Clarendon Township, near Shawville, Quebec, Canada · Ghosts and hauntings · September–November 1889.
The Dagg case is unusual among nineteenth-century haunting reports because its alleged agent did more than make noises: it argued, contradicted itself and reportedly sang. The principal visitor, Canadian artist Percy Woodcock, believed that an unseen intelligence was present. His published account also gives a reader reasons to distinguish carefully between what he heard, what the family told him and what happened after he left. Those distinctions are essential to assessing the story.
The report and the household
The substantial account reviewed here appeared as “Remarkable Spontaneous Manifestations” in Light on 28 December 1889, pages 620–623. The introduction says it was extracted, with slight abridgement, from the Recorder of Brockville. It places the events at George and Susan Dagg’s farmhouse in Clarendon Township, several miles from Shawville. The reported onset was 15 September 1889. Woodcock arrived in Shawville on Thursday 14 November and visited the house the following morning.
The household included young children and a girl named Dinah who had been placed with the family some years earlier. The article records that the troubles stopped when she was away. That association became important to both believers and sceptics, but it does not itself establish whether she produced the effects, experienced something unexplained or became the focus of other people’s expectations. The surviving public report does not justify a diagnosis or an accusation against a child.
What happened before Woodcock arrived
Much of the opening narrative was information Woodcock collected from the Daggs and neighbours. It describes missing money, spoiled food, broken windows, small fires, objects moved about the house and clothing or hair apparently interfered with. Initial suspicion fell on a boy who denied wrongdoing; later suspicion involved a neighbouring family. These are accusations within the historical report, not findings endorsed by this article.
One often-repeated episode concerns a black dog. The Light account specifically tries to correct other reports: Dinah said she had seen a large dog-like shape knock down a fence, but when others went outside they saw no animal. What they could examine was the fallen fence. That distinction changes the evidential claim. Several people seeing a damaged fence is not the same as several people watching an extraordinary creature cause the damage.
The conversation in the shed
On Saturday, Woodcock reported going with Dinah to an open shed behind the house. After she addressed the supposed presence, he heard a coarse voice seemingly nearby. It threatened him and used language the newspaper would not print. Woodcock said he challenged its behaviour rather than accepting its claimed identity. George Dagg joined them, and the account describes a conversation continuing for hours and later moving into the house as visitors arrived.
Woodcock considered a concealed speaker or ventriloquism. He described the shed as open, searched the premises and watched Dinah. He also asked her to fill her mouth with water; he said the conversation continued. These details show that he thought about ordinary agency and attempted checks. They remain his report of an informal examination, without a recording, a complete procedure or independent instrumental documentation from which the sound’s source can now be reconstructed.
He further claimed that a pencil stood up and wrote after he requested a sample of writing, then was flung across the shed. The account does not supply a tested surviving specimen here. Its importance is that Woodcock presented this as a personal observation, unlike the numerous preceding incidents he learnt about from others. His sincerity and his interpretation are separate questions from whether the published description establishes an unseen author. Read the scanned original report.
An unreliable voice and an accused neighbour
The voice attributed the trouble to a neighbouring woman amid an existing boundary dispute. Woodcock brought her to the house so the allegations could be challenged. According to the report, she denied them; the voice tangled itself in contradictions and admitted lying. Its allegations then shifted to her children, who also denied knowledge of the supposed practices. A promised demonstration involving a hidden book was not produced.
This episode is a reason not to repeat the voice’s claims as factual information about the neighbours. The report itself portrays the accusations as unreliable. Woodcock’s investigation was not simply an endorsement of everything said in the house: he pressed for consistency and rejected claims he considered false, while continuing to believe that the source of the speech was extraordinary.
On Sunday the voice reportedly claimed to be an angel rather than the abusive presence heard previously. Woodcock objected that it sounded the same and again drew attention to contradictory answers. Some visitors resisted his scepticism about the angelic identity. The disagreement is valuable evidence about how people in the room interpreted the performance; it is not independent confirmation of either supernatural identity.
The signed statement and the farewell
Woodcock prepared a collective statement dated 17 November. Seventeen names were printed beneath a summary of fires, projectiles, moved objects, the voice and the children’s reported visions. A signed collective account establishes that named people were willing to endorse a narrative. It does not tell us, incident by incident, exactly which signatory directly observed each event or how independently their accounts were obtained.
The article explicitly says Woodcock left around 10.30 on Sunday night. Its subsequent description of hymn-singing, prolonged farewells and a clergyman’s visit came from the Daggs and other people who stayed. That limit is often lost in shorter retellings. Woodcock returned on Monday morning and was present when the children came in excitedly describing a beautiful figure ascending amid fire. Adults went to the door but saw nothing unusual. He witnessed their account being given, rather than the alleged ascent.
The following year’s argument
A Light follow-up of 22 November 1890, transcribed in the Poltergeist Archive’s collection of original reports, contains Woodcock’s letter dated 29 October and a report from teacher Robert Grant. Woodcock said there had been no repetition after Dinah was removed and complained that further examination had not been permitted. His claims about other people’s reasons for restricting access are allegations in his letter, not independently verified explanations.
Grant described a retrospective enquiry, not a visit during the disturbances. He encountered a fraud-and-ventriloquism explanation among people farther away, then supportive accounts nearer the family. He asked what witnesses had actually seen. Arthur Smart, for example, described objects found elsewhere without having watched their entire movement; Mrs Dagg said she had seen a table turn. Grant found them convincing, but his interviews did not reproduce the phenomena.
The same archive preserves a March 1890 letter signed “John Jones” opposing the ventriloquism hypothesis on claimed technical expertise. It was a response to published descriptions, rather than a fresh demonstration at the farmhouse. Neither its confident rejection nor the contrary rumours can substitute for a controlled observation. The follow-up documents continuing disagreement and belief; it does not establish a final identification.
Assessment and map
Insufficient evidence. The original report contains specific observations and attempted checks, but also substantial hearsay, collective testimony and unrecorded speech. The reviewed follow-up does not supply a reproducible identification of the voice or demonstrate a paranormal cause. The reported cessation of the disturbances is an ending in the chronology, not a verified explanation of every episode.
The map uses an approximate Shawville town reference, 45.60°N, 76.48°W, rounded from GeoNames 6145563. It identifies the regional reference named in the original account. It is not the farmhouse, a residential address or permission to visit private land. The modern street photograph illustrates Shawville only.
Sources reviewed 13 November 2026. The 1889 report is linked as a complete scanned issue; the 1890 letters were read as transcriptions, with that access difference retained. For a different Canadian household tradition and its conflicting testimony, see the Baldoon Mystery.