An essay by Fukunaga Suiken titled "Nihonto no oremi" (Breakability of Japanese Swords) in the book Nihonto wazamono nyumon (Introduction to Japanese Swords of Noted Cutting Ability; 1972, pp. 23--82) is relevant to the above discussion. Fukunaga reviews the history of sword durability tests and their results.
In 1943 the Japanese Army devised a 3-fold test for new swords: 1 - impact test, 2 - straw bundle (makiwara) cutting, 3 - iron rod (3 mm by 10 mm thick) cutting.
The impact test consisted of placing a sword on a stand where it was supported only at each end and repeatedly dropping a 3-kanme (about 25 U.S. pounds) iron weight onto the center of the blade from an ever-greater series of fixed heights. First the weight would be dropped from a height of 15 cm, then 25 cm, then 35 cm, and so on until the sword broke. Blades from each smith were subjected to three kinds of impact tests. The weight was dropped on the side (hira), on the cutting edge (ha), and on the back (mune) of the blades. Fukunaga (pp. 40--44) analyzes the results of the first round of tests conducted in November 1943. He charts the performance of blades from each smith on a graph that shows the height from which the weight was dropped, how much the swords bent, and when the swords broke. Here are the results for side vs. edge vs. back.
=======
Side (hria): the worst results were from a blade from an unnamed smith which bent 40 cm (or 40 mm?*) on the fifth impact dropped from a height of 55 cm. The best results were from a blade made by Takefuji Hisahiro (Fukuoka Pref.) which finally bent 35 cm on the ninth impact dropped from a height of 95 cm.
* The text says "mm" but the charts say "cm" (same for all three results)
=======
Cutting edge (ha): the worst results were from a blade from an unnamed smith which bent 11 cm on the eighth impact dropped from a height of 85 cm. At that same point in the test a blade by Takefuji bent only 3 cm.
=======
Back (mune): the worst results were from several blades that broke when the weight was dropped for the first time from a height of 15 cm. The best results were from a blade made by Sakurai Shinkoku (Nagano Pref.) which bent only 8 cm on the sixth impact dropped from a height of 65 cm.
=======
The bottom line:
Objective tests performed by the Japanese Army showed that stationary sword blades supported at each end broke most easily when struck on the back of the blades. Sword blades withstood the greatest impact force when struck on the cutting edge.
Fukunaga notes that in spite of the fact that everyone always identifies the main defect of Japanese swords as being their inability to withstand blows to the back (mune uchi), Sakurai proved that it is possible to create a sword that can overcome this weakness.
After the tests were completed, the blades were cut in half and subjected to metallurgical analysis. The smiths had used a variety of construction techniques (i.e., makuri gitae, hon sanmai awase gitae, kofuse gitae, etc.). Interestingly, the construction technique did not seem to be a major factor in blade performance, although blades with hard skin steel on all four sides (shihozume gitae) tended to perform worst while all the best performing swords were simple makuri gitae or kofuse gitae. Fukunaga states that the critical factor was the depth of the core steel, particularly how close the core came to the hardened edge (ha). Blades in which the softer core steel extended down almost to the hardened edge steel performed best, while the worst performance was by blades with edge steel so thick that the core steel began further away from the edge. Since the cross sections of the blades are so small, minuscule variations mattered a great deal.
Fukunaga also reports on many other kinds of tests. I will just mention two more.
(1) Tokugawa Nariaki (1800--1860), the ruler of the Mito domain, devised the following series of tests for the swords used by his men. 1 - bo tameshi: one person holds a sword in a seigan stance while a second person strikes it forcibly on each side with a stick about as thick as a human wrist. 2 - makiwara tameshi: cutting straw bundles. 3 - tsuno tameshi: cutting deer antlers. (Fukunaga reports that it was very difficult to produce an edge angle that could cut both straw bundles and deer antler equally well.) 4 - mizu tameshi: striking the flat side of the blade against the surface of water. Supposedly this final test produced the most failures. Unfortunately, detailed results were not recorded.
(2) In 1937 the magazine Token kogei (Sword Industry) wanted to determine whether Japanese swords could withstand the rigors of warfare during the cold Mongolian winters. They took 12 swords and chilled them to minus 60 degrees Celsius. Then they asked Kurimoto Shinzo (who is identified as a "Yagyuryu" shihan) to cut with them. All 12 either broke or bent immediately.
By the way, the term wazamono was popularized in the book Kaiho kenjaku (1815) by Yamada Asaemon Yoshichika who personally rated the cutting ability of swords produced by about 180 smiths. The essays collected in Nihonto wazamono nyumon are not concerned with those traditional rating, but attempt to summarize the state of modern (ca. 1960s) knowledge about the cutting ability of Japanese swords. Other contributors include Yoshikawa Kentaro (on the relationship between polish and cutting ability), Tokuno Kazuo (on lineages of sword testers), Wake Yutaka (on metallurgy and durability), Nakamura Taisaburo (on swords that can cut and cannot cut), and Hyakutake Masayu (on standard methods for tameshigiri) as well as a round table discussion with several other swordsmen and smiths.
William Bodiford
Professor
Dept. of Asian Languages & Cultures
UCLA