The lute in its historical reality

source: http://www.aquilacorde.com/lutes.htm

2007 © Aquila Corde Armoniche S.a.s. All rights reserved
by Mimmo Peruffo, 2007

                                                                                                                                                  by Mimmo Peruffo, 2007

Our philosophy concerning gut stringings (but also their synthetic equivalents) for the lute family and 5 course guitars is simple:

to reproduce, as far as possible, the typical sounds of historic instruments as they were in use.

Sill, within those limitations, research in the field of historic stringmaking made some important progress in recent years and, although we cannot claim we know exactly what the sound of the dolce strumento  was like (a speculative point rather than a concrete one, anyway, since there must have been different opinions among lutenists in the past, too), we can fairly confidently define the acoustical region, common to all lutes, which was imposed by the stringmaking technologies of the past.

  First of all, let us rule out the materials whose sound definitely cannot match the characteristic sound qualities of the lute:

1. PVF (‘carbon’) strings: much too bright in comparison with any type of gut string.

2. Nylon: produces  a somewhat duller and darker sound than gut.

3. Nylgut: thin strings sound very close to gut, but does not quite compare by increasing diameters.
4. Wound on nylon multifilament: almost all the strings of this type are much too bright and possess too much sustain - the opposite of what revealed by research on 18^th century wound strings, which were fundamental-heavy and needed octaves for brightness, and had limited sustain.

 And then let’s consider some other parameters pertaining to the sound of the lute:
 5. Working string tension: today’s criteria, when working out lute stringings, rarely follow the idea of feeling of equal stiffness on every course, like advocated by the ancients. The modern rule is, in general, to calculate the string diameters by applying the same tension, expressed in kilograms, to each string (this criterion is first described by Maugine & Maigne 'Nouveau manuel complet du luthier' in Paris, 1869) and completely ignores the variability of some typical factors, such as the different amount of reduction of the string diameters under working tension and how different strings of different manufacture and/or length feel under the fingers.

 6. Octave strings: the modern tendency is to apply a noticeably lower tension than on  their respective fundamental strings (Virdung, 1511 seem to suggest that the diameter of the octave string should be half that of its fundamental).

 7. Trebles: when single strung, modern tendency is to apply too low a tension, giving an unbalanced feeling between the treble and the other courses.

 8. Stringing criteria: the principle of grouping the strings into three well defined Sorts (like advocated in the old treatises, in Trebles, Meanes and Basses) is usually ignored. Thus, we often see strings of one Sort  invading the field of another, thus altering the timbric and dynamic balance of the instrument (wound strings on the 4^th course, wound long diapasons &c).

In conclusion: the acoustical qualities of today’s lutes are, in general, remarkably brighter and have more sustain in the bass and, because of the wound strings, also in the mid-register, thus failing to achieve the timbric and dynamic homogeneity we believe was typical of the past.

At the top end, trebles can be much brighter (PVF or ‘carbon’) or duller (nylon) than gut. We have created a new sound that doesn’t have very much in common with that of the past. 

No criticism at all on this choice: the lute can well be played like this, too.

The three ages of the lute and the three Sorts of strings

(From here on, where we talk of string Sorts, we understand them in Dowland and Mace's sense, as in ‘Varietie of Lute Lessons’ and in 'Musik's Monument')

The history of the lute (meant as family of instruments), seen in relation to the string making technologies which were developed in then course of the 17^th and 18^th centuries, can be divided in three basic periods, which, generally speaking, are essentially connected to the types of available bass strings:

- Lutes from about the mid-15^th century to about 1570-80 (6 course lute and vihuela).
- Lutes from about 1580 to the end of the 17^th century (7, 8, 9, 10 course lutes, long and short extended archlutes, theorbos, 11 course and 13 course d-minor lutes with no, or short, extension and baroque guitars).
- 18^th century lutes (11 and 13 course d-minor lutes without extension, 13 course d-minor lutes with swan-neck extension, archlutes, theorbos, mandoras and baroque guitars).

We know that as from the early 17^th century (i.e. the time when the lute had an open string range of 2 octaves and a fourth) the ancients felt the necessity to identify  three Sorts of strings (see Dowland, 1610): Trebles, Meanes and Basses.

After a long period of study and practical experimentation we came to the conclusion that, far from being a simple commercial description, the scope of such distinction was to achieve some kind of switch thorough  the registers from trebles to lower bass. 
The acoustical and mechanical problems in the lower registers increase with the increasing string diameters and can only be solved by switching, at the right point, from one type (i.e.Sort) of string to the next. In other words, since it was not possible to unlimitedly increase the diameters, it was necessary to employ different  types of strings, each able to overcome the limits reached in the previous register.

Just like today when we have to work out a complete range of strings for the lute, we assume that ancient string makers followed, from the late 16^th century on, three different manufacturing processes in order to produce:

    - Treble strings (Dowland’s and Mace’s Trebles; i.e. Romans, Minikins etc), i.e. the first three courses of both Renaissance and Baroque lutes.
    - Mid register (4^th and 5^th courses, Dowland’s Meanes, which he divides in Small and Great Meanes; i.e. Gansars).
    - Low register (from the 6^th course down, the Basses; Lyons, Pistoys, Catlins).

That different manufacturing processes were not interchangeable is evident both in Dowland (1610) and in Mace (1676): the former says that Gansars (which in his opinion made excellent Meanes) could not be used as Trebles since they would immediately break under stress. On the other hand, had the Meanes been manufactured the same way the Trebles were, we believe they would have presented serious acoustical performance problems, since they would have been much too stiff: Trebles as described by Dowland were rather stiff and prickly to the flesh of the thumb pressing against the string's tip.

Also Thomas Mace, 66 years after Dowland, underlines the fact that the thin Minikins (treble strings) are so strong that if you pull  them with your hands they 'will many times endanger the cutting into your flesh, rather than it will break, although it be a small Treble-Minikin string'. On the contrary, 'your Venice-Catlins (i.e suitables for the 4^th and 5^th courses) will scarcely be broken, by a mans (reasonable) strength', in spite of being thicker.

In order to achieve this we must sacrifice suppleness. We find trace of this in some old sources: Dowland (1610), to quote him once again, stated that a good treble must feel stiff and prickly to the thumb; Baron (1727) claims that a good Roman treble can last up to 4 weeks. Could, say, a couple of weeks playing life have been the rule?

We believe that this term was only reserved to strings that underwent particular treatments (as reported in some historical sources, like Skippon’s description of a stringmaking workshop in Padua, c. 1660, for instance) apt to stiffen the gut (we do that only with strings of diameters thinner than .48mm).

By increasing its thickness, string length remaining equal, a string will gradually lose its acoustical qualities, until it becomes completely dull. This is due to the inner damping effect, called Inharmonicity. On the Renaissance lute the problem begins to appear as from the fourth course, becoming increasingly serious as we move down the registers. Pairing octave strings on the lower courses was the expedient the ancients employed to retrieve the lost harmonics (see Virdung, 1511).

In order to remedy this loss of acoustic capacity we try to achieve the highest possible degree of suppleness, which is here the most important parameter. This is obtained, no doubt, at the cost of tensile resistance but it is no real problem, since we are far away from the Breaking Frequency  (faq 14).

1. By specifically treating the fresh gut in order to reduce its stiffness as much as possible, before twisting. 

We believe that strings similar to our Venice type, i.e. a rope-like structure, were in use for the basses of the 6 course lute until about 1570. Recently acquired sources describe not only the presence of orditori  (i.e. wheels with three or four rotating hooks used to make ropes) in 16^th century roman stringmakers workshop inventories, but we also have evidence that strings of this type were already in use on musical instruments as from the end of the 15^th century (Patrizio Barbieri: Roman and Neapolitan gut strings, 1550-1590, GSJ, May 2006, pp 176-7.).

However, they were probably already in use well before that time:see here an example from the late Roman imperial period:

.

The new basses were probably developed to their best in a region between Florence and Bologna (which is where the Venice Catlins mentioned by Dowland in 1610 were produced).

Fact is, the lutenist Michelangelo Galilei, in a letter to his brother Galileo, asks to send him ‘...four thick strings from Florence to meet his own and his students’ needs...’. Michelangelo at the time was living in Munich, one of the most renown string producing centers. It would seem obvious that the local strings were no match for the Florentine basses.

In the light of all the information we have so far, we believe that the Spanish vihuela de mano was not strung with unison courses throughout.

    1. Italian and German string making technology before 1570 was not so advanced as to grant the production of efficient enough bass strings (octaves were needed to provide the harmonics), as made clear by Virdung and Galilei.

    3. Pisador (1552), talking about the 4th course, made it clear it ought to be strung in unison. Such a statement could imply that the use of octaves was standard but he did not like it, or it was not appropriate for his music. Hence the necessity to write down something that was outside the musicians’ common practice.

    4. Fuenllana (1554) prescribes playing only one the two strings in the couse in some passages (as does Dalza): this artifice is only limited to the 2^nd, 3^rd and 4^th course, though, another hint that at least the 4th would be strung with unisons. We know nothing about the 5^th and 6^th.

    5. Bermudo (1555) states that the guitar’s 4^th course has an octave, like the fourth of the lute, or Flemish vihuela. Here can be inferred that the 4th of the vihuela was a unison while the lute wasn’t, since he needs to refer to the lute, an instrument less familiar to him, while it would have been natural to refer to the vihuela. Again, we know nothing about the 5^th and 6^th.

    6. Bermudo also says that if you wish to turn a vihuela into a guitar (4^th with octave, all other courses in unison) you simply have to take off the 1^st and 6^th courses. This would suggest that the vihuela had a unison 4th (but sometimes also a paired octave, as implied by Pisador - see above 3.), i.e. guitar 3^rd, and the 5^th, i.e. guitar 4^th, with octave. It follows that the 6^th must also have had an octave.

    7. On top of that Bermudo also discusses slanting the bridge (ch. LXXXV), in order to compensate for the amount of space taken by the large knot of the 6^th string, which is always referred to in the singular, never in the plural. So the course must have had a paired octave. The larger amount of space taken by the knot (not by the knots!) and the resulting need to slant the bridge in order to keep the length of all strings equal, clearly indicate that the string must have been pretty thick.

If the basses were that thick, they could not, owing to their high Inharmonicity Index, have had such a good acoustical performance.The stringent consequence is that it needed an octave.

    8. The only source clearly mentioning unison stringing on the vihuela dates back to 1611, a fairly long time after the instrument had fallen into disuse. This source (Sebastian de Covarrubia’s Tesoro de la lengua castellana, 1611) does not specifically treat musical matters. It is a dictionary compiled at a time where the progress made in the string making technology already allowed to dispose of octave strings on the lute. So it is an anacronism to apply a piece of information from the early 17^th century to an instrument that was in use in the mid 16^th century. Applying the same principle we could assume, reading Dowland, that Francesco da Milano’s lute was strung with all unisons!

    9. Double treble and unison courses: the fact that the vihuela was generally (but not always) strung with a double treble led some scholars to take that as evidence in favour of all courses having been strung with unisons. We fail to grasp the logic of it. There is, on the other hand, evidence proving that the vihuela could have a single treble, whereas most Renaissance lutes where strung with double trebles.

  If, as by now proven, rope-like strings were already in use in the mid 15^th century, and the 6 course lute needed paired octaves in the bass register to compensate for the poor sound, what made it possible to extend the basses down another 4^th or 5^th?

We believe the problem was definitely solved when a new way was found to increase the weight of gut to twice its natural value.

(see Mimmo Peruffo: "The mystery of gut bass strings in the sixteenth and seventeenth centuries: the role of loaded-weighted gut", Recercare, v 1993, pp. 115-51).

Why twice that value?

The acoustical performance of a gut string (as understood in the concept of Inharmonicity) is a function of type and amount of twist, working tension, diameter and material employed. The sum total of these parameters, obviously each carried out the best possible way, resulted in the acoustical limit that was represented by the bottom string of the six course lute. Now, if that was the lowest limit of acceptable sound quality, a string that is manufactured the same way and is expected to work one fourth lower (i.e. the new acoustical limit) cannot exceed that same diameter. In order to achieve that - the calculation is quite simple - the specific weight of the material employed cannot be less than twice that of natural gut.
 A new development must have granted a noticeable reduction of the string diameter and a lower Inharmonicity limit, which until then had been an open string range of two octaves, by a fourth: ‘The Lutes of the newe invention with thirtene strynges, be not subiecte to this inconvenince, where of the laste is put be lowe: whiche accordyng to the maner now abaies, is thereby augmented a whole fowerth’  remarks Adrien Le Roy in his 'A briefe and plaine instruction...'  in 1574.

 Modern strings that can achieve that can present different shades of dark red, brown or blackish colour, but also light yellow - depending on the oxides or sulfides employed. Also metal powders like metallic-copper (which is what we use on our loaded strings because is not toxic) achieve the same goal: we still have ancient recipes describing how to produce the finest copper powder (we tried them quite successfully), like the one by Don Alessio Piemontese ‘I secreti...’, printed in Venice in 1555: the resulting colour, too, looks very much like what we see on iconographical sources:

                                                                                            7course lute by anonymous: detail

       

                                                                         Girolamo Martinelli, 2nd half of the 17th C: Concerto in casa Lazzari

Girolamo Martinelli: detail of the violone strings

             Girolamo Martinelli: detail on the bass-violin string          

Francois le Troy, 2nd half of the 17th C: detail on the brown basses

                                          

                                             Anonymous dutch painter, 2nd half of the 17th C: detail of the red bass strings on a 12 course lute

Anonymous dutch painter, 2nd half of the 17th C: detail of the red bass strings on a 12 course lute

Rutilio Manetti, Siena 1624: detail on the brown Lute bass strings

Rutilio Manetti, Siena 1624: detail of the Violin brown 3rd & 4th strings ("...best strings are Roman 1st & 2nd of Venice catlins: 3rd & 4th best be finest & smoothest Lyons, all 4 differ in size..." James Talbot's manuscript, 1695 ca)

    Incorporating finely insoluble powdered solid pigments into a matrix of different nature was a fairly common practice in the 16^th and 17^th centuries: several ancient recipes could have been easily employed for ‘loading’ gut (see, for instance, Giovanventura Rossetti’s recipes for dyeing fabrics, silk and leather in his 'Plichto de l’arte de tentori che insegna tenger pani, telle, banbasi et sede si per larthe magiore come per la comune', Venezia, 1568). Some of these describe how to incorporate cinnabar (red mercury sulphide) or lithargyrum (yellow lead oxide) into wax, leather, silk, wood, hair, inks &c.: indeed, only a short step away from gut.

Lead, Iron and Mercury oxides

 There are several points in favour of our hypothesis, and one against:

    a) Consistently small diameters of string holes in bridges regarded as original: over a period of ten years we carried out a thorough survey on some sixty lutes (and on some bowed instruments) from several European collections. About half of them have bridges we thought we could trust to be original.  The measuring of the bridge-holes was carried out with accuracy, using rods of increasing exact diameters thus we have verified the maximum passing diameter. It will be worth mentioning that by so doing we do not obtain the actual string-diameter but that of the hole, which was obviously drilled with a certain empirical oversize.

If we leave aside the loaded gut hypothesis, natural unloaded gut bass strings fitting such small diameters would have to work under a mean tension of about 1.2-1.3 kg: this is the equivalent of a modern lute strung with a tension of 3.0 kg per string and then tuned down 8 or 9 semitones - just try it once on your all gut strung-lute!

(see Ephraim Segerman: ' On Historical lute Strings  Types and Tensions', FOMRHI bull 77,  October 1994 pp54-7; in this work the actual string diameter was considered equal to the 85% of the maximum passing string hole-diameter).

There are only two options for such small holes:
    1) Only the basses worked at a much lower tension.
    2) Lutes were generally very low strung throughout.

The Mary Burwell lute tutor (ca. 1670):
“When you stroke all the stringes with your thumbe you must feel an even stiffnes which proceeds from the size of the stringes".

John Dowland ('Varietie of Lute Lessons', di Robert Dowland, 1610):

"But to our purpose: these double bases likewise must neither be stretched too hard, nor too weake, but that they may according to your feeling in striking with your thombe and finger equally counterpoyse the trebles".

It has to be borne in mind that with a tension of about 1.2 kg gut basses not only  hardly give any sound at all, but also feel more like rubber bands and are very hard to control by the thumb of the right hand. 

Point 2) is likewise not tenable: with a mean tension of 1.2 kg the first two or three courses would require such small diameters as to be technically impossible to produce (for example, the first three courses on D-minor baroque lute with a 70 cm string length at a-415 Hz pitch would be: 1^st = .25 mm, 2^nd = .30 mm, 3^rd = .40 mm).

In other worlds they are much more thinner than allowed by a fundamental string making rule in the 16^th century, i.e. one single whole lamb's gut must be employed to produce a treble string as described, for instance, by Athanasius Kircher in his ‘Musurgia Universalis’ (Rome 1650).

Our tests shows out that, starting from one single whole lamb gut (as A.Kircher suggested), gauges had just an average of  .45-.48 mm, not less.

Also in this case it must be noted that the basses would result too slack and dull.   

 b) The remarkable performance of all-gut basses in use towards the middle of the 17^th century as opposed to the poor quality of bass strings in use in the first half of the 16^th century (see Virdung, 1511, and Galilei, 1568): here is what we read in the Mary Burwell lute tutor (c.1670), about the all-gut basses on the lute with short extension: ‘...the confusion that the length of sound produce it alsoe..’  and ‘...every basse sound make a confond with every string...’ and, talking about the eleventh course ‘...the lutemasters have taken away that great string because the sound of it is too long and smothis the sound of the others’.

And finally Thomas Mace (Chap. XLII, p. 208): "This inconvenience [i.e. the power and persistence of sound of the basses which causes confusion and dissonances with the higher registers] is found upon French Lutes, when their heads are made too long; as some desire to have them...". 

c) Mersenne ('Harmonie Universelle', 1636): the 11^th bass of a lute (without extension) can ring up to 20 seconds: ‘...& et que le son des grosses chordes de Luth est apperceu de l’oreille durant la sixsieme partie, ou le tiers d’une minute...’  a  performance that’s hard to obtain even with a modern wound string, never mind by a thick rope-like string. Here, though, we can’t hide the feeling that Mersenne might have somewhat exaggerated!

d) Iconographical sources:

1) Lute bass strings are painted with very thin gauges: they fit quite well with the bridgehole diameters

2) Where we find coloured basses, they are always coloured in a homogeneous way and exactly where we have to emply, today, wound strings, i.e. from the 6^th course down. The chromatic transition is not a gradual one, i.e. strings do not get darker and darker according to the increasing thickness of the strings, but by sudden changes, from yellowish higher strings to completely different colours.

Such dyed strings must have also been quite supple; see the details of the bass string knots at the bridge:

e) Dowland (1610): the fact that he prescribes a unison 6^th is a strong suggestion that the basses of his time possessed a high acoustic performance, unknown before and unthinkable in a rope-like string. In practical terms, his 6^th course strings must have been thin enough to grant a lower Inharmonicity index and thus allow the use of unisons.
   

f) Mace (1676): the best lute bass strings in his time were the dark red Pistoys ('...dyed in a deep dark red colour...’). 

We believe that Dowland is not referring to the third Sort  (i.e. Basses) at all. He is describing the Meanes and explains that even if they are thicker than Trebles they are still translucent.

About Basses proper, which we shall treat later ('For the greater sorts or Base strings, some are made at Nurenburge, and also at Straesburge...) he says absolutely nothing. We should not oversee the fact that when he describes a given Sort he always uses a capital letter (i.e. Trebles, Meanes, Basses). This is not the case when he mentions 'greater strings', in the above passage, where he is referring to what comes just before the colon (and the colon, when it does not open a list of items, is explicative, to make clear a concept that has just been exposed), i.e. the Meanes.

The translucency question seems purely speculative, anyway: the real heart of the matter lies in the small diameters of historical lutes bridge holes and plausible working tensions.

So, what kind of sound did the basses of the time produce, then?

Here is the only testimony we know of (Edward Benlowes, 1603-76): ‘...still torturing the deep mouth’d Catlines till hoarse thundering diapason should the whole room fill...’.

   Mid 17^th century sources, as just seen, tell us that gut basses like Lyons and Pistoys  possessed a remarkable acoustic exuberance

- unknown on the six course lute Galilei and Virdung complain about - to the point of causing the serious problems of acoustical confusion, even on a lute with short extention, as described in the Mary Burwell tutor and by Mace, which had to be dealt with by giving up the extension and readopting the French lute without extension like the one in Charles Mouton’s portrait, for instance:

.

2) Finding the best suited twisting process to reduce the string stiffness to a minimum (2^nd dimension)
3) ‘Loading’ the gut with mineral compounds (which we regard as the 3^rd and last dimension)

With the development of the third dimention (i.e. increasing at pleasure the specific weight of gut) made around 1570-80, it became possible to open a new musical epoch, through instruments more capable of providing the fundamental, the new role of the basses of both plucked and bowed instruments.
  The appearance of wound strings, in the second half of the 17^th century, was no real revolution: seen from a technical point of view it was only a different and more efficient way to increase the weight of gut.

The all-gut bass strings we reconstruct today (in practice a Venice string loaded with insoluble metallic copper powder) is not only the perfect synthesis of the two different opinions shared by researchers in the field of all-gut bass strings, but also represents the logical evolution of the technological know-how of 16^th century string makers, which we believe, at least in part, to understand.

-----------------

All considerations so far expressed regarding the lutes without extension apply also to those with extension. The only difference lies in what sort of strings we choose for the extended basses.

Let us consider two basic types of instruments:

a) Theorbos and archlutes with long extension and single diapasons
The purpose of very long extensions is twofold: on the one hand we reduce the string diameters for a better acoustical performance (string length and thickness are inversely proportional), on the other hand - and this is probably the more precious advantage - we obtain a noticeably better sustain, an indispensable factor for continuo playing.
 No document gives us any clues about what kind of strings might have been  used as diapasons (apart from Piccinini, who mentions using silver wire for 5^th, 6^th and extended basses, but calls the instrument bandora), but we feel we can exclude loaded gut strings, both on organological (bridge holes diameters) and iconographical grounds .

Here are some 5^th and 6^th fingerboard’s course bridgeholes diameters:

-Chitarrone /archlute “Magno Diefopruchar a Venetia”, (C45) Vienna, Kunsthistorisches Museum: 5th course 1.7mm gauge both string holes of the course; 6th 1.9 mm both string holes of the course. Vibrating string lengths: 6x2=67 cm; 8x1=142 cm.

-Theorbo “1611/Padoua Vvendelio Venere”, (C47) Vienna, Kunsthistorisches Museum: 5th course 1.3 mm to the bass side string; 1.4 mm for the octave. 6th course: 1.5 mm for the bass side, 1.3 mm for the octave side. Vibrating string lengths: 6x2=76 cm; 8x1=121 cm.

-Chitarrone “Matheus Buechenberg/ Roma 1614”, (190-1882); London, Victoria and Albert Museum; 5th  and 6th course string-holes: no.64 drill (*) both strings of each course. Vibrating string lengths: 6x2=89 cm; 8x1=159 cm.

-Chitarrone /archlute “Andrea Taus, Siena 1621”, (5989-1859); London, Victoria and Albert Museum: 5th course both string-holes no.58 drill (*). 6th course: 1/16 of inch (~1.58 mm) to the bass side hole; no.58 drill (*) for the octave side. Vibrating string lengths: 6x2=67 cm; 8x1=143 cm.

-Chitarrone by anonimous, (7755-1862); London, Victoria and Albert Museum; 5th and 6th courses: all string-holes no.58 drill (*). Vibrating string lengths: 6x2=70 cm; 8x1=148 cm.

-Chitarrone “Christoph Koch zu dem Gulden Adtler/ in Veneding Jul. 1650”, (Kat. Nr. 3581); Berlin, Staatliches Institut […], from a letter sent to me by Dr. Annette Otterstedt in 1996 year: “The holes in the bridge look rather wide for metal strings…”. Vibrating string lengths: 7x2=83 cm; 7x1=167 cm.

* The equivalent gauge, in mm, was not specified

The choice falls between strings with natural specific weight, like our Venice or the traditional high twist. It is worth reminding that the Inharmonicity limit of the thickest diapason on a theorbo or archlute was pretty much the same as the 6^th on the 6 course lute. In other words, the product of frequency by string length results in a similar Acoustic Quality Index (faq 15) and long diapasons need not be of the third Sort.

b) Archlutes and d-minor lutes with short extension and paired octave basses
We have no historical sources to suggest what strings ancient lute players used as diapasons, we must therefore proceed by exclusion.
The use of octaves on extended basses would suggest that it was necessary to remedy a loss of acoustical quality. Logic would suggest non-loaded gut strings, at least as long as the string diameters fall within about 1.4 mm (i.e. an average 6^th on a 6 course lute).

 Information about the string disposition on 4^th, 5^th and 6^th courses is very scanty (courses below the 6^th always had a paired octave). Dowland prescribes unisons down to the 6^th course included. Iconographical sources, on the other hand, show the use of an octaved 4^th course even on 10 course lutes (see Terbruggen, ca. 1624, in the National Gallery in London), whereas some rare sources show a unison 4^th, while 5^th and 6^th have octaves (see Rutilio Manetti, ca. 1624, in Dublin).
William Barley (A New Booke of Tabliture, 1596) recommends using octaves on 4^th, 5^th and 6^th. John Johnson, Francis Cutting and Anthony Holborne hint that, in the second half of the 16^th century in England, the use of octaves was not at all uncommon.

b) 11, 12 and 13 course D-minor lutes (with and without extension)
All historical evidence we know of (e.g. Perrine’s Pièces de luth, 1680: '...les 3. 4. 5. sont doublées d’unissons, et 6. 7. 8. 9. 10 et onze sont doublées d’octaves'), both written and iconographical, show that octaves were in use from the 6th course (included) down. Presently we have no evidence whatsoever of unisons having been used on the 6th course.

c) Double strung theorbos 
To our knowledge, here are no written sources on the subject. A survey carried out on bridge holes shows that both holes on 5^th and 6^th courses have the same diameter. Unisons would be expected.

d) Archlutes with both long and short extension
Again, no written sources on the subject, as far as we know. Iconographical sources show both octaves (e.g. Anton van Dyck’s portrait of a lute player, ca. 1630 in the Prado museum in Madrid) and unisons on 5^th and 6^th courses (e.g. anonymous portrait of a lute player, North Italian School, ca. 1720, in the article by Robert Spencer).

Lute bass strings in the 18^th century   

 We believe that the late 17^th century lute was not affected, as a rule, by the appearance of wound strings, which were developed in the second half of that century. Surviving treatises (Thomas Mace, 1676, and James Talbot, ca. 1690, amongst others) point towards all-gut basses.

  The machine for making overspun strings. Encyclopédie, ou Dictionnaire raissonné des sciences, des arts et des metier [...], Briasson et al., Paris 1751-80.

Another source mentioning wound strings on the lute is François Alexandre Pierre de Garsault’s Notionnaire..., Le Luth, Planche XXXVI - ‘Accord des basses cordes simples filées...’, Paris, 1761:

    In the 18^th century, wound strings can be grouped into three categories, all built around a gut core (at least up to the second half of the century - the earliest mention of wound on silk known to date is after 1760):

    1. double wound (i.e. a first winding is covered by a second one)
    2. close wound
    3. open wound (called demifilé by the French).

    Type 1. was probably used for bowed instruments with particularly short string length and low pitch (violoncello da spalla &c.).
    Type 2. would seem to be the right one for the 13 course lute:

but we would rather opt for type 3. upon an important consideration: from what we know about the metallurgic technology of the time it seems that it was not possible, at least in the common practice, to produce wires thinner than about .12 mm (see for example James Grassineau 'A Musical Dictionary'  London, 1740 under the world 'wires'; see also the Cryselius's wire gauges and the 18^th Nuremberg's  wire gauge tables).

As a consequence we think that it was not possible to produce wound strings for the 6^th, 7^th and 8^th courses for the d-minor lute, even if we reduced the gut core to the point of completely unbalancing the Index of Metallicity and the mechanical stability of the string (faq 45).
 

An open wound string was simple and efficient: by spacing the winding it was possible to come around the wire diameter problem, with one limitaton: here, too, it was the thinnest available wire that had to be employed in the production of the 6^th string.
 What we are saying here is that open wound strings were not a transitional  phenomenon, in the sense of bridging over the gap between all-gut and close wound strings, they were a clever stratagem that made it possible to come around the technological limitations of the wire manufacture of the time.

How do we know that open wound strings were really used in the 18^th century lutes?

One piece of evidence and several probative elements point in that direction:

    a) The direct evidence comes from the pieces of strings on a Lute by Raphael Mest. Half wound strings were in use only in the 18th century and it is hard to imagine a later addition of this particular kind of string on an instrument that had already fallen into disuse.

b) A strong vertical ovalization of bass bridge holes and signs of abrasion on the upper plate edges on original 18^th century bridges: an open wound string does not run smoothly (not as smoothly as a close wound does) but acts like some kind of file on the hole edges. We hardly find this kind of wear on modern lutes, for instance, where we use close wound basses.

   

   c) The diameters of bass bridge holes on 13 course lutes with bass ‘rider’ are rather compatible with open wound strings, while holes for the noticeably thinner close wound strings would be expected to be smaller (a half wound string for the 13^th course with a working tension of about 3 kg presents a diameter of about 1.6 mm against a statistical average of 1.8-1.9 of hole- diameter as measured on original lutes). Unfortunately, this evidence do not work with the swan-neck lutes.

Table 2

Massimum passing hole diameters on some swan neck d-minor German lutes 

  d) German d-minor lutes with the bass rider or with swan-neck keep their octaves on the basses: half wound strings, experimentally produced according 18^th century instructions for guitar strings (core of the same diameter as the octave and spacing between the spires the same as, or slightly more than, the diameter of the wire - see Le Cocq, 1724), present an average specific weight comparable to that of the basses we think previously in use, i.e. about twice that of natural gut. Such a strings produce no particularly bright sound, nor do they possess a good sustain: hence the use of octaves.

 e) No original intabulation expressly requires damping the basses. It is reasonable to deduce that the strings did not possess a good sustain.

 f) In rare cases, iconographical sources from the 18^th century show lutes with white basses (silver wound?) as opposed to the yellowish colour of the upper courses and, at least in one case, we see something clearly looking like a half wound string:

  
Joahn Kupezky (1667-1740), luteplayer. In the original the last bass string seem to be an half wound tupe



  13 course lute with bass rider: detail on the white bass strings
(Courtesy by David van Edwards)



Antoine Pesne (1678-1758): portrait of Eleonoire von Kayserlingt, 1740 ca: in the original bass strings are deep red coloured (!)

It must be noted that the use of half wound strings fell into disuse exactly at the time when some specific types of instruments did, like the 5 course guitar, the 7 string bass viol and probably also the lute. Half wound strings simply became unnecessary on bowed
instruments tuned in ‘large’ intervals, i.e. in 5ths, where it was possible to switch directly from a plain gut to a close wound string.

   Compared with its Baroque predecessor, the 6 string guitar underwent a string length reduction of about 10-12 cm, while the working tension of each single string was increased to about the same as the sum of the two string in a course (thus, incidentally, keeping also the feeling just about the same) and gave access, for the first time, to a close wound 4^th string, this time on a silk core.

These were, we believe, the decisive steps towards modern close wound guitar stringing as we know it, and brought the use of half wound strings to an end. The adoption of a silk core, a superior material to gut both according to the sources of the time and to modern practice (being more supple and more resistant than gut, silk makes the use of thicker wire possible) opened, in our opinion, the way to the guitar’s 6^th string.

                                            Juan Guerrero, Paris 1760 : silk wound bass strings were better than those made with a gut-core

  But what were the typical features of a half wound string?

Again, let’s have a look at historical sources:

a) The space between the wire spires was the same as, or slightly wider than, the diameter of the wire used for the winding - hence demi : half. Here is Le Cocq’s description (Recueil des pièces de guitare composées paer Mr. François Le Cocq, Brussels, Bibliothèque du Conservatoire Royal de Musique, Ms Littera S, n. 5615, 1730. Ch. Des chordes, 1724): Se charge les deux octaves que se mets au quatrieme et cinquieme rang d’un fin filet de laiton ou d’argent, ce dernier en vaut miex ... se ne les charge qu’à demi: c’est à dire qu’il reste un espace vide à la corde, de la grosseur dudit  filet ou même un peu plus...'.

    These few but important indications fit perfectly the half wound string leftovers on Raphael Mest’s lute. Therefore strings with very open winding and/or  embedded wire in the gut core have no historical justification.
 

Wound on silk basses for the lute
Wound on silk basses could have been used after 1760, provided they were half wound, but unfortunately we have no historical evidence for that. In any case the limit set by the metal wire technology of the time is still valid even when a silk core is employed.

The discovery of  what we believe is the ‘true’ meaning of the role of the string Sorts led to a better understanding of what a correct lute stringing should be like.

The most remarkable point, certainly worth emphasizing, is the frequent lack or ‘relay’ in stringings with synthetic materials: strings of the third (wound) Sort  ‘invading’ the space pertaining to the second Sort, long theorbo diapasons strung with strings of the third (again, wound) Sort, mid register strung with strings produced with a method pertaining to the first Sort, i.e. rather stiff, which cause a noticeable loss of acoustical quality. Not to speak of the indiscriminate use of strings of the third Sort as octaves. In one word, if we exclude aluminium wound strings and carbon strings (with which we try and fill the mid register gap), present  lute stringings lack appropriate synthetic strings for the mid registers.
In the case of all-gut stringing we must once more stress that strings of different Sorts  and their maufacturing processes are absolutely not interchangeable.

The string maker has very limited leeway indeed: putting together a good set of gut string for the lute looks more like a tricky narrow path than a wide and easy highway.

After all, hasn’t this always been part of the fascination of the Dolce Strumento?

Zophany, 1770 ca. The Sharp family: see the white basses

Here the question arises: if basses were really made of plain gut, why was the extension limited to 95-100 cm, why did they not make them longer for better performances; say 1.20-1.30 cms?