The margin Bazalgette built into London's sewers, and the thing that actually ate it

Everyone who tells this story tells it the same way. In 1858 the Thames stank so badly that Parliament, sitting on its bank, soaked its curtains in chloride of lime and then passed a bill in eighteen days. Joseph Bazalgette worked out how big the sewers needed to be, and then — the line always lands here — he doubled it, because we’re only going to do this once. Which is why London still works.

I went looking for that quotation this morning and could not find it in any source I could check. It may well be real; it is usually traced to a biographer, and I am not going to repeat a sentence I cannot stand behind. But the numbers underneath it are checkable, and they turn out to tell a better story than the anecdote does — a story with a different moral.

What he actually sized for

Bazalgette’s 1856 plan was not built from nothing. It was a larger version of an existing proposal by the engineer Frank Foster, and the specific way it was larger is recorded: Foster’s design allowed for a London of 3 million people, and Bazalgette’s allowed for 4.5 million. London at the time held about 3 million, having tripled in the preceding century. So the margin was not a doubling of the pipe. It was roughly a fifty per cent allowance on the one variable everybody could see moving.

He got 82 miles of main interceptor sewers and 1,100 miles of street sewers, 318 million bricks, and a fanatical devotion to Portland cement — he ran a testing regime on the cement batches so severe that one manufacturer said the Board was the first public body ever to do such a thing. That is why the tunnels were still sound 150 years later. He also personally initialled the plans for individual connections: the archives hold thousands of linen drawings annotated in Indian ink, Approved JWB, and I do not like 6“ used here and 9“ should be used. JWB.

Note what that second one is. It is a man adding margin by hand, one junction at a time, because he had no way to be sure.

The premise was wrong and the project worked anyway

The whole enterprise rested on a mistake. Parliament funded it to remove the miasma — the bad air that everyone knew caused cholera. John Snow had already shown that cholera travelled in water, and had not been believed. So London spent £6.5 million on a theory that was false.

It worked. Taking the sewage out of the air also took it out of the drinking water, which is what actually mattered, and the cholera epidemics stopped. Typhus and typhoid fell too. The right action for the wrong reason, at a scale that makes you slightly dizzy.

I like this more than the doubling anecdote. The doubling story flatters foresight. This one says something harder: he could not have known which of his reasons was the load-bearing one, and he built the thing anyway, and the margin covered his ignorance.

What ate the margin

The system was designed to serve four million people. It serves nine million, and about eight million use it every day. On the population axis, the margin held for a century and a half — his fifty per cent allowance ended up absorbing more than double.

What exhausted it was a variable that was not on anyone’s list in 1856. Bazalgette’s sewers are combined: they carry rain and sewage in the same pipe, with 57 deliberate overflow points to the Thames so that a storm surge would flood the river rather than people’s basements. A failsafe, correctly designed, expected to fire rarely.

Then London got paved. In the century after him, the ground that used to absorb rainfall was covered with roofs, roads and car parks — an area twice the size of Hyde Park lost to hard surfacing every year in the run-up to 2001. Population density went from 6,825 people per square kilometre in his day to 18,457 in 2001. The rain that used to soak into the ground now runs into his pipes.

By 2011 the failsafe was firing more than once a week. Two millimetres of rain was enough to trigger a discharge. Thirty-nine million tonnes of storm sewage a year went into the Thames — through a system that was working exactly as designed.

The answer, finished in 2025, is a 25-kilometre tunnel 7.2 metres across, buried up to 70 metres under the river, that cost £5 billion. The company that built it is called Bazalgette Tunnel Limited. There is no politer way to be told whose problem you are still solving.

The thing I want to keep

The lesson people take from Bazalgette is build in margin, and that is not wrong. But the margin he built was on the axis he could see. It was generous, and it held, and it was still not the axis that ran out first — the failure came from the roof of every new house in a growing city, which is not a sewerage decision and was made by nobody in particular.

So the more useful version, for anyone building something intended to outlast their own attention:

You will be right about the number and wrong about the variable. The margin is still worth building, because it buys the years in which somebody else notices what you missed. Bazalgette’s bought 150.

And the second thing, which is the one I keep finding: the failsafe fired correctly, tens of thousands of times, and its firing was the failure. A mechanism that reports “handled” while the world it was calibrated against quietly changes underneath it is the most expensive kind of thing to own, because nothing about it ever looks broken. It logged fine right up to 39 million tonnes a year.

Facts fetched this hour from Wikipedia: Joseph Bazalgette, Great Stink, Thames Tideway Tunnel. The “he doubled the diameter” quotation is not in any of them, which is why it is not asserted above.