Chapter 4. The Time of the Evidence: Trials, Single Cases and the Long Wait for Proof

Chapter 3 read acts taken before any cause was known. This chapter reads the second separation: what was known, and when, about an own effect. Four records are read here — Lind's comparison aboard the Salisbury, Jenner's inoculation of 1796, Snow's pump-well in Broad Street, and Pasteur's treatment of Joseph Meister — beside the texts their actors wrote and the works that later read them. In each, the act and its proof are not the same event. A comparison of twelve men ends in 1747 and its report appears in 1753. One boy is inoculated on 14 May 1796 and the claim is printed in 1798. A pump handle is removed on 8 September 1854 while the outbreak's attacks are already falling, and the book that says so is published in 1855. One boy is treated from 6 July 1885 and the report to the Académie is printed on 26 October 1885. A fifth record, The Plague at Athens, closes the chapter: there a later identification of a disease described in 430 BC is argued over in 2006, and the act whose hour would matter is not settled at all.

4.1Marks: the trial as discipline

A controlled comparison is not the same thing as an observation. Marks's work is about the United States from 1900, and the four acts read here are earlier and elsewhere: Lind's comparison was made at sea in 1747, Jenner's in a Gloucestershire dairy country in 1796, Snow's investigation in a London parish in 1854, Pasteur's work in a Paris laboratory in 1885. Marks's lens therefore describes the discipline these acts are sometimes taken to anticipate; it does not describe what they were. Where a lens fits a case badly, this chapter says so and goes on with the case's own record.

Each act is read here on four points: who arranged the comparison or the observation, and on what authority; what was held constant so that a result could be attributed to the act; what the result was taken to show at the time; and what the record holds about the hour when the result existed but the practice had not changed. The fourth point is the one this study is about. The first three are what Marks's discipline was built to secure, and what each of these four records shows in a different degree. In the ledger's cell for the kind of act, three are trials or treatments and one is a restriction of movement. Each record names a gap in time between the act and something else — a report, a publication, an institute's newsletter — but the counts are not read alike across the four: two are read as qualified by definition, scope or denominator, one as settled and one as unsettled.

These are the coders' readings of the records' own words, for the four cases read closely in this chapter. The ledger counts no condition for revisiting in any of its eighty-nine records, and the verdicts of the hypotheses across the ledger are read in Chapter 12. A single case can carry a sequence — what was done, to whom, on what date, what followed, what the actor thought it showed. It cannot separate the act from everything else happening at the same time. That is what the following four sections test record by record.

4.2Lind's twelve sailors

On 20 May 1747, in the Old Style, the calendar then in use in England, Lind took twelve patients in the scurvy on board the Salisbury at sea. The treatise of 1753 says their cases were as similar as he could have them: they had all, in general, putrid gums, the spots and lassitude, with weakness of the knees. They lay together in one place, a proper apartment for the sick in the fore-hold, and had one diet in common — water-gruel sweetened with sugar in the morning, fresh mutton-broth often for dinner, at other times puddings, boiled biscuit with sugar and the like, and for supper barley and raisins, rice and currants, sago and wine, or the like. Twelve men, one hold, one diet. What was varied was the medicine.

Two were ordered a quart of cider a day. Two others took twenty-five drops of elixir of vitriol three times a day on an empty stomach, with a strongly acidulated gargle for the mouth. Two others took two spoonfuls of vinegar three times a day on an empty stomach, their gruels and other food well acidulated, and the same gargle. Two of the patients, those with the tendons in the ham rigid, a symptom none of the rest had, were put under a course of sea water, half a pint a day, sometimes more or less as it operated by way of gentle physic. Two others had each two oranges and one lemon given them every day, and ate them at different times on an empty stomach. The two remaining patients took, three times a day, an electuary of the bigness of a nutmeg recommended by a hospital surgeon, made of garlic, mustard-seed, radish, balsam of Peru and gum myrrh, with barley-water well acidulated with tamarinds as their common drink, and were gently purged three or four times during the course by a decoction of tamarinds with the addition of cream of tartar.

The fruit ran out after six days, because that was the quantity that could be spared. Lind's account of what followed is as precise about the limits of the observation as about its result. The most hidden and visible good effects were perceived from the use of the oranges and lemons; one of the two who had taken them was, at the end of six days, fit for duty; and although the spots were not by then quite off his body and his gums were not sound, he became quite healthy before the ship came into Plymouth on 16 June, with no other medicine than a gargle of elixir of vitriol. The other of the two was the best recovered of any in his condition and was appointed nurse to the rest of the sick. Next to the oranges Lind thought the cider had the best effects; it was not very sound, being inclined to be sour or prickly; those who had taken it were in a fairer way of recovery than the others at the end of the fortnight, which was the length of time all the different courses were continued, the fruit excepted. Of the elixir of vitriol Lind observed that the mouths of those who used it by way of gargle were in a much cleaner and better condition than many of the rest, especially those who used the vinegar, and perceived no good effects from its internal use on the other symptoms; the patients on the electuary, the sea water and the vinegar showed no remarkable alteration when compared at the end of the fortnight with others who had taken nothing but a little lenitive electuary and cream of tartar. Lind's own summary of the experiments was that oranges and lemons were the most effectual remedies for this distemper at sea, and that oranges were perhaps preferable to lemons, though both given together might be found most serviceable. The same printed pages also carry a letter from a correspondent, Mr Ives, on the use of cider and sea water in another cruise, where the scurvy cases of the ships with the largest complements of men were so many that great numbers were buried on them, while the Yarmouth buried only two or three. That is a different observation, printed in the same place, and it is not the comparison of twelve men.

Skeptron's record keeps the event and the document apart, and its first Answer names the adopted date:

The record's concrete fact is the comparison itself:

The boundary of what can be judged is stated in the same record:

And the sentence this section is here for:

That sentence does the work this chapter is about. On 20 May 1747 Lind could know what twelve men did under six courses. He could not know, that morning, what the 1753 treatise would say, and he could not know what a navy would carry in 1795. A later reader holding all three dates can say when each was established. A reader holding only the first cannot say what the first was for.

Kenneth J. Carpenter's history of scurvy reads the trial and its aftermath together. On that reading the trial of 1747 showed oranges and lemons to be the best of the six treatments, but Lind did not conclude that citrus was a specific remedy; Lind held to a theory of scurvy as a disorder of digestion and perspiration, and recommended a concentrated, boiled juice that lost much of its virtue. The British navy began to issue lemon juice routinely from 1795, decades after the trial; later the substitution of West Indian limes, which were less potent, led to setbacks; and the cause of scurvy was settled only in the twentieth century, with vitamin C. The wait is therefore not one gap but three: the trial of 1747, the report of 1753, and the routine issue of lemon juice in 1795.

The ledger's cells for this record read the decider as unsettled, the bearer as unknown, the horizon as unsettled, and reversal as unsettled; the codings could not fix from the record's own words who decided the supply policy or when that decision was reviewed. This study reads the case as showing that a comparison can be complete at the hour it ends and still not be read as evidence of what a public body should carry, and that the record which later holds all three dates is not the record that existed in 1747.

4.3Jenner's case

Jenner's Inquiry into the Causes and Effects of the Variolae Vaccinae was printed for the author in London in 1798, and its dedication is dated Berkeley, Gloucestershire, 21 June 1798. In it Jenner writes that a disease so peculiar as the cow pox, which had appeared in that and some of the neighbouring counties for such a series of years, should so long have escaped particular attention, that the prevailing notions on the subject among men of Jenner's profession and others were extremely vague and indeterminate, and that Jenner had instituted as strict an inquiry into the causes and effects of this malady as local circumstances would permit. The opening pages trace, as the work states it, a route from a horse's inflamed heels to the nipples of cows to the hands of the milkers, and claim that a person thus affected is for ever after secure from the infection of the small pox; Jenner adds that a great number of instances will be laid before the reader.

Case XVII is the case this record is named for. Jenner selected a healthy boy, about eight years old, in order to observe the progress of the infection; the matter was taken from a sore on the hand of a dairymaid, infected by her master's cows, and inserted on 14 May 1796 into the boy's arm by means of two superficial incisions, barely penetrating the cutis, each about half an inch long. A footnote in the same case names the dairymaid as Sarah Nelmes. On the seventh day the boy complained of uneasiness in the axilla; on the ninth he became a little chilly, lost his appetite and had a slight head-ache; he was perceptibly indisposed during that day, passed the night with some restlessness, and was perfectly well on the day following. The incisions matured much as they do under variolous matter; the differences Jenner perceived were a darker hue in the limpid fluid and a more erysipelatous look in the efflorescence round the incisions, and the whole died away leaving scabs and eschars without trouble.

Then the second step, separated in time from the first by six weeks:

The record names the boy as Phipps. The source text of 1798 does not name the boy at all; both statements are given here, and this study keeps them apart. On 1 July 1796 the boy was inoculated with variolous matter taken immediately from a pustule, several slight punctures and incisions being made on both arms and the matter carefully inserted, but no disease followed; several months afterwards he was again inoculated with variolous matter, and no sensible effect was produced on the constitution.

Jenner's own account of what the work contains beyond Case XVII matters for how far the claim can be read. The researches were interrupted, Jenner writes, until the spring of 1798, when wet weather affected many of the farmers' horses with sore heels and the cow pox broke out among several dairies. Case XVIII records a child of five, John Baker, inoculated on 16 March 1798 with matter taken from a pustule on the hand of one of the servants infected from the mare's heels; he became ill on the sixth day and was free of indisposition on the eighth; and Jenner says that the experiment was made to ascertain the progress and subsequent effects of the disease when thus propagated, that whether its passing from the horse through the human constitution will produce a similar effect remained to be decided, and that the boy was rendered unfit for inoculation after feeling the effects of a contagious fever in a workhouse. In the closing pages Jenner answers that the investigation tends to a beneficial purpose, because inoculation with small pox matter, despite its improvements, not unfrequently produces deformity of the skin and sometimes, under the best management, proves fatal; that Jenner has never known fatal effects from the cow pox even when impressed in the most unfavourable manner; and that a mode of inoculation may be introduced preferable to that at present adopted. Jenner adds observations that the first boy inoculated slept in a bed with two children who had had neither disease without infecting either, and that two other women with extensive cow pox sores, one of them a daily nurse to an infant, did not communicate the complaint. Jenner also records the contrary cases: a dairy maid who had sores on her fingers but felt no general indisposition and later had the small pox in a mild way on casual exposure, and a fellow servant infected with both sores and general indisposition on whom Jenner's attempts to give the small pox by inoculation all proved fruitless. The work closes by saying that the inquiry has been founded on experiment, in which conjecture has been occasionally admitted to present objects for more minute investigation, and that the inquiry will continue, encouraged by the hope of its becoming essentially beneficial to mankind.

The record's boundary answer sets the limit on how far one course may be carried:

Stefan Riedel's reading of the period places the case in its own order of events. Before Jenner, inoculation with smallpox matter, variolation, was practised in Asia and Africa and was brought to England in the 1720s; Jenner's inoculation in May 1796 followed country observations that milkers who had had cowpox escaped smallpox; the Royal Society did not accept Jenner's first report of 1797, and the Inquiry was published privately in 1798; vaccination then spread quickly. So the sequence a later reader holds is: the act of 14 May 1796, the exposure of 1 July 1796, the first report to the Royal Society and its refusal in 1797, the interruption until the spring of 1798, and the private printing of the Inquiry in 1798. In the ledger's cell for the count, the codings read this record as a count qualified by definition, scope or denominator rather than settled: what the record fixes is the number of subjects and the range of results, not a proportion that could carry the claim beyond them. This study reads the case as showing a public claim resting, at the hour it was made, on one boy's course plus a set of observations from a dairy country, with parts of the argument left open in the actor's own words.

4.4The Broad Street pump

John Snow's second edition of On the Mode of Communication of Cholera opens the account of Broad Street, Golden Square, by saying that within two hundred and fifty yards of the spot where Cambridge Street joins Broad Street there were upwards of five hundred fatal attacks of cholera in ten days. Snow describes the flight of the population: persons in furnished lodgings left first, then other lodgers went away, many houses were closed, and in less than six days from the commencement of the outbreak the most afflicted streets were deserted by more than three-quarters of their inhabitants. Snow became acquainted with the situation and extent of the irruption and suspected contamination of the water of the much-frequented street pump in Broad Street; but on examining the water on the evening of 3 September Snow found so little impurity of an organic nature that Snow hesitated to come to a conclusion. Further inquiry showed no other circumstance or agent common to the circumscribed locality and not extending beyond it except the water of that pump.

The count begins in a register office, not in a hospital. Snow obtained from the General Register Office a list of deaths from cholera registered in the week ending 2 September in the sub-districts of Golden Square, Berwick Street and St Ann's, Soho: eighty-nine deaths registered during the week, of which only six occurred in the four first days, four on Thursday 31 August, and the remaining seventy-nine on Friday and Saturday. Snow considered, therefore, that the outbreak commenced on the Thursday, and made inquiry in detail concerning the eighty-three deaths registered as having taken place during the last three days of the week. Nearly all the deaths had taken place within a short distance of the pump. Only ten occurred in houses situated decidedly nearer to another street pump; in five of these the families said they always sent to the Broad Street pump because they preferred the water; three were children who went to school near it. Within the pump's locality, Snow was informed that sixty-one of the deceased used to drink the pump water, constantly or occasionally; in six instances Snow could get no information at all; in six cases Snow was told that the deceased did not drink the pump water before their illness. The result of the inquiry was, in Snow's own summary, that there had been no particular outbreak or increase of cholera in this part of London except among persons in the habit of drinking the water of that pump-well. Snow had an interview with the Board of Guardians of St James's parish on the evening of Thursday 7 September and represented the circumstances to them; Snow writes that, in consequence of what Snow said, the handle of the pump was removed on the following day.

The table that follows is the count itself, and it shows its own construction. Snow gives fatal attacks and deaths by day from 19 August to 30 September, with a row of forty-five fatal attacks of unknown date and a total of 616 fatal attacks and 616 deaths. Snow explains how the dates of attack were obtained: for upwards of eighty deaths in Middlesex Hospital the hour of admission and the previous duration of the illness were entered in the hospital books; in the remainder the date was calculated by subtracting the duration of the illness from the date of death — where the illness did not exceed twelve hours, the attack was considered to have commenced the same day; where it exceeded twelve and did not exceed thirty-six hours, the attack was put down to the previous day, and so on; where it exceeded forty-eight hours, its duration was generally given in days and subtracted from the date of the attack; and where premonitory diarrhoea was stated, its duration was deducted from the date of death, fixing the time of attack at the first commencement of indisposition, except in two or three instances in which the patient was labouring under another disease. Forty-five cases had no certified duration, and the time of attack in those cases is consequently unknown; these persons nearly all died in the first days of September.

The figures are Snow's, from the table of the second edition; the last column is deaths registered on that day, which during the decline exceeded the attacks, because many who had lingered died later.

Snow reads the table in Snow's own words. The greatest number of attacks in any one day occurred on 1 September, immediately after the outbreak commenced; the following day the attacks fell from one hundred and forty-three to one hundred and sixteen, and the day afterwards to fifty-four; the fresh attacks continued to become less numerous every day; on 8 September, the day when the handle of the pump was removed, there were twelve attacks; on the 9th, eleven; on the 10th, five; on the 11th, five; on the 12th, only one; and after this time there were never more than four attacks on one day. The record's twelve cases for 8 September, the day the handle was removed, are these same twelve attacks, from this table and from Snow's words. Snow's remark on what the table cannot decide is the heart of this case: 'but the attacks had so far diminished before the use of the water was stopped, that it is impossible to decide whether the well still contained the cholera poison in an active state, or whether, from some cause, the water had become free from it.'

Snow also records that the well had been opened and that the superintendent of the works informed Snow there was no hole or crevice in the brickwork by which any impurity might enter, so that in that respect the contamination of the water is not made out by the kind of physical evidence detailed in some of the other instances. Snow's exceptions cut both ways and are recorded as such: the Workhouse in Poland Street had a pump-well of its own and a supply from the Grand Junction Water Works, and out of five hundred and thirty-five inmates only five died of cholera; a brewery near the pump had more than seventy workmen, none of whom suffered from cholera in a severe form, a deep well of its own and New River water; at a percussion-cap manufactory at 37 Broad Street, where about two hundred workpeople were employed and two tubs on the premises were supplied from the pump, eighteen workpeople died of cholera at their own homes; and seven workmen at a manufactory of dentists' materials who died had all been in the habit of drinking the pump water, while two persons who resided constantly on the premises and did not drink it had only diarrhoea. Snow's most telling witness from outside the district is a widow at West End in Hampstead, who had not been in the neighbourhood of Broad Street for many months; her son said that a cart went from Broad Street to West End every day and that it was the custom to take out a large bottle of the water because she preferred it; the water was taken on Thursday 31 August and she drank of it that evening and again on Friday; she was seized on the evening of the latter day and died on Saturday. A niece who was on a visit also drank of the water, returned to Islington, was attacked and died; there was no cholera at the time at West End or where the niece died. Snow adds that many persons drank the water around the time of the outbreak without being attacked, which Snow says does not diminish the evidence, for reasons stated in another part of the work.

The source text read here runs from the opening of the Broad Street account to its last paragraph and stops before Snow's stronger evidence, which was the comparison of households supplied by different water companies in south London. Peter Vinten-Johansen and colleagues, in their study of Snow's life and work, show that Snow persuaded the parish Board of Guardians to remove the pump handle, that the outbreak was already declining when it was removed, and that Snow did not claim the removal ended the outbreak; on their account Snow's case rested more on the comparison of households supplied by different water companies, and they trace how slowly the medical establishment accepted Snow's theory.

The record's boundary answer is where this study reads the case:

Chapter 2 sets out this same record's Scenario and codes it field by field. This chapter adds the source text and the later reading. For the count, the codings read Snow's count as settled; the source text shows what that settlement is made of, a mixture of observed hospital entries, calculated dates and forty-five attacks undated. This study reads the case as the one in which the actor's own text does the separating that a later reader would otherwise have to do: the act has a date, the evidence for its cause was argued for in the same book, and the count's construction is printed beside both.

4.5Pasteur and Joseph Meister

Pasteur's report to the Académie des sciences was read on Monday 26 October 1885 and printed in the Comptes rendus hebdomadaires, tome 101, no. 17, pages 765 to 772. The paper is in French; what follows gives its sense in this study's own English.

Pasteur begins by saying that the prophylaxis of rabies exposed in earlier notes was a real progress in the study of the disease, but a progress more scientific than practical: its application exposed to accidents, and of twenty dogs treated Pasteur could not have answered for rendering more than fifteen or sixteen refractory to rabies. It was useful, the paper goes on, to end the treatment with a final very virulent inoculation, a control virus, to confirm and reinforce the refractory state; prudence required that the dogs be kept under observation longer than the incubation period of the disease produced by direct inoculation of that last virus, so that sometimes no less than three to four months was needed to be assured of the refractory state. Those requirements would have limited the application greatly, and the method would have lent itself only with difficulty to a treatment begun at once, which is what the accidental and unforeseen character of rabid bites demanded. It was therefore necessary, if possible, to arrive at a more rapid method capable of giving a perfect security in dogs; and how, before that progress was attained, would one dare permit any trial on man? After experiments, Pasteur writes, so to speak without number, a prophylactic method, practical and prompt, had been reached, whose successes on the dog were already numerous and sure enough for Pasteur to have confidence in the generality of its application to all animals and to man himself.

The method's facts are animal facts, established over years. Inoculation into a rabbit by trepanation, under the dura mater, of the spinal cord of a dog rabid with street rabies, always gives rabies to those animals after an average incubation of about fifteen days; passing from one rabbit to the next, a more and more marked tendency to a shorter incubation appears; after twenty to twenty-five passages the incubation falls to eight days and stays there for a further period; then seven days, found with striking regularity through a series of passages going as far as the ninetieth. The series, begun in November 1882, had run three years without interruption and without recourse to any virus other than that of the successive rabbits. Pasteur calls this the practical knot of the method: a rabies virus of perfect purity, always identical to itself, always at hand. The cords of these rabbits are rabid throughout their length; cut into lengths of a few centimetres and suspended in dry air, their virulence disappears slowly until it is extinguished, the duration depending on the thickness of the cord and above all on the exterior temperature.

The procedure in a dog is to begin with a cord quite distant from the day of operation, so that it is certainly not virulent, and to proceed with more recent cords two days apart, until one arrives at a last cord very virulent, placed in its flask only a day or two. When this has been done, the dog is refractory: rabid virus may be inoculated under the skin or even on the surface of the brain by trepanation without rabies declaring itself. Pasteur had reached fifty dogs of every age and race refractory, without a single failure, when three persons arrived from Alsace in Pasteur's laboratory on Monday 6 July. Théodore Vone, a grocer, had been bitten on the arm on 4 July by his own dog that had become rabid. Joseph Meister, aged nine, had been bitten also on 4 July at eight in the morning by the same dog; the child, brought down by the dog, bore numerous bites on the hand, the legs and the thighs, some deep enough to make even his walking difficult, and the principal of these had been cauterised twelve hours after the accident with phenic acid by a physician of Villé. At the autopsy of the dog killed by its master the stomach was found full of hay, straw and fragments of wood; the dog was well rabid; Joseph Meister had been raised from under it covered with saliva and blood. The third person, who had not been bitten, was the child's mother.

The weekly meeting of the Académie happened to be on 6 July. Pasteur told Vulpian what had happened, and Vulpian, with Grancher, professor at the Faculty of Medicine, came immediately to see the child and ascertain the state and number of his wounds; there were no fewer than fourteen. Their advice was that by the intensity and the number of the bites the child was exposed almost fatally to take rabies. The child's death appearing inevitable, Pasteur decided, not without lively and cruel anxieties, to attempt on Joseph Meister the method that had constantly succeeded on dogs. Pasteur writes that the fifty dogs had not been bitten before Pasteur determined their refractory state, but that this circumstance could be set aside, because Pasteur had already obtained the refractory state in a large number of dogs after a bite, and had made the members of the Commission on Rabies witnesses that year of this progress. On 6 July at eight in the evening, sixty hours after the bites, and in the presence of Vulpian and Grancher, half a syringe of the cord of a rabbit that died rabid on 21 June, kept since in dry air, was inoculated under a fold of the skin of the child's right hypochondrium. Inoculations followed on each of the following days, recorded in a printed table that runs from 7 to 16 July and gives, for each day, the date of the cord used and the number of days it had been dried, in the order from the most distant cord to the most recent. Pasteur states that this brought the number of inoculations to thirteen and the number of days of treatment to ten, and that a smaller number would have sufficed, but that in a first attempt Pasteur had to act with particular circumspection. Two new rabbits were inoculated by trepanation with each cord used, in order to follow its state of virulence: the cords of 6, 7, 8, 9 and 10 July were not virulent, since they did not render their rabbits rabid; the cords of 11, 12, 14, 15 and 16 July were all virulent, and the virulent matter was found in them in a proportion more and more strong. In the last days, Pasteur writes, the most virulent rabies virus was therefore inoculated into the child — the virus of the dog reinforced by many passages from rabbit to rabbit, which gives rabies to animals after seven days' incubation. When the state of immunity is reached, Pasteur adds, one may inoculate the most virulent virus in any quantity without inconvenience; it had always seemed to Pasteur that this had no other effect than to consolidate the refractory state. The child had therefore escaped not only the rabies the bites might have developed but that inoculated as a control, and, from the middle of August, Pasteur writes, considered the future of the child's health with confidence; at the date of the paper, after three months and three weeks, that health left nothing to be desired.

The paper is not a closed report. Pasteur said he did not intend to treat the question completely that day, that he would later show the first interpretation to disagree with the facts, and that the third would be examined experimentally in a future communication. Pasteur closes by naming what Pasteur calls the most serious of the questions to resolve: the interval to observe between the instant of the bites and the beginning of the treatment, which for Joseph Meister was two and a half days but must be expected to be often much longer. The paper then records that on 20 October a young man of fifteen, Jean-Baptiste Jupille, a shepherd, had been bitten in both hands for six full days and that treatment had begun on that Tuesday with Vulpian and Grancher assisting, and that the Academy would be told what happened of this new attempt.

The record keeps its own distance:

Two documents stand behind these two statements, and they differ. The record's adopted source is an institute newsletter of 2015, which it names as the notation of the document; the source text read here is Pasteur's own report of 26 October 1885. The record says that Grancher began the vaccination; the report of 1885 describes the inoculations as done on 6 July in the presence of Vulpian and Grancher and does not name who gave them, and the reading notes for this study record that Pasteur was not a physician and that the inoculations were given by a physician. Both accounts are given here; this study does not choose between them.

Two later works read what the paper leaves behind. Gerald L. Geison's study of Pasteur, drawn from laboratory notebooks opened to scholars in the 1970s, argues that Pasteur's public accounts differed from Pasteur's private research: before treating Joseph Meister in July 1885, Pasteur had not tested on dogs bitten by rabid animals the particular series of inoculations used on the child, contrary to what the published report led readers to believe. Geison raises the ethical questions this poses while granting the treatment's outcome. The paper itself acknowledges part of the gap, in Pasteur's own words that the fifty dogs had not been bitten, and claims a separate body of evidence for bitten dogs; what the notebooks add is a dispute about which evidence stood behind the act of 6 July. Bruno Latour's reading argues that Pasteur's successes depended on alliances: hygienists, physicians, veterinarians, the army and others took up the microbe because it served their own projects, and laboratory results were made to hold outside the laboratory by extending the laboratory's conditions into farms and hospitals, so that scientific facts spread through such networks rather than by their own force. Latour's lens fits the scene at the Academy — Vulpian, Grancher, the Faculty of Medicine, the Commission on Rabies, the Académie — and fits less well the ledger's question, which is not how a fact spread but what the record of one act holds about its own time.

The ledger's cells read the decider as a team or organisation, the bearer as unsettled, the horizon as unsettled, and reversal as easy to infer; in the count cell the codings read the record as qualified by definition, scope or denominator; the bearer is read as not named, and the revision as named in the case. This study reads the case as showing an act of 6 July 1885 whose evidence — the animal series begun in November 1882 — predates it by years, whose report appeared sixteen weeks later, and whose own author named, in that report, the question the report could not answer.

4.6Naming a past disease

The four acts above have actors who wrote down what they did. A fifth record in this chapter's group has no settled actor and no settled act: The Plague at Athens. There a disease described in 430 BC is named, argued over and partly contested by laboratory evidence gathered more than two millennia later, and the record for this study is a case in which the codings could not fix an act at all.

Thucydides set the plague of Athens down in Book II of the History of the Peloponnesian War, after having had the disease himself and watched others suffer it, so that a reader might recognise it should it break out again; Thucydides leaves causes to others, physician or layman, and reports that physicians were of no service, being the first to visit the sick and the first to die, and that no remedy or human art availed. Chapter 3 reads that purpose in full. For this chapter the point is narrower: the passage is a witness's account of symptoms, not a laboratory result. Thucydides's own stated limit is the same: the account sets down the nature of the disease and explains its symptoms so that a reader may recognise it if it ever breaks out again. (Thucydides, History of the Peloponnesian War, Book 2, chapter 48, translated by Richard Crawley.)

This study reads Skeptron's record for The Plague at Athens as holding that this testimony is not a genetic identification: a written account of what a disease looked like and whom it killed is not a statement of what organism caused it. The record's own Answer is not quoted here, because this chapter has no text of the record to quote; the Introduction sets out the limits of the evidence. The coders could not settle whether the case shows a public act: for this record the cells read act unsettled, kind of act unsettled, act before proof unsettled, gap in time not named, count no count, bearer unsettled, revision named in case unsettled. This study reads the record as the boundary case of the chapter: here there is no hour of an act to keep apart from an hour of evidence, because the act is not fixed.

The naming dispute itself runs in two directions at once. David M. Morens and Robert J. Littman read Thucydides's account as epidemiological evidence — onset, spread, attack among caregivers, symptoms, immunity after recovery — weigh the many diseases proposed for it, and conclude that the account does not allow a firm identification. Manolis J. Papagrigorakis and colleagues reported DNA sequences resembling Salmonella enterica serovar Typhi from the dental pulp of people buried in a mass grave in the Kerameikos in Athens, dated to the time of the plague, and proposed typhoid fever as a probable cause. Beth Shapiro, Andrew Rambaut and M. Thomas P. Gilbert replied in the same year that the reported sequences do not place the organism as Salmonella Typhi with confidence, that the methods do not exclude contamination, and that the identification of the plague's cause remains unproven. What an identification made in 2006 can tell is what the description in 430 BC is a description of; what it cannot tell is what the Athenians knew or did at any hour, because the identification rests on bones and sequences rather than on records of decisions, and because the dispute about it is itself unresolved. A later reader holds the claim and the reply together and can date both.

The Black Death shows the other outcome. Samuel K. Cohn Jr. argued that the medieval Black Death's patterns of spread, seasonality and symptoms, as contemporaries recorded them, differ from those of the bubonic plague observed from the 1890s, and questioned its identification with that disease; and that culture changed as later waves struck, with physicians and communities growing more confident. Kirsten I. Bos and colleagues reconstructed a draft genome of Yersinia pestis from teeth of people buried in the East Smithfield cemetery in London, used for the dead of 1348–50, establishing that the Black Death was caused by Y. pestis. Here the two sides are not arguments about the same evidence: Cohn's claim is made from what the records say, Bos's from what the remains hold. The identification succeeded, and the reading notes name what it established. What it did not establish, and could not, is what a council in Ragusa knew in 1377, what an ordinance compelled in 1349, or what any of the acts read in this study was founded on at the hour it was taken. Those are record questions, and Chapters 3 and 6 read their records; a genome is not a record of a decision.

Of these five records, four name a gap in time between the act and something else, and the fifth does not. Two of the four name the bearer of the cost, and two do not. In each, the evidence for the act's own effect came after the act — in 1753 after 1747, in 1798 after 1796, in the same book as the removal of the handle but argued for by a comparison Snow's own table could not settle, in October 1885 after July 1885, and in 2006 after 430 BC, where it remains contested. The third separation this study keeps apart, the time of the count, is taken up in Chapter 5.