Flannan Isles: the three keepers and the empty lighthouse

Eilean Mòr rises from the Atlantic west of Lewis, the largest island in the Flannan group. The Northern Lighthouse Board describes seven rocky, uninhabited islands, traditionally called the Seven Hunters. The white lighthouse occupies the island’s highest ground, above landing places reached by steep paths and steps.

The station was established in 1899 and associated with engineer David Alan Stevenson. Its shore accommodation stood at Breasclete on Lewis, where families could live and relief crews could shelter. In 1900 there was no radio link between the station and the mainland. An observer on Lewis watched for signals and failures of the light; urgent information travelled onward by telegram.

That geography is central to the case. The tower, living quarters, landing equipment, relief vessel and shore station formed one working system, stretched across difficult water. A failure to see a light and the discovery of an empty building were separated by distance and the opportunity to land. Source: the lighthouse operator’s station history.

The missing keepers were James Ducat, the principal keeper; Thomas Marshall, an assistant keeper; and Donald McArthur, serving as an occasional keeper in place of William Ross, who was on sick leave. Joseph Moore, another assistant keeper, was ashore when the men disappeared.

National Records of Scotland’s account places the apparent loss on the afternoon of 15 December 1900. The station had completed its morning work, but the light was not seen that night. Its disappearance was discovered when the routine relief reached the island on 26 December.

The surviving correspondence was written by colleagues who had to maintain the station after discovering the loss. Read the archival introduction and identification of the keepers.

The tender Hesperus arrived to make the relief. Its master, Captain Harvie, found the usual signs of activity missing. Sound signals and a rocket brought no response. Moore was put ashore to investigate and returned with news that the station was deserted.

Harvie’s telegram, dated 26 December, reported the three missing keepers and the stopped clocks. It offered an immediate possibility—men blown over the cliffs or drowned while securing equipment—rather than the result of a completed inquiry. Darkness and the need to maintain the light shaped the response.

Moore went back with volunteers: buoymaster Allan MacDonald and seamen Campbell and Lamont. Harvie reported that they had been left to keep the light burning while further arrangements were made. The telegram captures the first alarm and the operational urgency, before the fuller search and superintendent’s inspection. Read Harvie’s telegram in the Board’s document collection.

Moore’s letter of 28 December gives a more personal account of the discovery. He described closed entrance doors, a kitchen fire that had not burned for days and empty beds. Alarmed, he hurried back down to the boat and brought others up for a further search.

In the lightroom, the equipment was in order: the lamp had been cleaned, the oil supply prepared and the windows screened. Moore and the replacement party lit the light that evening and continued to operate it while searching the island.

His observations also extended to clothing. Ducat’s sea boots and waterproof and Marshall’s sea boots and oilskins were absent, while McArthur’s usual outdoor coat remained behind. Moore inferred that McArthur might have gone out in shirt sleeves. It is a concrete detail from a colleague’s inspection, but it does not reveal the missing man’s precise movements.

Moore distinguished the orderly east landing from damage and displaced equipment on the western side. His letter is one of the strongest routes into what the relief party actually found. Read Moore’s letter, dated 28 December 1900.

Superintendent Robert Muirhead reached the island on 29 December and examined the station and its surroundings. His report, submitted on 8 January 1901, described a marked contrast between the intact eastern arrangements and damage around the western landing.

Muirhead found displaced iron railings, scattered ropes and a missing storage box. A heavy stone had been moved onto the path. A lifebuoy had been torn from its fastenings: remnants of canvas remained on ropes that appeared not to have been untied.

The heights made the damage striking. His account places the crane platform around seventy feet above sea level and some affected equipment around 110 feet up. He could not safely descend all the way to the landing because of the sea. These measurements belong to his inspection; they are not a measurement of the particular wave that may have taken the men.

The damaged landing provides the inquiry’s main physical setting. It explains why attention focused on an excursion outside the tower rather than a disturbance within the living quarters. Read Muirhead’s description of the western damage.

Muirhead proposed that the men had gone towards the west landing to secure the box of ropes or inspect damage, and that an unexpectedly large wave had swept them away. He considered the wind, but favoured the sea: the wind’s direction, he thought, made a direct blowing-over-the-cliff explanation less persuasive.

The prepared lamp and morning observations helped establish the likely date.

The unresolved part is the final sequence. No surviving witness describes the men leaving the station, reaching the landing or responding to one another. The reconstruction leaves questions about why all three were exposed at the same time. Read the archive’s account of the investigation.

Wilfrid Wilson Gibson’s Flannan Isle turned the disappearance into an atmospheric literary narrative. Its relief party finds a meal laid out, a fallen chair and an unsettling silence. Strange birds and an imagined history of earlier losses deepen the feeling of a place under a shadow.

Those details are powerful precisely because the poem invites readers into the searchers’ fear. They should be encountered as Gibson’s storytelling. Moore’s actual letter and Muirhead’s inspection have a different texture: doors, oil supplies, clothing, ropes and railings.

Reading the poem beside the correspondence shows how the remembered scene developed. A reader can enjoy its haunting imagery while keeping the original investigators’ observations clearly visible. Read Gibson’s poem and compare it with the contemporary documents.

Historian Mike Dash’s The Vanishing Lighthousemen of Eilean Mór returns to lighthouse records, newspaper accounts and later retellings. It includes transcribed documents and a detailed examination of the supposed dramatic log passages describing frightened men and prayer.

Dash traces those passages through later authors and questions their provenance, form and chronology. That investigation is a useful part of the case’s afterlife: a frequently repeated quotation can acquire an authority that its documentary trail does not support.

There is also a clear archival limit. Both the Northern Lighthouse Board and National Records of Scotland say they do not hold the station’s logbook. The surviving investigators’ references to its entries and the preparation slate should therefore be distinguished from a claim that a complete original log is available online. Read Dash’s archival study and the current NRS research guide.

For further research, NRS identifies the wider lighthouse records and the Flannan station series, NLC53, covering 1899–1971. Its guide explains that records are stored off site and need to be ordered before a visit. The disappearance correspondence is associated with NLC3/1/1 in the archive’s published account.

In 2026, Moore’s letter featured in the NRS exhibition For the Safety of All: Scotland’s Lighthouses. The exhibition ran in August and September and is now a past event. Its announcement provides another route to the letter and places the loss within the wider history of lighthouse work.

The light continued as a staffed station until automation in 1971. The operator lists no public access to the lighthouse. UKPX’s map pin identifies the station itself; it is a historical location marker, not a suggestion that the tower is open to visitors. Research guide · 2026 exhibition announcement · Station information.