Bartolomeo Cristofori/Giovanni Ferrini: two Florentine restorers and re-builders.  A Neapolitan connection between Cristofori and Ferrini in two Italian harpsichords in Britain.


Grant O'Brien



The use of simple geometry in the construction of Italian keyboard instruments

The anonymous single-manual harpsichord in the Russell Collection, Edinburgh, No. HS1-A1620.1

The anonymous harpsichord in the Royal College of Music, London, No. 175

The alteration of these two instruments by Cristofori/Ferrini




            Bartolomeo Cristofori is one of the greatest names in the field of keyboard instrument making of all time.  The design and the skill of execution of his piano escapement action were not really superseded from its invention in about 1700 until Erard patented his double escapement mechanism in 1821.  Many of his ideas such as the inverted wrestplank can still be found in pianos dating from a period 150 years after Cristofori's death, and indeed many of the principles of his action are still to be found in the modern grand piano of the present day.

            Although Cristofori’s name is well-known as the first to design a successful and efficient piano action and his harpsichords, cembali traversi, spinettoni, and spinette ovale are beautifully designed and exhibit an ingenuity and inventiveness found in no other maker of the historical period.  They are also executed with a hand capable of workmanship of unsurpassed quality, and this extends to Cristofori’s pupil Giovanni Ferrini.  But his skills as a harpsichord builder and as a restorer of old, earlier instrument in order to bring them up to date with contemporary instruments are less well recognised.  The re-working of earlier instruments which Cristofori and Ferrini altered in order to give them a compass, disposition and pitch useful for the contemporary musical environment which existed in Florence at the beginning of the eighteenth century.

            In this paper I want to discuss this latter aspect of Cristofori and Ferrini’s creative output.  I want to show that Cristofori, or perhaps his pupil Giovanni Ferrini, had a hand in the alteration of two unsigned and undated harpsichords in British collections.  One of these instruments is in the Donaldson Collection of the Royal College of Music in London and bears the catalogue number 175; the other is in the Russell Collection of Early Keyboard Instruments at the University of Edinburgh in Scotland.  I want to show first of all that both of these instruments were made in the southern part of Italy, and I then want to demonstrate that Cristofori or Ferrini had a hand in the alteration of these instruments.  Along the way there are some interesting developments and discoveries.



The use of simple geometry in the construction of Italian keyboard instruments

Click here to read my paper on the geometry of Italian keyboard instruments

            The work that I have done recently in this field and illustrated below shows that the maker began his design by drawing out the baseboard using dimensions which were simple integers or fractions of the local unit of measurement, and the case sides that he then applied to the sides of the baseboard in the workshop were cut to a height also equal to a simple number of units, or units plus simple fractions.  The combination of the fact that the case sides were hand thicknessed and therefore not all of exactly the same thickness (not even from one end of the board to the other) and the irregular geometry of both polygonal virginals and harpsichords, meant that the resulting final outside dimensions of the instrument were totally unrelated to the local unit of measurement used by the maker.  Therefore a maker starting with two identically-dimensioned baseboards constructed according to his design could end up with slightly differently-sized cases after the sides were added to the two identical baseboards.  Similarly it is the height of the case without the top cap moulding that the maker would measure in his local unit of measurement[1].  He would mark out a number of planks all of the same width in convenient units and then cut and apply these to the outside edges of the baseboard.  Experience measuring Italian instruments has shown that even here, the case-wall heights are often slightly less than expected in places where the top of the case has been planed down to equalise the level of the top of the case at the corners when these did not match exactly after the case sides were assembled.  It is therefore the maximum case-wall height that corresponds to the makers design and not the average case-wall height.  Similarly the position of the soundboard was located by choosing a simple distance from the top of the soundboard liner relative to the top or the bottom of the case boards and this would have been marked out with a marking gauge before the case sides were assembled and attached to the baseboard.  The bottom of the soundboard was therefore not positioned relative to the upper surface of the baseboard, and similarly the top of the soundboard (which was usually made to have a variable thickness for acoustical reasons) was similarly also unrelated in simple units of the local measurement to the position of the top or bottom of the case.

          The problem faced by an investigator is to find the unit of measurement used to design and construct any given instrument.  A harpsichord or a polygonal virginal has many different measurements and it is not at all obvious from looking at the millimetre measurements what the local unit used to arrive at them was.  This is further complicated by the fact that, being hand made, none of the measurements is perfectly exact.  Any method used to find the unit of measurement must therefore also be relatively independent of any inaccuracies resulting from the working methods of the maker. 

            I want to show here, first of all, how some of the basic principles used by Italian makers when setting out their design for the baseboards of both polygonal virginals and harpsichords were based on the local unit of measurement.  The method used by these makers is based on the way in which they used a simple geometrical construction to arrive at the corner angles of polygonal virginals and the tail angle of harpsichords.  Working in reverse, a study of the measurement of the angles and sides of these corners enables a calculation of the unit of measurement.  Establishing the unit of measurement used in the design of the instrument can then be used to determine the centre in which it was built[2].  The importance of such a method to the determination of the maker of an anonymous instrument is clear.  The method clearly does not pinpoint precisely who the maker was, but it does reduce the number of possible makers from the vast breadth of Italian harpsichord, virginal and spinet builders.  It suffices then to compare the anonymous instrument in question with other similar instruments by known builders from the same city or region. 

            I would like to illustrate the method that I have developed to arrive at the unit of measurement for a harpsichord which uses a very simple geometry, and then apply the method to the two harpsichords under study here.  Having established the unit of measurement used in the construction of these two harpsichords I then want to show that the alterations to these two instruments were executed using the Florentine inch or soldo.

            Unfortunately the usual catalogue measurements of Italian harpsichords and virginals do not normally enable one to make an analysis of the size of the baseboard from which the maker began the design and construction of the instrument.  It is therefore necessary to measure the baseboard (if possible by removing the keyboard and jacks and inverting the instrument) without the case sides and then to analyse these measurements. 


The anonymous single-manual harpsichord in the Russell Collection, Edinburgh, No. HS1-A1620.1

Figure 1 - ¾-view in its original case and with its original Neapolitan stand

Anonymous Italian single-manual harpsichord, Naples, c.1620

Russell Collection, Edinburgh, Cat. No. HS1-A1620.2


            The measurements of the baseboard and case show clearly that this instrument was made in Naples (click here for further details).  The anonymous harpsichord in the Russell Collection (Figure 1), although it has a few minor features in common with the 'Neapolitan School of Stringed Keyboard Instrument Making', is clearly built in a quite different style.  The only similarity with the ‘Neapolitan’ school is the panelled front of the nameboard.  Otherwise the case sides are of cypress and not of maple, and they do not project down below the lower surface of the baseboard.  The keyboard does not, and never did, slide in and out of the instrument like a drawer, and the tail has a blunt angle and is not pointed like the harpsichords built in this so-called ‘Neapolitan’ tradition.


The anonymous harpsichord in the Royal College of Music, London, No. 175

 Harpsichord by Onofrio Guarracino, Naples, c.1660, RCM 175

Figure 2 - ¾-view without its external case

Anonymous Italian single-manual harpsichord attributed here to Onofrio Guarracino, Naples

Royal College of Music, London, Cat. No. 175


 Click here to go to the RCM Instrument Museum site

            This is a fine example of an instrument built in the Neapolitan tradition (Figure 2) (Click here to go to a list of the characteristics of the keyboard instruments built in the  Neapolitan school).  It has case sides of maple, a wide lower outside case moulding, and the case sides project slightly below the level of the bottom of the baseboard.  The panelled nameboard, the jackrail mounting system, the keyboard that slides in and out of the case like a drawer, the pointed tail angle, etc. are also all characteristics ascribed to this ‘Neapolitan’ school. 

            The tail angle of this instrument was measured to be 36½º, and the tangent of this angle is tan 36½º = 0.74 is close to 0.7356 = suggesting that the sides of the triangle making up the tail angle have lengths of 10once and 14½ once.  Using these to estimate the size of the oncia being used here, the lengths in once of the other measurements of the baseboard and case height can be made.  These are shown in Table 1.


                                                                                        Measurement          Local                                Length of

                                                                                                    in mm          unit                                     oncia

                               Component of tail parallel the spine:               313    =    14½ once                            21.586

                 Component of tail perpendicular to the spine:               231    =    10once                            21.656

                                                               Spine (long side):             1817    =    84 once                               21.631

                                                               Baseboard width:               770    =    35½ once                            21.690

                                            Baseboard cheek (short side):               478    =    22 once                               21.727

                                                              Case sides height:            213½    =    10 once                               21.35   .

                                                                                   Total          3822½    =    176once         Average:   21.637

Table 1 - Baseboard measurements without the case sides

Anonymous Italian single-manual harpsichord attributed here to Onofrio Guarracino, Naples

Royal College of Music, London, Cat. No. 175



 Figure 3 - Baseboard measurements without the case sides where the angles and lengths as measured are given on the left and nominal angles and nominal lengths are given on the right.

Anonymous Italian single-manual harpsichord attributed here to Onofrio Guarracino, Naples

Royal College of Music, London, Cat. No. 175


             A look at a table of the units of measurement used in the various centres in Italy during the historical period shows that in Naples the palmo, divided into 12 units, had a length of 262.01mm[3].  Hence the oncia had a length of:


This value is close to the value of length of the oncia found above but is significantly different from it.  However, many of the features of this instrument are clearly typical of Neapolitan practice, and many are characteristic of Onofrio Guarracino.  But if this ‘textbook’ value were actually the length of the oncia being used by Guarracino the length of the spine side of the baseboard would have been 21.834x84 = 1834mm an easily-measurable 17mm difference from the 1817 mm found here.  However, the values of the oncia determined by me from some other instruments by Guarracino are given in the table below:


                                                                                                                           Instrument              Oncia

                                     Anonymous harpsichord, Collection of Fernanda Giulini, Milan              21.612mm

                                       Guarracino rectangular virginal, 1679, Rodger Mirrey, London              21.610mm

                                                         Guarracino harpsichord, 1651, Andrea Coen, Rome              21.607mm

                         Anon. Harpsichord, Gemeentemuseum, The Hague Cat. No. 1933.0543              21.649mm

                                                                                                                                    Average       21.620mm

Table 2 - The unit of measurement used in various instruments by Onofrio Guarracino, Naples, active in the second half of the seventeenth century


            The average of these values found in Table 2 above compares with the value of 21.637mm found above.  The difference is only 0.08% and therefore there is strong agreement among the five different instruments.  This therefore suggests strongly that Guarracino was the maker of the RCM175 instrument.  The results of the calculation of the unit of measurement on all of these different instruments shows that Guarracino must have had his own workshop unit which, although it was clearly close to the ‘standard’ text-book unit used in Naples (=21.835mm), was significantly different from it by an amount that would be distinguishable by straightforward measurement.  The discovery that Guarracino, although basically using the Neapolitan unit, had his own individual workshop unit that was slightly different from the ‘standard’ Neapolitan unit is one of most important factors that can be used to attribute instruments to him.  Any instrument with all of the usual features of Guarracino’s work and which uses this unit of measurement is therefore almost certainly by Guarracino and not by one of his contemporaries working in Naples.  The difference between the workshop unit used by Guarracino and this ‘standard’ unit of measurement is therefore one of the many indications that RCM175 is by Guarracino.  It has many characteristics of other Neapolitan keyboard instruments but is characterised by the use of the workshop unit of measurement used only in other instruments by Guarracino.  This therefore helps to tie it down to this one specific maker.


The alteration of these two instruments by Cristofori/Ferrini

            Before starting this discussion it has to be made clear that it is much more difficult to attribute the alteration of an instrument to a given maker than it is to attribute an unaltered instrument by the same maker.  The number of clues to the authorship of an alteration are clearly far fewer than if the instrument were built from scratch by a given builder.  What, then, are some of the characteristics of Cristofori and Ferrini that can be used to establish the presence of their hand in the alteration of these two instruments:

1.     The use of the Florentine soldo based on the Florentine braccio divided into 20 units.

2.    The use of construction characteristics typical of the Cristofori/Ferrini workshop.  The typical internal bracing system which they used can obviously not be considered here because the case structure of these instruments already existed.  However, the Cristofori/Ferrini instruments always have an extremely robust structure for hitching the bass strings and supporting the higher bass string tension.  Further, the jacks are guided, like other instruments of the Florentine school such as those of de Quoco, using upper and lower guides instead of the more ‘classic’ Italian boxslides.

3.    The key guiding system sometimes found with round wooden pegs between the tails of the keylevers instead of the more usual wooden slip sliding in a rack is a characteristic not found in the instruments of other builders.

4.    The use of short sections of bridge placed parallel to the main section of bridge in the bass, rather than the more common mitred-section characteristic of almost all other makers.

Figure 3 - The bass ends of the 8' bridge showing the 'added' section of bridge place parallel to the main bridge on the right.  A close look at the far ends of the bass strings shows they only pass what appear to be hitchpins, and are not attached to them.  Instead the strings pass through the tail of the case and are attached to hitchpins which are driven into uthe outside of the tail and into the bass liner below the soundboard liner and soundboard.  Notice how closely the bridge pins are placed to the ends of the bridge sections.


5.    The use of two angled almost vertical cuts at the ends of the bridges and nuts, placed very close to their ends and to the last bridge or nut pin.  This results in a maximum area of freely-vibrating soundboard near the ends of the bridge with a consequent improvement in the sound of the instrument at the extreme ends of the compass.

6.    The use of some of the mouldings also found on other instruments by Cristofori and Ferrini.

7.    The use of the double ovolo moulding on both the front and back edges of the bridge and nut.

            It was the last of these characteristics that first drew my attention to the fact that the two instruments just discussed might have been re-worked by Cristofori or Ferrini.  Although the bridges clearly belonged to the original states of both instruments, the nuts were replacements and seemed clearly to have come from the Cristofori/Ferrini workshop.  Also the jacks in both of these two instruments are guided in upper and lower registers and not in boxslides, and the mouldings on the outside edges of these registers are similar to those found on other harpsichords by Cristofori and Ferrini.  These are shown in Figure 4:


Nuts and Bridges

                                                                                                                                   Russell Collection                                                                  RCM175

Nuts and bridges



Upper Register Guides

                                                                                                                                          Russell Collection                                                              RCM175

Figure 4 - Some Cristofori/Ferrini features

A comparison of the Russell Collection and the RCM single-manual harpsichords


            In addition to the typical rounded mouldings along the front and back edges of the nuts, both instruments have the characteristic almost vertical angled cuts at the ends of both the bridges and the nuts.  The Russell Collection harpsichord also has an unusual system of attaching the bass strings.  Instead of being hitched to pins inside the case along the tail as is normal practice, Cristofori or Ferrini has cut neat small square holes just above the tail hitchpin moulding, and the bass strings pass through these holes and hitch to pins placed in the outside of the case about 30mm below the holes.  This extraordinary system seems to be typical of a pre-occupation with attaching the bass strings in a way which is completely secure and which will not result in a splitting of the bass hitchpin moulding or tail liner.  It is, to be sure, completely secure and free of any tendency to splitting.

             Although it does not have any ‘belt and braces’ method of attaching the bass strings, the RCM harpsichord has the typical key-guiding system of Cristofori and Ferrini with small wooden pegs placed between the tails of the keylevers and the keyframe itself slides in and out of the instrument like a drawer.  All of these features, along with the typical upper and lower jack guides, seem to me to indicate that either Cristofori or Ferrini had a hand in the re-working of both of these instruments.

Figure 5 - The keyboard of the harpsichord RCM175 with keyboards by Cristofori/Ferrini features showing the pegs used to guide the keylevers near the keylever tails.

Single-manual harpsichord, attributed to Onofrio Guarracino, RCM175, in the Royal College of Music, London


             In addition to this it seems clear that a number of features of these instruments are based on the use of the Florentine soldo measurement.  What I want to show is that Cristofori and Ferrini used a soldo = 27.53mm which was a twentieth part of the Florentine braccio = 550.64mm.  In this connection and before looking at the string scalings of the two instruments discussed above I would like to have a look first at the scalings of an instrument made from the start by Cristofori.  Figure 5 below shows the string scalings of the single-manual harpsichord by Bartolomeo Cristofori dated 1722 in the Musikinstrumentenmuseum of the University of Leipzig.  This shows that Cristofori paid close attention to ensuring that the string scalings of his instruments had Pythagorean scalings which doubled in length with each octave drop in pitch as far down in the compass as was possible, and in this case right down to tenor c.  It also shows that if the length of the soldo given above is correct, then the scalings are based on c2 = 10 1/3 soldi.

  Scalings of Leipzig Cristofori 1722

 Figure 5 - String Scalings

Italian single-manual harpsichord, by Bartolomeo Cristofori, Florence, 1722

Musikinstrumentenmuseum, Leipzig, Cat. No. 84


            Figure 6 below shows the string scalings of RCM 175.  Like the Russell Collection harpsichord, the nut is a replacement and the compass has been changed, and therefore the present position of the nut reflects a string scaling design appropriate to the re-working of the instrument, rather than to that of its original design.  In this case it seems clear that this design was based on a treble scaling of 9soldi, although in this case, because the instrument originally had an enharmonic compass with more than 12 notes to the octave,  it was possible to achieve Pythagorean scalings for only a part of the compass in the treble.


Figure 6 - String Scalings

Anonymous Italian single-manual harpsichord attributed here to Onofrio Guarracino, Naples

Royal College of Music, London, Cat. No. 175


          Similarly Figure 7 shows the string scalings of the Russell Collection harpsichord which seems to have scalings based on 9¾ soldi.  This is close to a semitone higher than RCM 175 and a semitone lower than the 1722 Leipzig harpsichord by Cristofori.  It therefore appears that Cristofori/Ferrini were working to at least 3 different pitch levels differing from one another by semitone intervals.

 Scalings of Russell Collection Anonymous Neapolitan harpsichord

Figure 7 - String Scalings

Anonymous Italian single-manual harpsichord, Naples, c.1620

Russell Collection, Edinburgh, Cat. No. HS1-A1620.2


          Figure 8 below shows a graph of the jackslot spacing of the front register of RCM 175.  Here the distance from the spine of the instrument to the edge of each jackslot is plotted against the note sounded by the jack whose jackslot is being measured.  The uniform spacing of the jackslots gives rise to a straight line plot whose mathematical characteristics can be calculated using the usual statistical analysis.

Figure 8 - Spacing of the jackslots in units of both the Neapolitan oncia and the Florentine soldo.

Anonymous Italian single-manual harpsichord attributed here to Onofrio Guarracino, Naples

Royal College of Music, London, Cat. No. 175


            Using the usual regression analysis by the method of least squares, calculation gives a correlation coefficient for this data of r = 0.999993572 indicating a very good fit of the measured data to a straight line.  The calculated slope of the line is m = 13.7675mm/jackslot.

             This slope = 13.7675mm/jackslot is equivalent to 0.50005soldi/jackslot, based on the Florentine soldo of 27.532mm, or to a spacing of 0.63154 Neapolitan once or to 7/11 = 0.63636 Neapolitan once according to the values found in the reference tables.  This therefore appears to be a spacing of exactly 50 jackslots in 25 soldi:


13.7675mm/jackslot = 25 Florentine soldi/50 jackslots = 0.5000 Florentine soldi/jackslot

or, in terms of the Neapolitan oncia:

13.7675mm/jackslot = 7 Neapolitan once/11 jackslots = 0.63636 Neapolitan once/jackslot


            Using the value of the Florentine soldo used by Cristofori to calculate the soldo gives:

1 soldo = (13.767x50)/25  = 27.535mm.

This compares with the value given by both Vega and Colonel Cotty[4] for the braccio divided into 20 soldi of 550.64mm, or of 1 soldo = 550.64/20 = 27.532mm.  This only 0.01% different from that estimated here and strongly suggests that the register slots were cut out by designing them to be exactly ½ of a Florentine soldo apart.

              Clearly because the spacings used by Guarracino in once are the same, Cristofori simply used his unit to mark out the registers for the up-dated, modernised instrument.

             Although different from RCM 175 because of the existence of different space constraints imposed by the re-working of instruments with different widths, the jack spacing the Russell Collection harpsichord is also clearly based on the Florentine soldo.  In this case, however, there are 51 jackslots in 25 soldi giving rise to a calculated Florentine soldo of 27.328, again only marginally different (error = 0.75%) from the value found in the tables.

             It would be an incredible coincidence if these two instruments, both of which have numerous characteristics of the instruments of Cristofori and Ferrini, where not altered by one or the other of these two Florentine builders.  The use of the Florentine soldo in the scaling design and in the design of the register spacing of both instruments certainly makes clear that they were altered in Florence.  This, in addition to the many other characteristics individual to Cristofori and Ferrini, make it almost a certainly that one of these two had a hand in the re-working of these instruments.

             I conclude by posing three questions:

  1. Why are there no early Florentine signed or anonymous instruments which have been re-worked by Ferrini or Cristofori both of whom worked in Florence where one would have expected numerous out-of-date instruments of Florentine origin? 

  2. Following on from this, did Cristofori and Ferrini engage in a commercial trade in old, out of date instruments which they bought in the south of Italy and then re-worked to sell on again?  Were the harpsichords made in Naples therefore especially valued by Cristofori and Ferrini to the extent that they engaged in such a trade to the exclusion of locally made harpsichords?

  3. What are the characteristics that distinguish Cristofori’s work from that of Ferrini, and therefore how can we distinguish the work of these two important Florentine master/apprentice harpsichord makers?  One distinguishing feature seems to be the use of round pegs to guide the key tails as a practice characteristic only of Ferrini’s work.  Are there any others?

Ars sine scientia nihil est.

Grant O’Brien

Imola, December, 1997, Revised Edinburgh, December, 2003, and January, 2018.




[1] In the North-European tradition where the case sides are much thicker than in Italian practice, the moulding is often cut into the top of the case side itself, and the case sides are usually (but not always) applied to the top of the baseboard.  It is therefore the case height less the thickness of the baseboard that the maker would measure out in a simple number of local units.

[2] Denzil Wraight, in his otherwise splendid work on the identification of Italian keyboard instruments, rejects the evidence provided by the local unit of measurement (see: Denzil Wraight, ‘The identification and authentication of Italian string keyboard instruments’, The Historical Harpsichord.  Volume Three, general editor Howard Schott, (Pendragon Press, Stuyvesant, NY, 1992) pp. 66-76).  Unfortunately he seems to reject the whole process of using the local unit of measurement as a method for determining the origin of an instrument on the basis of a quoted example of the confusion that has arisen because of the fact that the Frankfurt and Vicenza inches are fortuitously in the ratio of 3 to 4.

[3] See: Horace Doursther, Dictionnaire universel des poids et mesures anciens et modernes, (M Hayer, Brussels, 1840), Giovanni Croci, Dizionario universale dei pesi e delle misure in uso presso gli antichi e moderni con ragguaglio ai pesi e misure del sistema metrico, (Published by the Author, Milan, 1860) and my article ‘The use of simple geometry and the local unit of measurement in the design of Italian stringed keyboard instruments:  an aid to attribution and to organological analysis’, The Galpin Society Journal, 52 (1999) 108-171.  There is an abridged table of lengths arranged according to length and another arranged according to the centre in Italy on this site.

[4] See: Georg von Vega, Natürliches Mäß-, Gewicht- und Münzsystem, (Vienna, 1803) and Colonel Cotty, Aide-Mémoire a l’usage des officiers d’artillerie de France, 2 (Paris, 1819) 896-7 and the references quoted on this site in footnote 3.


Go back to the top of the document