F or almost a century, dendrochronology has been an essential tool in many fields of scientific research. The method of dating wood by studying tree-ring patterns has been employed by archaeologists, geographers and even astronomers, while historians

and art critics have adopted the science to find the ages of ships and panel paintings. It is only quite recently, however, that the method has been used for the dating of stringed instruments – even though, with a little detective work, the method can reveal much more than just the age of the wood. Increasingly, researchers are discovering fascinating correlations and relationships between various instruments and schools of making, as well as significant insights into the workings of the old masters.

Working with my colleague and fellow violin maker John Topham, I took Andrea Amati’s c.1566 ‘Charles IX’ violin as the starting point for the research. One of four surviving Amati instruments made for the French king’s court, the decorated violin is currently on display at the Palazzo Comunale in Cremona. In 2009 we obtained permission from the exhibition’s curator, Andrea Mosconi, to take a series of macro-photographs of the violin for a full, thorough dendrochronological study. Our investigation unearthed a number of remarkable surprises.

Tree rings are made by the living cells that grow underneath the bark. As they multiply, the cells form a concentric tree

The Tree of knOWledgeA blend of science and sleuthing, dendrochronology can reveal unsuspected facets of the maker’s craft. Peter rAtcliff explains how examining the tree rings on an Andrea Amati produced surprising results

ring each year, gradually enlarging the size of the trunk. The obvious boundary between dark and light areas results from an increased concentration of cells forming towards the end of each growing season, which usually terminates fairly abruptly. The distance between rings fluctuates according to many factors, environmental and otherwise (such as the age of the tree), and the relative size of each consecutive ring forms the basis of the process of ‘crossdating’ – a much more reliable method than simply counting the rings. Crossdating involves taking a sample of tree-ring measurements and comparing them, ring for ring, with samples from trees that grew in the same geographical region and time period. Hence, this method can determine the age of the wood to the precise year.

during our analysis, we discovered numerous correlations between the wood used in the ‘Charles IX’ and the data obtained from the front of another Amati instrument, also part of Charles IX’s set – the decorated viola in the Ashmolean Museum in Oxford. In fact, a comparison of the statistical and graphical results for both violins (figure 1) pointed to a ‘same-tree match’.

FIGURE 1 The close correlation between the tree rings on the ‘Charles IX‘ violin and the Ashmolean Museum’s Amati viola suggests the wood for both instruments came from the same tree

DENDROCHRONOLOGY

www.thestrad.com32 THE STRAD february 2012

We determined that the latest ring on the lower treble side of the ‘Charles IX’ grew in the year 1491,

while the one at the outer bass edge dated to 1493. With the high probability that both the c.1566 violin and the

c.1564 viola came from the same tree, we could then speculate that the earliest possible felling date for the tree (the ‘terminus post quem’) closely followed the date of the latest ring of the viola, which was 1500.

this exhilarating ‘same-tree match’ prompted us to examine closely the wood from other instruments

made by Andrea Amati. We gathered as much data as we could from instruments scattered throughout the world. To our delight, we soon

found a further three Amati violins beautifully fitting the existing pattern. The first we came

across was an undecorated example from c.1574, known as the ‘Payne’, ‘gonley’, owned

by the Chi Mei Culture Foundation.We soon also discovered that the grain

pattern of the ‘kurtz’ Andrea Amati violin, in the Metropolitan Museum of Art in new York, appeared to have similar visual characteristics to the others. Preliminary analysis confirmed excellent statistical crossdating results and graphed pattern

comparison. The ‘kurtz’ is another decorated violin, and may have belonged to a set that

was made for the marriage of Charles IX’s sister, elisabeth of Valois, to Philip II of Spain.

As that marriage took place in 1559, it would place the instrument around that time – making

it one of the earliest surviving violins in existence.

Andrea Amati’s c.1566 ‘Charles IX’ violin was the starting point for the research

String

S Co

lleCtio

n o

f the Pa

lazzo

Co

mu

na

le, Crem

on

a, ita

ly

DENDROCHRONOLOGY

february 2012 THE STRAD 33www.thestrad.com

The Metropolitan Museum has adopted this attribution, originally put forward by violin historian and maker Carlo Chiesa in 2007, having compared the violin’s decoration with that of another instrument from the period.

Another Andrea Amati instrument from the Chi Mei collection, known as the ‘ross’ and dated to c.1570, is an undecorated violin with a single-piece front. Totally unexpectedly, it too tested positive for a same-tree match with the others. Importantly, this one-piece front showed several extra rings that overlapped all the others, giving us a date of 1514. This latest ring date could therefore apply to all the instruments, and establish a new terminus post quem for the tree.

We then compared the statistical results for each of the five Andrea Amati instruments (figure 2). each line on the graph represents the ‘mean series’ (which combines the data for the bass and treble sides) of each instrument, except for the longest line at the bottom, which represents the one-piece front of the ‘ross’. Second up is the Ashmolean viola, followed by the ‘kurtz’, the c.1566 Charles IX, and the ‘Payne’, ‘gonley’ at the top. The excellent coincidence linking all five graphs was the concluding factor, confirming to us that a single tree was used for the bellies of the five Amati instruments.

during the course of the research, which took place over the past 18 months, several other instruments by Andrea Amati were tested or reassessed. Some formed part of the set made for Charles IX, including specimens from the collection at the national Music Museum in Vermillion, South dakota, the small ‘Charles IX’ violin housed in the Tullie House Museum in Carlisle, and the violin from the Ashmolean collection. Although a few dendrochronological relationships were established between some of their woods, none was found to relate directly to the tree from the matching quintet.

The Andrea Amati viola from c.1564, held by the Ashmolean Museum in Oxford

The distance between tree rings fluctuates according to many factors, environmental and otherwise

FIGURE 2 From top to bottom: the statistical results for the ‘Payne’, ‘Gonley’; the ‘Charles IX’; the ‘Kurtz‘; the Ashmolean viola; and the ‘Ross‘

aSh

mo

lean

mu

Seum

, un

iverSity o

f oxfo

rd

34 THE STRAD february 2012 www.thestrad.com

One of the oldest surviving violins in existence: the c.1559 ’Kurtz’

Having determined a close connection between these five instruments, we examined their methods of construction, beginning with the way the plates were oriented. When a log is split, it creates ‘cake-like’ wedges, with the larger outer section below the position of the bark. That area eventually forms the gluing surfaces for the instrument’s centre joint, which means the youngest tree rings will normally be situated there. This is the case for both the Ashmolean viola and the ‘kurtz’ violin. In the case of the ‘Charles IX’ violin, however, we have the opposite – the rings are closer to the centre of the tree, ending up at the centre joint.

The plates of the ‘Payne’, ‘gonley’ are arranged in another different fashion: the growth direction of the bass side runs from the centre joint outwards, while the treble side runs towards the middle. We also noted the presence of a jointed wing, about 25mm wide, on the lower treble bout. Finally, the front of the ‘ross’ – made from a single piece of wood – displays yet another method of working.

It would seem that Andrea Amati, having made five instruments with wood from the same log, did not think it essential, acoustically or visually, to follow strict book-matching practices as later became the norm. Alternatively, one could speculate that the maker was experimenting, trying almost every

Crossdating the wood of the ‘Charles iX’ violin

Crossdating involves comparing the tree-ring measurements of the wood under examination with a database of woods that have previously been examined — particularly those thought to be from the same region and time period. Nowadays, correlations are frequently detected through the use of statistics, which can identify possible relationships or matches and their level of importance. The correlation is often expressed by the‘Student’s t-value’ calculation: the higher the t-value, the more statistically significant it becomes, often achieving scores of twelve and higher for wood samples from the same tree.

a second valid method of comparison is the Gleichläufigkeit or ‘percentage of parallel agreement’. This is based on how closely the trees’ growth is matched year by year. for example, if sample a shows a tree ring that is wider by one year compared with the previous year, while sample b shows a ring that is narrower, the Gleichläufigkeit score goes down.

Our analysis showed an excellent correlation between the ‘Charles IX’ violin and the ashmolean Museum’s amati viola. These results were further confirmed by strong statistical evidence obtained with published references from Switzerland and austria, as well as correlating data with several instruments from the 16th century.

The statistical tests between the combined data from the two sides of the amati violin, against that of the treble side of the viola, yielded a highly significant t-value in excess of t=15.5. The Gleichläufigkeit was also extremely high, at 81.25 per cent. The treble sides achieved an exceptional, almost unheard of, t-value of over 19.

While indicating the alpine circle as the general area, the correlations with the Swiss and austrian references may be incidental, and do not automatically determine precise wood provenance. reference chronologies do not yet cover all areas, mainly owing to the lack of living trees that are old enough to compile such references. references for spruce from the 15th and 16th centuries, though scarce, are constantly improved by the addition of data from early instruments.

Co

urt

eSy

met

roPo

lita

n m

uSe

um

of

art

, new

yo

rk

The tree rings are clearly visible in this close-up of the c.1566 ‘Charles IX‘

peTe

r r

aTC

lIff

DENDROCHRONOLOGY

www.thestrad.com february 2012 THE STRAD 35

available combination. The occasional addition of wings, as in the ‘Payne’, ‘gonley’, suggests it may have been more important to him to make the best (and most economical) use of the timber available to him, rather than creating fronts of perfect symmetry – even though some of the instruments were made for royalty.

It also appears that some of the plates have been flipped over lengthways. This may have been done to avoid a defect

It seems Andrea Amati did not think it essential to follow strict book-matching practices

in the wood, such as an unsightly knot, being too prominent. The treble side of the ‘kurtz’ shows wavy grain toward the lower bout centre joint. The c.1566 violin shows a similar wave at the upper bass edge, and a slight wave at the lower bass bout edge. The Ashmolean viola has this similar wavy grain at the upper treble towards the centre. The presence of this slight defect in the grain, at the same period, strongly suggests that these particular plates were very close to each other within the log.

it is also possible that Amati had a whole section of a log, and that he made many more instruments with it than the five we have been able to locate so far. Many of the instruments made for Charles IX have sadly been lost, and we also know that not all surviving examples from that set were made from this  particular tree. Although some strong relationships were found between timbers from other instruments, no evidence of the use of another single tree has emerged.

Another interesting aspect of this research is the length of time for which this particular tree was used. If the ‘kurtz’ violin was indeed made prior to 1559, and the ‘Payne’, ‘gonley’ in 1574, the various wedges from the log would have been used over a

The most recent of the five instruments: the ‘Payne‘, ‘Gonley‘ violin from c.1574

Co

urt

eSy

Ch

i mei

Cu

ltu

re f

ou

nd

atio

n

DENDROCHRONOLOGY

www.thestrad.com february 2012 THE STRAD 37

minimum period of 15 years. With a terminus post quem of 1514, it could presumably have been used at any time after about 1520. Statistically, five of the thirteen surviving violins and violas being made from the same log raises interesting questions, and might suggest that Andrea Amati did not make that many instruments over the course of his career. The haphazard variety of plate orientation and their physical attributes – some with wings, squiggles in the grain, uneven growth – seem to point to a tree that was barely suitable for violin making, at least by modern-day standards. This might seem surprising, but it is easy to forget that we are talking about the first ever maker of violins. However, it raises questions regarding the choice and availability of suitable spruce tonewood at that time.

The conclusions drawn from our examinations, and the questions they raise, provide an insight into a part of violin making that has remained almost untouched by previous researchers. The connections and coincidences we unearthed prove the usefulness of dendrochronological analysis in this field. We hope our results will be universally accepted as scientific and conclusive material, though it remains to be seen whether they will encourage future exhibition curators to include serious dendrochronological research as part of their core investigation.

to read John dilworth’s analysis of the ‘Charles iX‘ from the June 2005 issue of The Strad, subscribe to the strad archive at www.thestrad.com/StradArchive.asp

how was the log split?Figure 3 shows one possible explanation of how the various plates were arranged within the log. Three sets of plates may have been cut or split from a thick, quarter-sawn or split block. The piece labelled a is split further to make the c.1566 violin front, while piece b is used for the ashmolean viola, and piece C to make the c.1574 ‘payne’, ‘Gonley’.

The centre joints on a and b find themselves at opposite ends of the growing period of the tree. pieces C1 and C2 will also have their centre joints at opposite ends from each other. The a1 and b1 plates, both used for the treble sides, would then show a particularly good statistical correlation. The statistical results between the various plates of both violins and the viola would suggest that they were probably in that position within the log.

It is possible that andrea amati considered the thinner end of the C2 plate to be inadequate (possibly due to the grain being too wide or too far off the quarter), and decided to replace it with a wing, now visible at the lower treble bout of the c.1574 ‘payne’, ‘Gonley’ violin.

FIGURE 3 How the plates may have been arranged in the log

Only the c.1570 ‘Ross’ violin has a one-piece front

The examination raises questions regarding the choice and availability of spruce tonewood at that time

Co

urt

eSy

Ch

i mei

Cu

ltu

re f

ou

nd

atio

n

DENDROCHRONOLOGY

38 THE STRAD february 2012 www.thestrad.com