Nailed Necks

Earlier this year I became curious about why early European luthiers - violin, guitar and lute makers - attached the necks of their instruments, not only with glue, but with forged iron nails driven through the neck block. I was also curious to find out if this practice - historically - applied to ouds (the oud, supposedly, being the direct ancestor of the European lute).

So far no evidence has come to light confirming nails being used for this purpose by oud makers. Instead - at least as far as the earliest surviving (19th C) ouds are concerned - necks are joined to the neck block with glue and a dovetail or a simpler dowel joint. The question is does the dovetail extension or dowel provide extra strength to the neck joint or is it just a device to ensure correct alignment of the joint when being glued - or a bit of both?

The same question applies to the nailed joint of the Western luthiers.

So, some brief trials were undertaken to test the relative strength of a glued/nailed joint - during the hot, humid conditions of this past summer. The results may be of some interest to the oud makers of this forum.

Here are a couple of images of an early 17th C lute neck block showing a large headed iron nail driven though the neck block and a similar image for a Baroque guitar - again with a single nail driven through the neck block. Images kindly provided by luthiers Daniel Sinier and Françoise de Ridder

http://sinierderidder.free.fr

Violin makers such as Antonio Stradivarius (17th C) used three or four nails (dependent upon the size of the instrument) to attach the neck.

Here are another two examples of nails used by the early lute makers,
The first is a sketch made from an X-ray image of a late
16th C lute by Michielle Harton - showing a single iron nail driven through the neck block into the neck. Note the irregularity of the blacksmith forged nail.

The second image shows a lute by Hans Frei (mid 16th C) that has been modified later in life (17th C) from its original 6 courses to 10 courses requiring a wider neck. Here the luthier - believing perhaps - that if one nail reinforced the strength of the neck joint then three nails would be even better!

Luthier Marcos Kaiser Mori has kindly provided some more examples of nailed neck joints in lutes.

http://www.marcoskaiser.com

The first shows a single nail in the neck block of a lute by Wendelin Tieffenbrücker, dated 1606, that has later been converted to a chitarrone by the addition of an extended neck.

The second image shows multiple nails used in a Theorboed lute (again an extended neck) by Martin Kaiser, 1609, instrument #E 24, Musée de la Musique, Paris. Three nails are used in the neck block and two nails to attach the extended neck.
Again the luthier appears to have faith in the power of the nail!

Note the irregularity of the hand made forged iron nails.

More to follow.

To judge from surviving instruments (19th C and later) oud neck joints are generally made perpendicular to the plane of the fingerboard whereas those of the early lute (at least beyond the Medieval period) slope at an acute angle. The reason is that lutes from the late 16th C had up to 10 courses of strings requiring a wider neck than the 5 or 6 course oud. In order to maintain a practical depth to the neck (D) at the neck joint the neck joint was cut at an appropriate angle to give the required fingerboard width at the neck joint.

Such a sloping neck joint presents difficulties when it comes to gluing the neck to the neck block as the neck will slide relative to the neck block as a clamping force is applied unless some means is taken to prevent this movement.

Although the situation in gluing a perpendicular joint is potentially less of a problem, the use of a sliding dovetail neck joint is often found on old ouds. One obvious advantage is that the neck may be fitted and adjusted as a 'dry run' prior to final gluing - always a good thing with glued joints.
Although the sliding dovetail joint would have been a practical solution to gluing a sloping lute neck joint - it was not used, as far as I am aware, by lute makers of the 16th and 17th C.

In Germany, the lute making tradition was revived in the mid 19th C, by popular nationalistic demand, in the form of guitar-lutes, instruments, however, that had no relationship to the lightly built lutes of the 16th and 17th C being just guitars that were lute shaped. By the mid 20th C - with a developing interest in rediscovering the huge surviving repertoire of lute music of the 16th and 17th C - heavily built instruments were being produced that were more authentic looking but were still designed essentially for use by classical guitarists. These instruments persisted into the early 1970's until authentic copies of original lutes started to become available.

'Die Guitar un Ihr Bau' by Franz Jahnel was published in 1963 and represents the modern state of knowledge, as it then was, about the construction of fretted instruments such as the guitar, mandoline, lute etc.
It is interesting to note that the detailed information presented about the lute shows a neck joint with sliding dovetail (see attached image) although the author does state that this type of joint is only one possibility. He also states that Medieval lutes had neck joints with a tenon (although no lutes of that period survive to confirm this - one way or another).
So was Franz Jahnel passing on some unwritten traditional information about lute neck joints of the 16th and 17th C? Unlikely, alas, as his scale drawing claiming to represent an early lute appears, by its design, to be nothing more than one of those heavily built, non authentic, instruments of 1960's vintage.

More to follow.

A modern method of gluing the neck of a lute to the neck block is to use a round headed wood screw to pull the glued joint together - acting as a clamping force until the glue has cured. Usually the screw is left in place after the glue has cured.
This method can be used for both slow setting synthetic glues or for traditional hot hide glues if the joint is first partially assembled, the glue run into the joint and the screw immediately tightened with a power operated driver.

A problem with modern wood screws is that that the unthreaded shank of the screw is smaller in diameter than the outside diameter of the threaded portion (not the case with the old fashioned cut thread screws where the unthreaded shank is the same diameter as the threaded portion).
This means that a clearance hole must first be drilled through the neck block to allow the threaded portion of the screw to pass unhindered. This in turn means that the neck must be restrained with a guide block - temporarily fixed to the neck block - to prevent the neck sliding or rotating relative to the neck block when being clamped in place by the wood screw.
Nevertheless, this arrangement like the sliding dovetail joint of the oud allows adjustments to be made to the neck assembly before being finally glued in place.

The question now is - how can a nail (or nails) be used to replace the function of a wood screw for this type of joint? What might have been the step by step procedure used by the ancient luthiers - perhaps not as straightforward as it might at first appear?

Only one way to find out!

The first step in this trial is to make the nail - or 'spike'. The blacksmith made nails of earlier times were hand forged by heating a strip of iron to bright red temperature - until soft - and then hammering the nail to the required shape on an anvil. The flat head of the nail was made by first 'upsetting' the end of the nail to form a 'bulge'. The nail was then placed in a 'heading' tool (after again heating to red heat) and the flat head then formed by hammering - like a rivet.

For convenience I have started with commercially available 'square' cut nails that were once standard (until about as late as the 1950's) until the now familiar round wire nails became generally available.

The head of the nail was heated with a propane torch and then formed in the heading tool. This is an original blacksmith's forged heading tool, picked up at auction, but a less fancy version can be easily made by drilling an appropriate diameter hole in a thick bar of steel.

The tapered cut nails of the 19th C were not pointed. The blunt end reduced tendency to split the wood.

The finished nail will now be tested in a trial neck/neck block assembly with vertical joint - as found on ouds and pre 16th C lutes.

The test neck/neck block assembly was made from pine/cedar with a vertical neck joint. The neck block was first drilled with a small diameter pilot hole to guide the nail perpendicular to the neck joint. The neck was not pre-drilled and the test was undertaken 'dry' without using glue

With the assembly positioned in correct alignment and with the neck held in a vice the nail was then hammered through the neck block and into the neck. The nail cut an oblong section hole in the neck block and did not split the neck. However, due to non uniformity of the nail and its tendency to follow grain direction in the neck - the neck joint was found to be slightly out of alignment across the joint face when the nail was fully driven home. This was corrected by withdrawing the neck slightly and applying force on the nail to bend it (the iron of the nail is quite soft) to bring the joint faces to a perfect fit. While it would then be possible to withdraw the neck again to allow glue to be applied to the joint before hammering the nail finally in place to complete the assembly, I doubt if this is the procedure that was followed by the ancient lute makers.

The nails used on the surviving lutes - to judge from the x-ray images - were thinner in section than the test nail and so more easily self bent when being driven into place. Also a preferred neck material was Lime wood (veneered with hardwood) - relatively soft and light with little pronounced grain that might cause a nail to wander out of alignment.

My guess, on the basis of this limited test. is that the old lute makers first drilled a pilot hole in the neck block. Then with the neck held in a vice and the assembly held in correct alignment, the nail was driven through the neck block into the neck for a short distance - just to register the nail position relative to the neck. The neck was then withdrawn, glue applied to the neck joint and the nail finally driven home to complete the assembly.
If a single nail failed to bring the joint faces together in perfect alignment then additional nails may have been used to correct the situation.

The primary purpose of the nail with its large diameter head is, therefore, to quickly apply a clamping force and at the same time keep the joint assembly in correct alignment. A secondary benefit is that, once in place, the nail will prevent joint movement and possible separation when the hide glue is subject to humidity cycling over time.

A second trial using a thinner 'square' section forged nail was undertaken this past summer to test the relative survival rate of a plain, un-reinforced, vertical neck joint made with hide glue to one reinforced with a nail - a destructive test, both joints being subject, under load, to extreme conditions of temperature and relative humidity.

More to follow.

For the second trial a thinner nail was prepared - again by reforging the head of a cut nail (for convenience). I use old style cut nails still made for authentic historical carpentry work by the Tremont nail company.

This time around a Rosehead Clinch nail 3 inches (75 mm) long was used the head being heated with a propane torch and the head increased in diameter by forging in the heading tool.

The finished nail is much thinner in section than version #1 so should allow easier seating of the joint faces when driven through the neck block.

Quote: Originally posted by jdowning

... but not as a means to reset a neck that had 'pulled up' under string tension over time.

Fine, I'm staying tuned

The test rig for the second trial was a straightforward affair with two simplified 'necks' of pine (cut from the same piece of wood) glued side by side to a pine 'neck block'. The 'necks' were glued to the block with high bloom strength (fast setting, high strength) hot hide glue in an inverted position - one being also nailed in place.

Prior to gluing the joints the end grain of the 'necks' was sealed as is usual practice. This is to prevent loss of joint strength due to the glue being absorbed into the end grain. My method for sealing end grain is to coat the joint surface with hot glue and then to immediately scorch the glue with a hot iron to form a brittle surface crust of glue. The crust is removed by scraping, filing or sanding, the joint verified again for precise fit, and then glued.

To obtain maximum joint strength the two joint faces must be brought into intimate contact with adequate clamping pressure - maintained until the glue has fully cured. Clamping pressure is provided by the nail the case of the nailed joint. However, for the other joint no external clamps were used the joint surfaces being held firmly together with strong hand pressure for a couple of minutes until the glue had gelled and then left undisturbed to fully cure for 24 hours.

For the nailed joint a pilot hole was first drilled through the 'neck block' and the nail driven part way into the 'neck' with the assembly held in correct alignment. The 'dry' joint was then separated, hot glue applied and the nail driven fully into the neck - the prepared nail hole providing correct registration of the joint and preventing the joint sliding out of alignment. The thinner section, more flexible nail allowed the joint to seat perfectly (unlike the nail in test #1).

The block was then bolted to a wall, under cover, in an exterior building and each 'neck' was loaded with a 1.5 Kilogram weight placed at a distance of 20 cm from the joint and the rig exposed to the fluctuating temperatures and relative humidity of a Canadian Summer (hot and humid!)

As this test was intended to be an accelerated, destructive test of the joints, the loading of the joints was intentionally designed to be significantly more severe than would be experienced in an oud (or lute).
More on that next.

Quote: Originally posted by jdowning

... each 'neck' was loaded with a 1.5 Kilogram weight ...

The attached simplified sketch compares the forces acting on a neck joint on an oud (or lute) with vertical neck joint and a lute with sloping neck joint. For this example string tension is taken as 40 Kilograms with a neck length of 20 cm and string height above the neck joint of 3 mm.

The string force can be resolved into two components at the nut one acting parallel to the fingerboard (FC) and the other vertical to the fingerboard (FB). The latter force, acting at a distance of
20 cm from the neck joint creates a bending moment on the neck joint tending to separate the joint at point B. However, due to the low angle of the strings relative to the finger board surface the force FB is quite small - about 0.5 Kg - giving a bending moment of about 10 kg.cm.
(Force FB is also equivalent to the shearing force FC acting on the joint surfaces tending to cause the joint to slide apart).
The predominant component force FC acting in a direction normal to the joint surfaces is just a bit less than the string tension and is a compression force tending to clamp the joint surfaces together.

In the case of the sloping neck joint of the lute the compression force FC of about 40 Kg acting parallel to the surface of the fingerboard is resolved into two component forces - a compression force of about 36 Kg (Fc) acting normal to the joint surfaces and a shear force of about 18 Kg (Fs) acting parallel to the joint faces.

For the second trial (which simulates the vertical neck joint case) the suspended load of 1.5 Kg (FB) is, therefore, 3X greater than the calculated load produced by a total string tension load of 40 Kg. Furthermore the tests joints, in addition to the resultant bending moment (30 kg.cm) experience a shear force of 1.5 Kg - all without the benefit of any sustained clamping force - except for that applied by the nail in the case of the nailed joint.
So the loading conditions on the glued joints are more severe than would be experienced by an oud or lute under string tension - so quite rapid failure of the joints under conditions of fluctuating high ambient conditions of temperature and relative humidity might be expected? And which joint will be first to fail?

Quote: Originally posted by jdowning

And which joint will be first to fail?

Quote: Originally posted by jdowning

What would be the advantage of a much longer nail ...

The #2 test rig was first set up in my granary loft where there is free air flow (good for seasoning wood) and daily temperatures were elevated. In order to keep Relative Humidity as high as possible (low to high 70's) the rig was covered with a plastic sheet to form a tent and a bowl of water placed underneath.
As no failure or impending failure of the joints was in evidence after several days, the rig was later moved to my barn (this time without the tent and water bowl) where temperatures were cooler and RH relatively higher (but still within the upper 70's range).

After a total of 27 days (August 02 to August 28), with still no sign of joint failure, a plastic tent was again placed over the rig with a pan of boiling water underneath for a period of two hours each day during the afternoon in an attempt to raise the RH to above 80% (an RH of 82 % was the highest achieved by this method). In order to further promote joint failure the suspended load on each 'neck' was increased to 2.75 Kg.
On September 02 the un-nailed joint was found to have failed overnight with no sign of failure of the nailed joint.

So although the nailed joint 'wins the contest' - as anticipated - what I found surprising was that the un-nailed joint survived for so long without failing during the period when ambient conditions ranged from 37% to 100% and temperatures ranging from 12°C to 25°.
The attached chart summarises the daily temperature and RH fluctuations taken from the Weather Bureau hourly records for August and September. Typically the lowest temperatures and highest RH occurred each day during the early hours of the morning reaching a minimum value around mid day when temperatures were at a maximum.

The attached image shows the failed joint and it can be seen that there is a significant area where failure had occurred in the glue film (no wood fibres in evidence). The joint face measures 50 mm wide by 25 mm deep.
The nailed joint was forcibly broken by hand and - apart from a small area lacking wood fibres - it can be seen that the wood of the joint failed before the glue (i.e. the glue was stronger than the wood).
It might be concluded from this single test that the nailed joint was inherently stronger because of the more effective, greater clamping force applied by the nail when the joint was first glued together. A likely secondary benefit of the nail was that it helped to stabilise the joint and prevent separation under fluctuating conditions of temperature and RH.

It should be noted that the critical condition of RH is 84% where hide glue loses about 33% of its strength. However this condition is only achieved in a joint if the glue and surrounding wood is saturated at a stable 84%. This condition is unlikely to occur under the short term excursions of RH as experienced by the joints under test. So this is good to know!

The next test will be to determine the practical methodology for nailing a lute neck joint with sloping faces. Will the nail alone provide correct registration of the joint when fully driven home or will temporary guide blocks be required as in the case of a screwed joint of this type?
More to follow

Other historical examples of nailed neck joints can be found in early violins where the neck is butt joined to the ribs (not constructed in the Spanish way where the ribs are glued into slots cut into an integral neck/neck block).
It is of interest to note that the early luthiers might construct both plucked and bowed instruments - so a lute maker would also make viols or violins. It is perhaps no surprise then that some techniques were applied by the luthiers to both types of instrument.

Roger Hargrave is a master violin maker with extensive first hand experience in the examination and measurement of some of the most famous of the surviving instruments of the violin family by makers such as Stradivarius and Guaneri del Gesù. Fortunately for the world of instrument makers, he is willing to share his extensive experience in a series of articles about the violins that he has examined, their characteristics, materials and methods of construction etc. - all written in a clear, easy to read manner. These articles are freely available - in English or German - at:

http://www.roger-hargrave.de

Click on the 'Library' button

Having briefly looked at the content of the articles it did not take long to find a reference to nailed neck joints in the article about the work of Guaneri del Gesù working in the early 18th C Italy.
The neck joint of a violin by Andrea Guaneri was shaped to fit the upper curve of the ribs with a gouge and so was not perfection in its fit - yet it did the job.
Stradivarius used from three to five nails dependent upon the size of instrument (violin or cello). Del Gesù used five on a violin but as Roger Hargrave proposes two of these may may have first been driven into the neck and with their heads clipped off prevented the neck from moving relative to the neck block when pilot holes were being drilled for three other nails.

The attached image by Hargrave shows a plan view of the nails driven through the neck joint. In his article about Stainer violins there is also an X-ray image showing the position of a nail in the neck block of a violin.
The third image shows the tips of three nails visible in the heel of the neck of the 'Harrison' violin by Stradivarius, 1693 in the Rawlins Gallery, National Music Musem.

Another twist to this tale is the article "The Case for Casein" where Roger Hargrave suggests that casein glue (a slow setting strong waterproof glue made from cheese and lime) might have been used by some luthiers instead of hide glue. Although known since ancient times a modern version of casein glue was - until recently - used for wooden airframe structures the most famous example being the De Haviland 'Mosquito' aircraft of World War 2

.... and here is another X-Ray image of three nails in the neck joint of a Guaneri viola - no other information is available at present to identify the source of the image.

The purpose of the third and final test is to determine the best method for nailing a sloping neck joint typical of a lute. The reason for sloping a lute neck joint is to achieve a wider fingerboard at the neck joint (in order to accommodate a greater number of courses) without increasing the depth of the neck. The slope chosen for the test is about 20° from the vertical giving a width of 60 mm for a depth of 25 mm (for a vertical oud style joint the width would be 50 mm for a depth of 25mm).

The 'neck block' for the test rig was cut from cedar and drilled with a pilot hole to guide the nail. The 'neck' was made from so called American Yellow Poplar (which is not a Poplar species at all but Magnolia, a very large, beautiful tree from the Southern States). It is a softer wood with less pronounced grain than Pine. I did not have any Limewood which would have been the most historically correct choice.
The 'neck joint' was left with a sawn finish as this was just a quick 'rough and ready' trial.

The nail used in Test #2 was modified by filing into a more uniform square section with a pointed tip - to simulate original nails seen in the previously posted X-Ray images. The modified nail was then heated to red heat with a propane torch and allowed to slowly cool to ensure that the nail was fully softened (annealed).

The first step was to hammer the nail through the neck block until the point just protruded. The neck was then held correctly positioned and aligned with the neck block and the nail hammered into the neck for a short distance in order to register the neck with the neck block. The joint was then pulled apart leaving the nail protruding. (This is the point in the procedure where hot hide glue would be applied to the joint faces. With the glue applied the joint would then be reassembled with the protruding nail providing correct register).

With the test joint reassembled the nail was fully hammered into place. The joint did not appear to move out of alignment and appeared to be correctly seated but - unfortunately - the neck split and so it was difficult to fully assess the outcome. Well seasoned Magnolia wood may be more prone to splitting than anticipated.

The test will, therefore, be repeated this time again using White Pine for the neck. The nail will also be shortened by about 20 mm (3/4 inch) as it is likely that the excessive length contributed to the splitting. From this experience the length of the nail should be about 2X the maximum thickness of the neck block.
Also the joint faces will be finished smooth to ensure that the rough sawn faces of this test did not contribute to the joint faces not sliding as the nail was driven home.

For the repeated test #3 a new neck block/neck assembly has been made up - this time with a neck joint slope of 30° to the vertical, both joint faces finished smooth and with the nail shortened to 57 mm in length (2X the maximum thickness of the neck block).
With a maximum depth of the neck of 25 mm the sloping joint gives a maximum width at the neck joint of 85 mm.

The procedure for nailing was as previously reported. With the nail driven fully through the neck block and into the neck the neck joint faces were perfectly seated but the neck had moved upwards relative to the neck block by about 2 mm.
Splitting the neck to reveal the nail shows that when being driven into the neck the nail had bent downwards relative to the top surface of the neck.
This suggests that a pilot hole must first be drilled for the full length of the nail as a guide to prevent the nail bending.
To drill the pilot hole in correct alignment, the pilot hole in the neck block will be drilled first. The neck and and neck block will then be temporarily glued together in the correct position with a spot of hot hide glue applied to the joint face. The pilot hole will then be completed (once the glue has set) using the pilot hole in the neck block to guide the drill bit into the neck. The joint will then be broken apart and prepared for reassembly, gluing and nailing.
This procedure will be tested next.

Drilling a full depth pilot hole did not prevent relative movement of the joint faces as the nail was hammered in place - if anything it made matters a bit worse with the joint being displaced by about 3 mm or so (but with the joint faces gain brought into perfect contact)

After the test, the neck was again cut open along the axis of the nail to reveal the position of the nail. It can be seen that as the point of the nail entered the pilot hole in the neck it made contact first with the top of the pilot hole - the wedging action of the point then initially pushing the neck upwards (point A). After the nail had entered about half way into the neck the nail then began to run straight again.
Had the wood grain run vertically in the neck (as it normally would) then it is likely that tendency of the nail to wander would have been minimised - largely unaffected by the difference in hardness between the early and latewood cells.

To conclude, the best results might be obtained by:
1) Using a wood with little or no pronounced grain such as Limewood for the neck (veneered on the outside with Ebony or other harder wood).
2) Drill a pilot hole in the neck block only - not in the neck.
3) The point of the square nail should be flattened to minimise chance of splitting the wood of the neck.

Although not tried in this test, a guide block temporarily glued to the neck block - to prevent relative movement of the neck joint faces as the nail is being driven into place - would likely be beneficial as it is in the case of a screwed joint).

That's a good question Chris but I am not sure if anyone knows the answer with certainty.

Nailed neck joints were employed until about the mid 19th C and then abandoned in favour of dovetailed neck joints for violins and guitars (i.e. for those instruments not built in the Spanish way). The attached image shows this type of joint in a late 19th C violin.
Violin maker Ed. Heron Allen - writing in 1885 ('Violin Making - as it was and is') confirms that "in olden days the neck was often fixed to the back with a nail or screw, an expedient extremely deleterious to the fiddle" (although he does not explain why this was, in his opinion, such a bad practice or why the early violins with nailed necks managed to survive for so long). So the knowledge associated with the early nailed neck joint technique has been lost - and it is not otherwise possible to tell from surviving instruments if pilot holes were used.

It is possible that nailed neck joints were once also used on ouds but this cannot be confirmed as no ouds survive prior to the end of the 18th C. Indeed, perhaps this practice originated in early oud construction and so was adopted by the lute makers of the 16th and 17th C?

I am not sure if any lute makers today use nailed neck joints so the only way to find out the best way to do it - is to try to do it!
Some violin makers, who are building replicas of early violins, report having tried the method - some drilling pilot holes in neck and neck block others just in the neck block to initially guide the nail(s).
The difference between a nailed violin and lute neck joint is that the joint on a violin is vertical and the neck made of hard maple whereas the lute neck joint slopes and the neck is made with a soft wooden core veneered on the outside with a harder wood.

I conclude from these trials that - for lutes - it may be best to drill a pilot hole in the neck block only and use temporary guide blocks to prevent any possibility of relative movement of the joint faces as the nail is finally hammered into place.

Quote: Originally posted by jdowning

That's a good question Chris but I am not sure if anyone knows the answer with certainty.

Yes, so here's the next question, that might be interesting

Since when were small thin nails available, that were suitable for this luthery task?

Were they already available in the middle ages? I fear, not ... only large thick nails. Or am I wrong?

Quote: Originally posted by jdowning

So, of course, small thin hand forged iron nails would have been available to the luthier in the middle ages.

Quote: Originally posted by jdowning

However it is impossible to say if oud makers did used nailed neck joints unless there is some confirmation ...

Quote: Originally posted by jdowning

... in Mediaeval times - wrought iron rod used for making nails may have been more difficult and expensive to obtain for those other than the rich or the military.

Multiple nailed neck joints on lutes can be found on surviving instruments of the extended neck type - the extended neck carrying long bass strings that are played open - total number of courses being typically 14 or more
The smallest of these lutes is the 'liuto attiorbato' with a stopped string length of about 60 cm with the largest - the theorbo - having a stopped string length of around 90 cm.

The attached engraving is from Syntagma Musicum by Michael Praetorius published in 'Germany' in 1619 described as "Lang Romanische Theorba or Chitarron" The scale included with the engraving shows that the instrument measured about 7 Brunswick Fuss in height (a Fuss (foot) being equivalent to 285.4 mm) or , in today's standard, about 6ft 6 inches or 2 metres.

Extended neck lutes have essentially two neck joints - the usual neck to neck block joint (B) and the extended neck to neck joint (A).
The bending forces due to string tension acting on both joints is magnified by the extended neck - the longer the neck the greater the bending moment at the joints. Due to the offset bass strings there is a horizontal (towards the bass side) bending moment as well as a vertical bending moment

The smaller extended neck lutes might have only a single nail in joint B and none in joint A. The larger theorbos might be found with two or more nails in both joints - the nails in joint A being driven in through the lower pegbox opening as shown in the attached sketch. It can be seen that the shape of this joint prevents the joint sliding apart when the nail is driven home.
The attached image of a 'Theorbo Neck' is actually the unfinished extended neck of a smaller 'Liuto Attiorbato' that I am currently constructing - just for clarification to show the bending forces and where the nails are positioned on a larger lute.

The use of multiple nails on the larger extended neck lutes tends to confirm that the function of the nails was not only to apply a clamping force to the joints when being glued but also to provide extra reinforcement of the joints to counter the horizontal bending moments at times when the glue had temporarily softened.

But, in practice, there is a downside to this.

The attached images of the extended neck joint of an original Theorbo by Martin Kaiser - kindly provided by London based luthier Marcos Kaiser Mori - illustrate a potential long term problem with the use of nails. I have taken the liberty to add text to the original images - for clarity.
The original instrument is damaged. The extended neck has broken from the neck at the joint (probably due to an accidental sideways impact). The attached x-ray image - in plan view - shows two nails in the joint but in fact there would appear to be three - two nails being applied one above the other which may have contributed to the joint separating at this point when the glue on the bass side of the joint, through age and humidity cycles, had become weakened and brittle (and not reinforced by the nails). Note also the corrosion and consequent disintegration of one of the nails.
Too much of a good thing perhaps?

Note that the pointed wooden sticks in one image are part of an apparatus used by Marcos to measure the original instrument see:

http://www.marcoskaiser.com