Follow-up Discussion of the 'Khotan Bow'

What would be the mechanical advantage, if any, of non-lift off siyahs?

Posted by: Dale Yessak 07/18/2002, 11:01:20

As I posted in a separate question concerning the Magyar bow design, what would be the mechanical advantage, if any, of non-lift off siyahs? I can see that these are essentially siyahs that are braced so as to have already pre-loaded when braced to beyond the lift off stage, i.e. leverage optimized while at brace height, yet such an arrangement would seem to be less efficient than siyahs of a greater angle that provide an actual "camming" action at lift off. In the Kotan bow design, we can see that the long, thin siyahs needed additional bone reinforcement -- perhaps as lateral stiffeners? -- so would this indicate that the design NEEDED to be non-lift off so as to make it more laterally stable?

A separate issue might be the question of whether or not this design petered out as a technological dead end because the benefits of such a complex design did not justify the additional work needed when compared to simpler composite designs.

Posted by: Adam Karpowicz 07/19/2002, 06:57:40

Good to see you here, Dale. You are right in saying the long siyahs would lead to the "high-braced" bow and non-contact siyahs. From my experience it is difficult to make a bow with angled, contact siyahs and at the same time to keep the siyahs long. Such bows are prone to twist and string shedding even at brace, due the leverages involved. Note, other designs with long siyahs, such as some Indian for example, have the angle between string and siyahs very small at brace, no more than some 15-20deg. Bows with angled siyahs, such as Turkish or Tatar have the siyahs short to minimize this leverage and stabilize the bows. Angled bows with longer siyahs, such as Manchu or Mongolian rely on overall stiffness to stabilize - this translates to more massive knees and siyahs.

The Khotan bow with the relatively slender and long siyahs required wide bending sections of limbs for stability plus it was braced high for non-contact. The longer and lighter the siyahs and the shorter the bending limb, the more efficient the bow will be, so the Khotan bow would not be a "dead end" as you suggest. The width of the limbs could be less, I believe, with no loss of stability, on the other hand such a wide limb distributes the stresses over more material, making the limbs less likely to creep (string follow) and allowing the bow to be strung for longer time with little loss of performance, definitely an advantage.

Posted by: Dale Yessak 07/19/2002, 09:19:34

I understand the mechanics (wide limb advantages, etc.), but I'm just curious why this design does not seem to appear elsewhere. It seems like it would be much more efficient and a lot quicker than what we term "conventional" composite designs. Yet it seems as though this particular bow design is a rarity among horn bows. Why would that be?

I wonder if the Khotan design really was a rarity?

Posted by: Stephen Selby 07/19/2002, 22:37:42

I wonder if the Khotan design really was a rarity? The period of the Khotan bow was around 300 AD. All the bows I have seen preserved from that time (four in number), as well as the 'Gansu' bow which came from another region, was of that design. Looking back at my translation of the 'Rites of Zhou' about making horn bows (which may pre-date the Khotan bow by 500 years), there is a lot of stress placed on limiting the amount of horn used to the minimum required to achieve the effect.

Seeing Asian bows of that age from anywhere is unusual; and nearly all the depictions I have seen of these bows in art take a sideways view onto the bow, so you can't really guess how wide the limbs were. I have seen just one Tang bow (around 800AD), and that did indeed have much narrower limbs.

Posted by: Adam Karpowicz 07/19/2002, 12:02:36

As to the bow: I think the Khotan bow was not perhaps so common, because it requires too much work to make. The wide limbs complicate the construction considerably, core is made of many pieces, also horn, not to forget about the bone overlays.

I think the extreme width of limbs is an exaggeration, unless the bowyer did not have good horn and wood to use.

Posted by: Stephen Selby 07/18/2002, 21:34:43

The only significant difference (at least in surface appearance) between the Khotan bows and other bows that I have examined used up until the end of the Yuan Dynasty (to 1368) was (a) that the very wide limbs were later reduced to a size that, to us now, would seem more conventional, and (b) the siyahs became more massive.

Posted by: Bede Dwyer 07/18/2002, 20:43:29

I don't think you can say a design "petered out" if is was the dominant style for at least 1200 years. The development of thin, stiff siyahs obviously offered some advantages that prevented the reappearance of recurves for a long time.

I have spoken to some bowyers who have made this type of bow, though not so extreme as this one, and they were all very happy with the performance. I think Tim Baker's comments about stability and economy of materials are right on the mark.

Posted by: Dale Yessak 07/19/2002, 09:44:57

Bede, it just seems to me that if the wide-limb-thin-siyah concept had a significant advantage that it would be seen in many more forms and in many more locations. But my primary area of inquiry is the siyah angle of bows like this one and the various Magyar-Hun-Avar reconstructions ... what performance advantage, if any, would such low angle siyahs provide? Adam has stated that these might be more stable as opposed to siyahs with TOO MUCH angle due to the latter's propensity to twist. Perhaps any loss of performance due to a lesser string angle is offset by the thinner tip's lighter weight. I don't know, it makes one wonder.

The main reason I'm curious about this is because I made an osage-sinew recurve a few months ago which I had problems with getting the string to track properly. As a last resort, I added Asiatic style string bridges along with longer loops on the string, and the bow became not only more stable, but also faster. I realized that the speed increase was due to the string bridges effectively increasing the amount of "string contact" with the limbs and also because the bow, being braced about an inch lower, gained a little more power stroke. But the whole experience got me curious about the mechanics of recurve/siyah angle and the effect it has on the performance of a bow. The Khotan and Magyar bows are examples of non-contact siyahs, and so my curiosity was piqued to see if anyone knew if these bows actually represented less, more, or equal performance advantages over other designs.


Posted by: Bede Dwyer 07/20/2002, 23:26:34

You wrote "it just seems to me that if the wide-limb-thin-siyah concept had a significant advantage that it would be seen in many more forms and in many more locations" which is pretty much what happened. Due to the use of bone and antler laths for strengthening the siyahs, quite a few sets of these fittings have been found in archaeological contexts from China to Roman Britain.

The ratio of bending limb length to siyah length varies over time and space. The Roman laths are very similar in size (some quite long pieces have survived in Europe and China that can be compared). However, over time, siyahs became shorter and solider. The angle at the base of the siyah changed from almost none in the Yrzi bow to the pronounced curves in Avar bows from Hungary.

Adam comments somewhere in this correspondence that the width of the limb was probably related to availability of suitable timber and lengths of horn. Old Chinese texts refer to the Northern Barbarians needing several strips of horn to make their bows because of the lack of horns of suitable length. The "Traditional Bowyer's Bible" points out how to make good bows out of inferior timber by making the limbs flatter and broader than usual.

Since the late 1800s, people in the West have been interested in constructing composite bows and now we have Internet forums like this on the subject. In the past on the steppe and among its neighbours, bows were important military weapons, but there was little thought of research and development (Dionysios of Syracuse was an exception in the West). Detailed testing was not really developed. The ad hoc testing method was flight shooting which is very old, but that soon turned into a sport. Target competitions tested the bow and the archer together, but that also developed into a sport.

Tribal groups probably each thought their method of bow-making was the best and only changed when forced to by external forces. In relatively modern times this happened in Mongolia when it was absorbed into the Qing Empire of China. Mongol Banner troops were made to use the Manchu bow and eventually stopped making their own. Modern Mongol bows are a variation of the Manchu bow.

Large empires like Rome and China could shop around for ideas. The Romans seem to have adopted the most effective bow in use on their borders. The Chinese were in such large scale production that military production probably had a life of its own separate from regional styles.

When a bow maker today makes a bow, he or she may easily incorporate ideas from different style of bows. Sometimes I think you start making one type of bow and end up with another. I don't think any of us knows all the answers so we are still being surprised. Also I don't think there is a "perfect" bow. Instead there are bows for different purposes: flight, target, hunting, war, and practice.

Am I the only one or is this discussion thread getting too large to keep track of all the information?

Posted by: Dale Yessak 07/21/2002, 23:10:46

Point well taken, Bede. I'm not nearly so familiar as you with the development of composites, and perhaps I've made an incorrect assumption.

There are obviously many differing design factors in composites, just as there are in simpler self and backed bows. I'm just trying to get a feel for them at this stage.

The Poisson Effect

Posted by: Adam Karpowicz 07/12/2002, 08:00:26

Excellent review of the bow, thanks. It is tempting to make a replica!

As to the "spoon" shape of limbs, I understand you mean the working sections cross-sectional shape with the concavity on the back side. This concavity (as I think wrote to you before at one time) will occur naturally, once a flat-made limb is flexed.

The explanation is in the so called "Poisson effect", where the materials stressed under tension will contract laterally, and stressed under compression expand laterally (easy to demonstrate with a piece of rubber). This is exactly what happens in this bow: the back side is under tension, so it contracts, and the belly is under compression, so it expands. The net effect is the cross-sectional concavity on the back, as we see.

The wider the limbs, the more it will show. I made a bow at one time with limbs about 5cm wide and only 6-7m thick (the proportion of width to thickness not too far from this bow). The limbs were of course flat when made. Once strung and shot, the bow acquired this feature immediately, and due to the natural creep in materials, the limbs stayed like this in the unstrung bow. This distortion of limbs is permanent.

There is no doubt in my mind the Niya bow shows the Poisson effect very well.

Posted by: Stephen Selby 07/12/2002, 08:43:19

That answers one question, but raises another.

The 'Gansu bow' is obviously a model. Unless it was once a working bow that was subsequently stripped of its horn and sinew (which would be a bit bizarre), then the poisson effect is present in the Gansu bow although that bow has never actually been stressed in a composite form.

Might a bowyer anticipate the poisson effect when carving the wooden core?

Posted by: Adam Karpowicz 07/12/2002, 11:19:20

A strung (not shot)bow will assume this profile too. The concavity will be more pronounced if the materials are allowed to creep more, for example in higher humidity and temperature. The Gansu bow could acquire the profile when left strung in the grave.

This profile is not normally noticed at the usual width of bows. Since you mentioned Tim Baker, I recall having a similar discussion with him not long ago - he had not been aware of this before. He then found such concavity in a 2" wide self-bow with the help of a straight edge placed at the back across the limbs.

Posted by: Pat 07/12/2002, 16:37:37

I have noticed this effect in reverse while sinewing an Elm bow with pronounced reflex. I strung the bow backwards and applied sinew. The belly of the bow took on a concave profile temporarily. I'm sure the moisture from the hide glue aided this. It was very pronounced. As the sinew matrix dried it no doubt underwent shrinking across the limb which drew the belly back to a flat profile. After shooting the limb was basically flat again. The limb was about two inches wide. Pat

Posted by: Bede Dwyer 07/12/2002, 22:08:33

Thanks for the explanation of the limb concavity. I have two questions:

Would a model maker naturally just copy the concavity because that is what all the bows, which he had seen, possessed? The artificial addition of the concavity in the Gansu bow could be seen by examining the grain near the uplift, if it is visible. The concavity on the Gansu bow seems more of a hollowing to me rather than the lateral flexure apparent in the Khotan bow. Of course, I am relying on photographs. Stephen is in the best position to check this.

Is the convexity of the back in many later bows an over-correction for the Khotan bow's concavity? I wonder if there is a connection, though I admit is very farfetched at this stage. When you consider how close the Khotan bow is to the Kum Darya bow and the Hungarian bows on one hand and the Yrzi bow on the other, we may be looking at a missing link in the evolution of bow design.

Posted by: Adam Karpowicz 07/13/2002, 06:19:10

It is indeed a possibility that they copied the profile in wood, although I would very much doubt it. There can be grain swirls in wood giving an appearance of such work, proving this would be very difficult.

I do not see any advantages from the mechanical point of view in the upturned edges of the bending sections. The edges are under more stress this way. The over-correction is quite probable, this would be done to compensate, as Pat said, for shrinkage of sinew as well.


Why the broad limbs?

Posted by: Tim Baker 07/13/2002, 18:29:40

Adam, thanks for the message.

Stephen, Adam: Yes, level-3 deja-vu. Part of the thinking behind the TBB design was that more reflex would store more energy, but that conventional-width highly reflexed bows exceed the elastic limit of materials, take much set, and end up with less efficient just-unbraced profiles, the solution being more horn-sinew to properly hold that extra energy, and that this horn-sinew would have to be added in width not length or thickness. The idea was that if thin/wide enough the limb could hold a more than tips-touching just-unbraced reflex. But I wonder if the Khotan bow was highly reflexed. It appears not to have had string-liftoff siyah action, a low energy storage design. So possibly the wide limb solved the problem of normal energy storage in low-elasticity materials—most likely belly material. On the other hand, a no lift-off design would be more stable, quite a valuable feature in the field, and if reflexed as per above, stored energy could be very high, and stack still low. Of course there are other possible explanation for the wide limbs. As for the back’s concavity: if the limb was originally rectangular, which such wide limb must largely be, then some portion of the spooning must be P-effect. Adam made me aware of this a few years ago, and the most simple of tests demonstrate it exists on all near rectangular limbs. But I imagine that most of the spooning here is due to sinew contraction. I think the P effect can be overcome by appropriately convexing the limb during construction, and/or possibly by inhibiting the effect with bands of sinew running across the belly. Both ideas I believe Adam introduced on another bow site. But even without such correction ultra-wide limbs work well and safely--Pyramid shaped limbs four feet wide [plywood] for example.

I sure would like to see someone make a more than tips-touching-when-just-unbraced composite as per TBBlll, possible needing four or so inch wide limbs. I think it could be the fastest natural-materials bow ever made.

Please set me straight on any incorrect assumption or conclusion I might have made here.

Posted by: Stephen Selby 07/19/2002, 22:42:02

Following up on the comments made about the Khotan bow design reducing string-follow: at Khotan, the maximum temperature range over a whole year would be -20 to +40 Degrees Celsius.

Posted by: Pat 07/13/2002, 22:34:38

If the bow tended to hollow towards the back due to the P effect would the linear joints in the horn strips not be prone to separating? Perhaps this would be more of a problem if the limbs were severely reflexed. It is interesting that bows with multiple horn strips do seem to exhibit a net reflex(handle setback and Siyah angle) but the working section of the limb may have(deliberate?) "string follow". This bow and the Persian bows in a past letter and Volume 2 of TTBB being notable examples. Any thoughts if this may be the reason for the lack of reflex in the working limb? I also wonder if the low reflex in the siyah (likely no lift-off during draw depending on brace height) may be because a very wide flat limb is quite prone to twisting. A highly reflexed siyah might make the limb very unstable. I know the reflexed wood/sinew bows I have made with a thin wide limb and siyahs are very "wobbly" (although I'm sure horn added would make for a more stable limb). Still, wide and thin in any material usually means it can be twisted and warped very easily. The bow shown looks like it would be quite stable due to subtle modifications.

Posted by: Adam Karpowicz 07/14/2002, 06:27:35

It is true the horn strips are indeed prone to separating in bows with multiple strips on the belly. Persian bows of this design were covered with sinew all around to alleviate this problem. On the other hand, if the bond between the wood and horn is good, the separations between the horn strips do not affect performance.

I made only a couple of bows with the strips so far and I am now sure the strips, as opposed to solid horn, make the limbs more flexible (at the same thickness), so less efficient. I also think the Khotan bow is overbuild, narrower limbs would be much better for a bow this length, given the resiliency of sinew/horn combination.

I believe the deflex in the limbs is not deliberate in the Khotan bow. The limbs had to be build in full reflex, possibly the bending sections were straight in this case. The deflex came with use.

From my observations, the "wobbly" problems with wide and thin limbs happens only if the bending sections are very long. In this bow, they are relatively short and I would not expect any such problems here. The short bending sections plus the non-contact siyahs make the bow very stable.

Limb-to-Siyah Transition Questions

Posted by: Bede Dwyer 07/17/2002, 09:48:09

I have attached a rough sketch made from the x-ray photograph of the siyah. It covers the area from the small fracture in the core of the siyah to the point after the siyah has joined the main limb.

I have guessed that the belly side of the wooden core of the siyah was reinforced with another wooden strip. I assume that it is broken up and has shrunken due to drying. In the Qum Darya bow, the horn that covers the belly-side of part of the siyah appears to be a separate, overlapped piece (but that is based only on a drawing).

I have some trouble interpreting the x-ray, but it was of immeasurable help in understanding the photographs of the siyah. From reading the Traditional Bowyer's Bible, I know that bows were made in America in the 20th century with static recurves reinforced on the belly.

Stephen, you are in the best position to comment on this as you have seen and handled the bow. What do you think?

Posted by: Adam Karpowicz 07/18/2002, 06:22:02

Bede, if I understand you correctly, you believe the wooden reinforcement in the siyah is in one piece (orange on the drawing). I think the broken belly fragment of the piece is actually another, short piece of wood, glued onto the belly of the core at the knee. It would be easier to manufacture, by gluing the first (long) piece on the siyah, then filing it smooth and gluing the next to build up the thickness there. Otherwise the bowyer would need to pre-bend a thicker piece of wood, not an always an easy task, depending on the length etc. It is quite possible both methods were used as well, maybe even on the same bow.

Posted by: Bede Dwyer 07/18/2002, 20:36:34

I agree with you, but I wasn't sure. I coloured the siyah reinforcements orange to distinguish them from the core. I did not think they were pieces split from the core by drying out of the wood, but I was not certain whether they were individual parts because of the degree of break up of the pieces.

Posted by: Stephen Selby 07/17/2002, 10:38:06

The bow is at the museum undergoing further tests. This picture is all I have to hand. The picture is a bit misleading: the darker brown wood has shrunk somewhat. The extension of the horn in the limb, visible in the X-ray, is not apparent in this view at all.