In the first ten days of September 1854, more than five hundred people died of cholera within a few streets of each other in Soho, London. A local physician named John Snow marked each death on a street map and saw that the deaths formed a ring around a single water pump on Broad Street.
The map is now reproduced in every textbook on data visualisation and epidemiology. The story that goes with it, in which a clever doctor draws a diagram, removes the pump handle, and stops the outbreak, is largely a simplification. The real work was longer, harder, and considerably more impressive.
What Everyone Believed
Cholera arrived in Britain in 1831 and killed tens of thousands in a series of epidemics. The accepted explanation was miasma: disease is caused by foul air arising from decaying matter, and outbreaks concentrate in poor districts because those districts smell worse.
The theory was not stupid. It correlates with the evidence, since crowded, filthy areas did suffer most, and it produced useful public health policy, because clearing waste and improving ventilation genuinely helped. It was held by nearly every serious authority, including Edwin Chadwick, who ran sanitary reform in England, and William Farr, the statistician at the General Register Office.
It also produced a catastrophic error. Under Chadwick’s reforms, London’s cesspits were emptied into the sewers, and the sewers discharged into the Thames, from which several companies drew drinking water. Removing the smell moved the infection.
Snow’s Alternative
John Snow was a Yorkshire surgeon’s apprentice turned London physician, and by the 1850s he was the country’s leading authority on anaesthesia. He administered chloroform to Queen Victoria for the birth of Prince Leopold in 1853, which did a great deal to make the practice respectable.
His interest in cholera dated from the 1831 epidemic, which he had seen as a young apprentice in a colliery village. In 1849 he published a pamphlet arguing that cholera is not airborne but is transmitted by something swallowed.
His reasoning was clinical rather than statistical. Cholera begins with violent intestinal symptoms, not respiratory ones, which is what you would expect from a poison entering by the mouth rather than the lungs. It spreads among people who handle the sick and their bedding. And a disease carried in the air should affect everyone breathing that air, which it plainly did not.
The pamphlet was largely ignored. It offered no mechanism anyone could see, since the organism responsible would not be identified for another thirty five years.
Broad Street
The outbreak that began on 31 August 1854 was, in Snow’s words, the most terrible that ever occurred in this kingdom. He lived nearby and began investigating within days.
He obtained the death records from the General Register Office and visited the addresses, asking where the household had drawn its water. He then plotted the deaths on a map of the district as stacked bars at each address, one mark per death.
The concentration around the Broad Street pump was obvious, but proximity alone proves little, since the pump sat in the middle of the worst affected streets by construction. What made the argument work was Snow’s treatment of the exceptions.
- The workhouse on Poland Street stood inside the affected area and held more than five hundred inmates. Only five died. It had its own well and its own supply, and its residents rarely used the street pump
- The brewery on Broad Street employed around seventy men, none of whom contracted cholera. The proprietor told Snow that the men drank the malt liquor they produced, and that the brewery had a deep well of its own
- The widow in Hampstead is the decisive case. She had lived in Broad Street years earlier and so liked the taste of the water that she had a bottle brought to her by cart regularly. She drank from it on 31 August, and died of cholera two days later, several miles from the outbreak, in a district with no other cases. Her niece, visiting, drank the same water and also died
A woman who never went near the district but drank its water died. Hundreds of men who lived in the district but drank beer did not. No airborne explanation accommodates both facts.
Snow also drew, on a later version of the map, a boundary line marking the area within which the Broad Street pump was the nearest by walking distance rather than straight line distance. It is an early example of what would now be called a Voronoi diagram, and it shows the deaths falling inside that boundary and stopping at it.
The Pump Handle
On the evening of 7 September, Snow presented his findings to the Board of Guardians of St James’s parish. They were sceptical and removed the pump handle the following day.
The outbreak ended shortly afterwards, and the two events are usually reported as cause and effect. Snow himself was careful about this, noting that the epidemic had already been declining, partly because a large proportion of the population had fled the district. He did not claim that removing the handle stopped the outbreak, only that it prevented a possible resurgence.
A later investigation by the Reverend Henry Whitehead, initially a sceptic who set out to disprove Snow and was converted by the evidence, identified the probable source. A baby at 40 Broad Street had fallen ill with cholera in late August, and the nappies had been washed and the water emptied into a cesspit less than a metre from the well, whose brick lining had decayed.
The Better Experiment
The map is famous. Snow’s more rigorous work is less well known and is the stronger piece of science.
South London was supplied by two water companies whose pipes ran along the same streets and served neighbouring houses, sometimes on opposite sides of a single road. In 1852 the Lambeth company had moved its intake upstream to Thames Ditton, above the sewage outfalls. The Southwark and Vauxhall company continued to draw from the tidal river in central London.
Snow recognised that this was an experiment already in progress, involving more than three hundred thousand people of every social class, mixed together with no factor distinguishing them except which pipe entered their house. Householders often did not know which company supplied them, so he developed a chemical test based on the higher salt content of the tidal water to identify the source.
The mortality rate among Southwark and Vauxhall customers was roughly eight to nine times that among Lambeth customers.
This is a cohort study in the modern sense: a large population, naturally divided, with a single differing exposure and a measured outcome. Snow published it in 1855 in the expanded second edition of his book. It is the foundation document of epidemiology.
Acceptance
He did not persuade the establishment. The official report on the 1854 outbreak concluded that the evidence did not support the water theory. Snow died of a stroke in 1858 at the age of forty five, with the miasma theory still dominant.
What changed matters was the Great Stink of the same year, when the smell of the Thames drove Parliament to fund Joseph Bazalgette’s sewer system, built for miasmatic reasons and effective for Snow’s reasons. William Farr, Snow’s most capable opponent, examined the 1866 outbreak with Snow’s methods and publicly accepted the water theory.
The organism was seen under a microscope by Filippo Pacini in Florence in 1854, the same year as Broad Street, in a paper nobody read. Robert Koch isolated Vibrio cholerae independently in 1884 and established it as the cause, by which time Pasteur’s work on microorganisms had made a germ theory of disease intellectually available in a way it had not been in 1854.
The Map as Argument
The Broad Street map is treated as the origin point of modern data visualisation, alongside the statistical graphics that Florence Nightingale was producing in the same decade and the isothermal maps Humboldt had introduced to show climate patterns.
What all three have in common is that the diagram is doing analytical work rather than presenting a conclusion reached elsewhere. A table of five hundred addresses conveys nothing. The same addresses plotted in space reveal a structure that no amount of staring at the list would produce, and the structure points at a cause.
That is also its limitation, and Snow understood it. A cluster around a pump is consistent with the pump being the source and equally consistent with the pump simply being in the middle of a bad neighbourhood. The map generated the hypothesis. The workhouse, the brewery, the Hampstead widow and the three hundred thousand water customers tested it.
What It Established
Snow demonstrated the cause of a disease, the route by which it travels, and the intervention that stops it, without any knowledge of the organism responsible, without a microscope, and against a professional consensus that held to the end of his life.
The method he used, gather the individual cases, map them, look for the exceptions that break the pattern, then find a natural experiment large enough to settle it, is what epidemiology has done ever since. A pump in Soho with its handle removed is a reasonable monument, but the argument that mattered was made with a chemical salt test and a list of water bills.
No products in the cart. 