Devil's Sea

Start with the sentence that ought to have ended the legend before it began. The Yomiuri article of 14 January 1955 that first grouped these losses together, and from which the Reuters cable and the whole English tradition descend, said in its own text that the Kaiyō Maru No. 5 had run into the Myōjin-shō explosion. Everything after that removed the volcano. The volcano is not a hypothesis. Robert S. Dietz and M. J. Sheehy published the reconstruction in the Journal of the Oceanographical Society of Japan in 1953 and again in the Geological Society of America Bulletin in 1954. Myōjin-shō is remote and was watched by very few eyes, so they worked from instruments: a tsunami recorder on Hachijōjima, 130 km to the north, an atmospheric-electricity recorder in Tokyo, and — decisively — the US Navy's Sofar hydrophone stations at Point Sur and Point Arena in California, about 8,600 km away, which logged well over a hundred explosions, one of them lasting more than an hour. Their abstract states the finding without hedging: from the tsunami data and the Sofar records it was independently concluded that the Kaiyō Maru No. 5 was destroyed by a large eruption at about 12:20 Japan time on 24 September, and that all thirty-one aboard were lost. The Maritime Safety Agency's own board of inquiry, chaired by the agency's commandant Yanagisawa Yonekichi, reached the same conclusion on 22 December 1952: the ship was struck by a submarine volcanic explosion from below and to starboard, her upper works were largely blown away on that side, and she capsized and sank at once. The Smithsonian's Global Volcanism Program records the loss in the same flat terms in its entry for Myojinsho (volcano number 284070): an explosive submarine eruption in 1952 destroyed a Japanese research vessel, killing all thirty-one on board. Nobody with access to the instruments has ever treated this as unexplained. The next claim to check is the one that makes the place old: that Japanese fishermen have feared these waters for generations. The Japanese record points the other way. The Asahi Shimbun of 12 January 1955 wrote that some people were now even beginning to call this sea area ma no umi. The phrase itself is a newspaper cliché rather than a place name: the same Japanese press had already applied "sea of the devil" or "devil's waters" to the Bay of Bengal in 1914, the Kurils in 1924, the Korea Strait in 1926, Lake Baikal in 1926, the coast of eastern Japan in 1930, Hainan in 1939, the waters off Britain in 1939, the Atlantic in 1940 and the Taiwan Strait in October 1954. It means treacherous water. Larry Kusche, the Arizona State University reference librarian who audited the Bermuda Triangle case by case, traced the English name to that Reuters cable, wrote to the Japanese bodies concerned, and reported that nobody he asked had heard of a "Devil's Sea". Two checks of my own, with their controls, point the same way. Sweeping 2,267 KB of the Japan Coast Guard's own marine volcano database — thirty-three volcano files covering the whole Japanese arc — the terms ma no umi, "Devil" and "Dragon" return zero hits between them; the control terms in the same sweep return 201 hits for Myōjin-shō and 830 for "eruption", and the file for the Bayonnaise Rocks carries the line recording that on 24 September the Kaiyō Maru No. 5 of the Hydrographic Department was lost while surveying, with thirty-one dead in the line of duty. In the Library of Congress's Chronicling America corpus of American newspapers, the phrase "dragon's triangle" returns zero hits for the whole period 1900-1955; the controls in the same corpus and window return fourteen hits for "Myojin", the earliest on 18 September 1952 — the day after the eruption was reported — and three for "Bayonnaise Rocks" going back to February 1946. The American press knew this volcano. It did not know a Dragon's Triangle. The nine ships themselves dissolve on contact. Apart from the Kaiyō Maru No. 5 at 211 tons and the fisheries patrol vessel Kuroshio Maru No. 3 at 145, they were fishing boats of 62 to 190 tons, in a country still poor enough that many carried no usable radio, and they were lost across a stretch of ocean running roughly 1,200 km from Miyakejima in the Izu chain down to the Volcano Islands. Several have documented causes in the contemporary press: the Seishō Maru No. 1 radioed that she was flooding in a gale and sent an SOS; the Kōchi Maru No. 16 was last heard from while sheltering from a typhoon; the Chōfuku Maru No. 5 reported taking water in wind and rain. As for the "special hazard area", what the Maritime Safety Agency had actually issued, in 1953, was a notice to keep 15 km clear of Myōjin-shō. Berlitz went further than exaggeration. In The Dragon's Triangle he misnames the ship, inflates her to 500 tons, dates her loss to 1955 rather than 1952, and presents her as a survey vessel sent to investigate why the area was dangerous — she was sent to chart an erupting volcano — while asserting that nothing lost in the triangle has ever left a trace. A life ring, a marked buoy and pumice-studded timber were recovered within seventy-two hours. One genuine uncertainty survives, and it is volcanological. In 1955, in issue 50 of the Hydrographic Department's Suiro Yōhō, Tsukamoto Yūshirō argued that the ship was destroyed not by Myōjin-shō itself but by a separate vent lying between the reef and the Bayonnaise Rocks, which he provisionally named Takane-shō, citing the 1915 near-loss of the fishing boat Takane Maru No. 3 at that intermediate position. A 1989 survey with the Hydrographic Department's self-propelled buoy Manbō, reported in 1990, found a central cone at 300-350 metres depth in about the place he predicted. Which vent killed the Kaiyō Maru No. 5 is still open. That is the only mystery these waters have, and it is a question about plumbing under the sea floor, not about anything waiting in it.