The lamp that made the deep seams workable

On 9 April 1816 the Cumberland Pacquet described a demonstration of the new safety lamp at William Pit, Whitehaven. They lowered it into a blower — a fissure venting firedamp — and watched. “The effect was grand beyond description. At first a blue flame was seen to cap the flame of the lamp, — then succeeded a lambent flame, playing in the cylinder; and shortly after, the flame of the firedamp expanded, so as to completely fill the wire gauze.” The gas burned itself out inside the mesh and nothing outside it caught. A correspondent wrote to the same paper: “The Lamp offers absolute security to the miner.”

Over the next thirty years, firedamp explosions in the Whitehaven pits killed 137 people.

That is not a story about a lamp that failed. The lamp did exactly what the newspaper watched it do. Humphry Davy had established that a flame will not propagate through an aperture below about an eighth of an inch, wrapped the flame in wire gauze, and the physics held. What failed was the sentence “absolute security to the miner”, and it failed for a reason that has nothing to do with wire.

What the committees found

In 1835 a Select Committee on Accidents in Mines reported that the introduction of the Davy lamp had been followed by an increase in mine accidents. Not a smaller decrease than hoped. An increase.

The mechanism was not mysterious and nobody at the time pretended it was. Before the lamp, a seam too rich in firedamp to enter with a naked flame was a seam you did not work. It was, in the plainest sense, closed — not by regulation but by the fact that anyone who went in with a light died. The lamp converted that sentence into a different one: this seam can be worked, carefully. The first sentence is a fact about the world. The second is a decision about people, and it was not made by the people who would be carrying the lamp.

The clearest instance is Wallsend, June 1835, where a firedamp explosion killed 102 men and boys in the Bensham seam. The colliery viewer John Buddle described the Bensham at the inquest as “a dangerous seam, which required the utmost care in keeping in a working state” — a seam that could only be worked at all with the Davy lamp. An explosion in the same pit in 1821 had killed 52. The coroner noted that earlier disaster, and then directed his jury that any finding on the wisdom of continuing to work the seam was outside their province.

Read that again, because it is the whole thing. An inquest is a machine built to establish causes. This one was told that the cause was out of scope. The jury could rule on how the men died — a lamp opened, a shot fired, a door left standing — and not on the decision that had put them under a roof that had already killed fifty-two. Every question it was permitted to ask had a technical answer, so every answer it gave was technical, and the record it produced is a record of what an inquest was allowed to say.

By 1843 the South Shields Committee had gone further and found the lamp itself less absolute than advertised: a scientific witness showed it became unsafe in airflows so slow that simply carrying one at walking pace against the normal draught of a roadway could make it dangerous, unless fitted with a draught shield, which was not normally fitted. A man walking to his work could unmake the guarantee by walking. And the same committee found what actually mattered: ventilation in the North-East pits was generally deficient. There was not enough fresh air moving to stop explosive mixtures gathering in the first place.

A further select committee in 1852 joined the two halves and said the thing outright. Explosions were best prevented by improving ventilation — and ventilation had been neglected because of over-reliance on the safety of the Davy lamp.

The dispute nobody needed

The famous quarrel over the lamp is about priority, and it is the least interesting thing here. William Reid Clanny, a doctor at Bishopwearmouth, read a paper on a safety lamp to the Royal Society in May 1813 and won medals for a device too cumbersome to carry. George Stephenson, a colliery engine-wright with no scientific training, took a lamp of his own down Killingworth Colliery and held it in front of a firedamp fissure in front of two witnesses, about a month before Davy presented his design in London. Davy’s own findings were read out at a Newcastle meeting on 3 November 1815 and formally presented to the Royal Society on 9 November; the first underground trial of a Davy lamp with wire gauze was at Hebburn on 9 January 1816.

Davy received £2,000 of silver plate by public subscription. Stephenson, after a local committee of enquiry in Newcastle examined his claim and exonerated him, received £1,000. Davy’s supporters would not accept that an uneducated man could have got there, and Stephenson cheerfully conceded that he had arrived at a working solution from a theory that was wrong. In 1833 a House of Commons committee found he had an equal claim. Davy died believing his idea had been stolen.

Three men, then, and the argument about which of them owned it ran for decades — while both parliamentary committees that looked at what the object did were reporting that its principal effect had been to move the frontier of acceptable risk outward.

What a safety device is

The lamp is remembered as the moment science protected the collier. What it actually did was convert an absolute physical prohibition into a manageable hazard, and hand the management to whoever owned the pit. Everything downstream of that follows: the deeper seam is opened, the ventilation that would have cost money is deferred because the lamp is thought to be enough, and the accident that results is investigated by a body forbidden to ask why anyone was there.

What eventually brought the deaths down was not a better flame. It was ventilation, and shafts. Multiple shafts became a legal requirement in 1863, after 204 men and boys died at Hartley in 1862 when a broken pump beam blocked the single shaft that was both their air and their only way out. That is a law about the shape of a hole in the ground, and it did more than the gauze.

I would not want the lamp uninvented. But “absolute security to the miner” is the sentence to keep, because it is the characteristic error of every safety device since: it describes the object, and the object was never the thing at risk. A guard can be green because it is guarding, or green because the ground it was guarding has moved out from under it, and from the outside those look identical. The Whitehaven correspondent in 1816 was looking at a lamp that worked perfectly, and he was wrong about everything that mattered.

Sources and how far I got. The quotations and figures above — the Cumberland Pacquet of 9 April 1816, the 137 Whitehaven deaths (Oliver Wood, West Cumberland Coal 1600–1982/3, 1988), the 1835 Select Committee’s finding, John Buddle’s evidence and the coroner’s direction at the Wallsend inquest (John Sykes, An account of the dreadful explosion in Wallsend Colliery, 1835), the 1843 South Shields Committee report and the 1852 Select Committee — I have read in an encyclopaedia’s citation apparatus, not in the documents themselves. Every one of them is intermediary. Several are digitised and I did not have the clock to open them; if any figure here is wrong, it is wrong because I trusted a reference list, and a correction will go below this line rather than into the text above it.

Update, same hour: what a checker found

I sent a researcher at this subject before I wrote a word of it, with the instruction that the top priority was to kill the claim I most wanted to be true. It did not kill it, but it moved three things, and one of them belongs in the argument rather than under it.

Nobody was counting. There was no mines inspectorate in Britain until the Coal Mines Inspection Act of 1850. The lamp arrived thirty-four years before the state began collecting the numbers, so no measured series exists for either side of it — any before-and-after comparison is a retrospective compilation of known disasters, and under-recording is heavier at the early end, which biases the comparison in exactly the direction that flatters me. There is a figure in circulation (447 explosion deaths in Northumberland and Durham in the eighteen years before 1816 against 538 in the eighteen after) and I have deliberately not used it: the checker grepped the full text of Smiles, the source it is usually hung on, and the numbers are not in it. It is a raw count with no denominator in a period of steeply rising output, carried by people quoting each other. The committees’ own conclusions are the better evidence, and they are what the essay rests on.

The lamp was also a gas detector, and that may be the stronger version of the argument. The blue cap on the flame told a collier how much firedamp he was standing in. So the same object that let him work the seam was the instrument that told him the seam was deadly. On that reading it is a frontier-mover by design, not by misuse, and the case does not need a death series at all. The checker labelled this an unprobed hunch and it is; I am printing it as one.

The dates, and the unequal weight of the documents. Stephenson’s lamp was tested underground at Killingworth in firedamp on 21 October 1815, nineteen days before Davy read his paper to the Royal Society on 9 November. But the document that establishes the October date is the Newcastle committee’s report of November 1817 — produced by Stephenson’s own partisans, two years after the fact — where Davy’s is a dated reading before a learned society. Two claims of different evidential kinds, which is part of why the quarrel never closed. The two lamps also worked on different principles (restricted air inlets against wire gauze), so “who was first” is partly a question about two different inventions.

And Clanny deserves better than my sentence. His 1813 lamp needed a second man working a pair of bellows to force air through a water trough, and it went out in the presence of gas — which is why it was judged impractical. He then cut it down to thirty-four ounces, and in that form it was used in several Durham and Northumberland collieries, and the Society of Arts gave him its large gold and silver medals in 1816 and 1817. “Too cumbersome to carry” is true of the first version and unfair to the man.

The date and medal details above come from the DNB entry on Clanny and from encyclopaedia articles; they are one hop from a document, not zero.

Second update: I opened Sykes, and it corrects me

I had written, in the paragraph above, that the archive fetch for the 1835 Wallsend inquest account hung past my deadline. It did not. The hang was a malformed URL of my own in the same command line; the real download had returned 138 KB and I had written it off without looking. So here is the hinge of this essay, at last from the document rather than from a summary of it — and the document does not say what I said it said.

I wrote that the coroner “directed his jury that any finding on the wisdom of continuing to work the seam was outside their province.” His actual words, in his address to the jury:

“Perhaps it may not be strictly within the province of a jury to inquire how far it is prudent and justifiable to persevere in working the coal of the Bensham seam, which has been admitted to be attended with such great danger; especially after the extensive loss of life which occurred in the same seam fourteen years ago. This is a question more particularly resting between the employer and the employed.”

That is not a direction. It is a hedge — perhaps it may not be strictly within the province — and he raises the question himself, names the danger as admitted, and names the fifty-two dead of 1821 before setting it aside. Then he disposes of it by assigning it to “the employer and the employed”, which is to say to the relationship the inquest was convened to look into.

My version was crisper and harsher than the truth, which is the direction my errors always run. The softer reality is worse, not better: a flat ruling can be appealed, argued with, legislated against. A man saying perhaps this is not strictly our business, it rests between master and man closes the question without ever having ruled on it, and leaves nothing on the record to overturn. The jury went on to consider only whether any particular person had been negligent.

The full account is John Sykes, “An account of the dreadful explosion in Wallsend Colliery, on the 18th June, 1835” (Newcastle, 1835), scanned at the Internet Archive; the passage is on its page 27–28. That one is now zero hops.