In 1994 the Metropolitan Museum of Art in NewYork staged an exhibition of works by the Cremoneseviolin maker Giuseppe Guarneri del Gesù. No equiv-alent representative collection of del Gesù’s instru-ments has ever been assembled in one place at onetime, and the opportunity this exhibition affordedfor comparative study was unique. In the past, theexamination of Guarneri’s instruments has been lim-ited to small numbers, on rare occasions. Even delGesù’s own workshop is unlikely ever to have con-tained a similar collection of violins spanning hiswhole career.

Gathering data about del Gesù’s life and works hasalways been something of a haphazard affair. Histor-ically, the greatest difficulty has been a lack of con-sistency: for any information to be useful, themethod of obtaining it and the conditions underwhich it is obtained need to be the same for every in-strument. The long-established use of photographyprovides a good illustration of the problems involved.In order to create a consistent photographic record,the same camera, lens, film, lighting and set are re-quired for each photograph. At a conservative esti-mate, it would probably take thirty to fifty years fora group of instruments comparable to those exhib-ited at theMetropolitanMuseum to pas through even

the busiest shop. Over such a long period of time itwould be impossible to maintain the necessary qual-ity in any photographic archive. Similar restrictionsapply to all other methods of recording and storinginformation. In this study, perhaps for the first time,consistency was the hallmark of the data collected.

Although it has long been recognised that individ-ually, photographs, outlines, descriptions and meas-urements have their shortcomings, collectively theycan be extremely informative. In order to obtain themaximum advantage from the following survey, it isnecessary to understand something about how thedata was compiled and what the strengths and weak-nesses of the various methods are. Because it is im-possible to reproduce the three-dimensionalcharacter of an instrument in two-dimensional formwithout compromising the end product, a completeset of technical drawings was never a consideration.Instead, the intention was to reproduce the availableinformation as it was collected, in as pure a form aspossible. Particularly in the case of the drawings, cer-tain features were omitted or were deliberately leftincomplete.1 Any personal interpretations of thesemissing details would have been misleading.

All the measurements in this survey were taken

11

TThhee WWoorrkkiinngg MMeetthhooddss ooff GGuuaarrnneerrii ddeell GGeessùù aanndd tthheeiirrIInnfflluueennccee uuppoonn hhiiss SSttyylliissttiicc

DDeevveellooppmmeennttText and Illustrations by Roger Graham Hargrave

Although the greatest care has been taken while compiling this site it almost certainly contains many mistakes. As such its contents shouldbe treated with extreme caution. Neither I nor my fellow contributors can accept responsibility for any losses resulting from information or opin-ions, new or old, which are reproduced here. Some of the ideas and information have already been superseded by subsequent research and de-velopment. (I have attempted to included a bibliography for further information on such pieces) In spite of this I believe that these articles arestill of considerable use. For copyright or other practical reasons it has not been possible to reproduce all the illustrations. I have included thetext for the series of posters that I created for the Strad magazine. While these posters are all still available, with one exception, they have beenreproduced without the original accompanying text.

with a caliper. Although caliper measurements areextremely accurate, the method has certain draw-backs when applied to violins. Just as mean averagesare not accurate measurements, accurate measure-ments are not mean averages. The measurementswere accurately taken from specific points. It is thepoints themselves which are subjective. Throughoutthis analysis, it is possible that even the slightest de-viation from these points would have resulted in dif-ferent readings.

The outlines of the instruments were taken usinga specially constructed box which contains a recessfor the arching. A non-shrinking transparent plasticdrawing foil was taped to the box and the outline wasdrawn around the instrument from the rib side of theback. The resulting outline is a mirror image of thatwhich is seen when the back is observed in the nor-mal way. This anomaly has been corrected for thebook. For technical reasons, no belly outlines weretaken.

Because the edges of the instrument are roundedand in almost every case the back plates are slightlytwisted, there may be some minor inaccuracies in theoutlines, especially in the centre bouts and at the cor-ners. Nevertheless, aberrations caused by the cam-era lens are probably greater than those present inthe outline drawings, especially in the curves of theupper and lower bouts. The outline drawings and thephotographs should be carefully checked against thecaliper measurements, which in the case of the upperand lower bouts were taken at the widest point. Inthe centre bouts, the narrowest point was used. Thelength of the back was measured from the base sideof the button to the centre of the lower bout curve.The belly length was taken from the treble side of theneck to the lower saddle. Variations between thecaliper measurements and the drawings are usuallyless than 0.5 mm over the length of the instrument.Where discrepancies occur, the photographs shouldbe carefully consulted, especially in the centre boutsand at the corners; otherwise, the outline should beconsidered pre-eminent.

Where circumstances allowed, a number of extrabelly arching templates were included. These weretaken on either side of the fingerboard and wheneverthe fittings could be removed, the central long archwas also recorded up to the end of the fingerboard.Close examination of the arching templates revealsthat the long arches, including those running along-side the fingerboard, do not always match theheights of the corresponding cross-arches. The ex-

planation for this is simple. The plates and the ribsto which they are attached were often warped andtwisted and this warping was occasionally quite ex-treme. Unfortunately, in the time available it provedimpossible to establish how much warping had oc-curred, and in which direction it ran. In most cases,the arching templates were recorded using a highlymodified form copier with pins of 0.8 mm. In general,the results obtained are good, but because of thecomplex nature of the edges, it was not possible toreproduce their profile with suitable accuracy. Forthis reason the edges have been indicated by a dottedline. Readers will need to refer to the descriptive text,the accompanying photographs and the measure-ments in order to obtain a more complete picture ofthe edge profile.

Particularly troublesome was the establishment ofmeasurements relating to the purfling. The distanceof the purfling from the edge was measured with acalibrated lens. For this purpose, a point judged to berelatively unworn was selected on the upper or lowerbouts. From the photographs it can be establishedthat this distance often varies widely, even on thesame bout. Worn edges usually account for most ofthese irregularities, but deviations in the tighter cor-ner curves and inconsistencies in the channel widthscan also contribute. Sometimes the distance of thepurfling from the edge is indicated by two measure-ments: one represents the greatest distance and theother the smallest. In each instance these measure-ments were taken from a point on the back where theedge appeared reasonably pure. As can be deducedfrom the given measurements, the thicknessing ofthe whites was fairly consistent. The blacks were dif-ficult to measure. This was partly because the indi-vidual strips varied slightly in thickness, even on thesame instrument, and partly because the channelsinto which they were set were often unevenly cut.The resulting gaps are filled, either with dirt, somekind of filler or varnish, or occasionally even tinyslivers of wood. Fortunately, the photography is ofsuch a high calibre that these details can generallybe discerned. From the photographs, a close exami-nation of the line of the purfling can help the readerestimate how much the original outline has beenworn, particularly on the belly.

The diagrams of the sides show the depth or thick-nessing of the edges. Occasionally, partial or com-plete doubling of the edges has diminished the valueof these measurements, especially on the bellies.Such technicalities are pointed out in the text or canbe discerned from the side photographs. With refer-

22

33ence to the tight upper and lower curves of the cen-tre bouts, the descriptions of the edges are of partic-ular importance. Not only are these curvesforeshortened in the photographs, they are also par-tially hidden by the corners.

Only four of the exhibition instruments have but-tons which are completely original: the “Kreisler”,the “King Joseph”, the “Kochánski” and Paganini’s“Cannon”. The dimensions for these buttons can befound in the tables and on the diagrams of the sides.In most cases, these original buttons are clearly ta-pered, the thickest point being furthest away fromthe purfling and edge.

In spite of some local deviations, the back over-hang measurements are fairly reliable. However, theback corners presented something of a problem. Thealmost vertical cut-off angles of some corners madeit extremely difficult to fix a point from which to takethe corner overhang measurement. This was finallytaken from the rib end to what was judged to be theapproximate centre of the cut-off line (Figure 1).Worn corners created extra difficulties in establish-ing this central position. In a few cases, the wear wasso extreme as to render these measurements practi-cally worthless.

Paganini’s “Cannon” is the only instrument forwhich the belly overhang measurements wererecorded. Such measurements are normally unreli-able, because bellies are removed more frequentlythan backs and their edges and corners are generallysubjected to more wear and tear.

The back, belly and rib thickness measurements

were taken with a “Hacklinger” magnetic caliper,which is accurate to within 0.1 mm. This process wascarried out with the instruments closed. In somecases, where the fittings were removed, it was possi-ble to take measurements in the area normally cov-ered by the tailpiece. In a few instances extrameasurements were also taken around the centre ofthe back. Unfortunately, no account could be takenof any patches in the plates, nor was it possible to es-tablish whether any instrument had been rethick-nessed. The presence of a label caused the caliper togive a slightly fuller reading and this should be takeninto consideration. (The “Ysaÿe” has an extra labelon the back which is noted in the text.)

Whenever a central pin was present and visible, itsposition is recorded, as is the plate thickness at thepin position. Further information about this pin canbe found in the text relating to individual instru-ments and in the section regarding del Gesù’s archingand thicknessing methods (p. 147-8). On many in-struments, the central pin can be seen on the photo-graphs emerging as a tiny dot in the central regionof the back.

The rib thicknesses were taken in the central areaof each bout. They are shown on the side diagrams.Although these dimensions are precise they may notpresent an entirely reliable picture. The centre boutribs of several of del Gesù’s violins have been thinnedto 0.5 mm or even 0.3 mm where they come into con-tact with the corner blocks (see p. 134). Because ofthe presence of the blocks, it was impossible to ex-amine the rib thicknessing in this area. Del Gesù fin-ished the inside surface of the ribs with acoarse-toothed plane iron, creating irregularitieswhich may also have affected the reading. The ribheight measurements can take no account of anylowering or raising of the ribs which may have takenplace, nor do they show where the ribs have been dis-torted by pressure from the chinrest, as is the casewith the “Kochánski”.

The soundhole measurements were taken with acaliper. Centuries of careless soundpost setting, wingcrack repairs, arching distortions and rare alterationof the soundhole nicks all need to be consideredwhen consulting these measurements. To obtain themost accurate picture, the photographs, the meas-urements, and the descriptions should be used incombination. The soundhole measurements do notnecessarily correspond with the dimensions of thesoundholes as they appear on the photographs: thisis because the camera angle has caused the sound-

Figure 1. Del Gesù’s rib corners usually sit centrally on the plate corners; the overhang measurement

was taken from this central position.

44hole bodies to appear slightly narrower. For technicalreasons, no facsimiles of the soundholes have beenincluded.

The outline drawings of the heads may requiresome explanation. The drawings represent the trebleside. With a few exceptions, wear to the heads, espe-cially in the area stretching from below the A peg tothe top of the scroll, has resulted in outlines whichare slightly smaller than they would have been orig-inally. In some instances this area was so worn thatno discernible outline exists. Quite apart from gen-eral wear to the head, around the scroll del Gesùoften undercut the turns. This and his propensity toleave tool marks created problems when positioningthe measuring calipers. Some consistency wasachieved by taking the measurements as close to theouter edge as possible, while remaining inside thechamfer and any worn surfaces (Figure 2). On thehead diagrams (Figure 3) measurement A is the nar-rowest point of the scroll. Measurement B marks thewidth of the back of the scroll, level with the throat.C and D mark the widths at the top and underside ofthe scroll turns, level with the centre of the eye.Measurement E is only recorded in the case of Pa-ganini’s “Cannon”; in every other case a new neckhas been grafted into the head.

The head portraits demonstrate explicitly theproblem of photo graphic distortion. If the side viewsof the scroll are projected across to the front and

back views, it can be seen that the turns do not al-ways correspond. Anyone wishing to reconstruct thehead must take all of the above factors into accountand attempt to overcome any discrepancies by com-bining and comparing all the available information.

TThhee AAmmaattii MMeetthhoodd

It is generally accepted that the Amati family cre-ated and developed the designs and constructionalmethods which classical Cremonese makers followedfor more than two centuries. Andrea Amati and histwo sons, Antonio and Girolamo I, were the first andsecond of four generations of Amati luthiers, span-ning Cremona’s entire period of classical violin mak-ing. Nicolò Amati, the son of Girolamo I, was perhapsthe greatest maker of the family and from an earlyage, Andrea Guarneri, the first of the Guarneri fam-ily of violin makers, is recorded as living and workingin his house. Andrea was the father of GiuseppeGuarneri, known today as Giuseppe Filius Andreæ,who was in turn the father of Bartolomeo Giuseppe,called Giuseppe Guarneri del Gesù.

Effectively an unbroken line of instruction existedbetween the Amatis and Guarneri del Gesù. From acareful examination of their instruments, there canbe little doubt that the method and techniques em-ployed by the Guarneri family were essentially thoseof the first Amatis. In this they were not alone: withinthe city of Cremona (and even beyond), successive

Figure 3. Head measurements.

Figure 2. Positioning of the calipers as they were applied to the turns of the scroll, inside the

area of the chamfer and ware, but outside the undercutting, or overcutting.

55generations of makers gradually extended and re-fined the early Amati designs until, in the first halfof the eighteenth century, their influence was barelyrecognisable. Nevertheless, concealed beneath a ve-neer of stylistic detail, the Amati designs flourishedand their basic rules of construction remainedlargely unchallenged and unchanged. Althoughchronologically and stylistically the early Amatis andGuarneri del Gesù represent extreme ends of the Cre-monese spectrum and almost two hundred years sep-arated their works, the key to understanding delGesù’s instruments lies in understanding the essenceof the Amati system. Fortunately, because the Amatimethod was universal in Cremona, clues to del Gesù'smethod can be found on all Cremonese violins.

It may be that a far greater affinity between themethod of the Amatis and that of Guarneri del Gesùwould become apparent, were it possible to examinetheir instruments in original, “baroque”2 condition.Unfortunately no full-sized violin by either makerhas survived unmodified. Some instruments, such as

del Gesù’s “Cannon” and “Alard”, are of value in thiscontext, as are the pochettes of 1735 and 1740, butthe evidence they provide is incomplete.3

Apart from the instruments themselves, the onlysubstantial information about baroque violin makingwhich has survived to the present day is containedin the enormous collection of tools, drawings, mouldsand templates housed in the Museo Stradivariano inCremona.4 Among these relics are complete sets ofworking drawings, templates and moulds made byStradivari for the construction of both tenor and con-tralto violas (Figure 4).These would have enabled himto mark out an entire instrument using no more thana pair of dividers, a straight edge and a marking tool.It is probable that similar drawings and templates ex-isted for each type of instrument made in the Stradi-vari workshop, but sadly no other complete set hassurvived. The majority of the artefacts in the MuseoStradivariano can be traced back to Antonio Stradi-vari himself, and as a result we know more about hisworking methods than we do about any other his-

Figure 4. Reproduction of the surviving drawings, templates and mould for a Stradivari contralto viola.

The Working Methods of Guarneri del Gesù

torical maker.5 Stradivari’s own reliance on theAmati tradition makes this information helpful, andwith some circumspection it can be applied to theanalysis of del Gesù’s violins.

TThhee SSttyylliissttiicc DDeevveellooppmmeenntt

While the Amati system, with its use of an insidemould, was fundamental to del Gesù’s method, sty-listically he owed much to his teacher, who was al-most certainly his father, Giuseppe Filius Andreæ. Hewould have provided del Gesù with the essentialgrounding in the classical tradition of Cremoneselutherie. From these beginnings, del Gesù’s workevolved dramatically in style and effect during hiscomparatively short working life, earning him a rep-utation as an artist and innovator in his own right.

Giuseppe Filius Andreæ was in some ways as idio-syncratic a maker as his son. In his early work, datingfrom around 1690, he was strictly bound by Amatiprinciples. After about 1710 a gradual appreciationof Stradivari’s ideas is accompanied by a looser tech-nique, derived at least in part from a freer and morespontaneous use of the mould. The general form ofthe outline shows frequent deviations of line from in-strument to instrument, and considerable asymme-try between front and back. Like most of hiscontemporaries, del Gesù’s father seems to have con-tinued the Amati practice of making both a large anda small model (the difference being mainly one ofbreadth rather than length).

Increasingly unproductive during the latter yearsof his career, Giuseppe Filius Andreæ appears to havewithdrawn almost entirely from violin making someyears before his death in 1740. His late work is diffi-cult to date accurately due to the rarity of authenticlabels but, as far as can be judged, it seems to haveundergone several changes after 1715 (the year of hisfinancial crisis) which are generally attributed to theassistance of his youngest son. These changes includea slightly bolder form to the edgework and sound-holes, and a more delicate finish to the heads. Pietro,the eldest son of Giuseppe Filius Andreæ, may havebeen active in the Guarneri workshop until 1717, butby the end of that year he had left Cremona to estab-lish himself in Venice. significantly, the slight butclearly discernible differences in style which occa-sionally appear in the late works of the elderGiuseppe remain consistently observable in the earlyexamples of del Gesù’s independent work.

Most important among the innovations for which

del Gesù was probably responsible was in the treat-ment of the arching. The work of the elder Giuseppeis quickly recognisable in the slightly pinched form ofthe central area of front and back, with a compara-tively wide channel cut around the edge that isclearly derived from the ideas of Nicolò Amati. It isremarkable how early in his career del Gesù rejectedthis type of arching and, albeit in his own distinctivestyle, adopted the lower, flatter modelling of Anto-nio Stradivari. Evidently the young violin maker waswell aware that the Guarneris’ leading competitorhad established his technical superiority largely as aresult of his fresh ideas on arching.

There are only a few reliable markers by which todate the contribution of the younger Giuseppe to theGuarneri workshop production. The near absence oforiginal labels in the instruments of this period posesan almost insurmountable obstacle. None of the pu-tative early works of del Gesù bears an original label,and the instruments considered to be jointly attrib-utable to father and son bear only the label of the fa-ther, if any. However, a group of the elder Giuseppe’sviolins, one with a rare Giuseppe filius Andreæ labelof 1719, have a particularly distinctive outline. Themiddle bouts are comparable with the small pattern,but the upper and lower bouts retain the full width ofthe larger model, and the upper corners projectwidely outward. (This style was also preferred byCarlo Bergonzi for his instruments.) The archingsspring full from the edges, with no inward scoop pastthe line of the purfling, to arrive quickly at a rela-tively low summit. The edges themselves arestronger than before, the line of the purfling set a lit-tle further in, and the soundholes are wider apart andcut more boldly. The soundhole wings are broader, inimitation of the style of Stradivari. Some or all ofthese changes may have been initiated by theyounger Giuseppe.

By 1722 del Gesù had left his father’s workshop,bringing to an end their first and perhaps closest pe-riod of collaboration. As a result of his departure,production in the family business slumped, never torecover. Del Gesù himself seems to have produced fewinstruments in this the first stage of his independentcareer (which began as early as 1721), but those as-cribed to him share several distinctive features. It isclear, for example, that he abandoned the narrow-waisted outline. The “Dancla” can be taken as repre-sentative of his model during this transitional period.The outline is slightly extended in the upper bouts,yet is not incompatible with some earlier violins ofGiuseppe filius Andreæ. Del Gesù’s working of the

66

scroll is quite different from that of his father. Al-though the “Dancla” has only an imitation label,dated 1726, this was generally agreed to be an ap-propriate attribution until dendro chronologicalanalysis established that 1726 was the earliest possi-ble date for the violin’s construction, and that it wasmore likely to have been made a year or two later (seep. 161).

Unfortunately, little is known of del Gesù’s where-abouts or activities between 1722 and 1728, when hetook up the lease of the Osteria dei Mori. Moreover,there is no recorded label dated prior to 1727 bearinghis name. The evidence of labels, where they do sur-vive, is perhaps the most crucial. For example, whiledel Gesù is not usually credited with having made anycellos, two such instruments exist bearing very lateexamples of his father’s label. The first is dated 1729,and is for the most part fairly typical of Giuseppe fil-ius Andreæ. The back is certainly comparable withothers he had made previously. It is scooped indeeply from the edges, extremely high and stifflysculpted to the usual pinched apex. In contrast thefront is low, with flat curves springing directly fromthe purfling. It is entirely different in concept to thatof the poplar wood back.

The Cremonese method of construction requiredthe back arching and thicknessing to be virtuallycompleted and for the back to be mounted upon thehead/neck/rib structure before any work on thebelly commenced. It is therefore conceivable thatGiuseppe filius Andreæ could have completed thehead, neck, ribs and back of a cello which was fin-ished by the addition of a belly by del Gesù, withsoundholes typical of his work during this period.(Even the belly outline could have been finalised bythe elder Giuseppe without affecting the eventualarching shape or the cut of the soundholes.) The var-ious characters of the archings would not have beenaltered in any significant way, even if the purfling ofthe front and back were completed at differentstages. The label was inserted before the belly wasfixed, and possibly before work on the belly had evenbegun. The 1729 cello, with its mixture of styles, plainpoplar back and beech ribs and scroll, provides a keyfor the dating of several extant violins which havethe appearance of hastily finished work, made fromrather poor and mismatched materials.

1731 is the date of the last label of the elderGiuseppe, and the earliest surviving example of delGesù’s own IHS label. The former appears in a secondcello, which seems to be entirely the work of del

Gesù, with his characteristic archings evident onboth plates. The scroll of the cello is also of differentworkmanship, matching those of del Gesù’s early vi-olins, in particular the “Kreisler”. The proportionsare quite different from any previously known workby the elder Giuseppe: the centre bouts are shorterthan usual, and this alone indicates the use of a dif-ferent mould. Despite the claim on the label, thewhole marks a complete break from the earlier styleof Giuseppe filius Andreæ, who was a fairly proliWcmaker of celli. It is perhaps surprising that havinggone to the effort of acquiring another mould, as faras we know, del Gesù never made use of it again. Thechange of design may indicate that although he waswilling to fulfil a commission for his father, who wasprobably by this time too frail to undertake the phys-ical effort of building a cello, he was neverthelessworking independently of him.

Del Gesù’s own label of 1731, bearing the IHS in-signia, is found in the “Baltic” violin. Made in thesame year as his move to the casa di San Bernardo,this instrument represents a change in conception:its more precise craftsmanship and delicate designestablishes the style of del Gesù’s work for the nexteight or nine years. During this period he seems tohave worked with considerable energy and consis-tency. A large number of instruments have survived,suggesting a high rate of production for a craftsmanapparently working alone. Alone, that is, apart fromhis father, who was evidently given sole responsibil-ity for making the scrolls (see p. 138). Delicacy ofform certainly appealed to him at this time, and heeven produced a charming pochette, the wonderfullypreserved “Chardon” of 1735. This has a shield-shaped final rather than a scroll, a feature difficult toattribute to father or son. Dating and identification ofinstruments from the 1730s is less problematic as agreater proportion of them carry authentic labels.After 1736 del Gesù’s industry appears to have de-clined, and he seems not to have made more than teninstruments in any one year.6

By 1740, dramatic changes were beginning to takeplace and during the last four years of his life delGesù clearly rejected many of the certainties and re-strictions of his former approach. He seems to havebeen experi menting constantly: closely related in-struments occur in groups of two or three, followedby a violent swing away to the next idea, which ispursued again for two or three attempts. The “Vieux-temps” of 1741 appears to stand alone, and bears lit-tle comparison with the “Wilton” and the “Alard”,which share many characteristics. Again from 1742,

77

88but very different in conception, come the “Dushkin”and the violin of Dr Sloan, which are closely linkedwith the “Sauret” and the “Hladik” (or “Canary Bird”)of the following year. These in turn are quite differ-ent in character from the “Cannon” and “Carrodus”of 1743. The “Doyen”, “Ole Bull” and “De Beriot” of1744 represent yet another leap in style, and againtheir appearance is quite distinct from the “LordCoke” and the “Leduc”.

The quantity and comparative diversity of the in-struments attributed to the final year of del Gesù’slife is striking. Besides those mentioned above, the“Hennel”, the “Gregoritsch” and the “Prince of Or-ange” are also attributed to 1744. In fact, only the“Ole Bull” and the “Prince of Orange” have labels ofundoubted authenticity, but we may be reasonablyconfident about the dating of the others. If the“Leduc” (with its apparently posthumous label of1745) is included in the list, the total reaches eightvery individual violins. The Hills estimate that froma working life of less than two dozen years, del Gesùproduced between 150 and 200 instruments. The factthat his productivity appears to have varied unpre-dictably from year to year may have a number of ex-planations, including perhaps a poor rate of survival.The only certainty is that until the end, del Gesù'simagination and creativity never left him, and itmight be argued that his final year was also his mostglorious.