On the morning of 6 December 1917 a railway dispatcher named Patrick Vincent Coleman stayed at a telegraph key a few hundred yards from a burning munitions ship and sent a warning up the line before the ship exploded and killed him. The story is told constantly in dispatch training, usually with numbers and a quotation that have grown in transit. This piece separates what the Halifax record supports from what retellings have added, and then asks the harder question about what an agency actually owes a person sitting at a communications post inside a hazard zone.

Nineteen minutes in the Narrows

The French steamer Mont-Blanc had loaded a wartime cargo in New York and arrived off Halifax on the evening of 5 December 1917, too late to pass the submarine net into the harbour, so she anchored outside and came up the next morning under a local pilot. The Norwegian steamer Imo, chartered for Belgian relief work, was outbound. The two ships met in the Narrows, the constricted channel between Halifax and Dartmouth, and collided at roughly 8:45 in the morning. The impact opened Mont-Blanc’s forward hold and knocked over drums of benzol stowed on deck, and the resulting fire could not be fought, so the master ordered the ship abandoned and the crew rowed for the Dartmouth shore.

The cargo is described in published histories as picric acid, TNT, guncotton and benzol totalling on the order of 2,900 tons, and the individual tonnages differ between accounts depending on which manifest entries are counted, so if you need figures you can defend, go to the published histories and the inquiry record rather than to a summary like this one. The burning ship drifted across the harbour and came to rest against Pier 6 in the Richmond district on the Halifax side, which was a working waterfront of railway yards, factories, a sugar refinery and dense housing.

For something close to nineteen minutes the fire burned in plain view, drawing dockworkers, firefighters and a great many ordinary residents to windows and doorways and to the waterfront itself, because a ship on fire in the harbour was a spectacle and almost nobody ashore knew what was in the hold. The explosion came at about 9:04 a.m. It flattened the Richmond district, generated a wave that swept the shoreline on both sides including the Mi’kmaq settlement at Tufts Cove, and is routinely described as the largest accidental man-made explosion before the nuclear era.

The man at the Richmond yard key

Coleman was a train dispatcher for the Canadian Government Railways, working out of the telegraph office at the Richmond railway yard with chief clerk William Lovett. His job that morning was the ordinary one of a dispatcher on a single-track railway, which is to know where every train on his territory is and to issue the orders that keep two of them from occupying the same piece of track. That work is done by telegraph, from a fixed desk, with a train sheet in front of you, and it is not a job you can do while walking away from the building.

A sailor reached the office and told the two men what Mont-Blanc was carrying. Accounts differ on who he was, with some versions identifying him as a member of the Mont-Blanc’s crew and others as a sailor from a naval vessel in the harbour, and I have not seen a version of that detail I would call settled. Lovett telephoned the terminal agent. Both men started to leave, and Coleman went back to the key because there was a passenger train inbound from the west that would run straight into Richmond unless somebody stopped it.

He sent his warning and he was killed when the ship exploded. Lovett died as well. The office stood close enough to Pier 6 that survival at either desk was not realistically in question once the detonation occurred, and that geometry matters for the lesson later in this piece, because Coleman was not choosing between a safe distance and a risky one. He was choosing between spending his last minutes moving and spending them sending.

The message, and why the wording is not settled

The wording most often reproduced runs along the lines of holding up the train, an ammunition ship afire in the harbour making for Pier 6 and about to explode, a line saying this will probably be his last message, and a goodbye. I am deliberately paraphrasing rather than quoting, because the versions in circulation differ from one another in phrasing, punctuation and length, and I have not found one that is established as the exact transmitted text rather than a reconstruction. The original telegraph form, if one existed at the receiving end, is not something I can point you to with a citation I would stand behind.

That distinction is worth making in front of a training class rather than skipping. A quotation attributed to a dead man, reproduced with quotation marks in a hundred retellings, acquires the appearance of documentary evidence without ever having been documented, and the same mechanism produces the invented last radio transmissions that circulate after fire and police line-of-duty deaths. The Nova Scotia Archives and the Maritime Museum of the Atlantic in Halifax hold the primary material on the explosion, and that is where the question of exact wording should be settled if it can be.

What is not in dispute is the substance of what he sent, because the operational content survived in the behaviour of the railway. A dispatcher’s warning went out over the railway telegraph, it identified the hazard as an ammunition ship about to explode, it named a location, and it directed that a train be held. Those four elements are the whole of a usable emergency message, and they were assembled and transmitted by one man in something under two minutes with a burning munitions carrier alongside the pier down the street.

What the train actually did

The train usually named is No. 10, the overnight express inbound from Saint John, New Brunswick, due into Halifax around 8:55 that morning and commonly described as carrying roughly 300 passengers. The widely published account has the train halted short of the city, at or near Rockingham, several miles from the blast, with its passengers unhurt. Whether it was Coleman’s message that produced that outcome is less certain than the retellings suggest, because the train was running behind schedule and accounts of how far behind it was differ, which means the counterfactual position of that train at 9:04 a.m. cannot be reconstructed from a timetable.

The number saved is where the story inflates most reliably. The figure of roughly 300 is the passenger complement usually given for that train, not a count of people who would otherwise have died, and those are different quantities. Some retellings raise the figure to 700, some describe multiple trains stopped, and some add hundreds of railway workers to the total. I would not put any of those numbers in a declarative sentence. The honest statement is that the message reached the railway telegraph network, that the inbound express did not enter the destruction zone before the explosion, and that the warning also put the railway on notice at points outside Halifax where the news would matter within the hour.

None of that diminishes what Coleman did, and the instinct to inflate it usually comes from a fear that the plain version is not impressive enough. A dispatcher who understood a hazard from a single verbal report, decided within seconds that it was credible, identified the one train his warning could still affect, and transmitted a clear hold order while knowing he would not survive it, performed the job correctly under conditions no procedure anticipates. The documented version is sufficient.

The toll, and the numbers that drift in retelling

The figure you will see most often is approximately 2,000 killed and around 9,000 injured, and that is the range published by general reference sources including the Canadian Encyclopedia and by the Maritime Museum of the Atlantic. It is a rounded estimate rather than a count. The Nova Scotia Archives maintains the Halifax Explosion Remembrance Book, which records victims by name from the death registrations and related records, and the number of named individuals in that compilation falls somewhat short of the round 2,000, which is entirely normal for a disaster in which bodies were destroyed, harbour workers were transient, and some households died entire with nobody left to report them.

So there are two defensible ways to state it, and they do not agree. If you want a documented count of named dead, the Remembrance Book is the source and the number is lower. If you want the historical estimate that includes the unrecovered and unrecorded, approximately 2,000 is the figure the museums and reference works publish. Give both and say they differ rather than picking whichever fits the sentence, and send anybody who needs a citable number to the Nova Scotia Archives directly.

The injury figures deserve the same treatment. Because the fire burned for close to twenty minutes in full view, a large share of the wounded were hurt by window glass while watching, and eye injuries were numerous enough that the medical response to them is part of the standard history and is connected to the founding of the Canadian National Institute for the Blind in 1918. Published counts of eye injuries and of permanent blindness vary between accounts, and I would not quote a specific number without checking it against the Halifax institutional sources first. Property figures vary as well, with the usual published description being the near-total destruction of the Richmond district and damage across a large part of the city, leaving thousands homeless in December.

How word got out of a city with no telephone lines

The blast wrecked local telephone and telegraph plant across the north end of the city, along with the yards, the poles and much of the rolling stock in the immediate area, which left Halifax in the condition every communications officer recognises from more recent events, namely a jurisdiction that cannot report its own situation. The path that worked was the intercity railway telegraph, whose offices and line continued to function outside the damage radius, and whose operators were already trained to pass traffic in a fixed form to named destinations.

Relief moved on that path because it was the path that could carry both the message and the load. A relief train from Truro, roughly sixty miles away, was moving toward Halifax on the morning of the explosion carrying medical personnel and supplies, and further trains followed from other Nova Scotia points. In Massachusetts, where the news arrived the same day, the state organised a relief expedition that left Boston on the evening of 6 December and reached Halifax on 8 December, delayed by a blizzard that struck the region on 7 December and buried the city under heavy snow while survivors were still being dug out of wreckage.

The structural lesson is one that shows up again in Katrina and in every hurricane since, which is that the surviving communications path out of a destroyed area is usually the one that was engineered for something else and happened to have its infrastructure outside the footprint. A railway telegraph existed to move trains, its circuits ran to towns nobody in Halifax was thinking about that morning, and it became the outbound warning and mutual aid channel for a city-scale mass casualty event because it was the thing still standing.

Staying at the post, and the part that does not transfer cleanly

Dispatch training uses Coleman as an example of remaining at the position under threat, and the example is sound as far as it goes, because telecommunicators do stay through tornado warnings, through civil unrest outside the building, through structure fires in their own facility and through events where the only alternative is leaving the field without a voice. The Halifax case is cleaner than most because Coleman’s decision cost him only the last two minutes of a life that the blast was going to take regardless of where in that yard he stood.

The cases your agency will actually face are messier, and they are the ones where staying costs a life that could have been saved. A telecommunicator who remains in a building that is on fire, or inside a plume, or in the collapse zone of a compromised structure, is not being heroic if the traffic could have been carried from a secondary position two miles away starting four minutes later. The decision that matters is made in advance by the people who write the continuity plan and who choose where the backup centre is, how long it takes to bring up, and who has the authority to order the move without calling anyone for permission.

That authority question is where most small agencies are weakest. Ask a shift supervisor at three in the morning whether she can order an evacuation and relocation of the centre on her own signature, and in a lot of places the honest answer is that she would call the director and hope he answers. If your plan requires an off-duty phone call before anyone leaves a hazardous building, you have written a plan that depends on the hazard being polite about its timing, and that paragraph needs rewriting this quarter rather than after the incident.

What a modern centre should have written down

Coleman had no procedure for what he did, and he improvised a correct emergency message from a single unverified verbal report because he knew his territory well enough to identify the one train at risk. A modern centre can do better than improvisation on two fronts. The first is the outbound warning itself, which in 2026 means knowing without looking who can push a Wireless Emergency Alert for your county, what the geographic targeting looks like, which siren or mass notification systems you control directly, and how long each of those takes from decision to delivery. If the person on duty has to read a binder to answer any of that, the answer arrives too late.

The second front is the credible single report. Coleman acted on one sailor’s statement with no verification and no time to get any, and he was right, but the same behaviour on a different morning produces a false alert. The useful design question is which categories of single unverified report your centre is authorised to act on immediately and which ones require a second source, and that decision belongs in policy written by your director rather than in the judgement of whoever happens to take the call. Hazardous materials release from a named fixed site, dam failure, and an active threat report from an on-scene officer are the usual members of the act-immediately category, with the wording of the public message pre-drafted so nobody composes one from scratch.

Both of those are unglamorous administrative work, and they are the parts of the Halifax story that actually transfer. The record does not show that Coleman was braver than the people around him, and it does show that he knew his system, recognised what the information he had just received meant for it, and used the one tool at his desk to act on it. Everything an agency can do in advance to make that easier for the person on duty is worth more than a poster of the story on the wall of the break room.

What to do at your agency

• Have your communications manager write down, on one page, the name and phone number of every person authorised to originate a Wireless Emergency Alert for your jurisdiction and the after-hours path to reach each of them, then verify the list by calling two of the numbers this month.
• Ask your director or 911 board for the paragraph in your continuity plan that names, by shift position rather than by person, who can order an evacuation of the communications centre without further approval, and if that paragraph does not exist, draft it and put it on the next board agenda.
• Have your shift supervisors time a real relocation to your backup position during a slow overnight, from the order to the first dispatched call from the new location, and record the actual elapsed time rather than the planned one.
• Have your GIS or planning staff plot your primary centre, your backup centre and your main tower sites against the hazard footprints already in your hazard mitigation plan, including rail corridors carrying hazardous materials, fixed chemical sites, dam inundation areas and flood zones, and report at your next safety committee meeting which facilities share a footprint.
• Write into your existing dispatch policy a short list of report types that are acted on immediately from a single unverified source, with the pre-drafted public message text for each, and have your training officer brief it at shift meetings this month.
• Have your training officer identify one communications path out of your jurisdiction whose infrastructure sits entirely outside your worst-case hazard footprint, whether that is an amateur or auxiliary net, a satellite terminal, a state agency circuit or a neighbouring county’s system, and test it on a scheduled date with the results written down.
• If you use the Coleman story in your telecommunicator orientation, correct the slide so it states that the quoted wording is traditional and unverified and that the number of lives saved is an estimate, and point students to the Nova Scotia Archives for the documented record.

Takeaways

• Mont-Blanc and Imo collided in the Narrows at roughly 8:45 a.m. on 6 December 1917, the resulting fire burned in view for close to nineteen minutes, and the explosion occurred at about 9:04 a.m. against Pier 6 in the Richmond district of Halifax.
• Vince Coleman was a Canadian Government Railways train dispatcher working a telegraph key a few hundred yards from Pier 6, and he returned to that key after being warned about the cargo in order to stop an inbound passenger train.
• The famous wording of his message exists in several differing versions and should be described as traditional rather than as an authenticated transcript, with the primary record held by the Nova Scotia Archives and the Maritime Museum of the Atlantic.
• The inbound express, usually identified as No. 10 from Saint John and usually described as carrying about 300 passengers, did not enter the destruction zone, although the train was running late and the number of lives the warning saved cannot be established.
• General reference sources including the Canadian Encyclopedia and the Maritime Museum of the Atlantic publish approximately 2,000 killed and around 9,000 injured, while the Nova Scotia Archives Halifax Explosion Remembrance Book documents a somewhat smaller number of victims by name, and those two figures do not agree.
• Local telephone and telegraph plant was destroyed, and the intercity railway telegraph carried the first word out because its offices and lines lay outside the damage radius, with a relief train from Truro moving toward Halifax the same morning and the Massachusetts expedition leaving Boston that evening and arriving on 8 December through a blizzard.
• Staying at a communications post is defensible when the traffic cannot be carried from anywhere else and the exposure is unavoidable, and it is indefensible when a functioning backup position exists and the plan failed to authorise anyone on duty to order the move.
• The transferable work is administrative, meaning known alert origination authority, pre-drafted warning text, a named shift position with evacuation authority, and at least one outbound path whose equipment sits outside your own hazard footprint.