Forms of mystery

The 1708 ‘Dancla’ Stradivari together with its most probable mould, the ‘G’ (MS49)

‘G’ MOULD COURTESY MUSEO DEL VIOLINO. ‘DANCLA’ PHOTO JAN RÖHRMANN

The first part of this article (‘Variations on a theme’, May 2019) provided readers with new interpretations of the twelve violin moulds attributed to Antonio Stradivari now housed in Cremona’s Museo del Violino. The study of these artefacts has been an ongoing process for the past 250 years: it became a real obsession for collectors such as Count Cozio di Salabue, who acquired the moulds through the descendants of Antonio Stradivari in 1776; for makers such as Giuseppe Fiorini, who purchased the relics from the heirs of Cozio in 1920, later donating them to the town of Cremona; and more recently some of the most knowledgeable scholars active during the past decades.

What all of us thought we knew about these items (which has been thoroughly explained by Simone Fernando Sacconi in his 1972 work I ‘Segreti’ di Stradivari) now requires some rethinking, especially after the publication in 2016 of the Museo’s catalogue of Stradivari relics. Scientific analyses of the inks and handwriting present on them has in fact shown that the majority of the inscriptions are of 19th-century origin, and are due to Cozio’s cataloguing efforts. Thus the letters (for instance, P for ‘Prima’, S for ‘Seconda’ and so on) and the dates inscribed on the moulds remain to be investigated with greater accuracy, always considering that Cozio might have been handing down to us previous knowledge about the moulds originating either from Stradivari’s workshop (rewriting fading letters?) or his heirs.

In this article, we would like to introduce an approach that, echoing the terminology used for dating tree-rings, we might call ‘morphochronology’. While dendrochronology can establish the earliest possible date (or terminus post quem) for an instrument to be made, morpho-chronology could possibly use the opposite approach by proposing a ‘terminus ante quem’ for a specific Stradivari mould to have been designed. The two techniques would, however, be similar to the extent that they can only limit the field of research towards one time direction. Just like wood dating can tell us that a violin couldn’t possibly have been made while the tree was still growing, finding an exact match between the rib outline of a Stradivari violin (if its dating can be considered certain) and one of the existing moulds would tell us that the mould itself pre-dates that specific instrument, even if we cannot establish by how many years through this technique alone.

At the end of this article is our first attempt at placing the moulds in chronological order, but this should be considered a very tentative initial effort. It also requires accepting a set of assumptions that, as in any scientific field, can be easily refuted by new experiences. In this context, ‘new experiences’ would be the opportunity for analysing more and more original examples, and thus requires the generous cooperation of institutions, collectors and dealers.

Our approach is based on two assumptions in particular. First, when a Stradivari violin outline is close enough to one of the moulds preserved in the Museo, we assume that the violin was made using that form. Second, when an undisputed violin by Stradivari is extremely different from any of the moulds, then we assume that its form was lost during the long history of this collection – either it was discarded by Stradivari, or modified to an extent that makes it noticeably different from the way it looks today.

Of course, the interesting part of all this lies in the expression we just used: ‘close enough to one of the forms’. In the previous article we explained some of the reasons why delicate and fragile structures such as violin ribs and plates could have undergone changes that make them now less adherent to a specific mould (which are themselves wooden workshop tools, in some cases heavily worn).

There are at least two methods that can be jointly used to achieve this purpose. The first, covered in the first article, is a statistical approach whereby we analysed outlines of c.250 Stradivari violins to de ne the individual features of each mould, and proposed theories about the relative frequency of their use.e second is more factual, where the rib outline of a given violin (obtained through micro-CT scanning with an accuracy of c.80 microns) is superimposed on to an accurate reproduction of the mould, showing how well the rib structure and blocks fit.is must also take into account various issues: shrinking and distortion, to begin with, but also the fact that upper and lower ribs which were originally in one piece were regularly severed and shortened during subsequent repairs, and blocks replaced.is is one of the possible explanations for a worse-than-expected fit (usually the ribs being on the short side with respect to the mould).

FIGURE 1 Ribs of the 1699 Auer, Benvenuti’ Stradivari overlaid on theB’ mould

ALL MOULDS COURTESY MUSEO DEL VIOLINO. ALL RIB SCANS VIOLINFORENSIC

IT IS VERY POSSIBLE THAT SEVERAL VERSIONS OF A SINGLE FORM WERE ORIGINALLY USED IN STRADIVARI’S WORKSHOP

A project at the Vienna University has recently micro-CT-scanned more than 20 Stradivari instruments. We would like to present here this relatively new technique to show ‘how good’ a fit between a rib garland and a mould can be on the basis of the instruments scanned.

The image on page 30 shows the ‘G’ mould (MS49) paired with the ‘Dancla’ violin of 1708 (as shown on The Strad ’s poster,bit.ly/2v0jlqP).The historical knowledge of scholars such as Sacconi and Pollens seems to hold very well against our analysis, as we have not yet been able to identify any violins made on the ‘G’ mould before 1708.

Figure 1 shows the 1699 ‘Auer, Benvenuti’ Stradivari ribs superimposed on to the ‘B’ mould (MS33). We think most violin makers would consider this to be a very good fit, even for a violin made by them. Incidentally, we already knew from Sacconi that this particular mould, unique in Stradivari’s production, was extensively used during the 1690s. Sacconi was convinced that this could be directly inferred from the date he thought Stradivari had written on the mould (3 June 1692); the date is now attributed to a later hand, though, whoever wrote it seemed to be quite aware of when this form was introduced, as in our statistical results too the rst violins built on MS33 do indeed date from 1692. Figures 2 to 4 present similar overlays for moulds MS 2, 44 and 38.

Next, we switched to an approach that is the exact opposite of the one above: how bad can a fit between an original violin and mould be? Figure 5 presents an overlay of the 1669 ‘Francalucci, Clisbee’, currently exhibited at the Museo, with the ‘T’ mould (MS11). is is the closest of the moulds in the museum’s collection (the only one where the width across the C-bouts is as narrow as that of the violin); however, the 2–3mm difierence in length on both the upper and lower bouts cannot be explained, certainly not by shortening the ribs, which are longer here.is is why we felt the need to introduce a new archetype in Stradivari’s production, which we thought it appropriate to call ‘Long T’. Incidentally, our results show that the number of instruments that could have been made using this elongated form of ‘T’ (four violins, all dating from the 1660s and 70s) is larger than that of those matching the actual ‘T’ (only the ‘Arányi’, which unfortunately has lost its original label). How do the actual ‘T’ and our ‘Long T’ relate? This has been explained convincingly in the past, by both Stewart Pollens and François Denis: when drawing his template, Stradivari seemingly just ‘moved’ a part of the violin to change its length, and left all other dimensions the same. This explanation differs substantially from the proportional design method that was used in classical Cremona and Brescia, as Denis has shown. Both Pollens and Denis also recognised an identical process to obtain the ‘G’ form from the ‘P’ and ‘PG’ mould group.

FIGURE 2 ‘S’ mould (MS2) overlaid on the 1698 ‘Rouse Boughton’

FIGURE 3 ‘P’ mould (MS44) overlaid on the 1707 ‘Brüstlein’

FIGURE 4 ‘B’ mould (short) (MS38) overlaid on the 1725 ‘Chaconne’

FIGURE 5 ‘T’ mould (MS11) overlaid on the 1669 ‘Francalucci, Clisbee’

FIGURE 6 Ribs of the 1718 Arnold Rose, ViottT overlaid on theP’ mould

FIGURE 7 PG’ mould overlaid on the ribs o the 1714Smith-Quersin’

To remain in the field of what we call the ‘Grand Pattern’ forms, we can speculate that the dating on some moulds is not as accurate as it proved to be in the case of MS33. Our studies seem to indicate that both ‘PG’ (MS21) and ‘P’ (MS44) might have been in use before their dating (now attributed to Stradivari’s grandson, also called Antonio) of 1689 and 1705 respectively. In fact we find several earlier violins, in some case very well-known ones such as the 1704 ‘Betts’, that pre-date the inscriptions on their moulds, in this case MS44. As we have seen in the first part, MS44 was the most extensively used of all Stradivari’s moulds (we estimate about half of his entire production), so it is very possible that several versions of the form were originally used in his workshop, being replaced as they gradually wore out. Or they may have co-existed to make two instruments at the same time.

If we are to speculate about a lost mould, one pre-requisite should be the presence of a distinct group of several instruments with similar outlines, which do not quite fit one of the existing moulds. It is dificult to apply this stricture when it comes to similar versions of the same mould: however, our independent researches seemed to agree that there was in fact another version of MS44, which we called ‘narrow P’ because of its slightly reduced dimensions. The group is quite widespread (21 examples in our charts) and its use reaches a peak after 1703, and here we show how these violins differ from MS44 through the 1718 ‘Arnold Rosé, Viotti’ (figure 6).

To come back again to the early years of Antonio Stradivari, it is interesting to observe that violins of large dimensions were introduced even before we can find instruments based either on the ‘P’, ‘PG’ or ‘G’ models. This is the case, for example, of the famous inlaid violin of 1679 known as the ‘Hellier’. The proportions of this instrument, and of several later ones which are notably often of the decorated type, are very close to Stradivari’s Grand Pattern, but they differ in that they have shorter and rounder C-bouts and stretched lower bouts. The ‘Hellier’ is the earliest violin of this type in our list, and since the category has been also identified in the past by experts as that of ‘Amatisé’ instruments, we decided to adopt the same name. Figure 7 (page 34) shows a comparison between the rib structure of a later instrument that could have been made using the Amatisé form, the 1714 ‘Smith-Quersin’, and the ‘PG’ mould, showing a significant difference in the C-bout length.

There are other instances where Stradivari’s taste for ‘moving around’ individual parts of his models can be seen at work. Some of them seem to support the groups that have already been defined in the field of expertise, such as the interesting ‘short Long Pattern’ violins, a category emerging in about 1694 (as described by Charles Beare in the catalogue of the 2013 Oxford exhibition). Stradivari seemingly abandoned his Long Pattern models in steps, and the first one might have been to reduce the length of ‘SL’ (MS28). In fact, most of his earlier violins made on ‘SL’ proportions would fit a stretched version of this mould (for instance, the 1690 ‘Stephens’), and we could identify only two violins towards the end of the 1690s fitting the existing version of it. Lastly the maker might have had, besides ‘T’ and ‘SL’, a long version of ‘PG’ as well, as a distinct group of instruments (and the reader might also want to check the outline of the ‘NN’ violin from 1721) seem to suggest.

Introducing the concept that the mould collection in Cremona is incomplete should be investigated further, but still it allowed us to take into consideration several instrument outlines that could not be explained otherwise. It also showed how Stradivari’s design approach remained consistent from one change to the next. His attitude mirrors a phase when the original appeal of traditional Cremonese geometry based, as Denis has shown, on the belief in the power of ratios, left room for a more empirical one. Stradivari apparently aimed at changing just one variable at a time (e.g. ‘let’s make the lower bouts longer and leave all the rest the same’) to learn about the difference. This resulted in an experimental approach (twelve surviving moulds, plus perhaps a few more) which was unprecedented in Cremona, and actually challenges every maker working today to present violins made in dozens of different shapes. Was Antonio still ahead of all of us?

Preliminary studies for this article and all high-resolution CT scans were made in the framework of the research projectA 21st-century approach to the study of historic violin bodies’, funded by the Anniversary Fund of the Austrian National Bank. The micro-CT scans were provided by the Micro-CT-Lab of the Department for Anthropology at the University of Vienna, headed by Dr Gerhard Weber

A first attempt to place Stradivari moulds in chronological order, based on photo analyses and overlays between moulds and micro-CT scans of the ribs. The five entries in red represent archetypes missing from the Museo del Violino collection (lost moulds’). The authors would like to thank Marion Pollard and Juliette Joanny for their assistance. Photos of the moulds are courtesy of the Museo del Violino and François Denis.