Click words in text to see background information or location in 1859 map of London.  Then click in browser to return.

Snow's investigation

The officials of St James's parish (where in Golden Square is located) were aware that their locality had suffered a cholera outbreak of unusual intensity, about 500 deaths having occurred in the 10 days from Aug 31 to Sept 9, 1854, in an area of only a few square blocks [1,2].  In November of that year, they appointed a cholera inquiry committee to examine the event, and asked Snow to join them. One of his tasks was to write a report on the local water supply, which he submitted on Dec 12, 1854 [4].  About a month later, Snow published a detailed monograph on the mode of communication of cholera, which included a discussion of the Broad Street outbreak [5].  Internal evidence suggests that Snow completed the account in his monograph before he wrote his report to the parish committee. In any event, the two accounts of his methods parallel are very similar, and accord with his brief, earlier account, published in Sept 23, 1854 [6].

Snow had observed cholera first-hand in 1831 as an apprentice surgeon-apothecary, but he seems not to have questioned prevailing theories of cholera pathology and transmission until 1848-49, when he formulated a theory quite at odds with the then-dominant views. Snow argued that cholera was a localized disease of the gut, and that its symptoms were entirely the result of fluid loss [7]. The causal agent, he reasoned, must enter by mouth, multiply within the gut, and then spread to others by the fecal-oral route. In his 1855 monograph he suggested that the structure of the unknown agent was that of a cell but, in the absence of microscopic evidence, he avoided speculation on its exact nature. While direct contact with contaminated bedclothes, for example, could explain cholera's spread within a household and other confined spaces, Snow deduced that the transmission of the disease across greater distances was due to drinking water contaminated with raw sewage containing the specific cholera poison. Snow described two local outbreaks of cholera in south London in 1849 that seemed to suggest strongly a water-borne route [7].

When epidemic cholera next arrived in England in 1853-54, Snow realized that he had an opportunity to test his hypothesis on a grand scale. Two competing water companies had laid pipes up the same streets in several south London districts. In the 1848-49 epidemic, both companies had drawn upon water from the River Thames that had been contaminated by sewage. By 1853, one company (Southwark and Vauxhall) still supplied contaminated water, while the other company (Lambeth) had moved its intake source upstream to a fairly clean section of the river. When Snow learned of his change in the water supply, he realized that this arrangement constituted a superb natural experiment.  He undertook door-to-door investigations to relate cholera mortality to the source of water and found greater mortality among those drinking the water supplied by the Southwark and Vauxhall company, and made these data the centerpiece of the 1855 edition of his monograph [5]. It was this investigation in south London that Snow suspended for several days at the beginning of September to investigate the severe localized outbreak near Golden Square.

Because he already had a hypothesis, Snow immediately investigated the water supply. The piped water was supplied by a company that drew its water from an unpolluted section of the Thames; and he knew from his own experience of living in the area that most residents preferred pump water. Moreover, his study of earlier outbreaks led him to conclude that a sharp localized outbreak pointed to a contaminated pump or well rather than a problem with the general water supply [6].

He turned his attention to the five pumps near Golden Square, finding that most of them contained impurities visible to the naked eye. By contrast, when Snow inspected the Broad Street pump water, on Sept 3, it looked clear; but a local resident reported that its water had smelled offensive just the day before. On Sept 5, Snow obtained from the General Register Office a list of the 83 deaths ascribed to cholera in the Golden Square area since Aug 31. In his own words:

On proceeding to the spot, I found that nearly all the deaths had taken place within a short distance of the [Broad Street] pump. There were only ten deaths in houses situated decidedly nearer to another street-pump. In five of these cases the families of the decreased persons informed me that they always sent to the pump in Broad Street, as they preferred the water to that of the pumps which were nearer. In three other cases, the deceased were children who went to school near the pump in Broad Street ...

With regard to the deaths occurring in the locality belonging to the pump, there were 61 instances in which I was informed that the deceased persons used to drink the pump water from Broad Street, either constantly or occasionally ...

The result of the inquiry, then, is, that there was been no particular outbreak or prevalence of cholera in this part of London except among the persons who were in the habit of drinking the water of the above-mentioned pump well.

I had an interview with the Board of Guardians of St James's parish, on the evening of the 7th inst [Sept 7], and represented the above circumstances to them. In consequence of what I said, the handle of the pump was removed on the following day [6].

Thus, Snow's initial suspicion of the Broad Street pump was a deduction based on his intensive earlier study of similar cholera outbreaks, not an induction arrived at primarily from the geographical facts of the case. While Snow was clearly thinking in topographical terms (visualizing the location of houses where deaths occurred in relation to the placement of pumps), in the case that he presented to the parish officials he neither said nor alluded to the idea that a map had been instrumental in the discovery of the source of the outbreak.

Snow first exhibited a cholera spot map (with small black bars indicating deaths) of the Broad Street area at a meeting of the Epidemiological Society of London on Dec 4, 1854, nearly 3 months after his first investigation [7]. He had in the interim made two more local inquiries in the Broad Street area, though one was hampered by the large number of local residents who had fled the region. By December, Snow had gathered data on 616 deaths, but because he did not know the exact address of some residents of St James's parish who had died outside the district, his map had only 574 bars.

Snow may have realized that a spot map would be a useful illustration for his report to the parish committee and for his own book. The first edition of On the mode of communication of cholera, published in 1849, contained no maps and only one table [8].  By 1854 Snow had seen the excellent map in Shapter's work on the cholera in Exeter, which Shapter included as a frontispiece but hardly discussed in his text [9].  Shapter's book, which Snow cited in the second edition of his own work, may have persuaded Snow of the value of a map as an illustration.

Snow published two slightly different versions of his map, of which the map published in his cholera monograph appears to be the earlier [5]. By the time the parish report was published 6 months later, Snow had made a few revisions. The most important was the addition of a dotted line enclosing the area that was closer to the Broad Street pump (in walking distance) than to any other street pump (see figure 1). Such an equidistance line dividing a map into regions is called a Voronoi diagram, and Snow's second map has been credited with being the earliest use of this device [10].  The map with the Voronoi diagram most closely mirrors Snow's narrative account of his investigation on the spot; and the text of his report to the parish committee is altered from the version that appeared in the monograph to refer to this new feature of the map [4].  This variation between Snow's two maps lends further support to the hypothesis that he compiled them for illustrative rather than investigative purposes [11].  As a result of his two later investigations on the spot, Snow had amassed even more data to show that most people who died of cholera, but who lived outside the Voronoi area, had nonetheless drunk water from the Broad Street pump [4].

Besides the Voronoi diagram, Snow made an apparently small change that later proved important. In his first map, Snow had located the Broad Street pump in the wrong place, in front of the Newcastle Arms, the public house on the corner (still there in what is now Broadwick Street but today named after John Snow) rather than adjacent to the house next door at number 40. If Snow ever saw the more accurate government maps that we will describe below, he would have discovered that the spot at which he showed the pump was actually occupied by a sewer grating. When he revised his map, he moved the pump 24 feet to the west, to its correct location.

Snow also modified the map in the parish report in several minor ways. He altered the political boundary lines to better show the area within which deaths had occurred, and added subdistrict division lines (see figure 1). He included one distant pump that had not been shown on his first map, raising the total number of pumps depicted from 13 to 14. He missed a more distant pump, in Nag's Head Yard, that was shown on the other maps we will discuss below.

Because Snow was preoccupied with his more important south London study, his initial investigation in Broad Street was brief. Another member of the parish inquiry committee, Rev Henry Whitehead, eventually discovered the probable cause of the contamination in the water of the pump. In April, 1855, Whitehead learned that a child at 40 Broad Street had become ill with "diarrhea" at the end of August, and died on Sept 2. Her mother had washed the soiled nappies and disposed of the water through the house drains -- drains that were later shown to run a few feet from the well of the pump. An earlier excavation of the Broad Street pump well, in the autumn of 1854, had found the brick lining apparently intact. But, in late April, 1855, a more careful excavation showed that both the house drain and the pump well had decayed brickwork that permitted percolation of fluid from the former to the latter [1, 4]. Snow, unaware of the death of this child, drew only four bars at number 40. Later maps incorporated Whitehead's data and showed five bars (see figure 2).