Historiographical Unpacking—31 August 1854
the participants, need to be unpacked if we are to be just in our assessments of actions taken dur- ing the point-source outbreak centered in Broad Street. Readers of the daily newspapers and major medical journals would have found many more theories than those I selected. It was a hot and contentious topic. Since the General Register Office regularly surveyed and assessed this litera- ture, I decided to narrow the field to the GRO’s report on the 1848-49 English cholera epidemic, Thursday, 31 August 1854 prepared under the direction of William Farr, the A normal day during a cholera epidemic in metro- examiner and compiler of abstracts and head of politan London. Figures from the seventh week of the statistical department. the epidemic, ending the previous Saturday, were Farr (1852) highlighted eight “theories available: 847 fatalities from Asiatic cholera in a and analogies,” from which I chose three fre- population greater than two and a quarter mil- quently cited by subsequent commentators who lion. Worrisome, but not as severe as the number eventually sought to explain the 1854 outbreak carried off by cholera during the seventh week of in St. James, Westminster parish (v). Farr’s the 1849 epidemic. In fact, there were hopeful multicausal zymotic theory and John Snow’s signs in the current epidemic’s weekly progress: alimentary-contagion theory are presented in one just 5 dead during the first week, followed by 26, extended scenario. The third theory, atmospheric 133, 399, 644, 729, and 847 in subsequent weeks infection, involved assumptions and reasoning (MTG 1854, 249). Was this visitation on the cusp common to those contingent contagionists who of peaking? believed that cholera could be transmitted person- to-person in extremely filthy “fever nests.” As an No such luck. To have the individuals I feature in exponent I chose the experiential-minded clini- the historical narratives engage in an exchange of cian, Alexander Stewart, M.D., assistant physician theoretical views after all hell broke loose the fol- to the Middlesex Hospital, where Florence Nightin- lowing day would have been artificial and stilted. gale tended cholera victims from Soho during the So I chose to begin the story on a day of hope in outbreak. order to discuss theories of cholera causation and Of course, first I had to confirm that she propagation that, while tacit knowledge amongst actually was there and done that.
1 Historiographical Unpacking—31 August 1854
to wait a few weeks for my next library trip and Was Florence Nightingale at the make a proper job of it, or (fatal error—the tem- Middlesex Hospital in 1854? porary measure I chose) to pop into the Health Sciences library close by the parking garage and My misadventure with Edward Cook. check their holdings. The only scholarly work on I’ll begin with a cautionary tale about Nightingale in that library was a biography by Sir how I initially deviated from my normal research Edward Cook (1913), longer in the tooth than methodology and wasted a lot of time. the the one I had consulted while researching The misadventure began with an impul- CC&SoM (Vinten-Johansen, et al. 2003), but fre- sive response to a reasonable thought. It oc- quently mentioned as a reliable work of scholar- curred to me that my historical narrative should ship. I located Middlesex Hospital in Sir Edward’s include descriptions of what was happening at the index, found the page, and read that in August Middlesex Hospital, where many cholera victims 1854 Nightingale was vacationing at Lea Hurst, were either carried or dragged themselves dur- the family summer house in Devonshire. ing the early days of the outbreak in St. James, Miss Nightingale cut short her holiday on Westminster. I had recently re-read two medical hearing that an epidemic of cholera had bro- journal pieces by Alexander Stewart in which he ken out in London. She volunteered to give described what happened at the Middlesex Hos- help with cholera patients in the Middlesex Hospital. She was up day and night receiv- pital during the “Soho outbreak,” as he called it ing the women patients — chiefly, it seems, (the mind boggles at the variety of place-names outcasts in the district of Soho — undressing for the same event). Then I vaguely recalled a them, and ministering to them. The epidem- ic, however, subsided, and she returned to biographer’s mention that Florence Nightingale her normal work in Harley Street (I:140). had served as a caregiver at that hospital shortly before departing for the Crimea. Worth looking Bingo! I copied the relevant pages with a hand into, I said to myself. Nightingale’s experiences scanner and headed home. might permit me to add a perspective from the The next day I transferred the pages from era prior to formal training of nurses. Cook’s biography to my computer and considered These musings occurred as I was walking my options. “Do nothing until you confirm Cook’s toward a parking garage on the Emory Univer- account in recent biographies,” whispered my em- sity campus. I didn‘t have time that afternoon to pirical angel; “put Flo on the back burner until you retrace my steps to the main library and under- make it back to a research library.” take a systematic literature search. My choice was “Nah,” countered my impatient angel.
2 Historiographical Unpacking—31 August 1854
Don’t be such a brick and mortar Luddite; follow residents well to the west of Broad Street during Cook’s leads with internet searches whilst the the second half of August. idea is fresh.” Point taken, there might be elec- On the other hand, perhaps Cook had tronic versions of scholarly works on the internet. meant the big one that began Friday 1 Septem- It would be two weeks before I made ber and simply gotten the date wrong. If so, how another research trip, and although I knew that would Nightingale have learned about it in a rural it would be much faster to pull more recent part of Devonshire? I looked for telling articles or biographies of Nightingale off library shelves notices in the Times via an online data-base; the than Google my way to confirmation, impatience first report of a major cholera outbreak under- won out. I spent far too much time the next two way in Soho appeared on Monday 3 September. I weeks pondering my reactions to the paragraph checked railroad routes, connections, and time- in Cook’s biography and following premature tables between London and Whatstandwell, the or inchoate lines of research. For example, I nearest railway station from Lea Hurst. The worst checked to see if Nightingale had mentioned of the epidemic was over by Tuesday afternoon, her experiences as a caregiver at the Middlesex so even if she took the train back to London on in one of her letters. Nothing came up in Hugh Monday and went straight away to the Middlesex Small’s online version of Sue Goldie’s calendar Hospital, she would have still have missed the (short summations) of Nightingale correspond- massive influx of cholera patients that appeared ence. during the previous weekend. Cook’s account But Cook had had unrestricted access to wasn’t panning out. We can all make mistakes, Nightingale’s papers when writing the biography, and perhaps I had stumbled upon one by Sir Ed- so I assumed he had come across something ward Cook. about her ministrations during a Soho cholera But mine was a mistake as needless as outbreak in August 1854. Cook’s time-frame it was dumb. I know better than to engage in didn’t jibe with the major outbreak for which the premature and stab-in-the-dark speculation like Broad Street pump turned out to be the culprit, a spinning top, bouncing helter-skelter from one but I nonetheless revisited a well-worn path topic to another. I had permitted my fascination through London medical journals for August and with the wondrous opportunities of online re- early September 1854, hoping to find something search to override a method I’ve used, and taught I had missed earlier that would substantiate others, for many decades: 1, Formulate an histor- Cook’s comments. The only suspect, an unlikely ical problem that sets up a line of research; 2, un- one at that, was a minor uptick in deaths among dertake a focused review of primary and scholarly
3 Historiographical Unpacking—31 August 1854 literature; 3, draft a preliminary thesis statement historical problem was a narrow one, so it made to guide me as I begin “writing up” the evidence sense to begin with secondary sources. selected from this literature; and 4, revise the A catalogue search yielded nine titles that thesis statement, if necessary, as my interpreta- seemed to be interpretive studies of Nighingale. I tion of the evidence evolves. Then repeat this pulled them all, found a free table in the stacks, process, over and over again, for each segment of organized the books in reverse order by pub- the entire argument (see “The Research Essay” in lication date (most recent on top, Cook on the Appendix A). bottom), and set up my laptop computer and The next section demonstrates the first portable scanner. I decided to approach the topic three steps in this method as applied to Nightin- from ignorance and allow the Secondary Way to gale’s potential involvement in the 1854 cholera inform me about the lay of this land and guide outbreak in St. James, Westminster. Jump to the me to supporting primary sources — in short, I next section if you find my method at odds with hoped for a clinching research-confirmation loop. what works quite well for you. There are many Ideally, scholarship builds on and cor- ways to do history. rects, as necessary, the work of our predeces- sors. I opened Mark Bostridge’s biography of 1. Historical problem Florence Nightingale (2008). A skim of the Was Florence Nightingale at the Middlesex preface, table of contents, notes and bibliography Hospital during the Broad Street cholera out- suggested a comprehensive and balanced study. break; if so, in what capacity? Chapter 8 dealt with Nightingale’s tenure as superintendant at the Establishment for Gentle- 2. Literature review women during Illness on Upper Harley Street, At my next research-library opportunity, London, which I read quickly along with the I undertook a review of the available literature, relevant endnotes. Suggestion confirmed; this is guided by my two-part historical problem. a serious, scholarly study. Bostridge devoted one First I had to decide whether to begin by paragraph to her “temporary leave of absence examining primary or secondary sources. Since from Upper Harley Street” as a volunteer at the eventually I would be looking at both, a blind Middlesex Hospital from 31 August until “the search of Nightingale’s published works and cor- intensity of the epidemic receded” (199). I made respondence seemed silly if other researchers had electronic copies of chapter and endnotes, jot- already found an answer that I could confirm later ted down bibliographical information on the two with a targeted search in primary sources. My sources he cited on her Middlesex Hospital so-
4 Historiographical Unpacking—31 August 1854 journ, and then turned my attention to the other then searched Florence Nightingale and identified books I had pulled from the shelves. five volumes containing selections from Nightin- I searched index and table of contents in gale’s works and letters during the time period every book for Middlesex Hospital, cholera, the that interested me. I checked them out to study Upper Harley Street institution for ailing gentle- at home. women, and the years 1853-54. I scan-copied Eventually I read and re-read everything everything I found, even seemingly trivial or I had found at the library and updated my chart insignificant comments, on these subjects; and with information gleaned from the five scholarly then preliminarily sorted my findings into four collections of Nightingale writings: three more “no categories: mentions” (Harthill 1996; McDonald 2004; Vicinus • No mention of Nightingale at the & Nergaard 1990) for a total of six; and two more Middlesex Hospital — 3 (Cope 1958; Hobbs 1997; undocumented assertions that she was at the Small 1999). hospital nursing cholera victims (Bishop & Goldie • Documented evidence that Nightingale 1962, 132; McDonald 2001, 27) for a total of nursed patients at the Middlesex Hospital dur- four. ing the St. James/Soho cholera outbreak — 4 I decided that the half-dozen no-mentions (Bostridge 2008, 199; Haldane 1931, 98; Huxley among the secondary sources did not in them- 1975, 53-55; Woodham-Smith 1950, 79-80). All selves constitute grounds to answer my histori- four cited one or two letters by Elizabeth Cleg- cal problem negatively. Two of the three were horn Gaskell from October 1854. monographs on specialized subjects, and the • Undocumented accounts that she was third was a brief biography. The fact that none of at the Middlesex Hospital during a London chol- these scholars chose to mention the 1854 London era epidemic — Cook (1913); Bishop & Goldie cholera epidemic did not seem decisive; it could (1962). simply mean that it wasn’t significant for their • Some documented evidence that Night- interpretations. Shifting to the Primary Way, I ingale was at the Middlesex Hospital, but she was disappointed that two collections of selected deliberately misled Mrs. Gaskell about what she Nightingale letters made no reference to this epi- did there — 1 (Smith 1982, 16-17). demic or the Middlesex Hospital, but, as before, All five secondary sources offering sup- I could not expect these scholars to share my porting documentation relied on Mrs. Gaskell’s research interests when they made their selec- letters; a quick search of the library catalogue tions. The absence of relevant keywords in the identified re-prints in two scholarly collections. I index of the volume on public health from Night-
5 Historiographical Unpacking—31 August 1854 ingale’s Collected Works was particularly perplex- ingale’s “speeches” to the family about nursing ing since the editor had stated, without citation, prostitutes admitted with pronounced cholera in the introductory volume that she had “nursed symptoms to the Middlesex Hospital (Chapple & patients at the Middlesex Hospital, notably ‘Soho Shelston 2000, 115). outcasts’” (McDonald 2001, 27). Were the latter I then reviewed the use of these letters in words Nightingale’s, and if so, where did they ap- the secondary sources I had consulted on Night- pear in the Collected Works? Were they the words ingale. Haldane (1931) appears to have been of an editor who did not wish to seem politically the first to print Mrs. Gaskell’s two letters nearly incorrect? Or had the editor also taken Cook at in full, including the descriptions of Nightingale his word, without the follow-up I was undertak- at the Middlesex Hospital (93, 98), although ing? O’Malley (1931) was close behind (207-08); Since five “she-was-there” commenta- either could have been the source for Woodham- tors cited Mrs. Gaskell, I shifted my attention to Smith’s (1950) uncited snippets in her biography her letters. In October 1854, Elizabeth Gaskell’s of Nightingale (79-80). It turns out that Cook first visit to Lea Hurst partially overlapped with (1913) was aware of the first letter Mrs. Gaskell the annual holiday to which Florence Nightin- wrote from Lea Hurst, and it was the likely basis gale was entitled as the Superintendant at the for his undocumented assertion about Nightingale Establishment for Gentlewomen during Illness. at the Middlesex Hospital; for he quoted other Mrs. Gaskell wrote two letters during her stay passages from this letter elsewhere in the biog- that included extracts from descriptions Florence raphy. Bostridge (2008) based “the story of FN’s Nightingale ostensibly gave her parents, sister, work at the Middlesex” on Mrs. Gaskell’s letters and house guest about her experiences at the (583). He dismisses Smith’s contention that her Middlesex Hospital. In the first, a letter started version is flawed on the grounds that Nightin- on Wednesday evening, 11 October 1854, Mrs. gale’s sister, Parthenope, confirms it in an unpub- Gaskell wrote that Nightingale began her “super- lished memoir, circa 1857 (583). intendance” at the hospital on 31 August, that a Whoops! This could be a she said, he said flood of cholera patients from “the Soho district, situation. Smith’s claim required closer scrutiny. Broad Street especially” overwhelmed her and Francis Smith’s (1982) thesis is that the rest of the staff for the ensuing forty-eight Nightingale could not have “supervise[d] the hours, and suggests she remained at the hospital admission of female cholera victims” at the at least a week (Chapple & Pollard 1966, 305). In Middlesex Hospital, as recorded by Elizabeth a second letter, Mrs. Gaskell elaborated on Night- Gaskell “later in August while both were on holi-
6 Historiographical Unpacking—31 August 1854 day at . . . Lea Hurst”; this claim is the invention own petard. First, negative evidence from the of a “titillating fabulist” (16-17). Smith selects records he consulted at the hospital archives only a passage from Mrs. Gaskell’s letter of 27 Octo- show that Nightingale was neither a regular staff ber 1854, where she summarized Nightingale’s member nor amongst those who responded to description of a constant stream of prostitutes the hospital’s call for volunteers on 2 Septem- suffering from cholera who “ ‘staggered off their ber. As to Smith’s second point, prostitution was beat’ ” to the hospital for treatment (16, citing not a registered vocation in Victorian England Chapple and Pollard 1966, 318). and it is unimaginable that any person admitting Smith (1982) then marshals evidence female patients to the Middlesex Hospital would from relevant Admission Books in the hospital have assigned such an appellation. Although this archives that disprove “this story” (16). First, hospital did normally limit admissions to pa- Smith argues that Nightingale’s name is con- tients bearing a letter of recommendation from spicuous by its absence from a report on how a subscriber, these were not normal times for the hospital coped with an unexpected volume any hospital whose catchment area included St. of cholera patients from late August through the James, Westminster and St. Anne’s, Soho. In this third week of September, 1854. Second, none instance, a simple newspaper-search would have of the female patients named in the hospital’s shown that the General Board of Health issued a Admissions Book were listed as prostitutes; ac- directive that all hospitals should admit anyone cording to Smith, Middlesex Hospital would not presenting symptoms of cholera and choleraic have jeopardized its reputation among its pri- diarrhea during this horrific cholera outbreak. mary clientele — male patients, “mostly ‘respect- Third, Smith was misled by a hair-splitting nicety able artisans’ from a neighbouring piano factory” in contemporary medical terminology. There — by admitting female prostitutes. Third, the was much disagreement about the symptoms first female admission with “ ‘undoubted [Asian] of Asiatic cholera. Whoever authored the re- cholera’ ” did not occur until 5 September, well port Smith cited had a different notion of what after the epidemic’s peak (presumably quoting constituted “undoubted cholera” than Septimus from an Admission Book). In short, Mrs. Gaskell Sibley, Middlesex Hospital registrar, who listed “story is not supported by the Middlesex Hospital four or five females (depending on the gender of archives” (16); the novelist was the unwitting a waistcoat maker) dying from “cholera maligna” source of an enduring fable that Florence Night- at the hospital on 1 and 2 September (UK, GRO ingale created for herself. 1854, 309). In short, Smith’s questionable evi- Francis Smith, however, is hoisted by his dence left me unconvinced that Nightingale was a
7 Historiographical Unpacking—31 August 1854
“titilating fabulist” or that Mrs. Gaskell’s account with Bostridge’s endorsement of Mrs. Gaskell’s is unreliable. account simply because it parallels what he found Had Sir Edward Cook inadvertently mis- in a manuscript by Parthenope Nightingale. After led Smith, as he did me? Smith (1982) has high all, Florence Nightingale’s sister seems to have esteem for Nightingale’s early biographer; “he is been at Lea Hurst throughout September 1854, accurate,” amongst a host of other accomplish- so whatever she wrote in her memoir about Flo’s ments noted in the preface (xii). In essence, doings in London then would have been second- Smith considers every subsequent commentary hand knowledge, at best. It was time to see if on Nightingale, including his own, essentially the Primary Way could sort this matter. derived from Cook. Smith follows Cook’s chro- nology, stating that Nightingale “moved” to the I was aware of a document that Bostridge does Middlesex Hospital in August 1854 and met Mrs. not cite, written by someone closer to the action Gaskell “later in August” at the family’s summer than either Mrs. Gaskell or Parthenope Nightin- house (16). By this reckoning, Nightingale would gale, which could be used to assess the reliabil- have been on holiday in Derbyshire, not at the ity of Mrs. Gaskell’s account. The Medical Times Middlesex, when cholera burst forth in St. James, and Gazette of 7 October 1854 contained an Westminster. Best I can tell, that’s why Smith article by Alexander P. Stewart, M.D., assistant thought the evidence he selected from hospital physician to the Middlesex Hospital. Stewart had archives was so telling; they reveal Nightingale prepared a report with assistance from his col- had been at the hospital in some undisclosed leagues containing “the complete statistics of the capacity but had left before this major outbreak. late fearful outbreak, so far as we have had to do Unlike Smith, I had found no reason why with it” (364). Cook’s chronology should trump Mrs. Gaskell’s. I compared a critical passage in Mrs. She wrote the letter Smith cites in October, Gaskell’s letter to Catherine Winkworth with four shortly after arriving at Lea Hurst for a visit that passages from Dr. Stewart’s article on how the only overlapped a few days with Nightingale’s Middlesex Hospital coped with cholera victims “fortnightly” holiday that ended 10 October coming from “the Soho district of St. James’s (Chapple & Pollard 1966, 305-07; Chapple and parish” during the first week of September (re- Shelston 2000, 115). That meant Nightingale produced on the next page). Mrs. Gaskell took could have been at the Middlesex Hospital during several evenings to write this letter. She states the height of the cholera outbreak and its after- that Nightingale went “on the 31st of August to math in September. But I was still uncomfortable take superintendance of the Cholera patients in
8 Historiographical Unpacking—31 August 1854
Elizabeth Gaskell to Catherine Winkworth, Wednesday evening, 11 October 1854; Lea Hurst (Chapple and Pollard 1966, 305).
Alexander P. Stewart, “Cholera in the Middlesex Hospital” (1854a, 364-65).
9 Historiographical Unpacking—31 August 1854 the Middlesex Hospital,” that an unexpected influx one of whom died and the other survived; (2) of a patients began arriving at the hospital the that the hospital discharged non-critical pa- following day, that Nightingale took part in nurs- tients on Friday 1 September to free up beds for ing them, and that she remained at the hospital the mass of new cholera patients who required long enough to see a reduction in virulence as the admission; (3) that the rush began mid-day on epidemic abated. Friday 1 September and continued for forty-eight Mrs. Gaskell composed this passage on hours; (4) that the therapeutic policy of this Wednesday, 11 October, the day after Nightingale particular hospital called for the employment of had cut short her holiday and returned to London. counter-irritants such as stupes — cloths mois- The context leads me to think that she had heard tened in hot turpentine; and (5) that very few of Nightingale describe her experiences the previous the cholera patients admitted to a hospital ward week, that is, before Stewart’s article appeared. the first week of the epidemic survived the dis- However, someone who wishes to compound ease. In short, I now felt comfortable in accepting Nightingale’s character flaws as a plagiarizing Mrs. Gaskell’s account that Florence Nightingale fabulist could claim that she cribbed from Stew- had been at the Middlesex Hospital during this art’s article if it could be shown that the 7 Octo- cholera outbreak. But in what capacity? ber issue of MTG had been sent to Lea Hurst prior Mrs. Gaskell wrote that Nightingale was to the account documented by Mrs. Gaskell. I at the hospital as a superintendent of patients, have found no evidence to that effect. which meant that her intended task was to man- There are sufficient parallels between age workers involved in patient care. Stewart is Mrs. Gaskell’s account of Nightingale’s first week explicit that “the ordinary staff of the Hospital” at the Middlesex Hospital and the selected pas- handled the unexpected influx of cholera patients sages from Stewart’s article to convince me that that began late Friday morning, 1 September, Nightingale must have worked at that hospital and lasted for twenty-four hours, at which point during the point-source cholera outbreak in St. they were so knackered that “a large temporary James, Westminster. Both authors are discussing addition was made” for an unspecified period. the same local outbreak: Mrs. Gaskell refers to Recall that Smith could not locate Nightingale’s “the Soho district, Broad St especially,” which lies name among those of regular staff members within the parish of St. James, Westminster men- and volunteers listed in a report on the outbreak tioned by Stewart. Only an “insider” in a position housed in the Middlesex Hospital Archives. Not of some authority would have known (1) that no surprising. Nightingale was not on staff; she was more than two nurses came down with cholera, superintendent of the Establishment for Gentle-
10 Historiographical Unpacking—31 August 1854 women during Illness on Upper Harley Street holiday in September, normally a slow period until mid-October 1854. She wasn’t a volunteer, since the medical students were still on summer either; Mrs. Gaskell is very explicit that Florence recess and the hospital deferred all but emergen- Nightingale arrived at the hospital on Thursday cy surgeries until their return by the beginning of 31 August, and, according to Dr. Stewart, the October. The Matron could make such an appoint- hospital did not send out an appeal for volunteers ment on her own. Subbing for one of the regular until mid-day on 2 September. What could Night- Sisters at the hospital falls within Mrs. Gaskell’s ingale have been doing at the hospital since 31 description that Nightingale went there to super- August that Dr. Stewart would have considered intend the care of patients. Nightingale’s experi- “ordinary”? ence as supervisor of salaried nurses for a year Nightingale’s experience at the Upper in what amounted to a small, private infirmary Harley Street establishment for ill gentlewomen matches what the matron of a large metropolitan only qualified her for two staff positions at a ma- hospital would expect from a locum Head Ward jor London hospital: as the Matron, who managed Nurse. domestic matters, and as a Sister, sometimes Although I had already come up empty- termed a Head Ward Nurse, who superintended handed during the literature review from a search salaried nurses. Domesticity had consumed so of Nightingale’s Collected Works about being at much of Nightingale’s time in her first year at the Middlesex Hospital during the 1854 cholera the Upper Harley Street establishment that she epidemic, I did find two suggestive summa- had recently given its steering committee notice tions by Sue Goldie (1983) in the “Calendar of that she would soon depart. So it seems highly the Letters of Florence Nightingale” (now avail- unlikely that she would have considered replac- able online, thanks to Hugh Small — URL in the ing the hospital Matron, even for a short period, bibliography for Goldie 1983; I searched for and equally unlikely that the medical and surgical “Middlesex”). In 1859 Nightingale recommended staff at Middlesex Hospital would have sanctioned the Middlesex Hospital as the best in London such a temporary appointment without a formal (FN to her mother, 20/3/59), and in 1862 she interview process (which chatty Flo would surely described it as a great London hospital (FN to have mentioned to someone in her family, and Douglas Galton, 9/11/62). Were these recom- Bostridge duly noted). mendations just based on a previous on-site Instead, I thought it very probable that investigation, part of the survey of metropolitan Nightingale took the place of one of the regular hospitals she undertook in the winter and spring ward Sisters who was scheduled for leave or a of 1854? Perhaps, but they could just as well re-
11 Historiographical Unpacking—31 August 1854 flect an intensive, personal experience during the lishment should function while Nightingale put in 1854 cholera epidemic, as described to her family some time at the Middlesex Hospital. in Mrs. Gaskell’s presence. The tone of the narrative reflects my sense that there was nothing untoward or under- 3. Preliminary thesis statement handed in Nightingale’s decision to double her Florence Nightingale was probably sub- superintendence responsibilities (as Smith, 16, stituting for one of the regular head ward nurses suggests there was). Although two rooms at the at the Middlesex Hospital when the Soho cholera establishment were set aside for her, she was un- outbreak began, and she remained in that role der no contractual obligation to live there, or even until the outbreak was essentially over. be on premises every day. The establishment’s I employ a preliminary thesis as the start- general council as well as the Ladies’ Commit- ing point for writing narratives and a lodestone tee, with whom she communicated directly, knew for choosing lines of research required to answer that she had a refuge in St. James Square, Pall questions that emerge as I write. The assump- Mall, and used it regularly; she shared the rent tion is that the preliminary thesis will surely be with an aunt who lived outside London but used tweaked, possibly discarded, depending on what the rooms when she came to the city (Bostridge happens as I write and research. If I end up with 2008, 193). something satisfactory, I revise the thesis state- Nightingale was superintendent from ment to reflect the argument imbedded in the late April 1853 until mid-October 1854, although narrative. she did not take up residence in London until the establishment, formerly at 8 Chandos Street, formally moved to 1 Upper Harley Street in early Crafting the Nightingale narratives August 1853. She wanted to select her own ma- The opening narrative episode is in two parts, tron; the committee agree on the condition that both about Florence Nightingale. It begins at the Nightingale paid the matron’s salary. She chose Establishment for Gentlewomen during Illness on Mrs. Clarke and gave her responsibility for daily Upper Harley Street. I settled on a meeting with domestic matters, maintaining housekeeping Mrs. Clarke, the establishment’s Matron, for this accounts, and control of the petty cash account imaginative scenario. No written evidence docu- (the following letters in Goldie 1983: FN to Lady ments such a happening on this day. But it seems Canning, 29/4/53; Add.MSS.45796.f.17, 3_681, historically probable that these two women would 29/4/53). Mrs. Clarke worked tirelessly beside have met at some point to discuss how the estab- her boss at first but apparently became less reli-
12 Historiographical Unpacking—31 August 1854 able thereafter (Bostridge 2008, 194); she seems 1990, 65-68; and FN to Lady Canning, which con- to have become homesick. Nightingale wrote her tains Nighingale’s specifications for rehabilitating mother at the end of July 1854 that Mrs. Clarke the private residence at Upper Harley Street into wanted to leave the establishment and return to a bona fide care-taking establishment (Add.MSS. Sheffield after completing their joint contractual 45796.f.39, in Goldie 1983, 3_690 6/1853). obligation of serving twelve months at Upper Har- Besides Mrs. Clarke, three nurses, a cook, ley Street (Goldie, 3_789 26/7/54). a male servant, and at least one female serv- Three nurses, one for each floor, handled ant assisted Nightingale at the establishment. routine care of between a dozen and a score of Of these, only Mrs. Clarke and the male servant temporarily invalided patients. Stays were sup- would have been involved in Nightingale’s de- posedly limited to two months, but they could be parture for the Middlesex Hospital. I thought the extended at the discretion of the Ladies’ Commit- male servant deserved to be mentioned by his full tee, and often were. Two medical men gave gratis name, but I did not have access to the establish- service to the establishment: Mr. William Bow- ment’s archive. However, Bostridge (2008) quotes man, assistant-surgeon at King’s College Hospital from an 1853 letter in which Nightingale men- with a residence in Golden Square; and Dr. Henry tioned retaining “John, the Cook & Nurse Smith” Bence Jones, physician at St. George’s Hospital from the Chandos Street staff (194), which I with a residence in Lower Grosvenor Street. Ei- confirmed by consulting Goldie; FN to father and ther medical man could be called in an emergen- mother, (Add.MSS.45796.f.50, in Goldie 1983, cy (Bostridge 2008, 194). I confirmed their affilia- 3_713 8/53). If this “John” had been employed tions by checking listings for “Metropolitan Hos- at the establishment in March 1851 and present pitals & Medical Schools,” Lancet (17 September when the enumerator appeared at 8 Chandos 1853):265-66 and (16 September 1854):234-35, Street, then his full name could be in the 1851 and I found their residences in the London Medi- England Census. cal Directory 1846, 17, 85. If an emergency did I used the online Ancestry Library Edition occur at the establishment while Nightingale was via Michigan State University Libraries. If you are at the Middlesex Hospital, she could have come unfamiliar with this resource, here are the steps I at the end of a shift since less than a kilometer took to navigate it to a successful outcome: separated hospital from establishment. • On the home page, I selected U.K. Cen- Other details in the imagined scenario sus Collection, then 1851 England Census. between Nightingale and Mrs. Clarke were taken • Since I did not know the surname, I had from Bostridge 2008, 188-98; Vicinus & Nergaard to find the particular enumerator’s page contain-
13 Historiographical Unpacking—31 August 1854 ing 8 Chandos Street, which meant that I had to Under Personal, I selected male, Servant as the “Browse this collection.” relationship to the head of household, and en- • In the drop-down menu for county, I tered 1811 for birth year (+/- 5). For Birthplace, selected Middlesex. I entered England, county Kent, and Deal as the • Next one must select a civil parish. The parish or place. simplest way to narrow the list is to locate the ad- • I clicked Search, and the first entry was dress on a contemporary map. I consulted London for a John Strachin who, other than the spelling, Sheet 61 of the Ordnance Survey (Alan Godfrey, matched what I had already located: 1986); Chandos Street was in the parish of St. Marylebone, situated just north of Cavendish Square. • The drop-down menu offered six pos- sible sub-registration districts, one of which was Cavendish Square, which I selected. • Twelve Enumeration Districts appeared. From here on, luck determines how long you have to slog. I began with District 1, and viewed the description page for successive districts until I found Chandos Street listed near the end of streets covered in District 7. • There are 92 enumerator pages for this district. I worked backwards and found 8 Chandos Street on Ancestry’s page 71 (Enumerator’s page 69). Barely legible on the bottom line was John Something-blurred, which I initially deciphered as John Strachim, 40, servant, born in Kent Deal. As a final check, I clicked to view the record for • I needed to clarify and confirm the enu- John Strachin, which brought me to the enumera- merator’s cursive script. I returned to the home tor’s page I had previously found. I now had a page for the 1851 England Census and typed in surname for John, the male servant at the Estab- my interpretation of the enumerator’s spelling. lishment for Gentlewomen during Illness. He de- Under Residence, I selected Middlesex for county serves to be remembered as much as his famous and typed St. Marylebone as the civil parish. employer.
14 Historiographical Unpacking—31 August 1854
Another purpose of the opening narrative other prior to making any commitments. A win- scenario is to suggest Nightingale’s probable state win opportunity. of mind on 31 August 1854 and pose an answer to the question: Why would she have wanted to 4. Revised thesis statement do something at the Middlesex Hospital while still Florence Nightingale had agreed to substi- officially superintendent of the ladies’ infirmary in tute as a head ward nurse at Middlesex Hospital Upper Harley Street? in order to suss the set-up and staff there in the We know Nightingale was a short-timer event that she decided not to pursue the pro- at the ladies’ infirmary. In early August she posed nursing superintendent position at King’s completed her commitment to serve as superin- College Hospital; shortly after she arrived, chol- tendent for a year, and in a quarterly report she era broke out in St. James, Westminster, and she gave notice of her intent to leave, probably within remained at the Middlesex, which admitted many three to six months (“To Committee of the Insti- cholera patients, until after the local epidemic had tution,” in Goldie 1983, 1854). She did not tell peaked. her bosses on the Ladies’ Committe that she had The revised thesis statement covers three already interviewed for the post of Matron/Nurs- elements: what I think happened, when, plus ing Superintendent at King’s College Hospital (FN how/why it came about. I’ll add a fourth, why to her mother, in Goldie, 26/7/54). However, after I think the experience was significant for Night- completing her investigation of London hospitals ingale, when unpacking her responses during the early in the summer of 1854, she had informed outbreak itself. selected individuals that she would consider leav- ing the institution for a post as superintendent The second imagined scenario in the opening of nursing at a major hospital willing to set up a episode, where the Matron at Middlesex Hospital training school under her administration. looks through a personnel folder, conveys infor- But relations with King’s staff during the mation about Florence Nightingale from a con- interview process had turned dicey at best, nasty temporary’s perspective. This approach seems at worst. It occurred to me that Nightingale might preferable to third-person narration, which some have considered Middlesex Hospital an alternative readers could interpret as authoritative “objective if negotiations at King’s fell through. Subbing as history.” Of course, the scenario is written by me, superintendent of a nursing ward at the Middlesex a commentator, distant in time and place, bor- would have permitted hospital staff and prospec- rowing a technique used by the Swedish mystery tive candidate alike many chances to assess each writer, Stieg Larsson (2005). Early on, the reader
15 Historiographical Unpacking—31 August 1854 receives information about the fictional journalist, Sidney Herbert actually wrote something in her Mikael Blomqvist, via a report prepared by an- behalf, however, the absence of evidence required other character, Lisbeth Salander (52-59). When me to fabricate an anonymous letter of endorse- Betty read the English translation, we discussed ment or drop the notion entirely. Since there is, the book and she singled out this section (48-57) in my view, sufficient historical probability that as a particularly interesting way to convey infor- Nightingale would have presented such a letter, mation. I decided that fabrication was justifiable as long At the time I had just begun drafting the as I did not attribute it to Herbert or any other episodes about Nightingale. I entered a square person. bracketed note [to differentiate between my views In the imagined letter of reference, and the evidence] to consider informing the read- information about Nightingale’s superintend- er about Nightingale’s background via something ence of the Upper Harley Street institute is taken being read by the Matron at Middlesex Hospital. from Bostridge (2008, 188-98), supplemented At the time I was also experimenting with differ- by Nightingale’s letter of 5 June 1853 to Lady ent ways to adapt James Wood’s notion of “free Canning (Vicinus and Nergaard 1990, 66-68). indirect speech” to historical narratives (Wood Additional background information came from 2008, 9; see my Historiographical Unpacking of Nightingale’s letter to her father and mother, in the Prologue and Appendix D). During one writing which she says she’s visited most hospitals in session, I reviewed my notes about Nightingale’s Paris (Goldie 1983, 3_691 6/1853; copy at the informal survey of nursing at metropolitan London Wellcome Institute). hospitals during the spring of 1854. It suddenly I did not have remote access to the occurred to me that an administrator such as the archives of Middlesex Hospital that would have Matron of a major hospital would have expected permitted me to locate the name of its Matron in a letter of reference before letting someone loose August 1854. Instead I located the name of the on the wards. My initial drafts of this scenario Matron at the time of the 1851 English Census: gave Sidney Herbert as the author of such a let- county Middlesex>St. Marylebone parish>sub- ter. For he and his wife, Liz, had been very inter- registration district of All Souls>Middlesex Hospi- ested in Nighingale’s findings for several reasons, tal>2 (H.O. 107, 1486, 2). Mary Jarrow is listed not the least being that as Secretary at War, Sid- as Matron, and I decided to use her, even though ney Herbert was responsible for what treatment I’m unsure if she was still matron when Nightin- rendered to British troops (Bostridge 2008, 197). gale was at the hospital. I did not find an entry When I could not find documentary evidence that for Mary Jarrow in the 1861 English census.
16 Historiographical Unpacking—31 August 1854
The narrative form: nor regular, unquoted, indirect speech where I The first sentence in the opening sce- would state who was doing the thinking either by nario — John would soon be knocking to say the name or pronoun. Instead, it is indirect style free cab was waiting — is not standard fare in histori- of flags to the source. The reader is immediately cal narration. Normally, one would encounter a transported into the mind of someone yet to be preliminary clause indicating who was doing the identified. My purpose in the opening scenario is thinking. For example, if there existed verbatim to convey the gist of Nightingale’s persona as the documentary evidence of what transpired at a administrator of a small nursing home who lacked meeting on that day between Florence Nightingale practical hospital and nursing experience. So I and Mary Clarke, it’s possible I could have em- left the superintendent nameless since the mere ployed direct, or quoted, speech such as — Flor- mention of Florence Nightingale might conjure up ence Nightingale said, “John will soon be knocking “lady with the lamp” images from her later work to say the cab is waiting.” If the extant evidence at the British army hospital in the Crimea. The was sufficiently detailed but short of verbatim, it remainder of the opening paragraph is in authorial would have been historiographically valid to use indirect style to indicate that the opening sen- indirect speech reported by myself as author, tence is from a female’s perspective as she listens based on that documented account (sometimes to the matron’s complaints. referred to as authorial style) — Florence Night- I begin the second paragraph in authorial ingale looked at the clock and thought, John will style to set up what the superintendent will say soon be knocking to say the cab was waiting. If thereafter via free indirect speech: the evidence was only suggestive on this point, I After a few moments the Superintendent could have made the indirect speech conditional began speaking calmly to the Matron. Every- — Perhaps Nightingale glanced at the clock and thing is sorted. . . . wondered if John would soon be knocking. Each of these examples flags an author as the source The superintendent’s instructions are presented of the speech: Nightingale in the first instance, without recurring indicators of indirect speech myself as the author/narrator in the next two such as “she said” or “she continued” because (see Appendix D for Wood’s explanation of these the opening sentence is sufficient to identify the styles, which I have borrowed). speaker and clarity isn’t advanced by flagging My first sentence, however, is free indi- myself as the author. rect speech or thought. It’s neither direct speech The rest of the opening scenario and most by Nightingale (there are no quotation marks) of the narrative in the second scenario where the
17 Historiographical Unpacking—31 August 1854
Matron of Middlesex Hospital, Mary Jarrow, re- from documented evidence which could also be views Nightingale’s file is presented in third-per- presented as indirect authorial style alone, with- son, indirect style reported by myself as author. out narration. I shift between such authorial indirect narration and free indirect thought to suggest Nightingale’s Additional documentation state of mind whilst Jarrow sits opposite her. Sometime in 1853-54, Both of the opening imaginary scenarios Nightingale sat for a photo- featuring Nightingale are historiographical exten- graphic portrait by Kilburn of sions of documented primary evidence confirmed Regent Street (Bostridge 2008, in my review of the literature. The evidence de- xii, plate 21). I scanned a copy cides. Follow it carefully, explicate it transparently, of Bostridge’s illustration (at and you have the makings of historical interpreta- right) and used Photoshop to tion. Get ahead of it and you have the makings, crop the portrait, eliminate the at best, of historical fiction. Once Smith (1982) shadow under her mouth, and challenged Mrs. Gaskell’s account of Nightingale’s simplify the woodwork of the doings at Middlesex Hospital, subsequent histori- chair for the setting — a hospital ans are encumbered to justify using it. Bostridge matron’s office. (2008) found confirmation in an unpublished The map insert show- memoir by Parthenope Nightingale, which was ing the corner of Upper Harley and Weymouth unavailable to me. Stewart’s report (1854a) con- Streets comes from Reynold’s Map of Modern tains details about happenings at the Middlesex London (1859), uploaded by Ralph Frerichs to Hospital during the cholera outbreak that are the web site he created and manages (
18 Historiographical Unpacking—31 August 1854 today is Brunswick Place on Reynold’s map. the writings of Thomas Sydenham (1624-89). The new premises for the Establishment He coined an expression, the “epidemic constitu- for Gentlewomen during Illness was a three-story tion,” for a congeries of environmental changes, house, plus attic and basement, at 1 Upper Har- chiefly seasonal atmospheric conditions, variable ley Street; a stable was in the rear. Today, this soil and vegetation conditions, and temperature property is numbered 90 Harley Street (Bostridge changes, believed to interact with stagnant bod- 2008, 191). ily humors to produce diseases not endemic to a locality. In other words, Sydenham used consti- Cholera Theories tution to mean two complementary things: the While roughing out a draft of Snow’s oral-fecal peculiar and ultimately unknowable combination version of the doctrine of contagion, I couldn’t of environmental factors that generated specific get William Farr and Alexander Stewart out of epidemic diseases; and the degree of physical my mind. Why those two chaps? Neither was a vitality of the individuals in affected localities, contagionist? And then it struck me how histori- interpreted essentially within the Hippocratic cally unfair and inaccurate it would be to ignore humoral framework. Everyone in a given locality two contingent contagionists who disagreed with with an “epidemic constitution” was exposed to the bulk of Snow’s views on the cause and trans- its morbid influence, but predisposition deter- mission of cholera during the 1854 epidemic. I mined which individuals came down with physi- decided to present Farr’s ideas as an explanatory ologically disruptive fevers and the extent of their interlude during a meeting with Snow on August infirmity. 24, and Stewart’s in a hypothetical meeting with The “epidemic constitution” was a con- Matron Jarrow and Nightingale later the same geries of unknown causal factors, beyond full day. The narratives and explanatory commentar- human understanding. Therefore, Sydenham ies about all three historical figures incorporated believed that medical men only had one option several predominant theories of epidemic dis- when epidemic diseases appeared: employ expe- eases in England during the middle third of the rience at the bedside and pharmacological em- nineteenth century. piricism in search of remedies effective for each unique disease. Sydenham’s prototype for dis- The epidemic constitution ease specificity was ague (later renamed benign The most commonly held view of cholera tertian malaria) since judicious administration as a disease amongst English medical men and of ground cinchona bark seemed to cure ague sanitarians was anticontagionism, associated with but have no beneficial effects in other epidemic
19 Historiographical Unpacking—31 August 1854 diseases (actually, it only suppresses symptoms). “epidemic constitution” model in reports sent Historiographically, it was not necessary from India. They pointed to descriptions by army for me to analyze Sydenham’s seventeenth- surgeons of instantaneous outbreaks, with the century notion, “epidemic constitutions of the greatest number of victims occurring in the first season,” in order to introduce the three major few days; cholera suddenly appearing in areas theories in play during the mid-nineteenth cen- where it had never been endemic, including local- tury. But it was necessary to introduce Syden- ities infamous for deleterious marshes; outbreaks ham. I located sufficient background information that could not be traced to human contact but on this esteemed clinician in two readily accessi- could be associated with the movement of pre- ble books: Porter (1997, 230) and Hamlin (2009, vailing winds; and most tellingly, medical person- 156-57). Hamlin, 2002 was especially valuable nel treating cholera victims rarely, if ever, came in helping me understand how Sydenham’s dual down with the disease themselves. Such evidence usage of constitution (epidemic and predispos- strongly suggested that cholera was a typical epi- ing) remained central to the thinking of various demic disease; the fever that occurred in people non-contagionists in the mid-nineteenth century. susceptible to its peculiar “epidemic constitu- I did need to investigate what mid- tion” never eventuated in a morbid, infectious nineteenth century medical men thought was still virus that could produce cholera when inhaled useful in Sydenham’s writings on epidemic fevers by healthy people (note the different meaning of and the treatments he recommended. I consider virus from that currently used in biomedicine). Gavin Milroy’s review (1847) especially helpful A few observers in India, however, in this respect. Rothstein (1972, 28-29) explains weren’t convinced that cholera was a common why Sydenham and his followers considered epidemic on the Sydenham model. There were cinchona bark a valid therapy for intermittent no seasonal differences in virulence; cholera ap- fevers. peared whether it was hot or cool, wet or dry. Critics of atmospherically induced epidemics also Asiatic cholera cited positive evidence that cholera was a conta- Cholera, long endemic in parts of India, gion. They offered evidence of outbreaks that had began a pandemic progress in November 1817 progressed from village to village, often in direc- near Calcutta that would eventually spread west- tions contrary to wind movements, and never ap- ward to Bombay, south to Ceylon, and as far east pearing in two distinct parts of a province at the as China. Many English surgeons and physicians same time (as one would expect if cholera were spied signature characteristics of Sydenham’s spread by something general in the atmosphere).
20 Historiographical Unpacking—31 August 1854
Cholera only occurred in villages after the arrival England in the autumn of 1831. Among primary of persons from areas where the disease already sources, I recommend Corbyn and Blaine’s entry prevailed, whereas isolated villages escaped the in Medico-Chirurgical Transactions available on scourge entirely. Some hospitals reported that JSA&RC. The Times, the Lancet, and the relatively most, sometimes all, medical attendants and newer London Medical Gazette, if accessible, are caregivers became ill with cholera. That could only especially helpful in assessing the general mood in happen if pathological processes followed the con- England as the pandemic of Asiatic cholera came tagion pattern in which a morbid virus produced in ever closer. victims finds fresh material for its reproduction in healthy people who are in close contact with the Contingent contagion ill. In 1831 another pandemic originating in But how does it pass from person to per- India brought widespread fear that Asiatic chol- son? So far, there was no evidence of the strict, era would reach England for the first time. Con- contact-based contagion that occurred with small- tagionists saw an ineluctable, north-westward pox. Too many caregivers had touched cholera progress of outbreaks which they interpreted as victims with impunity to invoke direct contact as proof that any place where humans travelled was a mechanism of communication, and the evidence susceptible; why should England be an excep- for indirect contact via fomites (clothing, bedding tion? Anticontagionists saw telltale geographical and other objects that had been in contact with gaps among the outbreaks, undeniable proof in diseased victims) was inconclusive, at best. Some their minds that Asiatic cholera was an epidemic wondered if the comparison with smallpox was disease spread by prevailing winds, not human misguided. The mortality from cholera was signifi- contact; its epidemic constitution would determine cant, but hardly comparable to smallpox. Perhaps if England was affected, and if so, its virulence cholera was a minimally contagious disease, like would likely be limited to those already predis- typhus; that might explain why it was so deucedly posed to epidemics of Sydenham’s cholera mor- difficult to trace its progress via specific individu- bus (summer diarrhea). als in many instances. James Johnson, however, an Irish born CC&SoM (165-72), Hamlin (2009, (for London physician, thought the two warring camps detailed context), and Durey (1979) offer sec- had ground in common. In June 1831, he sought ondary overviews of early British responses to to stiffen the British spine as rumors of an inevi- the first cholera pandemic that began in India table calamity of contagion roiled the body public. in 1817, as well as the second which reached He argued in a letter to the Times that Asiatic
21 Historiographical Unpacking—31 August 1854 cholera did have some signature characteristics of formidable in an English climate” (June 1831). a classic epidemic disease. Like other epidemics Many medical men considered Johnson’s on the Sydenham model, Asiatic cholera probably notion of contingent contagion nonsensical. For sprang from unknown chemical reactions in the them, Asiatic cholera was either contagious or it bowels of the earth; the disease seemed coun- wasn’t. Anything else simply gave a fancy name ter-intuitively pervasive when conditions were to medical uncertainty. But for this historian, optimal, whereas actual outbreaks were very contingent contagion is a viable interpretive cat- localized; and normally, the disease was non- egory for analyzing views on cholera transmission contagious and limited to predisposed individu- in England at least until the 1866 epidemic. For als. However, Johnson spied credible evidence of Johnson’s notion of contingent infectious effluvia person-to-person transmission where people lived jarred the neat, oversimplistic opposition between in especially filthy conditions at low elevations. In contagionists and anticontagionists, explains the such circumstances, the bodies of cholera victims rift that occurred in the anticontagionist camp produced a morbid material that was infectious; during the 1832 epidemic, and cleared the way effluvial vapors from the exhalations and skin for a detente with contagionists that evolved into of the ill suffused the room where they lay and the sanitary movement. anyone in close proximity for an extended period I have posted two of James Johnson’s let- of time was at risk of contracting cholera. A wor- ters to the Times on JSA&RC and a letter he sent risome, but very manageable problem, according to the editor of the Lancet; see the Bibliography to Johnson. With “proper attention to cleanliness, for details and URLs. The Westminster Medical ventilation, and separation,” contagious infection Society debates (1832, two listings) mentioned in can be halted and Asiatic cholera returned to its the Bibliography offer insights into the disagree- typical, non-contagious nature (June 1831). ments among contagionists and anticontagionists, Johnson repeated his views in October including contingent contagion, in the midst of 1831 that cholera is not contagious unless cer- the first English epidemic. I’ve posted samples of tain environmental contingencies existed. The contingent contagionists on JSA&RC who be- following month, a critic in the Lancet coined the lieved that the infectious nature of Asiatic chol- term, contingent contagion, which stuck. It did era depended, at least in part, on environmental so, perhaps, due to an ironic incongruity between circumstances and/or individual predisposition: Johnson’s erroneous predictions about the course Watson (1842); Fourcault (1849); and Fretten- of the disease and the actual outcome: “I ven- bacher (1849). ture to prophesy that it [cholera] will never prove
22 Historiographical Unpacking—31 August 1854
Infection as medium of contagion the rule rather than the exception. Environmental Contagionists who were troubled by the contingent contagionism gave a huge boost to an absence of evidence that cholera was directly incipient sanitarian movement that, in alignment transmissible on the smallpox model, by touch with Benthamite reformers, sought to improve or fomites, could embrace the intermediate no- the condition of the laboring classes, especially in tion of contingent contagion because infection rapidly industrializing towns and cities, via mas- assumed disease progress by person-to-person sive social engineering projects. Cholera, the new transmission of morbid material. Infection, the modern plague, was more intractable than small- notion that contagious matter entered the body pox; it could not be tempered by inoculation. But via the lungs, was the oldest proposed medium sanitarians believed there was a socio-economic of contagion, overshadowed in recent decades by vaccination available for the scourge of cholera; the skin because of inoculation and vaccination. reduce, perhaps even eliminate, its capacity to Although cholera infection and smallpox differed infect by preventively improving the living con- in the mode of communication, they had a com- ditions and changing the habits of the humans mon contagious signature: a morbid virus in both essential to its progress. diseases was produced in the victims’ bodies and Writings by strict contagionists, who be- transmitted to others. Infection in Asiatic chol- lieved that Asiatic cholera was always contagious era occurred when the virus emanating from a and only transmitted by human contact, posted victim’s skin and breath was inhaled by healthy to the JSA&RC include: the Editors of the Lancet persons, in whom the virus found fresh material (1842) on three modes of contagion; Copland for its reproduction and propagation. and the Editors of the Lancet (1846) on infectious Indirect transmission by atmospheric nature of pestilential cholera; James (1848) on infection had the added advantage of explaining communicability by infection; and the Editors of cholera’s progressive mode of communication LMG (1849) on diffusion of cholera by fomites. when direct contact could not be established, as well as contradictory evidence about the fate of Local miasmatists health care workers. If cholera was not an ex- During the 1831-32 English cholera tremely contagious disease, the threshold for epidemic, an increasing number of anticontagion- infection would be higher than most medical men ists found the general atmospheric explanation encountered during routine treatment of victims. at odds with their experience in local outbreaks, Contagionists could find common ground just as Johnson had in India. The usual suspects with local miasmatists simply by making infection for morbid sources quickly shifted from vegeta-
23 Historiographical Unpacking—31 August 1854 tive miasmata carried by prevailing winds to local Formerly healthy persons come down with cholera concentrations of miasma and, especially, effluvia after inhaling morbid matter produced in the bod- — the invisible, fetid vapors generated by decom- ies of cholera victims, a classic medium of conta- posing animal products and by-products, includ- gion. ing excrement. Whether or not a healthy person In short, local miamatists accepted the contracted cholera depended on the miasmatic argument for contingent contagion (whether or effluvial virulence of an offending place and or not they would salute Johnson’s flag), partly constitutional susceptibility to any epidemic fever. because it considered infection a rare excep- Typical Asiatic cholera was non-contagious; those tion, partly because it preserved the theory of who contracted it did so from something morbid “epidemic constitution.” An “epidemic constitu- about a particular place, not a fellow human. tion” could be retained as an unknown first cause A rare exception proved that rule. Expe- of epidemic cholera. Thereafter, environmental rience suggested that the cholera poison could circumstances determined the particular mani- transmogrify into an infectious virus in some festations of the disease — usually a non-con- victims who inhabited confined spaces. Certain tagious fever, exceptionally an infectious fever. environmental circumstances seemed especially For Sydenham loyalists, Johnson’s compromise conducive to initiate this pathological process in accounted for local cholera outbreaks in areas bodily constitutions. Again, the evidence suggest- contrary to prevailing winds, as well as excep- ed to this group of medical men and sanitarian- tional mortality in confined, overcrowded spaces minded lay persons that poverty, overcrowding, with no immediate source of organic decomposi- poor ventilation, low-lying habitations, and filth tion. Contingent contagion would explain why set in motion inscrutable changes in the bod- occasionally medical attendants did succumb after ily economy that heightened susceptibility to treating victims in fever nests or poorly ventilated cholera. Moreover, adults inhabiting such fever dispensaries over long stretches of time. Infection nests frequently compounded their situation by was fortunately the rarest contingency. engaging in debilitating habits (especially exces- The following examples of local mias- sive drinking of distilled liquor) and consuming matists who believed that Asiatic cholera is one an unwholesome diet. The most ardent of non- disease with local manifestations, perhaps includ- contagionists had to admit that in very enclosed ing rare instances of infection, are posted on the places, only a pathological explanation could JSA&RC: Laycock (1846) and Clark (1846). Also account for the massive morbidity and mortal- posted are examples of the increasingly outmod- ity from Asiatic cholera. Infection had occurred. ed general miasmatist notion among some anti-
24 Historiographical Unpacking—31 August 1854 contagionists and strict interpreters of Sydenham atmosphere when the meteorological conditions who believed that the “epidemic constitution” did were opportune. In Farr’s theory, elevation deter- apply to Asiatic cholera: A pithy formulation by mined the maximum dosage of poisonous London the pseudonymous Ero (1837); the editors of the fog one was likely to inhale; low-lying areas (one Medical Times in 1848; and Hingeston (1855) on of Johnson’s prime contingencies) environmen- the collateral science of meteorology and epidem- tally magnified individual susceptibility. Once ic cholera. inhaled, the zymotic material could (in susceptible persons) alter the blood, reproduce itself, and be Pragmatic contingent contagion passed to others as an infectious contagion. The 31 August narratives seek to validate William Farr, employing a fascinating the theoretical integrity of two dominant epidemic blend of Parisian medical and Benthamite positiv- disease theories in mid-nineteenth century Eng- istic influences, developed a sophisticated statisti- land. The first of these is Farr’s zymotic theory, cal nosology of diseases in the early 1840s. Later, straddling local non-contagion, contingent conta- he used mortality data from the 1849 London epi- gion, and infectious contagion. The second, and demic to develop an elevation theory of causation more mainstream at the time than Farr’s theory, based on distribution of cholera-ladened vapor is Alexander Stewart’s localized non-contagionism from the surface of the Thames, the very same with a rare contingent-contagious exception for data Snow had used to implicate impure water improperly ventilated and unsanitary spaces. supplied by private water companies. In August 1854, Farr’s zymotic and elevation theories posed Farr’s zymotic and elevation theories in the 1840s arguably the most significant challenge to Snow’s are elaborations of Johnson’s original notion. interpretation of increasing cholera mortality in Farr believed that the “epidemic constitution” the low-lying South London districts. Consequent- conducive to the generation of cholerine varied ly, it deserved inclusion in my cholera narratives. from place to place. Temperature, humidity, wind But how detailed should I make it? direction and speed, barometric pressure, and Farr was tangential to my main event precipitation were foreseeable contributors. In since he considered the local outbreak that oc- the London metropolis, Farr proposed that hu- curred in the higher-elevated residential areas of man sewage deposited into the River Thames St. James, Westminster and St. Anne’s, Soho an was associated with the production of poisonous exception to his elevation theory; being a multi- zymotic material which formed a bond with harm- causalist, he was quite willing to accept Snow’s less miasmata, became volatile, and rose into the explanation for this particular event. On cholera
25 Historiographical Unpacking—31 August 1854 mortality in low-lying South London, however, It isn’t important for my purposes, for example, Farr still held the interpretive advantage in 1854, whether Farr’s interpretation of Liebig was accu- and he seemed to think Snow’s SoLo inquiries rate or even fair. I just had to present what Farr would ultimately strengthen the elevation theory. thought about a very specific issue. Since I was limiting my analysis of the SoLo The inserts with two elevation tables and project to explanatory context for Snow’s involve- the equation are based on Eyler (1973, 89); Ey- ment in the Broad Street pump outbreak, explica- ler’s formulation summarizes and very much sim- tion of Farr’s theory should also be limited. plifies Farr (1852, lxii-lxiv [sequence 78-80]). The A narrow search of scholarly literature on quotation that cholera mortality varies inversely Farr yielded a book and several articles by John with elevation, is from Farr (1852, lxix [sequence Eyler. I began with the article (1973) that seemed 85]). The quotation in which Farr traces the no- most likely to give me the overview I was hop- tion of zymotic diseases to Sydenham appears in ing for. It was almost forty years old, but that Farr (1842, 121; reprinted in Farr 1852, lxxxii- wouldn’t matter for my purposes if the research lxxxiii [sequence 98-99]). And the quotation was solid. I tested Eyler’s trustworthiness by where Farr asks questions about the production comparing the article with a few primary source and propagation of cholerine is taken from Farr references he cited. All was well. (1852, lxxx [sequence 96]). I was lucky. A twenty-two page article contained most everything I would need to sum- Alexander Stewart, Assistant Physician to the marize Farr’s views of relevance to my topic. Middlesex Hospital, was also a local miasmatist After roughing out this portion of the narrative who believed in the possibility that cholera could scenario, I compared what I had written with be infectious under certain circumstances. Two of Eyler’s book (1979) published after the article, his publications — a letter to the editor of MTG made a few revisions that reflected new thinking, and the follow-up report about the hospital’s do- used two primary sources (Farr 1842, 119-22 for ings during the outbreak (extracts already pre- adaptation of Liebig to zymosis; and Farr 1852, sented in my analysis of Mrs. Gaskell’s account) lxi-lxv for the elevation theory) for additional de- — contain signature characteristics of contingent tails, checked each index entry for Farr in Hamlin contagionism. In his mind the localized outbreak (2009), and was done. Mind you, I only needed was caused by something “inscrutable” (1854) enough information to establish Farr’s zymotic that first appeared in a very contained part of theory and his take on the natural experiment in the parish of St. James, Westminster in the latter South London that he was facilitating for Snow. part of August. When the full eruption occurred, it
26 Historiographical Unpacking—31 August 1854 was near-instantaneous and pervasive in the local- a knight’s move in his thinking about the mode of ized area. More than 250 patients with symptoms communication of Asiatic cholera. Experience at of cholera and choleraic diarrhea were admitted Killingworth in 1832 had convinced him that chol- to well-ventilated wards in which the beds were era was an infectious contagion. Men came out widely separated from each other. Stewart believed of the pits, retching and with the trots; it seemed that this layout and a generous food allowance for more reasonable to assume they had inhaled hospital staff during the height of the epidemic effluvia given off by fellow miners than mias- largely prevented infectious transmission; only two mata carried by prevailing winds. Snow’s teetotal staff members contracted cholera during the entire address from 1836 and his last known com- month of September. One nurse ignored premoni- ments about cholera (21 October 1848) indicate tory symptoms for half a day and became too ill to that, during his early years in London as student save. The other recovered fully from severe diar- and general practitioner, he considered chol- rhea (Stewart 1854a, 365). Stewart does not explain era a febrile disease, regardless of subsequent why, given such policies, a disease he considered symptoms. That notion, as well as the ostensible typically non-contagious felled either staff member. parallel with asphyxia that Snow articulates at He emphasizes that the hospital had institutional the meeting of the WMS mentioned previously, fit guidelines for ward cleanliness, ventilation, and spa- well with Copland’s definition and description that tial separation of patients (Johnson’s preventatives cholera is “infectious . . . not by contact, but from for avoiding contingent contagion). By implication, the inhalation into the lungs, along with the air, of perhaps he believed that the chaos engendered by the morbid effluvium given out from the body or the overwhelming number of admissions during the bodies of the affected” (1846, 517). horrific Soho outbreak resulted in an unusual situa- Nine months later, the opening pages of tion where two nurses were in close contact for too MCC show how Snow advanced the contagion- long, with too many patients, given their individual ist argument with respect to cholera by rejecting predispositions. infection and moving sideways to an alternative contagionist mode of communication. He hopped Snow on cholera over contact (the acclaimed smallpox model) and As far as his contemporaries were concerned made a case for the alimentary canal via patho- in the summer of 1854, Snow had yet to make a logical analogies with other diseases; yet he still compelling case in support of what many considered gave contact an indirect role if cholera victims or an interesting but very peculiar theory. caregivers handle food without taking sanitary During the winter of 1848-49, he had made precautions. It was, by anyone’s measure, a bril-
27 Historiographical Unpacking—31 August 1854 liant inductive-hypothetico-deductive gambit. interest, however, the Secondary Way (reliance But the confirmation process since 1849 on existing interpretive studies) would certainly had been less than brilliant. Many of Snow’s ex- be a place to start. I’m still comfortable recom- amples foundered on probabilistic shoals in an era mending CC&SoM as a starting point for such a when most medical men and medical commenta- tack since it provides the gist of Snow’s ideas on tors used absolutist charts. If caregivers treating cholera from initial hypothesis (1849 [MCC]) to cholera victims also prepared food without first full-fledged and partially supported theory (1849a washing their hands, why didn’t every partaker [PMCC]) to expanded analysis of the 1854 epi- come down with cholera if the morbid matter was demic (1855 [MCC2]; and 1856a [SoLo56]). Of in the food? Snow offered several reasons, but course, one should always double-check specifics individual constitutional predisposition, the chest- in secondary works against the relevant primary nut for exceptions at the time, was not among documents, now accessible online at JSA&RC. The them. Snow’s carefully documented examples of ones I used are listed with URLs in the Bibliogra- water-borne oral-fecal communication were often phy. I also strongly recommend Hamlin (2009); easily trumped by an alternative suggestion that a “biography” of a disease rather than a single his data could be explained by clouds of infectious individual, Hamlin’s interpretation of Snow is able effluvia coming from filthy locations. The natural to be contextually more expansive than CC&SoM experiment he conceived in the autumn of 1853 (2003). to study cholera mortality in the intermixed sub- districts of south London was his best opportunity * * * to turn some minds his way. I chose the Primary Way (basing my rea- My discussion of cholera theories often melds soning mainly on the documentary evidence) in the argument in CC&SoM (165-198) with Hamlin developing my interpretation of Snow’s work on (2009, 152-62). Snow’s notion of disease speci- cholera because I wanted to establish something ficity and oral-fecal transmission is presented as we did not feature in the biography (CC&SoM) a contagionist position, pathologically and epide- and is not fully developed in other interpretive miologically. Farr’s position was more complicated studies – that Snow had used natural experi- and changeable: his Parisian training, statistical ments as confirming evidence since 1849, four bent, and initial assumption that epidemic cholera years before the Lambeth/S&V natural experi- emerged from a multiplicity of unknown causal ment came to light. factors (an “epidemic constitution”) first aligned For someone without this particular him with an orientation Hamlin terms positivist
28 Historiographical Unpacking—31 August 1854 anticontagionism. In the 1840s, Farr amended beginning 27 August (Snow 1854a). Why, I asked his views to include a mode of transmission, the myself, would Farr make such an offer? Had Snow zymotic hypothesis. The change was amenable encountered an unexpected difficulty? to a group Hamlin calls miasmatic anti-contagion- A bit of reverse literary-archaeological ists. By the early 1850s Farr’s hypothesis allowed reasoning about Snow’s writing method makes for the possibility of aerial contagious infection, me think something unexpected and catastrophic so I interpret him as a contingent contagionist in had occurred, necessitating a change in the scope his interactions with Snow. I interpret Stewart as of the natural experiment. I began with the fact a classic contingent contagionist, epidemiologi- that Victorian publishers often asked writers to cally, although he would probably fit the bill of a submit fair copy on the installment plan to facili- positivist anticontagionist in Hamlin’s typology; tate typesetting of long works, and John Churchill Stewart’s two reports (1854, 1854a) suggest was no exception. He seems to have struck an near-deification of numbers and statistics, drawn agreement with Snow early in 1854 for an ex- from careful, inductive analysis of experiential panded edition of MCC (Snow 1849). Long before knowledge gained at the bedside. Hamlin and I the second wave of the 1853-54 epidemic began differ, however, on the use of contingent conta- in July 1854, Snow replicated the structure of ar- gionism; he focuses on the convergence of the gument in the first edition and began submitting three positions after 1860, whereas I use the installments for MCC2 (1855) that contained the term as I read it to have been employed by con- fruits of additional research on cholera outbreaks temporaries in London medical journals from the he considered suggestive of oral-fecal transmis- concept’s inception in 1831 through 1854. sion. The installment scaffolding is especially Farr and Snow at the GRO evident in the section on point-source cholera Snow (1855) wrote that he and Farr met some- outbreaks caused by polluted neighborhood time during the week ending Saturday, 26 Au- water supplies. Snow essentially copied the MCC gust, when Snow presented preliminary results of discussion of Horsleydown and Albion Terrace door-to-door inquiries in two Kennington sub-dis- from 1849, then writes up additional cases as he tricts he had begun “about the middle of August” becomes aware of them via correspondence, per- (77). During this meeting Farr offered to have all sonal communication, government reports, medi- registrars in South London make inquiries about cal journals and newspapers. At least I can detect the water supply at each address where someone no argumentative reason why, for example, in would die from cholera in the current epidemic, this section his discussion of cholera at Ilford in
29 Historiographical Unpacking—31 August 1854
1849 is followed by the 1832 epidemic in New- re-used text from a letter to the editor, published burn and an 1814 outbreak in India. It appears in a medical journal the third week of September, to me that he sent them seriatim to the publisher detailing his analysis of 83 deaths during the first for typesetting (32-36). week of the epidemic (1854b; see the parallel- Whereas an archaeologist exposes suc- column word analysis in Appendix B). Subsequent cessive layers in a midden back in time, I follow text about the Broad Street outbreak was writ- Snow’s reasoning forward in time by examining ten later that autumn, also in installments, as installment segments in his published remains. he completed each phase of a more extensive Virtually no significant revision was possible once investigation than the limited inquiries conducted type had been set; only an infrequent, visibly in three days during the first week in September. squeezed footnote was reluctantly permitted. The resulting chronology of creation is often self-evi- Impure water and the propagation of chol- dent, sometimes embarrassingly so as in the first era in South London line of Snow’s (1855) introduction to the Broad Several weeks into the 1853 cholera Street pump episode: epidemic in metropolitan London, William Farr de- tected a possible pattern in South London. There The most terrible outbreak of cholera which ever occurred in this kingdom, is prob- appeared to be fewer deaths in districts supplied ably that which took place in Broad Street, with piped water by the Lambeth Company com- Golden Square, and the adjoining streets, a pared to the same stage during the 1849 epidem- few weeks ago (38). ic. His own elevation theory could not account for this difference in cholera mortality. Farr wondered MCC2 was published in January 1855 whereas the if purer water was a significant contributing factor outbreak in Golden Squared happened during the in rendering predisposed individuals less suscepti- first week of September 1854. “A few weeks ago” ble to cholera. He knew that the Lambeth Compa- suggests that Snow had given John Churchill the ny had shut its pumping station on the Thames at initial portion of his discussion of the Broad Street Lambeth in January 1852 when new works were outbreak in late September. The five paragraphs completed at Thames Ditton, three miles beyond following the introductory paragraph containing the tidal reach. But there were two other com- the sentence quoted above suggest that what panies serving South London, which, like other would turn out to be an erroneous chronological private water companies supplying the metropo- reference was an inadvertent outcome of Snow’s lis, were forbidden by an 1852 Act of Parliament economical writing method. For he essentially to draw water anywhere below the Teddington
30 Historiographical Unpacking—31 August 1854
Lock on the Thames or from any of its tributaries brought a reduction in cholera mortality. But Farr below the highest tidal point after August 1855 did not think it qualified as a defining study, an (unless special extensions were granted). What experimentum crucis, which would require identi- progress had they made in making this transition? cal factors except for water quality: The Registrar General approved a questionnaire to measure the effect of good or bad water that was sent to ten water company directors in supply, it is requisite to find two classes of mid-October (UK. GRO 1853, 402). inhabitants living on the same level, mov- Farr had answers from eight companies ing in equal space, enjoying an equal share of the means of subsistence, engaged in the in hand by mid-November, so he prepared a sup- same pursuits, but differing in this respect, plement to the Weekly Return of 19 November ―that one drinks water from Battersea, the entitled, “Cholera and the London water supply” other from Kew [far from sewage outlets and the Thames’ tidal reach]. (UK. GRO 1853, 401-06). The information on company watersheds (areas in which they had In his mind, “the circumstances of London” did laid pipes) showed that the districts supplied by not present the requisite conditions since “gener- both the Lambeth Company and the Southwark ally . . . the poorest and lowest, if not densest & Vauxhall (S&V) Company had two-thirds of districts, use the worst water,” whereas those the cholera mortality through week twelve of the in more fortunate social and economic circum- epidemic than districts where S&V was the sole stances inhabit higher terrain where the water is provider (406). S&V was still distributing Thames often the best available in the metropolis (401; water from its old Battersea works (a half mile CC&SoM, 260). John Snow, however, transformed above the Vauxhall Bridge), where tidal inflows Farr’s notion of crucial experiment into a work- brought metropolitan sewage released from outlet able concept when he discovered that S&V and pipes near London Bridge. In 1849, when the Lambeth were still active competitors in sixteen Lambeth Company drew impure Thames water sub-districts of South London. from a pumping station located near the Hunger- Snow’s installment approach in writ- ford Suspension Bridge (closer to the main sew- ing MCC2 (1855) reveals how this idea came to age outlets), cholera mortality in areas served by fruition. The germ was not the paragraph on the it had surpassed that in districts only receiving ideal experimentum crucis mentioned above but S&V water. a table that appeared in the following number of This data suggested that another natural the Weekly Return: specifically, a statement that experiment on the Exeter model was underway, Lambeth and S&V supplied the same districts in where an improvement in municipal water supply
31 Historiographical Unpacking—31 August 1854 parts of South London (68-69). Then occurred deaths/100,000 in the intermixed area, compared another example of Snow’s phenomenal memory to 114/100,000 in twelve sub-districts in which for salient facts: he located verbatim testimony S&V was the only supplier and 0/100,000 in three provided by Mr. Joseph Quick, an engineer with sub-districts where the Lambeth Company had S&V, to the Commission that produced the first no competition (73). Compelling evidence for the Health of Towns Report (a short-hand title) in credibility of his theory, but he knew it was insuf- 1844, to the effect that some districts in South ficient to seal the deal for many of his contempo- London had two or three sets of pipes in the raries. same streets (UK. Parliament 1844, 2:133; Snow One senses Snow’s excitement (1855) as 1855, 61, 68; Snow 1856a, 241. See also UK. he describes the making of a study that would Parliament 1845, 1:192-94). meet Farr’s criteria for an experimentum crucis: Quick’s testimony made Snow wonder if Although the facts shown in the above table there was a way to establish a clear demarcation afford very strong evidence of the powerful of the intermixed watersheds (which in 1854 had influence which the drinking of water con- been reduced via amalgamation to two compa- taining the sewage of a town exerts over the spread of cholera, when that disease nies, S&V and Lambeth). The second Health of is present, yet the question does not end Towns Report (UK. Parliament 1845) contained a here; for the intermixing of the water supply map depicting the distribution boundaries of all of the Southwark and Vauxhall Company with that of the Lambeth Company, over an private water companies in the metropolis (fac- extensive part of London, admitted of the ing page 1:137). It seems very likely, given the subject being sifted in such a way as to yield layout of Map 2 in MCC2 (Snow 1855, facing page the most incontrovertible proof on one side 74) that Snow used the 1845 map, updated to or the other (74). reflect changes in supplied areas reported to the It was too late in the epidemic and there was too GRO in 1853, as well as reported property rates little daylight at that point in the astronomical showing values of pipes owned by the two com- year to undertake anything systematic, but Snow panies, to determine that there were sixteen in- did make a few inquiries. He was encouraged to termixed sub-districts (Snow 1855, 72-74). When find that the Registrar-General’s office published a final list of fatalities from the cholera epidemic that in the sub-districts enumerated in the above extended from August 1853 until January 1854, table as being supplied by both Compa- nies, the mixing of the supply is of the most Snow recalculated these numbers by sub-districts intimate kind. The pipes of each Company and entered them into a table: there were 60 go down all the streets, and into nearly all
32 Historiographical Unpacking—31 August 1854
the courts and alleys. . . . In many cases a single house has a supply different from that If cholera returned to London while S&V still used on either side. Each company supplies both rich and poor, both large houses and small; the pumping station at Battersea to send piped there is no difference either in the condition water to its customers, Snow believed the next or occupation of the persons receiving the task was simple enough: “To turn this grand ex- water of the different Companies (74-75). periment to account, all that was required was to learn the supply of water to each individual house Snow realized that this was a natural experiment where a fatal attack of cholera might occur” unlike any he had come across before: (75). As there is no difference whatever, either in Note the conditional, “might” in the previ- the houses or the people receiving the sup- ous sentence. All three hundred thousand resi- ply of the two Water Companies, or in any of dents would not contract cholera, of course, even the physical conditions with which they are surrounded, it is obvious that no experiment fewer would actually die from the disease. Hence, could have been devised which would more Snow was confident when he wrote this passage thoroughly test the effect of water supply that he’d be able to complete such an investiga- on the progress of cholera than this, which circumstances placed ready made before the tion entirely on his own: “I was desirous of mak- observer (75). ing the investigation myself, in order that I might have the most satisfactory proof of the truth or He had used population data from the 1851 cen- fallacy of the doctrine which I had been advocat- sus to calculate cholera fatality ratios for table 6 ing for five years” (76). of MCC2 (73), according to which 301,149 people were registered in the sixteen intermixed sub- The SoLo denominator problem districts. That meant that this Another “might” was sorted when epi-
experiment . . . was on the grandest scale. demic cholera returned to London in the early No fewer than three hundred thousand peo- July 1854. By the fourth week of the epidemic, it ple of both sexes, of every age and occupa- was evident that the 1853 pattern in the South tion, and of every rank and station, from London intermixed sub-districts was re-occurring. gentlefolks down to the very poor, were divided into two groups without their choice, But where should Snow begin inquiries? He and, in most cases, without their knowl- waited another week. The Return for the week edge; one group being supplied with wa- ending 12 August, which was publicly available ter containing the sewage of London, and, amongst it, whatever might have come from the middle of the following week, showed that the cholera patients, the other group having cholera mortality was particularly extensive in the water quite free from such impurity (75).
33 Historiographical Unpacking—31 August 1854 two Kennington sub-districts of Lambeth. hoods within sixteen sub-districts was the de- fining feature in Snow’s South London (SoLo) I commenced my inquiry about the mid- dle of August . . . and I found that thirty- “grand experiment” (Snow 1855, 75). Within the eight of the houses in which these deaths intermixed area, however, the numbers of houses occurred were supplied with water by the each company supplied could vary considerably. Southwark and Vauxhall Company, four That meant Snow must calculate very specific houses were supplied by the Lambeth Company, and two had pump-wells on the exposure ratios ― numbers of houses with chol- premises and no supply from either of the era mortalities in each sub-district (the enumera- Companies. tor) divided by the total number of houses served As soon as I had ascertained these particulars I communicated them to Dr. Farr in that sub-district (the enumerator) ― for each . . . (Snow 1855, 76-77). company. Otherwise, the results could be badly skewed. The task turned out to be considerably more com- Take the two Kennington sub-districts as plicated than he had first imagined. Snow often an example. During the first five weeks of the found it difficult to locate the precise addresses 1854 epidemic, Snow found 9.5 times greater reported to the GRO due to absent or duplicated cholera mortality in S&V-supplied houses in Ken- house numbers. When he did find proper address- nington than those served by Lambeth; looking es, the residents were frequently unaware which good for Snow’s theory, eh? Although the two water company supplied their houses. It took companies had roughly equal market share at the some time to devise a near fool-proof chemical time in the two sub-districts, what if S&V were test, the outcome of a stroke of luck. During the sending water to ten times more houses in those weeks he undertook inquiries, there was nearly two sub-districts than Lambeth was? In that case, forty times the amount of salt per gallon in water the exposure ratios would have been roughly supplied by S&V compared to Lambeth water. De- identical. termining correct enumerators for his study “was The manner in which Snow (1854a) necessarily attended with a good deal of trouble” described his initial investigation in the two Ken- (77) and more time-consuming than he had an- nington sub-districts suggests that he initially ticipated, but he persevered. If only he had been flirted with exposure ratios based entirely on equally assiduous in locating proper denominators houses. He showed Farr figures on the number of for his study. houses in which a cholera mortality had occurred during the first five weeks of the current epidem- An intermixed water supply in similar neighbor- ic, organized by the source of water consumed in
34 Historiographical Unpacking—31 August 1854 those houses (247). As to denominators, Snow from Snow (1854a) in the form of a letter to the interpreted the information on estimated value editor of the Medical Times and Gazette (MTG) of pipes and other materials that S&V and Lam- dated “Aug., 1854”: beth had supplied Parliament to suggest that the Sir,--I have been engaged, during the companies supplied nearly the same numbers of last ten days, in an inquiry which promises houses in each of these four sub-districts (247). to yield very conclusive evidence respecting In other words, he arrived at the GRO on or about the mode of propagation of cholera...... th Thursday, the 24 August 1854, with the results I intend to continue the inquiry, ex- of initial inquiries in the “grand experiment” he tending it to the other sub-districts in which had envisioned during the winter of 1853/54: the two water companies are intermixed, and to bring it down to the 26th inst [Au- gust]. After this date, I am informed by Mr. • A study limited to the 16 Farr that the supply of water at the house sub-districts where S&V and Lam- in which every fatal attack of cholera may beth had laid pipes in the same occur, will be returned by the Registrars in streets, alleys, and mews. all the Districts on the South of the Thames • Snow anticipated that (247). he would be able, on his own, to investigate the water supply at This letter was published on 2 September and every house in this intermixed area included the results of inquiries Snow (1855) had where one or more fatal attacks of cholera was recorded for the full made in two Waterloo sub-districts; that is, more duration of the epidemic. data than he had shown Farr previously when • Exposure ratios calculat- the offer was actually made (76-77). The pas- ed for each company as the num- ber of houses with one or more sage quoted above shows that within ten days of fatalities in each intermixed sub- initiating house-to-house inquiries in Kennington, district ∕ total number of houses Snow had abandoned the notion that he, person- supplied in that sub-district. ally, would make inquiries for the duration of the current epidemic at every house in the intermixed area where someone had died of cholera. Few at Farr’s offer to involve GRO registers in that time would have known his original intention. Snow’s SoLo study was the first indication that The first public disclosure (Snow 1854a) of the something was awry. There is no need to specu- SoLo study outlined it as a joint venture: Part 1, late about when a passage in MCC2 was com- the first seven weeks, with inquiries undertaken posed; we have real-time documentary evidence by Snow in the intermixed sub-districts; part
35 Historiographical Unpacking—31 August 1854
2, from the eighth week onward, all registrars he felt comfortable in drawing preliminary conclu- in South London making inquiries about water sions: sources when assembling their regular weekly both Companies supply alike all kinds of reports. houses, — those of the rich and the poor in- It strikes me as unlikely that Farr would discriminately. It is evident, therefore, that, involve all South London registrars in part 2 un- in the sub-districts to which the inquiry has extended, the people having the improved less he and Snow had already broached the pos- water supply enjoy as much immunity from sibility of expanding part 1 beyond the intermixed cholera as if they were living at a higher lev- area mentioned in the letter. The narrative flash- el, on the north side of the Thames (247). back to 24 August suggests that this discussion Snow (1854c) wrote the Medical Times occurred at the same meeting when Snow pre- and Gazette on Monday 2 October that he sented his Kennington findings. Snow could state, had just completed inquiries into 642 chol- without equivocation, that he would conduct all era deaths in the intermixed sub-districts of inquiries about deaths in the intermixed sub-dis- South London registered during the first seven tricts during the first seven weeks of the epidem- weeks of the epidemic. Only 93 fatal attacks ic, and still have known that the absence of data took place in houses supplied by the Lambeth required for calculating exposure ratios for the Company, whereas a whopping 509 instances intermixed area necessitated eventual expansion occurred in houses supplied by S&V; the water of part 1. Perhaps, at the time, he hoped Farr was source in the remaining 40 cases came from as adept at magic as he was at statistics, and the pumps, ditches, directly from the Thames, or housing data would be divulged in rapid fashion. could not be determined. The letter to the editor Snow (1854a) But he could not seal the deal, as wrote at the very end of August side-stepped Snow (1854c) admitted publicly for the first the denominator problem he would describe five time, until he knew “the number of houses in weeks later. Although I believe he had just found each sub-district supplied by each of the Water out the problem existed, I’m not suggesting that Companies respectively” (365). That is, gross Snow misled anyone with preliminary, unadjusted totals for the intermixed area were only sug- exposure ratios for S&V and Lambeth. District gestive; they did not show what he believed and parish rate books (containing the value of a sub-district analysis would — “the effect of pipes and other property) indicated roughly equal the impure water [from S&V] in propagating market share by the two companies in the four cholera . . . in a very striking manner, and with sub-districts he had investigated. Consequently,
36 Historiographical Unpacking—31 August 1854 great detail” (365). He attached a table depict- sixteen intermixed sub-districts was to extend in- ing what he had discovered about water sourc- quiries into the twelve sub-districts served solely es and fatal cholera attacks at the sub-district by the S&V water company, as well as the four level, which meant he had the enumerators for Lambeth-only sub-districts. calculating detailed exposure ratios for both Snow had farmed out inquiries in the companies. He only lacked the proper denomi- twelve S&V-only sub-districts to a newly minted nator data to complete his part of the “grand apothecary Licentiate, John Joseph Whiting. Since experiment,” although he “hope[d] shortly” to there was no competition in this area with the receive this information (365). Without it Snow Lambeth Company, Whiting’s task was to deter- could not complete his analysis of the natural mine whether S&V ran pipes to houses where experiment underway in the intermixed area of someone had died of cholera during the first South London. seven weeks of the epidemic, or the residents Snow then described the back-up plan fetched water from alternative sources. Unfortu- that had been underway throughout the month nately, Whiting was only able to complete inquir- of September. “In the mean time” — by which he ies on cholera fatalities during the first four weeks meant, since shortly after learning that the neces- before he departed the metropolis for a country sary denominator data was unavailable: practice, forcing Snow (1854c) to “calculate” (high-falutin for guess) that the water sources for in order to be able to compare the mortality from cholera among the customers of each all cases in these twelve sub-districts for weeks Company, with the entire number of houses five through seven were “in the same proportions supplied by each of them respectively, I as those occurring previously” (365). Snow’s let- thought it desirable to extend the inquiry to ter contained a table depicting his and Whiting’s Rotherhithe, Bermondsey, Camberwell, and certain parts of Southwark, which are sup- findings by district, with asterisks next to the plied by the Southwark and Vauxhall Compa- S&V-only sub-districts that involved estimated ny alone. I was unable by myself to execute results. this part of the inquiry . . . (365). Had Snow ended the letter (1854c) at this The quote doesn’t mention four outlying sub- juncture, he risked giving the impression that the districts with little cholera supplied only by Lam- SoLo study was just another, suggestive-at-best, beth, which Snow had put on his own to-do list example of comparing cholera fatalities after the and completed in September. Snow and Farr had introduction of pure water in some areas to other decided that the only way to take advantage of areas with unchanged supply. He decided to give the natural experiment currently underway in the his audience a prefigurement of the analysis to
37 Historiographical Unpacking—31 August 1854 come when he received the sub-district supply the houses supplied by the former Company data from S&V and Lambeth that would permit as in those supplied by the latter (366). him to plot exposure ratios for the early part of The possible effects of pure and impure water on the epidemic within the intermixed area, before cholera mortality were even more striking in the normal propagation really set in and skewed the first four weeks, when other forms of propagation causative effect of impure water. Pending that weren’t as pronounced: ideal scenario, he used the only denominator data available to him at the time and calculated ex- During the first four weeks of the present posure ratios during the first seven weeks of the epidemic, 563 persons died of cholera in London. Of these it has been ascertained, epidemic for each company’s entire watershed: by a personal inquiry at every one of the houses in which the attack took place, that The entire number of houses supplied by the no less than 268 of the fatal attacks took Southwark and Vauxhall Company, accord- place in houses supplied with water by ing to a return made to the General Board of the Southwark and Vauxhall Company. . . Health in 1850, was 34,217, and the number . During these four weeks there were but supplied by the Lambeth Company, according ten deaths from cholera in houses supplied to the same return, was 23,396. The number with water by the Lambeth Water Company; of houses supplied by both Companies has although it has been shown above that they increased with the extension of the Metropo- supply fully two-thirds as many houses as lis, but it is pretty certain that the proportion the other Company. The cholera was con- continues nearly the same, and for the sake sequently eighteen times as fatal among of comparison, the number of houses may be the population supplied with the water from supposed to remain the same also (366). Battersea Fields as among that with the purer water from Thames Ditton, during Snow then divided these figures on the total num- these four weeks, although this latter popu- lation is intimately mixed with the former. ber of houses each company supplied in South London by the total of actual plus “calculated” A lot of guesswork, two dodgey numbers, and a cholera fatalities in houses traced to each com- slight over-reach in the last sentence; only 62% pany, then proposed of the population in the combined watersheds that while a death from cholera had oc- had an intermixed supply (Snow 1855, 84). But curred in 1 house in every 28 supplied by it was the best he could do with the information the Southwark and Vauxhall Company, a available. fatal attack of cholera had occurred in only 1 out of 251 of the houses supplied by the However, it was not the analysis of the Lambeth Company. The mortality, in short, “grand experiment” Snow had already described to August 26, was just nine times as great in for the forthcoming publication of MCC2.
38 Historiographical Unpacking—31 August 1854
26th on his own (76-77). The October letter to Snow’s SoLo analysis in MCC2 MTG (1854c) contained a long description of dif- October ended without Farr receiving the ficulties he encountered when he asked residents desired denominator data. Snow had a deadline which company supplied their houses; “It would, to submit the remaining sections of MCC2. He had therefore, have been impossible for me to com- delayed writing up the South London study as plete the inquiry if I had not found that I could long as he could. distinguish the water of the two Companies with We can imagine Snow’s chagrin. He had perfect certainty by a chemical test” (365). Snow in hand enumerator data on cholera fatalities reduced the entire passage in MTG when writ- in the intermixed sub-districts during the first ing MCC2, but some wording is nearly verbatim: seven weeks of the epidemic. He had person- “It would, indeed, have been almost impossible ally collected it, as promised in the letter to MTG for me to complete the inquiry, if I had not found published on 2 September. Granted, his findings that I could distinguish the water of the two com- did not cover the entire epidemic, but they con- panies with perfect certainty by a chemical test” tained neither guesswork nor “calculations.” Part (77-78). of the planned “grand experiment” was ready for The structure of Snow’s argument in analysis. But he could proceed no further without MCC2 often differs from that in the second letter sub-district level information on how many houses to MTG, but snippets of varying length frequently each of the private water companies supplied in reappear. He condensed a long paragraph on the the intermixed area. What was to be done? absence of sub-district housing supply and hiring Snow really had no option except to revise Whiting, deleting the phrase, “I hope shortly to and expand the two letters he had sent to MTG, learn . . .” but recopying verbatim that “I was for- an approach he had already used for the open- tunate enough to obtain the assistance of a medi- ing sections of MCC2 with the eponymous 1849 cal man,” and additional phrases about Whiting’s pamphlet (the title page states, “Second Edition, inquiries on fatalities during the first four weeks much Enlarged”). The South London portion of of the epidemic in S&V’s unmixed watershed MCC2 runs just over sixteen pages of text and (1855, 78-79). An entire paragraph in MCC2, with tables (1855, 76-92). Snow transformed the let- the exception of one word, comes from the MTG ter published in September (1854a) into a brief letter (1854c): discussion of his initial inquiries in Kennington, It may, indeed, be confidently asserted, that Farr’s offer to involve GRO registrars beginning 27 if the Southwark and Vauxhall Water Com- August, and Snow’s “resolve” to investigate chol- pany had been able to use the same expedi- era fatalities in the entire intermixed area until the tion as the Lambeth Company in complet-
39 Historiographical Unpacking—31 August 1854
ing their new works, and obtaining water table for the first seven weeks, added a new table free from sewage, the present epidemic of covering the first four weeks, and organized both cholera would have been confined in a great measure to persons employed among the tables to show sub-district fatalities in three cat- shipping, and to poor people, who get water egories: solely S&V, intermixed, and solely Lam- by pailsful direct from the Thames or tidal beth (1855, 84-85). Denominator data in MTG ditches (366). is from 1850, whereas in MCC2 it’s from 1853 (1855, 80); accordingly, the figures Snow used Snow changed “present” to “late” since the chol- for Lambeth’s percentage of market share com- era epidemic had ended when he wrote this pas- pared to S&V dropped from 68% to 65%. Snow sage (1855, 81). There are many other incidents used total watershed housing figures and popu- where he re-used, with minimal alteration, parts lation data to discuss fatalities in South London of sentences from MTG when writing parallel sec- for the fourteen weeks of the full 1854 epidemic tions in MCC2. (1855, 86-91). In addition, in MCC2 Snow toned Perhaps the most notable differences down his criticism of “ill-directed efforts of benev- between the October letter to MTG and MCC2 olent individuals among the non-medical part of are corrections Snow made to his discussion the community” that he believed had caused the of the number of fatal attacks he and Whiting sewage concentration of the Thames to increase had investigated and an accompanying table; a since the 1830s (1854c, 366). change in the data he used to calculate tentative exposure ratios; and addition of data gathered by I’ve carried the discussion of MCC2 well beyond GRO registrars from 27 August through 14 Octo- what transpires in the 31 August 1854 narrative ber. For the intermixed area during the first seven because of a common misperception that it con- weeks, Snow changed the numbers for S&V from tains Snow’s analysis of the “grand experiment” 509 to 525, Lambeth less drastically from 93 to during the 1854 cholera epidemic where the 94 (1854c, 365 and 1855, 85). For entire water- watersheds of the Lambeth and S&V companies sheds during the first four weeks, 268 deaths in intermixed in South London. The revised Lancet S&V-supplied houses became 286 and Lambeth obituary (2013) during the 200th anniversary went from 10 to 14 (1854c, 366 and 1855, 79); year of Snow’s birth is an example: the likelihood that the early S&V number was a typographical error at the press is lessened by It was his “Grand Experiment” that same the fact that Snow changed the fatality proportion year [1854] that secured his huge reputa- between the two companies from 18:1 to 14:1 in tion in epidemiology. . . . During 1848–49, the death rates for the two companies were MCC2. He also corrected figures in the singleMTG the same, but by 1854, after Lambeth’s
40 Historiographical Unpacking—31 August 1854
move, Southwark and Vauxhall’s rate was various phases of the epidemic. But one must between eight and nine times higher, and consult Snow’s article (1856a) for his analy- in the first 4 weeks of the epidemic, South- wark and Vauxhall customers had a 14-fold sis of the “grand experiment,” despite Wade higher risk. In 1855, Snow published a Hampton Frost’s dismissal of it as “not alto- much-expanded second edition of On the gether essential to Snow’s original argument, Mode of Communication of Cholera that which was already well established . . .” (1965, included these results. xvi) — yep, you guessed it, in MCC2 reprinted The 14:1 risk disparity (60% of S&V inquiries after Frost’s introduction. during this period were conducted by Whiting) refers to what Snow described in MCC2 (1855, Historical continuity vs. historical change 80) as total cholera fatalities during those One of my goals when teaching historiog- weeks in the entire watersheds of the two raphy has always been to make transparent the companies, not the intermixed area that con- tacit knowledge of a professional historian. This stituted Snow’s intended “grand experiment.” often necessitated a my-way-or-take-another- He was not able to undertake those calcula- class approach to counter default notions of doing tions until most of the sub-district level hous- history. I gave students worksheets containing ing data he required was published in a retro- intellectual rubrics they must follow when re- spective study lead by John Simon (1856). searching and writing until they mastered my In short, MCC2 does comprise a de- methodology. Appendix A is an example where scription of why Snow believed this natural ex- I asked them to decide, before drafting a thesis periment (in which happenstance determined if statement, if their interpretations involved his- one’s house received impure or purer Thames torical continuity or historical change. water) had the makings of an experimentum Nightingale’s sojourn on a Middlesex crucis. MCC2 does describe what had to be Hospital ward was an example of historical con- done to try and determine the water supply at tinuity. It was the actualization of a decision she each house where someone had contracted a had already made to leave the Gentlewoman’s fatal case of cholera, why it was necessary to Establishment for something else, some new post farm out some investigations, and the totals or experience yet to be determined. Hence, my he, Whiting, and the registrars were able to thesis statement for the opening narratives in amass in their respective assignments. MCC2 which she is featured reflected basic elements of does contain Snow’s risk-exposure calculations historical continuity: what she did, when she did per company for their entire watersheds during it, and how/why it came about.
41 Historiographical Unpacking—31 August 1854
My interpretation that a circumstance meeting has the characteristic feature of a con- beyond John Snow’s control forced him to alter tingent historical event as set forth by Stephen his intended focus on just the intermixed water- Jay Gould (1989, 283). The causal explanation sheds in South London is an example of historical and/or contingent event constitute the “Because” change. It requires a different thesis statement, component in the thesis statement. historiographically, than the one that underpins Fourth, identify the context necessary the opening narratives featuring Florence Night- for a reader to understand your narration of this ingale, a thesis statement based on a diagram historical change. For the 31 August narrative, I similar to what I would assign students to con- chose to summarize Snow and Farr’s respective struct with the following four c-words in mind: cholera theories and the serendipitous appear- First, depict a change in chronology as ance of a natural experiment during the London “From > To”; for example, from Snow’s initial vi- cholera epidemic of 1854. sion of the SoLo project as limited to intermixed My thesis statement for Snow and Farr watersheds > to the expansion to include all sub- at the GRO is an example of historical change: districts where either S&V or Lambeth was the The absence of denominator data essential to sole provider of piped water. analyzing just the intermixed Lambeth/S&V Second, propose a causal explanation for watersheds forced Snow to suspend his “grand this change, such as the denominator problem experiment,” then expand and share inquiries Snow encountered — the unavailability of data into fatal cholera attacks occurring in every South on the exact numbers of houses the two water London district served by the two companies. companies served in each of the inter-mixed sub- This unexpected change in plans affected every- districts. thing he did during the 1854 epidemic, including Third, present evidence of a contingent his approach to investigating the local outbreak event or factor that substantiates the proposed in Golden Square. “From > To” chronology and the hypothesized causal explanation. The contingent event for The narrative form Snow in the 31 August narrative occurs when his Farr and Snow’s ideas and actions are de- anticipated SoLo project changed on a dime: a veloped in a series of dated flashback-narratives, documented meeting in August 1854 where Snow some containing sub-titled explanatory sections, (1854a) showed Farr preliminary findings from as well as a final narrative on Thursday 31 Au- two inter-mixed sub-districts and Farr’s response gust. was to offer the services of his registrars. This Although I have found no documentary evidence
42 Historiographical Unpacking—31 August 1854 that Snow and William Farr actually met at the Snow could have found the gist of recent GRO on that day, it is very possible that they did cholera mortality in the metropolis from reading so. It was the eighth week of the London cholera extracts in the Times or procuring a complete epidemic. The Weekly Return no longer contained copy of the return; but to suss out specifics, details of each death; the numbers were too he had to walk to Somerset House. He had no enormous. So Snow would have had to visit the anesthetic administrations scheduled between 25 GRO to secure a list of everyone who had died of August and 1 September, so he could have made cholera during the week ending 26 August in the the visit on Wednesday or Thursday. Absent evi- South London sub-districts covered by the natu- dence to the contrary, I chose Thursday. Which- ral experiment. ever day he did visit the GRO that week, it is Farr had already instructed GRO staff to very likely that he would have popped into Farr’s provide Snow (1854a) each week with mortality office for a chat and to give him an update on lists containing information not available in daily the progress of the South London investigation. newspapers or printed Weekly Returns. Sub-dis- This was sufficient probability for me to compose trict registrars were supposed to complete re- an historical narrative with a setting at the GRO turns of births and deaths each Sunday and pass on the same day that Florence Nightingale began them on to their district registrar for review. Dis- her sojourn at the Middlesex Hospital. trict registrars then forwarded their reports to the As in the Nightingale episode at the insti- head GRO office in Somerset House, where Farr’s tute on Upper Harley Street, the narrative style team attempted to prepare complete copy for I employ for meetings between Farr and Snow each return by Monday evening for type-setting is predominantly free indirect speech. When the into pamphlet form; copy was also sent to major narrative contains conversational speech, it is newspapers. When all went smoothly, Weekly Re- always in free indirect style, not quoted speech, turns were available at District Registrars’ offices since the latter in an historical work denotes on Wednesday and extracts from it appeared in quotations from documentary evidence. Take the Wednesday editions of the dailies; the Times, opening lines of the 24 August flashback as an for example, published overviews, tables, and example: meteorological summations for the week ending So far, the experimentum crucis is yielding 19 August 1854 on Wednesday the 23rd (“The anticipated results! Health of London,” 8c-d), and on Wednesday 30 Farr looked perplexed. What was Snow August for the week ending the 26th (“The Public talking about? Health,” 9c-d).
43 Historiographical Unpacking—31 August 1854
This is free indirect style — Snow’s spoken state- starting 27 August. The Snow/Farr section ends ment (without attribution) and Farr’s reaction on 31 August, the narrative present so to speak, (his internal speech) as free of authorial flags (in with an hypothesized meeting that updates this instance, an historian’s seeming omniscient Snow’s findings with two Waterloo sub-districts knowledge) as I could make it. Had I used re- and shows his resolve to begin inquiries in the ported or simple indirect speech in this instance, other sixteen sub-districts on his expanded to-do I would have added, “said Snow” to the first sen- list in South London. Narratives about Snow’s tence and “thought Farr” to the third. My autho- hypothesized meetings with Farr on 24 and 31 rial presence is limited to providing context: how August are based on Snow (1855, 76-79; 1856a, Farr looked. 242-43; and 1854a. Left-justified dates in bold font designate encounters between Snow and Farr, beginning Additional documentation & unpacking with the flashback to 24 August 1854. This nar- The page preceding the Farr-Snow sce- rative is interrupted by explanatory sections (set nario contains four figures scanned from the 19 off by centered and bolded sub-titles in a smaller November Weekly Return (UK. GRO 1853, 389 font) that provide historical context and Snow’s [title page]; 391 [sample of individual deaths]; doings during the early stages of the SoLo inves- 401 [supplement title and extract, marginal tigation (also designated by left-justified dates check marks mine]). in bold font), before resuming at the point where Three figures on the opening page of the Snow shows Farr the results of preliminary in- Farr-Snow scenario: quiries in the Kennington sub-districts. When nar- • Detail from Map 2, “Water supply to rating Snow’s activities in the run-up to the SoLo Registrar-General’s districts for South London” project and the initial inquiries in Kennington, I (Snow 1855, after 74)
44 Historiographical Unpacking—31 August 1854
The figure on the tidal reach of the Thames and Lea rivers is an amended version of a map in Luckin (1986 frontispiece). I scanned
Luckin (1986) also has an illustration of London registration districts as of 1850 (71). I imported it into Photoshop and cropped the sec- tion containing the South London districts served by the Lambeth and S&V water companies. Luckin’s map into Photoshop, erased everything Then I erased Luckin’s registration district num- but the two rivers and the distance marker, and bers, altered used the Horizontal Type Tool to add place names the bridges to mentioned in the Snow-Farr narrative. comply with the The figure of the first published data detail covered documenting a reverse in cholera mortality in in the narra- districts served by Lambeth and S&V between the tion, and filled 1853 and 1849 London epidemics is an extract in the missing from the 19 November Weekly Return (UK. GRO district bounda- 1853, 406). This Return suggested that the 1853 ries on a printed London cholera epidemic contained the makings copy. I used of before-and-after natural experiment like those this enhanced occurring in previous epidemics in Exeter and printed copy as a base, along with the Brush and Hull. Horizontal Type Tools in Photoshop, to create the The 26 November Return contained the figure of the intermixed sub-districts, as well as game-changing evidence that there were still in- the subsequent three figures depicting aspects termixed watersheds in South London (UK. GRO of the SoLo natural experiment. An experienced 1853, 409).
45 Historiographical Unpacking—31 August 1854
graphic designer Sloane Square & would, of course, Nine Elms, 1869” produce more (Godfrey, London professional figures sheet 88). Unlike that I can manage. the previous fig- But one purpose of ure, in which any my narrative is to changes at the show what anyone lock that occurred with an elementary during the forty facility in editing years since 1854 images might have done. are immaterial to Shortly after the first of the SoLo figures the narrative, the 1869 map depicts a functioning comes the prepared-mind reference to cholera in railroad bridge, rails for four different railways, the Baltic Sea. The narrative is based on Snow’s and the Battersea Park Station — none of which letter to the editor of MTG (Snow 1854). The appear on Map 2 of Snow (1855, after 74). Con- PDF at
46 Historiographical Unpacking—31 August 1854
respect chronology; never explain something that Stewart and Nightingale at the happened earlier by something that happened Middlesex Hospital later. The narrative scenario in which Nightingale Alexander Stewart was a clinician who believed first meets Dr. Stewart was crafted to emphasize that cholera was not directly contagious, but that on 31 August 1854, Nightingale was a medi- could be circumstantially infectious. Huh? How cal novice. The gentlewoman’s institute she had can something be infectious but not contagious? founded in Upper Harley Street was a short-term Is this chap a total nutter? nursing home, not a hospital. She would have Hardly. It’s all about a very common known very little about many diseases that might phenomenon, the changing meaning of words eventuate in admission to a female ward at a ma- over time. For example, in a report detailing the jor metropolitan hospital. sudden increase in ill residents who flocked to the Consequently, it would be historically Middlesex Hospital in early September, Stewart inaccurate to depict Nightingale on the afternoon (1854a) remarked that “a very large proportion of her first day as a ward sister on the basis of of these being, on admission, [were] far beyond the hospital reforms she instituted after travel- the reach of remedial skill. Such being the case, ling to the Crimea in November 1854. In fact, the it is not wonderful, that of nearly 90 cases admit- reforms for which Nightingale is justly renowned ted during the fifty hours ending at one p.m., on had much in common with what was already in Sunday, the 3rd of September, forty were at that place at the Middlesex Hospital during the cholera hour already dead” (364; italics mine). Of course, outbreak in September, nothing at all in common Stewart did not consider that many deaths either with how she had organized and managed the fa- admirable or unusually good; Stewart chose a cilities in Upper Harley Street. This point is impor- word that in his age meant astonishing, a usage tant when considering the historical significance infrequent now. Similarly, if we parse Stewart’s of Nightingale’s brief sojourn at the Middlesex viewpoint about cholera in terms of the cholera Hospital during the Golden Square cholera out- theories in play in mid-nineteenth century Eng- break. land, it was quite reasonable for a London physi- I chose to use an imagined (but likely) cian in 1854 to consider cholera a non-contagious meeting with Dr. Stewart to have him explain how infectious disease. Middlesex Hospital was set up to facilitate good nursing practice and minimize contingencies that The narrative form could promote infectious transmission of epidemic A cardinal rule in historiography is to diseases such as cholera. In the mid-1850s, ac-
47 Historiographical Unpacking—31 August 1854 cording to Stewart’s October 1854 report, the en- edu/snowcontextual.php>). “Contextual Mate- tire medical staff considered typical Asiatic chol- rial,” a data base of medical journal articles and era a non-contagionist epidemic fever, although letters to the editor (in PDF format), was particu- hospital infection was a rare, but very realistic, larly relevant in preparing this narrative scenario. possibility. In short, Stewart and his colleagues If you wish to locate other citations in the data were contingent contagionists. base, please employ the “Find” function on your Although Stewart had been concerned for browser. several weeks about a dramatic increase in the If you have access to a research library, number of dispensary outpatients who present- be sure to run at least a few of my suggestions ed with diarrhea, he did not foresee the major to ground in their respective issues. Just as one outbreak of cholera that began the night after often finds interesting stuff around a desired item his hypothesized meeting with Nightingale. No on a stack shelf, paging through an issue (wheth- one expected it, especially not a severe outbreak er electronically or getting your fingers dirty) may in any of the northern districts of metropolitan yield a fuller harvest than just reading what I London because most sickness and mortality from posted on the web site. If available to you, James the current epidemic, underway since early July, Copland’s Dictionary (an encyclopedia, really) is was occurring elsewhere, particularly south of the useful for a contemporary’s viewpoints, although Thames. keep in mind that he was a staunch contagionist. The OED often establishes earliest usage. Additional documentation & unpacking: A faster, secondary-source route is avail- In the twenty-first century, historical nar- able to the same end. CC&SoM (Vinten-Johansen ratives set in the mid-1850s require contemporary 2003, 165-98) and Durey (1979, 101-21) cover terms and concepts that are made accessible to much of this terrain. However, if one chooses present-day readers. In order to explain and vali- this option I recommend at least using Durey’s date Alexander Stewart’s contingent-contagionist endnotes (231-33) to supplement those given in perspective on cholera, I chose to begin with CC&SoM above for targeted searches of the pri- primary sources. It was the simplest approach for mary literature as checks on interpretations in the me because I had already compiled detailed notes secondary literature. when researching CC&SoM and entered them At the beginning of this scenario, the into various data bases (many available in the figure depicting the location of Middlesex Hospital research schemas sub-menu of the Online Com- is a detail from Godfrey, “The West End, 1870.” panion in JSA&RC 48 Historiographical Unpacking—31 August 1854 at Middlesex Hospital, but the only one I have statement attributed to her: “. . . every day this located is from early in the nineteenth century. Al- week . . . those who live ten feet above a pesti- lential river will die, and those who live forty feet above will live” (Small 1999, 129). The source, however, was not cited, so I sent an e-mail mes- sage to the author requesting clarification. Hugh Small kindly replied with the desired bibliographi- cal information: Goldie, 3_794, 25/8/54 (a lesson to all researchers; nothing untoward in asking). The information Small provided suggests to me that the source of Nightingale’s comment may have been an abstract from the Weekly Return for 19 August which appeared in the Times the following week. Nightingale’s mother sent her though the sanitary advantages of hospital wards issues when they were done with them (FN to as described by Stewart are clearly evident in this mother, 5/10/53, in Goldie 1983, 3_730 5/10/53). lithograph, I decided not to place it in the nar- Consequently the narrative reflects my interpre- rative because the clothing styles depicted were tation that Nightingale had a basic awareness of outmoded by the 1850s. the current cholera epidemic and Farr’s statistical Stewart’s overview of diseases repre- analysis and zymotic theory that mortality was sented on the female ward is based on currently associated with elevation relative to the Thames. prevalent zymotic diseases listed in each Weekly Dr. Stewart’s general therapeutic orien- Return, including the one issued 2 September tation and treatment of cholera victims will be 1854 (UK. GRO 1854, 318). unpacked in the next chapter. The narrative assumes that Nightingale Stewart (1854) accused the Vestry of St. had at least as much knowledge of epidemic chol- James, Westminster of having wilfully ignored a era and the current situation in metropolitan Lon- significant outbreak of diarrhea and cholera in don as any educated person who read the Times. the greater Soho district for a fortnight prior to 1 Even though I could find no mention of cholera September. In the narrative I have Stewart men- prior to September 1854 in Goldie’s summa- tion this outbreak in Soho and St. Giles, as well tions of Nightingale letters, during the literature- as the death of a dispensary surgeon who had research phase I had come across the following been treating victims. Charles Tilly’s death was 49 Historiographical Unpacking—31 August 1854 listed in the Times (31 August 1854, 1a); he was name, just that she was a widow, with dependent 34. The Westminster General Dispensary, where children, who resided in a “small house” in which Tilly worked, put an announcement in the Times, eight people had died by 4 September, including requesting public support to help it cover unex- the widow and some of her children. pected expenses during the second half of August First I had to find the actual address of (2 September 1854, 6d). this small house. Three houses in Whitehead’s street-by-street accounting listed eight dead (1854, 4); but only Hopkins Street contains a Whitehead and the Hopkins St. Widow suggestive description: “only 3 small houses; Henry Whitehead was the senior curate at St. population, about 70; deaths, 13, of which 8 Luke’s, Berwick Street, the official parish church were in one house” (1854, 4). I then checked for nine thousand persons. He was three weeks a “plan” containing mortality data produced in short of his twenty-ninth birthday when he made 1855 for the Board of Health, and found eight an evening call at 9 Hopkins Street. My reading of deaths recorded at 9 Hopkins Street (UK. GBH. his published writings, especially the sermons he Medical Council 1855). An enumerator for the delivered at St. Luke’s, makes me certain that he 1851 census listed a Mary Ann Plynn, widow, had excellent rapport with his parishioners, which with five children at 9 Hopkins Street (HO I wanted to demonstrate in this short narra- 107/1483/132. Online: Middlesex>Westminster, tive scenario. Another purpose was to set up the St. James>Berwick Street>District 3/21). I con- opening narrative for the following day. firmed the spelling of her name by a search of the Historical detection led me to the address census data base, which brought me to the same and the names I use in this scenario. I began with page of the census. An extract from the GRO’s Whitehead’s (1854) pamphlet, The Cholera in Weekly Return for the week ending 2 September, Berwick Street, written shortly after the conclu- published in the Morning Chronicle on 6 Septem- sion of the point source outbreak in his parish dis- ber 1854, noted the death of three residents of 9 trict but while epidemic cholera was still plaguing Hopkins Street – a mother and two of her sons. the metropolis. Whitehead states that “the first I had the name of the person Whitehead visited intimation which the writer received” that an out- and where she lived. The census enumeration also break was underway “came in the form of a sum- gave me names and vital data on other occupants mons to the death-bed of one with whom he had at this address. I made the assumption that all of cheerfully conversed at a late hour on the preced- them were still lived there in the summer of 1854, ing evening” (2-3). He does not give the person’s although that’s clearly a stretch and, at best, his- 50 Historiographical Unpacking—31 August 1854 torical possibility. their mind-sets without further flagging devices The map detail is from the plan men- (he thought, she thought, etc.). After the edi- tioned above; I erased mortality data in Photo- tor’s friends, it’s one paragraph in third-person shop. (indirect style) before ending the scenario in free indirect style. The narrative form Whitehead wrote that he spent part of Conclusion the evening with ”a patient, gentle widow”; my We know what began to happen late in the even- description of the room at 9 Hopkins Street for ing of Thursday 31 August 1854. Farr and Snow, which the 1851 census listed her as head con- Nightingale and Stewart, Whitehead and the Hop- veys what I imagine such a person would do to kins Street widow did not. As explained earlier, make it livable for herself, five children, and four this disjunction is the essence of dramatic irony, lodgers. I decided to have mugwort hanging in which I have incorporated into all the historical the room to reflect the widow’s interest in cleanli- narratives for this day. This was no master plan ness, regardless of the family’s poverty. Mugwort from the outset, however. (from muggia, meaning midge) is an old Saxon My decision to apply dramatic irony was word and common name for Artemisia vulgaris, an unexpected outcome of writing up Florence a widespread plant attractive to some insects. Nightingale’s doings at the Middlesex Hospital Bunches of the plant were hung for a period of on 31 August. The two short narrative scenarios time, then taken down and placed in a bag, which I composed reflected Mrs. Gaskell’s perception was beaten, thereby killing the insects trapped in- that Nightingale and the hospital staff she met side (MacKinnon, et al. 1992, 105). Sprigs might that day had no inkling an outbreak was brewing. be used again, or discarded, depending on their I also knew from having read Whitehead (1854) condition and one’s financial circumstances. that he was similarly inclined. So, I asked myself, I constructed the purpose of Whitehead’s why not establish the same feeling in a narrative visit to be discussing arrangements for the parish about Snow and Farr on the 31st? Sunday School outing that was initially scheduled That decision also solved a vexing or- to occur two days hence (Whitehead 1856, 58). ganizational problem. How, I had frequently The opening two paragraphs are autho- wondered, could I present Snow’s initial cholera rial third-person narrative. The next two para- hypothesis from 1849 and its theoretical updat- graphs are in indirect and free indirect styles; I ing in the years thereafter without disturbing the introduce the speakers and immediately shift into narrative integrity of the Broad Street cholera 51 Historiographical Unpacking—31 August 1854 outbreak? The same concern applied to Farr’s zymotic view of diseases and the multitude of cholera causation theories. The solution I devised was to interrupt narrative interchanges with ex- tended explanatory sections and then incorporate the information presented there into subsequent narratives. 52