Content uploaded by Tom Koch
Author content
All content in this area was uploaded by Tom Koch on Jul 29, 2014
Content may be subject to copyright.
Essential, Illustrative, or . . . Just Propaganda?
Rethinking John Snow’s Broad Street Map
Tom Koch
Department of Geography / University of British Columbia / Vancouver / BC / Canada
Ken Denike
Professor Emeritus / Department of Geography / University of British Columbia / Vancouver / BC / Canada
Abstract
For more than a century John Snow’s iconic map of an 1854 cholera outbreak in the Broad Street area of Soho, London,
has been the very definition of how to discover the source of a disease. Some now argue, however, that the map was
merely an illustrative and not very imaginative graphic. Here we argue that this position is incorrect. Snow’s mapping
of the Broad Street outbreak produced a spatial argument that was a critical evidentiary statement. This position requires
us to ask, If that is true, is the map in part responsible for Snow’s inability to convince contemporaries of his argument
that cholera was water-borne and not airborne? In doing so, we use mid-nineteenth-century methodologies to
demonstrate that the problem was not in the map but in Snow’s handling of the data. This review of a seminal study
in the history of disease studies not only informs historical perspective but, in its conclusions, speaks to the utility of
medical mapping in contemporary disease studies, where spatialization of a disease event remains a critical method
of investigation.
Keywords: cholera, disease studies, epidemiology, history of cartography, John Snow, medical cartography, public health
Re
´sume
´
Pendant plus d’un sie
`cle, la carte iconique de John Snow portant sur une flambe
´e de chole
´ra en 1854 dans le secteur de
la rue Broad du quartier Soho, a
`Londres, a de
´fini la manie
`re de de
´couvrir la source d’une maladie. Toutefois, certains
affirment maintenant que la carte n’a servi que d’illustration, sans vraiment e
ˆtre imaginative. Selon nous, cette prise de
position est incorrecte. La carte de Snow repre
´sentant l’e
´pide
´mie de la rue Broad a e
´te
´au cœur des observations : la carte
a fourni un argument spatial, un e
´nonce
´de preuve essentiel. Cette position nous oblige a
`nous demander : « Si cela e
´tait
vrai, la carte est-elle en partie responsable de l’incapacite
´de Snow a
`convaincre ses contemporains de son observation
selon laquelle le chole
´ra se propageait par l’eau et non par l’air? » On utilise des me
´thodologies du milieu du XIX
e
sie
`cle
pour de
´montrer que le proble
`me n’e
´tait pas la carte, mais la manie
`re dont Snow a utilise
´les donne
´es. Cette revue d’une
e
´valuation se
´minale de l’histoire de l’e
´tude des maladies ne sert pas seulement a
`donner une perspective historique, mais,
dans sa conclusion, explique l’utilite
´du mappage me
´dical lors de l’e
´tude des maladies contemporaines, dans lesquelles la
spatialisation d’une maladie demeure une me
´thode essentielle d’investigation.
Mots cle
´s: chole
´ra, e
´tude de maladie, e
´pide
´miologie, histoire de la cartographie, John Snow, cartographie me
´dicale, sante
´publique
For more than 100 years John Snow’s iconic maps of an
1854 cholera outbreak in the Broad Street area of Soho,
London, have been the very definition of how to discover
the source of a disease outbreak (Snow 1855a, 1855b). A
standard image in public-health texts since W.T. Sedg-
wick’s landmark publication in 1901 (Koch 2005a, 132–
Cartographica (volume 45, issue 1), pp. 19–31 doi: 10.3138/carto.45.1.19 19
40; Sedgwick 1901), Snow’s Broad Street map has been a
foundational example advanced at varying times by carto-
graphers (Robinson 1982), graphic experts (Tufte 1983,
1997), geographers (Koch and Denike 2004; Monmonier
1996), public-health officials (Sedgwick 1901), and histo-
rians of science, medicine and technology (Shapin 2006).
Nor is interest in Snow’s mapping limited to professional
communities; the story of John Snow and his maps is also
the subject of popular books (S. Johnson 2006; Hempel
2006) and Web sites (Frerichs 2008).
Yet some modern commentators dismiss Snow’s mapping
as merely illustrative (Brody and others 2000), at best
‘‘merely preparation’’ for Snow’s real analysis (Vinten-
Johansen and others 2003, 336). At worst, the maps are
denigrated as mere propaganda, and not very good pro-
paganda at that (Monmonier 2002, 155). This is also the
perspective of some Snow biographers who, in devaluing
Snow’s Broad Street mapping, dismiss medical cartogra-
phy and geography in general as fields for those ‘‘those
who won’t do the hard work of building a careful and de-
cisive argument about the nature and origin of a disease
outbreak’’ (Vinten-Johansen and others, 397–98).
Here we consider this curious division, one in which the
Broad Street map is either incidental and irrelevant or, as
we insist, central to Snow’s fundamentally spatial argu-
ment correlating mortality in the Broad Street cholera
outbreak with proximity to the Broad Street pump and
well. We argue that in denying the importance of Snow’s
maps, the modern critics ignore the essentially spatial na-
ture of his argument and, by extension, of disease map-
ping in general. Asserting the evidentiary importance of
Snow’s maps raises an interesting question: If the map
was central to Snow’s argument, was his inability to con-
vince contemporaries of his argument a failure of the
maps themselves? At the least, it is useful to ask whether
Snow’s maps, the traces of his argument, give some in-
sight into the nature of Snow’s argument and its popular
failure in mid-nineteenth-century medicine and public
health. Asking these questions raises the more general
question of the nature and role of mapping in disease
studies.
Snow’s Cholera
Snow published a short monograph, On the Mode of Com-
munication of Cholera (1849a [MCC1]), that argued the
then radical thesis that cholera was not, as most believed,
a miasmatic disease borne on the foul airs of the city but
instead a water-borne disease that was ingested, not in-
haled. Most agreed with an anonymous reviewer, proba-
bly E.A. Parkes, who wrote in the London Medical Gazette
that while Snow’s theory of water-borne disease was intri-
guing, the data he presented were inconclusive ([Parkes]
1849). Almost immediately, Snow attempted to bolster
his argument, publishing two supporting papers within
months (Snow 1849b, 1849c) to reinforce his thesis with
new data.
Snow was not alone in his search to unravel the nature of
cholera. Scores of researchers across Europe and North
America were simultaneously advancing different theories
of cholera, most advocating an airborne disease agent
within the then current miasmatic disease theory (Koch
2005a). Some researchers saw water as a potential cholera
source, either alone or as part of a theory of air- and
water-borne cholera. This was the argument advanced by
William Farr, compiler of abstracts for Britain’s General
Register Office (GRO), who in 1852 published an encyclo-
paedic 400-page study of the cholera epidemic of 1848–
1849 in Great Britain. In 100 pages of text supported by
300 pages of maps, graphs, and tables, Farr argued a com-
plex, multi-factorial theory of cholera that included air
and water as elements contributing to its generation and
dissemination (Farr 1852).
As the medical historian P.E. Brown put it in 1961, albeit
without conscious irony, ‘‘All the elements of his [Snow’s]
theory were already in the air’’ by 1849. In 1848, for ex-
ample, R.D. Thomson and William Farr expressed their
concerns about the wells of Glasgow in The Lancet (Brown
1961, 30); the next year William Budd argued that sewer-
contaminated well water was a source of typhoid and
likely of cholera as well. ‘‘The frightful fatality of the dis-
ease in particular parts of infected towns,’’ Budd con-
cluded, was caused by the sewer-contaminated drinking
water found in that single well (1849, 9). Elsewhere, a
researcher in Cincinnati, Ohio, argued that well water
was the likely source of a local cholera outbreak and pro-
moted drinking rainwater as a prophylaxis (Lea 1850). He
believed, however, that the mineral content of the water,
and not contamination, was the cause of cholera.
A reviewer in The Lancet argued that conclusive proof of
the nature of cholera was almost impossible without defi-
nite proof of its invisible agent (Lancet 1853). That proof
would have to wait until Robert Koch’s identification of
the bacterium in 1883. Snow, however, believed that he
could craft a conclusive argument, even without that evi-
dence, on the basis of three studies he carried out simul-
taneously during the 1854 cholera epidemic, investigating
cholera in South London’s registration districts, a local-
ized outbreak in the town of Deptford, and, finally, a fero-
cious outbreak in the Broad Street area of St. James,
Westminster, in central London (Snow 1854). In 1855
these separate investigations were published together in a
greatly expanded second edition of Snow’s 1849 mono-
graph (Snow 1855a [MCC2]).
In it pride of place was devoted to the ambitious South
London cholera study. This was, Snow asserted, work
‘‘on the grandest scale,’’ a natural experiment involving
‘‘no fewer than three hundred thousand people, of both
sexes, of every age and station’’ in which cholera mortality
would be correlated with water provided by one or another
Tom Koch and Ken Denike
20 cartographica (volume 45, issue 1)
local water supplier (Snow 1855a, 75). Snow sought to
make his case at three distinct scales: that of the local
water-supply companies (first categorized by Farr in
1852), that of registration districts, and the very fine scale
of registration sub-districts within them.
‘‘All that was required,’’ wrote Snow, ‘‘was to learn the
supply of water to each individual house where a fatal
attack of cholera might occur’’ (Snow 1855a, 75). Alas,
those data were unavailable until well after MCC2 went
to press. As Snow admitted later, ‘‘I was unable at the
time to show the relation between the supply of houses
in which fatal attacks took place, and the entire supply of
each district and sub-district, on account of the latter cir-
cumstance not being known’’ (1856, 7). Without those
data, Snow’s 1855 study could only argue cholera at the
coarse scale of local water suppliers, one that did not rule
out either the geographic or the socio-economic variables
that Farr and others had argued might be determinants
of the disease. The result was, as one reviewer correctly
noted, at best suggestive but certainly not definitive
(Parkes 1855). Nor, when the necessary data became
available (Simon 1856), was Snow able to successfully cal-
culate a conclusive argument based on these new data
(Koch and Denike 2006).
The Deptford study was similarly suggestive but inconclu-
sive. The Broad Street outbreak, therefore, was the only
study in which Snow could rigorously construct a precise
and fine-grained argument, one based on street-by-street
mortality, that cholera could only be a water-borne dis-
ease. In addition to the version published in MCC2,
Snow prepared a second report for a local parish investi-
gation into the outbreak centred on Broad Street (Snow
1855b).
Broad Street: Case Evidence
For his study Snow collected two separate types of evi-
dence. The first employed short case histories of cholera
victims in an attempt to demonstrate that all drank water
from the Broad Street pump. Most of these cases were not
collected by Snow but reported to him by other physicians
practising in the stricken area, most notably Dr Marshall
of Greet Street (Snow 1855a, 43) and Dr Fraser of Oakley
Square (44). In addition, Snow could draw upon the case
histories collected by the local curate, Rev. Henry White-
head, who in 1854 published a detailed analysis of cholera
in his parish.
While certainly suggestive, the circumstantial evidence
presented by these histories was less than conclusive to
Snow’s contemporaries. Dr John Simon, for example,
questioned the scale of the neighbourhood study as a basis
for a general theory of cholera (Simon 1856). Rev. White-
head’s cases pointed to the possibility of divine provi-
dence, and, as E.A. Parkes would later point out in a re-
view of MCC2, many of the cases could be explained in
other ways (Parkes 1855, 456–57). The problem was not
specific to Snow but a general problem of arguing on the
basis of anecdotal case histories: they could be used to
prove anything. In 1855, for example, Dr George Johnson,
an assistant physician at King’s College Hospital in
London, published a 294-page treatise with 54 case his-
tories (including autopsy reports) to insist that cholera
must be pulmonary, and therefore inhaled (G. Johnson
1855). An anonymous reviewer praised the study for the
thoroughness of its case presentation, if not for its treat-
ment protocols promoting the ‘‘eliminative plan of treat-
ment’’: castor oil (‘‘Bibliographical Record’’ 1855).
Broad Street Evidence: The Maps
The second class of data presented by Snow was carto-
graphic. In his maps Snow used mortality records col-
lected by the GRO to create a class of cholera incidence
related in the map by proximity to public water sources.
The purpose of the map was not ‘‘hypothesis generating’’
(Monmonier 2002) but ‘‘hypothesis testing.’’ By creating a
mapped class of all cholera deaths located in relation to
the Broad Street well and pump (see Figure 1), Snow
crafted a spatial argument in which one class of events,
cholera deaths, was positively correlated on the map with
one member of a class of suspected contagion sites.
In the MCC2 map, a dotted line defines a principal study
area, based on registration sub-district boundaries, within
which cholera deaths reported to the GRO were located
by address. The overwhelming centrality of the Broad
Street pump in relation to the majority of GRO-reported
deaths was the fine-grained, spatially precise argument
Snow had promised but had been unable to deliver in
previous studies, including the South London study. The
combination of case histories and the map that made of
them a single analytic field is what in 1855 distinguished
Snow’s argument from those of others, such as Johnson
of King’s College Hospital.
Antecedents
There was nothing exceptional about Snow’s construction
of a spatial argument lodged in maps of a specific disease
event. In nineteenth-century disease studies generally
(Koch 2005a, 2005b) and in the cholera literature speci-
fically (Koch 2008), disease mapping was an accepted
medium of analysis. The use of maps as a medium in
which spatial arguments related disease incidence to sus-
pected origin sites was an eighteenth-century innovation
that by the mid-nineteenth century was generally accepted,
even expected. Among the earliest surviving studies of this
type are Valentine Seaman’s 1796 maps correlating the in-
cidence of yellow fever with sites of ‘‘furry miasmata,’’
odiferous human and animal waste, in New York City.
These ‘‘hypothesis-testing’’ maps attempted to prove the
Essential, Illustrative, or . . . Just Propaganda?
Cartographica (volume 45, issue 1) 21
airborne nature of yellow fever through the demonstrable
proximity of yellow-fever deaths to sites of odiferous
waste (Koch 2005a, 28–33).
In the first decades of the nineteenth century, the use of
such maps in disease studies grew rapidly (Koch 2010).
Certainly, from the first report of cholera’s incidence in
India in 1819 (Jameson 1819), cholera was continually
mapped at the scale of the neighbourhood (Lea 1850),
the city (Hammett 1832; Reese 1833; Hellis 1833), the
nation (Farr 1852), and the world (‘‘History of the Rise’’
1831). Scores, perhaps hundreds, of books, papers, pam-
phlets, and reports published in England, Europe, and
North America used maps to explore various theories
of cholera through the analysis of patterns of disease in-
cidence in relation to potentially causal environmental
elements.
Even those whose reports did not include maps recog-
nized the importance of the medium in disease research.
For example, Dr John Simon, medical officer of health
for the City of London, wrote in a report on cholera mor-
tality in 1854 that ‘‘when the 211 deaths are mapped upon
a house-plan of the City (as may conveniently be done by
stamping a black ink mark at each place where one of
them has occurred) the broad features of the epidemic
are rendered visible at a glance’’ (10–11). While Simon
included no maps in this report, the map ‘‘thinking’’
(Brody and others 2000, 68), and its importance as a
methodology, was clearly stated in his work. Absent a
robust microscopy, the spatial analytic of the map was
perhaps the central medium in disease studies in this era.
Within this tradition, Snow’s maps were neither inciden-
tal nor merely illustrative. They were the workbenches on
which Snow constructed a fundamentally spatial argu-
ment relating clusters of disease mortality in relation to a
suspected environmental disease source. Other researchers
simultaneously investigating the Broad Street outbreak –
including the London Sewer Commission engineer Edmund
Cooper (1854) and parish curate Rev. Henry Whitehead
(1855) – similarly mapped GRO mortality records in
separate investigations of what Snow called ‘‘the most ter-
rible outbreak of cholera which ever occurred in this
kingdom’’ (1855a, 38).
Snow’s Maps
Snow hired C.F. Cheffins, a prominent London engraver
best known for his transportation maps, to modify an
existing street map of Snow’s study area to create the
Figure 1. John Snow’s famous map of the 1854 Broad Street epidemic attempted to positively correlate disease intensity
with proximity to a single water source, the Broad Street well and pump.
Tom Koch and Ken Denike
22 cartographica (volume 45, issue 1)
cartographic arena in which Snow might make his case.
Measuring 415 384 mm, the foldout map included a
dotted line that created out of official registration sub-
district boundaries a single study area in which Snow’s
argument could be inscribed. To this Snow had added
the location of 14 public wells, as well as those of 596
deaths reported to the GRO. Each death was symbolized
by a black rectangular mark, signifying the home of each
decedent. In Snow’s second map, for the parish inquiry,
he included additional deaths for which he had earlier
had no home address.
These postings, as Denis Wood and John Fels (2008)
would call them, permitted Snow to demonstrate a posi-
tive correlation between increasing mortality and decreas-
ing distance from the Broad Street well and pump at the
epicentre of the outbreak. Snow’s evidentiary use of the
map, which he described as ‘‘a diagram of the topography
of the outbreak’’ (1855a, 45), was limited to the observa-
tion that deaths appeared to be clustered in the area of the
Broad Street pump: ‘‘It may also be noticed,’’ he wrote,
‘‘that the deaths are most numerous near to the pump
where the water could be more readily obtained’’ (1855a,
47). To a second map prepared for the St. James Parish
inquiry committee Snow added an irregular polygon
based on walking distance from the Broad Street pump
(Snow 1855b; see Figure 2). In the first map Snow there-
fore created a class of mapped cholera deaths across the
study area; the second map’s polygon identified a subset,
‘‘Broad Street pump deaths,’’ defined by their relative
proximity to that water source. The apparent density of
that cluster, which was not analysed in any other fashion,
was the centrepiece of Snow’s spatial argument.
In the main, Snow’s contemporaries found his conclu-
sions suggestive but certainly not definitive. One problem
was that the Broad Street pump at the epicentre of the
outbreak was well known for the quality of its water and
therefore seemed an unlikely source for the cholera out-
break. England’s best-known microscopist, Arthur Hill
Hassall, examined a sample of the well water and declared
it ‘‘relatively bereft of [contaminated] microscopic animal
life’’ (Vinten-Johansen and others 2003, 247). Snow ad-
mitted that the absence of observable contaminates forced
him to ‘‘hesitate to come to a conclusion’’ on the pump
water’s complicity in the outbreak (1855a, 39). He could
demonstrate the centrality of the well and pump but
could only assert deductively its pollution: ‘‘Whether the
impurities of the water were derived from the sewers, the
drains, or the cesspools, of which latter there are a num-
ber in the neighborhood, I cannot tell’’ (1855a, 53).
Another problem was that some researchers were reluc-
tant to accept as definitive the results of a single large-
scale neighbourhood study. Simon (1856), for example,
argued that because most London neighbourhoods were
served by a plethora of local wells, there would always be
a water source near the centre of a neighbourhood out-
break. Other researchers similarly mapped other local
Figure 2. In a second map prepared for a local parish inquiry committee, Snow created an irregular polygon based on
walking distance to identify a subset of cholera deaths located nearest the Broad Street pump.
Essential, Illustrative, or . . . Just Propaganda?
Cartographica (volume 45, issue 1) 23
outbreaks and their water sources, only to arrive at con-
clusions very different from Snow’s – for example, John
Lea (1850) in Cincinnati, and Thomas Shapter (1849) in
Exeter.
A third problem was Snow’s wholly non-quantitative
treatment of his data in a period when mortality ratios
were the common mechanism for describing disease mor-
tality. In his 1849 monograph Snow had insisted that ‘‘the
subject is capable of being decided by exact numerical
investigation’’ (1849a, 16; Paneth 2004). Elsewhere – in
MCC2 and in other papers – Snow made frequent use of
mortality ratios. Yet in the Broad Street study, the only
study in which precise mortality data could easily have
been crafted, Snow made no effort to quantify his visual
argument. But the mapped cluster of cholera deaths
meant little without data on the population of the streets
in which the cholera victims lived. If the population of
Broad Street was 10 times that of, say, Rupert Street,
then 10 times as many deaths might be expected in the
former as in the latter.
Finally, many objected to Snow’s declarative style, his stri-
dent insistence that his data and its treatment were suffi-
cient to prove his theory (see, e.g., Parkes 1855). In the
same vein, many objected to Snow’s often dismissive re-
jection of other theories of cholera and the arguments
that supported them: ‘‘Snow’s colleagues did not so much
oppose his theory as they objected to his dismissing other
explanations for cholera’s occurrence’’ (Eyler 2001, 26). It
was not that Snow was wrong but that he had not proved
his thesis that water, and only water, was the source of the
outbreak. The problem was not in the maps, we argue,
but in Snow’s handling of mapped data that were carto-
graphic and spatial.
Quantifying Broad Street
To satisfy his contemporaries Snow would have needed
three things. First he would have needed to transform his
visual argument into ‘‘exact numerical investigation’’
based on his mapped data (Snow 1849, 16; Paneth 2004).
Second, he would have needed a mechanism permitting
him to compare the Broad Street water-service area with
others in the affected registration sub-districts. Third,
Snow would have needed to demonstrate, within the
limits of the science of his day, that his solution was
more likely than those advanced by other researchers as
suspected sources of cholera. Below we demonstrate that
with very little extra effort Snow could, within those
limits, have completed these tasks using data that, like
his, were cartographic and spatial.
Snow could not count the number of deaths in the
observed cluster in the famous MCC2 map (Figure 1)
because its boundaries were unclear. Did Carnaby, King,
or Marshall Street define its western boundary? Where
did the cluster begin and end to the south? Without
some form of geographic boundary, Snow could only
speak generally about cholera incidence in his study area.
This problem was solved in Snow’s parish inquiry map,
with its irregular polygon based on walking distance.
Within Snow’s Broad Street service area, a total of 381
deaths were mapped in 223 houses; in other words,
approximately two-thirds of Snow’s 596 mapped deaths
were located in the Broad Street pump’s service area.
Calculated another way, more than half of all houses in
which cholera occurred in the general study area (N¼
403) were situated in the Broad Street water-service poly-
gon (see Table 1). Even this level of quantification of the
mapped data would have transformed Snow’s argument
based on a visual impression into a forceful, numerically-
based conclusion.
The data required for a rigorous consideration of Broad
Street mortality were available to Snow. We can see
this in the map produced for the Sewer Commission by
Edmund Cooper (1854; see Figure 3) and in another sub-
mitted by Rev. Henry Whitehead (1855) as part of his
report to the St. James Parish Cholera Inquiry Committee.
Whitehead’s map, in turn, modified one produced for the
Board of Health’s report on cholera and the Broad Street
outbreak (Board of Health 1855). Cooper’s map posted
only 351 deaths, those occurring in the first two weeks of
the outbreak, while Whitehead’s map and that of the
Board of Health included a total of 684 deaths occurring
across the outbreak. Snow’s maps used a data set of the
outbreak that included only deaths through September
1854. In Snow’s mapping, cholera deaths were assigned
to streets that included no address. In Cooper’s and
Table 1. Three researchers created three different maps for simultaneously studies of the Broad Street outbreak. In
addition to Snow’s maps (given here as one) were another by Rev. Henry Whitehead and a third produced for
the London Sewer Commission by engineer Edmund Cooper.
Total Map Area Study Map Area Total Streets Total Deaths Total Pumps
m
2
m
2
nnn
Snow (MCC2) 867,849 538,077 304 596 14
Whitehead 588,096 442,903 190 684 11
Cooper 673,654 494,473 190 351 9
Tom Koch and Ken Denike
24 cartographica (volume 45, issue 1)
Whitehead’s maps of the outbreak, each street in their
respective study areas included the number of houses on
each street and located mortal cholera cases by house
address.
Like Snow’s, the study areas of Cooper’s and Whitehead’s
maps were bounded to the west by Regent Street and to
the north by Oxford Street. Snow, however, included
several deaths and water sources outside that boundary in
his maps, whereas Cooper and Whitehead confined their
street-address and mortality-incidence data to the study
area. The data included in these other maps provided the
platform Snow could have used to make his case more
forcefully within the parameters of the science of the day.
To demonstrate this, we transferred data from White-
head’s map to a photocopy of Snow’s second map and,
in the nineteenth-century manner, did the necessary cal-
culations by hand.
The numbers of houses per street recorded in White-
head’s and Cooper’s maps were manually counted (they
were the same) and then added to the copy of Snow’s
map. Taking the total number of deaths in Snow’s irregu-
lar polygon as numerator and the total number of houses
in the area as denominator, we fashioned a crude mor-
tality ratio of deaths per house. Given a reported average
occupancy of 10 persons per house, based on 1851 census
data (Farr 1852), the population (10 persons/house) of all
Figure 3. Edmund Cooper located cholera by house number. Thick black lines indicate homes in which cholera deaths
occurred; shorter black bars stacked under the black lines signify multiple deaths at a single location. Evident in the map
is the erroneous location of the 1665 plague burial site.
Essential, Illustrative, or . . . Just Propaganda?
Cartographica (volume 45, issue 1) 25
255 houses in the irregular polygon as the denominator
represented a better divisor (2550), being a mortality ratio
of 149.41 deaths per 1000 persons.
To be convincing, however, Snow would have required
comparative mortality ratios for adjacent pump catch-
ments. He could have created catchments similar to his
single polygon based on walking distance for other pumps
in his study area; however, he did not. Here we faced
a methodological problem. The difference between the
London of the mid-1850s and the modern, automobile-
dominated city makes it impossible to create equivalent
service catchment areas based on pedestrian walking dis-
tance today. Street traffic and access have changed too
much.
To demonstrate the potential for a nineteenth-century
districting analysis, therefore, we drew a set of Thiessen
polygons centred on other mapped water pumps in the
study area. The lines joining all these individual service
areas created a Voronoi network, a continuous set of
polygons each centred on a single public well and pump.
Formally, the result is called a Dirichlet tessellation, after
Snow’s contemporary, the nineteenth-century mathemati-
cian P. Lejeune Dirichlet, who first described the proce-
dure (Bailey and Gattrell 1995). The procedure requires
that points be located midway between all subject loca-
tions, in this case water wells and pumps. These points
are then connected to create polygons whose edges are
each equidistant between two and only two centroids
(wells). All subjects (deaths) in each of the resulting poly-
gons (catchments) are nearer to its centroid than to any
other in the set (of wells and pumps).
The polygons created by hand in this manner served as
water-service catchments for all pumps in the study area.
Polygons were not created for pumps whose areas ex-
tended beyond Oxford and Regent Streets and thus
beyond Cooper’s or Whitehead’s mapped data set of
house locations and numbers. Together, however, the
eight water-service areas for which polygons were con-
structed contained slightly more than 93% of all deaths
mapped by Snow. For this study, deaths were counted to
create numerators for each polygon, and the number of
houses in each polygon was used to construct a mortality
denominator.
The results, summarized in Table 2, would have strongly
supported Snow’s thesis, with 28.71 deaths per 1000 in
the Rupert Street pump-service area, 25.88 deaths per
1000 in the Little Marlborough Street pump-service area,
and 12.34 deaths per 1000 person in the Warwick Street
pump-service area. Importantly, the number of deaths
per house, a simple nineteenth-century measure of inten-
sity, diminishes as one moves outward from the Broad
Street service area.
Using this very nineteenth-century methodology, Snow
also could have discounted the likelihood that other, fre-
quently mentioned sites of potential contagion were the
source of the outbreak. Consider, for example, the 1665
plague burial site. Inexplicably, Cooper located it in the
northwest quadrant of the study area, and Snow accepted
that assignment based on Cooper’s assessment and the
word of Cooper’s London Sewer Commission boss.
‘‘Non-medical people’’ thought the old ‘‘pest-field’’ was
more central, Snow wrote; ‘‘the situation of the supposed
pit is, however, said to be Little Marlborough Street, just
out of the area in which the chief mortality occurred’’
(1855a, 54).
Unfortunately, Cooper was wrong, and the ‘‘non-medical’’
people were right. Whitehead correctly described the loca-
tion of the former plague burial site as both larger and
nearer the epicentre of the outbreak, its southwest boun-
dary a block’s distance from the Broad Street well and
pump (see Figure 4). To confirm this, we checked his-
torical maps of the district by engraver Richard Blome
Table 2. Calculating mortality in Broad Street water-service areas*
Pump Locations Snow
Cholera
Deaths
Snow
Cholera
Houses
Intensity per
House
(Deaths/Houses)
Total Houses
in Service
Area
Population for
Service Area
Mortality per
1000 persons
Broad Street 381 223 1.71 255 2550 149.41
Rupert St. 38 60 0.63 209 2090 18.18
Little Marlborough St. 59 43 1.37 228 2280 25.88
Briddle St. 30 28 1.07 135 1350 22.22
Warwick St. 19 16 1.19 154 1540 12.34
Marlborough Mews 7 5 1.4 48 480 14.58
Berniers St. 13 14 0.93 115 1150 11.3
Newman St. 21 14 1.5 N/A N/A
*A mortality ratio is defined as the total number of deaths divided by the total population of a pump’s service area. Population was
defined as the number of houses per service area multiplied by an average of 10 persons per house.
Tom Koch and Ken Denike
26 cartographica (volume 45, issue 1)
(1720). Snow certainly knew that others believed the old
‘‘pest-field’’ to be proximate to the outbreak’s centre. We
know, too, that during this period Snow worked in close
association with Whitehead, the parish curate, and thus
would presumably have had access to his materials (W.
Whitehead 1865).
Using Whitehead’s map as a guide, we drew the boun-
daries of the old plague burial site on a photocopy of
Snow’s MCC2 map. To this we added an approximately
10-m buffer to allow for the winds that some believed
carried miasmatic odours, mentioned in popular reports,
into neighbouring streets. We then counted all deaths
that occurred within the heavily settled buffered area, as
well as the number of houses on all street segments within
it. There were 113 cholera deaths in 57 affected houses (of
a total 209 houses), a mortality ratio of 104.63 deaths per
1000 persons – sufficiently high to explain the concern of
Snow’s contemporaries, but far lower than the mortality
ratio centred on the Broad Street well and pump. Within
the parameters of mid-nineteenth-century science, we be-
lieve this would have been very convincing evidence indeed.
Calculating mortality for the sewer lines is a more com-
plex task that today would require some sophistication. A
crude method that would have served in Snow’s day
would be to count deaths on streets proximate to the sus-
pect post-1850 sewer lines, as well as the houses that lined
the streets along which those sewers ran, to create a sim-
ple mortality ratio for those streets. We counted 219
houses on sewer-suspect streets on Snow’s map. Among
them were 139 cholera deaths in 73 houses, about 10%
of the total number of houses. Mortality per house was
1.51, far lower than in Broad Street, and general mortality
for these streets was calculated as 63.47 per 1000 persons.
Analysed in this fashion, the apparent culpability of the
sewer lines that Londoners commented on in the news-
papers of the day could have been definitively denied.
The results appear to be definitive: the mortality ratio of
the Broad Street service area was so much higher than
that of any other catchment, or possible contagion site,
as to strongly argue the likelihood that the Broad Street
well was indeed the source of the outbreak. We believe
these findings would have been convincing to Snow’s con-
temporaries. Certainly they would have been more con-
vincing than the argument Snow made in MCC2.
Discussion
It took us four or five days of intermittent labour to count
the houses per street, transfer the old cemetery site to
Snow’s map, create the buffer, and trace the critical sewer
lines to Snow’s maps. With these data in place, the calcu-
lations, carried out by hand (but checked with an electric
calculator), took perhaps two hours. We believe that
Snow’s contemporaries would have seen these results as,
if not definitive, then certainly far more convincing than
those Snow presented. We further believe that the ques-
tions asked by Snow’s contemporaries were correct and
deserved to be seriously considered.
Figure 4. Rev. Henry Whitehead included in his map both Cooper’s incorrect location of the 1665 plague burial site and a
correctly sited area whose southwest corner is a block from the Broad Street pump.
Essential, Illustrative, or . . . Just Propaganda?
Cartographica (volume 45, issue 1) 27
The methodology we used in this study was one current
in Snow’s era and – given his frequent use of mortality
ratios elsewhere in his work from 1849 through 1858 –
clearly one he was familiar with. Why, then, did Snow
not take the time to do the work that would have been
convincing in the science of his day? We can only specu-
late. It may be that Snow believed his maps presented so
clear and compelling a statement that no further calcula-
tion was necessary. Mapping was an accepted methodol-
ogy whereby classes of disease events were first con-
structed and then compared to proximate environmental
contaminants. In his maps Snow could see the suspect
well and pump at the very epicentre of the mapped class
of deaths. If Snow thought nothing more was needed to
make his case, however, the reservations of his critics
make clear that in this assumption he was wrong.
A more practical reason for Snow’s reluctance to quantify
his spatial data may be that he did not want to take the
time. Simply, he was overextended: in 1854 Snow attempted
simultaneously to investigate three separate cholera out-
breaks (Deptford, Broad Street, and South London) while
still practising as a physician, a specialist in anaesthesiology,
during a period of ferocious epidemic disease. Because he
tried to do too much, perhaps he had insufficient time to
complete the analysis in any of his study projects.
Finally, it was Snow’s habit to publish early and leave to
later papers the accumulation of additional data to bolster
an argument. He did this, as previously noted, with his
1849 pamphlet (MCC1) and in 1856, once the data were
made available, in an attempt to improve his South
London study (Koch and Denike 2006). Perhaps the real
question is not why Snow published the Broad Street
study without adequate calculation but why, given the
concerns of critics, he did not use the mapped data pre-
sented by Cooper and Whitehead to go back and finish
the ‘‘exact numerical investigation’’ later.
essential or illustrative
It is probably more important today to ask why some
modern cartographers (Monmonier 2002) and public-
health researchers (Brody and others 2000; Vinten-Johansen
and others 2003) deny the evidentiary importance of the
Broad Street map in which Snow’s fundamentally spatial
argument was lodged. Again, we speculate.
The insistence by some that Snow’s maps did not serve
important evidentiary and argumentative functions is a
reflection not on Snow’s map but on those commenta-
tors’ lack of knowledge of mid-nineteenth-century medi-
cal science and perhaps, more generally, their ignorance
of the history of maps as tools of substantive analysis and
argumentation.
More important, perhaps, is the denigration by Snow’s
biographers, who dismiss both Snow’s map and medical
mapping in general. Epidemiology and public health
are twentieth-century constructs that social epidemiolo-
gist Nancy Krieger (2000, 155) defines as ‘‘preconditional
on the emergence of quantitative population sciences’’
and the ‘‘fundamental beliefs that intimate relations exist
between mathematics and material reality.’’ This ‘‘quanti-
tative population science’’ has been relatively ignorant of
spatial data in its analysis of epidemic incidence (Gould
and others 1991). That ignorance includes both the his-
tory of disease mapping and the spatial analytics that,
since at least the nineteenth centuries, have accompanied
it (Koch 2009). Given Snow’s centrality in the mythology
of these disciplines and his failure to quantify his study, it
is, we suspect, easier to blame the map than to blame the
progenitor himself. Dismissing Snow’s map as at best an
illustrative afterthought allows them to dismiss mapping
and the spatial analytic it presents as well. Clearly this dis-
missal of the map and its critical argument serves neither
in this case nor across several hundred years of disease
study, then to now (Koch 2004, 2005a, 2009). The dis-
missal of Snow’s Broad Street map, in other words, says
more about the critics than about Snow or the science of
his day. Of course, not all public-health experts dismiss
cartographic analytics; an example is John Krieger’s treat-
ment of the history of census tracts in health research and
of modern disease mapping as a resource in epidemiology
and public health (Krieger and others 2006; Krieger 2006).
map or legend
It also may be that over the last century Snow’s map has
come to be just a graphic, that the substantive instrument
that Snow fashioned in his maps has been lost in its
various appropriations (Koch 2004). In 1901 William T.
Sedgwick redrew Snow’s parish inquiry map to emphasize
its irregular polygon for use in a textbook on ‘‘sanitary
science,’’ as public health was then called. His map, ‘‘After
John Snow,’’ was used to illustrate an inferential process
of deduction that students were to employ in consider-
ing the logical relationship between spatially grounded
data on disease incidence and environmental, principally
water-borne, sites of suspected contagion.
Since 1901 a progression of illustrations based on Sedg-
wick’s ‘‘After John Snow’’ that purport to be John Snow’s
have been constructed (see Koch 2005a). In the 1950s, for
example, geographer E.W. Gilbert published as ‘‘Dr. John
Snow’s Map (1855) of Deaths’’ a simplified version of
Sedgwick’s rendition (he quotes Sedgwick, but not Snow)
to argue the importance of medical cartography and geo-
graphy. In this map Gilbert changed the symbology and
removed streets – the essence of Snow’s topography – to
emphasize visually the centrality of the Broad Street
pump. In the 1960s Gilbert’s map was used as the tem-
plate for another ‘‘Snow’s Map’’ published by another
geographer, L.D. Stamp (1964). This map, in turn, was
the basis for another ‘‘Snow Map’’ by E.R. Tufte (1983)
that, with Stamp’s, was the basis for Mark Monmonier’s
Tom Koch and Ken Denike
28 cartographica (volume 45, issue 1)
‘‘Snow’s Dot Map’’ (Monmonier 1996, 2005). Most egre-
gious, perhaps, was the pruning of Snow’s database of
more than 100 cholera deaths in a digital version of
Snow’s Broad Street maps produced by the US Centers
for Disease Control’s EpiInfo mapping software (CDC
2000; see, e.g., Lang 2000).
It is typically these later versions, rather than Snow’s own
maps, that are referred to by many who critique (and
many who laud) Snow’s Broad Street study (see Granados
2009 for a recent example). These later versions were
graphics, not elements of a sustained research program;
they were crafted as advertisements promoting cartogra-
phy, geography, and public health as disciplines, rather
than as spatial arguments in which disease data are first
categorized and then analysed. The analytic and eviden-
tiary value of Snow’s maps was lost in the propagandistic
intent of these maps by later authors whose principal
argument has been the importance of their own disci-
plines rather than a methodology of disease study.
Snow was correct: cholera is a water-borne disease. Science,
however, is not about being ‘‘right.’’ Rather, it is about con-
vincing a jury of one’s contemporaries of the correctness of
a thesis on the basis of commonly accepted methodologies.
This is what Snow attempted and failed to do. Snow’s
failure to quantify his spatial data or to give rigorous con-
sideration to other possible sources of the Broad Street
outbreak left questions unanswered.
This was not a failure of Snow’s map, however. The map
provided a crucial evidentiary base for a spatial argument
that, given several additional days’ consideration, could
have been more convincing. Nor is it correct to lay the
blame for Snow’s failure to convince his contemporaries
at the door of mapping generally. The means by which
Snow could have answered his critics was lodged in other
maps of the same outbreak to which he had access.
Finally, to dismiss medical mapping on the basis of
Snow’s failure to adequately answer his colleagues signals
a lack of both historical acumen and contemporary
knowledge and awareness. As computerized cartographic
approaches to spatial analysis of health have advanced,
the mapped analysis of health data has become increas-
ingly important (see, e.g., Krieger and others 2006). In
precisely the same way that Snow ignored his contempo-
raries and the evidence they presented, those who insist
upon the map as simply a graphic addendum ignore the
reality of Snow’s scientific method and the importance of
the map in its analysis. To argue otherwise is to deny both
history and the wealth of contemporary studies in which
spatial arguments embedded in maps are presented in the
arena of disease investigation.
Acknowledgements
The authors wish to thank the anonymous peer reviewers
for their careful and constructive criticism of early drafts
of this manuscript. We believe their suggestions have im-
proved the quality of this paper and are obliged to them
for their attention.
Author Information
Tom Koch is adjunct professor of medical geography at
the University of British Columbia. Mailing address: 136
Hammersmith Avenue, Toronto, ON M4E 2W6 Canada.
E-mail: tomkoch@kochworks.com or
tomkoch@internet.ubc.ca.
Ken Denike is professor emeritus of geography at the
University of British Columbia, Vancouver, BC V6T 1Z2
Canada. E-mail: kdenike@geog.ubc.ca.
References
Bailey, T.C., and A.C. Gattrell. 1995. Interactive Spatial Data
Analysis. Harlow, UK: Longman: 156.
‘‘Bibliographical Record.’’ 1855. British and Foreign Medico-
Chirurgical Review 16: 148–49.
Blome, R. 1720. ‘‘The Parish of St. James Clerkenwell [as]
Taken from Ye Last Survey with Corrections.’’ British Library,
Crace Collection. Available at http://www.bl.uk/onlinegallery/
onlineex/crace/t/007000000000015u000370a0.html
Board of Health. 1855. Report of the Committee for Scien-
tific Inquiries in Relation to the Cholera-Epidemic of 1854.
London: Her Majesty’s Stationary Office [HMSO].
Brody, H., M.R. Rip, P. Vinten Johansen, N. Paneth, and S.
Rachman. 2000. ‘‘Map-Making and Myth-Making in Broad
Street: The London Cholera Epidemic, 1854.’’ The Lancet
356: 54–58. doi:10.1016/S0140-6736(00)02442-9
Brown, P.E. 1961. ‘‘John Snow – The Autumn Loiterer.’’ Bulle-
tin of the History of Medicine 35: 519–29.
Budd, W. 1849. Malignant Cholera: Its Mode of Propaga-
tion and Its Prevention. London: John Churchill.
Centers for Disease Control and Prevention [CDC]. 2000.
EpiInfo Manual. Washington, DC: US Department of Health
and Human Services.
Cooper, E. 1854. Report on an Enquiry and Examination
into the State of the Drainage of the Homes Situated in
that Part of the Parish of St. James, Westminster. Metro-
politan Commission of Sewers 478/21. London Metropolitan
Archives.
Eyler, J.M. 2001. ‘‘The Changing Assessments of John Snow’s
and William Farr’s Cholera Studies.’’ Sozial Pra¨ventivmedizin
46: 225–32. doi:10.1007/BF01593177
Farr, W. 1852. Report on the Mortality from Cholera in
England, 1848–1849. London: HMSO. (Also published as
Farr, W. 1852. Registrar General’s Report on Cholera in
England 1849–1850. London: W. Clowes.)
Frerichs, R. 2008.‘‘Mapping the Broad Street Outbreak. John
Snow: A Historical Giant in Epidemiology.’’ Available at
http://www.ph.ucla.edu/epi/snow.html
Gould, P., P.J. Kael, W. Geoff, and A. Golub. 1991. ‘‘AIDS:
Predicting the Next Map.’’ Interfaces 21/3: 80–92.
Essential, Illustrative, or . . . Just Propaganda?
Cartographica (volume 45, issue 1) 29
Granados, S. 2009. ‘‘Gra
´fica Publicada por J. Snow en 1854.’’
Periodismo Gra
´fico (9 May). Available at http://periodismo-
grafico.blogspot.com/search/label/pioneros (last accessed 9
May 2009). doi:10.1287/inte.21.3.80
Hammett, J. 1832. The Substance of the Official Medical
Reports Upon the Epidemic Cholera. Which Prevailed
Among the Poor at Dantzick Between the First of July and
the First part of September, 1831. London: S. Highley, Fleet
Street.
Hellis, C.-E. 1833. Souvenirs du chole
´ra a
`Rouen et dans le
De
´partement de la Seine-Infe
´rieure en 1832; Orne
´s d’un
Plan de Rouen, indiquant la marche de l’epide
´mie dans la
ville. Paris: Ballie
`re et Delaunay.
Hempel, S. 2006. The Medical Detective: John Snow and the
Mystery of Cholera. London: Granta.
‘‘History of the Rise, Progress, Ravages, etc. of the Blue
Cholera of India.’’ 1831. The Lancet 17/429: 241–84.
Jameson, J. 1819. Report on the Epidemick Cholera Morbus,
as It Visited the Territories Subject to the Presidency of
Bengal, in the Years 1817, 1818 and 1819. Drawn Up by
Order of the Government, Under the Superintendence of
the Medical Board. E. Balfour: Government Gazette Press.
Johnson, G. 1855. On Epidemic Diarrhoea and Cholera,
Their Pathology and Treatment; With a Record of Cases
(BL No. 001885198). London.
Johnson, S. 2006. The Ghost Map: The Story of London’s
Most Terrifying Epidemic – and How It Changed Science,
Cities, and the Modern World. New York: Riverhead.
Koch, T. 2004. ‘‘The Map as Intent: Variations on the Theme
of John Snow.’’ Cartographica 39/4: 1–13. doi:10.3138/
B123-8124-4390-5792
———. 2005a. Cartographies of Disease: Maps, Mapping and
Medicine. Redlands, CA: ESRI Press.
———. 2005b.‘‘Mapping the Miasma: Air, Health, and Place in
Early Medical Mapping.’’ Cartographic Perspectives 42: 4–27.
———. 2008. ‘‘Cholera Mapping from 1819 to 1854: Before
John Snow and the Broad Street Outbreak.’’ Paper read at
Cambridge Seminars in the History of Cartography, Emma-
nuel College, Cambridge University, 6 May 2008, Cambridge,
UK.
———. 2009. ‘‘Social Epidemiology as Medical Geography:
Back to the Future.’’ GeoJournal 74: 99–106. doi:10.1007/
s10708-009-9266-9
——— . 2 0 1 0 . Mapping Science, Making Medicine: Disease,
that Epidemic Thing. Chicago: University of Chicago Press,
forthcoming.
Koch, T., and K. Denike. 2004. ‘‘Medical Mapping: The Revo-
lution in Teaching – and Using – Maps for the Analysis of
Medical Issues.’’ Journal of Geography 103/2: 76–85.
doi:10.1080/00221340408978578
———. 2006. ‘‘Rethinking John Snow’s South London Study:
A Bayesian Evaluation and Recalculation.’’ Social Science
and Medicine 63/1: 271–83.
doi:10.1016/j.socscimed.2005.12.006
Krieger, J. 2000. ‘‘Toward a Critical Re-engagement in the
Twenty-First Century.’’ Epidemiology and Social Sciences
22/1: 155–63.
———. 2006. The Public Health Disparities Geocoding Pro-
ject Monograph. Available at http://www.hsph.harvard.edu/
thegeocodingproject/
Krieger, J., J.T. Chen, P.D. Waterman, D.H. Rehkopf, and S.V.
Subramanian. 2006. ‘‘Painting a Truer Picture of US Socio-
economic and Racial Ethnic Health Inequalities: The Public
Health Disparities Geocoding Project.’’ American Journal of
Public Health 95: 312–23. doi:10.2105/AJPH.2003.032482
‘‘The Lancet: London: Saturday, October 22, 1853.’’ 1853.
The Lancet 62/Supp. 573: 393–94.
Lang, L. 2000. GIS for Health Organizations. Redlands, CA:
ESRI Press.
Lea, J. 1850. Cholera, with Reference to the Geological
Theory: A Proximate Cause – a Law by Which It Is
Governed – a Prophylactic. Cincinnati: Wright, Ferris, & Co.
Monmonier, M. 1996. How to Lie with Maps, 2nd ed.
Chicago: University of Chicago Press.
———. 2002. Spying with Maps. Surveillance Technologies
and the Future of Privacy. Chicago: University of Chicago
Press.
———. 2005. ‘‘Critic, Deconstruct Thyself: A Rejoinder to
Koch’s Nonsense of Snow.’’ Cartographica 40/3: 105–8.
doi:10.3138/N2Q6-0582-0254-7157
Paneth, N. 2004. ‘‘Assessing the Contributions of John Snow
to Epidemiology 150 Years after Removal of the Broad
Street Pump Handle.’’ Epidemiology 15: 514–16.
doi:10.1097/01.ede.0000135915.94799.00
[Parkes, E.A.] 1849. ‘‘Review.’’ London Medical Gazette 9:
466.
Parkes, E. 1855. ‘‘Review: Mode of Communication of
Cholera by John Snow.’’ British and Foreign Medico-
Chirugical Review 15: 449–63.
Reese, D.M. 1833. A Plain and Practical Treatise on the
Epidemic Cholera, as It Prevailed in the City of New York,
in the Summer of 1832. New York: Conner & Cooke.
Robinson, A.H. 1982. Early Thematic Mapping in the History
of Cartography. Chicago: University of Chicago Press.
Seaman, V. 1796. An Account of the Epidemic Yellow Fever,
as It Appeared in the City of New-York in the Year 1795:
Containing, Besides Its History, &c., the Most Probable
Means of Preventing Its Return, and of Avoiding It, in
Case It Should Again Become Epidemic. New York: Hopkins,
Webb & Co.
Sedgwick, W.T. 1901. Principles of Sanitary Science and the
Public Health with Special Reference to the Causation and
Prevention of Infectious Disease. New York: Macmillan.
Shapin, S. 2006. ‘‘Sick City.’’ The New Yorker (e). Available at
http://www.newyorker.com/archive/2006/11/06/
061106crbo_books
Shapter, T. 1849. The History of Cholera in 1832 in Exeter.
London: John Churchill.
Simon, J. 1854. Report on the Cholera Epidemic of 1854 as
It Prevailed in City of London. London: G. Gawson.
———. 1856. Report on the Last Two Cholera-Epidemics of
Tom Koch and Ken Denike
30 cartographica (volume 45, issue 1)
London as Affected by the Consumption of Impure Water.
London: George E. Eyre & William Spottiswoode.
Snow, J. 1849a. On the Mode of Communication of Cholera.
London: Churchill.
———. 1849b. ‘‘On the Pathology and Mode of Transmission
of Cholera [Part 1].’’ Medical Times and Gazette 44 (2
November): 745–52.
———. 1849c. ‘‘On the Pathology and Mode of Transmission
of Cholera [Part 2].’’ Medical Times and Gazette 44 (30
November): 923–20.
———. 1854. ‘‘The Cholera Near Golden-Square and at Dept-
ford.’’ Medical Times 9: 321–22.
———. 1855a. On the Mode of Transmission of Cholera, 2nd
ed. London: Churchill.
———. 1855b. ‘‘Dr. Snow’s Report.’’ In Parish of St. James
Inquiry Committee, Report of the Cholera Outbreak in the
Parish of St. James, Westminster, during the Autumn of
1854. London: J. Churchill. 97–120.
———. 1856. ‘‘Cholera and the Water Supply in the South
Districts of London in 1854.’’ Journal of Public Health and
Sanitary Review (October): 239–57. Available at http://
johnsnow.matrix.msu.edu/work.php?id=15-78-56
Stamp, L.D. 1964. Some Aspects of Medical Geography.
New York: Oxford University Press.
Tufte, E.R. 1983. The Visual Display of Quantitative Infor-
mation. Cheshire, CT: Graphics Press.
———. 1997. Visual Explanations. Cheshire, CT: Graphics
Press.
Vinten-Johansen, P., H. Brody, N. Paneth, S. Rachman, and
M.R. Rip. 2003. Cholera, Chloroform, and the Science of
Medicine: A Life of John Snow. Oxford: Oxford University
Press.
Whitehead, H. 1854. The Cholera in Berwick Street, 2nd ed.
London: Hope & Co.
———. 1855. ‘‘Mr. Whitehead’s Report.’’ In Parish of St. James
Inquiry Committee, Report of the Cholera Outbreak in the
Parish of St. James, Westminster, during the Autumn of
1854. London: J. Churchill. 120–67.
Whitehead, W. 1865. ‘‘The Broad Street Pump: An Episode
in the Cholera Epidemic of 1854.’’ MacMillan’s Magazine
(December): 113–22.
Wood, D., and J. Fels. 2008. The Natures of Maps: Carto-
graphic Constructions of the Natural World. Chicago: Uni-
versity of Chicago Press.
Essential, Illustrative, or . . . Just Propaganda?
Cartographica (volume 45, issue 1) 31
