The Map as Intent:Variations on the Theme ofJohn Snow
TOM KOCH
Department of Geography / UnivlrrsiQof British Columbia / Vancouver/BC / Canada
Abstract
"Critical geographers" concerned with cartography insist
maps are first and foremost social artefacts that must be
"read" as authorial documents rather than perceived as
data statements. Their argument is typically dismissed as
trivial because the relation between intent and output has
rarely been demonstrated in a critical way. This article
seeks to demonstrate the degree to which authorial intent
defines map content and appearance through an
analysis of a single set of maps. All are based on the original
data collected in 1854by DrJohn Snow as part of his
study of the cholera outbreak in London's Soho district.
Snow's original map is included as baseline for a study
that includes versions of the Snow map by a range of authors,
including Cliff and Haggett; the US Centers for
Disease Control; Gilbert; Tufte; and Monmonier. The resulting
appropriations bear progressively less resemblance
to the original work, despite the use of the same
data set and its clear availability. The result is a cautionary
tale of the distance between maps and the data they represent.
The article also insists upon the close relation behveen
authorial intent and mapped result irrespective of
the data available. Finally, the article concludes that because
mapping is not value-free,dangers occur when professional
cartographers and geographers attempt to map
data from fields in which they are ignorant.
Keywords:cartography, cholera, epidemiology, GIS, medical
mapping, John Snow
Resume
Les << gkographes critiques ,, qui s'intiressent a la cartographie
soutiennent que les cartes sont d'abord et avant
tout des artefactssociaux que I'on doit N lire comme des
documents Cmanant d'auteurs plut6t que percevoir comTom
Koch is adjunct professor of geography at the University of British
Colun~biaand adjunct professor of gerontology at Simon Fraser
University. Address correspondence to Professor Tom Koch, Department
of Geography, University of British Columbia, 1984 West Mall,
Vancouver, BC V6T 122 Canada. Tel.: (604) 714-0348. E-mail:
tokoch@attglobal.net.
me des knoncks informatifs. On considere en gknkral leur
argument comme futile vu qu'on a rarement prouvk de
f a ~ o ncritique le rapport entre intention et rksultat. Cet
article cherche a dkmontrer le degrk auquel l'intention
de l'auteur dkfinit le contenu et la presentation d'une
carte par le biais de l'analyse d'une skrie unique de cartes
gkographiques. Ces dernieres sont toutes klaborkes a
partir des donnkes d'origine recueillies en 1854 par le
docteur John Snow, dans le cadre de son ktude sur
l'ipidkmie de cholera dans le quartier londonien de
Soho. On inclut la carte originale de Snow comme base
de rifkrence pour une Ctude qui comprend des versions
de cette carte ktablies par toute une gamme d'auteurs,
dont Cliff et Haggett, le Center for Disease Control
amiricain, Gilbert, Tufte et Monmonier. Les appropriations
qui en dkcoulent ressemblent de moins en moins a
l'oeuvre originale, bien qu'elles aient recours au m6me
ensemble de donnkes et qu'elles puissent facilement yaccider.
Cela se solde par un rkcit Cdifiant au sujet de l'kcart
entre les cartes et les donnkes qu'elles reprksentent. L'article
insiste kgalement sur le rapport ktroit entre l'intention
de l'auteur et la carte qui en rCsulte et qui ne tient
pas compte des donnkes disponibles. Pour finir, l'article
conclut que, la cartographie n'ktant pas une science objective,
il existe un reel danger lorsque les cartographes
et gkographes professionnels tentent de cartographier
des donnkes dans des domaines qu'ils connaissent mal.
Mots cles : cartographie, cholkra, epidkmiologie, SIG,
cartographie midicale, John Snow
Introduction
I
t is easy to dismiss the so-called "critical geographers"
(Lemann 2001) who write about map-making
as non-mapping, armchair theorists whose view of
cartography is social and theoretical rather than practical
or real (Harley 2001). Even when their central point
is ceded ("maps are social documents"), the result seems
a trivial point, irrelevant to the greater issues of m a p
making. With Brian Harley (1989) we can "deconstruct"
the map as a social artefact. But does his (or another's)
interpretation lead to different cartographic techniques
or outcomes? We may know, as Denis Wood writes, that
CARTOGRAPHICA,VOLUME 39, # 4, WINTER 2004
2 TOM KOCH
"what the map's about - what is really at stake - is whatever
the discourse facilitated by this pointing [mapping]
is about" (2002, 145).But unravel that, and the lesson is
that this or that map is about the subject we choose to
describe 011 its surface.
In the main, working cartographers have treated this
class of social theorists as irrelevant - or, worse, treated
them with condescension - because their conclusions do
not seem to affect the physical reality of the maps we
make. Theory, or at least relevant theory, concerns itself
with issues of symbolization and design that improve a
map's communicability. As Mark Monmonier puts it,
"I'm especially concerned that proponents of social criticism
of cartography don't really seem to be very committed
to communication" (Crampton 2002,638).
Until it can be shown that intention and the rest of social
theory matter in a concrete way to the maps that are
daily drawn, the arguments of the critics from social theory
will remain ephemeral to the greater issues of mapmaking
perceived by those, like Monmonier, who assume
that cartography is about the representation of
specific, presumably objective, phenomena. To take seriously
the stance of critics such as Harley and Wood is to
accept that mapping is, first and foremost, manipulation;
that social and conceptual prejudices and perspectives
are determining elements in mapmaking.
Generalization is always difficult. Wood and Harley
are distinct voices in the broad and sometimes crowded
field of cartographic (and geographic) theory. So, too, is
Monmonier, albeit in a different tradition. In the main,
however, the so-called critical geographers writing on
cartography can be distinguished by their rejection of
the general theory of maps as first and foremost representative
artefacts. They stand in opposition to the main
line of cartographic (and geographic) theory that the
map is a tool of concrete presentation, a method of
graphic representation of spatial relationships and forms
(Robinson and others 1984;Wood in press). For the critics,
that pretence to representative objectivity, as
Stephen Turner has put it in another context, "exists
only as a product of human activity" (1991, 23). From
the perspective of the critical geographers concerned
with mapping, maps are the externalization of subjective
decisions in a construction that owes everything to authorial
choice. "Each map is made from a particular perspective
and for a particular purpose" (Fleersink 2003,
136). The implication is that this purpose defines the
subjectivelygrounded result (Wood in press).
This article offers a case in point, one whose authorial
transformations can be traced, their importance demonstrated.
It applies the general perspective of these "critical
geographers" in cartography to one of the most
famous of all nineteenth-century maps: John Snow's
map of the 1854cholera outbreak in Soho, London (Tobler
1994).Its approach follows that of Wood (1992) in
its reliance on Roland Barthes' system of semiotic analysis.
One reason for this is Wood's position on the use of
semiology in the analysis of mapping and my own fimiliarity
with it as a tool for the understanding of text and
image (Koch 1990). Another is Wood's role as a commentator
on other "new geographers" concerned with
cartography, such as P.D.A. Harvey and Brian Harley
(see Wood 1993, 1994, 2002). The article thus serves, at
one level, as a critical study of the transformation of
Snow's work, and, at another, as both a new contribution
to the critical discourse of map-making and a practical
application of semiotics to its understanding.
The History
A series of maps of the 1854 cholera outbreak in Soho,
London - historical and contemporary - serve as an
example of the manner and degree to which a mapmaker's
intent defines the context that determines the
content of the resulting map. The maps include John
Snow's original maps; E.W. Gilbert's 1958 version of
Snow's map; Andrew D. Cliff and Peter Haggett's 1988
maps; Edward Tufte's 1983 revision of Gilbert's 1958
map; Monmonier's 1990s revision of the Gilbert-Tufte
map; and the US Centers for Disease Control's (CDC)
2001 map based on Snow, Gilbert, and Tufte. Not only
does the interpretation of the Snow map change, but the
map itself is twisted, turned, and truncated - violated,
for want of a better word - by each map-maker's mindset
and point of view. Even in its most "scientific"mode - a
GIS version by the US CDC designed for teaching purposes
- the context of its making defines the map - its appearance
and content - in ways that are demonstrably
false and inaccurate.
The cholera maps begin with the John Snow's 1854
map of the cholera outbreak in Soho, London. The image
of cholera deaths Snow produced is now invoked as
an icon so well known that it is assumed to require neither
analysis nor discussion. It serves, for example, as the
cover of a recent text titled GpographicInformation Analysis
(O'Sullivan and Unwin 2003), not as an example of
material the book covers (it does not) but because it is so
well known - so emblematic of what mapped analysis can
do - that no discussion is necessary.
That Snow's work has been mythologized almost beyond
historical redemption is generally accepted
(McLeod 2000). In this article, however, the focus is on
the mythic icon, Snow's 1854 map, and the ways it has
been distorted beyond recognition in the service of this
or that author's personal agenda. 111this way the concerns
of McLeod (2000) and Brody and others (2000) the
general myth of Snow - are extended to consider the
appropriation and transformation of Snow's mapping to
serve interests Snow himself would not have recognized
as legitimate.
THE SNOW MAP
John Snow was a medical assistant in Newcastle-area coal
mines during the first of four cholera pandemics to
sweep through England in the 1800s (Winterton 1980).
CART0C;RAPHICA. VOI,UME 39, # 4, WINTER 2004
THE MAP AS INTENT: VARIATIONS ON THE THEME OFJOHN SNOW 3
During the second great epidemic that began in the late
1840s,Snow was a physician in London, where a total of
14,600 deaths, or 6.2 deaths per 1000 persons, were recorded
(Winterton 1980). The epidemic was especially
severe in the Berwick Street area of Soho, where Snow
lived. In 1849 Snow published a long two-part paper on
cholera in the Medical Gazette and Times (Snow 1849a,
1849b) as well as a short monograph, 0 1 1 the Mode of
Trarlsmission of Cholera (Snow 1849~).
In these works Snow argued that cholera was waterborne
rather than. as most then believed, airborne. His
hypothesis was clinical rather than spatial: airborne diseases
affect the lungs, but cholera is an intestinal disorder
causing extensive and sometimes fatal diarrhoea. It
was, therefore, he argued, caused by something ingested
rather than inhaled (Snow 1849a). Snow further believed
the vector of transmission was either personal
contact with an infected person or the drinking of contaminated
water in which some materies morbi travelled
(Snow 1849c;McLeod 2000). Unfortunately, he wrote in
his review of the literature. "as we are never informed in
works on cholera what water the people drink I have
scarcely been able to collect any information on this
point" (Snow 1849b,926).
The 1854 cholera epidemic gave Snow a chance to
gather data that would redress that signal failing. In that
year he undertook two separate studies. One considered
the correlation between water sources and the incidence
of cholera in South London. Another famously examinrd
a localized outbreak in London's Soho District. The
results of both studies were published in a second and
greatly expanded edition of On theMode of 7ransmission of
C h o h (Snow 1854/1936). In that publication was the
first appearance of Snow's large-scale cholera outbreak
map, the one that, in the twentieth century, became emblematic
of the potential of medical cartography.
Snow's map was not a self-consciousexercise in analytic
cartography, however. It was instead, he said, a "diagram
of the topography of the outbreak," a graphic
summary of the elements he analysed textually (Snow
1854/1936, 45). It is useful to think of Snow as an early
ecologist seeking to describe a range of elements contributing
to a phenomenon. He was not a medical cartographer
using maps to uncover the epicentre of a
disease outbreak or the vector of a specific pathogen. Indeed,
he was not a map-maker at all.
A portion of Snow's 1854 Soho map that includes the
Broad Street pump is presented here as Figure 1. It appears
to have been a commercial map sufficiently detailed
to include the work places and other "anomalous"
locations - a brewery on Broad Street and a workhouse
on Poland Street, for example - to which Snow referred
in his text. On this map, which includes the location of
13 pumps that were the primary sources of drinking water
in the district, he marked each individual cholera
death occurring during the study period.
"All the deaths from cholera which were registered in
the six weeks from lgthAugust to 3othSeptember within
this locality, as well as those of persons removed into
Middlesex Hospital, are shown in the map by a black line
in the situation of the house in which it occurred, or in
which the fatal account was contracted" (Snow 1854/
1936,46).Snow marked each death on the map as an individual
event, rather than simply marking the houses
where deaths occurred. Where there were multiple
deaths at a single address - a frequent occurrence near
the Broad Street epicentre of the outbreak - each was
marked separately,one below the other.
In both the 1854 monograph and an 1855 report,
which included a second version of the map with 14
pumps, Snow wrote that "it might be noticed that the
deaths are most numerous near to the pump in Broad
Street" (Snow 1855, 109).This was the extent of his analysis
based on the map itself. Indeed, he warned against
assuming a simple mapped relationship between the location
of cholera deaths and simple distance from or
proximity to individual pumps. With regard to the
Rupert Street pump, for example, he cautioned "that
some streets which are nearer to it on the map are in fact
a good way removed on account of the circuitous road to
it" (Snow 1854/1936; 1849a,45-46).
Indeed, Snow was as interested in what the map did
not show as in what it appeared to portray. His text (Snow
1854/1936) pays careful attention to anomalous cases:
locations near the central pump that appeared to be
cholera free and deaths that occurred at a distance from
the epicentre of the outbreak. Despite its proximity to
the Broad Street Pump, for example, the workhouse in
Poland Street was relatively free of cholera cases: "The
Workhouse in Poland Street is more than three-fourths
surrounded by houses in which deaths from cholera occurred,
yet out of five hundred and thirty-five inmates
only five died of cholera, the other deaths which took
place being those of persons admitted after they were attacked"
(Snow 1854/1936, 42). Snow visited the workhouse
and learned it had its own pump-well on the
premises and that "the inmates never sent to Broad
Street for water." Had the mortality in the workhouse
been equal to that of the houses and shops surrounding
it - those whose water was drawn from the Broad Street
pump - Snow estimated that more than 100 inmates
would ha\e died.
Similarly, more than 70 workers employed at the
Broad Street brewer): near the epicentre of the outbreak,
remained free of the disease that affected the residents
of the area. When Snow visited, the manager told
him there was a deep, private well on-site whose water
was available to all the workers. Mostly the employees
drank either water from that well or the brewery's malt
liquor ("the men were allowed a small quantity of malt
liquor"), never from the Broad Street pump (Snow
1854/1936,42).
The map also served to locate for readers the deaths
within the district but relati\ely distant from the Broad
CARTOGRAPHICA,VOLUME 39, # 4, WINTER 2004
4 TOM KOCH
Figure 1. Detail ofJohn Snow's 1854 map ofthe Soho area cholera outbreak,from Snow (184%).The m'gznal is 415 by 384 mm.
CARTOGRAPHICA,VOLUME 39,# 4, MTNTER 2004
THE MAP AS INTENT: VARIATIONS O N THE THEME OF JOHN SNOW 5
0 yards 25
t '7
0 metres 25
0 yards 100
I 1
0 metres 100
Figure 2. Two views of Snow: (a) Cliff and Haggett (1988) used Snow'J cholera map as a tool to describe distrirting using a Lbronoi
network of Thiessen polygons. Numbers in individualpolygons represent deaths occurring within each. In his 1855 map, (b) Snow
used an irregular distance measure to create a single boundary around the Broad Street pump. Rrproduced by permission of the
authors.
Street epicentre, anomalies that appeared to challenge
the centrality of the Broad Street pump. In each instance,
however, interviews with relatives revealed that
the deceased either worked or studied near the Broad
Street pump - from which each was known to drink - or
had been with cholera patients in the Broad Street area
immediately prior to the onset of their own fatal illness.
These included, for example, the cases of a young girl
from Ham Yard (south of Brewer Street) and another
from Angel Court, off Great Windmill Street. Both typically
drank from the Broad Street pump on their way to
or from a school off Broad Street. So, too, he discovered,
did another school child from Naylor's Yard, off Silver
Street. The seemingly anomalous death of a Noel Street
boy, who lived north of Portland and east of Wardour
Streets, was similarly explained by his attendance at the
National School at the end of Broad Street.
Nor did all anomalies involve schoolchildren. Prior to
contracting cholera a tailor at 6 Heddon Court, west of
Regent Street, had spent most of his time on Broad
Street;and prior to her death, a woman from 10Heddon
Court had been nursing a Broad Street friend who also
died of the disease. Snow carefully investigated each
case;each was represented as a mark on his map of cholera
deaths.
Snow was perhaps the first to carefully use proximity
as a measurement in analysing the intensity and diffusion
of a disease at this scale. In an attempt to define a
"cholera area" for his 1855 report to parish officials, he
created an irregular boundary, centred on the Broad
Street pump, that included the majority of deaths occurring
in the Soho outbreak. "By the most careful calculation,"
Snow (1855) wrote, he had drawn a line marking
the median distance between the Broad Street pump
and the others that were nearest to it in the district. The
resulting boundary (Figure 2b) describing the area of
greatest intensity of occurrence (the "cholera area") was
an eccentric and irregular line approximating the median
distance between the Broad Street pump and other
local pumps.
This irregular line is the source of the erroneous but
entrenched myth of Snow as the originator of Thiessen
polygon analysis creating what is sometimes called a
Voronoi network.' This is central to the myth of Snow as
a father of analytic cartography. It is also a critical misreading
of Snow's maps and work. Snow neither drew a
polygon network nor used one to calculate relative rates
of death for the 13 pumps in his study area. The source
of the myth may be Cliff and Haggett's use of Snow's
data in 1988 to demonstrate how a Voronoi network of
CARTOGRAPHICA,VOLUME 39, # 4, WINTER 2004
6 TOM KOCH
CARTOGRAPHICA, VOLUME 39, # 4, U'INTER 2004
Yords
SO 0 so 100 150 200
convinced by his argument, were challenged
to consider his theory. The
Reverend Henry Whitehead (1855),
for example, was sufficiently engaged
by Snow's work - and sufficiently sceptical
- that he independently collected
data on the 1854 cholera deaths and
later became a strong supporter of
Snow's theory.
While Snow was by no means the
first to use maps to study disease - as
Gilbert (1958) and Robinson (1982),
among others, have pointed out - his
Soho map was perhaps the most comprehensive
study of a large-scale outbreak
attempted to that date. That he
was generally umuccessful at convincing
his peers of cholera's water-borne
vector does not detract from his real accomplishments.
Snow's conclusion
flew in the face of then accepted medical
wisdom, and his data and analysis
were insufficient to convince his contenlporaries
otherwise. It was not until
after Robert Koch identified the bacterium
Vibn'o chokrae in 1883 that most
physicians accepted Snow's argument
that the disease was water-borne rather
Figure 3.E.W!Gilbert'sz~ersion(withoriginal caption) ofJohn Sn,ozo's 1855 map of the than carried on the "miasma"of the air.
Soho chokra outbreak ( I 958, 174). Reproduced 4y permission ofBlnckwel1Publishing.
The Modern Turn
In the 1950s a "lively interest in rriediThiessen
polygons might be created (Figure 2a). In their cal geography" led E.W. Gilbert to publish a paper on
work, however, they were careful to distinguish between the Pioneer Maps of Health and Disease in England (1958).
Snow's original mapping and their use of his data to il- In it he reviews "early writers on 'medical topography"'
lustrate modern analytic approaches. and presents a small set of nineteenth-century maps purWhat
Snow contributed was, perhaps, the idea of dis- porting to describe the incidence or diffusion of various
tricts based on a calculation of proxirnity. His contempo- diseases. The first map in his paper is "Dr. John Snow's
raries suggested other solutions to defining the disease map (1855) of deaths from cholera in the Broad Street
outbreak's boundaries, based primarily on density of oc- area of London in September 1854." In the text Gilbert
currence. These included manually drawing boundary describes Snow as the man "largely responsible for demlines
around the affected area and a circle thatwould en- onstrating the water-borne origin of cholera" (174). Gilcompass
the majority of deaths (General Report 1855: bert's reproduction (Figure 3) of Dr Snow's 1855 map is
Whitehead 1855). Snow's insight was an important ad- used to show how the nineteenth-century physician devance,
but it was certainly not the polygonal network that fined a "cholera field" based on disease occurrence and
some contemporary writers attribute to him. To state then identified the Broad Street pulnp as the centre of
Snow's failure to develop this modern spatial analytic in the disease outbreak.
no way diminishes his importance as a critical thinker The map Gilbert presented, however, was not Snow's
whose work contributed enormously to the emerging map but, instead, one broadly based on it. Individual
fields of epidemiology and public health. deaths are marked, not by Snow's bars, but by small dots
Snow framed a hypothesis based on a clinical insight (an innovation first introduced by public health expert
and investigated it at a range of scales with all the data W.T. Sedgwick in 1914; Vinten-Johansen and others
that he could collect. He used maps in his 1854and 1855 2003); the pulnps are not circles but x's. Absent are the
reports to summarize what he described as a topography breweries, the workhouse, and the other "anomalous" loof
the localized outbreak in a way that would permit cations whose investigation was critical to Snow's study.
readers to see the landscape of the illness he described. Gone, too, are many of the roads, streets, and mews. The
His mapped approach influenced others who, while not major streets that are retained (Oxford, Carnaby, Re-
THE MAP AS INTENT: VARIATIONS ON THE THEME OFJOHN SNOW 7
Signifier Signifed
I 1
r'
4
3
2
Figure 4. The manner in which myth appropn'ates language or
images in the creation of myth. Adapted from Barthes (1983,
100 and Koch (1990, 26).
,
Sign - Myth
Signifier Signified
Sign
gent, etc.) serve less to locate the deaths than as a frame
highlighting the relationship between central deaths
and the Broad Street pump. There is little resemblance
to the environment that was Snow's mapped subject.
There were 12 water pumps identified in Snow's 1854
map, and 14 are seen in his 1855 map, but only 11 are
marked on Gilbert's map. All in all, the ecology is denuded
of myriad elements that made the original richly useful
and informative.
Gilbert's version "updates" Snow by converting the latter's
idiosyncratic Victorian image into a standard 1950sstyle
dot map. Snow's map self-consciously summarized
research that considered a hypothesis about the precise
pattern of diffusion of a specificdisease. It was dense and
complex, an artefact illustrating a detailed study. Gilbert's
abridgement - drawn either by him or at his direction
- self-consciously sought to assign to Snow a simple,
very legible correlation: many deaths, one pump, and
therefore one source for the outbreak. Gilbert's changes
to Snow's map served to emphasize that single theme.
Gilbert - or a cartographer he hired, perhaps, to do the
work at his direction - removed everything in Snow's
map that would not contribute to this goal in the creation
of his own "Snow" map. We know this was a selfconscious
revision because Gilbert cites the 1936 reproduction
of Snow's 1854monograph On the Mode of Transmission
of Cholma as well as the 1855 Snow map as the
source of his 1958 "Snow"map.
Gilbert's map of Snow's work is clear, simple, and extremely
legible. It reproduces more easily than the original,
but if ease of reproduction were the essential
purpose of Gilbert's adaptation he would have noted the
changes made and described the map as one based on
Snow's. His failure to do so forces the question of why
Gilbert presented his map as Snow's work.
The answer is that Gilbert clearly was not interested in
the complex ecological portrait that Snow crafted in his
effort to make sense of the 1854outbreak. Gilbert appropriated
Snow's map - there is no better word except,
perhaps, "stole" - so that his own would serve the then
evolving myth of Snow as the man who "discovered" the
cause of cholera through mapping a simple correlation.
In advancing the myth of John Snow as a self-conscious
and analytic map-maker (not a mapper graphically summarizing
data to help illustrate an argument) Gilbert
turns the 1854and 1855 Snow maps into an icon that falsified
history.
At this level of understanding, the distinction between
"mapper" and "map-maker"is critical (Wood 1993;Koch
1990). Snow mapped elements pertinent to the cholera
outbreak in Soho in an attempt to advance an argument
about the nature of the disease through an investigation
of the Soho outbreak. The map was not a stand-alone analytic
tool but one summarizing (and locating) a wealth
of data. Gilbert-the-map-makercreated a graphic, a map
designed to show a clear correspondence between the
incidence of a disease outbreak and a single source of
contagion. The elements of Snow's map that did not
contribute to this goal were eliminated by Gilbert to
transform Snow's complex investigation into what Roland
Barthes (1983) calls a "simple signification."
Figure 4 is a basic schematic of the process by which,
at level 1, individual marks (signifier) together create a
dedicated map (signified) generating a synthesis at level
2 of a sign (a true representation of "x"). This, in turn,
serves at level 3 as a signifier as the raw material by which
objects or words (maps, paintings, photographs, etc.) are
further transformed into general myth (level 4). The
arrow in Figure 4 shows how the level 2 sign, created by
level 1's signifier and signified, is appropriated and
transformed, at levels 3 and 4, into myth. The whole is a
second-order system in which signified and signifier operate
at the level of language (verbal or visual) to create
the signifiers and signs that become a mythic (or iconic)
statement.
At one level Gilbert's map signifies "map reveals disease
source"; globally, that "maps serve medicine" - or,
more precisely, "medical science." As a mapmaker, Gilbert
crafted a myth precisely as Barthes defines it, as a
"kind of speech better defined by its intention than its
literal sense" (1983, 103).The resulting image exemplifies
the "naturalization of the cultural" (Barthes 1972,
131) in support of a system of values that in this case revolves
around the potential of medical mapping to identify
the source of an epidemic outbreak. It represents not
a lie but an "inflection," in this case an emphasis in the
service of a view of maps as tools in the service of a type
of science (Barthes 1983). This emphasis required that
Snow's complex researches -which did not convince his
peers - be transformed into the myth of the independent
nineteenth-century researcher who mapped the outbreak
and so discovered a simple correlation proving
that cholera was water-borne. In the process, the myth
continues, Snow single-handedly introduced analytic
mapping to nineteenth-century medical researchers.
How do we know this? We have Snow's map a la Gilbert
CARTOGWHICA, VOLUME 39. # 4, WINTER 2004
8 TOM KOCH
as proof. That Snow was one of a phalanx of persons John Snow. Infinitely superior to his earlier work, this
mapping the incidence of cholera in the mid-nineteenth version bases its analysis on a reading of Snow and incentury
(Robinson 1982) is a fact politely ignored in the cludes detail of the original map rather than one based
myth-making. on Gilbert.
THE TUFTE CONTRIBUTION
Gilbert's map was reproduced as "the famous dot map of
Dr.John Snow"in Edward R. Tufte's influential The Visual
Display of QuantitativeInformation. Tufte presented Gilbert's
version as Snow's work, "an early and most worthy
use of a map to chart patterns of disease" (Tufte 1983,
24). Snow did not "chart" the patterns of disease in his
map, however: that term implies a rigour and completeness
his map did not present and Snow himself did not
claim. Because there were pockets of cholera distant
from the pump and areas near the Broad Street that
were cholera-free (the Brewery, the workhouse, etc.),
Snowwas careful not to assert that his map offered a convincing
proof of his theory or a complete chart of the
disease on which conclusions resulting from a simple relation
or equivalence (proximity to pump = deaths)
could be drawn. For Snow himself, the result was suggestive
but not conclusive.
Tufte, whose principal interest (as the title of his 1983
book indicates) is the "visual display of quantitative information,"
appropriated Gilbert's map, itself an appropriation
of Snow's map. In his text, Tufte emphasizes
Snow's map as a technical advance, but in his presentation
he uses Gilbert's map to signify not simply that
"maps serve medicine" but that the degree to which they
do so depends on the map-maker's clarity of presentation.
Tufte thus talks about Snow not as a mapper but as
a map-makerwho created a dot map to "chart" the disease
to reveal a simple correlation between cholera and
water, and between the Soho cholera deaths and the
Broad Street pump. Although Gilbert changed Snow's
symbolization, in his text Gilbert (1958) describes the
original Snow map's symbol system. Tufte, however,
writes that in Snow's map (meaning Gilbert's), "deaths
were marked by dots and, in addition, the area's eleven
water pumps were located by crosses" (1983,24).
Tufte's focus is neither epidemiology nor the mapping
of disease. Even had he been aware of it, his interest
was certainly not Snow's hard thinking about the Soho
cholera outbreak. Rather, Tufte's interest was in the visual
display of quantitative information. His intent was to
promote the best graphical portrayal of data to maximize
its utility, and for this purpose, Gilbert's simplification
served better than Snow's original. Tufte's (1983)
appropriation serves his purpose but does not reflect the
reality of Snow, the "map thinker," considering a problem
whose solution was unknown. What Tufte gained in
the process was a simple argument represented by a simple
icon; what he lost were the complicated reality Snow
considered and his efforts to understand it. Tufte implicitly
acknowledged these shortcomings in his later book
Visual Explanations (1997),which included a section on
THE MONMONIER CONTRIBUTION
Tufte's 1983book became a standard, and his appropriation
of Gilbert's appropriation of Snow became, in turn,
the standard referent for Snow's map. Indeed, the default
assumption among medical cartographers and geographers,
who frequently riproduce it, is that Gilbert/
Tufte is Snow's map.' And, since Tufte, other authors
have felt free to modify the Gilbert/Tufte icon for their
own purposes and call it "Dr Snow's."In How to Lie with
Maps, for example, Mark Monmonier (1996) uses the
Tufte/Gilbert map to create his own version of "Snow's
Dot Map" (Figure 5). In this version, however, the emphasis
is not on the many deaths but on the pumps,
which Monmonier has again re-symbolizedfor emphasis.
The vastly enlarged circles exist upon the attenuated
field of streets, whose width has been made uniform.
The width of remaining streets (e.g., Oxford Street) is altered
in a manner subtly emphasizing the centrality of
the Broad Street pump.
Lest anyone miss the point, an arrow and legend are
used to locate the Broad Street pump. The emphasis has
shifted in the process, and the point of the whole is the
importance of the pumps themselves. The point becomes
"Snow discovered the pump that caused the outbreak,"
not "Snow mapped a correlation between deaths
and water sources" or the earlier "Snow discovered
through mapping that cholera is water-borne."
At this stage in the transformation of Snow's map into
an iconic image his map-making has become an instrument
of discovery, a critical analytical tool. The global
signifier in this iteration (mapsserve medicine) is here subordinated
to the more local maps reueal disease source. By
invoking Snow in this appropriation, Monmonier simultaneously
invokes Gilbert and Tuftes' message of maps as
important analytic tools in medicine and in the study of
specific disease events. Monmonier was free to emphasize
how maps serve medicine through his version of their
map and the myth it served to promote: Snow identified
one pump out of many through brilliant mapmaking.
Certainly nothing Snow wrote - and nothing in the
data Snow collected and mapped - make the Soho cholera
outbreak inherently superior as a case study for Monmonier's
thesis in How to Lie with Maps. But by the early
1990s,the time of Monmonier's writing, the question was
notwhy but...why notusewhat had become the "famous,"
iconic map alwaysused to talk about mapmaking. Snow's
original interest had been wholly lost in the now abstract
icon at the heart of the myth. Simply, Snow's data and
map were now up for grabs, and anyone could do anything
they wished without attention to the original work
by Snow the medical thinker and researcher.
With the signifier secured (maps serve medicine) and
C:ARTOGRAPHICA, VOLUME 39, # 4, WINTER 2004
THE MAP AS INTENT: VARIATIONS O N THE THEME OFJOHN SNOW 9
1 Snow's Dot Map
Figurr 5. Mark Monmonier's vrrsion of the Tuftr/Gzlbert map (Monmonier 1996,
158).Image courtesy of University of ChicagoPress.
the potential of mapping for epidemiology encoded
(map reveals disease source),not only the original map but
also the Snow myth and the tnap icon themselves could
be dispensed with. Thus Monmonier was able to dismiss
Snow's work as largely irrelevant to contemporary epidemiology
or public health: "Real epidemiology isn't like
that, at least not in late-twentieth-centiiryAmerica. Cholera
is rare, if not extinct, and contagious diseases like
pneumonia and influenza are less troublesome than
heart disease, cancer, stroke, and numerous degenerative
ailments once ascribed to old age" (1997, 263).
But heart disease, cancer, influenza, and stroke are all
diseases that aremapped (asMonmonier's own examples
illustrate) by epidemiologists attempting to define clusters
and patterns in a manner similar to the one Snow pioneered
(US DHHS 1997). Nor is cholera any more
extinct than pneumonia and influenza, both of which remain
extremely serious diseases. In fact, at the time of
Monmonier's writing in the 1990s, the world was in the
midst of the seventh international cholera pandemic,
which began in 1961 and by the early to mid-1990s was
diffusing through the Americas (CDC 2000a). The Pan
American Health Organization (PAHO) reported 391,751
cases in the Americas in 1991 and 85,802 cases in 1995
(Arbona and Crum 1996).Extinct? Hardly.
In the 1990sthe US Centers for Disease
Control and Prevention (CDC),
the World Health Organization (WHO),
and PAHO all presented data, often in
map form, on epidemics and pandemics
of AIDS,flu (influenza), hepatitis,
and the West Nile Virus, to name several
examples. All were complex regional,
national, and global health
problems. Data for these and more localized
outbreaks of specific diseases
(e.g., meningitis and salmonella) were
(and are) frequently mapped by international,
national, regional, and local
health organizations concerned with
precisely the type of disease investigation
that John Snow pioneered in the
1850s.
THE US CENTERS FOR DISEASE CONTROI,
AND PREVENTION
Which is why the US CDC now distributes
a simple mapping program, Epi
Map, with its epidemiology software,
Epi Info 2000 (cnc 2000b). Training
modules provided by the CDC include.
not surprisingly, data from Snow's
1854study of the Soho outbreak in the
CD-ROMthat accompanies its analysis
and mapping programs.
By 2000 this version of Snow's work
was 'frequently reproduced in introductory
books on medical mapping, especially those using
GIS (Lang 2000). Unfortunately, maps based on the
CDC data (Figure 6), and described asJohn Snow's, bear
even less resemblance to the original than the Gilbert/
Tufte or Monmonier versions.
On a severely truncated, undifferentiated street network
that ,draws no distinction between wider or narrower
streets, deaths are represented by georeferenced
house location (street number and name assigned). Unlike
previous iterations, the CDC map does not represent
deaths from cholera but, instead, 206 homes at which
cholera fatalities occurred. The total database of deaths,
including multiple deaths at many of the locations, totals
only 456 deaths from cholera. The result is that 372 symbols
representing cholera deaths (578 deaths on the Gilbert
map minus 206 homes where death occurred) are
omitted. Even more egregiously, 122deaths are not only
not represented on the map but also wholly excluded
from the database. So excised are all deaths occurring at
homes situated west of Carnaby Street, east of Berwick
Street, or south of Brewer Street: deaths whose location
outside the Broad Street epicentre required Snow's careful
investigations to determine the relation between the
deceased (students and workers) and the Broad Street
pump.
CARTOGRAPHICA, VOLUME 39,# 4, WINTER 2004
10 TOM KOCH
But because this version advocates implicitly for map- street network. From these, each chose which data to
making using GIS, the message subtly changes. It is now would include and then manipulated it for his own purnot
simply that "maps reveal disease sources" or that poses. All except Cliff and Haggett called the result
"maps serve medicine" but that "CISmapping reveals dis- "John Snow's map."
ease sources" and thus, at another level, "CISserves medi- The result is a series of very different maps, each a
cine." The mapping of individual deaths is rejected in function of its author's intention. In Gibert's and then
Vlgo Streat Pump
------'7
\
- - favour
of mapping houses where deaths occurred, presumably
to facilitate the georeferencing of cases, electronically
collapsing them at a single street address.
Outliers (deaths nearest to otherpumps) are excluded to
avoid the necessity of a case-by-case analysis that, while
critical epidemiologically, is not easily persuasive graphically.
This version is all about the visual display (a la
Tufte) and analysis (a la Monmonier) of electronically
stored health data, not about the investigation of the
cause of the deaths in Soho in 1854.
This is not an inevitable consequence either of the
digitization of data or of the use of GIS (Vinten-Johansen
and others 2003). Figure 7 shows a map based on a digitized
version of Snow's data obtained in 1998 from the
Environmental Sciences Resources Institute (ESRI). Internal
evidence strongly suggests that the ESRI data set
was digitized from a map originally prepared by Cliff and
Haggett (1988). On that map, the number of pumps is
returned to the original 13 of Snow's
1855 map and the registry of deaths
(N=578) is undiminished. Absent,
however, are the institutions Snow investigated
(brewery, schools, workhouse,
etc.) and a system of street
assignments that would show their relative
width, data Snow presumably
used in making his eccentric "cholera
district." Nor is the "plague pit" that
some believed to be the origin of the
epidemic included. But then, while
the plague pit appears on the CDC map
as a transparent rectangle, it was absent
in Snow's own work, which the
CDC map purports to present and discuss.
Still, the whole serves better than
the CDC version, and most of the other
appropriations, not only because Cliff
and Haggett self-consciously distinguished
between maps based on Snow's
and those they themselves made but
because the base data of deaths,
Tufte's, the deaths - individual dots - are predominant
and the pumps almost invisible. Gilbert but not Tufte emphasized
the map-maker's ability to display a simple correlation
in an objective fashion in the service of medical
and health science. In Monmonier's version the pumps
dominate the landscape with large, circular symbols, emphasizing
the outcome of the analysis. Building on the
myths fanned by these map-makers, the CDC created an
abstract landscape whose resemblance to Snow's is minimal
and wildlymisleading. Its purpose is to serve as an advertisement
for GIS mapmaking as an instrument of
medical science. ESRI similarly sought to promote its GIS
product through a use of Snow's data -and the myth of its
efficacy - but did so in a manner that, while less visually
Figure 6. CLIC map of the Soho cholera outbreak with deaths aggregated by location pumps, and streets more closely reflect
rather than individually displayed. The map resultsfrom data prepared and mapped in own.
t h ~CDC'S EpiInfo and Epi Map programs. It is here rendered in black an,d white, with
itsfont size increasedfor kgability,from a map @ Lung (2000,15). Discussion
Gilbert, Tufte, Monmonier, Cliff and
Haggett, the folk at the CDC, and those
What results is a clear, visually precise cartographic at ESRI: all had access to data from the original Snow
product that fails by any standard of epidemiological ac- map, which included (a) deaths from cholera occurring
curacy. The CDC product builds not upon Snow's work in the Soho area in 1854; (b) the location of public water
but on the iterations of Gilbert, Tufte, and Monmonier. pumps in that district in that year; and (c) the Soho-area
satisfying,is less egregiously misleading.
Some will insist, as did one peer reviewer of this article,
that while later authors should, perhaps, have identified
the manner of their appropriation, the sin itself is
WTOGMPHICA, VOLUME 39, # 4, WINTER 2004
THE MAP AS INTENT: VARIATIONS ON THE THEME OFJOHN SNOW 11
Broad Street Pump area
0 yards 50
P
0 metres 50
Figure 7.A (;IS-generatedmap with data based on Cliffand Haggett (1988).Map drawn 4y a.uthor,f;romdata provided 4y ESH.
CARTOGRAPHICA,VOLUME 39, # 4, WINTER 2004
12 TOM KOCH
Marks On Cholera
I "'P I I
Figure 8. Illustrat~.~the steps @ which Snow's origtnal work is
appropriated to creak the myth that GZS(or mapping generally)
serves objective science. It is the rpecijic verrion of Barthes'
semiotic map, summarized inFigure 4.
trivial. "Of course they simplified in order to instruct,"
said the reviewer. "This didn't imply they were obfuscating
the message of Snow, just that his data provided a
compelling example for other points." The destruction
and manipulation of data go well beyond simplification
in the service of pedagogy, however. When the data
themselves are diminished - 13or 14 pumps reduced to
11; deaths removed from the database; the street network
denuded and homogenized; pertinent locations
like the brewery excised - obfuscation (or worse) is what
occurs.
The point is that, as M700dwrites, 'What the map's
about - what is really at stake - is whatever the discourse
facilitated by this pointing [mapping] is about" (2002,
145). Each map-maker was engaged in a different discourse;
each discourse resulted in a very different map.
These maps were not objective and ''value-free" representations
of Dr Snow's map - both the data and its
mode of representation were altered -but reflections of
the individual map-makers' intentions. Snow was interested
in creating a "topography of cholera" to summarize
a wealth of data all of which were potentially
pertinent to an understanding of the Soho outbreak. He
worked in ignorance of the specific cause of the disease
(t'ibn'o cholerne) and included as much information as he
could to illustrate his arguments. His map was designed
to assist a reader interested in the problem (the mode of
transmission of cholera) but unfamiliar with the study
area. It complemented a text that paid special attention
to areas close to the pump that were free of the disease
as well as to those distant from Broad Street where cholera
was present.
Gilbert's interest was not in cholera (or epidemiology
in general) but in "pioneers" of medical cartography, the
very idea of which serves a myth of the map-maker-ashero.
In service of this interest his map emphasized the
correlation between the incidence of disease and its
single source. This polemically advanced the potential of
map-making for epidemiology and, more generally,
medical science. Tufte built upon Gilbert's treatment of
Snow's work in a discourse about the visual display of
quantitative data and, more specifically,mapped displays
of that data. Monmonier emphasized the pumps to legitimize
his own work. The CDC map simplified things even
further to transform the myth of map-making as medical
science into one about GIS, map-making, and medical
science. The ESRI version did the same thing, although it
did not insist that the result was "Dr Snow's map" and its
manipulation of the data set~wasfar less egregious.
The result is that "Dr Snow's map," in its many modern
forms, is no longer about understanding the 1854
cholera outbreak but about the power of GIS, graphic
analysis, and map-making that is presumed to result
when map-makers work in the service of science. Figure
8 presents this in a manner similar to the general semiotic
paradigm summarized in Figure 4. The level 1 marks
on the map (points, lines, polygons) together signified
"Cholera in London." The conjuriction of signifier
(marks) and signified together became "Snow's Map"
(level 2). This was appropriated, in turn (level 3), as a
signifier unrelated to the original event, one that signified
"Science," or at least a scientific form of investigation.
The result is the myth that "GIS Serves Medicine"
(or "mapping serves medicine") in a manner that owes
at once everything and nothing to the marks Snow made
on his map as part of his consideration of the 1854 cholera
outbreak in this district of London.
Gilbert, Tufte, Monmonier, and the CDC were all free
to appropriate Snow's work and transform it into whatever
they wanted because it had become an artefact divorced
from its purpose. They did so because Snow's
map had become an icon in the public domain and they
were therefore free to do with it as they would, just as
anyone may bend a myth to his or her own individual
purpose. That is their function, after all; in myth's plasticity
and adaptability lies its attraction.
The failure here lies primarily in appropriating and
manipulating the data without acknowledging the act.
Secondarily, it is in the individual authors' insistence
(implied where not stated) that each is presenting "Dr
Snow's map" when, in fact, the resulting map is their
own work, designed to facilitate their own discourse, not
Snow's. The problem is that Gilbert, Tufte, Monmonier,
and the CDC mapmakers do not say why they left out
some details and emphasized others in a way Snow himself
would not have countenanced. But that is what happens
in myth: it becomes so naturalized, its presence so
taken for granted, that one is invited to forget that it is a
myth. The myth appropriates the work on which it is
based.
At another level, the point is that Harley (1989),
Wood (1992), and others are right: maps are as much
about what we want them to mean as they are about the
data they purport to represent. Map-making is manipulation,
a process dependent on authorial background,
intent, and perspective. Intention determines which eleC-mTOGRAPHIC,4,VOLUME
39, # 4, WINTER 2004
THE MAP AS INTENT: VARIATIONS ON THE THEME OFJOHN SNOW 13
ments of what data set will be included in a map and how
those data are symbolized for editorial emphasis. None
of these maps are value-free representations of reality,
pieces of "neutral science." All are self-consciousmanipulations
of a rich ecology. The conclusion seems unavoidable:
map-making - including computer-based mapmaking
- results first and foremost from intention. It is
the mapmaker's intent that defines the artefact that
results.
The most egregious departure from the original and
from the data it presents - is the CDC version of
Snow's work. Even in the design of a simple teaching
case, the absence of graphically troublesome but epidemiologically
critical deaths is egregious. It is an extreme
version of the maps of Gilbert, Tufte, and Monmonier,
whose relative ignorance of epidemiology, and of Snow's
cholera studies, permitted progressively radical distortions
of the work they purported to present. The CDC
version may thus be read as a cautionary example of
what happens when the myths of map-making and the
requirements of visual clarity overwhelm the mapmaker.
This consideration leads directly to the problem of
what might be called "specialized ignorance." Map-makers
lacking detailed knowledge of the subjects they seek
to represent will be prone to mistakes. Electronic mapping
is relatively easy, and with GIS, some believe that the
map-maker may stand forth as an expert scientist (Schuurman
1999), able to solve complex spatial problems,
without a deep understanding of the subject, through a
process of data manipulation and representation.
The Snow maps suggest that it isn't that easy, that spatially
related data do not "speak for themselves." Where
the relation between cause and effect - between contaminated
water and incidence of cholera - are known, then
mapping may serve as an analytic. But such problems are
essentially trivial, the outcome a foregone conclusion.
When a clear causal relationship is unknown, the problem
is returned to the level of complexity Snow confronted.
His map struggled with ecological data, the
relationship between its elements uncertain. It did not
solve the problem -"What causes cholera?"- but contributed
to Snow's thinking on the subject. Gilbert, Tufte,
Monmonier, and the CDC sacrificed Snow's brilliantly
complex and amazingly thorough thinking - which included
mapping - in service of a myth whose ultimate
message is that mapmaking is a science that, without further
research, can solve complex problems with nothing
more than the possibility of a simple spatial correlation.
Many with expertise in medical history and modern
epidemiology view modern mapping with suspicion precisely
because it advances the assumption of simple correlations
as an answer to complex phenomena. As Brody
and others put it, "One sees an echo of Snow the mapmaker
without the corresponding appreciation of Snow
the thinker in today's 'desktop mapping revolution"'
(2000,66). Until the thinking is taken into account, their
caution will be justified. But when that happens, the
myth itself will have been transformed, and "thinking
serves medicine" will be the message. Should that happen,
attention will return to Snow, not his map alone.
Conclusion
Critical geographers argue that maps are not value-free
representations of spatial phenomena but meditations
by map-makers on those phenomena. Some cartographers
have dismissed this argument because it seemed irrelevant
to the practical task of graphically presenting
specific data. Here a series of maps based on data first
collected and mapped in 1854 by Dr John Snow has
been used to demonstrate the degree to which intent
matters in cartographic representation. Depending on
the map-maker's purpose, elements have been excluded,
networks truncated, and pertinent cases removed from
the original and widely available data set. In every case
the changing map can be seen to result from the interplay
between authorial intent and the data the mapmaker
manipulates.
The conclusion seems inescapable: maps reflect specific
phenomena of interest to map-makers, who choose
from the available data to fashion idiosyncratic interpretations
of those phenomena. Each map results from the
selection of data by the map-maker from a greater set of
potentially relevant data. Map-making is not a value-free
science that somehow stands apart from social, cultural,
economic, and professional prejudices. Like all other sciences,
and other forms of exposition, map-making is
mired in the myths and assumptions of the individuals
who promote this or that map within the culture(s) the
map-makers serve.
Acknowledgements
The author acknowledges the assistance of the peer reviewers
and Cartographica editor who read and commented
on earlier drafts of this paper. He also gratefully
acknowledges the editorial review of Denis Wood, whose
comments on an earlier draft of this paper added greatly
to its final form.
Notes
1. See. for example, McLeod (2000); O'Sullivan and Unwin
(2003,931-32).
2. See Meade and Earickson (2000,447).They, in turn, cite as
a source the Snow map not by Tufte hut by Howe (1972,
178).
References
Arbona, S., and S. Crum. 1996. "Medical Geography and Cholera
in Peril." Geographers' Craft Project, Department of Geography,
The University of Colorado at Boulder. Available at
http://~~~.c~lorado.edu/geography/gcraft/warmup/
cholera/cholera.html.
Barthes, R. 1972. Mythologies.New York: Hill & Wang.
. 1983. "Myth Today." In A Bnrthes Readc7, ed. S. Sontag.
New York: Hill & Wang.
CARTOGRAPHTCA,VOLUME 39, # 4, WINTER 2004
14 TOM KOCH
Brody, H., M.R. Ripp, P. Vinten-Johansen, N. Paneth, and S.
Rachman. 2000. "Mapmaking and Myth-making in Broad
Street: The London Cholera Epidemic, 1854." Lancet 356:
54-58.
Centers for Disease Control and Prevention [CDC]. 2000a.
Weekly Epidaiological Record 31 (4 Aug.), C h o h , 1999:
249-56.
. 2000b. Epi Info Manual. Atlanta, GA: CDC.
Cliff,A.D.,and P. Haggett. 1988.Atlas ofDiseaseDistributions:AnalyticApproaches
toEpidemiologicalData. London: Blackwell.
Crampton,J.W. 2002. "Maps,Politics, and History: An Interview
with Mark Monmonier." Environment and Planning D: Society
and Space 20: 637-46.
General Report. 1855. In Repmt of the Cholera Outbreak in theParish
of St.Jams, Westminsterduring theAutumn of 1854. London:
Churchill. 16.
Gilbert, E.W. 1958. "Pioneer Maps of Health and Disease in
England." GeographicalJournal 124: 172-83.
Harley,J.B. 1989. "Deconstructing the Map." Cartographica 26/
2: 1-20.
. 2001. TheNew Nature of Maps: Essays in theHistory of Cartography.
Baltimore:Johns Hopkins University Press.
Heersink, P. 2003. Review of Lie of the Land: The Secret Life of
Maps (ed. A. Carlucci and P. Barber). Cartographica 38/1&2:
136-37.
Howe, G.M. 1972.Man, Environment, and Disease in GreatBritain.
New York: Barnes & Noble.
Koch, T. 1990. The News as Myth: Fact and Context inJournalism.
Westport, CT: Greenwood.
Lang, L. 2000. GIsfor Health Organizations. Redlands, CA: ESRI
Press.
Lemann, N. 2001. "Atlas Shrugs: The New Geography Argues
That Maps Have Shaped the World." Ntw Yorkq 9 April:
131-34.
McLeod, K. 2000. "Our Sense of Snow: The Myth ofJohn Snow
in Medical Geography." So&l Science & Medicine 7&8:
923-36.
Meade, MS., and R.J. Earickson. 2000. Medical Geography, 2nd
ed. New York: Guilford.
Monmonier, M. 1996. How to Lie with Maps, 2nd ed. Chicago:
University of Chicago Press.
. 1997. Cartographiesof Danger Chicago: Universityof Chicago
Press.
O'Sullivan, D.. and D.J. Unwin. 2003. Geographic Information
Analysis. Hoboken, NJ: Wiley.
Robinson, A.H. 1982. Ear4 Thematic Mapping in the History of
Cartography.Chicago: Universityof Chicago Press.
Robinson, A.H., R. Sale, J. Morrison, and P. Muehrcke. 1984.
Ebments of Cartograph) 5th ed. New York: Wiley.
Schuurman, N. 1999. Critical GIs: Theorizingan EmergingScience.
Cartographica36/4: Monograph 53.
Segwick, W.T. 1914. Pn'ncipbs of Sanitary Science and the Public
Health, with Special Reference to the Causation and Prevention of
InfectiousDiseases, 4th ed. New York: MacMillan.
Snow,J. 1849a."On the Pathology and Mode of Transmission of
Cholera." Medical Times and Gazette44: 745-52.
. 1849b. "On the Pathologyand Mode of Transmission of
Cholera." Medical Times and Gazette44: 923-29.
. 1849c. On the Mode of Transmission of Cholera. London:
Churchill.
. 1855. "Dr. Snow's Report." In Repmt on the Cholera Outbreak
in theParish of St.James, W~stminstqduring the autumn of
1854. London: Churchill. 97-120.
. 1936. On the Mode of Transmission of Cholera, Second
Edition. 1854.Reprinted in Snow on C h o h :Bang a Reprint of
Two Papers John Snow, M.D. London: Oxford University
Press.
Tobler, W. 1994. Snoru's Cholera Map. National Center for Geographic
Information and Analysis. Available at http://
www.ncgia.ucsb.edu/pubs/snow/snow.html.
Tufte, E.R. 1983. The Visual Display of Quantitative Information.
Cheshire, CT: Graphics Press.
. 1997. VisualExplanations. Cheshire, CT: Graphics Press.
Turner, S. 1991. "Social Constructionism and Social Theory."
S0ci0l0gzcal7'hpor:y 9/ 1: 22-33.
United States Department of Health and Human Services [US
DHHS]. 1997. Atlas of United States Mortality. Washington,
DC: National Center for Health Statistics,CDC.
Vinten-Johansen, P., H. Brodv, N. Paneth, S. Rachman, and M.
Ripp. 2003. Cholera, Chlorofmm, and the Science of Medicine: A
LifeofJohn Snozu. London: Oxford LJniversityPress.
Whitehead, H. 1855. "Mr. Whitehead's Keport." &port on the
Cholera Outbreak in theParish of St.James, Westminstq during the
autumn of 1854. London: Churchill. 120-67.
Winterton, W.R. 1980. "The Soho Cholera Epidemic of 1854."
History of Medicine 8/2: 11-20.
Wood, D. 1992. ThePower of Maps. New York: Guilford.
. 1993. "The Fine Line Between Mapping and Mapmaking."
Cartographica30/4: 50-60.
. 1994. "P.D.A. Harvey and Medieval Mapmaking: An
Essay Review." Cartographica 31/3: 52-59.
.2002. "The Map as a Kind of Talk: Brian Harley and the
Confabulation of the Inner and Outer Voice." Visual Communication
1/2: 139-61.
. In press. Introduction. In What Makes a Map a Map?
Lynchberg, VA: Center for American Places.
CARTOGRAPHICA, VOLUME 39, # 4, WINTER 2004