Bowstring materials in kyudo: Hosoya 2001 study

Ten archers from a university kyudo club, eight men and two women, shot the same glass fibre bow with four strings of different materials: the two aramids Kevlar and Technora, Vectran and hemp. For every shot the authors recorded the bow's vibration, the arrow's speed and a free comment from the archer on the feel at release.

The efficiency with which the energy of the drawn bow passes into the arrow is practically identical for the three synthetic fibres: 69.0 ± 3.3% with Kevlar, 69.1 ± 3.5% with Technora, 69.4 ± 3.4% with Vectran. Hemp sits lower, at 67.7 ± 3.9%, and the difference between hemp and each of the three synthetics is statistically significant, while the differences among the three synthetics are not.

Damping runs the other way. The logarithmic decrement of bending vibration rises from 0.61 ± 0.01 with Kevlar to 0.64 ± 0.13 with Technora, 0.65 ± 0.14 with Vectran and 0.70 ± 0.11 with hemp; for twisting vibration the order is Technora 0.32 ± 0.06, Kevlar 0.35 ± 0.08, Vectran 0.42 ± 0.07, hemp 0.51 ± 0.06. A higher decrement means an oscillation that dies away sooner, so hemp damps more. Vibration frequencies, analysed with a fast Fourier transform, do not change appreciably from one material to another.

In the free comments, sorted into four categories, the answers are distributed across the strings in a way that is not random (chi-square on 6 degrees of freedom of 12.69 against a critical value of 12.59, p < 0.05). Residual analysis points to two cells: string sound with Vectran, mentioned less often than expected, and vibration or impact at release with hemp, mentioned more often than expected. Hemp is also the string that stands furthest from the others in the physical measurements, and the authors read that coincidence as a sign that those quantities reach the archer's own sensitivity.

Everyone shot the same glass fibre bow, of similar strength to those the subjects used in training, treated as a linear spring within the 0.8 metre draw: its spring constant is 198.6 newtons per metre. The arrow has a duralumin shaft 0.96 metres long with a mass of 0.026 kilograms. The initial distance between string and grip was fixed at 0.15 metres for every subject and every trial, and the strings were shot about twenty times beforehand to let them settle. Shooting was at straw targets in a dedicated room.

Bending and twisting vibration were read by strain gauges bonded to the front and back of the bow near the grip. The instant of release was detected with a makeshift circuit using conductive tape on the string and on the hand. Arrow speed just after release comes from a ceramic magnet inside the shaft, which passes through two coils and induces a signal. Analogue-to-digital conversion sampled every 0.1 milliseconds.

The logarithmic decrement was not computed from the ratio of two successive maxima, as is usual, but from a curve fitted through all the maxima; the frequencies come from a Fourier transform over 200 milliseconds after release. Efficiency is the ratio of the arrow's kinetic energy, half mass times velocity squared, to the potential energy of the bow at full draw, half spring constant times displacement squared, with the displacement derived from the bending strain gauge calibrated in advance. The authors explain that they chose efficiency rather than raw speed because the draw length varied between subjects and between trials by the same subject.

Each mean is based on fifty data points per string, that is ten archers times five trials each. The free comments were collected by questionnaire after every trial and then grouped by the authors into four headings: vibration and impact, string sound, how promptly the string recovers, and other. Kevlar was left out of the chi-square because eight of the ten subjects used it habitually and mostly answered "as usual", so the test compares only Technora, Vectran and hemp.

The sample is one club: ten university kyudo archers, with one bow, one arrow type and fifty shots per string. There is nothing on score and nothing on impact position, so the study links neither efficiency nor damping to the outcome of the shot.

The bow is treated as a linear spring, a choice the paper states but which simplifies the force-draw curve of a Japanese bow; the potential energy computed as half spring constant times displacement squared inherits that simplification, and efficiency inherits it in turn. Absolute arrow speeds are not reported, so the reader sees only the ratios. The mass of the four strings was measured but stays in a table that cannot be read in the available copy.

The questionnaire is the weakest part. The comments were free and were sorted afterwards, by the authors themselves, into four categories of their own definition; no blinding is declared, and an archer shooting a hemp string rather than Kevlar can both feel it and see it. Leaving Kevlar out of the chi-square is reasonable, since eight of the ten already used it, but it removes the reference material from the comparison. And the test value is marginal, 12.69 against a critical value of 12.59.

There is also a point the paper does not explain. Hemp damps more than the synthetics, yet it is also the string the archers most often associated with vibration and impact at release. The two findings do not conflict, because the logarithmic decrement measures how quickly an oscillation dies away rather than how large it is to begin with, but the paper sets the two results side by side without saying so.

Finally the discipline. The shooting is with the Japanese bow, with kyudo technique and kyudo strings: none of these measurements was made on an Olympic recurve, and the step from one to the other is not in the data.

For a kyudo archer, and with some caution for anyone thinking about strings in general, the measurable finding is that Kevlar, Technora and Vectran return the same efficiency to within half a percentage point, while hemp returns one and a half points less. Half a percentage point of energy is not a difference that shows on the target; one and a half points is worth knowing if hemp is chosen for other reasons.

The second piece of information concerns feel. The bow's vibration frequencies do not change with the string material, so the difference an archer notices between two strings lies not in the pitch of the oscillation but in how quickly it dies away and in the sharpness of the impact. Hemp sits at the extreme on both counts: it damps faster and it is also the string archers named most often for vibration and impact at release.

What the study does not authorise is the jump to scoring. Nobody counted points, nobody looked at where the arrows landed, and the comments on feel come from archers who knew which string they were holding. A recurve archer will not find a transferable figure here either, because the bow, the thumb release technique and the materials are those of kyudo.

The contribution worth remembering is methodological: the authors tried to set the physical measurements of a string alongside the judgement of the person using it, and found that the string furthest from the others in the measurements was also the one that stood out in the comments. That is a hint about how a shooting material can be assessed, not a rule for choosing one.