Chapter 5. Paris, July 1885: one boy, two physicians and an academy

This section argues, as a historical claim, that Pasteur's report to the Académie des sciences presented the treatment of Joseph Meister as the application of a method proved on dogs, while conceding, a few paragraphs on, that the fifty dogs on which it had been proved had not been bitten; that the report placed two physicians, a commission and the laboratory's own controls at the points where the argument needed support; and that it left the method's effect on the boy, as distinct from his survival, unproved in its own terms. For the debate, the section asks what the report allows each of Geison and Latour to claim. Since Pasteur's text cannot be quoted here, the reading follows his paper in close paraphrase, in his order, and this article's English is its own.

5.1The problem the method had to solve

Pasteur opens with the shortcomings of the prophylaxis he and his collaborators had already described in earlier notes. It was, he says, a real advance in the study of the disease, but more scientific than practical. Its use exposed the animals to accidents: of twenty dogs treated, he could not have answered for making more than fifteen or sixteen refractory to rabies. The earlier method also called for a final, very virulent control inoculation, and prudence required keeping the dogs under watch for longer than the incubation of the disease that control produced, so that sometimes three or four months passed before one could be sure of the result. Such demands, he writes, would have greatly limited the method's use, and it could not easily be started at once, which the accidental and unforeseen nature of rabid bites required.

He then states the standard he set himself. A faster method was needed, one able to give a security on dogs he ventures to call perfect. And he asks, as a rhetorical question, how anyone could permit himself any trial on a human being before that progress had been reached.

That question is the report's own measure of what evidence would justify an act on a person. It is worth holding on to, because it is the measure against which Geison's argument, as summarised in its note, is cast: that what the report presented as having been reached, on dogs bitten by rabid animals, had not been reached with the particular series used on the boy.

5.2The method and the fifty dogs

After experiments he calls practically countless, Pasteur says, he arrived at a method, practical and prompt, whose successes on the dog were numerous and sure enough for him to trust its general application to all animals and to man himself. Its basis is a rabies virus passed from rabbit to rabbit, by inoculation under the dura mater, until its incubation settles at seven days; the series, he says, had run without a break since November 1882, giving a virus always identical to itself or nearly so. Pieces of the spinal cords of these rabbits, hung in dry air, slowly lose their virulence until it is extinguished.

On a dog, the procedure is a daily injection under the skin of broth containing a fragment of one of the drying cords, beginning with a cord old enough to be surely without virulence and moving day by day to fresher ones, until the last is very virulent. The dog is then refractory, even to rabies inoculated on the surface of the brain.

By this method, Pasteur reports, he had made fifty dogs of every age and breed refractory to rabies without meeting a single failure, when, unexpectedly, three people from Alsace presented themselves at his laboratory on Monday, 6 July.

5.3The arrival

The three were a grocer, Théodore Vone, bitten on the arm on 4 July by his own dog, which had gone mad; Joseph Meister, aged nine, bitten the same morning by the same dog, thrown down by it and bitten many times, on the hand, the legs and the thighs, some of the bites deep enough to make walking difficult; and Meister's mother, who had not been bitten. The principal wounds had been cauterised with phenic acid by a Dr Weber of Villé, twelve hours after the attack. At the autopsy of the dog, killed by its owner, the stomach had been found full of hay, straw and pieces of wood; the dog, Pasteur writes, was certainly rabid, and the boy had been pulled from beneath it covered in slaver and blood.

Vone had heavy bruises but assured Pasteur that the dog's teeth had not gone through his shirt. Pasteur told him there was nothing to fear and that he could go home that day, which he did. He kept the boy and his mother with him.

Two features of this paragraph matter for the debate. Pasteur made a triage on the evidence before him, sending one bitten person home on the strength of a shirt. And much of what he reports about the dog and the wounds is not his own observation: the autopsy of a dog its owner had killed, a cauterisation done by Dr Weber, and the state in which the boy had been pulled from beneath the dog. The report does not say how these facts reached Pasteur. As printed, the decision rests in part on events he did not see, as Snow's did.

5.4The consultation

The Académie met that same Monday. Pasteur went, saw his colleague Dr Vulpian there, and told him what had happened. Vulpian, and Dr Grancher, a professor in the Faculty of Medicine, came at once to see the boy and to establish the state and number of his wounds, of which there were no fewer than fourteen. Their opinion, as Pasteur reports it, was that by the intensity and number of his bites Meister was exposed almost fatally to rabies. Pasteur then told the two physicians of the new results he had obtained in the study of rabies since a lecture he had given at Copenhagen a year before.

The order is again worth keeping. The two physicians gave an opinion on the boy's prognosis. Pasteur then told them of his results. The report does not say what the two physicians thought of those results, or that they approved the treatment. It records that they examined the boy, that they judged his danger grave, and that they heard from Pasteur about his work.

5.5The decision and the paragraph about the dogs

The next paragraph is the decision. The boy's death appearing inevitable, Pasteur writes, he resolved, not without keen and cruel anxiety, as one may well imagine, to try on Joseph Meister the method that had constantly succeeded with him on dogs.

The paragraph that follows is the one on which the evidence side of the debate turns. Pasteur concedes that his fifty dogs had not been bitten before he made them refractory. But he knew, he says, that this circumstance could be set aside, because he had already obtained the refractory state against rabies in a great number of dogs after a bite; and he had, that year, made the members of the Rabies Commission witnesses of this new and important progress.

Read on the page, the paragraph does three things. It states openly that the fifty dogs of the earlier paragraph, the ones made refractory without a single failure by the report's own account, were not bitten dogs. It then asserts a different body of evidence: refractory states obtained in many dogs after a bite. And it names a body that had witnessed that evidence. What it does not say is which method, or which series of inoculations, had been used on the dogs treated after a bite. The two bodies of evidence are placed side by side, the first described in detail and the second only asserted, and the reader is left to join them.

Geison's argument, as its note gives it, is that Pasteur had not tested on dogs bitten by rabid animals the particular series he used on the boy, contrary to what his published report led readers to believe. The report's own wording is consistent with that reading in a precise way: it does not state that the series given to Meister had been tested on bitten dogs, and its structure invites a reader to assume it. This article cannot say from the report what the notebooks show; that is what Geison's book adds and what the published page cannot supply. It can say that the evidence on which the act was presented as resting is, in the report's own text, two different things, only one of which is described.

The paragraph also shows where the alliance enters. At the point where the evidence is thinnest, the report calls a witness: the Rabies Commission, whose members had seen the progress on dogs after a bite. The guarantee for the unstated part of the argument is a body of named observers, not a described experiment.

5.6The inoculations

On 6 July, at eight in the evening, sixty hours after the bites, and in the presence of Drs Vulpian and Grancher, the boy was inoculated under a fold of skin on the right side of the abdomen with half a syringe of cord from a rabbit dead of rabies on 21 June, kept since then in dry air, that is, for fifteen days. In the following days new inoculations were made, always in the same region, according to a table the paper prints and the machine reading has garbled. Pasteur's prose states the totals: thirteen inoculations over ten days of treatment. He adds that he will say later that fewer inoculations would have sufficed, but that in this first trial he had to act with particular caution.

The report's grammar at the moment of the first inoculation is impersonal. It says that the boy was inoculated, in the presence of the two physicians, and does not say by whose hand. The reading note on the paper adds that Pasteur was not a physician and that the inoculations were given by a physician. The report itself is silent on who held the syringe.

5.7The laboratory follows the boy

The next paragraphs describe a control that ran alongside the treatment. Each of the cords used on the boy was also inoculated by trepanation into two fresh rabbits, to follow its virulence. The rabbits showed, Pasteur reports, that the cords used in the first days were not virulent, since they did not give their rabbits rabies, while the cords used from 11 July onwards were all virulent, increasingly so; rabies appeared after seven days' incubation in the rabbits given the cords of 15 and 16 July, after eight days in those of the 12th and 14th, and after fifteen days in those of the 11th.

In the last days, therefore, Pasteur writes, he had inoculated Joseph Meister with the most virulent rabies virus, that of the dog reinforced by many passages through rabbits, a virus that gives rabies to rabbits after seven days and to dogs after eight or ten. He states that he was authorised in this undertaking by what had happened with the fifty dogs. Once immunity is reached, he says, the most virulent virus can be inoculated in any quantity without harm; it has always seemed to him to have no other effect than to consolidate the refractory state.

The conclusion he draws is that Meister escaped not only the rabies his bites might have developed, but also the rabies Pasteur himself had inoculated into him as a control of the immunity due to the treatment, a rabies more virulent than that of the street dog. The final very virulent inoculation, he adds, also limits the time of anxiety: if rabies were going to break out, it would do so sooner from the stronger virus than from the bites. From the middle of August he looked on the boy's future health with confidence; at the time of writing, three months and three weeks after the accident, that health left nothing to be desired.

Two things in this passage bear on the debate. The first is that the authorisation Pasteur claims for giving the boy the most virulent virus is, explicitly, the fifty dogs: the dogs an earlier paragraph had said were not bitten. The authorisation for the boldest step in the treatment rests on the body of evidence the report describes, not on the one it only asserts.

The second is Latour's. Each day's cord given to the boy was also given to rabbits, and the measure of what he had received was taken in those animals; the report presents the treatment and the animal controls as one series. Latour's note speaks of laboratory results made to hold outside the laboratory by extending the laboratory's conditions into farms and hospitals. Here the extension runs to a person: the laboratory's procedure, materials and controls accompany the boy's treatment. The note itself does not mention rabies or Meister.

5.8What the method means, left for later

Pasteur then asks how the new method should be interpreted, and says he does not intend to treat the question completely that day. Of three interpretations, he says he will later show the first, that drying simply lowers the virulence of the cords, to disagree with the facts; the second, that a virus of constant virulence given in small but daily increasing quantities could produce the refractory state, he is studying experimentally; the third, which he calls strange at first sight, that the virus might consist of a living substance and a non-living one able to stop it, he will examine experimentally in a future communication.

The report, in other words, describes a treatment whose mechanism its author had not settled at the time he gave it, and says so. He promises to show one interpretation wrong, is testing another, and reserves a third. For the evidence question, this is a second disclosure of the same kind as the one about the dogs: the method was applied before its working was understood.

5.9The second patient

The paper closes on what Pasteur calls perhaps the most serious question to be resolved: the interval to be observed between the bites and the start of treatment. For Meister it had been two and a half days; it will often, he says, be much longer. On Tuesday, 20 October, with the help of Vulpian and Grancher, he had begun to treat a youth of fifteen, Jean-Baptiste Jupille, a shepherd bitten six full days before while defending younger children from a dog, and he promises to report the result.

The ending looks forward. The report of 26 October is, in its own last paragraphs, the report of an experiment in progress: one patient past the period of anxiety, another under treatment, a key variable named and untested. Vulpian and Grancher appear again, at the second treatment as at the first.

5.10The record beside the text

Skeptron's record of the case rests not on Pasteur's paper but on an account published by the Institut Pasteur in 2015. Its scenario reads:

The record differs from the report in one particular. It names Grancher as the person who began the vaccination. The report says only that the boy was inoculated in the presence of Vulpian and Grancher, and does not name the hand. The reading note on the report agrees with the record that a physician gave the inoculations, but does not name which. The record's attribution comes from the institute's later account; the report of 1885 is silent. This article gives both and does not decide between them.

The record dates its own source plainly:

The modern summary thus rests on an account of 2015, not on the report read to the Académie in October 1885, and the record states both dates.

On the boundary of judgment, the record asks whether the case can be turned into something larger:

Its answer, after a refusal that belongs to Skeptron's recurring frame, gives a sentence of its own:

The four things this sentence separates are the terms of the debate. The doctor's actions are what was done to the boy, which the report describes and the record attributes to Grancher. The historical case is the outcome, the survival Geison grants. Later recollection is the layer of accounts, from the report of October 1885 to the newsletter of 2015; Geison's argument concerns the distance between one such account and the notebooks. Verification of group effects is what a single case cannot provide; Marks traces how the controlled clinical trial grew from reformers' distrust of the individual physician's judgment (Marks 1997). The record takes no side. It names the terms.

What the record omits is, again, the alliance side. It says Pasteur "consulted with doctors" and names one. It says nothing of the Académie, the Rabies Commission, the rabbits, the fifty dogs, or the paragraph about the bitten dogs. A reader of the record alone would learn that the case is one person's record and must not be generalised, which the report's own interpretive caution supports, and would not learn what evidence the report put forward or whom it called as witness.

5.11What the report gives each side

For Geison's side, the report supplies the paragraph about the dogs. It concedes that the fifty dogs were not bitten, asserts a separate success on bitten dogs without stating its method, and then rests its authorisation for the most virulent inoculation on the fifty. A reader looking for the gap between the public account and the private research will find, in the public account itself, the place where the gap would have to be.

For Latour's side, the report supplies named allies at each step: a country physician who cauterised the wounds; two physicians, one a professor of the Faculty of Medicine, who examined the boy, judged his danger and were present at the first inoculation and at the second patient's treatment; a commission that had witnessed the work on dogs; the academy to which the report was read; and the laboratory's animals, run in parallel with the boy. The report takes care to show that the act was not one man's alone.

What the report gives neither side is a demonstration that the treatment caused the boy's survival. Pasteur's argument for that rests on the prior dogs, on the control inoculation, and on the boy's health three months and three weeks later. The record's warning that one survival case does not convert into a group effect is the modern statement of a limit that the report, by deferring its interpretation and by turning at its close to the untested question of delay, already concedes.

The section began from the claim that the report presented the treatment as resting on proof in dogs while stating that the dogs were unbitten, placed allies at the points of weakness, and left the method's effect unproved in its own terms. The report bears out each part. What it adds to the debate is that Geison's reading and Latour's are anchored in adjacent sentences of one paragraph: the concession about the fifty dogs and the call on the Rabies Commission. Neither reading can be had without the other in view.