Gibraltar Point Lighthouse has two overlapping histories: the death of its first keeper, John Paul Radelmüller, and a ghost tradition attached to the Toronto Islands landmark. They require different kinds of evidence. Historical records can establish a death, a criminal proceeding and a verdict without establishing the familiar details of a haunting. John Ross Robertson’s 1908 account is particularly useful because it preserves an older version of the story while openly doubting parts of it.
The later ghost tradition
Toronto Public Library’s 2021 historical picture essay explicitly records the lighthouse’s haunted reputation. Its account distinguishes versions involving a fatal quarrel, concealed remains or a keeper who vanished. It also reports the tradition that his ghost searches for his bones. These are documented forms of a story, rather than an authenticated account of an apparition. The library’s treatment supplies a clear cultural context, but not a dated witness statement with the circumstances needed to assess a particular sighting.
What Robertson recorded in 1908
In Landmarks of Toronto, fifth series, Robertson described the lighthouse, its keeper’s accommodation, lamps and household fittings before turning to the death. His setting matters because the narrative concerned a working lighthouse community, rather than an isolated tower imagined without its supporting buildings. He discussed the first keeper under the shorter names Muller or Miller and identified him as German. His account of that man’s habits and character belongs to the tradition he collected; it should not be treated as an independently verified biography.
The familiar drinking episode appears in this early printed version. Robertson said soldiers visited the keeper and obtained beer from him. In the fatal encounter, a group arrived after drinking in York. The keeper initially supplied them but refused further drink, a fight followed and they beat him with belts and a stick. Robertson presented this as a story handed down over generations. He did not identify himself as an eyewitness, reproduce testimony from the confrontation or establish how each detail had survived.
The number in this version is also worth preserving accurately: Robertson described three visitors. A later court record concerns two defendants. Those are different categories of information. A storyteller’s group size is not a criminal finding, and the existence of named defendants cannot automatically be used to confirm every act attributed to unnamed soldiers in an oral narrative. The account is valuable as evidence of what was being told, even where its details cannot be established as events.
A collector who doubted his own story
Robertson did not simply endorse the tale. Under separate headings about whether the murder had happened, he considered the possibility that the account had been embellished or invented. At the same time, he identified a chain of transmission: the later keeper George Durnan said he had heard it from his father. That attribution is more useful than an anonymous modern assertion, but it still records family recollection at a considerable distance from 1815.
Durnan’s other contribution was a reported discovery in 1893. According to Robertson, George and his uncle Joe found pieces of a coffin and part of a human jaw beneath the sand. The search was motivated by a tradition about where the first keeper had been buried. The book does not supply a forensic identification, a documented match to Radelmüller or an independent examination establishing whose remains they were. Finding something compatible with a burial story would not, by itself, establish the identity of the person buried.
The distinction is especially important for a haunting account. An uncertain identification of human remains does not become evidence of a ghost because the search was prompted by a legend. Nor does Robertson’s report show that the alleged apparition led anyone to the remains. The preserved narrative connects a family story and a purposeful search; stronger supernatural details require their own sources.
Robertson also mentioned two young men who visited Durnan in 1903, describing themselves as the first keeper’s nephews and asking about his fate. He reported that Durnan repeated the same account. This passage illustrates the story’s continuing circulation, but it supplies no new witness to the death and does not resolve the identification of the bones. It is another episode in the history of inquiry and remembrance.
The documentary correction
Robertson believed he had found no relevant newspaper notice or trial record. Historian Eamonn O’Keeffe subsequently located contemporary documentation contradicting that negative claim. His published source selection includes a 5 January 1815 administrative letter reporting the keeper’s death, the York Gazette notice dated 14 January identifying 2 January as the date of death, and court minutes for the subsequent prosecution. The 31 March 1815 verdict for the two defendants, John Blowman and John Henry, was not guilty. They must not be presented as proven killers.
This correction changes the historical footing of the case. Robertson’s unsuccessful search cannot be used to dismiss the death as invented. Equally, a record of death and an acquittal cannot authenticate every later detail about drink, weapons, dismemberment, burial or an apparition. The evidence must be followed separately for each proposition. The surviving record does not turn a ghost narrative into an eyewitness deposition.
Assessment and location
The paranormal element is classified as insufficient evidence. There is a well-established historical setting and a traceable storytelling tradition, but the sources reviewed here do not provide a dated, independently corroborated ghost encounter. This assessment concerns the haunting claim; it does not deny the documented death or imply that an acquittal proves that no fatal violence occurred. Historical uncertainty and evidence of paranormal activity are different questions.
The map uses an approximate Centre Island locality reference, based on GeoNames entry 5919186 and rounded to 0.01°. Toronto Public Library identifies the lighthouse with Centre Island. The marker represents the island setting, not a surveyed tower position, a burial site or directions to any keeper’s former dwelling. The incident-year filter remains unassigned: 1815 is the keeper’s death year, while the reviewed ghost tradition supplies no reliably dated apparition.
Sources and related reading
John Ross Robertson, Landmarks of Toronto, fifth series (1908), especially pages 378–385; Eamonn O’Keeffe’s primary-source selection with archival references; Fife and Drum, December 2015, pages 3–4; and Toronto Public Library’s lighthouse history and picture essay (2021). The 1908 text, later archival correction and modern curatorial account serve different evidential purposes.
For a Canadian investigation centred on reported physical disturbances rather than a landmark’s ghost tradition, see Caledonia Mills, 1921–22.