17 May

John Snow’s On the Mode of Communication of Cholera was published as ‘a slender pamphlet’ in August 1849, during a major cholera epidemic. It was republished, ‘with much new matter’, early in 1855, now four times its original length. The pamphlet of 1849 outlined a hypothesis; the book of 1855 demonstrated that hypo- thesis to be true. Reading it now, it is a breathtaking, an astonishing performance. Yet Snow’s arguments were rejected by the leaders of the medical profession, and when, a decade or so later, they came to be generally adopted, they were accepted for the wrong reasons. The response to Snow was a test of his contemporaries, a test they failed. Snow was the first doctor in the long history of medicine to understand his enemy. He worked out how to prevent cholera, but he did not discover a cure for it. He gave an accurate account of the causal agent, but he never saw it through a microscope. (The first to do so was Filippo Pacini in 1854, though his discovery was largely ignored at the time, and Snow never heard of it.) Snow conducted no experiments, and therefore it is not at first apparent that he was a virtuoso exponent of the experimental method and that the streets of London were his laboratory. Cholera was a disease unknown to Western Europeans until the late eighteenth century. It began to appear outside Asia, its likely continent of origin, in 1817, and first reached England in 1831, when it killed around 23,000 people. A second outbreak in 1849 killed 53,000. A third, in 1853–4, killed 23,000. A fourth and last, in 1866, killed 14,000. Similar outbreaks occurred across Europe until the end of the nineteenth century. Snow’s account of the cause of cholera originated during the second English epidemic and was tested during the third; the adoption of arguments similar to his limited the impact of the fourth epidemic and prevented future epidemics –– there were only 135 deaths in England in 1893, when cholera was ravaging much of Europe. The conventional account of cholera in the 1840s and 1850s held that it was spread as a poison through the air. Like all epidemic diseases, cholera was held to be caused by a miasma, by bad air. (Malaria means ‘bad air’. Malaria was only shown to be caused by a parasite transmitted by mosquitoes in 1897, so that it remained a disease of bad air long after other diseases, such as cholera, had been shown to be caused by germs.) The ultimate source of bad air was rotting organic matter, and the best way of preventing it therefore was to eliminate the sources of foul odours by improving sewers and drains. Nineteenth-century English reformers such as Edwin Chadwick thus recommended exactly the same improvements that had been urged by their sixteenth-century Italian predecessors. The key difference was that where Renaissance doctors believed that epi- demics could under certain circumstances be transmitted directly from one person to another, and so advocated the quarantining of those infected, their nineteenth-century successors held that quaran- tine measures were pointless, and so allowed the free movement of people and goods even during epidemics: as Snow himself remarked, there were ‘great pecuniary interests’ that would have been damaged by any recourse to the precautions that had been adopted against the plague. Snow’s starting hypothesis was very different from those of his contemporaries. Because he was a pioneer in the use of anaesthetics (in 1853 he was chosen to administer chloroform to Queen Victoria during the delivery of Prince Leopold), Snow had a great deal of experience in what happened to people and animals when a poison entered the bloodstream as a result of inhalation. (He frequently tried out possible new anaesthetics, first on animals, and then, if the results were promising, on himself.) This experience provided no analogues to cholera. Moreover poisonous gases affected everyone exposed to them, while cholera struck some and passed others by. And poisonous gases obeyed the law of the diffusion of gases:


27. This drawing by George John Pinwell, entitled Death’s Dispensary, published in an English magazine during the cholera epidemic of 1866, marks the belated triumph of John Snow’s account of the mode of transmission of cholera. As the gases given off by putrefying substances become diffused in the air, the quantity in a given space is inversely as the square of the distance from their source. Thus a man working with his face one yard from offensive substances would breathe ten thousand times as much of the gas given off as a person living a hundred yards from the spot. Yet miasmatic theories claimed to be compatible with the fact that people at a considerable distance from the source of a smell often fell ill when those closer to the source were unaffected. Cholera attacks the gut, causing violent diarrhoea, so Snow concluded it most likely had its origin in something ingested, and this ingested substance probably came more or less directly from cholera sufferers (as smallpox came from smallpox sufferers). Snow was a vegetarian and a teetotaller, and this prepared him for his line of thinking. He himself drank distilled water because he was well aware that one could often find in tap water material that had passed through the human gut. Tap water was not, strictly speaking, vegetarian, since it commonly contained microscopic amounts of half-digested meat. At the age of 17 Snow had read John Frank Newton’s The Return to Nature: A Defence of the Vegetable Regime (1811), in which the water drunk by Londoners was shown to be full of ‘septic matter’. As an anaesthetist, Snow could make no sense of the miasmatic theory’s account of how poisonous gases operate; as an admirer of Frank Newton, Snow was predisposed to think that the water people drank made them ill, but he did not develop his alternative theory until the last months of 1848. In the first edition of The Mode of Communication of Cholera, Snow recounts the case of John Barnes, a labourer in Moor Monkton near Yorkshire, a village untouched by cholera. Barnes’s sister had died of cholera in Leeds; two weeks later her clothes were bundled up and sent to Moor Monkton by the carrier. ‘The clothes had not been washed; Barnes had opened the box in the evening; on the next day he had fallen sick of the disease.’ From John Barnes, the communica- tion of the disease could be traced through twenty individuals (with only one unexplained link), and thirteen deaths. Clearly the disease had travelled in the box of clothes. Snow had to hand other accounts of the propagation of cholera from person to person. These cases were not compatible with the modification of the miasmatic theory that some advocated, that chol- era, originally caused by rotting matter, was also caused by ‘effluvia given off from the patient into the surrounding air, and inhaled by others into the lungs’. Snow’s alternative hypothesis was that cholera was spread through excreta. The first mode of transmission he identi- fied was from hand to mouth: John Barnes had touched his sister’s soiled clothes, had failed to wash his hands, and had conveyed the source of the disease to his own gut. By similar means, his sickness had been conveyed to those who visited him and cared for him. It was easy to show that the disease had an incubation period of twenty-four to forty-eight hours. From the beginning, Snow’s argument implied that cholera was an ‘animalcule’ –– although he insisted that all his argument required was that it should be an ‘organized particle’, one ‘capable of multiplying in the human body’. Snow was exploring a germ theory of cholera at a time when germ theories were generally rejected in favour of chemical theories which attributed diseases to poisons (or, to use the Latin term, viruses). He did not want to dismiss out of hand the possibility of some sort of chemical account of the cholera poison, comparable to contemporary accounts of fermenta- tion, which maintained that it was an inorganic process, but his whole argument implied a germ theory, and by 1853 he had committed himself to the view that diseases are caused by living agents. Snow identified a second mode of transmission, which did not require direct contact with the patient or their personal effects. He looked closely at two identical alleys of houses in Horsleydown, near London, that stood next to each other. Cholera had struck one alley- way, called Surrey Buildings, killing eleven people, and had almost entirely spared the other, killing only one. In the alleyway where the infection had spread overflow from the drains ran back into the well from which the inhabitants drew their water. The soiled clothes of the sick had been washed in water that their neighbours had then drunk. Cholera had spread through the pollution of the water supply. In another case, a London suburban development called Albion Terrace, Snow had identified a row of seventeen houses where a severe rainstorm had caused the cesspools to overflow into the water supply. Here twenty-four people had died. In a neighbouring house, supplied with the same water, a gentleman who had always refused to drink the water was untouched. In 1849, Snow had five other examples of disease hotspots that could only be explained by the hypothesis that cholera was entering the drinking water. One was of a landlord who had dismissed his tenants’ complaints that their water stank. Cholera was frequent amongst the tenants, but not in the dis- tant village where the landlord lived. One Wednesday he drank a glass of his tenants’ water to show there was nothing wrong with it; he died the following Saturday. The official account of the deaths at Albion Terrace blamed an open sewer 400 feet away, which caused an unpleasant smell when the wind was in the wrong direction, together with a disagreeable smell from the sinks in the houses and some smelly rubbish in the basement of one of the houses. In other words the orthodox explan- ation was that the disease was airborne, and that it was caused (as all epidemic diseases had been believed to have been caused for centur- ies) by the smell of putrefaction. The solution was improved hygiene. Snow pointed out that most of London was exposed to exactly the same sorts of smells that were to be found at Albion Terrace, but on other such streets nobody at all had died. Snow had rejected the conventional view that cholera was trans- mitted through the air and was primarily caused by putrefaction. Instead he argued that it was transmitted through the water supply and by direct contact, and was carried in the faeces of cholera suf- ferers. He was thinking in terms of a germ theory of the disease. The great advantage of this was that he could explain why cholera seemed to strike in a random fashion. If it was dissolved in the air (as Haygarth thought the poison that caused smallpox was) or in the water, then it should affect everyone, or everyone who used that water, more or less identically. But if it was some sort of living creature, however small, then one glass of water might contain the infection, while another, equally impure in other respects (Snow found bits of skin and hair in the piped water supply), might not. ‘The eggs of the tape-worm must undoubtedly pass down the sewers into the Thames, but it by no means follows that everybody who drinks a glass of the water should swallow one of the eggs.’ The germ theory was thus crucial to explaining the erratic or seemingly random incidence of the disease. Snow’s appeal to tape-worms seems obvious and straightforward to us, but the question of whether parasitic worms were spon- taneously generated was still being debated in the 1840s. A key text arguing against spontaneous generation was J. J. S. Steenstrup’s On the Alternation of Generations (1845), which gave an account of the life cycle of the liver fluke. Snow’s contemporaries had thus only just caught up with Leeuwenhoek and Swammerdam, who had been convinced that the microscope opened the way to a refutation of spontaneous generation. It is also significant that infection by tape- worms had been used as an analogy for germ infection by Jacob Henle in Pathologische Untersuchungen (1840), the most important defence of germ theory before Pasteur, precisely in order to explain the apparently random incidence of diseases. Snow never refers to Henle directly, but he may well have read him. Snow saw that it was possible to test his hypothesis that cholera was transmitted through drinking water on a larger scale. In some parts of London water was supplied from the Thames, into which most sewers emptied. The Thames is a tidal river, so sewage would have been carried up the river as well as down; even where water was drawn from a source above the sewage outfall it would have been polluted with sewage so long as it was not drawn from a point above the highest reach of the tide. Some companies filtered the water they drew from the Thames or passed it through settling pools, some did not. Yet other companies drew their water from uncontaminated springs. It was thus a simple matter to compare the incidence of cholera in particular districts with the supply of drinking water in those districts. In Southwark, in 1832, water was taken straight from the Thames without filter or settling reservoir. The death rate was 110 per 10,000. In equally poor Shoreditch, water came mainly from the New River and the River Lea. The death rate was 10 per 10,000. Exactly comparable results applied in 1849. The careful case studies Snow produced of transmission from person to person and from house to house were highly suggestive, as was the comparative work he did on the incidence of cholera amongst the customers of different water companies, but since his conclusions were directly opposed to the long tradition of Hippocratic medicine, with its single-minded emphasis on miasma or bad air, it is hardly surprising that his contemporaries found them unconvincing. As far as Snow was concerned the return of cholera in 1853 was to provide a perfect opportunity to test his arguments. Snow’s most important study during the epidemic of 1853–4 was a comparative analysis of the incidence of cholera amongst the cus- tomers of different water companies. Since one of the companies had changed its source of water, he could show that the general pattern was comparable to those for 1832 and 1849, with the major change that his hypothesis would have predicted, that an improvement in the purity of water had led directly to a decline in the incidence of cholera –– and he could show that similar changes had taken place in Glasgow and Hull when water companies altered their sources of supply. In general, in London different companies served different areas. But in some areas two companies were in direct competition with each other, with one house drawing water from one company, and its neighbour from another. This, Snow said, was the equivalent of a laboratory experiment to test his theory: As there is no difference whatever, either in the houses or the people receiving the supply of the two Water Companies, or in any of the physical conditions with which they are surrounded, it is obvious that no experiment could have been devised which would more thoroughly test the effect of water supply on the progress of cholera than this, which circumstances placed ready made before the observer. The experiment, too, was on the grandest scale. No fewer than three hundred thousand people of both sexes, of every age and occupation, and of every rank and station, from gentlefolks down to the very poor, were divided into two groups without their choice, and, in most cases, without their knowledge; one group being supplied with water con- taining the sewage of London and the other group having water quite free from such impurity. Snow and an associate set out to visit every house in which there had been a cholera fatality and establish which company supplied its water –– Snow cut back on his practice, effectively giving up his income, in order to pursue his enquiries. Between 8 July and 5 August 1854 there were 563 deaths from cholera in London, or 9 deaths per 10,000 houses. In 40,046 houses supplied by the polluted water of the Southwark and Vauxhall Co. there were 286 fatalities, or 71 per 10,000 houses. Amongst the identical houses intermingled amongst them but supplied with the clean water of the Lambeth Co. there were 14 fatalities, or 5 per 10,000. Note that these fourteen fatalities did not present a problem for Snow’s argument: it was to be expected that some customers of the Lambeth Co. would visit friends who were customers of the Southwark and Vauxhall Co. and drink their water; would purchase drinks made with Southwark and Vauxhall water in pubs and cafés; and would visit and nurse those who had fallen sick from drinking Southwark and Vauxhall water. Snow’s use of the customers of the two water companies as a randomized trial of his hypothesis resulted in a brilliant vindication of his arguments. (Snow’s survey was later repeated and extended with much less striking results. But it is Snow’s results that are to be trusted. He published a list running over twenty-five pages of every death and every address in his study; and he pointed out two major difficulties: there was often more than one house with the same street address, so that one had to make careful enquiries to make sure one was at the right address; and people often did not know the name of their water supplier, since that was a matter for their landlord –– a problem Snow had circumvented by devising a chemical test to distinguish the water supplies of the two companies.) Snow also showed that an analysis of the occupations of those who died from cholera was highly revealing. Sailors and ballast-heavers were accustomed to drink water direct from the Thames –– one in twenty-four of them died of cholera in the epidemic of 1848–9; those who worked in breweries were said never to drink water at all, and indeed none of them died. Contemporaries often implied that it really did not matter whether Snow was right or not. Both the miasmatists and Snow believed that human faeces helped spread cholera, so one could con- clude from both their arguments that improved sanitation was the answer. Snow was impatient with this response, as he held that the activities of those who advocated improved sanitation had had the opposite effect from the one they had intended. In 1854, publishing the preliminary results of his enquiry into the water supply, he wrote: The persons who have been more instrumental in causing the increase in cholera are precisely those who have made the greatest efforts to check it, and who have been loudest in blaming the supineness of others. In 1832 there were few water-closets in London. The privies were chiefly emptied by night men, a race who have almost ceased to exist; or a portion of the contents of the cesspool flowed slowly, and after a time, into the sewers. By continued efforts to get rid of what were called the removable causes of disease, the excrement of the com- munity has been washed every year more rapidly into the river from which two-thirds of the inhabitants, till lately, obtained their supply of water. While the faeces lay in the cesspools or sewers, giving off a small quantity of unpleasant gas having no power to produce specific diseases, they were spoken of as dangerous and pestilential nuisances; but when washed into the drinking-water of the community, they figured only in Sanitary Reports as so many grains of organic matter per gallon. Thus the difference between his own account of the transmission of the disease and that of the miasmatists was fundamental. His priority was clean drinking water; theirs was flush toilets. Just as Snow had studied both the customers of water companies and particular hotspots in 1849, so now in 1854 he did a detailed study of one particular hotspot, in Broad Street, by Golden Square. There, within a circle with a radius of two hundred and ten yards, five hundred people died of cholera within ten days, between 31 August and 9 September. In the first two days, 197 people died. The next day Snow interrupted his survey of water company customers to study this outbreak. He immediately looked for a possible source of con- taminated water, and his attention focused on a pump in Broad Street, by Golden Square, which stood roughly at the centre of the outbreak. He had no direct proof that the water was polluted (it was not until six months later that Henry Whitehead was to show that the first case had been that of a baby girl, and that the water in which her nappies had been washed had gone into a drain which leaked into the well). What he could do was trace the 83 people who had died in the immediate neighbourhood during the previous three days and whose deaths had been registered (of the 197 deaths others had not yet been registered or had occurred in hospitals outside the neighbourhood); of these he could show that 77 had almost certainly drunk water from the pump. Four days after he arrived in the neighbourhood, a week after the beginning of the outbreak, Snow asked the local authority, the parish vestry, to remove the handle from the pump. This they did; the outbreak was already diminishing (for the water was probably no longer polluted), and within twenty-four hours it was effectively over. In the 2nd edition of The Mode of Communication of Cholera Snow published a map of the fatalities and their relationship to the Broad Street pump and to the other pumps in the neighbourhood. In another version of the map he even drew a dotted line which enclosed all the places which were closer to the Broad Street pump than to any other: the line follows a complicated path, allowing for the fact that pedestrians, unlike birds, cannot take the shortest route between two points, but must follow streets and turn corners. (This is now called a Voronoi diagram, and Snow’s map is the first such diagram.) Nearly all the deaths fell within the dotted line. The story of the pump handle and the map that illustrates it have entered the folklore of epidemiology, though the often careless retelling of this story has opened the way to a fundamental misunder- standing of what Snow had accomplished. No map showing Broad Street and its immediate area to be a hotspot, it is said, could prove Snow’s account of how cholera was transmitted; for other con- temporaries produced similar maps, and they were convinced that such maps showed that the disease was most likely disseminated through the air. They hypothesized some unidentified source of miasma at the centre of the circle within which the deaths fell. Con- sequently, it is argued, Snow had failed to prove his case, and his opponents had at least as good an argument as he had. But this is to make the elementary mistake of imagining that the map was a full presentation of Snow’s material –– ignoring not only the other evi- dence we have so far surveyed, but the fact that Snow had crucial additional evidence relating to the Broad Street epidemic itself. The workhouse in Poland Street was near the epicentre of the

28. The map of the fatalities in the neighbourhood of the Broad Street pump from the second edition of Snow’s The Mode of Communication of Cholera. hotspot. It had 535 inmates but only 5 fatalities. The inmates breathed the air of Broad Street, but did not drink water from its pump. On Broad Street itself a brewery employed 70 men; none of them died. All drank beer not water; in any case the brewery had its own well. At a percussion cap factory a few yards away 18 of 200 workers died; they were supplied with water from the pump. Both those who worked in the brewery and those who worked in the percussion cap factory breathed the same air, but they did not drink the same water. One person, who came from Brighton, spent twenty minutes in the house of someone who had died of cholera, drank a glass of brandy diluted with water from the Broad Street pump, and died the next day. A woman who lived in Hampstead, an area untouched by cholera, and who never went to Broad Street, but who was regularly supplied with a bottle of water from the Broad Street pump (she had lived in the area in the past and thought the water particularly delicious) died, as did her niece, who drank from the same bottle –– Snow describes this case (which of necessity does not appear on his map) as ‘perhaps the most conclusive of all in proving the connexion between the Broad Street pump and the outbreak of cholera’. The point being that the man from Brighton, the widow from Hampstead, together with her niece who was from Islington, did not breathe Broad Street air, but died just the same because they drank Broad Street water. No hypothesis, other than Snow’s, could account for their deaths. The map was not a complete representation of Snow’s argument; it was merely an illustration of one part of it. If any epidemiological study could be conclusive, Snow’s was –– the standard modern objection to epidemiological studies is that they are not randomized, but Snow’s study of the area of London where two water companies supplied water in competition with each other amounted to a randomized study. Henry Whitehead, the local cler- gyman, who set out on a house to house enquiry in order to refute Snow’s argument, ended up being converted to it. And yet his con- clusions were not accepted. Two enquiries –– one instituted by the Metropolitan Commission of Sewers, and the other by the Commit- tee on Scientific Enquiries of the General Board of Health, represent- ing the national government –– ruled in favour of airborne contagion. The Committee on Scientific Enquiries was particularly impressed by a study by William Farr, which showed a very close correlation in London and to some degree elsewhere in England between the death rate from cholera and proximity to sea level –– Farr even produced a mathematical formula to show that at half the elevation the mortality doubled, and so forth: thus the mortality was in the inverse ratio of the elevation. This formula would have resulted in the death of each and every person who happened to live exactly at sea level, but Farr pointed out that buildings raise us above the ground, so that the inhabitants of buildings at sea level, he claimed, actually lived (pre- sumably he meant slept) at a height of 13 feet above sea level. He combined such mathematical exactitude with much pure nonsense. ‘On the high lands,’ he writes, ‘men feel the loftiest emotions . . . Man feels his immortality in the hills.’ Hence all religions associate their gods with hills, not with swamps. Farr feared that the English were increasingly living in cities close to sea level, which would result in the degeneration of the race and the collapse of the population. The government should force people to build on higher ground. Snow had already effectively demolished Farr when the Commit- tee on Scientific Enquiries declared its support for him: ‘Dr Farr’, says Snow, was inclined to think that the level of the soil had some direct influence over the prevalence of cholera, but the fact of the most elevated towns in this kingdom, as Wolverhampton, Dowlais, Merthyr Tydvil, and Newcastle-upon-Tyne, having suffered excessively from this disease on several occasions, is opposed to this view, as is also the circumstance of Bethlehem Hospital, the Queen’s Prison, Horsemonger Lane Gaol, and several other large buildings, which are supplied with water from deep wells on the premises, having nearly or altogether escaped cholera, though situated on a very low level, and surrounded by the disease . . . Farr’s low-lying towns struck by cholera were, any reader of Snow will recognize, towns that drew water from tidal rivers. Farr was not wrong to think there was some sort of correlation between height above sea level and the incidence of cholera; his mistake lay in assum- ing that the difference was in the air people breathed not the water they drank. Of all the objections made to Snow there is only one that presented any difficulty for Snow’s argument. If cholera could be spread from hand to mouth, why did doctors (who often conducted autopsies on cholera victims) and undertakers (who handled their bodies) not fall ill as frequently as the relatives of the sick? Snow’s answer was that doctors and undertakers took care to wash their hands –– for him this constituted evidence that the simplest of precautions could serve to keep cholera at bay. There were two people whom Snow needed to convince. One was his opponent William Farr, who was superintendent of the statistical department of the General Register Office, and the acknowledged expert on the use of statistics in the study of diseases. As we have seen, in the 1850s he had an alternative theory; by 1866 however he was a convert to Snow’s germ theory, and he played a leading role in showing that the 1866 cholera epidemic was water- borne. The other was John Simon, who was Medical Officer of Health to the Privy Council, and the authority on whom the gov- ernment relied. Simon was happy to accept that water played a role in the dissemination of cholera, but he thought it was only one factor amongst many. He too converted in 1866, claiming that the decisive evidence was not the ‘popular’ experiment (that is, ‘the experiments which accident does for us’) ‘performed on half a million human beings in South London by the commercial water companies’ but the ‘scientific’ experiment of Karl Thiersch. Thiersch, working in Germany, took fresh intestinal liquid from cholera victims. He allowed it to decompose by exposure to air. He then fed small quantities to mice. He found that fresh faecal matter was harmless, as was faecal matter that had stood around for a week or so, but faecal matter that was three or four days old was deadly. The mice, Thiersch claimed, had successfully been infected with cholera. Thiersch’s work was immediately publicized by Justus von Liebig, the greatest biochemist of the day, who held the view that diseases were caused by poisons. From Simon’s point of view the great attraction of this experiment was that it explained why doctors and undertakers, who encountered relatively fresh faecal matter, were not infected. But there are several problems with Simon’s conversion as a result of Thiersch’s experiment. The first is that the experiment had been conducted as early as 1854, and indeed had been rejected by Snow in his 2nd edition. Snow was sure the experiment was faulty because he knew the incubation period of the disease was only twenty-four to forty-eight hours –– there was no time for the three or four day period of putrefaction posited by Thiersch. The second is that we now know you cannot give mice cholera by feeding them cholera germs. So whatever Thiersch had done, he had not infected his mice with cholera. Nevertheless, in 1866 Simon had Thiersch’s experiments repeated in England and declared them decisive. Why? What had changed was not the evidence regarding the transmission of cholera, which was exactly the same as it had been when Snow published his 2nd edition. What had changed was the acceptability of miasmatic theories of disease. Villemin had shown that tuberculosis could be transmitted from one animal to another. Sanderson had done similar work with cattle plague. Pasteur’s work was beginning to be widely discussed –– it was known to Farr. The medical profession was on the point of abandoning its commitment to miasmatic theories of epidemics and of adopting the view that each disease had a specific agent, even if it had yet to make a choice between Leibig’s view, that the agent was chemical, and Pasteur’s, that it was living. Farr and Simon were simply dedicated followers of fashion. This time, for a change, the fashionable view happened to be the right one.

Random Posts

Comments are closed.