Materials of the Kyoto bow: Hamada 2017 study

The author, who studies the tacit knowledge of traditional craftsmen at the Kyoto Institute of Technology, describes work done with Shibata Kanjūrō, a maker of Japanese bows in Kyoto and the twenty-first of his family. The bow, normally about 220 cm long, has its grip set low and is built in three layers of natural materials: bamboo on the back and the belly, with a core of different woods in between. The paper sets what Shibata says about his materials against laboratory tests.

In bending tests, haze wood (from the wax tree) broke in a brittle way while maple withstood larger deformation, as the bowyer had said; the bamboo strips of the core (higo), which he described as hard and brittle, did not behave as a brittle material. The bamboo chosen for the belly of the bow had a higher modulus of elasticity than that chosen for the back and, in cross-section, contained 18% fibre against 9%, and the modulus rose in proportion to fibre content. In vibration tests the belly bamboo oscillated with a shorter period, meaning it was stiffer, while damping was the same.

Finally, bows glued with nibe, the traditional glue made by boiling deer hide for three days, were compared with bows using synthetic glue. Peak arrow speed was almost the same, but with nibe the arrow accelerated more in the last part of the push and reached the speed of the other arrow 0.027 s after release.

The group usually works with analysis of the work stages, infrared cameras, eye tracking, force sensors, heart rate and EEG, and records what the craftsman says word for word; in this paper, though, the measurements reported are laboratory ones only. The three core woods were tested in three-point bending on 5 mm specimens, with 80 mm between supports at 50 mm per minute; the back and belly bamboo with 400 mm between supports at 1 mm per minute. For the vibration tests, bamboo strips 23 by 3.3 mm and 140 cm long were clamped at one end, deflected by 50 cm at the other and released, and filmed by a camera at 250 frames per second. Fibre content was read from photographs of the cross-section.

Arrow speed was obtained from high-speed footage with the camera facing the archer, and the mean curve for each type of bow was divided into 0.01 s intervals to compare acceleration. The paper does not say how many archers, bows or shots took part in the test, and the speeds appear only in graphs.

This is a five-page explanatory article for a general readership, summarising earlier work presented at conferences, with no statistics, sample sizes or uncertainties: the results have to be read off graphs and cannot be checked. The materials tests concern pieces chosen by a single bowyer, so they show that his choices are consistent, not a general rule about bamboo or woods. For the glues we do not know whether the bows were identical in every other respect (draw weight, shape, age), and the difference in acceleration does not translate into a difference in peak speed, which is what governs the trajectory. The idea that nibe gives the arrow more push just before it leaves the string is the author's interpretation.

For Olympic recurve archers nothing changes. For kyudo practitioners, or anyone curious about traditional bows, the paper puts some numbers to practices handed down by word of mouth: belly bamboo is chosen to be more fibrous and stiffer than back bamboo, and shaved down to about 4.5 mm so that only the part near the skin, where the fibres are densest, is kept. It also shows that the master's judgement of his materials partly stands up to laboratory testing and partly does not.

On glue the finding calls for caution: with nibe the arrow reaches the same peak speed, with a different pattern of push. Whether this changes the feel of the release or the accuracy was not measured.