Midnight Signals Case Files
Companion to Episode: The Lighthouse That Went Silent: The Flannan Isles Mystery And Three Missing Keepers
Two days after finding the lighthouse empty, Joseph Moore was still on Eilean Mòr. He had been sent back to keep the light burning, and on 28 December 1900 he wrote down what had happened. His letter began with the loss of his fellow keepers. Before it ended, he was describing the coat Donald McArthur had left behind.
Moore knew the place and the men. When the relief vessel Hesperus arrived on Boxing Day, he had expected the usual signs that someone ashore was preparing to receive it. There was no flag. Repeated blasts on the ship’s horn brought nobody out, and eventually a boat carried him to the east landing. He went up alone, leaving the others to wait below.
The gate was closed, as were the entrance doors. Inside, the kitchen door stood open. Moore looked at the fireplace, then went through the rooms and found the beds empty. He did not linger. His own account has him hurrying straight back to the boat, already certain that something serious had happened. He returned with the second mate and other men to search properly.
James Ducat, Thomas Marshall and Donald McArthur were gone. The lamp had been cleaned and its oil reservoir filled. The work needed to prepare the light had been done, but the men who should have lit it were nowhere on the island.
That evening Moore had to light it himself. Three volunteers remained with him, and the following day they searched from one end of Eilean Mòr to the other. By the time he wrote his letter, he could describe damage at the west landing, missing outdoor clothing and an intact crane. He could not tell anyone where his colleagues were.
The episode follows that search and the questions it left behind. The surviving correspondence also shows what happened after the first shock: keepers hurried from other stations, a superintendent visiting the widows, and a seaman asked to sit with Moore during his watch. Further back lie accounts of people who approached these islands under a quite different set of obligations, careful about where they stepped, what they said and how they treated one another.
The paid companion includes those earlier visits, a closer reading of the letters and landing evidence, three historical illustrations, a chronology and an illustrated case file to keep.
Before anyone kept the light
Long before a lighthouse stood on Eilean Mòr, boats came out from Lewis for the birds, eggs and sheep. Martin Martin described these journeys in his account of the Western Islands, published in 1703. Two of his informants had been to the Flannans the previous year. The men had rights to the islands and visited in summer, bringing home down, feathers and quills as well as food. A successful trip meant returning with a well-loaded boat.
Before landing, a newcomer had to learn the customs of the place. He was paired with someone who knew them, who would continue instructing him while they worked. The visitors believed that even a small breach might bring trouble upon them, so they arranged the whole crew in these partnerships.
The precautions began before landing. The crew approached with an east wind. If it turned west, even as they were about to go ashore, they headed home. Once the boat was secured against the rock, they put a wooden ladder in place and weighted its foot with a stone. After climbing up, the men uncovered their heads, turned sunwise and thanked God for bringing them safely across.
The rock where they had first set foot was treated with particular care. Relieving themselves there was forbidden, an offence believed to endanger everyone in the boat. Martin explained that the men respected this patch of rock because it was where they had escaped the ocean.
At the chapel, they put their outer garments on a stone provided for the purpose. Their morning observance involved three prayers: one while approaching on their knees, one while going around the chapel, and another close to it. They prayed again in the evening. Fowling had to wait until the prescribed arrival on the island was complete, and it stopped after evening prayers. Killing a bird with a stone was also forbidden.
Even ordinary speech changed. St Kilda was not to be called Hirt, but the high country. The Flannans themselves became the country. Familiar words for water, rock and shore were replaced by other terms. Martin recorded several of these substitutions, along with rules about the things that could be taken home. Sheep suet had to remain behind, however many sheep the crew killed.
One rule concerned a man’s partner directly: he must not secretly take or eat something without the other’s knowledge. Anyone who did, the men said, would vomit it up.
Martin called much of this superstition. His informants did not describe their conduct that way. When he asked one man whether he prayed as often or as fervently at home, the answer was no. The man said these remote islands possessed an inherent sanctity and that people arriving there found themselves more inclined to devotion than anywhere else.
Martin was recording what these men told him, not ceremonies he had witnessed himself. Their testimony does not establish that the keepers of 1900 followed the same customs. It places the island’s sacred reputation much earlier, among people who returned each summer to take what it provided.
A doorway three feet high
Joseph Anderson included the old building in Scotland in Early Christian Times in 1881, using the description and illustrations of T. S. Muir. The stones were fitted together without mortar. Outside, the little structure was roughly rectangular, though its opposite walls were not quite the same length. Inside, the chamber measured about seven feet by five. Its entrance was only three feet high, and there was no window.
The roof was especially distinctive. Slabs crossed between the walls to cover the interior, while from outside the roof resembled the bottom of a boat. The drawing preserves that heavy, rounded shape above the low opening. Anyone passing through the doorway would have had to stoop well down.
Anderson found its size and lack of an east window difficult to reconcile with an ordinary church. At the same time, its construction reminded him strongly of buildings with religious associations. He considered whether it might have been the dwelling of a hermit seeking seclusion on an island. He could not settle the question from the structure alone.
Neither his description nor Martin’s account supplies a secure construction date or identifies its first occupant. By the time Anderson was writing, visitors had been associating the place with prayer for at least as long as Martin’s book had been in print.
The naturalists who tried to land
A quite different expedition appears in J. A. Harvie-Brown and T. E. Buckley’s A Vertebrate Fauna of the Outer Hebrides, published in 1888. Harvie-Brown and the geologist Dr Heddle wanted to examine the Flannans, but their attempts were repeatedly governed by the swell around the rocks.
At Carloway they consulted Mr Macaulay, the postmaster and fisherman, who had often landed there. He agreed to accompany them if the weather allowed it. By that evening he was warning that the new swell would need two or three days to subside. They went out again nevertheless, hoping that a closer look might reveal a chance to land.
Off Eilean Mòr, the landing looked impossible. The party sailed around the group, watching the bird colonies and examining what they could from the vessel. They could see the remains of the chapel on the grassy slope. Eventually they tried the neighbouring island of Eilean Tigh, taking provisions into the small boat in case they got ashore and could not get off again.
Their first approach carried them between the island and a stack into apparently sheltered water. There was still nowhere practicable to land, so the boatmen pulled out again and waited. On a later rise of the water Harvie-Brown was able to get ashore with a bag and geological hammer. Heddle remained in the boat as the waves grew troublesome again.
The visit lasted no more than about an hour. Those who had landed crossed the top of the island and collected eggs, aware that a change in the tide might make re-embarking difficult within minutes. Heddle subsequently went ashore to chip a piece from a quartz vein. His leather-soled boots slipped, and he fell onto his knee.
The book records that the injury troubled him for six weeks. It also adds a teasing footnote: since the visitors had not followed Martin’s old observances on their first landing, might this have been retribution? The mishap was on Eilean Tigh, not at the future keepers’ landing on Eilean Mòr. Heddle left with an injured knee; the account of his fall carried the old customs into print again.
The work at the west landing
By December 1900 the station had been operating for just over a year. Ducat was its principal keeper, Marshall its second assistant, and McArthur was filling in for William Ross, who was on sick leave. Superintendent Robert Muirhead had recommended the men for the new station and knew them personally.
Their duties extended well beyond the lamp room. Stores and equipment had to be moved between the landings and the buildings above. The position of the west landing exposed its gear to seas that could reach astonishingly far up the rock. A crane there had already been swept away in the preceding winter. Its replacement had been put in during the summer of 1900.
When Muirhead arrived on 29 December, he could inspect only part of this damaged ground. The sea prevented him from reaching the actual landing place. He did reach the crane platform, which he put at about seventy feet above sea level. The replacement crane was in order, securely lashed, and showed no indication that the men had been working with it.
Higher up had stood a box holding mooring ropes and tackle. Muirhead estimated its position at 110 feet above sea level. The box was gone. Coiled ropes lay about, still in their coils, and none had been made fast to the crane or railings. Ironwork had been torn out or twisted. A stone estimated to weigh more than a ton had been displaced. The remaining ropes and fragments of canvas indicated that a lifebuoy had been torn from its fastening rather than simply untied.
Moore had noticed this damage too, but his letter was cautious about what it proved. He had no direct indication that the men had died at the west landing. He was trying to connect the condition of the place with the clothes missing from the station. Ducat’s sea boots and old waterproof coat were gone. Marshall’s sea boots and oilskins were missing. McArthur’s usual coat was still there.
Because he knew what McArthur habitually wore, Moore thought he had gone out in his shirt sleeves. That inference has helped later storytellers imagine someone calling him urgently from the building. It remains an inference about clothing. It cannot establish who went out first, whether anyone shouted for help, or whether all three men reached the same spot together.
Muirhead thought they had gone down to secure the rope box and been overtaken by an exceptionally large sea. In his reconstruction, water rose above the men and swept them away as it came down. The damage made that a serious explanation. His report did not provide a witness to the wave or establish the men’s last movements.
The missing box held equipment the keepers needed to receive a boat. Securing it would have been part of looking after the station. But the investigation could not follow the men from the building to the damaged ground, or say what had brought the last man out of the lighthouse.
Keeping Moore on the island
The first replacements had to be found while the Board was still learning what had happened. Muirhead arranged for a principal keeper to move from Stornoway to Tiumpan Head, freeing John Milne to take charge at the Flannans. Donald Jack was sent as well. They landed with the superintendent on 29 December and joined Moore.
The light had already been burning again for three nights. Allan MacDonald, the buoymaster, and the seamen Campbell and Lamont had volunteered to remain with Moore after the discovery. His letter states that they lit it at the proper time on 26 December and every night afterwards. In between came the search across the island.
Muirhead first checked that the incoming men could maintain the light. Only then did he turn to his inquiry. He took statements from the captain, the second mate, Moore and MacDonald, and examined what he could reach. By Saturday afternoon he was back at Breasclete, sending the result of his investigation and calling on the widows of Ducat and McArthur.
Moore’s condition needed a separate arrangement. He was very upset and appeared nervous. If he did not recover, Muirhead thought he would have to be transferred, though he was reluctant to recommend it. With new men taking over, he wanted at least one keeper who already knew the work of the station.
For the time being, Lamont would stay behind for a week or two. The seaman was to go to the lightroom with Moore and keep him company during his watch.
The people looking from Lewis
Across the water, Roderick MacKenzie had been appointed to watch for signals and note whether the light could be seen at night. When Muirhead went to his home on 30 December, MacKenzie was away. His two sons, about sixteen and eighteen, were there. They were the ones who actually carried out the observations, and Muirhead questioned them and examined the return book.
Their entries showed the light on 7 December, then several nights without seeing it, before it appeared again on the 12th. It was not seen afterwards until Moore relit it on the 26th. The tower itself had not been visible through the telescope between 7 and 29 December.
MacKenzie had grown anxious. For the two nights before the light reappeared, he had enlisted other people to help look for it. Muirhead thought a trained lightkeeper ashore might have raised the alarm sooner. He also wrote that earlier notice could have restored the light sooner, not prevented the men’s deaths.
The report went further than the often-retold discussion of visible signals. Someone had raised wireless telegraphy. Muirhead considered what a failure to make contact by wireless would have meant to him. His first suspicion, he said, would have been faulty signalling apparatus. The disappearance of all three keepers would have been the last thing to occur to him.
On his last visit, on 7 December, Ducat had talked about testing how quickly an appeal for assistance would be noticed and acted upon. Muirhead had considered a daily signal that all was well, but decided against it. Haze could conceal it for so long that the keepers’ families would become alarmed. The tower’s absence from the boys’ December observations shows the difficulty he was considering.
The room in the poem
Wilfrid Wilson Gibson’s poem Flannan Isle gives its search party a different arrival. Three strange black birds sit near the water, unnervingly like upright sailors, before vanishing beneath it. Inside the lighthouse, a meal is untouched and a chair has fallen over. The visitors also recall a history of death and madness attached to the light.
These are details within Gibson’s poem, not Moore’s statement. They make a powerful scene, with the missing men seeming almost present in the birds and the abandoned places at the table. The actual inquiry described the kitchen as tidied after a meal and the station as prepared for its next night’s work.
The episode already separates the poem’s table from the contemporary record. Reading the whole poem shows how much else belongs to Gibson’s imagined visit. It can be appreciated as a haunting response to the disappearance without being used to furnish the lighthouse for Joseph Moore.
After the inquiry
Ducat’s career had begun in 1878. By the time he went missing he had spent more than twenty years in lighthouse service. His wife Mary was left with four children, the youngest six. Marshall, who had entered the service in 1896 after working as a seaman, was unmarried. McArthur left a wife and two children. These family details survive through registers and records brought together by the National Records of Scotland.
Muirhead ended his report by remembering the men at work in the summer of 1899. He had helped choose them for an important new light, and he had seen the effort they put into getting it ready. His last recollection was of shaking their hands and saying goodbye on 7 December 1900.
The station remained manned until automation in September 1971. The papers from December 1900 preserve the attempts to keep it working while the men on the island and the families ashore waited for news.
On the evening the loss was discovered, the captain of Hesperus sent his telegram and told the Board he would not return to Oban until he heard from them. He had left men on the island to keep the light going. He would wait at the telegraph office until it closed, in case a reply came.
The dates in the surviving record
1703
Martin Martin publishes his account of the Western Islands, including the customs described to him by visitors to the Flannans.
Publication date, not the date of every custom or visit he records.
1881
Joseph Anderson publishes the small building on Eilean Mòr in Scotland in Early Christian Times, reproducing T. S. Muir’s description and illustrations.
1888
Harvie-Brown and Buckley’s A Vertebrate Fauna of the Outer Hebrides appears, including the Flannan map and the account of a difficult expedition to the group.
The book was published before the lighthouse existed; its expedition account is not evidence of the 1900 disappearance.
7 December 1899
The Flannan Isles light is first exhibited.
7 December 1900
Muirhead visits at the relief and speaks with Ducat about signals. It is the superintendent’s last meeting with the three men.
13 to 15 December 1900
The log is written up through 13 December. Details for the following day and the morning of the 15th remain on the slate, including the morning weather readings and the time the light was extinguished.
Muirhead infers that the men disappeared on the afternoon of the 15th. This is not an eyewitness time of death.
26 December 1900
Hesperus arrives for the relief. Moore discovers the empty station. He and the volunteers restore the light that evening; the captain sends his telegram.
28 December 1900
Moore writes his account from the lighthouse, describing the search, the landing damage and the missing clothing.
29 and 30 December 1900
Muirhead inspects the station and west landing, visits the widows at Breasclete, then questions MacKenzie’s sons and examines the observation book the following morning.
8 January 1901
Muirhead submits his report. He proposes that a large sea swept the men away and records the temporary staffing and support arranged for Moore.
1912
Gibson’s Flannan Isle is published, giving the disappearance an influential literary treatment.
28 September 1971
The lighthouse is automated, ending the resident keepers’ watches.
Sources and image credits
The episode and the December papers
The Midnight Signals episode. The Lighthouse That Went Silent: The Flannan Isles Mystery And Three Missing Keepers, released 31 August 2026. The full public transcript associated with the released video was the baseline for this companion. The correspondence and earlier books below supply the additional material.
Moore, the captain and Muirhead. The Northern Lighthouse Board’s Flannan Isles document collection reproduces the captain’s telegram of 26 December 1900, Moore’s letter of 28 December, and Muirhead’s report dated 8 January 1901. Reading them in that order shows how an immediate report became an investigation. The clothes and damaged equipment should be read with each writer’s qualifications intact. The captain’s name appears as both Harvey and Harvie in the material, and McArthur also appears as MacArthur in retellings.
The keepers’ families. Jocelyn Grant’s National Records of Scotland article, 12 December 2023, brings together employment, census and death records. It is the source for the brief family and service details above, and provides archival references for further enquiry. Its photographs and register reproductions have not been reused here.
The station’s history. The Northern Lighthouse Board’s Flannan Islands lighthouse record supplies the commissioning and automation dates. For the December investigation, the full contemporary papers take precedence over compressed modern summaries.
Earlier journeys and a later poem
Martin Martin, A Description of the Western Islands of Scotland (1703). The Islands off Lewis section preserves the account of the Flannan visitors and their observances. It is particularly useful for the partnership rule and the distinction between what Martin was told and what he witnessed. The work is public domain; modern page additions are separate.
Joseph Anderson, Scotland in Early Christian Times (1881). The digitised volume, pages 120 to 122 and figures 50 and 51, contains the chapel description drawn from T. S. Muir. Anderson’s uncertainty over its use is part of the account, not a modern discovery that it was never a chapel.
J. A. Harvie-Brown and T. E. Buckley, A Vertebrate Fauna of the Outer Hebrides (1888). The digitised book, introductory section “The Flannan Isles,” pages xxiii to xxxiv, includes the map, the boat approaches, Heddle’s injury and the footnote about disregarding the old customs. The original print scan was checked alongside its imperfect machine-readable text. These are historical collecting practices, not a guide to visiting or taking wildlife today.
Wilfrid Wilson Gibson, Flannan Isle (1912). The complete poem is a useful final reading after the correspondence. Its birds, meal and imagined earlier tragedies belong to the poem. The publication date is also recorded in the poem’s bibliographic entry. No lines from the poem are reproduced here.
Image credits and rights
1. Flannan Isles map. Mapmaker not identified in the accessed plate; published in Harvie-Brown and Buckley, A Vertebrate Fauna of the Outer Hebrides, David Douglas, Edinburgh, 1888. Digital reproduction uploaded by Roger Griffith (Rosser1954), with contrast adjustment by PawełMM. Image record and public-domain declaration; original book and Public Domain Mark. The Commons description gives 1898, but the map was verified in the 1888 book. The earlier publication date is used here. No new crop or alteration has been made.
2. Teampull Beannachadh, west end. Illustration from T. S. Muir’s earlier work, reproduced by Joseph Anderson in Scotland in Early Christian Times, David Douglas, Edinburgh, 1881, figure 50, page 120. Digital reproduction uploaded by Kognos. The item’s rights record identifies public-domain status in the country of origin under the expired life-plus-70 term and in the United States through early publication. The drawing depicts the actual Eilean Mòr building. No alteration has been made.
3. Admiralty Chart 3331, Flannan Isles. United Kingdom Hydrographic Office, first published 7 April 1903, with soundings credited to H. C. Otter (1857) and A. M. Field (1902). This scan includes later corrections and is not presented as the exact chart available to the keepers in 1900. Digital scan uploaded by Kognos. The item record and rights basis identify the work as public domain through expired Crown copyright, including worldwide application of that expiry. No alteration has been made. Not for navigation.
