On 31 January 1918, two British K class submarines were sunk and four others damaged off May Island in the Firth of Forth. These losses resulted not from enemy action but from collisions during an exercise with no enemy forces present.
The K class resulted from a faulty concept, which was for a submarine able to operate with the battle fleet with speeds of 24 knots on the surface and 9.5 knots submerged. Submarines were then powered by diesel engines on the surface and battery powered electric motors submerged. This combination allowed the highly successful E Class to make 15 knots on the surface and 9 knots submerged. Erroneous intelligence reports that some German U-boats were capable of 22 knots led to the J class, whose diesel engines were supposed to give a surface speed of 19.5 knots. Problems in heavy seas needed modifications that cut this to 17 knots.[1]
In order to achieve the require speed, the K class were powered on the surface by oil fired steam turbines. They were armed with ten 18 inch torpedo tubes, 18 torpedoes, two 4 inch guns and a 3 inch anti-aircraft gun. The use of steam meant that funnels and hatches had to be closed before they could dive. This could be done in 30 seconds, but there was a risk that one would be jammed open by a small obstruction. Eighteen were built, with another nine cancelled. K13 sank on trails in the Gareloch on 29 January 1917. She was raised and repaired but renumbered K22.[2]
The K boats were really submersible destroyers. The intention was that, like destroyers, they would operate in flotillas rather individually. Each flotilla would be led by a light cruiser until they were close enough to the enemy to fire their torpedoes, after which they would retire.[3]
Rear Admiral O. W. Phillips, at one time Chief Engineer of HMS K4, later wrote that 13 openings had to be closed before diving but that the K boats ‘were a wonderful effort and no more dangerous than any other submarine – if properly handled.’[4]
The editors of Conway’s All the World’s Fighting Ships, 1906-21 argue that the K boats were not a poor design, but were actually a ‘remarkable technical achievement.’ Their problem was not their design but the ‘faulty conception which had led them to being built.’ They were expected to operate surfaced at high speeds in close company with surface ships, often at night without lights, but their bridge facilities were not adequate to do so safely.[5]
By January 1918 two flotillas of K boats, the 12th (K3, K4, K6 and K7 led by the light cruiser HMS Fearless) and 13th (K11, K12, K14, K17 and K22 led by the flotilla leader [a large destroyer] HMS Ithuriel) were based at Rosyth in the Firth of Forth. On the evening of 31 January they were part of a force of 40 ships, including three battleships and four battlecruisers that sailed to join the rest of the Grand Fleet at Scapa Flow, Orkney, in an exercise codenamed EC1.The K boats were showing only a small blue stern light. All ships increased speed as they approached May Island because a U boat was reported to be in the area.
Two of the K boats changed course to avoid two minesweeping trawlers that suddenly appeared and moved across their course. Ithuriel’s helm jammed as she changed course. K14 and K22 then collided after the former’s helm jammed and the latter was later hit by the battlecruiser HMS Inflexible.
Ithuriel and the other three boats of the 13th Flotilla turned back to give assistance. They were narrowly missed by the battlecruiser HMAS Australia but the two submarine flotillas became mixed. Fearless rammed K17, which sank, K3 just avoided K4, K6 nearly hit K12 head on but then struck K4. The two submarines became locked together. K6 escaped by running full astern but K4 was almost cut in half and sank after being struck by K7.
Over 100 men died: they are listed on naval-history.net. Some of them got into the water but were run down by destroyers that did not know what had happened. There were no survivors from K4 and only eight from K17. Two submarines were lost, with four and a light cruiser damaged despite the complete absence of enemy forces.
The subsequent court of inquiry blamed the officers of the K boats involved rather than the concept behind their design or their flawed employment.[6]
Three more K boats were lost, all accidentally. K1 collided with K4 off Denmark on 18 November 1917 and was scuttled to avoid capture. K5 failed to surface after diving during exercises in the Bay of Biscay on 20 January 1921 and was lost with all hands. K15 sank at her moorings in Portsmouth on 25 June 1921, but all the crew survived thanks to prompt action by her captain. The closest that they came to sinking an enemy ship was when K7 hit a U-boat with a torpedo that failed to explode.
As well as the books footnoted and websites linked in the text, the following websites have been used as sources:
Steam Submarines: The Navy’s Dive to Disaster by John Watts on the website of the Submariners Association Barrow-in-Furness Branch. This source says that K7 passed over K4 as she sank without hitting her, but I have taken the version of events from the MoD linked below.
Battle of May Remembered, published by the Ministry of Defence on the 84th Anniversary of the disaster, when a memorial to the dead was unveiled at Anstruther, Fife.
Battle of May Island on Wikipedia.
[1] Gray, Conway’s 1906-1921, pp. 87-90.
[2] Gray, Conway’s 1906-1921, p. 91.
[3] Liddle, The Sailors War, 1914-18, p. 187.
[4] Quoted in Liddle, The Sailors War, 1914-18, p. 188.
[5] Gray, Conway’s 1906-1921, p. 91.
[6] Liddle, The Sailors War, 1914-18, p. 190.
War and Security 
I think there’s little doubt that the original idea behind the K’s as ‘fleet submarines’ was good, and I suspect that like many new technologies of the day, analysis of the theory, ahead of actually committing to construction, had to be run as a thought experiment because there was almost no practical experience to draw on (the same was actually true of warship construction generally at the time – it took the First World War to provide practical experience of how all the new industrial-age systems actually worked together). I suspect the main technical problem with the K’s, however well managed, was still the issue of having a steam plant aboard – shutting it down and sealing the funnels made submerging a lengthier process for the K’s, and complicated the processes. As you say, it all worked fine when properly handled, but the issue was properly handling it all – it only took one slip-up by a matelot, or a moment’s inattention, for things to go wrong. There was also the practical issue of the residual heat in the steam plant inside the boat underwater, which could only radiate into the boat itself. Probably they were no more dangerous than some of the peroxide-engined boats tried at the end of the Second World War and in the years afterwards, but the diesel-electric boats always were a more practical option at the time.