailpiece). First
C) Label on left says "Joh. Christ. Leidolff 1758". Johann Christoph tailpiece uses metal the same way gut would be tied. C) Variety of
Leidolff (1690-1758) was a violin maker in Vienna [7]. Some of the early tailpieces, some are ebony and some are dyed (notice wear on
pegs in the collection were identified with labels that only signified third).
names of cities or countries, such as "englishe", "beljique" or
"Salzbg' (most probably Salzburg), highlighting the regional nature
of early peg designs. Woods appear to be mostly boxwood. First A. B.
peg in B is ebony; other darker pegs may be plum or other
fruitwoods. Fourth peg in A. could be pernambuco.

A.

B.

C.

CASJ Vol. 4, N0.7 (Series II), May 2003 45

Eric Meyer — Early Tuning Pegs and Tailpieces

DISCUSSION to the black stains appearing on their throats. In addition, a harder,

finer grain wood such as ebony absorbs no sweat, which made gut
Tuning Peg Woods strings last longer [B]. Purfling ornaments and inlays were common
Salabue remarks that early makers used pearwood, which along before 1800, although Salabue notes that they were no longer made
with a steep taper tended not to crack the pegbox or abrade the because "ornaments easily become unstuck and spoil the tone. "[B]
holes [B]. Salabue also mentions that the "very best wood" is Much of the choice in woods could easily be attributed to economics,
"Zanzuino" (a species of dogwood, Sanguinea Cornus languinea) and the availability of exotic species, tied to the cost of importing.
[B]. Andrew Dipper (oral commun., 2003) believes that Salabue was One sees few rosewood tailpieces in the early collections. Boxwood
probably confusing this dogwood with Zissiphus or Giuggiolo. This was certainly used and probably pear, but it is difficult to identify
small fruit tree is in the buckthorn family and is referred to as the type of wood beneath stained surfaces.
Chinese Date Plum or Jujube. It was brought to Europe much earlier
and grew in Northern Italy. Turkish boxwood was also available to Tailpiece Designs and Trends
the Baroque makers. The Mantegazza brothers made pegs of ebony Three main changes appeared in the 18 century. These are:
th

"but of a sort especially black and therefore softer" [B]. This modification of simple string holes into slotted "keyhole" shapes,
Mauritius ebony supply was quickly exhausted. I have seen early the addition of a raised ridge to support strings after emerging from
pegs made from other woods, notably rosewood and hornbeam, the holes at the top of the tailpiece, and the evolution of the system
but I have been unable to confirm their species. The set of pegs at that suspended tailpiece from end button. Originally tailpieces had
the National Music Museum that first piqued my historical simple holes to anchor the strings (Figs. 2, 4, 9A). Changes in string
imagination (Fig. SE) appeared to me to be made from pernambuco. technology and tension required slots for the string ends. Knots
strong enough to hold the tension made changing broken strings
Tuning Peg Designs and Trends much easier, especially during a performance. The musician could
Unlike the several classic peg head shapes that are mostly mass- drop theknotted end through the hole in the top of the tailpiece and
produced today, the early pegs covered the gamut of shapes. Overall cinch it into the slot—a much quicker and easier procedure than
the heads appear harder edged, with more pointed corners and tying them.
steeper peg shaft tapers than modern, ones. The plethora of designs
is mirrored in the many shapes and types of instruments developed Raised bars of wood or bone were added for the strings to rest on as
and played. Accordingly, it is the obscure and less successful they emerged from the tailpiece. I have seen tailpieces with this
instruments, such as the viols and dance-master's kits, which seem ridge or "saddle" carved from the tailpiece material itself, but mostly
to afford the best view of what original pegs lookedlike. Since these this change was accomplished by a separate piece of inlayed wood.
instruments were in vogue for a shorter period, their pegs were
replaced less often. The two simple holes that held the loop of gut at the other end
originally allowed the gut to sit above the surface of the wood. This
As evolution ofviolins progressed pegs turners tended to use harder must have been both uncomfortable and dangerous in the days
woods, which became more available and less expensiveas advances before chinrest development. Acidic sweat eventually weakens
in transportation made the world smaller. I previously postulated natural gut. Sudden breakage can harm both the player and the
that pips were originally used to hide the dimple caused by the instrument. A depression was carved into the top of the tailpiece
lathe tailstock. Another explanation was suggested to me by Andrew that held this loop below the surface level in an intermediateremedy.
Dipper, restoration expert and instrument historian. Most early In Germany, at least, for a time in the 19th Century this connection
instruments were tuned using a ratchet-like tool called a was bridged by rigid metal wire that was fixed to a plate inlayed in
tournegauche, which fit over the peg head. The pip may have kept the top of the tailpiece. Adjusting the distance from the bridge to
the peg centered in this device to facilitate tuning. As time the tailpiece is difficult with this system but sound quality may
progressed, heads became less decorative and more comfortable to have been improved. Local cello makers in Nuremberg even sunk
the fingers. A notable exception is the übiquitous heart shaped peg metal, right-angled flanges into the end block of the instruments on
th
that was made famous by W.E. Hill and Sons at the end of the 19 to which the tailpieces were riveted. Eventually the modern system
Century. The shaft taper decreasedfrom 1/20 to 1/30. This may be of attachment developedrequiring a thicker piece of wood to allow
explained by the standardization of machine tapers and chucks, or for a niche carved into the underside of the tailpiece, which held the
the use of harder woods, which need less taper to impede their knots in the end of the tail gut.
travel through the two holes in the peg box.
Early tombstone shaped tailpieces, plain or decorated (Figure 2, 4
Tailpiece Woods and 9A), became streamlined with an accentuated waist and more
Early tailpieces were commonly made of maple and other light graceful curves. The top profile became rounder to mirror the shape
colored woods, sometimes dyed black [9]. However, ebony came of the top of the bridge. One early tailpiece in the National Music
into use by around 1800, at least partly because musicians objected museum collection (Fig. 6, left) has the rooftop style generally

46 CASJ Vol. 4, No. 7 (Series II), May 2003

 Eric Meyer — Early Tuning Pegs and Tailpieces

associated with the Hill tailpiece of the early 20th Century, yet still REFERENCES
has the square bottom and simple tail gut holes of the early period. [1] Schuback, P., 1999, and Meyer, E., The relationship of fittings
Many of the Germanic tailpieces that I saw in Nuremberg were and tone: Violin Society of America Journal, p. 109-146
scalloped or "shield" shaped above the nut. It appeared to me that [2] Boyden, D.D., 1969, Catalogue of the Hill collection of musical
the makers were relying on the shape to decorate the work, as instruments in the Ashmolean Museum, Oxford: London,
opposed to the fanciful puffing designs that we see on earlier Oxford University Press, 54 p.
examples. It is difficult to posit a time line for these changes due to [3] http://chemserv.o.pph.univie.ac.at/ska/ziplant.htm
the lack of records. As with pegs, tailpiece shapes evolved from [4] Sacconi, S.F., 1979, The 'secrets' of Stradivari: Cremona,
many fanciful and different designs to the few classical shapes we Libreria del Convegno, 284 p.
find commercially available today. Gut has been replaced by [5] Hill, W.H., Hill, A.F., and Hill, A.E., 1963, Antonio Stradivari:
threaded nylon, although some modern tailpiece makers use metal His life & work (1644-1737), New York, Dover Publications,
wire as the tailpiece connection. Some modern makers have reverted Inc., 315 p.
to gut for acoustical reasons. [6] Sherrington, U., and Oldham, G.(eds), 1961, Music, libraries
and instruments (Hinrichsen's Eleventh Music Book: papers
CONCLUSIONS read at the Joint Congress, Cambridge, 1959, of the lAML and
Little is recorded about the history of this aesthetic and functional the Galpin Society. London, Bach House.
part of the anatomy of the violin family. Perhaps the oversight has [7] Henley, W., 1973, Universal dictionary of violin and bow
to do with the perception of fittings as "used tires" that we wear out makers: Amati Publishing Kent, 1268 p.
and replace. For some time they have been made in factories and [8] Dipper, A., and Woodrow, D., 1987, Count Ignazio Alessandro
sold through catalogues, which has only increased this lack ofrespect. Cozio di Salabue: Observations on the construction of stringed
It is hard to know who made the early fittings, whether they were a instruments and their adjustment 1804, 1805, 1809, 1810, 1816:
specialized group of artisans or the luthiers and apprentices. As Redwood Burn Limited, Trowbridge, Wiltshire, 85 p.
with the rest of our endeavors to push musical instruments towards [9] http: //www.botanical.com/botanical/mgmh/1/
that last percentage of quality in sound and function, we can learn logwoo39.html
something from these artisans of the past. [10] http://www.publicbookshelf.com/public_html/
The_Household_Cyclopedia_of_General_lnformation/
wooddyes_cbh.html

CASJ Vol. 4, N0.7 (Series II), May 2003 47

 Curtate Cycloid Arching in Golden Age Cremonese Violin
Family Instruments

Quentin Play fair

99 Ferrier Aye
Toronto, Ontario Canada M4K 3H6
[email protected]

ABSTRACT York in 1994 [3]. I rejected personal notes and outlines because I
The possibility that arching patterns of Cremonese violins, violas, wanted my arguments based on commonly accessible materials. A
and cellos conform to curtate cycloid curves is considered in light common style emerged for cross-archings [4], although I found no
of the familiarity of Renaissance and Baroque mathematicians and solution for the longitudinal arch. Clearly, the exact way the
artisans with the curve, close matches obtainable with old arching system was used varied from maker to maker. It should be
instruments, and the simplicity and elegance that seem to be the stressed that what I noticed was far more of an observation than a
hallmark of Cremonese work. Ideal curate cycloid curves produced solution— although it seems to work well to my eye, it is open to
by a computer program are compared with cross-arching profiles evaluation by a wider public. Above all, I would like to avoid any
of seven important Cremonese instruments. A visual comparison link with ancient "secrets". Mysterious parchments and malodorous
suggests in most cases a favorable correlation, although it should be compounds gratify modern appetites for quick solutions to complex
acknowledged that correlation does not necessarily indicate cause problems. However, I do not think that they help us reconstruct the
and effect, and there is no written evidence proving that Cremonese working methods of the violin makers of 17th century Cremona.
makers used curtate cycloid curves.
GEOMETRY OF CYCLOID CURVES
INTRODUCTION There are three kinds of cycloid curves, if we confine ourselves to
If we approach the great Cremonese violin makers with a view to the path traced by a point along the radius of a circle rolling along
investigating their similarities, rather than their differences, it a flat surface: the pure cycloid, the curtate cycloid, and the prolate
becomes increasingly clear that a common system of design is cycloid. These can all be expressed as the locus of a point P, at a
evident from Andrea Amati to Guarneri del Gesu. It does not distance h from the centre of a circle whose radius is a, and whose
dominate the individual style of the maker, but it does provide
some insight into the basic framework that wouldhave been taught
to a 1 7th century apprentice. The corners of a Cremonese violin, for Figure 1. Geometrical variables of the cycloid curve. Oocenter
example, will all fit on a circle whose centre is the mid-point of the of circle; P=point used to draw curve; h=distance from O to P;
instrument. This does not impose uniformity, but it does contribute a=radius of circle.
towards a family likeness. Likewise, for all the variety of ff-holes,
the distance between the centers of the upper and lower holes is a
predictable 63-64 mm on a full-size violin. That, it seems, is how it
was done, and I am still waiting to see a significant variation. Of
course, having measured and pierced the initial holes for the ffs,
personal taste could, and did, take over. The Cremonese makers
seem to have hit on the perfect balance between formal structure
and individual expression, and I feel that if we could understand
and apply the original system of design we would be better off than
we are in our present position, where we often copy instruments
without grasping the frame of thought behind them.

With this in mind, I began to examine the cross-archings of all

—
instruments for which I had plans essentially, the Strad posters
[I], technical drawings from the National Music Museum [2], and
arching profiles taken from del Gesu violins displayed in New

48 CASJ Vol. 4, No. 7 (Series II), May 2003

 0

Playfair — Curtate Cycloid Arching in Golden Age Cremonese Violin Family Instruments

centre is at O (Fig. 1). The varying angle of the line OP in relation to of string. It is also possible to scribe a series of elegant shapes using
the fiat surface on which the wheel moves is represented by t. a wheel — these shapes being known to geometricians as cycloids.
Starting with the least interesting (at least, to violin makers), we can
Thus: x=at-h sin(t) roll a wheel along a straight line, while tracing the path of a point on
y=a-h cos(t) the perimeter (Fig. 2A). However, curves familiar to the violin-
When h=a, the result is a pure cycloid. maker emerge when the tracing point is moved in from the perimeter
When h<a the result is a curtate ("curtate" is Latin for "short) of the wheel (Fig. 2B).
cycloid.
When h>a the result is a prolate ("prolate" is Latin for "extended") The shape of the curtate cycloid curve shown in Figure 2B is defined
cycloid (a curve with loops, technically called cusps). by two factors: the size of the wheel, whose circumference defines
the period of the curve, and the distance of the tracing point from
The cycloid curve is also known as a trochoid or roulette, although the center of the wheel, which varies the height of the curve. Figure
this is not strictly accurate, since a trochoid is specifically either a 3 illustrates how variations in wheel size and tracing point location
prolate or curtate cycloid, and roulette refers to all curves created affect the character of the resulting curtate cycloid curves.
by moving circles, be those circles moved against a flat or curved
surface. However, the terms are often used loosely, a problem which Two dimensions, then, are needed: the horizontal width, and the
is complicated by the general non-mathematicalreference works, vertical height of the arch itself. A downloadable program on the
which have based their definitions on the complete Oxford English Internet [6] will print curtate cycloids in Adobe Acrobat (.pdf) format
Dictionary, or the Encyclopedia Britannica, both of which got their based on these two figures (Fig. 4). However, it is also easy to use
definitions identically muddled a hundred years ago. This allows wheels. I used V. inch plywood, and a hole cutter. The hole cutter is
the curious reader to follow a chain of plagiarism right through to adjusted to produce a circle whose circumference is the desired
Microsoft Encarta. For accurate information on this family of distance between the two low points. (Circumference/
curves, I recommend the St. Andrews University mathematics site 3.l4=diameter). The height of the arching is obtained by drilling a
[5]- hole Vz the desired distance from the centre. In practical terms, I cut
the wheels slightly small, and put an elastic band around them. I
DRAWING CURTATE CYCLOID CURVES also glued 80 grit sandpaper to a length of Vz inch plywood, which
Many traditional ways of drawing geometric shapes have been used I used as a straightedge to roll the wheels along. In this way, the
over the years by woodworkers, stonemasons and others. Apart problem of slippage is reduced.
from the obvious straight line and circle, ellipses and even ovoids
are possible using nothing more than a couple of nails and a length
Figure 3. Variations in curtate cycloid curves depending on wheel
size and tracing point location. In the examples shown, the highest
Figure 2. Cycloid curves. arching (middle) is made by a small wheel with a large tracing
A. A simple cycloid curve is traced by a point located on the edge point distance, whereas flatter arching (bottom) is defined by a
of a turning circle. larger wheel with a smaller distance.
B. Curtate cycloid curve. The curve is traced by a tracing point
located within a turning circle.

CASJ Vol. 4, N0.7 (Series II). May 2003 49

 Playfair — Curtate Cycloid Arching in Golden Age Cremonese Violin Family Instruments

Two things must be borne in mind when drawing curtate cycloids. shows back cross-archings that conform fairly well to cycoid curves
First, the arches produced are external. Generally, when we talk of (Fig. 7). Variations occur in the starting points of cycloid curves for
a violin with 16 mm arching, we are including the thickness of the the back plate, from near the edge in the upper, middle, and lower
wood. Assuming that the wood at the lowest point is 3 mm thick, bouts to far within the purfling line at the upper and lower corners.
the external arch is 13 mm, since that is the distance that it rises This sort of variation in low point location is typical.
from the lowest point to the center. Second, the lowest point of the
arching curve is not necessarily identical with the purfling, and the The violino piccolo made in 1613 by Hieronymus and Antonio
position of the low point affects the overall appearance of the Amati is an exceptional example of an early Cremonese instrument
instrument. A low point positioned near the plate edge causes the in unmodified condition. The original finish is intact and the arching
appearance of low, full arching (Fig. 5), whereas a low point shows little, if any, distortion (Fig. 8). Arching templates conform
positioned further in from the edge causes the appearance of higher, approximately to the arching (Fig. 9), showing low points near or
steeper arching (Fig. 6). It is easy to characterize an instrument as inside the purfling line for some profiles, and near the edge for
highly arched when in fact it is no higher than one having arching others. On the top plate, the same cycloid template (98,6.4) appears
that starts near the purfling, and whose contours are more gradual. to fit both the upper corner and the lower corner profiles, raising
the possibility that the same template had been used for both
COMPARATIVE CURVES profiles.
If the idea that Cremonese makers used curtate cycloid curves for
their cross-archings is to have any credibility, two things must be The back plate ofNicolo Amati's "Alard" violin of 1649 shows a
demonstrated: First, it must be shown that cross-archings on existing reasonable correlation with curtate cycloid templates (Fig. 10). Low
historic instruments are reasonable matches with wheel- or points are positioned well within the purfling line, giving the visual
computer-generated curves. Secondly, it must be established that impression of steep, high arching (e.g. Fig. 6).
th
16th and 17 century Europeans were familiar with the cycloid
family of curves. Andrea Guarneri's tenor viola of 1664 (Fig. 11) shows considerable
variation in the position of the low point, from near the edge in the
CROSS-ARCHING PROFILES middle bout of the table, to near the purfling line in the bouts, and
Cross-archings from a variety of Cremonese instruments can easily far within the plate margins in the upper and lower corner profiles
be compared with ideal curves produced by a wheel, or, in this (Fig. 12).
case, by computer [6]. Cycloid curves were overlaid on the cross-
archings of seven well-known instruments (1 violino piccolo, 3 Antonio Stradivari's "Archinto" viola of 1696 (Fig. 13) shows high,
violins, 2 violas, and 1 cello; table 1). Curves are presented as half robust arching with a consistent location of the low point within
templates, as would probably have been used by makers. For the purfling line.
example, Hargrave suggests that half templates were used on
Guarneri del Gesu's violins (See [3], vol. 2, p. 145). In contrast, Joseph filius Andrea Guarneri's c.1705 violin (Fig. 14)
shows considerable variation in the position of the low point, from
Most cycloid curves (shown starting in Fig. 7 using a heavy dashed far within the plate margins, to outside of the plate edge.
line) show a reasonable fit, especially considering that many plates
have distorted with age. The only exceptions that I can find are in Finally, in del Gesu's "Canon" violin of 1743 (Fig. 15), cycloid
the bridge area of tables, where a different, barrel—like shape seems templates project to the edges in the middle bouts, whereas the low
to predominate. points are farther within the plates in the upper and lower bouts.

The famous "King" cello of 1572, one of the set of decorated In summary, reasonable visual correlations are obtained for curtate
instruments that Andrea Amati made for Charles IX of France, cycloid templates superimposed on cross arching profiles of seven
Cremonese instruments. Variations are what might be expected
from differences in the original use of half templates, combined
Figure 4. Ideal curtate cycloid curve generated by computer with distortions caused over time. These half dozen examples are
program "Cycloid" [6]. User specifies width (W) between low
simply for illustration. Most, if not all, Cremonese instruments
points, and height (h) of curve. Horizontal line spacing is 10 mm. that I have tested seem to fit these cycloid templates, at least to my
satisfaction.

HISTORY
Most people subscribe to the snowball theory of knowledge that —
we becomebetter informed as time goes by, and moreis discovered.
W=1 30,00, h=10,50 Violin makers especially, have a hard time with such a concept,

50 CASJ Vol. 4, No. 7 (Series II), May 2003

 Playfair — Curtate Cycloid Arching in Golden Age Cremonese Violin Family Instruments

Figure 5. End views of arching on a viola made by Peregrino di Zanetto, Brescia, after 1564 (left, top plate; right, treble side of back plate).
The arching extends to edges of the plates, producing distinctive "full" character (photo by A.T. King; NationalMusic Museum, no. 3367).

Figure 6. End views of arching on a violin made by Nicolo Amati, 1628. The low point of the arching is inside the purfling line, which
results in a fairly steep visual impression (photo by A.T. King; National Music Museum, no. 3356).

since we are routinely humiliated in our attempts to equal, let alone heavenly bodies moved in epicycloids, which derive from the path
exceed, objects made three hundred years ago. I suspect that the of a wheel moving over a surface that is also curved. Over the
truth is that while our libraries may bulge with an accumulation of centuries, it became necessary to modify Ptolemy's system to bring
knowledge, our minds remain confined to those areas that help us it into conformity with the growing body of observed -information,
in the present. The fact that cycloid curves remain almost unknown but the basic principle was never challenged, especially after it had
and untaught today does not mean that this has always been the been absorbed by the Church, which made its position almost
case. There is no need to construct elaborate arguments as to how unassailable. The cycloid-based movement of the universe would
an obscure geometric construction came to the notice of ordinary have been as routine in Medieval and Renaissance education as the
17th century craftsmen. The construction of cycloid curves had been elliptical shape of an orbiting body is in ours.
common knowledge for centuries, and, at the time of the flowering
of the Cremonese^violin trade, was the subject of intense The successful mathematical analysis of cycloids did not occur until
mathematical investigation. the 17th century. Quite simply, a sufficient mathematical foundation
had not yet been generated. This does not mean that the subject was
It is uncertain whether the first discussion of cycloids was Greek, not of interest. Cardinal Nicolas of Cusa attempted the discovery of a
Arabic, or from further east. However, there is no doubt that formula in 1451, and in Germany the artist Albrecht Durer (1471-1528)
Ptolemy was the first to incorporate the construction into a major included cycloids in the book he wrote in the last years of his life in
work. In the 2nd century he developed a cosmic system that was to — —
which he explored the relationship critical, in his view between
last for 1300 years. He proposed a geocentric universe in which the geometry and success in craftsmanship and art [7].

CASJ Vol. 4, N0.7 (Series II), May 2003 51

 Playfair — Curtate Cycloid Arching in Golden Age Cremonese Violin Family Instruments

Table 1. Dimensions of curtate cycloid templates used to test cross-archings of Cremonese instruments. All dimensions are given in
millimeters. Width and height of curtate cycloid templates (see Fig. 4) are given as "width, height". Templates were made using
computer program "Cycloid" [6].

Plate Length Upper Bouts Upper Corners Middle Bouts Lower Corners Lower Bouts

Andrea Amati, "King" cello (1572)

table 757 320, 15.5 249, 24.5 412, 20.0
back 757 342, 20.0 276, 25.5 241, 27.0 308, 27.0 408, 22.0
Brothers Amati, Violino piccolo (1613)
table 266 113.4, 6.0 98, 6.4 81, 8.0 98, 6.4 134.6, 6.5
back 266 111, 5.3 94, 9.0 74.6, 10.0 101.2, 9.6 156, 6.0
Nicolo Amati, "Alard" viokn (1649)
back 351.5 132.4, 9.2 85, 14.7 175.4, 10.0
Andrea Tenor viola (1664)
table 482 200, 11.7 155.6, 18.0 272, 16.7
back 482 180.8, 9.0 171, 15.5 143, 17.0 191.2, 18.0 222, 11.0
Antonio "Archinto" viola (1696)
table 413.5 164, 9.5 146, 14.0 114, 15.0 192.4, 15.2 225.6, 9.5
back 413.5 163.4, 9.0 150, 14.0 113, 16.0 179, 15.8 229, 9.0
Joseph filius Andrea Viokn (c 1705)
table 354 148, 10.5 92.6, 14.5 187, 11.0
back 355 153, 9.6 14.8, 14.2 94, 13.5 130, 14.2 201, 9.6
Guarneri del "Canon" violin (1743)
table 353 155, 9.6 134.4, 11.5 107, 11.7 170, 11.5 197.2, 11.0

Figure 7. Comparison of curtate cycloid arching templates (dashed lines) with cross-archings for the "King" cello, Andrea Amati, 1572
(cross archings from technical drawing by John Pringle; used with permission of National Music Museum; shaded area indicates
instrument plate). Scale is different for bottom, left profile. W and H refer to width and height of cycloid template. Cycloid curves are
positioned slightly above cross-arching line. Light vertical and horizontal lines are frames for cycloid curve (see Fig. 4); horizontal lines
are 10 mm apart.
Andrea Amati "King" cello, 1572
Andrea Amati "King" cello, 1572 Back plate
Table

Lower bouts W=4oo, H=22

Lower bouts W=4l H=2o

52 CASJ Vol. 4, No. 7 (Series II), May 2003

 Playfair — Curtate Cycloid Arching in Golden Age Cremonese Violin Family Instruments

Figure 8. Neck views of curtate cycloid arching on a violino piccolo made by the Brothers Amati, 1613 (A.T. King photo; National Music
Museum, no. 3361)

Figure 9. Comparison of ideal curtate cycloid arching templates (dashed lines) with cross-archings for the violino piccolo, Brothers
Amati, 1613 (cross archings from technical drawing by John Pringle; used with permission of National Music Museum; shaded area
indicates instrument plate). See Fig. 7 caption for full explanation.

Brothers Amati Brothers Amati

Violino piccolo, 1613 Violino piccolo, 1613
Table Back plate

Lower bouts W=134.6 H=6.5 Lower bouts W=ls6, H=6

CASJ Vol. 4, N0.7 (Series II), May 2003 53

 Playfair — Curtate Cycloid Arching in Golden Age Cremonese Violin Family Instruments

Galileo [8] may have played a major role in destroying the Ptolemaic
system, but he was fascinated by the cycloid, which in 1599 he gave
its modern name (cycloid: "based on the circle, or the wheel"). He Figure 10. Comparison of ideal curtate cycloid arching templates
failed to produce a formula for the shape, but did suggest that a (dashed lines) with cross-archings for the "Alard" violin, Nicolo
Amati, 1649 (cross archings from Strad poster series; used with
cycloid would make an efficient basis for a bridge arch, which was
in fact carried out in the construction of the Ponte di Mezzo over permission; shaded area indicates instrument plate). See Fig. 7
the Arno River in Florence, some time after his death in 1642. caption for full explanation.

By the mid-1 600's there had been a surge of mathematical activity

embracing all fields, and the cycloid was attracting the interest of
Nicolo Amati
many of the leading thinkers of the time. The French priest, "Alard" violin 1649
mathematician, and musician, Marin Mersenne; Pascal (whose last
work was on cycloids); Christopher Wren (a mathematicianbefore
Back plate
he became an architect); Leibniz, and Newton all studied the shape,
and it is interesting to see the gradual absorption of the cycloid into
the repertoire of the engineer as the century progressed. Christiaan
Huygens used properties of the cycloid to develop clocks of
remarkable accuracy [9], and a French engineer named Desargues
established their value in the design of gears.

By 1725 Jonathan Swift, who expressed his satires in the clearest,

most accessible language, has no hesitation in using the term cycloid
as he pokes fun at the Royal Society, the focal point of English
science. He knew that his audience, which was literary rather than
scientific, would still get the joke as he describes mathematicians
sitting down to a rational meal, consisting of '_. shoulder of mutton
cut into an equilateral triangle, a piece of beef into a rhomboides, and a
pudding into a cycloid. However, the word was not to remain
commonplace for long. Since Galileo and the other astronomers of
the 16th and 17th centuries had replaced the Ptolemaic system, Lower bouts W=1 75.4, H=lo
knowledge of cycloids and epicycloids would no longer be
fundamental to an understanding of the universe. And the
mathematicians had devised formulae for every possible aspect of
the cycloid family. It was old news now, and there were new fields From about 1750, the old system of violin design seems to have
for them to conquer. .. disappeared. A system that defined structure and framework, while
allowing latitudeand individuality in the details, died, to be replaced
and simplified by copies of the past masters of Cremona.
". .. the most convincing ideas strike me as Duplication was substituted for evolution. Careful duplication
being those that are simple and direct." would have been a pivotal part of an apprentice's education and
guarantees an acceptable product. Its disadvantage is that it takes
the most unusual maker to rise above the problems of copying, as
Musically, too, times had changed. The Baroque period (ca 1600- this involves producing something fresh and interesting out of a
-1750) coincides almost exactly with the period of Cremonese violin basis of familiar details.
making greatness. When Bach died in 1750, Peter Guarneri of Venice,
the last of the Cremonese-trained masters, was coming to the final CONCLUSION
years ofhis career. The worldhad simplified for both musician and Cremonese violin makers had a product that was the result of an
luthier. The wide instrumental palette employed in the 1600's by efficient and pragmatic style of work and design. In recent years,
composers like Monteverdi, in which the violin, viol, and lute there has been quite a lot of work on recreating that workshop
families combined, was a thing of the past, as was the exploration of technique, and the most convincing ideas strike me as being those
harmonic possibilities as found in the motets of Gesualdo, or that are simple and direct. Since I first raised this subject in 1999 [4],
practical possibilities like the curious, radical violin tunings of the idea of curtate cycloids as an arching system has received far
Biber. Except for a few French viol players, the new world worked more acceptance than rejection, but I feel that it should still be on
with the violin family, and with the sonata form. probation. Thewholeissue ofrecreating a forgotten system of design

54 CASJ Vol. 4, No. 7 (Series II), May 2003

 Playfair — Curtate Cycloid Arching in Golden Age Cremonese Violin Family Instruments

Figure 11. End views of curtate cycloid arching on a tenor viola made by Andrea 1664 (A.T. King photo; National Music
Museum, no. 3354).

Figure 12. Comparison of ideal curtate cycloid arching templates (dashed lines) with cross-archings for tenor viola, Andrea
Guarneri, 1664 (cross archings from technical drawing by John Pringle; used with permission of National Music Museum; shaded
area indicates instrument plate). See Fig. 7 caption for full explanation.

Andrea Guarneri Andrea Guarneri

Tenor Viola, 1664 Tenor Viola, 1664

Upper corners W=l 7l, H=15.5

Middle bouts W=155.6, H=lB

Middle bouts W=l 4o, 1-I=l7

Lower corners W= 19 1.2, H=lB

Lower bouts W=272, H=16.7 Lower bouts W=222, H=ll

CASJ Vol. 4, N0.7 (Series II), May 2003 55

 Playfair — Curtate Cycloid Arching in Golden Age Cremonese Violin Family Instruments

Figure 13. Comparison of ideal curtate cycloid arching templates (dashed lines) with cross-archings for the "Archinto" viola, Antonio
Stradivari, 1696 (cross archings from Strad poster series; used with permission; shaded area indicates instrument plate). See Fig. 7 caption
for full explanation.

Antonio Stradivari Antonio Stradivari

the "Archinto" viola, 1696 the "Archinto" viola, 1696
Table Back plate

Upper bouts W=l64, H=9.5 Upperbouts W=1 63.4, H=9

Lower corners W= 192.4, H=15.2

Lower corners W=l79 H=15.8

Lower bouts W=225.6, H=9.5 Lower bouts W=229, H=9

Figure 14. Comparison of ideal curtate cycloid arching templates (dashed lines) with cross-archings for a violin by Joseph Guarneri
filius Andrea, c. 1705 (cross archings from Strad poster series; used with permission; shaded area indicates instrument plate). See Fig. 7
caption for full explanation.

Joseph Guarneri filius Andrea Joseph Guarneri filius Andrea

Violin, c. 1705 Violin, c. 1705

Lower bouts W= 187, H=l l Lower bouts W=2ol , H=9.6

56 CASJ Vol. 4, No. 7 (Series II), May 2003

 Playfair — Curtate Cycloid Arching in Golden Age Cremonese Violin Family Instruments

Figure 15. Comparison of ideal curtate cycloid arching templates (dashed lines) with cross-archings for the "Canon" violin, Guarneri
del 1743 (cross archings from Strad poster series; used with permission; shaded area indicates instrument plate).

Guarneri del Gesu Guarneri del Gesu

the "Canon" violin, 1743 the "Canon" violin, 1743
Table
Back plate

Upper bouts W=lss, H=9.6

Upper bouts W=l 62, H=7.4

Upper corners W=1 33.2, H=10.5

Lower corners W=ls7, H=l l

Lower bouts W=197.2, H=l l Lower bouts W= 195, H=7.4

is risky, to say the least. In its favor are the close matches obtainable [6] 'Cycloid' computer program, written by Stephen Mann for
with old instruments, and the simplicity and elegance that seem to Quentin Playfair. Free download available from http://
be the hallmark of Cremonesework. Against it is the fact that there www.cgl.uwaterloo.ca/~smann/ccycloid/
seems to be no written evidence to support my claim. As such, it [7] Encyclopaedia Britannica, 1911, 1978, Articles on Cycloids,
must remain a hypothesis, although I like to think of it as a feasible Pascal, Durer, and Wren.
and useful one, and well in accord with Cremonese tradition. [8] Boyer, 1967, Galileo's place in the history of
mathematics: Galileo, Man of Science, edited by McMullin,
in
E.: New York, Basic Books, p. 232-255.
REFERENCES CITED [9] Huygens, Horologium oscillatorium (Paris, 1673), trans.
[1] Orpheus publications, Strad poster series. Available from Richard J. Blackwell, Christiaan Huygens' The Pendulum
www.thestrad.com/ Clock or Geometrical Demonstration Concerning the Motion
[2] Pringle, J., Technical drawings of stringed instruments at the of Pendula as Applied to Clocks (Iowa State Press, 1983); for
National Music Museum, Vermillion, South Dakota, available an edition and translation of the Horologium of 1658, see
from http://www.usd.edu/smm/giftshop.html#drawing Ernest L. Edwardes, The Story of the Pendulum Clock
[3] Biddulph, P., Carlo C, Dilworth, J., Hargrave, R., Klein, P., (Altrincham, 1977), p. 60-97
Pollens, S., Rosengard, D., and Wen, E., 1998, Giuseppe [10] Pigoli, T., 1984, Evoluzione delle curve di bombatura:Liuteria,
Guarneri del Gesu: Peter Biddulph, London, 339 p. no. 10, April, 1984.
[4] Playfair, Q., 1999, Cremona's forgotten curve: The Strad,
Orpheus publications, London, vol. 110, no. 1315, Nov. 1999,
p. 1194-1199. ADDITIONAL REFERENCES
[5] University of St. Andrews, History of Mathematics (Note: Camus, C, 1733, Sur la figure des Dents des Roues, & des Ailes des
one of the best web sites I have come across in any area): Pignons, pour rendre les Horloges plus parfaites: Histoires et
http://www-groups.dcs.st-and.ac.uk/ memoires de l'Academie des Paris.

CASJ Vol. 4, N0.7 (Series II), May 2003 57

 Playfair — Curtate Cycloid Arching in Golden Age Cremonese Violin Family Instruments

Farouki, R.T., and Rampersad, J., 1998, Cycles upon cycles: An My thanks to Roger Hargravefor his permission to use archingprofiles
anecdotal history of higher curves in science and engineering: published in "The Strad", and the Guarneri book, which were the reason
Mathematical methods for curves and surfaces II (M. Daehlen, I started to think in terms of a Cremonese family likeness in the
first
T. Lyche, and L. L. Schumaker, eds.), Vanderbilt Univ. Press, place. The Stradposters have become an integral part the lives
of of
p. 95-116. many of us. My thanks to Naomi Sadlerfor allowing the Catgut Society
Walks, J., 1659, De Cycloide: Oxford. to use themhere. Dr. Andre P. Earson of the NationalMusic Museum
in Vermillion, South Dakota has kindly givenpermissionfor the use of
arching profiles measured by ]ohn Pringle, andphotographs of various
ACKNO WLEDGMENTS instruments in the collection. A. Thomas King took the photographs,
Shortly after I completed my first piece on cycloid curves in 1999, I and Jeff Eoen drafted several of the charts. Dr. Steven Mann, of the
became aware that the late Tullio Pigoli of Cremona, Italy had written department of Computer Sciences at the University
of Waterloo helped
on this subject in 1984 [10]. Although I reached my conclusions me make the transitionfrom wheels to computer printouts by writing a
independently, there is no doubt that he was thefirst to explore this area. program that even I could use, and has allowedthe use of his website as
It is unfortunate that Liuteria had such limitedcirculation. a source of thatprogram.

58 CASJ Vol. 4, No. 7 (Series II), May 2003

 Working Methods of Early Classical Violin Makers:
Implications of Recent Dendrochronological Studies

John Topham
114 Mid Street, South Nutfield, Redhill, Surrey, RHI 4JH, UK
e-mail: [email protected]

ABSTRACT Dendrochronology relies on the identification and assessment of

Dendrochronology, an established technique to precisely date wood, the growth pattern of the grain of the wood. A date is obtained by
has been successfully utikzed to analyze the spruce fronts of many measuring and recording the pattern, and comparing it to reference
musical instruments of the violin and viol families. Not only can material using pattern matching, visual, and statistical methods.
dates be established but wood from each side of fronts of the same The fundamental strength of dendrochronology in contrast with
or different instruments can be shown to match significantly other methods, such as carbon dating, is its precision. If a positive
allowing an analysis of relationships between instruments and dendrochronological date is obtained for a piece of wood, it is
makers. This analysis can be extended to assess the nature of the absolute—there is no margin of error, and no probability of an
makers' choice of wood for the fronts of instruments as well as the earker or later date. Dendrochronological dating is based on the
way makers made use of the wood. principle that the widths of the year-rings of a tree, which is
influenced by ckmatic and environmental conditions, is common
Many classical makers appear to have used fairly young wood, and in most trees of the same species that grow over a large area. These
paired spruce pieces with little concern for bookmatching or conditions include constants such as latitude and altitude, and
symmetry. Evidence from Guarneri del Gesu's violins indicates variables such as rainfall and temperature. The history of the tree's
that wood as kttle as 3 years old was sometimes used, and Stradivari growth is therefore, in a sense, encoded in the pattern of the tree
commonly used wood less than 20 years old. Spruce pieces in rings.
Stradivari's vioUns are generally not bookmatched, but come from
different parts of the same tree used during a specific period. He The similarity of growing conditions in a particular area, often
may have taken logs or billets of wood, split them into thin, single reflected in similar grain patterns, allows the matching of wood
pieces, and then later selected often non-adjacent pieces for building from different trees. If the felling date of a particular tree is known,
two-piece fronts. Spruce pieces in two-piece fronts are rarely an exact pattern match with an undated neighbouring tree can be
bookmatched, and often came from different trees in different found and a date for the wood can be estabkshed. Wood from a
regions. Finally, no evidence was found that Stradivari or del Gesu variety of musical instruments can be subjected to similar analyses
used special wood from a secret source, because significant cross using long sequences of measurements known as reference
matches occur between makers in different regions. chronologies, and dates for the fronts of these instruments can be
fixed.
INTRODUCTION
Since the 1980's, the science of dendrochronology or tree-ring DENDROCHRONOLOGY APPLIED TO THE VIOLIN FAMILY
analysis, which establishes absolute dates for old wood, has been The construction of most bowed stringed instruments including
providing new information on classical musical instruments. violins involves the use of wood from the maple family (Acer
Dendrochronology's appkcation in general has been used in many platanoides L.) for the back, sides and neck. The front, however, is
fields of research including archaeology, climatology, forest generally made from Norway spruce (Picea abies (L.) Karsten), which
management, ecology and art-history for the past eighty years or so.because ofits stiffness/weightratio is thought to possess the optimum
An American scientist, Andrew Douglass, pioneered the technique acoustic properties. The making process usually results in a two-
in the early 1930's and applied it to the study and dating of early piece front with a longitudinal join in the centre of the instrument,
American settlements in the Midwestern US. In the field of musical with the youngest year-rings coming together at the centre join
instrument research, dendrochronology has not only been shown (figure 1).
to be an incisive method for dating wood used on musical
instruments, particularly those from the violin family, but it has The spruce front is used for dendrochronological investigation
also provided a gHmpse of the hitherto unrecorded choices of makers because maple is unsuitable, due largely to its erratic growth. The
concerning wood [1,2]. remarkable thing about spruee—with its very clear annual growth

CASJ Vol. 4, N0.7 (Series II), May 2003 59

 John Topham — Working Methods of Early Classical Violin Makers: Implications of Recent Dendrochronological Studies

Figure 1. Diagrammatic representation of how the front of a viokn Figure 2. Schematic representation of the cross-matching of a ring-
is assembled. width sequence taken from a viokn against a reference chronology
constructed from localized material.

j
_
EnDnnimTiiiiii'
Reference Chronology
' DID] ,:: If!) : -

behaviour, particularly when it grows at high altitude—is that this

Instrument Sequence
similarity of grain pattern can extend for hundreds of kilometres
allowing wood taken from large regions to be matched. It has been
traditionally known that most violin makers in Europe obtained
their wood from the Alpine region, which covers many thousands
of square kilometres. For that reason, most research has fe"<^^raST~:::^^l_^'
concentrated on that area. Chronologies of spruce are compared
with tree-ring series takenfrom instruments and if a matchis apparent
then a date can be estabkshed (Figure 2). In recent years, I have
been able to measure the fronts of over a thousand instruments,
mainly from the violin family (violins, violas & cellos). The list death by the notable Piedmont dealer, Count Cozio de Salabue in
also includes instruments of the viol family, lutes and early guitars. the late eighteenth century. Some of his drawings also exist, as do
All instruments originated from countries in Europe, but, because some of the wooden forms on which Stradivari constructed the rib
of their availability, most of the instruments are either English or assembkes of his viokns and which were the basis of Pollens' book
Italian. [s]. Tools and templates of other makers are rare and certainly are
not in the pubkc domain. In essence, all we have left are the highly
WORKING METHODS OF THE OLD MASTERS —
prized results of the makers' labours instruments in various
Few makers have ever written down their working methods. conditions that have been vigorously played and are often heavily
Notable amateurs such as Edward Heron-Allen in the latenineteenth repaired and restored. The elegance of these instruments is plain
century wrote in detailof all aspects of the process of making viokns, to see, although the makers' initial decisions on beginning the
primarily with a view to help others make them, as well as construction of the instruments are difficult to assess. Many of the
explaining his own methods and sources of materials [3]. More decisions would have been based on empirical experience —
recently, Otto Mockel in 1930 (and updated in 1954 in association thickness of the plates, softness of varnish, etc. Many of these
with Dr Fritz Winkel) also provided a highly descriptive account of decisions are now a mystery; however, dendrochronology may be
making a violin, including an introduction of its history and able to assist in re-discovering some of the makers' methods and
scientific experiments in sound production [4]. Although both thereby uncover some of the makers' decisions.
Heron-Allen and Mockel are very descriptive, few of their
instruments have survived, and proof of the success of their methods POSSIBLE INFORMATION AVAILABLE FROM
is effectively unavailable. On the other hand, the large numbers of DENDROCHRONOLOGY
beautiful instruments by classical makers such as Stradivari, Unlike in other fields of tree-ring research, musical instrument
Guarneri del Gesu and Guadagnini that have survived have few or dendrochronology is limited in many ways. In fields such as
no written descriptions of their construction. Records of working ckmatology and forest management, core samples taken from kving
methods of these makers, something that would be most valuable and dead trees can be analysed in great depth. The density of the
to the modern maker, hardly exist at all. The Stradivari Museum in wood, which varies much kke the yearly growthrings, can be studied
Cremona has some of Antonio Stradivari's tools retrieved after his using x-rays to indicate such things as the health of the tree as well

60 CASJ Vol. 4, No. 7 (Series II), May 2003

 John Topham — Working Methods of Early Classical Violin Makers: Implications of Recent Dendrochronological Studies

woods (ebony, varieties of rosewood, etc), or show very erratic

Figure 3. Method of dendrochronologic study of spruce. yearly growth, such as with maples.
A. The ring widths measured from the front of an instrument are
plotted evenly along the bottom axis with each ring representing Ring-width information, however, still provides us with some idea
one year. of what the maker had in mind when he or she made the instrument
B. The width of each ring is represented by how high the point is [Note: For ease of observation,ring widths are represented as graphs.
plotted against vertical axis cakbrated in hundreds of a milkmetre. Figure 3 shows how these graphs are constructed]. One of the first
C. The points are joined with straight lines in order to see the properties that can be examined is the general grain density. This is
pattern of the ring widths more clearly. simply the number of rings per unit length (inches or centimetres)
and is usually indicated by the average ring width per series
(Figure 4). Another aspect of grain structure is its evenness, which
in dendrochronological terms is indicated by the grain's 'sensitivity.
The more sensitive the grain is the more varied the ring widths are
(Figure 5).
A. The grain as it would be seen on the front of an instrument.
Another property found is the variation of the widths of the rings
over the whole width of each piece used on fronts. Often pieces
„ 120 have been picked which show a gradual narrowing of the width of
.1 ♦
M.

so |
♦
T
; ♦
rings from the oldest part of the piece ofwood to the youngestpart.
. ♦ 4 This is a natural feature of the structure of wood in a tree. A young
JT _o .; a 4
tree has a small diameter and depending on its leaf crown it lays
__ ° ■ !
>s
f|li
4

20
♦ A
#
♦ ♦
♦
!
♦ #
:
down relatively wide rings. As the tree gets older, the tree's diameter
: ! #♦
gets larger and although there is an increase of the generation of
0 1 2 3 4 5 S 7 8 . 10 11 12 13 1. 15 16 17 18
wood mass as the leaf crown expands, this volume is distributed
Years
more thinly around the ever-increasing circumference of the tree's
B. Ring widths are plotted as points. trunk. This gives rise to ever-decreasing ring widths as the tree
grows. This effect is in general quite common but can be very
noticeable on some instruments, for example Stradivari's famous
1716 'Messiah' viokn. The widths of the oldest rings on the front of
this instrument (situated at the outer edges) are between 2 to 2.5
millimetres whereas the widths of the youngest rings (at the centre
of the front) are approximately 0.3 to 0.4 mikimetres. In general,
makers may not have much choice in the matter; however, some
makers appear to choose wood with more variation, others with
Years less. On a viola made by Wilkam Baker in c.1690 the grain of the
C. Lines are added for clarity. bass side of the front varies from just over 3 millimetres to just over
a third of a millimetre (Figure 6). In contrast, although by not
much, on the bass side of a front from a viokn by Carlo Bergonzi
made in 1723 the grain width varies from 1.57 millimetres for the
as the nature of the physical stress underwhich the tree kved. The oldest part to 0.77 millimetres for the youngest (Figure 7).
variation of different carbon isotopes found in each year ring can
also be studied to gain some idea of the sun's activity, which would ESTIMATES OF TIME ALLOWED FOR WOOD SEASONING
have strongly influenced the atmosphere at the time. This has Dendrochronological studies allow us to make some general
allowed researchers to accurately model the climate in particular conclusions about the time allowed for wood seasoning in classic
regions overmany years. Most of these techniques, however, involve instruments. The time period between the youngest ring and the
the destruction of the sample under study. attributed date is determined by two factors: seasoning time before
manufactureand removal of some of the outermost rings during the
For obvious reasons, such information cannot be extracted from planing process prior to joining the front. This number varies widely,
the front of a viokn. The only information available is a series of from a minimum of three years for some instruments by Guarneri
ring-width measurements. For various reasons only the spruce del Gesu [6] to several tens of years for most other makers. The
fronts ofmusical instruments (both plucked and bowedinstruments) frequent short intervals between the youngest ring and date of
can be used for dendrochronological investigation. The other woods manufacture suggest that Cremonese makers made no systematic
used generally either do not show yearly growth, such as tropical attempt to remove sapwood. The short intervals also indicate that

CASJ Vol. 4, N0.7 (Series II), May 2003 61

 John Topham — Working Methods of Early Classical Violin Makers: Implications of Recent Dendrochronological Studies

Figure 4. Ring-width sequences taken from two viokns of approximately the same width. The measurements are plotted on the same
vertical scale and positions.

Figure 5. Tree-ring series taken from the fronts of a Barak Norman bass viol and a Giovanni Francesco Pressenda viokn. The viol series
shows an example of relatively sensitive grain and the viokn series shows an example of less sensitive or complacent grain. Both graphs
are plotted to the same vertical scale although they have been displaced vertically to allow better observation.

I Barak Norman bassviol: centre strip I !!!!!! !

! Giovani Francesco Pressend a vioIn: one piece !!!!!!!

s
-;e_ Oi 50 60 70 00 90 100 '1 10 'i 00 000!
Years (relative)

Figure 6. Sequence taken from the bass side of the front a viola by Wilkam Baker of Oxford, c.1690. The grain widths vary from 3.02mm
to 0.31mm.

— _- William Bak er viola, 0.1690: bass side

2.0 ____: !
VI !
m i r
I 1.0 L
;.
1 i
0.5 !■
1

10 20 35 <n 50 60 70 SO 90 100 110 120 130

Years (relative)

62 CASJ Vol. 4, No. 7 (Series II), May 2003

 John Topham — Working Methods of Early Classical Violin Makers: Implications of Recent Dendrochronological Studies

Figure 7. Sequences taken from the bass side of a front from a viokn by Carlo Bergonzi, 1723.

—
— Bergonzi 1723:bass side

0T

0 T 00 40 m 60 00 00
Years (relative)

Figure 8. Sequences from fronts of the 1711 'Parke' violin, the 1714 'Dolphin' viokn and the 1715 'Marsick' viokn. The bass side is
represented by the top curve for each instrument,

1600 1620 ■;_ "■_ if!__ ■;s;_. 1700

Years (AD)

in many cases it was not thought important to allow a long seasoning PROPERTIES OF WOOD FOUND ON INSTRUMENTS BY
period. Guarneri del Gesu has a reputation as a hasty and not overly CLASSICAL MAKERS
careful workman and perhaps the short seasoning time is not Wood properties vary considerably from maker to maker and may
unexpected. However, numerous instruments by Stradivari (noted indicate the resourcefulness of the craftsman. The moie affluent
for meticulous craftsmanship) seem to be made with rather young the maker, the more kkely that he or she may have access to the
wood. The interval between the youngest ring and the date of quality of wood they regard as the best, i.e. wood with perhaps
manufacture of 79 Stradivari instruments varies widely, from 3 to narrow, even year-rings. It may also indicate a particular preference
60 years, with 14 instruments having an interval of between 7and 15 the maker is seeking—one maker perhaps bekeving an open, wide
years [7, B]. This, and other evidence, suggests that extended grain structure will produce better tone, others preferring a denser
seasoning periods were not considered important by classical makers structure.
PI-

CASJ Vol. 4, N0.7 (Series II), May 2003 63

 John Topham — Working Methods of Early Classical Violin Makers: Implications of Recent Dendrochronological Studies

Figure 9. Sequences from fronts of the 1719 'Alba' viokn, the 1722 'Conte de Chaponay' viokn and a viokn made in 1717. The bass side
is represented by the top curve for each instrument.

Some makers appear to have either deliberately or inadvertently One thing that could account for this is the way Stradivari may have
adopted particular strategies in the use of their wood. For example, stored the pieces he spkt up. It is possible he spkt logs into thin
the most famous maker of all, Antonio Stradivari, may have been single pieces suitable only for one side. This would have shortened
quite systematic in his use of wood. From about 70 viokns I have the seasoning time, allowing the wood to be used sooner. The
measured so far, it appears that he may have taken logs or bikets of pieces may then have been stored all together but in no particular
wood, spkt them into pieces and used them for a particular period. order. When Stradivari came to join his fronts, he would have
Two periods that most clearly follow this trend are from 1711 to picked out any pair of pieces in no particular order, resulting in the
1716 and 1717 to 1723. Examples of pieces he used from the 1711-16 mix we see today. Since the pieces came from the same log, the pair
period are represented in Figure 8. The graph shows sequences would have looked reasonably well matched but
from three viokns, the 1711 'Parke', the 1714 'Dolphin' and the 1715 dendrochronologically it is easy to see the differences.
'Marsick' viokns. These curves look very similar and upon closer
inspection four prominent narrowrings can be seen at 1639, 1677,
1685 and 1689. Similarly, examples of pieces he used from the 1717- "Stradivari ... would have picked out any pair
-23 period are represented in Figure 9. The graph shows sequences
of pieces in no particular order, resulting in the
from three viokns, the 1719 'Alba', the 1722 'Conte de Chaponay'
viokns and a violin dated 1717. Again the curves are similar to each
mix we see today. "
other, and, upon closer inspection, it can be seen that the most
prominent common characteristic of these curves is the group of In stark contrast, it appears Giuseppe Guarneri del Gesu utilised
narrowrings between 1620 and 1630. What is perhaps not so easy to his wood in a very different way to Stradivari. With evidence from
see is that the statistics do not support the idea that pairs in both 19 instruments in my database, not much of a systematic approach
groups are 'book-matched', i.e. pieces from each pair from each appears to have been adopted. Typical examples from three
instrumentwere taken from adjacent positions in the tree. In fact, it instruments are represented in Figure 10. Each instrument has a
appears that one piece from one instrument is closer to a piece on two-piece front as indicated on Figure 10 by bass and treble sides.
another instrument than they are with their paired piece. For However, not only do none of the pairs match each other, but none
example, the treble side of the 1715 'Marsick' is more similar to the of the pieces very significantly cross-match pieces from other
treble side of the 1711 'Parke' suggesting these are more closely instruments either. On close inspection they cross-match enough
adjacent. Equaky, the bass side of the 1722 'Conte de Chaponay' to fix their positions in time but the differences in the cross-match
more closely matches the treble side of the 1717 violin again strongly indicate that all six pieces came from different trees from
suggesting an adjacent pair. different regions.

64 CASJ Vol. 4, No. 7 (Series II), May 2003

 John Topham — Working Methods of Early Classical Violin Makers: Implications of Recent Dendrochronological Studies

Figure 10. Sequences taken from three viokns made by Giuseppe Guarnei del Gesu. It appears that aU six pieces came from different
trees.

treble side

1550 1560 1570 1580 1590 1600 1610 1620 1630 1640 1650 1860 1670 1680 1690 1700 1710 1720 1730 1740
Years (AD)

There are two Guarneri del Gesu instruments in the database that POSSIBLE DIFFERENCES BETWEEN MODERN AND OLD
do cross-match very well, a 1735viokn and the 1738 D'Egvike viokn, WOOD
and the match suggests that aU four pieces came from the same tree. On a final note, with respect to the way classical makers chose
The sequence of ring-width measurements of the two instruments wood, a recent long discussion has taken place on the musical
are represented in Figure 11. Overall, though, the wood used by instrument web site 'Maestronet' prompted by the question, 'What
Guarneri is quite variedand his choice ofwood appears to be erratic. if Stradivari made a viokn from modern wood?'. The people who
took part appeared divided into two camps, those who bekeved the
In all the dendrochronological studies of instruments, particularly wood used by Stradivari was 'special' in some particular way, and
those made in the classical period in Cremona and elsewhere in those who believed Stradivari could produce an instrument with
northern Italy, no pattern has appeared which may indicate a outstanding tonal qualities, characteristic of his instruments, with
methodical approachrelating to the best configuration for optimum modernwood. Little mention was made of recent objective research,
sound production. It has been considered, particularly in the modern so kttle formal analysis took place. However, new results from the
era, that 'book-matched' pieces especially for the front of the dendrochronology of Stradivari's and other classical Italian
instrument are the best arrangement for good sound production or instrument can provide part of the answer to the above question.
more desirable tone. As shown above with some instruments this
has indeed been the case, but, in general, and more noticeably with
the classical instruments, most makers appear to have paired pieces
with kttle concern for symmetry. In the kght of this knowledge, it
"it may be worth reconsidering the ethos of
is surprising that these instruments have suffered kttle tonaky. On symmetry, and instead opt for a more
the contrary, they have gained the reputation of being some of the abandoned method of pairing pieces"
best instruments ever made. With this in mind, it may be worth
reconsidering the ethos of symmetry, and instead opt for a more
abandoned method of pairing pieces where any two pieces will do According to my recent studies, on a dendrochronological level the
providing they 'look' right together. This variation may add to the wood used by northern Italian violin makers including Stradivari,
idea that either a good sound is in the ear of the beholderrather than other Cremonese and Venetian makers, particularly in the
any particular specific property provided by the wood, or that eighteenth century, is very similar. The wood from instruments
craftsmanship is more influential in achieving a good sound quakty from different makers often cross-matcheach other very significantly
than was previously thought. and is close enough occasionally to suggest the pieces came from

CASJ Vol. 4, N0.7 (Series II), May 2003 65

John Topham — Working Methods of Early Classical Violin Makers: Implications of Recent Dendrochronological Studies

Figure 11. Sequences from two Guarneri del Gesu viokns that show very similar characteristics which suggest that all four pieces came
from the same tree.

bass side /\_

trebleside

bass side

trebte sMe

1650 1660 1670 1680 1690 1700 1710 1720 1730 1740
Years (AD)

the same tree. In addition, many of the pieces from Italian wood used by major European makers. From this, one can deduce
instruments, including those made by Stradivari, match wood on that the wood Stradivari used had properties that were similar, if
instruments from different countries such as England and France, not identical, to modern spruce growing under the same conditions.
sometimes very significantly. This suggests that no particular maker In this respect, Stradivari's instruments have no particular tonal
had a secret source of wood known only to him. advantage.

CONCLUSION
"...no particular maker had a secret source of More work is needed to gain better insights into how makers of the
past chose their wood. Research has found that some makers appear
wood known only to him. " to be moderately consistent in the choice of their wood, length of
seasoning, and the way they joined pieces together, yet others appear
to be very erratic. The idea is emerging that the instruments from
According to most research I have seen, most, if not ak, fronts of this period are generally perceived to be of very high quality with
instruments today are made from one species of spruce (Norway respect to tone, appearance, and craftsmanship, almost irrespective
—
spruee Picea abies (L.) Karsten). Sacconi [10] mentioned Stradivari of the choices the maker made for his wood, at least as far as the
using Picea excelsa and, in fact, an unpubkshed paper by Prof. Eko spruces fronts are concerned. These inferences question long held
Corona of the University of Tuscany in Italy suggested that over views, and perhaps this gkmpse of the working practices of some of
thirty species of conifer this might have been used for the the greatest makers can provide a firmer basis for making better
construction of instrument fronts of one kind or another, although instruments in the future.
personally I have seen little evidence of this. According to all the
wood catalogues I have seen Picea excelsa is just another name for REFERENCES
Picea abies, and in spite of descriptions of the use ofmany softwood [1] Klein P., 1989, Dendrochronological analysis of European
species, according to the current dendrochronological evidence it string instruments: CIMCIM Newsletter, p. 37-41.
appears overwhelmingly that Picea abies is the only conifer wood [2] Topham J., and McCormick D., 1998, A dendrochronological
to have been used. investigation of British instruments of the violin family:
Journal of Archaeological Science vol. 25, p. 1149-1157.
Just this one species of spruce grows in the Alps today, as it has for [3] Heron-AUen, E., 1885, Viokn-Making as it was, and is: Ward
thousands of years and the Alps that have provided most of the Lock & Co. Ltd, London, 366 p.

66 CASJ Vol. 4, No. 7 (Series II), May 2003

 John Topham — Working Methods of Early Classical Violin Makers: Implications of Recent Dendrochronological Studies

[4] Mockel, 0., & Winkel, F., 1954, Die Kunst dcs Geigenbaues: [8] Topham, J., 2002, A dendrochronological study of instruments
Verlag Bernh. Friedr. Voigt. Berkn, 332 p. by Antonio Stradivari: unpubkshed paper available at http:/
[5] Pollens, S., 1998, The violin forms of Antonio Stradivari: /www.btinternet.com/~j.topham/dendrola.htm
Biddulph, London, 34 p. [9] Topham, J., and McCormick, D., 2000, A Dendrochronological
[6] Klein P., and Pollens S., 1998, The technique of Investigation of Stringed Instruments of the Cremonese School
dendrochronology as applied to violins made by Giuseppe (1666-1757) including "The Messiah" violin attributed to
Guarneri del Gesu: In Giuseppe Guarneri del Gesu (Edited by Antonio Stradivari: Journal of Archaeological Science, vol. 27',
Biddulph P.), Peter Biddulph, London, p. 159. p. 183-192
[7] Topham, J., 2001, The dendrochronology of Stradivari's [10] Sacconi, S.F., 1979, The 'secrets' of Stradivari: Libreria Del
viokns: Journal of the Viokn Society of America, Proceedings, Convegno, Cremona, 285 p.
vol. XVII, no. 3, p. 133-179.

CASJ Vol. 4, N0.7 (Series II), May 2003 67

DYNAMICS OF THE BOWED STRING Museum of Vienna (TMW) has lent equipment, including the
Jim Woodhouse; email [email protected] museum's Polytec laser vibrometer. Dipl. Ing. Wolfgang Vogl of
Paul Galluzzo; email [email protected] Thomastik-Infeld has lent equipment and technical support.
Cambridge University Engineering Department
Trumpington Street, Cambridge, CB2 IPZ, U.K. Project Status:
Paul's web site: http://www2.eng.cam.ac.uk/~pmg26/ The two-year project will conclude in Summer of 2003. Results
wik be presented at SMAC '03
Work in Cambridge on the dynamics of bowing a string has been
going on for many years. Computer models have been developed STRING ACOUSTICS WORK AT KTH
which allow bowing to be simulated, taking account of several Erik Jansson; email [email protected]
complicating factors of real strings, bows and instruments. The Department of Speech, Music, and Hearing
models can be used to ask "design" questions: "will this note be TMH Drottning, Kristinasv. 31
easier to play if I change such-and-such a feature?" However, to be 10044 Stockholm Sweden
confident in the answers it is necessaryto be sure that the simulations
are accurate. The current project is to conduct systematic Twenty years of teaching material "Acoustics for Viokn and Guitar
experiments using a computer-controlled bowing machine to give a Makers" wik soon be on the web http://www.speech.kth.se/music/
wide range of bowing gestures. Results can then be compared with acviguit4/index.html free to copy and use (a slower way is via http:/
the various simulationmodels in the hope of fine-tuning the models. /www.speech.kth.se ckck on music acoustic, news, and end of the
Results so far show encouraging agreement with the trends, but compendium line). The present main research project is to
various details of the bowed transients do not agree with the understand and explain the broad peak, "the BH-hiU" at 2500 Hz,
corresponding simulations. The next challenge is to adjust the which is most important for the quakty of a viokn. It has hitherto
simulation model, particularly the part of it which accounts for the resisted a solution but is clearly not confined to a bridge resonance
frictional behaviour of rosin, to remove these disagreements. The motion alone.
result should be a PhD for Paul Gakuzzo in 2003.

DO FLAT-BACK BASSES SOUND DIFFERENT FROM NUMERICAL OPTIMIZATION OF VIOLINS AND BOWS
ARCHED-BACK BASSES? Assistant Prof. Mats Tinnsten; e-mail [email protected]
Andrew Brown; email [email protected] Assistant Prof. Peter Carlsson; e-mail [email protected]
Musical Acoustics Research Institute, (Institut fuer WienerKlangstil) Mid Sweden University
University of Music and Performing Arts, Vienna, Austria Dept. of Information Technology and Media
SE-831 25 Ostersund Sweden
Purpose/goals:
This project attempts offer an explanation why the double bass is Project Status: Ongoing projects. Time two years.
the only modern bowed orchestral instrument that has a flat back.
The research starts from current surveys among instrument makers 1. Determination of wood propertiesfor blanks for violin tops
and musicians, and also considers some of the psycho-acoustical How do different parameters affect the vibration properties and the
and room-acoustical characteristics of the double bass that set it characteristics of sounds emanating from viokns (i.e. what makes a
apart form other instruments of the violin family. Finally, good viokn good)? Numericalmodels (FEM, finite element method;
experiments using standard analysis methods and kstening tests are and BEM, boundary elementmethod) of whole viokns and/or parts
being used to look for measurable and audible differences in are created and studied. In the proposed method, a FEM-code is
instrument radiation and tone colour between the two types. linked with a stochastic optimization algorithm in order to
automatically determine material parameters. The method requires
Participants: geometrical dimensions, density, and measured normal modes for
Heinz Fischbach, Ohlstadt, Germany, cello and bass maker, has the blank and it considers the fact that the Young's modulus in the
loaned five new basses that were made especially for this study. longitudinal direction varies in the radial direction.
The design is practically identical except for the backs, three flat
and two round, which faciktate the exchange of the backs as a control 2. Numerical optimization ofviolin top plate thickness and
if desired. The materials of one pair of basses, made of spruce and arching
maple, are quite similar, as are those of a second pair, made of Wood for musical instruments exhibits large variation in the
spruce and poplar. The firm AKG has generously supplied six material parameters, directly influencing the properties of the
good quality microphones. Dr. Peter Donhauser of the Technical vibrating structure. The objective of this project is to show that it is

68 CASJ Vol. 4, No. 7 (Series II), May 2003

possible to compensate for differences in material parameters of impulse will have died away. However the force applied by the
violin top plates by changing the distribution of plate thickness and string is applied not only quasi-periodically but at positions along
arch height, thus keeping the eigenfrequencies unchanged. The the hair that vary continuously. The bow is continually vibrating in
thickness and arch height compensation is determined using FEM response to these excitations, which in turn react on the string, thus
analysis linked with a stochastic optimization algorithm. providing a mechanism for the influence of tone. A succession of
essentiaky static measurements would seem unkkely to reveal very
3. Numerical optimization of violin bows with varying wood much about the bow influence on the instrument's tone.
densities
Pernambuk wood (Guilandina echinatd) for violin bows has large In a paper published in the Journal of the Acoustic Society of
density variations. Higher density is preferred for high quality America ( 108, 357-369, (2000) July, 2000 issue) Jim Woodhouse,
bows, but since Pernambuk is rare and expensive, we investigate Steve Garoff, and I derived and illustrated a method of measuring
the possibility of compensating for density variations in the wood the force driving the string at the bowing point. In principle ak that
while keeping constant first eigenfrequencies, mass, and location is required for a bow measurement is a set of two string transducers
of the center of gravity. A numerical model is constructed in the and at least one accelerometerattached someplace on the bow.
finite element program ANSYS. The calculations consider only That requires a minimum of three signal channels. The JASA article
the elastic properties of the bow (damping is not included). In order describes the use of the method on a violin E string with rigid
to perform automatic optimization of the geometry of the bow, the terminations, which vastly simpkfies the reconstruction of the bow
FEM-program ANSYS is coupled to an external optimization force on the string from the end forces on the string. However, the
routine (the stochastic Simulated Annealing method). sampkng rate using an E string must be high enough that the band
width for each end force signal be at least 50 kHz. That currently
requires special and quite expensive equipment.
A WAY TO MEASURE THE BOW
Robert T. email [email protected] I have submitted an abstract for a paper at the 2003 Stockholm
Department of Physics Musical Acoustics Conference (SMACO3) in August 2003, in which
Carnegie Mellon University I show that one can get sensibly the same results as with the full
Pittsburgh, PA, 15213 USA reconstructed force by using one of the terminationforces and the
http://info.phys.cmu.edu/people/faculty/Schumacher_RT/ accelerometer signal—only two input channels required. The data
is analyzed by using a sonogram method, in which the center address
The bow can be hardly described as peripheral to tone production of a Fourier transform is successively stepped one string oscillation
in stringed instruments. But the physical study of bows has not period through the several hundred periods of a bow stroke. One
even begun to approach the effort, the experimental, and the can then follow the amplitude of the bow response signal, and
theoretical studies devoted to the corpus of stringed instruments. normakze it to the force by the string that drives the bow, either at
Perhaps the reason is twofold. One reason may be that the effects the same frequency, or at any other frequency response peak.
of bows on tone, while universaky bekeved to be important, are in Results show that indeed the bow's response depends on the contact
'fact subtle, and perhaps obvious only to the player. Secondly, the position. The method can be used on a viokn (viola, cello) being
analogous study of normal modes of this seemingly simple object played in the normal way, with the advantage over a bowing machine
should be relatively trivial, and hence perhaps not very interesting. (used in my laboratory) that any pitch on any string can be used, so
Affix a sensor - typically an accelerometer - anyplace on the bow string frequencies can be explored in order to match string harmonic
and tap it anyplace, and from that get a frequency spectrum. Upon frequencies to resonant frequencies of the bow.
further thought, that does not mimic the use of the bow, since the
force is appked to it by the string. So the impulse should be appked I hope that my kmited laboratory explorations wik interest at least
via the bow hair. Perhaps achieving a full useful spectral range that one bow maker to bring his or her knowledge of bows and their
way is not reaky easy - the impulse must be very short. Worse, the peculiarities to this technique. Knowing the questions that must be
string applies the force at varying distances along the hair, and answered is essential to being able to make one's way sensibly
perhaps the response depends on the place of application. But then, through the vast thicket of data that is generated. After feedback
repetitive appkcation of the impulse at different places along the from attendance at SMACO3, I intend to pubksh more details in this
hair takes time, and the appkcation at necessarily discrete places journal, and/or on the web, along with some illustrative figures.
along the hair will occur after the response from the previous This publication would be an elaboration of the conference
proceedings. For seriously interested bow makers I will also provide
at that time the MatLab programs I have written for data analysis.

CASJ Vol. 4, N0.7 (Series II), May 2003 69

♦ JACOB STAINER „... IMPERIAL SERVANT AND VIOLIN- includes violins, violas of different sizes, two violoncekos and
MAKER FROM ABSAM" three viols. A highlight is a violin in original condition from
Special exhibition by the Kunsthistorisches Museum Wien in 1668. Other instruments show Stainer's influence on
Ambras Castle vioknmakers in Austria, Germany, Bohemia, Italy and England
from the late 17th to the early 19th century.
June 2-October 31, 2003
Curator: Dr. Rudolf Hopfner, Director of the Collection of
Ancient Musical Instruments Lenders: Vienna's Kunsthistorisches Museum; National Mu-
sic Museum, Vermilkon, Musee de la Musique, Paris; Re-
gional Museums in Innsbruck, Salzburg, Linz; Private collec-
tors in Austria, Germany and USA.

0 VSA-CATGUT-OBERLIN VIOLIN ACOUSTICS

WORKSHOP
This workshop, chaired by Fan Tao, examines the state-of-the-
art of musical acoustics, concentrating on practical techniques
to aid the instrument and bow maker.
July 7-11, 2003
Oberlin College, Oberlin, Ohio
www.vsa.to

www.catgutacoustical.org

/ 668 violin by Jacob Stainer (National Music

Museum; photo by A. T. King)
m"
#|VETENSKAPI&
iw_

The exhibitionis dedicated to the kfe and work of Jacob Stainer I OCH 1
(c. 1617-1683), the preeminent violin maker of the German-
speaking region. The exhibition will feature instruments by
Stainer, and items that give an impression of his cultural and KTH
social environment and the musical kfe in Innsbruck in the 17th
century. Part of the exhibit will concentrate on new biographi-
cal information about Stainer. More than 120 newly discov- ♦ STOCKHOLM MUSIC ACOUSTICS CONFERENCE2003
ered documents shed light on his trial for heresy, and how (SMAC 03)
Stainer defended himself. August 6 - 9,2003
The Catgut Acoustical Society is pleased to be a sponsor of the
One room will be dedicated to methods of documentation. Stockholm Music Acoustics Conference 2003. Organized by
Simple methods (drawings, templates, measuring devices) are the Music Acoustics Group at the Dept. of Speech, Music and
opposed to more sophisticated (UV-light, dendrochronology) Hearing, Royal Institute of Technology (KTH) Stockholm,
and high-tech (laser-vibrometry, sound analysis). An audio- Sweden.
guide will provide oral information together with music
samples recorded with Stainer's instruments. The catalogue SMC 03 will consolidate the tradition initiated by SMAC 83
will comprise articles of scholarly nature, descriptions, and and SMAC 93 with a major conference on musical acoustics in
color photographs. Articles will be in German with English Stockholm every 10th year. Like earker SMACs this conference
summaries. A CD will be recorded with mainly solo-pieces, will present contributions from the entire field of musical
played on four instruments: two viokns, violoncello, and viola acoustics.
da gamba.
Further information is available at:
A total of 160 items are on display, including 57 musical instru- http://www.speech.kth.se/smaco3/
ments (20 by Stainer). The selection of Stainer's instruments email: [email protected]

70 CASJ Vol. 4, No. 7 (Series II), May 2003

♦ VIOLIN SOCIETY OF AMERICA 2003 CONVENTION ancient capital and the municipal city of Nara Prefecture, lo-
cated in the central part of the main island of Japan. ISMA2OO4
November 6-9, 2003
Timonium (Baltimore), MD will take place March 31 -April 3, just before the 18th ICA in
http://www.vsa.to/nextconvention.htm Kyoto (April 4-9). ISMA2OO4 is a satelkte meeting of the ICA.
The scientific program wik cover a broad spectrum of musical
acoustics such as musical instruments and singing voices, hu-
♦ 146TH MEETING OF ACOUSTICAL SOCIETY OF man perception and psychoacoustics, tone analysis and syn-
AMERICA
thesis, music and computers, and others.
November 10-14, 2003
Austin, TX USA
As with past ISMA meetings, the main objective of the sympo-
http://asa.aip.org/ sium is to promote interdisciplinary communications among
musical acoustics researchers in institutions and industries
♦ 4TH WSEAS INT. CONF. ON ACOUSTICS, MUSIC, worldwide. Papers of recent works, in theoretical, experimen-
SPEECH, AND LANGUAGE PROCESSING tal, and practical domains wik be considered for presentation.
(FORMER ACOUSTICS AND MUSIC: THEORY AND Also, several workshops are planned for special topics. For
APPLICATIONS) (ICAMSL2OO3) information & preliminary registration, please visit ISMA2OO4
December 19-21, 2003 Home Page: http://www2.crl.go.jp/jt/a132/ISMA2OO4/
Udine, Italy,
http://www.wseas.org ♦ ICA2004
1 8th International Congress on Acoustics
♦ BRITISH VIOLIN MAKING ASSOCIATION 2004 April 4-9
VIOLIN AND BOW MAKING COMPETITION, 2004 Kyoto, Japan
March 26-April 4, 2004 http://www.ica2oo4.or.jp
Royal Academy of Music, London
The competition will be held in London during the Genius of ♦ 147TH MEETING OF ACOUSTICAL SOCIETY OF
the Violin Festival. This new two-week festival, based on the
AMERICA
collaborationbetween the Royal Academy of Music and the
May 24-28, 2004
Yehudi Menuhin International Competition for Young Viokn-
New York, NY, USA
ists and presented in collaboration with many Arts organiza-
http://asa.aip.org/
tions, wik be the first European festival devoted solely to the
viokn and its rich variety.
http://www.geniusoftheviolin.org
♦ THE BTH INTERNATIONAL CONFERENCE OF MUSIC
PERCEPTION & COGNITION
August 3 - 7, 2004
♦ ISMA 2004
Northwestern University, Evanston, 11, USA
The Acoustical Society of Japan and the Catgut Acoustical http://www.northwestern.edu/icmpc/
Society will jointly host the International Sympo-
sium on Musical Acoustics. The conference venue is Nara, an

CASJ Vol. 4, N0.7 (Series II), May 2003 71

 Impressions from ISMA'O2
Mexico City, December 9-13, 2002

Knut Guettler

The ISMA'O2 in Mexico City took place at The National Academy there a good way of displaying how the sound is produced in, say, a
of Music (ENM, Escuela Nacional de Musica) in Coyoacan (near trumpet or clarinet?", etc., etc.
to the Frida Kahlo museum). The Academy is wek suited for events
of this kind, and it was nice to observe that music students had been Several inventive solutions to these and similar questions were
invited to attend the presentations, an opportunity many of them presented in the course of the symposium. To mention but two
found appealing throughout the symposium. In that respect the names: Ana Barjau of the Polytechnical University of Catalunya,
symposium's official theme: "Musical Acoustics and an Interactive (Barcelona, Spain) who presented analyses and computer-aided
Musical Instruments Museum" often provided more accessible, less design techniques for wind instrument reconstruction—even for
technical presentations than are normaky experienced at the ISMA missing parts, and Murray Campbell of University of Edinburgh
meetings. Two other pronounced topics were music education, (U.K.) who showed real-time split-screens of waveforms, spectra
and neuroscience related to music perception and imagery. This and envelopes while his instruments were played by artificial lips
widening of ISMA's scope was in many ways reflecting the made from water-filled party balloons! Museums are traditionally
impressive multifaceted qualifications of the two organizers. Mguel facing a tradeoff between the conversation discipline—wanting to
Zenker—who besides being an educated luthier, is professor of keep the instruments unused—and the interests of musicians,
music at ENM, and one of the directors of its Laboratory for Musical musicologists and others who want preserveits sound by maintaining
Informatics and Electronic Music. Eduardo Castro-Sierra is head its playing functionality and a certain accessibility. Several papers
of the Laboratory of Psychoacoustics & Auditory Physiology at were devoted to this problem on a more philosophic basis.
Hospital Infantil de Mexico Federico Gomez, a health institute
providing special education, and medical and surgical care to
children of kmited economic means. Besides being an M.D. with a Our amiable hosts, Miguel Zenker (left) and Dr. Eduardo Castro-
Ph.D. in the Hearing and Speech Sciences, Castro-Sierra is a contract Sierra created a relaxed atmosphere and provided a variedprogram,
professor of Psychoacoustics, Musical Acoustics and Comparative including performances of Mexican music in classical and
Phonetics and Phonology at ENM, with a Ph.D. in music from that traditional styles.
same academy.

The theme "Musical Acoustics and an Interactive Musical

Instruments Museum" does, however, in itself call for expertise
from a number of disciplines, a fact wonderfully demonstrated
through the many creative approaches presented during the week.
An instrument displayed in a museum is potentially an interesting
object for a variety of groups: kids, historians, ethnologists,
musicians, musicologists, luthiers, and acousticians, to name but a
few. All with different needs as far as presentation is concerned.
Even when making the instrument's playing qualities accessible
through replicas, one faces questions kke: "Should the replica be
tuned to a pitch that enables it played together with other
instruments, and if so: old or modern ones?", "To what extend can
we credibly reconstruct the acoustic qualities of a broken or
—
incomplete historic instrument? not to mention an instrument of
which only a written description or visual image still exists?", "Is

72 CASJ Vol. 4, No. 7 (Series II), May 2003

EDITORIAL PROCEDURES FOR PUBLICATIONS sends the paperback to the author with suggestions for modification.
The CAS Journal is a peer-reviewed pubkcation. This means that Once the paper has attained final form, it enters the queue for space
the Editor wik seek the opinion of people knowledgeable about the in one of the next issues of the CAS Journal.
topic and quakfied to review the paper.
COPYRIGHT
If a paper is not deemed suitable for pubkcation in the CAS Journal, Submitting a paper for editorial consideration is a representation
the Editor wik so advise the author and explain the reasons. If the that the paper has not been copyrighted or published elsewhere,
paper is acceptable in principle but changes are needed, the Editor and is not under consideration for publication elsewhere.

GUIDELINES FOR SUBMITTING PAPERS kne drawings are acceptable; low-resolution graphic images copied
We invite submission of technical papers, short notes, book reviews, from web pages and low-quality photocopies of illustrations are
obituaries, letters, and comments. Articles should not exceed 6000 unacceptable. Graphic images should be formatted to near standard
words, equivalent to approximately 6 Journal pages, including column widths (3.5 or 7.0 inches) and saved or scanned at high
graphics. Book reviews and notes should be kmited to 1000 words. (300 dpi) resolution, preferably in .tiff format. Please include
Please consult a recent copy of the Journal for guideknes as to style. graphics files on disk separately from text; do not imbed images in
word-processing documents such as Word or Word Perfect.
Initial submissions can be in the form of hard copy, fax, 3V2—inch
floppy disk, CD-ROM, or email attachment. Please note that final Please send 3 paper copies of submitted articles to CAS Office, 55
copy must be provided as a TEXT file (.txt, text only) rather than as Park Street, Montclair, N.J. 07042 U.S.A., fax (973) 744-0375,
a native file from a word processing program. However, hard copy e-mail [email protected]. The Journal Editor can be contacted at
should be provided that reflects all special characters and desired [email protected].
formatting, superscripts, etc. so that our printer can accurately
reproduce those features. Clean black and white photographs and

OFFICERS TRUSTEES ADVISORY COUNCIL INTERNATIONAL

President Gregg T. Alf Dennis Flanagan VICE PRESIDENTS
Julius VandeKopple Pamela J. Anderson Frank Lewin Canada: Daniel Ling
Executive Vice President Christopher Chafe Gabriel Weinreich France: Voichita Bucur
Joseph Regh Sam R. Compton,
Compton, JJr. Germany: Volkmar Tetzner
Treasurer Joseph Curtin COMMITTEES Hong Kong: Anton Sic
Paul B. Ostergaard Knut Guettler Research Italy: Domenico Stanzial
Secretary A. Thomas King Oliver E. Rodgers Japan: Isao Nakamura
Elizabeth McGilvray Daniel Ling Journal Netherlands: Adrian M. Houtsma
Permanent Advisor Jeffrey S. Loen
Jeffrey Lo Gregg T. Alf, Associate Editor Scandinavia: Anders Askenfelt
Carleen M. Hutchins Okver E. Rodgers Virginia Benade, Associate Editor UK: Jim Woodhouse
General Manager John P. Soloninka
Soloninka Evan B. Davis, Associate Editor
Deana R. Campion Gary Scavone Robert T. Associate Editor
Robert T. Schumacher Catgut Musical Acoustics
Fan-Chia Tao Research Library
Joan E. Miller