Position codes: the Japanese Navy's kana grid

Japanese search planes reported an enemy's position as four characters: 「メキ四テ」, 「ヤソ3ス」. The two kana name a 20′ square of ocean, but the letters are scrambled and change with the base the search flew from. The code tables don't survive, so each key is being rebuilt letter by letter from reports whose position is known some other way.

How to read one

ホク三テへ

A 901st Air Group flying boat, 12 October 1944, 0210: "enemy task force, apparently, detected, ホク三テへ". ホ is the column: 123°40′–124°00′E. ク is the row: 21°20′–21°40′N. 三テ places the point inside that square; how is not yet worked out. へ names the base point, here 鵝鑾鼻 (Garanbi), Formosa's south cape. Read with it, the code means a square about 20 nautical miles a side, some 170 nm east of the cape.

The squares sit on whole degrees, three columns and three rows to a degree, like a chessboard laid on the chart. The letters along each edge aren't in any alphabetical order, so neighbouring columns can be ホ, ネ, ヘ. A key is the set of letters for one base point. The 901st's flights from Formosa signed their codes with へ and its flights from Manila with ヤ, and the same square has different letters in the two. The searches from Kanoya in April 1945, the ones that hunted TF 58 off Okinawa, add no suffix and use a third key.

A column or row whose letter is placed: solid good, dashed probable, dotted tentative Anchor: a code whose position is known Anchor with a letter not yet placed Where the flight itself put the contact Base point

Scroll or pinch to zoom · drag to pan · zoom in to see each square's two kana

Each strip is one placed letter: a column runs north–south, a row east–west, and a code can be read wherever a column and a row of the same colour cross. Hover an anchor to light up the square its code names; click it to open its source. The white gaps are letters nobody has found yet.

Try one

Code

A code ending in へ or ヤ is read on that key. Without a suffix, pick the key. A letter not yet placed leaves only a strip lit on the map: that much of the position is known.

The keys so far

Each box is one placed letter, with the edge of its strip. A dot stands for a 20′ strip whose letter isn't known yet.

Where the letters come from

Manila 1944 began with a figure in the official war history (戦史叢書 vol. 56, fig. 34), which draws three columns and three rows off Samar on 25 October 1944, the morning of the Battle off Samar. The 901st's charts and signals add six more.
Formosa 1944 comes from the 901st Air Group's night-search report for 10–28 October 1944 (JACAR C13120398700). It ends with two charts on a printed 1° graticule where the staff plotted every contact as a boxed ☒ with its code. Each box was measured against the graticule, to about ±5′, and each code checked against the signal on the sortie's own page. The war history's "Ishigaki 203° 152 nm" for ヘリ4セ, 14 October, lands in the same column and row.
Okinawa 1945 rests on a pencil track chart from 21 April 1945 with three contacts sketched beside the islands, and on the 6 April reports that must include Haynsworth's task group (see her track). Two night scouts on 7 April coded their own positions (JACAR C13120385000), dead-reckoned from Kanoya along their search lines. They agree with サ and キ and suggest ケ.

What isn't known

The last two characters. A digit and a kana place the point within the square. The plotted boxes are only good to a quarter of a square, and they show no more than hints: every 一 sits in the north-west of its square, イ near the top, セ near the bottom.
Whether the chart is right. The 901st's chart is its staff's reading of the signals. For five contacts the flight's own track gives a second position (the dashed lines on the map). Three agree. Two don't: コヌ四エへ (11 Oct) is plotted about 230 nm from where the flight reported itself an hour later, and レス四セへ (17 Oct) 2° north of the flight's dead reckoning.
Whether rows repeat. Rows ツ, キ, テ start at 12°N in the Manila key of 1944 and at 27°N in the Okinawa key of 1945, 15° (45 rows) apart. The へ and ヤ keys share no row letters but do share a column letter, ツ, at different longitudes. So the base point may change only the columns, with one row alphabet repeating every 15°. If so, row エ in April 1945 would be 28°00′N and イ 29°00′N. Nothing has tested that yet.

Method, every anchor and the open questions: guides/JAPANESE-GRID.md. The letters are reference/jgrid/keys.csv, the anchors reference/jgrid/anchors.csv, and tools/jgrid.py decodes and checks them.