The Only Way Out Was On Fire

Watch: New safety breakdowns every week on YouTube — youtube.com/@GetSafetyIntel

Share: Got a near-miss you can't put your name on? We publish them anonymously. QR + link at the bottom.

Seventy men and boys went down the Mauricewood pit that morning. Seven came back up.

Mauricewood was a Shotts Iron Company colliery on the edge of Penicuik, a few miles south of Edinburgh, working coal several hundred feet under the Midlothian farmland. Around noon on 5 September 1889, fire took hold in the pit's ventilation shaft — by the record, the timber lining itself caught, and the fire spread into a nearby coal seam.

Understand what that means, because it is the entire story. The ventilation shaft was not a side system. It was the machine that made the mine survivable — the thing that pushed clean air down to the men and pulled the bad air out. On 5 September 1889, that machine became a chimney pointed at seventy people.

Smoke from the fire in the return airway drove into the workings. The way out was on fire, or filling with the products of it. There was no second egress worth the name, no self-rescuer on any man's belt, no rescue brigade with breathing apparatus, no way to fight the fire at its seat. Sixty-three men and boys died.

The fire burned four days. The pit had to be sealed and flooded to kill it. The last of the dead were not brought up until 16 March 1890 — more than six months after their families watched the smoke come out of the shaft.

The cause of the fire was never determined. And here is the line every safety leader should sit with: the Inspector's report found no evidence of any want of care by the management, and concluded that the smoke and hot air reaching the workings from the return was not anticipated, and could not reasonably have been anticipated.

Not anticipated. That is the verdict. Not careless. Not reckless. Nobody imagined the air could come for them.

What actually failed. Four things, none of them a person.

The ventilation system was built out of fuel — timber lining sitting directly in the airway, in a mine full of coal.

The air path and the escape path were the same physical asset, so one fire took both at once. The first failure was also the last.

There was no detection, no suppression, and no survival gear. Once smoke entered the workings, the outcome was fixed.

And nobody could say with certainty who was underground or where. That is why recovery took six months.

The HOP read. The Inspector handed us the systems lesson in 1889 and we are still relearning it: it was not anticipated. This was not a behavior problem, and no amount of toolbox talks would have moved the outcome an inch. Sixty-three people died because the system had exactly one barrier between normal operations and a fatality, and that barrier was made of wood and doubled as the exit.

HOP thinking says error is not a cause, it is a consequence — and here nobody even had to make an error. The system failed on its own terms. The men had no options left sixty seconds in, which means the decision that killed them was made years earlier, by whoever specified a timber-lined shaft and called one way out sufficient.

Second lesson, quieter: emergency response capability is a design decision made months in advance, not during the event. Mauricewood had no rescue capacity because nobody had built one.

The standards that carry this today (current as of September 5, 2026 — verify current text at ecfr.gov before you build a program on it):

Underground coal falls under MSHA. 30 CFR 75.380 requires two separate escapeways from each working section to the surface — separate, so one event cannot take both. 30 CFR 75.1714 puts a self-contained self-rescuer within reach of every miner. 30 CFR 75.1502 requires a mine emergency evacuation and firefighting program of instruction, because the plan has to exist before the smoke does.

Underground construction falls under 29 CFR 1926.800, which reads like a direct answer to Mauricewood: a check-in/check-out procedure so you always know who is below ground, ventilation requirements, fire prevention and control, and emergency provisions covering self-rescuers and trained rescue teams.

General industry: 29 CFR 1910.36 requires at least two exit routes, located as far apart as practicable. 29 CFR 1910.37 requires them kept unobstructed, lit, and clear of flammable furnishings. 29 CFR 1910.38 requires an emergency action plan that includes accounting for employees after evacuation. 29 CFR 1910.134 covers respiratory protection, including escape-only respirators, and 29 CFR 1910.146 covers permit-required confined spaces — monitoring and rescue.

Read 1910.36 again with Mauricewood in mind. "As far apart as practicable" is not an aesthetic preference. It is the rule that exists because two exits sharing one shaft are one exit wearing a disguise.

Environmental. To kill the fire they sealed the pit and let it flood, then pumped the water back out months later to reach the bodies. Nobody in 1889 asked where that water went. Today the question has a permit number attached: the water you use to put a fire out becomes contaminated firewater, and containing it is its own engineered system. If your emergency plan ends at "extinguish," it is half a plan. Ask where ten thousand gallons goes on your worst day, and whether it reaches a storm drain.

Health. The men at Mauricewood were killed by smoke and afterdamp — carbon monoxide, overwhelmingly. Fire still kills mostly by atmosphere, not by flame, and CO is still the reason. The health story also runs past the body count: seven survivors, a village that spent six months waiting, and recovery crews who went back down repeatedly into a burned mine. We have a word for that exposure now. They did not.

Safety. Covered above — and all three threads trace back to one design decision about one shaft.

1. Maine pulp mill: the bill arrives for a hazard the company already knew about.

On January 27, 2026, a hydrogen sulfide release at the Woodland Pulp mill in Baileyville, Maine killed two people — a 20-year-old co-op chemical engineering student and a 26-year-old chemical engineer — and exposed ten more employees. In a July 14, 2026 investigation update, the CSB reported the mill had no personal H2S monitors, no fixed H2S detectors or alarms in the bleach plant, no building ventilation during normal operations or shutdowns, and no way to track where people were inside the Kraft Mill building. CSB Board Member Sylvia Johnson: "Woodland Pulp was aware of the hazards associated with hydrogen sulfide gas forming in the acid sewer piping, but despite this knowledge, the company did not have adequate systems in place to monitor or mitigate the hazards." OSHA cited the mill more than $700,000 in a July 30, 2026 news release; trade reporting puts the proposed total at $794,456 across four willful and eight serious violations.

So what for safety leaders: the finding that should scare you is not the missing monitors — it is "was aware." Documented knowledge of a hazard with no matching control is how a serious citation becomes a willful one. Pull your last three hazard assessments and ask what physically changed after each. If the answer is "we trained on it," you have Woodland's file.

2. West Virginia: the hazard did not retire when the plant did.

On April 22, 2026, an H2S release at Ames Goldsmith Corporation's Catalyst Refiners facility in Institute, West Virginia killed two workers and seriously injured four. The CSB's August 13, 2026 update found the site had four full-face respirators but not all were fitted with filters effective against hydrogen sulfide; no written procedures for disposing of certain chemicals through the wastewater treatment system; and no personal gas monitors provided or required. Workers believed respirators were unnecessary because production had stopped — the facility was being decommissioned. CSB Chairperson Steve Owens: "This tragic incident underscores the need to have clear procedures and safely manage serious chemical hazards during facility decommissioning."

So what for safety leaders: two things. First, decommissioning is the highest-hazard, lowest-attention phase of any asset's life — the process safety program winds down while the inventory is still in the pipe. Second, and worth a standing meeting item: an air-purifying respirator is not IDLH equipment. Under 29 CFR 1910.134, IDLH atmospheres require supplied air. NIOSH puts the H2S IDLH at 100 ppm and OSHA's limit sits in 29 CFR 1910.1000 Table Z-2 as a ceiling — verify both against current text. If your shutdown plan hands someone a cartridge respirator for an H2S space, your shutdown plan is the hazard.

3. Heat: the federal program lapsed, and California moved anyway.

OSHA's heat National Emphasis Program was extended in January 2025 to run until April 8, 2026 "or until superseded by an updated directive," and as of now there is no public indication it was extended or replaced. The federal Heat Injury and Illness Prevention rule — proposed August 30, 2024, hearing held summer 2025, post-hearing comments closed in fall 2025 — has not advanced, with no target date for final action. Meanwhile Cal/OSHA released a revised draft on August 14, 2026 that aligns indoor and outdoor heat language, adds rhabdomyolysis as a recognized heat illness (including the tell of abnormally dark, tea- or cola-colored urine), and removes the exemption that had limited high-heat procedures to agriculture, construction, landscaping, oil and gas extraction, and certain transportation. Comments are due September 21, 2026.

So what for safety leaders: a lapsed emphasis program changes inspection targeting, not your liability. Section 5(a)(1) of the OSH Act — 29 U.S.C. 654(a)(1) — did not lapse, and heat is a recognized hazard by any reading. If you operate in California outside those five industries, the draft may pull you into high-heat procedures for the first time; you have until September 21 to say something about it. And add rhabdo to your supervisor training regardless of state. Dark urine after a hot shift is not dehydration to push fluids at. It is an emergency room.

Two mechanics are replacing a duct section in a mezzanine penthouse. One door in, one door out — except the second door has had a pallet rack parked in front of it since spring, when the stockroom overflowed and somebody solved that problem the fast way. Hot work permit current. Fire watch posted. Extinguisher at arm's length. Every box on the checklist ticked.

The grinder throws sparks behind the duct run into a pocket of dried-out insulation nobody knew was there. The penthouse fills with smoke in under thirty seconds. Both men make it out the one working door. Singed sleeves, nobody hurt.

The HOP read: nobody in this story did anything wrong on the day. Every control the system asked for was in place and working. The decision that nearly killed two people was made five months earlier by someone solving a storage problem, who had no idea they were also deleting an exit.

So the question is not "why did they do hot work near insulation." It is: who in your building is authorized to make a change that removes a barrier, and who reviews it before it sticks? A rack in front of a door is a management-of-change event. It just does not look like one, because the person stacking the pallets was not thinking about fire. Mauricewood in miniature — and nobody anticipating it.

Got a fail or a near-miss? Hit reply, or scan the QR below — we'll feature it anonymously — no names, no company, no blame. Just the lesson, so the next crew doesn't learn it the hard way.

Pick one occupied space today — a control room, a mezzanine, a pump house, a vault, a walk-in — and answer three questions on your feet.

One: are there two ways out that do not share the same corridor, stairwell, or shaft? Two: does either exit route pass through or alongside the hazard the space contains? Three: right now, without picking up a radio, can you say who is inside?

If any of the three comes back wrong, you have found today's action item, and it outranks whatever was already on your list. Mauricewood was one shaft, one exit, and no headcount. Sixty-three men.

Watch: youtube.com/@GetSafetyIntel

Share your near-miss, anonymously: scantoconfess.com — or scan the QR above. No names, no company, no blame.

Forward this to one person on your crew. The lesson only works if it travels.

Please stay Safe & Hydrated!!!